Escolar Documentos
Profissional Documentos
Cultura Documentos
lished in 1960 by the Ford and Rockefeller Foundations with Toriyama K, Heong KL, Hardy B, editors. 2005. Rice is
the help and approval of the Government of the Philippines. life: scientific perspectives for the 21st century. Proceed-
Today IRRI is one of the 15 nonprofit international research ings of the World Rice Research Conference held in Tokyo
centers supported by the Consultative Group on International and Tsukuba, Japan, 4-7 November 2004. Los Baños (Phil-
Agricultural Research (CGIAR – www.cgiar.org). ippines): International Rice Research Institute, and Tsukuba
IRRI receives support from several CGIAR members, (Japan): Japan International Research Center for Agricul-
including the World Bank, European Union, Asian Devel- tural Sciences. CD-ROM. 590 p.
opment Bank, International Fund for Agricultural Develop-
ment, International Development Research Centre,
Rockefeller Foundation, Food and Agriculture Organization
of the United Nations, and agencies of the following coun-
tries: Australia, Austria, Belgium, Brazil, Canada, Denmark,
France, Germany, India, Iran, Japan, Malaysia, Netherlands,
Norway, People’s Republic of China, Republic of Korea,
Republic of the Philippines, Spain, Sweden, Switzerland,
Thailand, United Kingdom, United States, and Vietnam.
The responsibility for this publication rests with the
International Rice Research Institute.
ISBN 971-22-0204-6
Contents
FOREWORD xiv
ACKNOWLEDGMENTS xv
Keynotes
The power of settled life: rice farming as a lifestyle 2
Ken-ichi Matsumoto
The role of rice in the Japanese diet 3
Yoshiko Kagawa
A greeting 4
Yoshinobu Shimamura
Statement by Jacques Diouf 4
Address by Imperial Highness Crown Prince Naruhito 5
New technologies for rice production 6
Gurdev S. Khush
The changing economics and politics of rice: implications for food security, 7
globalization, and environmental sustainability
Joachim von Braun and María Soledad Bos
Feeding the world: How much more rice do we need? 21
Vaclav Smil
Development of sustainable agriculture from rice, water, 23
and the living environment
Riota Nakamura
Research strategy for rice in the 21st century 26
Ronald P. Cantrell and Gene P. Hettel
SESSION 1
The genus Oryza, its diversity, and its evolution
Molecular phylogeny and divergence of the rice tribe Oryzeae, with special 40
reference to the origin of the genus Oryza
Song Ge, Ya-long Guo, and Qi-hui Zhu
Eco-genetic diversification in the genus Oryza: implications for sustainable 44
rice production
Duncan Vaughan, Koh-ichi Kadowaki, Akito Kaga, and Norihiko Tomooka
Toward a global strategy for the conservation of rice genetic resources 47
Ruaraidh Sackville Hamilton and Ruth Raymond
Genetic architecture and complexity in wild and cultivated rice 49
Y. Sano
Who was the mother of cultivated rice? Differentiation of chloroplast genome 52
structure among cultivated rice and ancestral wild species
Koh-ichi Kadowaki
Diverse mechanisms of low-temperature stress response in rice 54
Ryozo Imai, Jiangqi Wen, Kentaro Sasaki, and Kiyoharu Oono
Genetic diversity of Myanmar rice landraces 56
Ye Tint Tun, K. Irie, T. Nagamine, Jhon Ba Maw, F. Kikuchi, and H. Fujimaki
Contents iii
QTL analysis for eating quality in japonica rice 60
H.G. Hwang, J.P. Suh, Y.C. Cho, S.J. Kwon, I.S. Choi, H.C. Hong, Y.G. Kim,
M.K. Kim, H.C. Choi, and Y.T. Lee
SESSION 2
Structure and function of the rice genome
The complete rice genome sequence and its application to breeding 66
and genetics
Takuji Sasaki
Exploitation and use of naturally occurring allelic variations in rice 69
Masahiro Yano, Yasunori Nonoue, Tsuyu Ando, Ayahiko Shomura,
Takehiko Shimizu, Izumi Kono, Saeko Konishi, Utako Yamanouchi,
Tadamasa Ueda, Shin-ichi Yamamoto, and Takeshi Izawa
Allelic and functional diversity of stress-tolerance genes in rice 73
Hei Leung, Hehe Wang, Jianli Wu, Ma. Elizabeth Naredo,
Marietta Baraoidan, Alicia Bordeos, Suzette Madamba, Gay Carrillo,
Jatinder Sangha, Zenna Negussie, Jill Cairns, Bin Liu, Yolanda Chen,
Darshan Brar, Il Ryong Choi, Cassiana Vera Cruz, Renee Lafitte,
Luca Comai, and Kenneth L. McNally
Functional genomics by reverse genetics 76
Gynheung An
Tissue culture-induced mutations and a new type of activation tagging 78
as tools for functional analysis of rice genes
Hirohiko Hirochika
Toward genome-wide transcriptional analysis in rice using MAS 80
oligonucleotide tiling-path microarrays
Lei Li, Xiangfeng Wang, Xueyong Li, Ning Su, Viktor Stolc, Bin Han,
Jiayang Li, Yongbiao Xue, Jun Wang, and Xing Wang Deng
Candidate gene characterization at the Pup1 locus: a major QTL increasing 83
tolerance of phosphorus deficiency
Matthias Wissuwa, Kristy Gatdula, and Abdelbagi Ismail
Wrap-up of Session 2 85
SESSION 3
Opportunities and challenges of transgenic rice
Overproduction of C4 enzymes in transgenic rice: an approach for improved 88
photosynthesis and crop yield
Mitsue Miyao-Tokutomi and Hiroshi Fukayama
The uptake and translocation of minerals in rice plants 90
Naoko K. Nishizawa
Improving drought and cold-stress tolerance in transgenic rice 94
Kazuko Yamaguchi-Shinozaki and Kazuo Shinozaki
Broad-spectrum disease resistance in transgenic rice 97
Motoshige Kawata
Golden Rice and improvement of human nutrition 99
Swapan Datta, Vilas Parkhi, Mayank Rai, Jing Tan, Niranjan Baisakh,
Lina Torrizo, Editha Abrigo, Norman Oliva, Md. Alamgir Hossain,
Russel Julian, Anindya Bandyopadhyay, and Karabi Datta
Health-promoting transgenic rice suppressing life-related disease 102
and type-I allergy
Fumio Takaiwa
Producing rice plants with a site-specific base change in the acetolactate 105
synthase gene by chimeraplast-directed gene targeting
A. Okuzaki and K. Toriyama
Transgenic rice plants expressing wheat catalase show improved 108
tolerance for chilling-induced damage in membranes
Haruo Saruyama, Hidenori Onodera, and Matsuo Uemura
Wrap-up of Session 3 110
SESSION 5
Broadening the gene pool and exploiting heterosis in cultivated rice
Developing aerobic rice in Brazil 154
B. da S. Pinheiro, E. da M. de Castro, O.P. de Moraes, and F. Breseghello
Broadening the gene pool of rice through introgression from wild species 157
D.S. Brar
Mutation in seed reserves and its use for improving grain quality in rice 160
Hikaru Satoh, Ken-ichi Ohtsubo, and Yasunori Nakamura
Heterosis in rice for increasing yield, production efficiency, 162
and rural employment opportunities
Sant S. Virmani
Contents v
Harnessing molecular markers in hybrid rice commercialization in the Philippines 166
E.D. Redoña, L.M. Perez, L.R. Hipolito, V.E. Elec, I.A. Pacada, L.M. Borines,
R.O. Solis, S.A. Ordoñez, and J. Agarcio
Genetic evolution of Rf1 locus for the fertility restorer gene of BT-type CMS rice 170
Tomohiko Kazama and Kinya Toriyama
A polygenic balance model in yield components as revealed by QTL analysis 172
in rice
Wilhelm E. Hagiwara, Kazumitsu Onishi, Itsuro Takamure, and Yoshio Sano
Wrap-up of Session 5 174
SESSION 6
Trends in crop establishment and management in Asia
Trends in crop establishment methods in Asia and research issues 178
Sushil Pandey and Lourdes Velasco
Direct seeding and weed management in Korea 181
Soon-Chul Kim and Woon-Goo Ha
Direct-seeding cultivation of rice in Japan: stabilization of seedling establishment 184
and improvement of lodging resistance
Satoshi Yoshinaga
Direct seeding of aerobic rice in China 186
Guang Hui Xie, Jun Yu, Jing Yan, Huaqi Wang, and Xiurong Zhu
An overview on direct seeding for rice crop establishment in the Philippines 189
Jovino L. de Dios, Evelyn F. Javier, Myrna D. Malabayabas,
Madonna C. Casimero, Alex J. Espiritu
Rice establishment in drought-prone areas of Bangladesh 193
M.A. Mazid, M.A. Mortimer, C.R. Riches, A. Orr, B. Karmaker,
A. Ali, M.A. Jabbar, and L.J. Wade
Emerging issues in weed management of direct-seeded rice in Malaysia, 196
Vietnam, and Thailand
M. Azmi, D. V. Chin, P. Vongsaroj, and D.E. Johnson
Changing from transplanted rice to direct seeding in the rice-wheat 198
cropping system in India
Y. Singh, Govindra Singh, David Johnson, and Martin Mortimer
Seedling recruitment in direct-seeded rice: weed biology and water management 202
A.M. Mortimer, O. Namuco, and D.E. Johnson
The crop protection industry’s view on trends in rice crop establishment 205
in Asia and their impact on weed management techniques
Jean-Louis Allard, Kee Fui Kon, Yasuo Morishima, and Ruediger Kotzian
Improved anchorage and bird protection with iron-coated seeds 209
in wet direct seeding of rice crops
Minoru Yamauchi
Issues for integrated weed management and decision support 211
in direct-seeded rice
D.E. Johnson and A.M. Mortimer
Control of Leptochloa chinensis (L.) Nees in wet-seeded rice fields in Sri Lanka 215
Anuruddhka S.K. Abeysekera and U.B. Wickrama
Village-level modeling of environment-friendly and appropriate 217
technologies and practices for direct seeding
Rowena G. Manalili, Bernard D. Tadeo, Emmanuel R. Tiongco,
Wilfredo B. Collado, Rodolfo V. Bermudez, Constancio A. Asis,
Jovino L. De Dios, Marvin F. Adap, Mario dela Cruz, Ulysses G. Duque,
Leonardo V. Marquez, Cheryll B. Casiwan, Roy F. Tabalno,
Placida C. Lanuza, and Belen C. Tejada
Wrap-up of Session 6 220
SESSION 8
Improving rice quality
The chemical basis of rice end-use quality 246
Kshirod R. Bhattacharya
Improving rice grain quality in Thailand 248
Kunya Cheaupun, Sunantha Wongpiyachon, and Ngamchuen Kongseree
New tools for understanding starch synthesis 251
Melissa Fitzgerald, Jeffrey Castro, Rosa Paula Cuevas, and Robert Gilbert
Textural differences between indica and japonica varieties in cooked rice 253
Keiko Hatae, Sonoko Ayabe, and Midori Kasai
Biological efficacy of consuming rice biofortified with iron 256
J. Haas, J. Beard, A. Del Mundo, G. Gregorio, L.M. Kolb, and A. Felix
Radical-scavenging activity of red and black rice 256
Tomoyuki Oki, Mami Masuda, Saki Nagai, Miwako Take’ichi,
Mio Kobayashi, Yoichi Nishiba, Terumi Sugawara, Ikuo Suda,
and Tetsuo Sato
A rice mutant with enhanced amylose content in endosperm derived 260
from low-amylose variety Snow Pearl: isolation and characterization
Yasuhiro Suzuki, Hiro-Yuki Hirano, Yoshio Sano, Kazuo Ise,
Ushio Matsukura, Noriaki Aoki, and Hiroyuki Sato
The role of the water-soluble fraction in rice pasting behavior 261
Tadashi Yoshihashi, Eizo Tatsumi, Vipa Surojanametakul,
Patcharee Tungtrakul, and Warunee Varanyanond
Wrap-up of Session 8 265
SESSION 9
Developing new uses of rice
Overview of rice and rice-based products 268
Bienvenido O. Juliano
Sterilization effect of electrolyzed water on rice food 270
Seiichiro Isobe, Chang-yong Lee, and Kyoichiro Yoshida
Current status of varietal improvement and use of specialty rice in Korea 272
Hae Chune Choi
Contents vii
Processed novel foodstuffs from pregerminated brown rice 275
by a twin-screw extruder
Ken’ichi Ohtsubo, Tomoya Okunishi, and Keitaro Suzuki
High-pressure food processing of rice and starch foods 278
Rikimaru Hayashi
Developing novel processes for incorporating the unique nutritional 280
and functional properties of rice into value-added products
Elaine T. Champagne, Harmeet S. Guraya, Frederick F. Shih,
and Ranjit S. Kadan
Dehydrin proteins in rice bran 283
Michiko Momma
Physicochemical properties of modified rice flour and its use 285
for processed food
T. Takahashi, M. Miura, N. Ohisa, K. Mori, and S. Kobayashi
Introducing soybean β-conglycinin genes into rice to improve nutritional 288
and physiological value
Takayasu Motoyama, Nobuyuki Maruyama, Takahiko Higasa,
Masaaki Yoshikawa, Fumio Takaiwa, and Shigeru Utsumi
Wrap-up of Session 9 291
SESSION 10
Postharvest technology for efficient processing and distribution of rice
Postharvest technology for rice in India: a changing scenario 294
Pallab Kumar Chattopadhyay
Development of a far-infrared radiation dryer for grain 296
Yasuyuki Hidaka, Kotaro Kubota, and Tomohiko Ichikawa
Status of rice milling and use of by-products 299
Naoto Shimizu, Yuji Katsuragi, and Toshinori Kimura
Advanced application technology of rice bran: preparation of ferulic acid 301
and its applications
Hisaji Taniguchi, Eisaku Nomura, and Asao Hosoda
New movements regarding the safety of rice and residual agrochemical inspection 304
Yukio Hosaka
Impact of infrastructure on profitability and global competitiveness 307
of rice production in the Philippines
Rowena G. Manalili and Leonardo A. Gonzales
Development of an on-farm rice storage technique using fresh chilly air and 310
preservation of high-quality rice
Shuso Kawamura, Kazuhiro Takekura, and Kazuhiko Itoh
Life-cycle inventory analysis of local parboiling processes 313
Poritosh Roy, Naoto Shimizu, Takeo Shiina, and Toshinori Kimura
Using genetics to create changes in rice cooking, processing, storage, 315
and health-beneficial properties
Christine Bergman
The role of proteins in textural changes in aged rice 318
Toshihisa Ohno, Takahiro Kaneko, and Naganori Ohisa
Postharvest technology of rice: the role of farm women in storing grains 320
with different storage practices
P. Sumathi and M.N. Budhar
Wrap-up of Session 10 322
SESSION 11
Enhancing the multifunctionality of rice systems
Multifunctional roles of paddy irrigation in monsoon Asia 324
Takao Masumoto
Accounting for culture in paddy cultivation: toward a broader definition 327
of “livelihood”
David Groenfeldt
SESSION 12
Conservation of soil, water, and environment in rice culture
Paddy soils around the world 352
K. Kyuma
Sustainability of paddy soil fertility in Vietnam 354
Ngo Ngoc Hung, Nguyen Bao Ve, Roland J. Buresh, Mark Bayley,
and Takeshi Watanabe
Managing soil fertility for sustainable rice production in northeast Thailand 357
Kunnika Naklang
Site-specific nutrient management and the sustainability of phosphorus 360
and potassium supply in irrigated rice soils of Asia
C. Witt, A. Dobermann, R. Buresh, S. Abdulrachman, H.C. Gines,
R. Nagarajan, S. Ramanathan, P.S. Tan, and G.H. Wang
Ecological engineering for sustainable rice production and the restoration 363
of degraded watersheds in West Africa
Toshiyuki Wakatsuki, Md. Moro Buri, and Oluwarotimi O. Fashola
Nitrogen cycling under the rice-wheat rotation and environmental effects 367
Jian-guo Zhu, Xiaozhi Wang, Zu-cong Cai, Ren Gao, and Yasukazu Hosen
Urea deep placement increases yield and saves nitrogen fertilizer 369
in farmers’ fields in Bangladesh
W.T. Bowen, R.B. Diamond, U. Singh, and T.P. Thompson
New fertilizer management to maximize yield and minimize environmental 372
effects in rice culture
Masahiko Saigusa
Does anaerobic decomposition of crop residues impair soil nitrogen 374
cycling and yield trends in lowland rice?
D.C. Olk, K.G. Cassman, M.M. Anders, K. Schmidt-Rohr, and J.-D. Mao
Influence of the paddy-upland rotation on soil physico-chemical properties 377
and crop growth in the Honam Plain of Korea
Lee Deog-Bae, Yang Chang-Hyu, Ryu Chul-Hyun, Lee Kyeong-Bo,
and Kim Jae-Duk
A decrease in soil fertility and crop productivity by succession of 379
the paddy-upland rotation
Hirokazu Sumida
Contents ix
Promising technologies for reducing cadmium contamination in rice 381
Satoru Ishikawa
N uptake inhibition of the rice plant in flooded soils receiving wheat straw 384
Fukuyo Tanaka
Stable isotope ratios of hydrogen and oxygen in paddy water affected 386
by evaporation
Yohei Hamada, Shiho Yabusaki, Norio Tase, and Ichiro Taniyama
Erosion control by sawah in comparison to other land-use systems 388
Fahmuddin Agus and Irawan
Wrap-up of Session 12 390
SESSION 13
Farmers’ participatory approaches to facilitate adoption of improved technology
How does a farmer accept a new technology? 394
Fujihiko Tozawa
Farmer participatory evaluation of nitrogen management technology: the case 395
of the leaf color chart in West Bengal, India
B. Bagchi, M.Z. Abedin, and S.K.T. Nasar
A farmer participatory approach in the adaptation and adoption of controlled 397
irrigation for saving water: a case study in Canarem, Victoria, Tarlac,
Philippines
F.G. Palis, M. Hossain, B.A.M. Bouman, P.A.A. Cenas, R.M. Lampayan,
A.T. Lactaoen, T.M. Norte, V.R. Vicmudo, and G.T. Castillo
Companion modeling and multi-agent systems for collective learning 401
and resource management in Asian rice ecosystems
F. Bousquet and G. Trébuil
Participatory approaches for improving rice breeding in the Mekong Delta 404
of Vietnam
Nguyen Ngoc De and Kotaro Ohara
IRRI’s approach to participatory research for development: advances 408
and limitations
Thelma R. Paris and M. Zainul Abedin
A participatory approach for building sustainable rice-farming systems 411
in the reclaimed farmland of Ogata, Japan
Yoshimitsu Taniguchi and Satoru Sato
Rice farmers’ participatory research has played a key role 412
in implementing the System of Rice Intensification
Dandu Jagannadha Raju
Wrap-up of Session 13 415
SESSION 14
Potentials for diversification in rice-based systems to enhance rural livelihoods
Agricultural diversification in Asia: opportunities and constraints 420
Prabhu Pingali
Consequences of technologies and production diversification for 422
the economic and environmental performance of rice-based farming
systems in East and Southeast Asia
Huib Hengsdijk, Marrit van den Berg, Reimund Roetter, Wang Guanghuo,
Joost Wolf, Lu Changhe, and Herman van Keulen
Rural poverty and agricultural diversification in Thailand 425
Alia Ahmad and Somporn Isvilanonda
Sustaining higher efficiency in rice production 428
Amelia S. delos Reyes, Arelene Julia B. Malabayabas,
and Mercedita A. Sombilla
Determinants of agricultural diversification in Vietnam: changes 432
at the farm level in the Mekong and Red River deltas
Magnus Jirström and Franz-Michael Rundquist
SESSION 15
Challenges to expanding rice production in unfavorable environments
In vitro selection of somaclonal and gametoclonal variants for salt tolerance 442
in rice
Nguyen Thi Lang, Dang Minh Tam, Hiromi Kobayashi, and Bui Chi Buu
Submergence damage in rice and challenges in expanding the crop’s 445
adaptability to submerged conditions in West and Central Africa
Koichi Futakuchi
Ecological, morphological, and physiological aspects of drought adaptation 448
of rice in upland and rainfed lowland systems
Shu Fukai and Akihiko Kamoshita
Managing iron toxicity in lowland rice: the role of tolerant genotypes 452
and plant nutrients
Kanwar L. Sahrawat
Soil acidity and related problems in upland rice in the tropics 454
Kensuke Okada and Matthias Wissuwa
The physiological foundation of crop breeding for stress environments 456
A. Blum
Expression of a serine protease during microsporogenesis in rice 458
Kentaro Kawaguchi, Naoshi Dohmae, Shuichi Matsuba,
Hideyuki Funatsuki, Yutaka Sato, and Masao Ishimoto
Responses to chilling temperature at the early stage of development 460
in rice: geographical clines and genetic bases as revealed by QTL analysis
Kazumitsu Onishi, Noriko Ishigoh-Oka, Mieko Adachi, and Yoshio Sano
QTL analysis on plasticity in lateral root development in response 463
to water stress in the rice plant
Hong Wang, Yoshiaki Inukai, Akihiko Kamoshita, Len Wade,
Joel Siopongco, Henry Nguyen, and Akira Yamauchi
Wrap-up of Session 15 465
SESSION 16
Pest management with minimal environmental stress
Integrated biodiversity management (IBM) in rice fields 468
Keizi Kiritani
Habitat manipulation in sustainable pest management in the rice ecosystem 470
of the Yangtze River Delta
Xiaoping Yu, Jianming Chen, Zhongxian Lu, Xusong Zheng, Hongxin Xu,
Juefeng Zhang
Evaluating augmentative releases of the mirid bug Cyrtorhinus lividipennis 473
to suppress brown planthopper Nilaparvata lugens in open paddy fields
Masaya Matsumura, Satoru Urano, and Yoshito Suzuki
Developing a rice production system for sustainable pest management 475
T.W. Mew
Evaluation of a leaf blast simulation model (BLASTMUL) for rice multilines 477
in different locations and cultivars, and effective blast control using
the model
T. Ashizawa, M. Sasahara, A. Ohba, T. Hori, K. Ishikawa, Y. Sasaki,
T. Kuroda, R. Harasawa, K.S. Zenbayashi, and S. Koizumi
Association of candidate defense genes with quantitative resistance to rice blast 479
and in silico analysis of their characteristics
G. Carrillo, J. Wu, B. Liu, N. Sugiyama, I. Oña, M. Variar, B. Courtois,
J.E. Leach, P.H. Goodwin, H. Leung, and C.M. Vera Cruz
Transferring resistance genes among different cereal species 482
Bingyu Zhao, Shavannor Smith, Jan Leach, and Scot Hulbert
Contents xi
Screening of allelopathic activity from rice cultivars by bioassay and field test 484
Yoshiharu Fujii, Hiroshi Araya, Syuntarou Hiradate, and Kaoru Ebana
Evaluating near-isogenic lines with QTLs for field resistance to rice blast 487
from upland rice cultivar Sensho through marker-aided selection
Norikuni Saka and Shuichi Fukuoka
Wrap-up of Session 16 490
SESSION 17
Rice supply and demand
International trade in rice: recent developments and prospects 492
Concepcion Calpe
Rice in the world at stake 494
Shoichi Ito
Rice consumption in China: Can China change rice consumption 497
from quantity to quality?
Chien Hsiaoping
Surplus rice supply in Asia 499
Vichai Sriprasert
Review of existing global rice market models 502
Eric J. Wailes
Household rice consumption in Japan: quantity and price by income 505
while controlling for household types
Kimiko Ishibashi, Yoshinobu Kono, and Yuuji Ooura
Wrap-up of Session 17 507
SESSION 18
Impact of globalization on rice farmers
Impacts of distorted trade policies on rice productivity 510
Manitra A. Rakotoarisoa
A new strategy for group farming in Japan 513
Kyoichi Miyatake
The role of the rice economy after the implementation of agricultural 515
policy reform and trade liberalization from the perspective of farmers:
the case of rice farmers in Java, Indonesia
Roosiana
El Niño sensitivity, resource endowment, and socioeconomic characteristics: 517
the case of wetland rice in Java, Indonesia
Shigeki Yokoyama and Bambang Irawan
Global competitiveness of medium-quality Indian rice: a PAM analysis 520
B.V. Chinnappa Reddy, M.S. Raghavendra, and Lalith Achoth
Behavior and strategies of Japanese rice producers under globalization 523
Masaki Umemoto
The future perspective of upland rice farmers in Indonesia in the era of 525
globalization
Yusman Syaukat and Sushil Pandey
Impact of globalization on rice farmers in Thailand 527
Rangsan Pitipunya
The rice economy and rice policy in China 530
Li Ninghui
Wrap-up of Session 18 532
SESSION 19
Climate change and rice production
Effects of elevated atmospheric CO2 concentration and increased temperature 536
on rice: implications for Asian rice production
T. Horie, H. Yoshida, S. Kawatsu, K. Katsura, K. Homma, and T. Shiraiwa
Monitoring rice growth and development using a crop model: the case 539
of northern Japan
Masaharu Yajima
SESSION 20
Improving rice productivity through IT
Data mining using combined yield and quality maps of paddy fields 566
Tadashi Chosa
A wireless sensor network with Field-Monitoring Servers and MetBroker 568
in paddy fields
Masayuki Hirafuji, Tokihiro Fukatsu, Hu Haoming, Takuji Kiura,
Tominari Watanabe, and Seishi Ninomiya
Prediction of airborne immigration of rice insect pests 571
Akira Otuka, Tomonari Watanabe, Yoshito Suzuki, Masaya Matsumura,
Akiko Furuno, and Masamichi Chino
Distance education and eLearning for sustainable agriculture: 574
lessons learned
Robert T. Raab and Buenafe R. Abdon
The Rice Knowledge Bank 576
Mark Bell and David Shires
Integrated information systems for crop research and improvement 579
Graham McLaren, Arllet Portugal, Alexander Cosico, William Eusebio,
Teri Ulat, May Ann Sallan, Victor Jun Ulat, and Richard Bruskiewich
Using APAN for content delivery: possibilities for the CGIAR 583
Paul O’Nolan
A decision support system for site-specific nitrogen management 587
of paddy rice
Ryouji Sasaki, Kazunobu Toriyama, and Yoichi Shibata
Wrap-up of Session 20 589
Contents xiii
Foreword
Just after World War II, rapid population growth with limited rice production led
experts to predict starvation in Asia. On its own, the Food and Agriculture Organiza-
tion (FAO) of the United Nations had declared 1966 the Year of Rice and numerous
countries took measures to improve production, marketing, milling, and nutrition.
Conferences were organized and scientific research and technology were stimulated.
Rice is now the staple food for over half of the world’s population but its production
faces many constraints under conditions of increasing world population and dimin-
ishing resources of water and land.
The United Nations launched the International Year of Rice 2004 on 31 Octo-
ber 2003. This is the second time that the United Nations has paid such a special
tribute to rice, the only food crop honored twice. Rice is the single most important
employment and income source for the rural poor. Rice will play a significant role in
meeting the important UN Millennium Development Goal of poverty reduction in
the world. Besides being an essential food, rice is also an important factor in enrich-
ing culture, lifestyles, and ecosystem functions. It is thus fitting that the United Na-
tions pronounced 2004 the International Year of Rice to emphasize the important
roles rice plays in the livelihoods and culture of humankind. Rice is a symbol of
cultural identity, global unity, and life.
To mark the Year of Rice 2004, the Ministry of Agriculture, Forestry, and Fish-
eries (MAFF) of Japan, research organizations affiliated with MAFF, and the Inter-
national Rice Research Institute (IRRI) jointly financed and organized the World
Rice Research Conference (WRRC) that was held in Tokyo and Tsukuba, Japan, 4-7
November 2004. The WRRC had two parts: the Tokyo Opening Ceremony and Sym-
posium in the Akasaka Prince Hotel graced by His Imperial Highness Crown Prince
Naruhito of Japan on 4 November and the International Rice Research Conference at
the Tsukuba Congress Center on 5-7 November.
The WRRC attracted 1,274 participants from 43 nations, who presented 190
scientific papers and 302 posters in 20 sessions and 6 workshops. Scientific themes
covered genomics to climate change and involved scientists with expertise in genes
to ecosystems. Such a broad range of subjects discussed during the conference makes
the WRRC one of the most significant scientific events of the last few decades. Sci-
entists presented their latest concepts, research findings, and products, which are
captured in the topics presented in this proceedings. The proceedings contains the
state of the art in rice science and production that we hope will be useful to rice
scientists, extension specialists, development agents, and policymakers who will use
this to better the lives of all humans, but especially those of poor farmers and con-
sumers.
AKINORI NOGUCHI
Vice president of the Japan International Research Center
for Agricultural Sciences and chairman
of the WRRC 2004 Organizing Committee
The success of the World Rice Research Conference (WRRC) was due largely to the
constant endeavor of many individuals and organizations. The framework of this
conference was set up by the executive committee composed of the Division of Inter-
national Research in the Secretariat of Agriculture, Forestry, and Fisheries Research
Council of the Ministry of Agriculture, Forestry, and Fisheries (MAFF) and the presi-
dents of co-sponsoring research institutes as well as the delegate from the Interna-
tional Rice Research Institute (IRRI).
The organizing committee consisted of A. Noguchi and Y. Morokoka (Japan
International Research Center for Agricultural Sciences, JIRCAS) as chairs, with
seven representatives from the co-sponsoring research institutes: T. Imbe (National
Agricultural and Bio-oriented Research Organization, NARO), K. Okuno (National
Institute of Agrobiological Sciences; NIAS), M. Oka (National Institute for Agro-
Environmental Sciences, NIAES), Y. Tsutsui (National Institute of Rural Engineer-
ing, NIRE), T. Imai (National Food Research Institute, NFRI), Y. Yunoki (MAFF),
and K.L. Heong (IRRI). Special thanks are due to M. Iwamoto, former president of
JIRCAS, and E. Miwa, president of NARO, for their valuable suggestions and con-
stant encouragement.
The secretariat was successfully handled by JIRCAS personnel headed by O.
Ito, with R. Ikeda, T. Kumashiro, K. Toriyama, K. Yasunobu, T. Uetani, H. Fujii, S.
Tsuchiya, R. Saito, H. Tanaka, T. Hatta, Y. Hamada, and H. Miura. For Web manage-
ment, E. Hettel (IRRI) and H. Miura (JIRCAS) contributed significantly. Special
appreciation goes to IRRI personnel: K.L. Heong for coordinating with IRRI and the
JIRCAS secretariat and giving much useful advice for the success of the WRRC, D.
Macintosh for press releases, and B. Hardy for editorial work on the manuscripts of
this large proceedings.
The conveners of each session organized and provided a first review of the
papers from their respective session contributers. Without their strong efforts, the
WRRC could not have succeeded. These are K. Okuno and R.S. Hamilton for S1; T.
Sasaki, H. Hirochika, and H. Leung for S2; F. Takaiwa, S. Oka, and S. Datta for S3;
M. Kondo, H. Ikehashi, and S. Peng for S4; Y. Fukuta, D. Mackill, and R. Ikeda for
S5; M. Yamauchi and D. Johnson for S6; N. Itokawa and J. Rickman for S7; U.
Matsukura for S8; K. Otsubo for S9; T. Kimura and J. Rickman for S10; Y. Tsutsui
and K. Yamaoka for S11; M. Saito, S. Ito, and T. Nozoe for S12; T. Paris and J.
Caldwell for S13; M. Hossain for S14; S. Tobita and R. Lafitte for S15; C. Vera Cruz
and Y. Suzuki for S16; O. Koyama and D. Dawe for S17; M. Umemoto and D. Dawe
for S18; T. Imagawa, T. Hasegawa, Y. Hayashi, and J. Sheehy for S19; and S. Ninomiya
and M. Bell for S20.
This conference was supported by different academic societies and it attracted
many participants of different disciplines. These were the Japan Society of Breeding,
Crop Science Society of Japan, Japanese Society of Soil Science and Plant Nutrition,
the Japanese Society of Irrigation, Drainage, and Reclamation Engineering, the Japa-
nese Society of Agricultural Machinery, the Japanese Society of Applied Entomol-
ogy and Zoology, the Phytopathological Society of Japan, Pesticide Science Society
of Japan, the Farm Management Society of Japan, the Japanese Society for Food
Contents xv
Science and Technology, the Society of Agricultural Meteorology of Japan, Japanese
Society for Tropical Agriculture, and Science Council of Japan (National Committee
for Rural Planning).
Financial support was given by the Japan Science Promotion Society. Special
thanks are also due to the Tsuchiura Tsukuba Convention Bureau and the Interna-
tional Congress Center (Epochal Tsukuba).
Folklorist Yanagida Kunio (1875-1962), known as the father The Japanese call the community system work kou. Simi-
of Japanese folklore study, once raised the question, “What is lar systems developed in Okinawa are called yui, and ke in
Japan?” In his view, insularity and rice cultivation are the two Korea and ban in South China.
key factors composing Japanese ethnicity. In other words, In these community work systems, innovations and pro-
Yanagida sees our Japanese way of life as the result of being cesses to increase productivity develop to improve ancestral
rice farmers on an island. farming land. In other words, the whole community is encour-
I would like to add another factor to these two: settled aged to exert an effort to breed crops and control the crops’
life, rooted in a community and a plot of land. quality, as well as construct drainage and irrigation systems
Some may argue that rice cultivation, either paddy rice for the community and weed the rice paddies.
or dry rice, is naturally coupled with a settled life and is too As another example, to harvest more crops on the same
obvious to mention as a new factor. However, this is not so. areas of land that their ancestors reclaimed, farmers must breed
If you see rice cultivation as a part of life deeply con- rice stocks more resistant to extreme temperature, drought, and
nected to ethnicity, being a rice farmer means living a settled disease. They also carry out thorough quality control by re-
life on specific parcels of land. However, if you see rice farm- moving pests and weeds from the rice field, improving the
ing as an industry that generates profits, farmers may choose rice paddy to create efficient irrigation and drainage condi-
to plant strawberries instead of rice when profits are low, as tions, and figuring out what the most suitable spaces are be-
happens in California. Furthermore, they may decide to sell tween rice plants.
their land to Disneyland and move to other areas. In this case, This lifestyle is very different from that of Western coun-
rice cultivation is not inevitably linked to a settled lifestyle. tries, where productivity increased through territorial expan-
California rice farmers are not the only ones who choose sion. The Western ranchers’ way of adding new pastures out-
alternate lifestyles. Another good example is Indonesia. Indo- side the towns, in my opinion, demonstrates the principle of
nesia achieved 100% rice self-sufficiency in 1997. Soon, the the “power to expand.” This naturally extends expansion to
farmers recognized that shrimp cultivation would triple their products and innovations in fields such as traffic technology,
income and they turned their paddies into shrimp cultivation transportation, communication, and military affairs.
ponds. As a result, the self-sufficiency rate of rice dropped to On Earth today, we do not have any new land to expand
80%. onto. The natural environment is worsening, the population is
Now, let us focus on settled rice farming as an ethnic exploding, and a food crisis is just around the corner.
lifestyle. The reward of your life as a farmer is to sustain your If we recognize this reality, isn’t this the time for us to
paddies and paddy terraces passed down from your ancestors. remember “the power of inner accumulation,” traditionally cul-
Issho-kenmei, a Japanese idiom for “do your best,” literally tivated by rice farming? It is the wisdom of various ethnic
means “maintain a piece of land for all your worth.” This idea groups to live on the same land for generations, thus increas-
has been supported as an ethnic ethos of Japanese. This de- ing wealth in a village, and in a rice paddy, in a single rice
scribes why Japanese pay more respect to the process of “do- grain. Rice farming will provide the core motive for the settled
ing your best” than to getting good results or high profits. lifestyle we now seek.
Sharing the ethos of issho-kenmei, people living in an
area work together to dig drainage ditches, build irrigation sys-
tems, and weed the whole village. The working group unavoid-
Notes
ably encourages the development of a community system. Author’s address: Professor, Reitaku University.
Rice is the starting point of Japanese food culture. In a tradi- The next point we must keep in mind in a rice-based diet
tional Japanese diet, the carbohydrates contained in rice are is the imbalance in nutrition when one eats rice too much. White
the main energy source. The average Japanese eats about 594 rice is most appetizing when accompanied by salty dishes.
kcal of rice and rice products per day, which is 31% of the Because of this, one is more likely to develop unhealthy eat-
average energy intake. Rice contains 6.1 g of protein per 100 ing habits—consuming a small amount of salty food and eat-
g, and, though this level is not so high, rice protein is superior ing a lot of white rice, thinking that white rice is the principal
in quality to wheat protein. Therefore, rice alone can support a food. This eating pattern results in an overconsumption of salt.
person if that person eats enough of it. Even today, the Japa- Sodium causes high blood pressure and a higher chance of
nese receive 12% of their daily protein from rice, and decades stroke. In addition, this diet pattern often lacks food made from
ago many sustained their lives by eating only rice. That is why animal products that supply a certain amount of high-quality
the Japanese call it shu-shoku, the principal food. protein necessary to avoid a range of diseases.
The conventional Japanese diet also contains a variety In the 1930s, physicians Shozo Kagawa and Aya Kagawa
of fuku-shoku, the side dishes served along with rice. These treating beriberi patients found a method to produce “rice with
wholesome foods are economical, too. Rice at breakfast re- the embryo.” They promoted “rice with the embryo” to pre-
duces the day’s food expense compared to popular Western- vent beriberi, and founded an educational institute that later
style breakfasts. Furthermore, a rice-based diet is healthier in became the Kagawa Nutrition University. To help people re-
nutritional balance. member the benefits of a balanced “rice with the embryo” diet,
Although rice is an excellent food, a large problem in- they made a campaign slogan: the principal food is “rice with
volving a rice-rich diet exists: beriberi, a thiamine deficiency. the embryo,” accompanied by one fish, one bean, and four
This disease emerged with the grain polishing that produces vegetables. This reminded people to eat 100 g of fish, 100 g of
white rice, intended to improve the digestion and taste of bean and bean products, and 400 g of vegetables a day.
cooked rice. The process removes the rice germ that contains The idea of a food guideline based on food groups was
most of the naturally occurring vitamin B1. Some 0.42 mg of new at the time. This eating method was widely accepted by
vitamin B1 is required for the metabolism of 1,000 kcal of the educated homemakers of Japan. After World War II, milk
carbohydrates. Therefore, for people who try to sustain their and milk products were introduced to school lunch programs.
diet mainly by rice, the relative shortage in vitamin B1 devel- Children’s better health proved its effectiveness and milk was
ops into an absolute deficiency and causes beriberi. added to the list. This is how the well-known and popular “four
An adult male needs 1.1 mg of vitamin B1 per day, and food groups” method we use today developed.
an adult female needs 0.7 mg. This is why young men who A rice-based diet has another advantage: rice is eaten as
consume more food and energy are more likely to develop a form of grain. Grains are digested and taken into the body
beriberi. In the 1920s and ’30s, when the Japanese diet de- more slowly than milled cereals. Therefore, in spite of low-fat
pended much more on rice, beriberi killed more than 10,000 and low-calorie meals, one has a long-lasting feeling of satiety
people and terrified Japanese called it a “nation-ruining dis- and less insulin is consumed. The key to the lower rate of dia-
ease.” Beriberi is common to ethnic groups that eat rice as betes in the Japanese population in the past seems to be rice.
their principal food, and has often occurred in Southeast Asian Rice is a delicious and handy food. I would like to emphasize
countries, including Indonesia. that it is an excellent food if you avoid too much polishing and
It is easy to solve the thiamine problem: just change to pay close attention to good nutritional balance.
“rice with the embryo,” produced by polishing the grain al-
most white but leaving the embryo. If one prefers white rice,
Notes
eating vitamin B1-enriched rice and/or accompanying dishes
rich in vitamin B1 prevents beriberi. Getting enough vitamin Author’s address: President of Kagawa Nutrition University, Japan.
B1 is the first and major issue in a rice-based diet.
Keynotes 3
A greeting
Yoshinobu Shimamura
The United Nations has designated 2004 as the International tion that suffers from hunger and poverty lives in Asia and
Year of Rice (IYR) and countries around the world are joining Africa. New research findings related to rice are expected to
hands to hold events to raise awareness about rice, which plays help eradicate hunger and poverty among farmers and others
a major role in eliminating hunger and poverty in developing in developing nations.
nations. Japan kicked off this special year with the Interna- To address these issues, Japan has provided cooperation
tional Symposium to Celebrate IYR, which was sponsored by to improve rice varieties and establish cultivation technolo-
the Ministry of Agriculture, Forestry, and Fisheries in January. gies in Asia. Our nation also plays a leadership role in rice
A variety of other events, spearheaded by the Japanese Com- research. In addition to working alongside ten other regions
mittee for the International Year of Rice 2004, are being staged and countries to sequence the rice genome, Japan is engaged
throughout the nation this year. in research on New Rice for Africa (NERICA) in cooperation
Rice is not only the staple food of monsoon Asian; it with international research institutes. To achieve the UN Mil-
also plays an important role in regional culture because of the lennium Development Goals, Japan will continue to promote
idyllic agricultural villages that form around the beautiful pad- research to meet international expectations.
dies. By holding the World Rice Research Conference during
The Japanese people’s lifestyle and society are based on this memorable year, I am confident that we will strengthen
rice and rice cultivation and have advanced thanks to the tire- cooperation among rice researchers throughout the world and
less efforts of farmers and researchers. Rice research has pro- further stimulate research to eradicate world hunger and pov-
duced many technologies and varieties of rice and contributed erty and solve environmental problems.
greatly to the furtherance of our nation’s agriculture.
Rice is the staple food of more than half of the world’s
population, and about 1 billion households depend on rice cul-
Notes
tivation for employment and their main source of livelihood. Author’s address: Minister of Agriculture, Forestry, and Fish-
At the same time, approximately 800 million people in devel- eries, Japan.
oping nations suffer from hunger. Most of the world popula-
Keynotes 5
eliminating hunger and poverty. I believe that this conference By bringing together the wisdom from all over the world
is very significant as an opportunity for experts to discuss the and promoting active discussions from a wide range of per-
production and use of rice and the many roles of rice paddies spectives, I hope that this conference will make a great contri-
from the perspective of science and technology, and to con- bution to solving the world’s food problems and protecting
sider the best avenues for future rice research. the global environment.
Major advances have taken place in rice production during the fine-tune the strategies for managing natural resources. Fertil-
last four decades because of the adoption of Green Revolution izer is the most important input in crop production. Yet, area-
technology. Rice production increased 130% from 257 mil- specific or regional fertilizer recommendations are based on
lion tons in 1966 to 598 million tons in 1999. Average rice short-term trials, which do not reflect the long-term impact of
yield increased from 2.1 to 3.9 t ha–1 during the same period. fertilizer use on the resource base. The chlorophyll meter is an
In 2000, average per capita food availability was 18% higher excellent technique to measure the leaf N status of rice and to
than in 1966. Whereas rice production increased at 3.0% and synchronize N application with crop demand. Water efficiency
2.5% per annum during the 1970s and ’80s, respectively, the can be improved through its judicious use and zero tillage.
increase was only 1.5% during the 1990s. During the last four Vast lands are unsuitable for crop production or have poor
years, production has been stagnant. According to the Interna- productivity because of soil problems such as salinity and al-
tional Food Policy Research Institute, rice production must kalinity. Gypsum can be used to reclaim acidic and saline soils.
increase 38% by 2025 to feed 4 billion rice consumers. Crop diversification is necessary to improve the productivity
The area planted to rice is declining because of the pres- of intensively farmed soils with cereal crops. Crop rotations in
sure of urbanization and industrialization. Availability of wa- addition to improving the soil structure show allelopathic ef-
ter for agriculture is declining and labor is moving to industry. fects through the release of chemical compounds directly in
To meet the challenge of producing more rice under those con- the soil or indirectly through microbial decomposition of resi-
straints, we need new technologies. These include rice variet- due, which affect the growth of another species. Integrated
ies with higher yield potential, greater yield stability, and pest management (IPM) approaches have led to a reduction in
adapted to changing global climate, as well as more efficient the overuse and misuse of pesticides in rice production in a
management practices. few countries. Universal adoption of IPM should lead to a re-
In addition to conventional hybridization and selection duction in the cost of rice production and more healthy rice.
procedures, other strategies for developing rice varieties with
higher yield potential include ideotype breeding, exploitation
Notes
of heterosis, wide hybridization, molecular breeding, genetic
engineering, and apomixis. Similarly, marker-aided selection Author’s address: Adjunct Professor, University of California, Davis.
and genetic engineering approaches are useful for developing
rice varieties with more durable resistance to diseases and in-
sects and tolerance of abiotic stresses such as drought, sub-
mergence, and salinity.
Precision agriculture should lead to more sustainable in-
creases in rice production, It is important to improve the effi-
ciency of all production inputs such as nutrients and water and
For centuries, rice has been one of the world’s most important O Rice markets are generally characterized as thin, con-
food crops. For the people of Asia in particular, rice has been centrated, and volatile. Because most rice is consumed
the main source of calories and an important source of income where it is produced, only a small share of produc-
and employment throughout most of their history. Furthermore, tion is traded internationally. In the fastest globaliz-
rice has shaped societies and cultures. Today, the cultivation, ing region of the world, Asia, it is surprising that the
marketing, and consumption of rice are changing faster than most important crop, rice, is not globalizing—its in-
ever before, yet there are also strong forces working to stabi- tegration into the international traded economy re-
lize and conserve rice systems. The economics and politics of mains limited, lagging far behind other commodities.
rice are changing because of globalization, technological trans- What are the reasons behind this phenomenon and
formation of agriculture driven by science, and changing diets can we expect this situation to change in the future?
and tastes due to prosperity and urbanization. At the same time, O Rice policies are partly shaped by culture, in particu-
rice remains the lifeline of many poor producers and consum- lar as they affect consumer and producer behaviors,
ers. Deeply embedded in established food systems, rice re- but for how much longer will that be the case as soci-
mains local and traditional and appears to be partially excluded etal transformations progress? The political economy
from globalization trends. Thus, rice is a most modern food of rice is changing, and that will shape rice produc-
and a very traditional one at the same time. tion and consumption in the future. Even though the
As 2004—the International Year of Rice, as declared by long history of state engagement in rice stockholding
the United Nations—comes to a close, it is pertinent to ad- and trading is coming to an end, rice remains a strate-
dress key issues about the emerging changes in the economics gic food security crop for policymakers and voters.
and politics of rice and to look into the future of rice and the What will be the future role of governments and the
key role it plays in the livelihood of millions of people. private sector in rice systems?
O Rice consumption plays a crucial role in assuring food
security to people in Asia, and parts of Africa and We approach these questions with deep respect for rice and
Latin America. However, diets are diversifying and what it means to the people and cultures in Asia and around
other crops are emerging as a result of growing ur- the world, and are conscious that our brief review can only
banization and rising incomes. Moreover, demand for address selected aspects of the above-mentioned issues and
rice is shifting from lower-quality rice to higher-qual- themes.1
ity rice. What will be the patterns of demand for rice
in the future?
The culture of rice
O Rice production and consumption remain closely
linked. Only a small proportion of the rice produced In many societies, when people consume or produce rice, they
is traded. Furthermore, production of different quali- are not just consuming calories or producing grain, they are
ties of rice remains tied to the tastes and preferences also engaging in practices that have an intrinsic cultural value.
of the region where it is produced. What are the rice While prices and incomes are important drivers of demand,
production patterns across the world and how will other factors such as traditions are also significant. For centu-
they change? ries, rice has shaped Asian societies and their cultures. Asian
O Rice systems are highly sustainable, if well managed. cultures are partly cultures of rice, and many Asian societies
However, it remains a permanent challenge to main- relate to rice beyond the satisfaction of basic needs. Rice is
tain the environmental sustainability of the diverse mentioned in the scriptures of the ancient civilizations of Asia.
rice ecosystems. The environmental externalities re- Its cultivation was considered as the basis of the social order
lated to rice production are complex. Moreover, rice and occupied a major place in Asian religions and customs
is increasingly grown in irrigated ecosystems; thus,
as water becomes increasingly scarce, priority should
be given to research aimed at producing water-sav-
ing rice varieties. What is the future of rice technol- 1We gratefully acknowledge comments on an earlier draft by Ron Cantrell,
ogy? Susumu Matsuoka, Rajul Pandya-Lorch, and Keijiro Otsuka.
Keynotes 7
(Hossain 1998). Hamilton (2003) describes the integral and US$ per ton
vital role of rice as follows: “A key tenet of rice culture is that 90
rice is a sacred food divinely given to humans that uniquely
sustains the human body in a way that no other food can.”2 80
Rice is so embedded in Asian cultures that even the languages
70
reflect the special nature of rice as the primary food for hu-
mans. In many Asian languages, there is no general word for 60
“food” other than the word for rice and an invitation to “eat”
implies the eating of rice. Moreover, growing rice used to be 50
considered an ideal form of human labor and reflective of a
well-ordered moral society (Hamilton 2003). Beyond the cul- 40
tural influence of rice, its cultivation defines the landscape in
many Asian countries. In China, Korea, and Japan, there is a 30
well-established concept of an ideal landscape, consisting of
20
rice fields spread out over a valley floor at the base of a moun-
tain. Furthermore, in Japanese art, there is a strong tradition of 10
depicting rice plants or rice agriculture as part of an idealized
natural environment (Hamilton 2003). Japanese “rice culture” 0
is based on a social infrastructure accumulated over centuries. 1970 1980 1990 2001
Just half a century ago, Japanese agriculture and rural areas Year
were not much different from those in present-day Vietnam or Cambodia Indonesia Japan Brazil
Thailand or southern China (being subsistent to a consider- Bangladesh Philippines India Nigeria
able degree), and, even today, many Japanese farm families Vietnam China Côte dIvoire
eat their own rice, a practice unheard of in the U.S. or Europe.
Fig. 1. Evolution of the share of calories originating from rice, se-
lected countries, 1970-2001. Source: FAOSTAT (September 2004).
Changing demand for rice
Rice is crucial for food security, especially in Asia, where more Table 1. Share (%) of calories originating from rice, by region,
1970-2001.
than 500 million people are undernourished and where rice
provides on average more than 30% of total calories. The im- Region 1970 1980 1990 2001
portance of rice is even starker in countries such as Cambodia,
Vietnam, and Bangladesh, where rice represents more than 70% Asia 38 36 35 31
of the daily calorie intake (Fig. 1). South America 11 11 12 11
Africa 5 7 7 8
Although the average share of total calories originating
United States 1 1 2 2
from rice has remained fairly constant during the last three European Union 1 1 1 2
decades for the world as a whole, in Asia this share has fallen World 20 19 21 21
from 38% to 31%, primarily due to diversification of diets
caused by urbanization and increasing incomes. Within Asia, Source: World Rice Statistics, International Rice Research Institute (Septem-
ber 2004).
it is notable that the share of calories from rice has declined in
China and Japan but this trend is not observed in other Asian
countries; instead, the share of rice has remained fairly stable.
In Africa, however, the share of rice in total food consumption The key factors that affect demand for rice are incomes,
is growing (Table 1). prices, population growth, and urbanization, but they do so in
The world’s biggest consumer of rice is China, followed different ways. As incomes rise, consumers tend to diversify
by India and Indonesia. While Asia has been and continues to their diet, with the share of rice declining and with consump-
be the main consumer of rice, consumption in other regions of tion of wheat initially increasing and of livestock and other
the world is increasing. Outside Asia, Nigeria and Brazil are products later increasing (Barker and Dawe 2002). Further,
the most significant consumers; in fact, Nigeria is a leading once calorie needs have been met, people express a prefer-
rice importer. ence for higher-quality rice as their incomes rise. In Japan,
even though overall consumption of rice has declined in the
past few decades, the type of rice demanded by consumers has
changed from standard-quality rice to high-quality rice.
2In
Outside Asia, patterns of rice consumption vary consid-
some Asian cultures, the traditional understanding goes further: “Since
humans live by eating rice, the human body and soul are regarded as being
erably. In African countries, such as Côte d’Ivoire or Nigeria,
made of rice. Therefore, it is by eating rice that humans are defined” (Hamilton rice consumption is increasing both in quantity as well as share
2003). of total food consumption, partly because of savings in time
Keynotes 9
Table 2. Poverty impact of rice research in India and China.
India China
Table 3. Total rice production, percentage of world production, and percentage increase of production by region, 1970-
2002.
Production (1,000 mt, paddy), 576,280 523,030 19,846 11,880 17,034 3,194
2002
Percentage of world production, 100 91 3 2 3 1
2002
Percentage increase in production, 82 80 90 126 133 79
1970-2002
Source: World Rice Statistics, International Rice Research Institute (September 2004).
expansion of the nonfarm sector and rising labor productivity these recent developments in world rice production. Stocks of
have pushed up nonfarm wage rates, which motivates migra- rice followed patterns similar to those of production, increas-
tion of labor—particularly of the young generation and chil- ing during the 1990s and declining more recently. Dawe (2002)
dren—from rural areas to cities and from farm to nonfarm ac- offers three reasons for the plunge in world prices: the signifi-
tivities within rural areas. Since traditional rice farming is a cant reduction in Indonesia’s imports after 1998, the devalua-
highly labor-intensive activity, increases in wages have pushed tion of the Thai baht during the Asian financial crisis, and the
up the cost of rice production and reduced the incomes and increase in rice production in Bangladesh, Vietnam, Pakistan,
profits of farmers (Hossain and Narciso 2004). In response, and India.
mechanization of harvesting and threshing has occurred in
Malaysia, Thailand, and, in recent years, China. Direct seed- Productivity and sustainability issues
ing has also replaced transplanting to further save on the labor around rice ecosystems
costs of crop establishment. These labor-saving techniques help Rice ecosystems vary widely across countries and regions. In
to restore profitability in the face of rising wages. Europe, the United States, and Australia, rice is cultivated solely
under irrigation. In Africa, on the other hand, the majority of
rice is cultivated upland, with less than 20% of rice cultivated
Rice production and technology under irrigation. A similar situation is found in Latin America.
Developing countries account for 95% of global rice produc- In Asia, more than 50% of rice is irrigated, with the majority
tion. Asia alone produces 90% of world rice, with China and of the remaining 50% being cultivated in rainfed lowlands (Fig.
India accounting for over half of the world’s output. World- 3).
wide, rice production has increased by more than 80% in the Successful agricultural development requires the diver-
last three decades (Table 3). The most striking increase is ob- sification of agriculture away from staple crops, for which
served in Africa, where rice production increased by 133%, demand gradually declines. In the case of Asia, however, di-
albeit from very low levels, reflecting the growing importance versification away from rice production is difficult because
of rice in this region. Other nontraditional rice-producing ar- the surface irrigation systems have been designed to provide
eas, including South, North, and Central America, have in- an adequate water supply for rice but not for other crops.
creased rice production at a higher rate than traditional rice Moreover, because of the monsoon climate, other crops are
producers. not easily adopted.
World rice production increased steadily during the The transformation to irrigated rice production has var-
1990s and started to decline only in the past four years (Fig. ied considerably in Asian countries during the past three de-
2). The decline in rice prices may be a main factor behind cades. Countries such as Pakistan and Japan have always cul-
% %
80 80
60
60
40
40
20
20
0
Latin America
Australia
Africa
Europe
World
USA
Asia
0
1970 1980 1990 2001
Year
Fig. 4. Evolution of the share of rice in the irri-
Region gated ecosystem, selected countries, 1970-
Fig. 3. Distribution of rice ecosystems, by regions and selected 2001. Source: World Rice Statistics, IRRI (Sep-
countries, 2001. Source: World Rice Statistics, IRRI (September tember 2004).
2004).
tivated rice under the irrigated ecosystem and continue to do tions (David and Otsuka 1994). In addition, the main techno-
so, largely because of the limited rainfall in these countries. logical breakthroughs in rice cultivation, which occurred as
Another group of countries that includes the Philippines, In- part of the Green Revolution, were in the cultivation of irri-
dia, and Bangladesh has experienced a steady increase in the gated rice; as a consequence, the yield of rice cultivated under
use of irrigation for rice cultivation in the last three decades. this ecosystem is now much higher than yields that can be
Most surprising is the case of Thailand, however, where the achieved under other conditions.
proportion of rice cultivated under irrigation declined from Concerns regarding the adverse environmental effects
the 1970s to the 1990s to a low of less than 10% before turn- of irrigation and flood control projects on waterlogging, salin-
ing around, although it has still not returned to the level of the ity, fish production, and the quality of groundwater have been
1970s (Fig. 4). growing (Hossain and Narciso 2004). Water is becoming in-
A move toward irrigated rice cultivation is desirable not creasingly scarce and there is growing competition for it be-
only because it increases rice yields but also because it de- tween agricultural and nonagricultural uses. The water-use ef-
creases the vulnerability of rice cultivation to weather condi- ficiency of irrigated rice is low. For example, rice production
Keynotes 11
requires about twice as much water as other crops such as maize Table 4. Rice yields in different countries and ecosystems.
and wheat. In Asia, irrigated rice consumes 150 billion m3 of
Country Ecosystem Year Rice yield
water, which corresponds to a water-use efficiency of approxi- (t ha-1)
mately 20,000 m3 ha–1. Assuming an average yield of 5 t ha–1,
the water productivity of irrigated rice is only 0.15 kg of milled Bangladesh All 2000 3.6
rice per m3 of water. This level of productivity is low com- Burkina Faso Rainfed 1987-90 2.5
pared with that of other crops and makes rice noncompetitive India Irrigated 1995-96 5.2
Rainfed 1995-97 2.3
compared to other uses of water in the face of growing water United States Irrigated 2001 7.0
scarcity (Sohl 2002). Japan Irrigated 1999 6.4
To improve the productivity of water use, it is important Philippines All 1999-2000 3.1
to identify sources of water loss and minimize them. For ex- Thailand Irrigated 2000 4.2
ample, reducing seepage and percolation will increase water Rainfed 2000 2.2
South Korea Irrigated 1999 6.6
efficiency on farms.3 Given the growing constraints to water Vietnam Irrigated 2000 4.2
use for rice, the future of rice production relies on the devel-
opment and adoption of water-saving technologies, a research Source: Hossain and Narciso (2004).
area currently being pursued at the International Rice Research
Institute (IRRI).4
Beyond the negative externalities associated with ineffi- Yield patterns and changes
cient use of water in rice cultivation, population increases are Rice yields vary enormously across ecosystems and countries.
putting increasing pressure on the land to be more productive. Yields of 4–6 t ha–1 are common in irrigated settings, as are
In marginal areas, intensification of land use may lead to deg- yields of 2–3 t ha–1 in rainfed ecologies (Table 4). The poten-
radation of resources through loss of biodiversity, deforesta- tial to increase yields in rainfed ecosystems is still high. This
tion, pest infestations, depletion of natural soil fertility, and ecosystem is dominant in the low-income countries of South
soil erosion. Furthermore, the use of pesticides and fertilizers Asia, Southeast Asia, West Africa, and Central America. Rice
to increase the output of a given piece of land will likely result farming in rainfed ecosystems is subject to natural hazards such
in degradation, environmental pollution, and adverse effects as droughts, floods, and typhoons. Where rainfall is unreliable
on human health (Cantrell and Hettel 2004). The lands that and drainage is poor, farmers still grow traditional varieties
are most threatened in Asia are the fragile rainfed and upland and use fertilizers in suboptimal amounts because of the un-
environments where the poor are forced to use whatever re- certainty of obtaining adequate returns from investment in in-
sources are available to produce the food they need. puts. This is one of the main reasons for the low yield and
Further, rice ecologies interact with regional and global large yield gap in countries with predominantly rainfed eco-
ecosystems. A case in point is the interaction among climate systems (Hossain and Narciso 2004).
change, land use, and production systems. Some lowland rice The scope for further conversion of rainfed ecosystems
systems in coastal zones and river deltas are potentially threat- into irrigated ecosystems is becoming increasingly limited. In
ened by rising sea levels caused by global warming. At the addition, the cost of introducing irrigation to new areas has
same time, irrigated rice contributes to greenhouse gases. These increased substantially, partly because areas that are amenable
complex interactions, as well as the appropriate responses in to irrigation tend already to have irrigation systems in place,
terms of ecological, technological, and policy measures, need and partly because of drastic declines in investment in the de-
further study. velopment and maintenance of large-scale irrigation projects
in many Asian countries (Hossain and Narciso 2002). More-
over, energy costs for water lifting are rising.
Scientific breakthroughs have enabled less-favored lands
to improve their productivity. For instance, research at IRRI
3Ifthis water is recovered for crop consumption at some point downstream, has produced rice varieties that are more drought-tolerant, and
then there is no loss in the irrigation system and it does not affect water effi- the Chinese agricultural research system has produced hybrid
ciency at the basin level (Guerra et al 1998). rice strains with higher yields. Another good example is the
4Guerra et al (1998) suggest four strategies to increase farm irrigation pro- development of the “New Rice for Africa” (NERICA) variety,
ductivity: (1) increasing production per unit of evapotranspiration, which in- a cross of African and Asian rice cultivars, by the West Africa
volves developing new varieties with higher yields and better fertilizer and
weed management; (2) reducing water use in land preparation; (3) adopting a
Rice Development Association (WARDA). NERICA promises
water-efficient method of rice establishment, mainly by moving from trans- more tolerance of most African stresses, including weeds and
planting to direct seeding to reduce irrigation inflow during land preparation; drought, shorter growing cycles, and higher yield potential.
and (4) reducing surface runoff, that is, reducing seepage and percolation This technological development provides an opportunity for
during the crop growth period. The authors also suggest some strategies at the farmers to stabilize and intensify low-input upland farming
irrigation system level; in brief, they include changing the crop and irrigation
schedule to use rainfall more effectively, improving water distribution strate-
systems. Faster adoption of these varieties may lead to a sub-
gies and rehabilitation, and modernizing irrigation infrastructure, among oth- stantial increase in production in Nigeria, Guinea, Côte
ers. d’Ivoire, Sierra Leone, Liberia, and Uganda, where upland and
Keynotes 13
US$ t1
400
350
300
250
200
150
White broken rice, Thai A1
100 Super, f.o.b. Bangkok
White rice, Thai 100% B
50 second grade, f.o.b. Bangkok
0
1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003
Year
Fig. 5. Evolution of Thailand’s (main rice exporter) rice price, 1983-2004. Source: FAO Com-
modities and Trade Division (September 2004), World Rice Statistics, IRRI (2004).
functions, to conduct public-service activities and to conduct buy and transport rice from surplus areas to deficit areas, un-
business as a commercial organization. Under the former func- der the Public Distribution System, to ensure food for the poor
tion, BULOG still distributes “Rice for the Poor” (Raskin) and and to stabilize prices (Nigam 2004). Nevertheless, although
maintains national stock reserves to counter food shortages the private sector negotiates and handles the export of rice,
during emergencies. As a commercial organization, BULOG the government still makes key decisions that determine the
can enter into trading activities as a profit-making organiza- quantity of rice exported (Dawe 2004).
tion.
The role of governments in the international rice trade Developments in and determinants
has also diminished, as has the influence of large trading com- of the price of rice
panies. Today, numerous smaller trading companies are en- The world price for rice has shown a declining trend over the
gaged in the international trade of rice, which has increased past 50 years. After adjusting for inflation, world rice prices
competition and eroded trading margins. However, the degree today are 77% lower than the average from 1950 to 1981. The
and type of policy changes that have occurred in recent years main reason for the decline in prices is the Green Revolution,
vary greatly from country to country. We consider two ex- which led to an increase in yields and a decrease in unit pro-
amples: Vietnam and India. duction costs, and consequently to an increase in worldwide
Vietnam’s rice market is characterized by a high level of supply (Dawe 2004). World prices for other grains have also
commercialization and a large role of the private sector in the declined in real terms, but to a lesser extent than has occurred
production and marketing stages. The private sector plays a for rice. Rice prices continue to fluctuate substantially, as can
key role in procuring rice from farmers, supplying rice to ex- be seen in the plot of rice prices for Thailand over the period
porters, and distributing rice within the country. At the same 1983-2004 (Fig. 5).
time, however, the private sector is highly underdeveloped and The price of rice fluctuates significantly and is sensitive
regionally imbalanced (Golleti and Minot 1997). Until recently, to variations in supply and demand. The main cause for this is
all rice exports from Vietnam were conducted by SOEs, but the small amount of rice traded internationally; less than 7%
over the last few years private firms have gradually been al- of the world’s total rice production is traded. Furthermore, six
lowed to export rice. Although SOEs buy from farmers, the countries (Thailand, Vietnam, China, the United States, Paki-
importance of these transactions is not sufficient for SOEs to stan, and India) supply 85% of the rice that is traded. This
stabilize producer prices, which is considered one of the main concentration in world rice markets implies that changes in
functions of these organizations. Thus, although SOEs still play production or consumption in major rice-trading countries have
an important role in Vietnam’s international trade, gradual lib- a strong effect on world prices. In recent years, there has been
eralization is occurring (Tolentino and Bruce 2003, Dawe a diversification of world importers and exporters that could
2004). further stabilize rice prices. Before the Green Revolution, Asia
In India, the government still intervenes in trade trans- (excluding the Middle East) accounted for 64% of world rice
actions. The Food Corporation of India is one of the largest imports, and an identical share of world exports. Today, a more
organizations procuring food grains on behalf of the govern- diversified group of countries accounts for rice imports.
ment of India. One of the main roles of this organization is to
Keynotes 15
Table 6. Producer support estimates for India, Indonesia, and Vietnam.
million tons in 1995-96, making India the second-largest ex- Food Policy Research Institute (IFPRI). The key policy as-
porter of rice in that year. Since then, India has become a ma- sumptions for each scenario are briefly described below.7
jor supplier of common and basmati rice. Vietnam’s PSEs for
rice show that rice farmers have tended to be protected, espe- Progressive policy scenario
cially at the end of the 1980s and after 1996, although from This scenario assumes a new focus on agricultural growth and
1990 to the mid-1990s they were taxed. This trend was further rural development. Developing countries’ public investments
accentuated by a dramatic decline in public stockholding and and government expenditures on agriculture and rural devel-
fertilizer prices. In general, as indicated by the PSEs, the trend opment, appropriately supported by official development as-
has been toward increasing support for rice (Nguyen and Grote sistance (ODA), increase between 2005 and 2015 and stabi-
2004). lize thereafter. Investments in education, social services, and
We now return to one of the initial questions posed above: health increase. The rate of agricultural technology improve-
Why is the integration of rice into international trade so lim- ment is high owing to increased investment in agricultural re-
ited? Of critical importance to this issue is inelastic supply search and development. Irrigation efficiency and water-use
and demand, that is, both supply and demand do not respond efficiency improve in this scenario, and the rate of irrigation
quickly and strongly to price changes. An additional factor is expansion is moderate to high. Furthermore, producer support
the vulnerability of production to occasional disruptions due to farmers in wealthy countries declines substantially, drop-
to climatic phenomena. These characteristics lead to large price ping to half of current levels by 2010, and to half of the 2010
fluctuations. Market interventions are then used to stabilize level by 2020.
prices and policies are applied to transfer income from con-
sumers to producers. Policy failure scenario
This scenario assumes trade and political conflicts, with no
progress on global agricultural trade negotiations and increased
Future scenarios for rice: toward 2050 levels of protectionism worldwide. Today, we cannot exclude
In view of the close linkages among production, markets, and political-economic forces from producing such outcomes. The
consumption of rice, any outlook on rice must take a holistic policy failure scenario assumes decreases in yield growth for
perspective, and the future of rice cannot be assessed without all crops and fish, and decreases in numbers growth for all
due consideration of the broader economic context and devel- livestock. It also shows trade policies that lead to an increase
opments related to other foods. Global and regional rice con- in protection in many countries, contributing to stagnating
sumption and production will be determined by changes in world trade and slow growth in the net imports of developing
consumption behavior and policy actions related to technol- countries. This impasse in agricultural trade liberalization fur-
ogy (research) and trade. An analysis of the effects of these ther contributes to the growing food deficit in developing coun-
variables on rice consumption, production, and market devel- tries. As a result of political conflicts, investments in social
opments can be useful for identifying appropriate policies that services and agricultural R&D are low. Producer support to
serve growth, sustainability, and poverty reduction. Below, we farmers in wealthy countries triples from current levels by 2020
present three policy scenarios using the International Model and remains steady through 2050.
for Policy Analysis of Agricultural Commodities and Trade
(IMPACT)6 developed by Mark Rosegrant of the International Technology and natural resource management
failure scenario
This scenario is characterized by water mismanagement, de-
clining irrigation efficiency, lack of adaptation to climate
change, and pest problems in agriculture. Low agricultural in-
6IMPACT covers 36 countries and accounts for almost all of the world’s com-
modities. The IMPACT model explores the potential implications of policies
in key variables such as prices, demand, yields, production, and net trade.
This model assumes competitive agricultural markets for crops and livestock 7For further details, see von Braun et al (2004); for a detailed model descrip-
(Rosegrant et al 2002). tion, see Rosegrant et al (2002).
470,000
430,000
3,000
390,000
350,000
1997 2015 2030 2050 2,000
Year 1997 2015 2030 2050
Fig. 7. World rice demand projections (1,000 mt, milled rice). Year
Source: IFPRI IMPACT projections (September 2004). Fig. 8. World rice yield projections under different policy sce-
narios. Source: IFPRI IMPACT projections (September 2004).
Keynotes 17
000 t US$ t1
40,000 Progressive policy actions
Progressive policy actions 600 Policy failure
Policy failure Technology and resource
30,000 Technology and resource 500 management failure
management failure
300
20,000
200
10,000 100
0
0 1997 2015 2030 2050
1997 2015 2030 2050
Year
Year
Fig. 10. Sub-Saharan Africa rice production projections under dif- Fig. 11. Projections of rice price under different policy scenarios.
ferent policy scenarios. Source: IFPRI IMPACT projections (Sep- Source: IFPRI IMPACT projections (September 2004).
tember 2004).
Keynotes 19
Hossain M, Narciso J. 2004. Long-term prospects for the global rice Schreinemachers P. 2004. Yield gaps and their determinants. Bonn
economy. Rome (Italy): FAO Rice Conference. (Germany): Center for Development Research. (In mimeo.)
Huang J, Bouis HE. 1996. Structural changes in the demand for food Siamwalla A, Haykin S. 1983. The world rice market: structure, con-
in Asia. 2020 Vision Discussion Paper No. 11. Washington, duct, and performance. IFPRI Research Report No. 39. Wash-
D.C. (USA): International Food Policy Research Institute. ington, D.C. (USA): International Food Policy Research In-
IRRI. World rice statistics, various years. Los Baños (Philippines): stitute.
International Rice Research Institute. Sidik M. 2004. Indonesia rice policy in view of trade liberalization.
Kaosa-ard M, Juliano BO. 1991. Assessing rice quality characteris- Paper presented at the FAO Rice Conference, 12-13 February
tics and prices in selected international markets: rice grain 2004.
marketing and quality issues. Selected papers from the Inter- Slayton T. 1999. The outlook for the rice trade in the new millen-
national Rice Research Conference, 27-31 August 1990, Seoul, nium. World Rice Commerce Conference.
Korea. Sohl M. 2002. Issues and challenges in rice technological develop-
Kerr J, Kolavalli S. 1999. Impact of agricultural research on poverty ment for sustainable food security. Keynote Address at the
alleviation: conceptual framework with illustrations from the 20th Session of the International Rice Commission, Bangkok,
literature. Environmental and Production Technology Divi- Thailand, 23-26 July 2002. Rome (Italy): FAO.
sion Discussion Paper No. 56. Washington, D.C. (USA): In- Thomas M, Orden D. 2004. Agricultural policies in Indonesia: pro-
ternational Food Policy Research Institute. ducer support estimates 1985-2005. Markets, Trade, and In-
Maclean JL, Dawe D, Hardy B, Hettel GP, editors. 2002. Rice alma- stitutions Division Discussion Paper. Washington, D.C.
nac. Third edition. Los Baños (Phillipines), Bouaké (Côte (USA): International Food Policy Research Institute. (Forth-
d’Ivoire), Cali (Colombia), and Rome (Italy): IRRI, WARDA, coming.)
CIAT, and FAO. Tolentino V, Bruce J. 2003. Successful transition to the market
Mullen K, Gulati A, Orden D. 2004a. Agricultural policies in India: economy: overview of the experience in Vietnam. Yangon
producer support estimates 1985-2002. Markets, Trade, and (Myanmar): FAO Seminar on rice sector policy: neighboring
Institutions Division Discussion Paper. Washington, D.C. countries’ experiences.
(USA): International Food Policy Research Institute. (Forth- USDA. 2004. Producer, supply and distribution online database.
coming.) Washington, D.C.: United States Department of Agriculture.
Mullen K, Sun D, Thomas M, Orden D, Gulati A. 2004b. Agricul- www.fas.usda.gov/psd/.
tural policy interventions in developing countries: mapping von Braun J, Rosegrant MW, Cohen MJ, Cline S, Brown, Bos MS,
the nature, degree and progress in reforms. Selected paper Pandya-Lorch R. 2004. New risks and opportunities for food
prepared for presentation at the American Agricultural Eco- security. 2020 Discussion Paper. Washington, D.C. (USA):
nomics Association Annual Meeting, Denver, Colorado, 1-4 International Food Policy Research Institute.
August 2004.
Nguyen NV, Grote U. 2004. Agricultural policies in Vietnam: pro-
ducer support estimates 1986-2002. Discussion Paper. Wash- Notes
ington, D.C., (USA) and Bonn (Germany): IFPRI and Center Authors’ address: International Food Policy Research Institute
for Development Research (ZEF). (Forthcoming.) (IFPRI), e-mail: j.vonbraun@cgiar.org.
Nigam S. 2004. Rice: challenges in production and marketing in
India. Paper presented at the FAO Rice Conference, 12-13
February 2004.
Pingali PL, Hossain M, Gerpacio RV. 1997. Asian rice bowls: the
returning crisis? New York and Manila: CAB International
and International Rice Research Institute.
Rosegrant M, Meijer S, Cline S. 2002. International Model for Policy
Analysis of Agricultural Commodities and Trade (IMPACT):
model description. Washington, D.C. (USA): International
Food Policy Research Institute.
The key question that underlies efforts to increase rice yields, Declining population growth rates, annual growth rate (%)
to improve the crop’s nutritional quality, and to cut its 2.5
postharvest losses is how much more of the grain will we need
to assure global adequacy of its supply? I will look just a gen- Asia
2.0
eration (25 years) ahead, but, even then, a precise answer is
impossible as well as unnecessary: what we need is a good World
approximation of where we are headed. Key determinants of 1.5
future rice consumption are the rates of population growth,
the process of population aging, the increased mechanization
1.0
of labor, the need to improve inadequate diets, and the extent
and pace of dietary transitions. The complex interplay of these
factors will determine potential demand, the total that should 0.5
give us a better appreciation of how high we should aim our
efforts at future yield increases and agronomic improvement.
0
Keynotes 21
lation older than 60 years will increase by roughly 15% to Annual rice consumption (kg capita1)
about two-thirds, or nearly 600 million people, by 2025. Com- 200
plexities of human metabolism preclude any accurate calcula-
tion of this aging effect (Smil 2000), but, assuming that, on
average, people above age 60 consume 15–20% less food en- Myanmar
150
ergy per day per capita than the mean for the adult population, Vietnam
then aging would reduce 2025 demand by no more than 2–3% Indonesia
compared to a population with a stationary age structure. Bangladesh
The Green Revolution based on the development tial environment for high yield, so that improved bred variet-
of irrigated agriculture ies of crops can be fully used. From 1961 to 2002, global irri-
gated agricultural land roughly doubled from 139 million to
Wisdom developed from soil and water has continuously been 277 million ha, while total land for arable and permanent crops
the fundamental element of all human activities since ancient expanded slightly from 1,357 million to 1,534 million ha (Fig.
civilization. The tragedies and ruins of the Mesopotamian/ 1). Global population and cereal production have also doubled
Sumerian civilization remind us of the significance of build- from 3.08 billion to 6.23 billion and from 877 million metric
ing sustainable agriculture and society. The drama of the col- tons to 2.03 billion metric tons. Irrigated land, which accounts
lapse of Mesopotamia started with a gradual rising of the sa- for about 18% of agricultural land area, produces about 40%
line groundwater level. The collapse advanced rapidly when of the food for the global population, contributing consider-
this level exceeded a threshold. It is well known that this prob- ably to the alleviation of global poverty and starvation.
lem was caused by faulty irrigation. Can we be confident that Sound and sustainable irrigated agriculture is indispens-
the current threshold is high enough when we look at the glo- able for humankind to survive in the future. Now, about 70%,
bal expansion of commercialized agriculture and hazardously or 2,504 km3, of the world’s annual freshwater usage of 3,572
exploited water resources, such as in the Aral Sea basin, the km3 is for agriculture, and, of this, about 70% is used mainly
Ogallala aquifer (Nebraska) and Central Valley (California) in for rice paddy agriculture in Asia.
the United States, Punjab and Haryana in India, and Northern Our generation is primarily responsible for assuring sus-
China? tainable and efficient agriculture through wiser governance and
A stable water supply was a strong driving force behind management of soil and water resources.
the Green Revolution. Irrigated agriculture provides an essen-
Keynotes 23
Index (1961 = 100)
250
240 Total cereal production (1961 = 100)
Population (1961 = 100
230
Irrigation area (1961 = 100)
220 Land for arable and permanent crops (1961 = 100)
210 Rainfed area (1961 = 100)
200
190
180
170
160
150
140
130
120
110
90
1973
1987
1965
1961
1963
1967
1969
1971
1975
1977
1979
1981
1983
1985
1989
1991
1993
1995
1997
1999
2001
2002
Year
World water issues and conversion of policies ments of developing countries faced financial pressure for the
for irrigation—from expansion to increasing efficiency operation and maintenance of overaged water facilities.
of existing systems In the 21st century, the century of water and life, another
option has been recognized as a better solution. That is to in-
Fresh water existing on the global land surface in readily us- crease the efficiency of the use of water in existing systems. In
able forms such as lakes, swamps, and rivers accounts for only many countries, attempts to increase this efficiency have been
about 0.0075% (104,620 km3) of all the water existing on our made through (1) renovating irrigation water facilities such as
planet. This percentage is equivalent to the ratio of two tea- lining canals with concrete, (2) introducing/reinforcing volu-
spoonfuls of water (15 cm3) in a typical household bathtub metric water pricing, and (3) introducing participatory irriga-
full of water. Of the annual global rainfall on land, which sup- tion management (PIM).
ports the hydrologic cycle of fresh water, only about 40%
(45,000 km3) becomes potential water resources after exclud-
Efficiency of agricultural water use in different regional
ing evaporation. Humankind has to share this available water
conditions—arid and humid
with other water uses such as industry, domestic use, and
biodiversity while producing food for the more than 6 billion Improving water-use efficiency in agriculture is a key issue
people living on the planet. during international water discussions nowadays. Many ex-
In the 20th century, the century of fire and machines, we perts have reported case studies and mentioned success sto-
strived to develop water resources mostly through construc- ries. These reports are helpful for improving water-use effi-
tion technologies such as large reservoirs, until the early 1980s. ciency in regions where it is reasonable for farmers to con-
Not only drastic growth in the world human population but stantly use a minimum amount of irrigation water to secure
also worldwide trends of economic growth and expansion of good crop growth. The concept of water-use efficiency in this
cities, especially in developing countries, resulted in sharp in- definition is typically applicable to agriculture in arid/semi-
creases in demand for domestic and industrial water use, and arid regions. This concept comes from the idea that all water
put strong and continuous pressure on the increase in water should be consumed in crop fields in the form of evapotrans-
supplies. Water resource development by construction of “hard- piration, allowing no water to be lost elsewhere.
ware” was promoted in many regions around the world. However, in humid regions blessed with abundant pre-
However, after the 1980s, this type of development met cipitation, the shortcomings of this concept have now come to
with certain limitations. Although the most effective way to be widely recognized. These shortcomings are mainly caused
increase water resources, speaking from an engineering view- by neglect of the following factors accompanying rice paddy
point, was to build large reservoirs, appropriate construction farming in these regions: (1) the highly substitutable charac-
sites for new reservoirs became limited. Moreover, govern- teristic of water usage and labor investment, (2) the dynami-
cally fluctuating competitiveness among water users over the ture, duck raising, washing, cleaning, bathing, cooling, gar-
short term, and (3) the enormous value of socioeconomic ele- dening, and fire fighting; (2) multifaceted socioeconomic ben-
ments other than food/fiber production; this value is defined efits to the public such as protecting the aqua-ecosystem, en-
as an “externality.” Factor 1 allows for reduced labor invest- hancing the water-related environment, forming landscape,
ment costs where farmers use much more than the minimum recharging the groundwater aquifer, and stabilizing downstream
amount of water required for meeting crop water requirements. river flow by return flow; and (3) various other nonuse-values
Factor 1 also allows for a drastic reduction in water use through in cultural-religious activities (Fig. 2).
farmers’ collective efforts during severe dry spells that occur The socioeconomic situation of countries in the Asian
unexpectedly under the conditions of factor 2. monsoon region is diversified, including tropical developing
A considerable part of the water irrigated into rice paddy and temperate developed countries. However, predominant rice
fields is not consumed but drained into the basin downstream paddy agriculture and its enormous external value are com-
during the rainy season from vast paddy fields in the Asian mon to them. Note that the external values are not only impor-
monsoon region. This is because (1) the value of water is quite tant for farmers and the economy in developing countries but
low because of abundant precipitation and low competitive- are also very valuable for citizens of developed countries in
ness among water users, and (2) immersion cultivation with a providing multifaceted socioeconomic benefits.
larger amount of water than the equivalent to evapotranspira- Water productivity, often advocated by the catch phrase
tion reduces labor investment cost. “more crop per drop,” is a newly conceptualized term as an
This behavior seemingly decreases water-use efficiency, indicator of water-use efficiency in rainfed agriculture, using
but, on the contrary, the affluent water use provides diverse about 16,000 km3 of rainwater annually, as well as in irrigated
and enormous value for socioeconomic externalities. The agriculture. This concept, however, unjustly underestimates the
drained return flow also helps to preserve nature, resulting in water productivity of rice paddy agriculture in the Asian mon-
any negative externality counting for nothing. Note that, in soon region. Therefore, it must be reevaluated after incorpo-
rice paddy agriculture in humid regions, the irrigation water rating the high value of socioeconomic externalities.
automatically serves as an externality, whereas, in upland fields
and relatively arid regions, the water has to be set aside and
preserved, especially for externalities.
Traditional participatory irrigation management
and governance (PIM/G) in rice paddy systems
Recent challenges to establish PIM in Asian monsoon coun-
Socioeconomic externalities in rice paddy agriculture tries show difficulties in transferring the operation and man-
in humid regions agement (O&M) of irrigation from government agencies to
The enormous external value generated by irrigated rice paddy farmers. Lessons learned from these challenges include how
agriculture can be grouped in the following categories: (1) to establish effective incentives for farmers. Initially, strong
multiple use of water by farmers and residents for aquacul- economic incentives superior to disincentives such as labor
Keynotes 25
contributions of O&M and water charge are necessary to launch composting, and joint ventures for purchasing and shipping.
modernized operations. Afterward, the diverse values of so- In Japan, agriculture has lost its substantial position in
cioeconomic externalities for farmers become driving forces the country’s economy. However, farmer group activities
to improve their performance. The diverse values are realized through PIM/G create new values for socioeconomic exter-
step by step through various group activities. Well-sustained nalities as a safety net against natural disasters and as a gen-
traditional PIM shows many examples of built-in incentives erator of resources for other sectors. At the WRRC, I can men-
related to externalities as well as agricultural production. tion examples of providing water from agriculture to domestic
When we look around the world, we find many current use during severe dry spells, increasing the performance of
irrigation systems that have been fulfilling both requisites of paddy fields while protecting reservoirs during extreme floods,
food production and externalities, such as the Muang Fai in creating winter sanctuaries for migratory birds, and restoring
Thailand, Kanna in Sri Lanka, and Subak in Indonesia. The groundwater levels for downstream water use in cities in win-
Dujiangyan irrigation system in Sichuan, China, was estab- ter. The value of socioeconomic externalities cannot be com-
lished in 250 B.C. and is still working for about 670,000 ha of mercially exchanged, whereas agricultural products are widely
farmland. In Japan, many old irrigation systems have lasted traded in the international market. Irrigation systems and wa-
for centuries, with technology advancing in stages. It has to be ter accompanied by the value of externalities are social over-
noted that these systems always consist of weirs, canals, and head capital and commons. Therefore, development programs
ponds operated and managed by indigenous farmers’ organi- for irrigated agriculture should take into account external value
zations, namely, Land Improvement Districts (LIDs). Most of as a benefit vis-à-vis the cost of the investment.
them are historically as old as their weirs and canals. The LIDs I believe that improving water-use efficiency by fully
are responsible for O&M of nationwide canals extending a taking into account the value of socioeconomic externalities
total length of about 400,000 km (equivalent to ten times around can create a future with more sustainable agriculture through a
the globe). rice-based system. Sustainability of this agriculture will be
We found it very interesting that irrigation systems that strongly supported by interactions between human activities
achieve a good balance between externalities and food pro- through PIM/G and broad-banded products, including various
duction are observed mostly in old systems. In presenting this economic values, the living environment, and cultural-religious
paper at the WRRC, I wanted to mention representative canal activities, as well as the agricultural product rice.
systems that have completed the difficult task of satisfying both The proposition I have maintained in this paper might
requisites. be a claim for the renaissance of old systems because water
management in the 21st century requires a holistic approach
in harmony with various factors, including externalities, as
Renaissance of efficient and sustainable rice production some old irrigation systems have achieved. These old systems
with socioeconomic externalities generated by PIM/G make a good guide for future sustainable development.
Group activities through PIM/G can generate much more value I especially wish to acknowledge the contributions of
for externalities than individually disconnected activities. A Mr. Kazumi Yamaoka of the National Institute for Rural Engi-
good example is shown by a full cultural and religious perfor- neering in co-authoring this document and preparing the pre-
mance in Bali, Indonesia. Such a group activity works as a sentation materials.
platform for consultation in a local community and sometimes
develops new activities such as mutual farming aid, group
Thank you for this opportunity to speak Why rice research must continue
to you as one of the keynoters of this
week’s World Rice Research Confer- Before we go into the details of what we think should be the
ence, which—as the culminating sci- research strategy for rice in the 21st century, I would like to
entific event of the International Year discuss, briefly, why indeed rice research for developing coun-
of Rice 2004—has brought to Tsukuba tries must continue and be reinvigorated.
the planet’s leading rice scientists to ex- Since the dawn of the Green Revolution—which began
change the latest research information in Asia with IRRI’s release in 1966 of IR8, the first modern,
on key rice-related issues. high-yielding semidwarf rice variety—the global rice harvest
has more than doubled, racing slightly ahead of population
growth. This increased production and the resulting lower
Keynotes 27
serve are starved for information. They need new management
skills and technical know-how as soon as possible.
To help close this knowledge gap by getting technology
off the shelf and into farmers’ hands, scientists and informa-
tion and communication technology (ICT) specialists at IRRI
have placed a Rice Knowledge Bank on the Web
(www.knowledgebank.irri.org) aimed at extension workers and
farmers. At the speed of a keystroke, it can provide the latest
information—from diagnosing field problems via RiceDoctor
(www.knowledgebank.irri.org/riceDoctor_MX/default.htm) to
making crop management decisions via TropRice
(www.knowledgebank.irri.org/troprice/default.htm). Usually,
only more prosperous farmers are currently going online. But,
sooner than some might expect, even the poorest farmers will
be going online to access field problem and crop management
information near where they toil in their rice fields. For ex-
ample, in the Philippines, the Open Academy of Philippine
Agriculture predicts that within 2 years it will be common-
place for Filipino farmers—through strategically located
Internet kiosks in farming villages—to be diagnosing crop pests
by accessing the Web and sharing information with peers any-
where in the country using electronic mail and cell phone tech-
nology (Marquez 2004).
In addition, continued on-farm research will be needed
to fine-tune new technologies if the yield gap is to be nar-
rowed. For example, IRRI has
a long tradition of listening to
farmers (Hettel and Dedolph
1996) and planning farmer
participatory research (FPR).
FPR is an approach that in-
volves encouraging farmers to
using the best management techniques could lead to yield im- engage in experiments in their
provements of 18% in the coming decades if the negative ef- own fields so that they can
fects of climate change can be avoided (Sheehy et al 2005). learn, adopt new technologies,
The incidence of weather extremes appears to be increasing and spread them to other farm-
and there is a need to breed new rice cultivars that can tolerate ers. With scientists acting as
the worst aspects of climate change. facilitators, farmers and scien-
Recent IRRI research has brought both good and bad tists work closely together
news. The good news is: It appears that improved crop and from initial design of the re-
nitrogen management has effectively closed the yield gap be- search project to data gather-
tween simulated potential yield and actual yield of existing ing, analysis, final conclu-
elite cultivars in long-term plots at IRRI (Kropff et al 2003). sions, and follow-up actions.
The bad news is: Even when these improved crop and nitro- A great example of this is the ongoing FPR in Bangladesh
gen management practices are employed, there can still be yield where farmers’ livelihoods are being improved by closing rice
fluctuations, which are largely associated with climatic fac- yield gaps caused by weeds. Farmers in Bangladesh grow rice
tors (Dobermann et al 2000). Recent fluctuations in grain yield over a wide range of rainfed and irrigated lands and weeds are
trials at IRRI have been associated with reduced solar radia- a major constraint. In this project of the Irrigated Rice Re-
tion and increased seasonal mean minimum temperature, and search Consortium (www.irri.org/irrc/weeds/closing.asp), re-
not related to crop management (Kropff et al 2003, Peng et al searchers are examining two contrasting farming systems in
2004b). Comilla and Rajshahi districts to learn how different rice-crop-
In many countries, new technologies are not making the ping systems affect weed abundance and methods for their
impact that they should be in closing the yield gap. To make management. This work has shown the benefit of integrated
any major breakthroughs in the immediate future will first re- management involving herbicides in combination with manual
quire the bridging of another gap—the knowledge gap weeding. Application of herbicides or use of a hand-pushed
(Schiffrin 2001). Extension specialists and the farmers they weeder in transplanted rice results in similar yields—and a
Keynotes 29
We scientifically tested several variations of the concept and
improved it. Now we are disseminating our findings with con-
fidence that the practice not only reduces farmers’ reliance on
chemical pesticides, thereby protecting the environment, but
also improves yield and income to give farmers the options
they need to break out of the poverty trap. Word-of-mouth is
already leading to the technique’s wide adoption on nearly 1
million hectares across 10 provinces in China.
Vietnam. Our Vietnam study offers valuable lessons on
how to reduce pesticide applications. We converted our find-
ings into one simple rule: “Don’t spray for the first 40 days.”
We launched a media campaign to deliver the message to farm-
ers, stressing the cost savings and health benefits of reduced
spraying. The result? In the test area of 21,000 households,
after an 18-month interval, we recorded a 53% reduction in the
number of insecticide applications—without affecting yield!
Many farmers reduce input costs by US$30–50 per season— Improved management of N, P, and K fertilizers through SSNM
equal to a month’s income in Vietnam. Eventually, this effort is now increasingly being shown in Asia to reduce disease and
persuaded almost 2 million rice-growing households in the insect damage (Buresh 2004, Jahn 2004, Wang et al 2004)—
Mekong Delta to cut back on using harmful and unnecessary yet another avenue on which farmers can reduce their need for
farm chemicals. Underscoring the significance of this work, pesticides and usher in the Doubly Green Revolution in rice.
the project received the St. Andrews Prize for the Environment When does MR begin? There are some interesting new
from Scotland’s St. Andrews University in 2002 and the Inter- twists on the importance of maintenance research and Dr.
national Green Apple Environment Award from the U.K.-based Dalrymple has been recently looking at the interface between
Green Organization in 2003 (IRRI 2003). MR and what he calls productivity-enhancing research (PER).
He assumes that there is an area of overlap between the two,
and that, in some cases, particularly significant advances in
MR could actually contribute to increasing yield. This is likely
to be true in the case of resistance to plant diseases and insects
(D. Dalrymple, USAID, 2004, pers. comm.). Clearly, the line
between MR and PER is fuzzy and it raises some definitional
questions, such as, When does MR actually begin to kick in?
Otsuka et al (1994) concluded that the Green Revolution would
not have been very revolutionary without the development and
the diffusion of second-generation modern rice varieties with
multiple pest and disease resistance. So, depending on where
MR is placed on the rice-improvement breeding continuum, it
can be argued that, without it, there might not have been a
Green Revolution.
Keynotes 31
Until recently, this amount of water has been taken for granted,
but this cannot continue.
To meet the water crisis head-on for rice, the Interna-
tional Platform for Saving Water in Rice (www.irri.org/ipswar/
about_us/ipswar.htm) has been created as a mechanism to in-
crease the efficiency and enhance the coherence of research
on water savings in rice-based cropping systems in Asia. The
overarching goal is to conserve water resources, which will in
turn safeguard national and household food security and alle-
viate poverty.
About 75% of the world’s rice is produced in irrigated
fields, where farmers can control the amount of water applied.
In the tropics, these systems can produce yields up to 9 t ha–1, the paddy for 3 to 5 days, and, as the water soaks into the soil,
but they typically use from 1,000 to 2,000 mm of water per the surface is then allowed to dry for a few days (usually from
hectare. This means that from 1,000 to more than 3,000 L of 2 to 4) before getting re-flooded.
water are used to produce each kg of rice. In contrast, rainfed Another alternative is dry-field or aerobic rice, where
rice can produce yields of up to 5 t ha–1 with a water input of the rice is sown directly into dry soil, like wheat or maize, and
only 800 mm of well-distributed rainfall. irrigation is applied to keep the soil sufficiently moist for good
Approaches to improving water productivity in both ir- plant growth, but the soil is never saturated. Both of these sys-
rigated and rainfed rice systems involve decreasing the water tems allow for substantial water savings of from 30% to 50%.
requirement of irrigated systems without affecting yield and Our strategic research on water will be looking closely
increasing the yield of rainfed systems without increasing their at both of these systems. For example, we have already estab-
water use. These approaches are complementary, with spillover lished an Aerobic Rice Working Group, involving breeders,
benefits expected across the irrigated and rainfed systems. The physiologists, and water and soil scientists, that is striving to
recent advancement of crop sciences, especially in rice overcome the many difficulties in taking rice out of its natural
genomics (see below) and linkage of conventional breeding aquatic environment. By developing a completely new man-
with molecular tools, has shown some promising possibilities agement system, the new aerobic rice should be able to yield
for making such innovations a reality. We are already aiming up to 6 t ha–1 using only half the water!
our strategic research to attack the water problem from sev- Human nutrition. And finally, the world’s poor will
eral angles, using modern tools of molecular biology, environ- achieve household food security only when—in addition to
mental monitoring, simulations, and database integration. being available in sufficient quantity—the food is of ample
Two types of water-saving systems can be used to re- quality as well. Although rice supplies adequate energy in the
place the traditional irrigated rice production schemes that are form of calories and is a good source of thiamine, riboflavin,
now under threat. One is alternate wetting and drying (AWD). and niacin, it is lacking as a source of vitamin A and other
In this system, the field is irrigated with enough water to flood critical vitamins, iron, zinc, and other micronutrients and amino
Keynotes 33
rgp.dna.affrc.go.jp/ nization. An interim steering committee for the consortium was
IRGSP/index.html), a formed at the Plant and Animal Genome Conference in San
leading member of which Diego in January 2003, leading to the formal inauguration of
is the Rice Genome Re- the International Rice Functional Genomics Consortium
search Program (http://rgp.dna.affrc.go.jp), based right here (www.iris.irri.org/IRFGC).
in Tsukuba. The IRGSP, a consortium of publicly funded labo- Currently, the IRRI-coordinated consortium is guided by
ratories in both developed and developing countries (Japan, a 21-member steering committee, which includes 17 institu-
USA, China, Taiwan-China, Korea, India, Thailand, France, tions from 10 countries—including Japan—and two Future
Brazil, and the United Kingdom), was established in 1997 to Harvest Centers—CIAT and IRRI. The collective goals are to
obtain a high-quality, map-based sequence of the entire 12- share materials, integrate databases, seek bilateral and multi-
chromosome rice genome using the cultivar Nipponbare of lateral partnerships, implement initiatives for the cooperative
Oryza sativa subsp. japonica. elucidation of gene function, and accelerate delivery of research
Five years ahead of the original schedule, the IRGSP, in results to benefit rice production (Phillips et al 2004). Again,
December 2002, announced the completion of a high-quality this is an effort that no one institution, country, or region could
draft sequence of the genome (Feng et al 2002, Sasaki et al take on in isolation. For example, through the IRFGC, IRRI is
2002). This great achievement could not have been accom- investing its USAID Linkage Program funds to support eight
plished in any one country or region. The rice genome sequence, laboratories in the United States to identify genes for stress
now freely available on the Internet (see http:// tolerance (www.iris.irri.org/irfgc/researchstrength.shtml). This
rgp.dna.affrc.go.jp/cgi-bin/statusdb/irgsp-status.cgi) with data unique consortium embraces and encourages the role of de-
integrated from Monsanto and Syngenta, is already inducing veloping countries in gene discovery and applications of the
innovative research worldwide in the new science of functional new science (Leung and An 2004). IRRI will continue to work
genomics. With 50,000 or more genes predicted in the DNA closely with the international community to pursue initiatives
sequence, the task ahead requires the development of many that will accelerate gene discovery and take advantage of the
genetic resources to unravel the function and interaction of new knowledge to make a true impact.
these genes and their relationship with important traits. To celebrate the completion of the rice genome and learn
about the latest developments in functional as well as struc-
International Rice Functional Genomics Consortium tural and evolutionary genomics research, IRFGC members
Enter the International Rice Functional Genomics Consortium. will gather during the 2nd Annual Symposium on Rice Func-
Parallel to the initiation and activities of the IRGSP, IRRI rec- tional Genomics (www.rfg2004.org) later this month (15-17
ognized the need to bring together diverse expertise to capi- November) in Tucson, Arizona. We’re truly excited about what
talize on the wealth of ge- the consortium will be coming up with in the years to come.
nomic information for func-
tional analysis. Informal Other important rice-related consortia
meetings were held in 1999 Within the CGIAR, IRRI has long taken the lead using the
and 2000 that led to the for- consortium concept to link research capacity of national agri-
mation of an International cultural research and extension systems (NARES) with that of
Rice Functional Genomics IRRI’s to solve important problems through multi-country col-
Working Group. This Work- laboration. We are using consortia to provide ways to conduct
ing Group brought together primarily strategic research by sharing research responsibili-
research groups from around ties according to each partner’s interests and capabilities. The
the world to discuss mutual key to success here is that we have, in consortia, truly peer
interests and promoted col- relationships among NARES and with IRRI for much of the
laboration in anticipation of strategic research we mentioned earlier, with a substantial com-
the completion of the rice mitment of staff and resources. The special feature of collabo-
genome sequence. During ration in these consortia is that IRRI scientists put substantive
this time, members of this re- components of their own strategic research at the consortia
search community developed useful materials to form a base sites with NARES partners.
for collaboration in functional genomics studies in rice involv- Some key consortia we would like to mention briefly
ing the genome sequence itself, gene cloning and arrays, and are the Irrigated Rice Research Consortium (IRRC), the new
mutants (Phillips et al 2004). Consortium for Unfavorable Rice Environments (CURE), and
After two years of discussion, the idea of creating a more the Rice-Wheat Consortium (RWC).
structured consortium was discussed at a Working Group meet- IRRC (www.irri.org/irrc). Some 2.2 billion Asian rice
ing in Beijing in September 2002, and at the November 2002 farmers and consumers depend on the sustainable productiv-
meeting in Canberra, Towards building a global rice gene ity of irrigated rice ecosystems. The IRRC, composed of
machine (www.pi.csiro.au/grgm02/home1.htm), organized by NARES from Bangladesh, Cambodia, China, India, Indone-
the Commonwealth Scientific and Industrial Research Orga- sia, Lao PDR, Malaysia, the Philippines, Thailand, and Viet-
Keynotes 35
agricultural production and rural livelihoods involving—but nance research and strategic research on its agenda. There re-
not necessarily confined to—rice. We are confident that the ally can’t be one without the other in the long term. The dual
new consortium will serve as a platform and forum for identi- challenges are producing new products to solve future prob-
fying and prioritizing the rainfed research needed to generate lems caused by diminishing resources and climate change and
impact in farmers’ fields. maintaining and securing what we have already achieved.
RWC. Bangladesh, India, Nepal, and Pakistan have de-
voted to agriculture nearly half of their total land area of 402
References
million hectares to feed and provide livelihoods for 1.8 billion
people. Rice and wheat, the staple food crops for these people, Buresh RJ. 2004. Specific benefits: farmers earn more from their
contribute more than 80% of the total cereal production in these rice crop by scientifically optimizing fertilizer use. Rice To-
countries. day 3(4):32-33. www.irri.org/publications/today/pdfs/3-4/
Celebrating its 10th anniversary this year, the Rice-Wheat Ricetoday_3-4.pdf.
Cantrell RP, Hettel GP. 2004a. New challenges and technological
Consortium (www.rwc.cgiar.org/rwc), a NARES-driven initia-
opportunities for rice-based production systems for food se-
tive of the four countries, has truly emerged as an innovative
curity and poverty alleviation. In: Proceedings of the FAO
model for regional and international collaboration. Operating Rice Conference, 13-14 Feb. 2004. Rome (Italy): Food and
as an inter-institutional and intercenter multidisciplinary net- Agriculture Organization. www.fao.org/rice2004/en/pdf/
work facilitating system-based farmer participatory research cantrell.pdf.
in the rice-wheat ecology of the Indo-Gangetic Plains, it has Cantrell RP, Hettel GP. 2004b. The doubly green revolution in rice.
successfully addressed the issues of productivity enhancement Presentation at the World Food Prize Symposium: Rice,
of rice and wheat in a sustainable fashion. These issues in- Biofortification, and Enhanced Nutrition, Des Moines, Iowa,
clude tillage and crop establishment, crop diversification, 14-15 October 2004.
postharvest, and environmental benefits. Through the RWC’s Dalrymple DG. 2004. Thoughts on context and reform in agricul-
efforts, tens of thousands of resource-poor farmers in South ture. Workshop on Technology Development in the Life Sci-
ences sponsored by the Program on Science, Technology and
Asia have been able to increase their income by using prac-
Global Development, Center on Globalization and Sustain-
tices that save water, fuel, and other inputs; facilitate timely
able Development, The Earth Institute, Columbia University,
planting; reduce tillage needs and crop residue burning; and 20-21 May. www.earthinstitute.columbia.edu/cgsd/events/
allow them to diversify their cropping systems (Ladha et al documents/dalrymple_000.doc.
2003). Datta K, Baisakh N, Oliva N, Torrizo L, Abrigo E, Tan J, Rai M,
Last year, external reviewers pointed out that “the effec- Rehana S, Al-Babili S, Beyer P, Potrykus I, Datta S. 2003.
tiveness of partnership among IRRI and CIMMYT and their Bioengineered ‘golden’ indica rice cultivars with beta-caro-
NARES partners as well as within and between the four na- tene metabolism in the endosperm with hygromycin and man-
tional systems is one of the most important achievements of nose selection systems. Plant Biotechnol. J. 1(2):81-90.
the RWC”—and truly this is what consortia are all about and www.blackwell-synergy.com/links/doi/10.1046/j.1467-
why they have the potential to achieve so much. 7652.2003.00015.x/abs/
Dobermann A, Dawe D, Roetter RP, Cassman KG. 2000. Reversal
of rice yield decline in a long-term continuous cropping ex-
Summary and conclusions periment. Agron. J. 92:633-643.
FAO (Food and Agriculture Organization). 1999. Regional expert
We are happy to have been able to keynote this scientific meet- consultation on bridging the rice yield gap in the Asia and
ing of the WRRC—this culminating conference of the Inter- Pacific Region. FAO Regional Office for Asia and the Pa-
national Year of Rice—to suggest what the strategy should be cific, Bangkok, Thailand, 5-7 October 1999. www.fao.org/ag/
for rice research in the 21st century. As we have just outlined, AGP/AGPC/doc/field/commrice/session/54.htm.
we feel that strategy must include Feng Q, Zhang Y, Hao P, et al. 2002. Sequence and analysis of rice
O Renewing and invigorating efforts to close the yield chromosome 4. Nature 420:316-320.
gap in rice; Hettel GP, Dedolph C, editors. 1996. Listening to the farmers. Los
O Remembering the critical importance of maintenance
Baños (Philippines): International Rice Research Institute. 97
p. www.irri.org/publications/annual/ar1996.asp.
research;
IRRI. 2003. Innovative response to pesticide misuse. IRRI press re-
O Using cutting-edge tools to enhance strategic thrusts
lease. www.irri.org/media/press/press.asp?id=79.
to (1) break the yield barrier, (2) face head-on the Jahn GC. 2004. Effect of soil nutrients on the growth, survival and
crises of impending global warming and water short- fecundity of insect pests of rice: an overview and a theory of
ages, and (3) improve human nutrition; and pest outbreaks with consideration of research approaches.
O Employing the consortium model whenever and wher- Multitrophic interactions in soil and integrated control. Inter-
ever possible for researchers to work together to ac- national Organization for Biological Control Bulletin 27
complish the important goals we just discussed. (1):115-122.
If you remember anything of what we’ve discussed to- Kropff MJ, Cassman KG, Peng S, van Laar HH. 2003. Yields at
day, we hope it is that, for any agricultural research institution IRRI research farm are still close to the climatic potential level.
to be successful, it must have a harmonious mix of mainte- Int. Rice Res. Notes 28(2):19-21. www.irri.org/publications/
irrn/pdfs/vol28no2/IRRNmini3.pdf.
Keynotes 37
SESSION 1
The genus Oryza, its diversity,
and its evolution
CONVENER: R. Okuno (NIAS)
CO-CONVENER: R. Sackville Hamilton (IRRI)
Molecular phylogeny and divergence of the rice tribe
Oryzeae, with special reference to the origin
of the genus Oryza
Song Ge, Ya-long Guo, and Qi-hui Zhu
The rice tribe Oryzeae consists of 12 genera and more than 70 mum parsimony (MP), neighbor-joining (NJ), and Bayesian
species distributed in tropical and temperate regions world- (BI) approaches. A comparison of the phylogenies inferred
wide (Tzvelev 1989, Vaughan 1994). Species in the genus from two nuclear genes shows that they are largely congruent,
Oryza and other genera closely related to Oryza have been with only one area of disagreement involving the placement of
extensively studied either because of their agronomically use- the monotypic genus Hygroryza (see below). The partition-
ful traits in rice genetic improvement (wild species in Oryza homogeneity test (PHT) indicated that two data sets (Adh2
and Porteresia) or because of their economic value as part of and GPA1) were statistically incongruent (P<0.01) when all
cuisine (Zizania) and forage (Leersia) (Vaughan and Morishima the species were included, whereas the topological incongru-
2003). However, phylogenetic relationships among genera in ence was not significant (P = 0.12) when the monotypic genus
this tribe have not been well studied, and the circumscription Hygroryza was excluded. Therefore, phylogenetic analysis was
and taxonomic position of some genera have remained con- conducted based on the combined data sets, excluding
troversial for decades. In addition, the origin and diversifica- Hygroryza, and the strict consensus tree using the MP method
tion of this tribe, in particular the origin of the genus Oryza is presented in Figure 1. Essentially, the NJ and BI methods
and its divergence, remain largely unclear. Recently, sequences generate topologies almost identical to that of MP, except for
of low-copy nuclear genes have been successfully used for slightly different statistical support for a few clades. As Figure
addressing phylogenetic questions and, in combination with 1 shows, the species of the rice tribe Oryzeae form a strongly
chloroplast DNA fragments, have provided especially power- supported monophyly and the tribe diverged early into two
ful markers in terms of phylogenetic reconstruction and bio- major clades with high bootstrap support (over 95%). These
geographic inference (Sang 2002). two clades correspond to the traditionally recognized subtribes
The alcohol dehydrogenase (Adh) gene is the most widely Oryzinae and Zizaniinae (Pyrah 1969). One clade includes three
used low-copy nuclear gene, whereas the chloroplast matK genera, Leersia, Oryza, and Porteresia, with Leersia as the
gene, trnL intron, and trnL-trnF spacer, as well as mitochon- basal lineage. It is obvious that Porteresia coarctata is nested
drial nad1 intron, are widely used in phylogenetic studies. into Oryza species, suggesting that it should be treated as a
Nuclear GPA1 that encodes a G protein α subunit is also a member of the genus Oryza rather than as a separate mono-
source of choice because it is a single copy in higher plants typic genus (Ge et al 2002). The other clade consists of three
and well characterized in function and structure. To better un- highly supported monophyletic groups involving the remain-
derstand the evolutionary and biogeographic history of the rice ing genera, except Hygroryza. The first group includes east-
tribe, we sequenced two chloroplast (matK and trnL-trnF) frag- ern Asian Chikusichloa, Australian monotypic Potamophila,
ments and one mitochondrial (nad1 intron) fragment and por- and African Prosphytochloa, whereas the second group con-
tions of two nuclear genes (Adh2 and GPA1) from 35 species tains the American Zizaniopsis and Luziola, which is a sister
representing 12 genera in the tribe, as well as Ehrharta and to the group including the American monotypic Rhynchoryza
Phyllostachys (subfamily Bambusoideae) as the outgroups. In and the eastern Asian–North American disjunct Zizania (Fig.
addition, to clarify the phylogenetic relationship among the A- 1). The combined data sets of matK, trnL-trnF, and nad1 pro-
genome species, which include cultivated rice (O. sativa), we vide similar topology but the resolution is slightly lower (not
have chosen to sequence introns of three anonymous nuclear shown).
single-copy genes (OsRFCD001283, OsRFCD017357, and It is noteworthy that Hygroryza aristata forms a clade
OsRFCD009971 on chromosomes 1, 4, and 2, respectively) with the Leersia/Oryza group on the Adh2 tree but clusters
that evolve much faster than commonly used ITS and cpDNA with the Rhynchoryza/Zizania + Luziola/Zizaniopsis group on
fragments. Based on sequences of these introns, we recon- the GPA1 tree (Fig. 1). The combined matK/trnL-F/nad1 phy-
structed the phylogeny and dated the origin and divergence of logeny presents a topology similar to that of GPA1 and the
the A-genome species. previous matK phylogeny (Ge et al 2002). However, because
this monotypic genus is a diploid (2n=24) endemic to south-
ern and southeastern Asia, the incongruence does not likely
Molecular phylogeny based on multiple sequences arise from hybridization and subsequent polyploidization,
from three genomes which may result in reticulate evolution. Whether this incon-
Phylogenetic analyses of the aligned sequences of the above gruence reflects “real” underlying biological phenomena or
fragments from three genomes were conducted using maxi-
O. meridionalis
O. punctata
O. officinalis
5
O. rhizomatis
O. australiensis
*Porteresia coarctata-7
8
*P. coarctata-4
O. brachyantha
O. granulata
Leersia oryzoides-1
14
L. perrieri-2
L. hexandra-2
L. perrieri-6
L. hexandra-3
L. oryzoides-4
L. tisserantti
*Hygroryza aristata-Adh2
20
*Prosphytochloa
*Potamophila parviflora
Chikusichloa aquatica
*Rhynchoryza
35 Zizania aquatica
Fig. 1. Strict consensus tree of the rice
tribe Oryzeae generated from the com-
Z. latifolia
bined Adh2 and GPA1 sequences. The
thick lines indicate the nodes with the
Zizaniopsis villanensis bootstrap support and Bayesian pos-
terior probability over 95%, while the
Luziola leiocarpa thin lines indicate the nodes with the
bootstrap support and Bayesian pos-
terior probability from 75% to 94%. The
L. fluitans
numbers on the node represent diver-
gence times (approximate) in million
*Hygroryza aristata-GPA1 years ago (MYA) as estimated by mo-
lecular clock approaches. Broken
Ehrharta erecta lines indicate the placement of
Hygroryza on Adh2 and GPA1 trees, re-
spectively. The numbers following the
Phyllostachys aurea species name represent the clone se-
10 changes quenced. The asterisks indicate the
monotypic genera.
0.4
japonica
+
rufipogon
+
nivara
0.6
glumaepatula
glaberrima
+
barthii
2.0
longistaminata
meridionalis
punctata
australiensis
5 changes
Fig. 2. Strict consensus tree of the A-genome species in Oryza generated from the
combined intron sequences of three nuclear genes. The thick lines indicate the nodes
with the bootstrap support and Bayesian posterior probability over 95%, while the thin
lines indicate the nodes with the bootstrap support and Bayesian posterior probability
from 75% to 94%. The numbers on the node represent divergence times in million
years ago (MYA) as estimated by molecular clock approaches. Accessions were omit-
ted from the figures.
Rice is a component of some remarkably productive and sus- grows were targeted for collection. Many samples of O.
tainable farming systems. However, many factors threaten sus- rufipogon were collected from different parts of the Mekong
tainable rice farming, such as economic development leading Delta, including Phup Hiep and Thap Muoi districts, where
to labor, land, and freshwater constraints to rice-farming sys- the soils were bright red and known to be highly acid sulfate.
tems. In addition, abiotic stresses, including the consequences Passport data of the collections reported this. Consequently,
of climate change, and ever-changing biotic stresses require germplasm from these locations could be rapidly evaluated in
diverse approaches to maintain and improve rice yields. In this standardized conditions to determine the type and degree of
paper, the contributions that diversity in the gene pools of rice abiotic stress resistance it had. Three accessions (106412,
can make to rice improvement are considered. 106423, and 106424)1 were found to be useful sources of re-
sistance to aluminum toxicity. One accession (106424) was
crossed to rice (cv. IR64) and recombinant inbred lines devel-
Primary gene pool oped from this cross were analyzed. Using root length to evalu-
Since Oryza species having the AA genome can hybridize natu- ate the response to stress, five QTLs from O. rufipogon were
rally with one another, they can all be considered part of the found that explained aluminum tolerance and these were
primary gene pool of rice. AA-genome Oryza species have mapped to rice chromosomes 1, 3, 7, 8, and 9 (Nyugen et al
astonishing eco-genetic diversification in relation to environ- 2003). Since this accession (106424) has useful traits and is
ments with different hydrological regimes. During domestica- being used in rice improvement programs, it is one of the ac-
tion, particular ecotypes of wild rice were selected for increased cessions chosen for analysis by the Oryza Map Alignment
yield. Subsequently, because of selection and accompanying Project (www.genome.arizona.edu/BAC_special_projects/).
human migrations, diversification of the cultivated rice gene
pool extended the climatic and geographic range of rice to Biotic stress resistance in O. rufipogon
areas where its wild relatives do not grow (Vaughan et al from the Central Plains of Thailand
2004a,b). The wild relatives of rice in this gene pool have been Tungro is the most serious virus disease of rice in South and
the main source of useful genes for rice improvement from Southeast Asia (Azzam and Chancellor 2002). Sources of re-
wild Oryza. Recent examples of the use of the wild species in sistance to tungro viruses are found in cultivated rice but some
this gene pool are given below. of these have broken down as new forms of these viruses have
emerged. Generally, field collection notes cannot help guide
Abiotic stress resistance in O. rufipogon
from the Mekong Delta
In December 1990, a collaborative collecting mission for the
wild relatives of rice was undertaken by the Vietnam Agricul- 1Accession numbers mentioned are materials in the International Rice
tural Sciences Institute and the International Rice Research Genebank at IRRI and refer to germplasm collected directly by the first au-
Institute (IRRI) in the Mekong Delta. During that mission, ar- thor with collaborating colleagues in national programs while a staff member
eas of adverse soils where the wild rice O. rufipogon Griff. at IRRI.
Twenty-two species of wild rice of the genus Oryza occur as In International Year of Rice 2004, we are celebrating
natives in 81 countries, in all tropical and subtropical regions the world’s most important crop for food security. It is highly
of the world. Rice is believed to have been first cultivated appropriate to start developing a global rice conservation strat-
around 6,000 to 14,000 years ago, with probably at least two egy as one of the early initiatives of the Global Crop Diversity
independent domestication events of O. sativa (with separate Trust.
events for indica and japonica rice) in Asia and a third of O.
glaberrima in Africa. Rice is now cultivated in 114 countries,
The history and diversity of rice
also in all tropical and subtropical regions, with distinctive
forms occurring in different parts of Asia, Africa, Europe, and Cultivated rice has at least three centers of origin (Fig. 1). Oryza
Oceania. sativa spread from its origins in South and East Asia to the
The genetic diversity encompassed by the genus is re- south, north, east, and west, eventually to cover all continents
markable. More than 400,000 accessions are conserved ex situ except Antarctica. The precise time line of the migration of
in genebanks. The task of efficiently conserving and using all rice remains obscure in many countries.
this diversity is formidable, and requires a globally integrated, Primary centers of diversity are clearly associated with
multinational effort. Yet, rice genebanks are working largely the centers of origin. An additional primary center of diversity
independently and ignorant of the achievements of sister in the mainland and islands off Southeast Asia appears to be
genebanks. The resulting scientific and economic inefficiency, associated with the intercrossing of the indica and japonica
with uncontrolled duplication and ineffective targeting of con- forms of O. sativa, encompassing a wide range of intermedi-
servation and use, is clear. ate forms and tropical japonica.
In this paper, we present the first steps toward a plan to On its westward migration, O. sativa reached Europe
rectify the situation, through development of a global strategy more than 2,000 years ago, and secondary centers of diversity
for the efficient conservation and use of rice diversity. We be- can be recognized with distinctive forms in the Indian Ocean,
gin by describing the new Global Crop Diversity Trust, through western Asia, and Europe. More recent distinctive secondary
which it has become possible to consider an efficient global centers of diversity are also apparent in Africa and South
strategy. We then consider the required elements of the strat- America.
egy, and the process toward implementation of those steps. For wild rice, the centers of species diversity and ge-
nomic diversity are the islands from Southeast Asia into the
Pacific Ocean. Nine of the 22 wild species occur in Indonesia,
The Global Crop Diversity Trust and 7 of the 10 genome types are found in the Asian-Pacific
The Global Crop Diversity Trust (www.startwithaseed.org) has islands. In addition, distinctive sets of species assemblages are
been set up to establish and support a rational, efficient, and associated with South Asia, Africa, and the Americas.
sustainable global system for the conservation of crop diver- Given so much genomic, species, and genetic diversity
sity. In the short term, it will assist in the development of ratio- within and between regions, a comprehensive strategy for con-
nal and efficient systems, by providing both expert technical serving the diversity of rice must be truly global in extent.
support and financial support. It will fund the upgrading of
selected national and regional genebanks and the establish-
Ex situ conservation of rice
ment of infrastructure for coordination. In the longer term, it
seeks to provide a permanent source of funding for the main- According to the FAO report on the State of the World’s Plant
tenance of crop diversity collections around the world. Genetic Resources, more than 400,000 accessions of rice are
It is developing the global system through two comple- conserved in genebanks around the world. However, the ma-
mentary strategies. It is developing a series of regional strate- jority of the accessions are in only a small number of genebanks
gies, based on identification of regional priorities building on (Table 1). Over 25% are held in IRRI and more than 60% of
the economic efficiencies of multicrop genebanks. At the same the remaining accessions are held in just five national Asian
time, it is developing a series of global crop strategies, which genebanks in China, India, Japan, Thailand, and the Republic
will be implemented within the context of the regional strate- of Korea.
gies and which will seek the genetic efficiencies of global co- This conservation effort is far from coordinated or glo-
ordination. bal. The extent of duplication across the major genebanks is
unknown. There are strong indications of gaps in coverage
Table 1. Summary of approximate numbers of accessions outside South and East Asia, especially in wild species but
of rice held in major rice genebanks around the world. also in cultivated rice. The absence of coordinated informa-
Genebanka Approximate number of accessions
tion makes it impossible to assess the extent of duplication or
gaps, or to fully and rationally analyze the distribution of di-
IRRI 107,000 versity or target the worst gaps.
China-CAAS 64,000
India-NBPGR 54,000
Japan-NIAR 36,000 The global strategy outlined
Thailand-RRI 24,000
Korea-RDA 23,000 A major element of the rice conservation strategy must be a
WARDA 20,000 global network, through which to coordinate conservation ac-
USA-NSGC 17,000 tivities worldwide. Yet, most existing rice networks are for rice
Brazil-CENARGEN 14,000 breeding and research (e.g., INGER, the International Network
Laos-NARC 12,000
IITA 12,000
for the Genetic Evaluation of Rice), not for genetic resources.
Russia-VIR 5,500 Through the GCDT, it is planned to develop such a network.
Vietnam-PGRC 4,800 Yet, any such network would be ineffective without an effec-
Pakistan-PGRI 2,400 tive information resource on which to base decisions and strat-
Australia 1,400 egies. Indeed, the lack of coordinated information is the major
Indonesia-RIR 1,000
Total 398,100
hindrance to the development of an efficient and rational glo-
bal strategy for the conservation of rice genetic resources.
aIRRI = International Rice Research Institute, CAAS = Chinese Acad- Therefore, the heart of the network must contain a global in-
emy of Agricultural Sciences, NBPGR = National Bureau of Plant Ge- formation system documenting, as far as possible, the relevant
netic Resources, NIAR = National Institute of Agrobiological Resources,
RRI = Rice Research Institute, RDA = Rural Development Administra- data of all accessions conserved in all participating genebanks.
tion, WARDA = Africa Rice Center, NSGC = National Small Grains
Center, CENARGEN = Embrapa Recursos Genéticos e Biotecnologia,
EMBRAPA = Empresa Brasileira de Pesquisa Agropecuária, NARC = Next steps
National Agricultural Research Center, IITA = International Institute for
Tropical Agriculture, VIR = Vavilov Research Institute of Plant Industry, Development of a global information resource will be the first
PGRC = Plant Genetic Resources Center, PGRI = PARC Plant Genetic objective of the global rice conservation strategy.
Resources Institute, PARC = Pakistan Agricultural Research Council,
RIR = Research Institute for Rice. There is a risk, however, that we are planning to avoid,
associated with assembling these data into a common system.
Early attempts to integrate data from different genebanks were
based on standardizing subsets of diverse data from different
genebanks into a common format and collating them into a
central database. Considerable effort was required to populate
the central database with data. Therefore, data in the central
A major challenge to modern geneticists and breeders is to 2002), suggesting that phenotypic diversity has not been suffi-
understand the nature of quantitative trait loci (QTL) and en- ciently exposed to date.
vironmental factors causing variation in quantitative traits, since Generally, isolating barriers appear between species but
adaptive traits are not determined by a single mutation under not within species, raising the question of how such a repro-
changing environments. Genetic complexity results from the ductive barrier could develop between species. We detected
combined effects of multiple genes whose expression is greatly cross-incompatibility in advanced generations of backcross-
influenced by the environment. We present here our recent stud- ing between wild and cultivated rice strains although no cross-
ies on the significance of genetic diversity in wild and culti- ing barrier had been reported within the primary gene pool
vated forms of rice. (Matsubara et al 2003). Genetic analysis revealed that an alien
segment of chromosome 6 from a wild strain caused a failure
of early endosperm development and that the cross-incompat-
Hidden variation in the primary gene pool ibility resulted from the sexual affinity between the cross-in-
The offspring of two plants can contain novel combinations of compatibility reactions in the female and male. Further, the
alleles that produce more extreme phenotypes than those of postfertilization barrier was controlled by at least three linked
either parent. This phenomenon (transgressive segregation) has genes on chromosome 6, showing a different distribution of
been shown to be common (Rieseberg et al 2003). This indi- interacting alleles among taxa (Fig. 1). The results suggest that
cates the importance of gene flow for the emergence of nov- differences leading to species-specific barriers might be partly
elty in crops under man-made habitats, possibly showing a rapid preexisting within the gene pool even though they are not ob-
change in their genetic architecture responding to the environ- served by conventional methods.
ment. Recently, the wild ancestor of maize (teosinte) was shown A profound isolating barrier tends to be species-specific.
to have a high potential to produce phenotypic diversity al- We analyzed a hybrid sterility gene, S1, on chromosome 6 be-
though epistasis had tended to conceal the expression of ge- tween Asian and African cultivated rice species. The hybrid
netic differences in natural populations (Lauter and Doebley sterility gene has the strongest effect among plant hybrids. The
S1 gene is called a “gamete eliminator” since it induces abor-
RG264
RM204 RZ588 Hd1 R32
wx RZ516 RZ398 S1520 OsC1 G200 C235 R111 G2028
Cif cim
Su-Cif
Su-Cif
Incompatible cross
Compatible cross
Fig. 1. (A) Three linked genes (Cif, cim, and Su-Cif) responsible for the unidirectional cross-incom-
patibility detected between wild and cultivated rice strains. (B) The genetic model proposed by
these genes, all of which act sporophytically. Cross-incompatibility occurs when gametes with the
cross-incompatibility reactions in the female (Cif) and the male (cim) are fertilized. Cim and cim are
predominantly distributed in indica and japonica types, respectively. + indicates the cross-compat-
ibility reaction. The locations of the three genes are shown by thick solid lines.
tion of both the male and female gametes possessing its allelic be partly built up on a chromosome by selection, and repul-
alternative S1-a only in the heterozygote. Surveys in their dis- sion or coupling linkages along a chromosome have been dem-
tribution showed that the S1 and S1-a alleles were specific to onstrated in Drosophila (Mather and Jinks 1982). Stabilizing
African and Asian rice species, respectively, although the sen- or normalizing selection could encourage the development of
sitivity to distortion in the female differed between indica and hidden variation within and between populations. Such a linked
japonica types. In addition, fine mapping revealed that only group of polygenes could be reshuffled through recombina-
the female gametes became viable when the flanking region of tion, resulting in the formation of newly recombined pheno-
the S1 gene was deleted, showing a preferential transmission types or transgression.
only in pollen grains. Such a meiotic drive (MD) system vio- Grain characteristics were examined between the two
lating Mendel’s rules has been detected in natural populations subspecific rice strains regarding QTLs on chromosome 6.
of a variety of organisms. The results in rice species suggested Grain characteristics are known to be quantitative traits and to
that a gamete eliminator could be derived from the male MD be diagnostic between them. To enhance the power to detect
by accumulating responder-like elements affecting the female QTLs with small effects, recombinant inbred lines (RILs) were
gamete and that this system could be involved in reproductive made after introgression of chromosome 6, so the segmented
barriers. fragments could be precisely compared, minimizing the re-
sidual variation. The resulting RILs showed a distinct trans-
gressive segregation in the seed dimension. Multiple linked
Polygenic balance theory QTLs were detected on the segment, showing positive and
Phenotypic variation is central to evolutionary adaptation un- negative values in additive effects as well as epistatic interac-
derlying natural and artificial selection. Polygenes, which are tions between the detected QTLs, although such QTLs on chro-
very sensitive to residual variation and the environment, might mosome 6 were not reported before. The results confirmed
RU1 Rp (W1944)
RU2 Rp (W1943)
80
77
SA7 In (160, 706, N303, C6172)
85
RU4 RP (W120)
75 56 B
RU3 Ra (W107) Rin (W1819)
78
RU5 Ra (W1865)
RU6 Rp (W1294)
95
96 64 RU7 Rp (W593) C
66 RU8 Ra (W2002)
Fig. 2. Neighbor-joining (NJ) reconstruction of the genealogical relationships among OsC1 haplotypes
of 43 strains in wild and cultivated rice strains. Oryza glaberrima, O. barthii (AF), and O. glumaepatula
(GLU) were used as outgroups. The maximum-parsimony (MP) method gave the same topology. Boot-
strap values for nodes supported in >50% of 1,000 bootstrap replicates are shown above (for NJ) and
below (for MP) the branches. Jp, Jv, and In represent japonica, javanica, and indica types of O. sativa,
respectively. Ra, Rin, and Rp represent annual, intermediate, and perennial types of O. rufipogon,
respectively. AF includes O. glaberrima (W025) and O. barthii (W1468, W1647).
Notes
Author’s address: Graduate School of Agriculture, Hokkaido Uni-
versity, Sapporo, Japan, e-mail: rysano@abs.agr.hokudai.ac.jp.
Phylogenetic analysis of Oryza species based DNA sequence variation (Fig. 1). The information was used to
on the mitochondrial and chloroplast genomes gain new insights into phylogenetic relationships among 22
Analysis of variation in the cytoplasmic genome is valuable species in the genus Oryza (Vaughan et al 2003). Five chloro-
for taxonomic analysis (e.g., Ge et al 1999). Simple sequence plast SSR loci (Ishii and McCouch 2000) and seven mitochon-
repeats (SSR) and their flanking regions in the mitochondrial drial SSR loci equal to or longer than ten mononucleotide re-
and chloroplast genomes were sequenced in order to reveal peats were chosen from known rice mitochondrial (Notsu et al
GG AA
Related genus
BB
EE
O. officinalis complex
CC
BBCC
Wheat, CCDD
barley,
maize, HHJJ FF
etc.
Genus Oryza
O. ridleyi complex
Fig. 1. A model for evolution of the cytoplasmic genome in Oryza species based on
SSRs and their flanking regions.
2002) and chloroplast genome sequences (Hiratsuka et al 32 accessions of Oryza species carrying the AA genome. The
1989). A total of 50 accessions of Oryza that represented six 8K was originally found as ORF100 by Chen et al (1993) and
different genomes of diploid Oryza species and three different Kanno et al (1993). The three regions showed a different ex-
allopolyploid genomes were analyzed. Many base substitutions tent of sequence variation. In the 8K region, most indica
and deletions/insertions were identified in the SSR loci and (76.1%) had a deletion type of polymorphism, whereas most
their flanking regions. Of mononucleotide SSR, G (or C)-re- japonica (98.5%) had a nondeletion type. In contrast, in the
peats were more variable than A (or T)-repeats. Results ob- 76K region, no length polymorphism was observed at all within
tained by chloroplast and mitochondrial SSR analyses showed the cultivated rice examined. In the 57K region, various length
similar phylogenetic relationships among species, although polymorphisms, so far six types, were identified by DNA se-
chloroplast SSR were more informative because of their higher quencing. These results showed that the extent of polymor-
sequence diversity. The CC genome is suggested to be the phism is variable depending on the chloroplast locus and that
maternal parent for the two BBCC-genome species (O. occurrence of an insertion/deletion event is independent re-
punctata and O. minuta) and the CCDD species (O. latifolia) garding the three regions. Sequence variations at the 8K and
based on the high level of sequence conservation between dip- 57K regions were also present in some of the O. nivara and O.
loid CC-genome species and these allotetraploid species. rufipogon but not in other wild species, suggesting that muta-
tions at the two regions happened relatively recently and prior
to the domestication of Asian cultivated rice. These results
Polyphyletic domestication of indica, temperate japonica, suggest that indica of type IV and VII were domesticated ei-
and tropical japonica from O. rufipogon: evaluation ther from a local population of O. rufipogon via O. nivara or
of genetic diversity by the chloroplast genome directly from a local population of O. rufipogon separately.
It is believed that Oryza sativa ecospecies indica and japonica Regarding japonica, japonica of type VI and IV (tropical type)
were domesticated in Asia from O. rufipogon and/or O. nivara, were derived from a different local population of O. rufipogon
but that domestication process is still not well understood. without O. nivara, polyphyletically and independently. A pre-
Complete chloroplast genome information is available for one vious study using nuclear element SINEs also showed that O.
japonica variety (Hiratsuka et al 1989). Recently, we have de- sativa has been derived polyphyletically from O. rufipogon
termined a whole chloroplast genome of O. nivara, an annual (Ohtsubo et al 2004). Our study strongly suggests that there
type of wild rice and possible progenitor to O. sativa (Masood are at least four major lineages for the domestication process
et al 2004). Several insertion/deletion events were identified of Asian cultivated rice (Fig. 2).
between the two species. To investigate the domestication pro-
cess of rice, three polymorphic loci, 8K, 57K, and 67K, were
examined using 303 accessions of Asian cultivated rice and
Type IV-AA
1
Type IV-TT 16
2
Hypothetical
ancestor 15
Type VII-TT
17
107
Type VI ?
unidentified
indica
japonica
O. rufipogon 81
Fig. 2. A model for the domestication process of cultivated rice from wild species.
Numerals in the figure show the number of accessions.
Rice is widely cultivated in a large number of different natural It has been reported that the young microspore stage in pollen
environments. Compared with other cereal crops such as wheat development was the most sensitive to low temperature (Satake
and barley, rice is much more sensitive to low temperature, and Hayase 1970). Exposure of rice plants at the tetrad stage
probably as a result of its subtropical origin. In terms of im- to a moderate low temperature (12 ºC) for 4 days resulted in
pact on rice production, male sterility is the most damaging male sterility in 80% of the spikelets (Satake and Hayase 1970,
impact of chilling. The developmental stages from pollen for- Nishiyama 1984). Microscopic observation of developing rice
mation to fertilization are the most vulnerable to low tempera- anthers suggested that one possible reason for male sterility
ture throughout the life cycle of rice plants (Nishiyama 1984). after low-temperature treatment was the failure of anther de-
Myanmar is a tropical country in monsoon Asia and is included rice. Therefore, agroecological zoning was attempted by us-
in the center of diversity of cultivated rice along with Yunnan ing monthly meteorological data accumulated at 34 observa-
Province (China), east Nepal, Bhutan, Assam (India), and north- tory sites throughout Myanmar for more than 30 years of rain-
ern Thailand (Nakagahra and Hayashi 1977, Nakagahra et al fall; maximum, minimum, and mean temperatures; differences
1984, Glaszmann 1987). Myanmar has varied natural envi- between day and night temperatures; evapotranspiration; and
ronments, various social and cultural customs, different types sunshine, compiled by FAO (1987). Eight agroecological zones
of farming and cropping systems, and tribal diversity. Conse- were elaborated by the results of the principal component analy-
quently, various ecotypes such as rainfed, irrigated, deepwater, sis: Northern Mountainous (NM), Eastern Plateau (EP), Semi-
floating, and upland rice are cultivated and a number of arid (SA), Western Hilly (WH), Western Coastal (WC), South-
landraces are differentiated in each ecotype. A number of ern Coastal (SC), Southern Plain (SP), and Ayeyarwady Delta
landraces are still cultivated locally but are being rapidly re- (AD) (Fig. 1). The NM, EP, and WH zones had the most com-
placed by improved cultivars. It is thus necessary to collect plex topography, with high mountains and deep valleys, but
landraces for ex situ conservation and to characterize and evalu- the others are rather flat and plain zones.
ate them for future rice breeding. In Myanmar, rainfed rice is the most dominant ecotype
Our study deals with a thousand or more rice landraces and it is grown extensively in SP and AD zones. In the moun-
collected throughout Myanmar in order to investigate their ge- tainous zones, NM, WH, and EP, upland rice is preferentially
netic diversity and characterize them. grown in the slash-and-burn farming system. Irrigated rice is
predominantly cultivated in the central SA zone and deepwater
and floating rice are grown exclusively in delta areas in AD
Agroecological zoning of rice-growing environments and WC zones.
The characterization of rice-growing areas was first needed
for analyzing the diversity of landraces of Myanmar cultivated
NM
WH
EP
WC n = 260
n = 130 EP
SA
WC
SP
SC SA
SP AD n = 125
n = 144
SC
AD n = 217
n = 149
Low
Medium
High
Myanmar
n = 1,391
45 45
40 WC 40 EP
35 n = 130 35 n = 260
30 30
25 WH 25
20 20
EP
15 15
10 SA 10
5 WC 5
0 0
45 SP 45
40 SP 40 SA
35 n = 144 35 n = 125
30 AD SC 30
25 25
20 20
15 15
10 10
5 5
0 0
45 45
40 AD 40 SC
35 n = 149 35 n = 217
30 30
25 25
20 20
15 15
10 10
5 5
0 0
0 4 8 12 16 20 24 28 32 36 40 0 4 8 12 16 20 24 28 32 36 40
2 6 10 14 18 22 26 30 34 38 2 6 10 14 18 22 26 30 34 38
No. of landraces
180
160 Myanmar
n = 1,391
140
120
100
80
60
40
20
0
0 4 8 12 16 20 24 28 32 36 40
2 6 10 14 18 22 26 30 34 38
Amylose content (%)
Fig. 2. Regional variation in amylose content of endosperm starch in Myanmar rice landraces. WH = Western Hilly, NM = Northern
Mountainous, EP = Eastern Plateau, SA = Semiarid, WC = Western Coastal, SC = Southern Coastal, SP = Southern Plain, and AD
= Ayeyarwady Delta.
In Korea, the most important goal in japonica rice breeding is Materials and methods
to improve the eating quality of cooked rice. The eating qual-
ity of rice is a complex characteristic whose many components Plant materials
are difficult to evaluate for selection in a rice breeding pro- Suweon365 possesses medium eating quality and is highly re-
gram. The quality of cooked rice has been examined by a sen- sistant to blast. On the other hand, Chucheongbyeo is known
sory test. Recently, we have been using an indirect rapid evalu- for its good eating quality and susceptibility to leaf blast. A
ation method by detecting the glossiness of cooked rice for total of 231 recombinant inbred lines (RILs) derived from a
eating quality. However, this is difficult to apply for early-gen- cross between Suweon365 and Chucheongbyeo were devel-
eration selection in the rice breeding program because of size oped using the single-seed descent method. The RIL popula-
constraints. tion was field-evaluated by the National Institute of Crop Sci-
The use of molecular markers has facilitated the under- ence, Suwon, for three years (1999 to 2001). To develop a
standing of quantitative trait loci (QTLs) and marker-assisted series of QTL-NILs (near-isogenic lines), four desirable RILs
selection (MAS). Many QTL analyses for grain quality of rice carrying QTL segments for glossiness of cooked rice (GCR)
have been reported (Bao et al 2002, Tan et al 1999, Wan et al from a Toyo taste meter value were successively backcrossed
2004, Zhou et al 2003). QTLs related to glossiness of cooked with Suweon365 for selection using linked markers until the
rice were identified at different loci in Ilpumbyeo—a high- BC4F2. The allele effect of QTLs related to GCR was evalu-
quality parent (Cho et al 2004, Lee 2003). In this study, we ated in BC3F1 and BC4F2 generations.
report on QTLs associated with physicochemical traits and
eating quality by glossiness of cooked rice using a japonica × Chemical properties
japonica cross. Also, we evaluated the allelic effect of QTLs The amylose content of milled rice was determined by the rela-
associated with glossiness of cooked rice in BC3F1 and BC4F2 tive absorbency of starch-iodine color in digested solution of
generations developed by MAS. 100-mesh rice flour by the modified method of Juliano (1971).
Protein content was calculated by total nitrogen multiplied by
qGCR-11
OSR01
8.6 cM
RM180
RM20B
qGCR-7-1
3.2 cM
RM214
qGCR-8
3.8 cM
RM547 RM332
RM500
em623 RM72 RM552
em723
em724 em445
em127
em145
em225
em525 RM044
em524 RM404
qHV-8
em425 em1217
em5211
em424 em7215
em429 qGIT-11
qADV-11 em429
1031R25
em4210 em2210
em7210 em4214
1031R08
em123 RM229
em522 RM021
RM320
RM342A RM206
RM560
OSR22 RM342B em426
RM342D RM515 em812
em443 em427
1031R28
em243 em726
em143 em129
RM346
RM182 em526
RM281
OSR04 em442
RM336 RM264
RM144
RM070
em529 em242
em738 dasng08
em1210 1031R23
qGCR-7-2
em1211
4.2 cM
em428
em229
em627
dasng04
RM505
em864
RM473a
RM234
RM47
RM478
Fig. 1. Chromosomal location of QTLs associated with glossiness of cooked rice. qGCRs are QTLs for glossiness of cooked rice
from the Toyo taste meter.
aHomo of CC: one band for Chucheongbyeo; hetero of CS: two bands of Chucheongbyeo and Suweon365; homo of SS: one band
for Suweon365. bAllelic effects in BC3F1 and BC4F2 are the value of CS-SS and CC-SS, respectively.
Lee JS. 2003. QTL analysis for grain quality properties in a japonica Zhou PH, Tan YF, He YQ, Xu CG, Zhang Q. 2003. Simultaneous
rice combination. Ph.D. thesis. Yeungnam University, Korea. improvement for four quality traits of Zhenshan 97, an elite
64 p. parent of hybrid rice, by molecular marker-assisted selection.
Tan YF, Li JX, Yu SB, Xing YZ, Xu CG, Zhang Q. 1999. The three Theor. Appl. Genet. 106:326-331.
important traits for cooking and eating quality of rice grains
are controlled by a single locus in an elite rice hybrid, Shanyou
63. Theor. Appl. Genet 99:642-648. Notes
Tian R, Jiang GH, Shen LH, Wang LQ, He YQ. 2004. Mapping quan- Authors’ addresses: H.G. Hwang, J.P. Suh, Y.C. Cho, I.S. Choi, H.C.
titative trait loci underlying the cooking and eating quality of Hong, Y.G. Kim, M.K. Kim, H.C. Choi, Y.T. Lee, National
rice using a DH population. Mol. Breed., on-line, accessed Institute of Crop Science, RDA, Suwon 441-857, Republic of
on 30 Sept. 2004. Korea; S.J. Kwon, National Institute of Agricultural Biotech-
Wan XY, Wan JM, Su CC, Wang CM, Shen WB, Li JM, Wang HL, nology, RDA, Suwon 441-857, Korea, e-mail:
Jiang L, Liu SJ, Chen LM, Yasui H, Yoshimura A. 2004. QTL yccho@rda.go.kr.
detection for eating quality of cooked rice in a population of
chromosome segment substitution lines. Theor. Appl. Genet.,
on-line, 10.1007/s00122-004-1744-3.
Rice is one of the main staples in the world and is cultivated genetically and reverse-genetically (NIAS 2004a, Sasaki 1998).
mainly in Asia, Africa, and Latin America using irrigated and This paper focuses on the rice genome sequencing efforts, with
upland ground for paddy fields. The major limiting factors for emphasis on the analysis and utility of the genome sequence in
rice cultivation are climate, soil, and water. Except for the the future.
upland rice in Africa, rice plants grow mainly in tropical and
semitropical areas characterized by a rainy season. The wild
How to sequence and decode the information of heredity
relatives of cultivated rice, Oryza sativa, are also widely dis-
tributed within nearly 20° north and south latitude (IRRI 2004). The term “genome” refers to the total genetic information that
Archaeological evidence suggested that rice cultivation began defines a species. The size of the genome corresponds to the
about 10,000 years ago in southern China, with wild rice as amount of nucleotide that comprises half of the chromosome
the main resource. Although the exact origin of both O. sativa pairs or haploid. This can be experimentally measured by flow
subsp. indica and subsp. japonica is still unknown, rice has cytometry and expressed as a nuclear C-value. The genome
been well domesticated even in such a “short” period and rice size of rice with 12 chromosomes has been estimated to be
cultivation has become widespread, extending to 40° north and 430 Mb (Royal Botanic Gardens 2004). Rice has the smallest
south latitude. The domestication of rice is undoubtedly one genome size when compared with other major cereal crops
of the most important developments in human history and, at such as maize, with 2,500 Mb, and wheat, with 16,000 Mb.
present, about half of the world population relies on rice for However, even with the most advanced method of nucleotide
its daily main food source. sequencing, the size of rice is not an easy target for complete
Despite the long history of breeding and improvement sequencing with high accuracy. In the case of rice, it is also
of rice varieties adapted to a wide range of environmental con- important to assign the location of the sequence to its genomic
ditions as well as varieties highly preferred by both farmers position so that the information can be directly useful to breed-
and consumers, novel rice varieties are still necessary to meet ers. This approach in sequencing analysis facilitates the iden-
the demands of increasing population pressure coupled with tification of gene function by forward and reverse genetics and
decreasing resources for rice production. In addition, variet- comparison of specific phenotypes among rice varieties based
ies with tolerance of biotic and abiotic stress, and high yield as on the difference in nucleotide sequence. This is extremely
well as good nutritional and cooking quality, will always be important considering that there are about 120,000 rice variet-
targeted by breeders and farmers. The breeding strategy used ies worldwide and each variety is characteristically defined by
so far for developing new varieties is based on the selection of the nucleotide sequences. Thus, an accurate and exact stan-
desirable phenotypes among progenies generated after cross- dard genome sequence is indispensable for subsequent goals
ing elite varieties. Since most agronomic traits are composed in rice genomics.
of more than two genes and, in extreme cases, ten to 15 genes, The rice genome sequencing effort has been an interna-
this strategy is similar to a “hit or miss” approach, making it tional collaboration since 1998. The International Rice Ge-
extremely difficult to select the most preferable allele for each nome Sequencing Project (IRGSP) is currently composed of
of the genes corresponding to a phenotype. We must be inno- 10 participating countries, led by Japan (Sasaki and Burr 2000,
vative in improving our rice breeding strategy for such com- NIAS 2004b). A clone-by-clone sequencing strategy using a
plex traits mainly contributed to farming and agribusiness by standard japonica rice variety, Nipponbare, has been adopted.
introducing up-to-date scientific knowledge. Although each country’s contribution differs, each group has
Recent advances in molecular biology made it possible made significant progress in decoding the rice genome se-
to acquire extensive knowledge of rice genetics such as the quence. Japan has contributed in sequencing 55% of the entire
fundamental factors of heredity and the nucleotide sequence genome, or an equivalent of 220 Mb genome sequence. In
of the genome. In Japan, the Ministry of Agriculture, Forestry, December 2002, the IRGSP finished the high-quality draft se-
and Fisheries (MAFF) started the rice genome project in 1991 quence of the entire genome, which is now available in public
with the aim of analyzing the structure of the rice genome, databases. It is expected that by December 2004 the entire
understanding the rice plant based on its genetic composition, Nipponbare genome, covered by nearly 3,400 BAC (bacterial
and using the information for innovative rice breeding strate- artificial chromosome) and PAC (P1-derived artificial chro-
gies. From 1998, the project has served as a launching pad for mosome) clones, will be sequenced with 99.99% accuracy
sequencing the entire genome through international coordina- without any sequence gap. The genomic positional assignment
tion and elucidating its function based on this information both of tiled BAC and PAC clones has been established based on
Over the last decade, a tremendous amount of progress has ing. This paper describes our recent progress in the marker-
been made in rice genome analysis, including in whole-genome assisted genetic and molecular dissection of several complex
sequencing (Sasaki 2003). This progress has provided power- traits and our efforts to develop plant materials that are, or will
ful tools—DNA markers—for plant genetics and breeding. be, valuable in the analysis of complex traits in rice.
DNA markers have been used in the genetic analysis of agro-
nomically important traits, such as disease and pest resistance,
Genetic and molecular dissection of complex traits in rice
plant height, and grain quality. Many phenotypic traits of eco-
nomic interest are controlled by naturally occurring allelic Flowering time (heading date)
variations and environmental conditions. These traits often Flowering time is a major determinant of the regional and sea-
show complex and quantitative inheritance. Recent progress sonal adaptation of rice cultivars and is one of the important
in rice genomics has had a large impact on our understanding traits in breeding programs. Analysis of natural allelic varia-
of such complex traits. DNA markers are used in the genetic tion has contributed significantly to the exploration of genes
analysis of quantitative traits, such as heading date, environ- involved in flowering time in rice. In quantitative trait locus
mental stress tolerance, and yield-related traits. This marker- (QTL) analysis, 15 QTLs for heading date have been detected
assisted approach has also made it possible to exploit hidden genetically by using different types of progeny derived from a
genetic factors contributing to the improvement of traits with cross between Nipponbare and Kasalath (Yano et al 2001).
agronomic importance and to clone genes involved in such Near-isogenic lines (NILs) for each QTL have also been de-
complex traits at the molecular level (Yano 2001). Tightly veloped by successive backcrossing and MAS. Consequently,
linked DNA markers or causal genes can be used in marker- eight QTLs have been mapped as single Mendelian factors by
assisted selection (MAS) in breeding programs. The advent of using advanced backcross progeny (Yano et al 2001). Further-
DNA markers has allowed us to practice MAS in rice breed- more, four QTLs (Hd1, Hd3a, Hd5, and Hd6) have been cloned
Fig. 1. Chromosome segment substitution lines (CSSLs) derived from a cross between Koshihikari and Kasalath. (A): Flowchart of
the process of selection of CSSLs. (B): Graphical representation of genotypes of the CSSLs. The black segments indicate regions
homozygous for Kasalath alleles; the white segments indicate regions homozygous for Koshihikari alleles; the hatched segments
indicate heterozygous regions. Genotype classes of the 129 RFLP markers for the CSSLs can be obtained at the Web site of the
Rice Genome Resource Center (www.rgrc.dna.affrc.go.jp/ine39.html).
Table 1. Chromosome segment substitution lines (CSSLs) derived from crosses within the cultivated species
Oryza sativa.
Cross combination
Population/ Number of Genotype data Reference
Recurrent Donor generation lines
fects in detail. Because CSSLs normally have one chromosomal Use of exploited natural allelic variations in rice
region substituted, they can be used as NILs themselves or as breeding
starting material to develop NILs. The advent of DNA marker technology has made it possible to
Development of NILs for QTLs is necessary for verifi- establish new breeding strategies, such as marker-assisted
cation and characterization of the QTLs detected (Tanksley breeding (Peleman and van der Voort 2003). That strategy re-
1993, Yano and Sasaki 1997). Once a QTL for a particular trait quires comprehensive dissection and understanding of com-
is detected, a CSSL can be used as a NIL itself or as a base plex traits of interest. To that end, even if a large number of
material for NIL development. As only one chromosome frag- DNA markers are available, novel plant materials (mapping
ment has been substituted, only one additional backcross will populations) such as ILs or CSSLs are necessary for marker-
be required for NIL development. Such NILs allow us to com- assisted strategies to be used in practical breeding (Zamir 2001,
bine two or three QTLs in one genetic background to clarify Peleman and van der Voort 2003). As mentioned above, novel
the epistatic interaction among the QTLs. mapping populations and plant materials have already been
developed. To enhance the exploitation and use of novel al-
leles for rice breeding, these CSSLs should be grown under
HHKK
CCDC
CCDC
BBCC
BBCC
IR64
IIIJJ
NIP
GG
GG
BB
CC
CC
CC
CC
CC
CC
CC
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
EE
1 5 10 15 20 25 30 35 40 45
B
1 5 10 15 20 25 30 35 40 45 I/N 1 5 10 15 20 25 30 35 40 45 I/N
C
1 5 10 15 20 25 30 35 40 45 I/N 1 5 10 15 20 25 30 35 40 45 I/N
Fig. 2. EcoTILLING of DREB1 on wild Oryza. DREB1 amplicons of a similar size were produced from
47 wild Oryza species using primers designed from Nipponbare (A). Mixtures of these products
were combined with either IR64 (left set of lanes) or Nipponbare (right set of lanes). The IRD700
channel is shown in panel B and the IRD800 in C. Circles indicate clear differences present in
both channels. Arrows indicate a mismatch present in japonica but absent from most wild and
indica rice.
Cohen JC, Kiss RS, Pertsemlidis A, Marcel YL, McPherson R, Hobbs Till BJ, Reynolds SH, Greene EA, Codomo CA, Enns LC, Johnson
HH. 2004. Multiple rare alleles contribute to low plasma lev- JE, Burtner AR, Young K, Taylor NE, Henikoff JG, Comai L,
els of HDL cholesterol. Science 305:869-872. Henikoff S. 2003. Large-scale discovery of induced point-
Cooper B, Clarke JD, Budworth P, Kreps J, Hutchison D, Park S, mutations with high-throughput TILLING. Genome Res.
Guimil S, Dunn M, Luginbuhl P, Ellero C, Goff SA, 13:524-530.
Glazebrook J. 2003. A network of rice genes associated with Zeng L, Qu S, Bordeos A, Yang C, Baraoidan M, Yan H, Xie Q,
stress response and seed development. Proc. Natl. Acad. Sci. Nahm BH, Leung H, Wang GL. 2004. Spl11, a negative regu-
USA 100:4945-4950. lator of plant cell death and defense, encodes a U-Box/ARM
Hirochika H, Guiderdoni E, An G, Hsing Y, Eun MY, Han CD, repeat protein endowed with E3 ubiquitin ligase activity. Plant
Upadhyaya N, Ramachandran S, Zhang Q, Pereira A, Cell 16:2795-2808.
Sundaresan V, Leung H. 2004. Rice mutant resources for gene
discovery. Plant Mol. Biol. 54:325-334.
Laitinen T, Polvi A, Rydman P, Vendelin J, Pulkkinen V, Salmikangas Notes
P, Mäkelä S, Rehn M, Pirskanen A, Rautanen A, Zucchelli M, Authors’ addresses: Division of Entomology and Plant Pathology,
Gullstén H, Leino M, Alenius H, Petäys T, Haahtela T, Laitinen Genetic Resources Center, Division of Plant Breeding, Ge-
A, Laprise C, Hudson TJ, Laitinen LA, Kere J. 2004. Charac- netics, and Biochemistry, Division of Crop, Soil, and Water
terization of a common susceptibility locus for asthma-related Sciences, International Rice Research Institute, DAPO Box
traits. Science 304:300-304. 7777, Metro Manila, Philippines; Department of Biology, Uni-
Lazzaro BP, Sceurman BK, Clark AG. 2004. Genetic basis of natural versity of Washington, Seattle, Washington, USA.
variation in D. melanogaster antibacterial immunity. Science Acknowledgments: This work is supported in part by the Swiss
303:1873-1876. Agency for Development and Cooperation (SDC), Rockefeller
Liu B, Zhang S, Zhu X, Yang Q, Wu S, Mei M, Mauleon R, Leach J, Foundation, and Generation Challenge Program.
Mew T, Leung H. 2004. Candidate defense genes as predic-
tors of quantitative blast resistance in rice. Mol. Plant-Mi-
crobe Interact. 17:1146-1152.
With the completion of genomic sequencing of rice, rice has recently summarized by Hirochika et al (2004). Here, I sum-
been firmly established as a model organism for both basic marize recent progress in insertional mutagenesis in rice using
and applied research. The next challenge is to uncover the func- Tos17 and its application to forward and reverse genetics stud-
tions of genes predicted by sequence analysis. Knockout of ies. In addition, I also mention our new challenge, the devel-
genes by insertional mutagenesis is a straightforward method opment of activation tagged lines using full-length cDNA.
for identifying gene functions. In Arabidopsis, whose entire
genomic sequencing has been completed, mutant populations
Forward and reverse genetics using Tos17
covering almost all the genes have been produced by using
insertion elements, such as T-DNA, Ac/Ds, and En/Spm. While The unique features of Tos17 that make it a powerful genetic
these mutant lines are important resources for forward genet- tool for forward and reverse genetics studies are summarized
ics studies of gene function, their applications in reverse ge- as follows. (1) Transposition can be regulated since Tos17 is
netics are even more important with the available Arabidopsis activated by tissue culture and becomes silent in regenerated
genome sequence. Two primary approaches for reverse genet- plants. (2) Highly mutagenic during tissue culture, Tos17 trans-
ics analysis have been established. One is PCR screening of poses preferentially into gene-rich, low-copy regions and about
DNA pools from mutants using primers corresponding to a ten loci on average are disrupted in each plant regenerated
target sequence. The other is flanking sequence tag (FST) analy- from 5-month-old culture. (3) Integration target loci were
sis, a strategy in which sequences flanking the insertion ele- widely distributed over the chromosomes, so that random in-
ments are determined in each mutant line to develop a data- sertion for saturation mutagenesis is feasible. (4) Induced
base of knockout genes that can be searched electronically. mutations are stable and germinally transmitted in the next
Because of the labor- and time-demanding nature of PCR generation. (5) The original copy number is quite low, one to
screening, FST database development is now the preferred five depending on varieties, so that it is easy to identify the
strategy. Although investment in the FST database is high, transposed copy responsible for the specific mutation. (6) Re-
screening of mutants is easy once the database is established. arrangements are very rare at junctions between Tos17 ends
In several Arabidopsis laboratories (http://signal.salk.edu/ and flanking host sequences, so that screening of mutants by
tabout.html, http://atidb.cshl.org, and http://genoplante- PCR and analysis of disrupted genes can be carried out with
info.infobiogen.fr), a large collection of FST data, totaling more high efficiency. (7) The transposon is endogenous, so that
than 100,000, has been established and opened to the public. screening and characterization of mutants in the field are pos-
A large-scale insertional mutagenesis has begun also in rice sible without any environmental concern. However, some fea-
using T-DNA, Ac/Ds, and the endogenous retrotransposon tures might also be disadvantageous. For example, transposi-
Tos17. The current status of rice mutant resources has been tion via a copy-and-paste mode means that no revertants can
As the international efforts to completely sequence the rice Current status of rice genome annotation
genome are nearly completed, an immediate challenge and
opportunity for the plant community is to comprehensively and Estimation of the total gene number in rice based upon the
accurately annotate the rice genome. This will lay a solid foun- initial draft sequences of Oryza. sativa L. subsp. japonica and
dation for rice functional genomics and proteomics. We report indica ranged widely from 30,000 to 60,000 (Goff et al 2002,
here our current strategy to use integrative approaches cen- Yu et al 2002). Finished sequences of chromosomes 1, 4, and
tered on whole-genome nucleotide tiling-path microarray and 10 allowed a much fine-tuned estimation that placed the total
array-guided full-length cDNA analysis for conducting ge- gene number of rice between 57,000 and 62,500 (Feng et al
nome-wide transcriptional analysis. Preliminary results sug- 2002, Sasaki et al 2002, The Rice Chromosome 10 Sequenc-
gest that these experimental approaches are useful in verifying ing Consortium 2003). Although the fast-accumulating finished
predicted rice gene models and in identifying a significant sequences will ultimately come to a conclusion for the total
number of new transcriptional units that were simply missed gene number estimation, it should be noted that about half of
from current annotation. The pilot results further indicate that those annotated gene models lack experimental support. Fur-
array data-guided RT-PCR cloning of full-length cDNA is ther, between one-third and one-half of the predicted genes
emerging as an effective protocol to provide the much-needed appear to have no recognizable homology in Arabidopsis and
experimental verification for those rice genes that lacked pre- features that exhibit striking deviations from experimentally
vious cDNA support. Thus, success of our effort described verified genes in other species. Therefore, experimental ef-
herein is likely to result in a more accurate annotation of all forts complementary to computer-based genome annotation are
rice genes in the genome. needed to verify predicted genes and to discover new genes in
rice.
Fig. 1. A proposed workflow for using tiling microarray to improve rice genome annotation. Current rice
genome annotation efforts focus on ab initio gene prediction, comparative genomics, and other
nonexperimental methods (block arrow on the right). Experiment-oriented approaches such as whole-
genome tiling microarray and full-length cDNA analysis are emerging as complements to the computa-
tion-based methods (block arrow on the left). Arrowheads indicate the direction of information flow. See
text for more details.
It has been realized from the pilot experiments that new com-
Integration of tiling array data to improve rice genome putational tools need to be developed and validated to inter-
annotation pret tiling data and facilitate incorporation of tiling data to
The hybridization data of the tiling microarrays can be used to improve rice genome annotation. For example, algorithms to
further rice genome annotation (Fig. 1). For example, the hy- ascertain transcription need to be improved by incorporating
bridization data can provide support to or verification of pre- more statistical parameters. Reliable tiling data in turn will
dicted genes without prior experimental support, as detection provide better training data to improve gene prediction and
of hybridization signals is strongly indicative of RNA synthe- other rice bioinformatic tools to enhance genome annotation
sis directed by the genome segment represented by the inter- (Fig. 1). The tiling data can be used in conjunction with sev-
rogating oligos. Thus, the information generated from tiling eral other public-funded rice genomics projects to maximize
arrays can be used to improve genome annotation by means of the detection of rice genes expressed at different developmen-
array-guided RT-PCR cloning and analysis of corresponding tal stages or under diverse environmental conditions. Conse-
cDNAs. Likewise, the array data are also expected to reveal quently, improved genome annotation will also permit better
novel transcriptional units (i.e., those that were not included array design to probe subtle transcriptional events such as al-
in any annotation but were detected by tiling arrays). In this ternative splicing, differential initiation and termination, etc.
regard, the array data could serve as a guidepost for cloning Finally, to present the tiling data in an accessible and informa-
the corresponding cDNA by means of RT-PCR (Fig. 1). Com- tive form, an interactive database is required where the tiling
pared with sequencing library-based full-length cDNA clones, data will be correlated with the complete genome sequence
which is considered the standard of gene annotation (Kikuchi with all the annotated rice genes and other genomic features
et al 2003), array-assisted cDNA cloning and analysis will be linked to cDNA/EST sequences and proteomic and structural
targeted, and thus be more cost-effective and inclusive to cover information. Applying these approaches in rice should aid in
the remaining portion of the rice genome that lacks correspond- the current genomic efforts to provide a complete and accu-
ing expressed sequences. rate rice genome annotation that holds the key to unravel the
biology of the agriculturally important cereal crops.
The current research on japonica rice genome tiling ar-
Compilation of reference genes for tiling array analysis ray analysis in the laboratory of Xing Wang Deng at Yale Uni-
For the sake of simplicity, a set of BGI (Beijing Genome Insti- versity is supported by a grant from the National Science Foun-
tute) japonica gene models was used as the reference genes in dation plant genome program (DBI-0421675). The collabora-
the pilot trial (Stolc et al 2004). To effectively decode the til- tive effort in China was supported by the 863 rice functional
ing microarray hybridization data, however, reference genes genomics program from the Ministry of Science and Technol-
are critically important, for which a comprehensive set of rice ogy of China.
gene models is precisely anchored in the genomic sequences.
A unique advantage in rice is the availability of multiple an-
References
notations such as the TIGR (The Institute for Genomic Re-
search) japonica annotation (www.tigr.org/tdb/e2k1/osa1/) and Feng Q, Zhang YJ, Hao P, Wang SY, Fu G, Huang YC, Li Y, Zhu JJ,
the BGI indica and japonica annotations (http:// Liu YL, Hu X, et al. 2002. Sequence and analysis of rice chro-
rise.genomics.org.cn). While each annotation has its own forte mosome 4. Nature 420:316-320.
and weakness, comparison and analysis of these annotations Goff SA, Ricke D, Lan TH, Presting G, Wang RL, Dunn M,
Glazebrook J, Sessions A, Oeller P, Varma H, et al. 2002. A
should yield a more comprehensive inventory of gene models
draft sequence of the rice genome (Oryza sativa L. ssp.
to be evaluated by tiling arrays than any single annotation can
japonica). Science 296:92-100.
offer. Reference gene compilation also takes advantage of the Kapranov P, Cawley SE, Drenkow J, Bekiranov S, Strausberg RL,
EST and full-length cDNA collections in the TIGR and the Fodor SPA, Gingeras TR. 2002. Large-scale transcriptional
Kikuchi full-length cDNA data sets (http:// activity in chromosomes 21 and 22. Science 296:916-919.
cdna01.dna.affrc.go.jp/cDNA/). For example, the full-length Kikuchi S, Satoh K, Nagata T, Kawagashira N, Doi K, Kishimoto N,
cDNA-confirmed genes could naturally serve as positive con- Yazaki J, Ishikawa M, Yamada H, Ooka H, et al. 2003. Col-
trols to test array data and be used as training data to improve lection, mapping, and annotation of over 28,000 cDNA clones
rice gene-finding algorithms (Fig. 1). Moreover, rice sequences from japonica rice. Science 300:1566-1569.
have been subjected to extensive comparative genomic analy- Sasaki T, Matsumoto T, Yamamoto K, Sakata K, Baba T, Katayose
sis with other cereals and other plant species. Closer and fur- Y, Wu JZ, Niimura Y, Cheng ZK, Nagamura Y, et al. 2002.
The genome sequence and structure of rice chromosome 1.
ther inspection of those data should help to identify common
Nature 420:312-316.
and unique cereal genes and to integrate multiple annotations
to provide a more comprehensive representation of the rice
genome content (Fig. 1).
Phosphorus (P) deficiency is a major yield-limiting factor for Candidate gene identification
rice, particularly under upland or rainfed lowland conditions
(Kirk et al 1998). P deficiency is frequently not due to low Efforts at IRRI are now directed toward identifying the gene(s)
soil-P content in absolute terms but due to tight binding of at the Pup1 locus. Further fine mapping of Pup1 has advanced
soil-P in forms that are not readily available to the plant. The considerably and Pup1 has now been mapped to a 195-kb in-
development of rice cultivars capable of using a higher por- terval spanning three BAC clones on chromosome 12 (Fig. 1).
tion of this fixed P already present in soils could be an attrac- Gene annotation in the Pup1 region identified 31 putative
tive and cost-effective approach to increasing rice yields on P- genes. Only one of those is a known gene, whereas some se-
deficient soils. quence similarities to known genes exist for an additional four
One promising step toward the development of more P- genes. The remaining annotations were for hypothetical pro-
efficient cultivars was the identification of the Pup1 locus in a teins (13) and transposable elements (13). None of the sequence
QTL mapping population that was derived from rice cultivars similarities suggested an association with processes involved
Nipponbare (low P uptake) and Kasalath (high P uptake) in P uptake or metabolism. This would suggest that Pup1 is
(Wissuwa et al 2002). The effect of Pup1 was confirmed using most likely a novel gene. It is also possible, however, that the
a near-isogenic line (NIL) carrying the positive Pup1 allele gene is simply absent or highly distorted in Nipponbare, which
from Kasalath in the Nipponbare genetic background. NIL- would make annotations based on Nipponbare sequence data
Pup1 had three to four times the P uptake from a highly P- impossible.
fixing volcanic ash soil than Nipponbare with equally high Gene-specific primers have been developed for the pu-
effects on biomass accumulation and grain yield (see Okada tative genes located in the Pup1 interval with the aim to inves-
and Wissuwa, this volume). Further analysis suggested that tigate P-deficiency-induced expression patterns. Nipponbare
roots of NIL-Pup1 were slightly more efficient in extracting and NIL-Pup1 were grown in nutrient solution with three lev-
soil-bound P and that this advantage helped to maintain higher els of P supply: excess P (50 µM), low P (1 µM), and zero P.
root growth rates, with additional benefits for P uptake RNA was isolated from roots and shoots of 5-week-old plants.
(Wissuwa 2004). RT-PCR performed on transcribed RNA samples identified
two candidate genes that showed differential expression de-
pending on genotype and P supply (Fig. 2). Candidate gene #1
M18 50
B
A
C
1
HG28 134
HG29 141
Pup interval
M50
A
227
C
2
M59 273
T20 329 B
A
C
M69 354
3
Fig. 1. Fine mapping of Pup1 to a 195-kb interval on chromosome 12, using se-
lected recombinant lines. The phenotype of lines A and B places Pup1 upward of
marker T20, while line C places Pup1 downward of HG28. The Pup1 interval spans
three BAC clones and gene annotation identified 31 hypothetical genes in that
region.
Pup1
Pup1
Pup1
Pup1
Pup1
N N N N N N
Fig. 2. Expression patterns (RT-PCR) for two Pup1 candidate genes that are
up-regulated by P deficiency in NIL-Pup1. In Nipponbare (N), gene #1 (at
110 bp) was not expressed at all, whereas gene #2 (at 540 bp) was consti-
tutively expressed. Expression analysis was conducted using root RNA
sampled from 5-week-old plants that were grown in nutrient solution under
three levels of P supply (50 µM P, 1 µM P, zero P).
Wrap-up of Session 2
Extensive analysis of the rice genome in the last decade has led since rice is widely cultivated from the equatorial region to about
to the development of valuable tools that have contributed a great 40º N and S. This suggests the existence of several QTLs for rice
deal toward significant progress in rice molecular genetics. With heading date. M. Yano identified at least 14 QTLs for heading
the completion of a high-quality map-based sequence of the rice date and found out that about half of them were photoperiod-
genome, it is expected that rice research will take on a new di- sensitive genes. Using a series of chromosomal substitution lines
mension. As reviewed by T. Sasaki of NIAS, the sequencing effort as a resource for the cloning of each targeted gene of the photo-
was organized in 1998 as a collaboration of 10 countries or re- period-sensitive genes, several, such as Hd1, Hd3a, Hd5, Ehd1,
gions, which made up the International Rice Genome Sequenc- and Lhd4, have been isolated. The difference in sequences among
ing Project (IRGSP). By adopting a common strategy and sharing these genes and their combination in rice varieties could explain
resources, the IRGSP successfully completed the sequencing of the phenotype of heading date at each cultivation area. This strat-
japonica rice variety Nipponbare at the end of 2004. This ses- egy relies on genetic analysis and requires accurate genome se-
sion highlighted research based on information derived from the quence information to facilitate designing DNA markers that can
genome sequence. be used to narrow down the candidate region of a targeted gene.
M. Yano of NIAS described studies on map-based cloning The identification of genes or allelic diversity by reverse genetics
of QTLs for rice heading date or flowering time. Heading date is a methods also requires access to genome sequence information.
crucial characteristic for adaptation to different cultivation areas
Most terrestrial plants, including many important crops, such C4 plant C4-specific gene C3-specific gene
as rice, wheat, soybean, and potato, assimilate CO2 through
the C3 photosynthetic pathway and are classified as C3 plants.
However, some plants, such as maize and sugarcane, possess Enhancer elements
Cell-specific elements
the C4 photosynthetic pathway in addition to the C3 pathway, Modification of kinetics
and these are classified as C4 plants. The C4 pathway acts to
Ancestral
concentrate CO2 at the site of the reactions of the C3 pathway, C4 plant C4-like gene C3-specific gene
and thus inhibits photorespiration. This CO2-concentrating
mechanism, together with modifications of leaf anatomy, en-
Light-responsive elements
ables C4 plants to achieve high photosynthetic capacity and
high water- and nitrogen-use efficiencies. As a consequence, Ancestral gene
the transfer of C4 traits to C3 plants is one strategy being adopted
for improving the photosynthetic performance of C3 plants. Fig. 1. A schematic representation of the evolution of C4-specific
genes.
The C4 pathway consists of three key steps: (1) the ini-
tial fixation of CO2 in the mesophyll cell cytosol by phos-
phoenolpyruvate carboxylase (PEPC) to form a C4 acid, (2) ties of enzymes (Fig. 1; Miyao 2003). This hypothesis is based
decarboxylation of a C4 acid in the bundle-sheath cells to re- on previous observations that the promoters of the maize C4-
lease CO2, and (3) regeneration of the primary CO2 acceptor specific PEPC and PPDK genes can drive high-level expres-
phosphoenolpyruvate in the mesophyll cell chloroplasts by sion of a reporter gene in transgenic rice plants in an organ-
pyruvate, orthophosphate dikinase (PPDK). The decarboxy- specific and mesophyll-cell-specific manner as in maize
lation reaction is catalyzed by one or more of the three en- (Matsuoka et al 1993, 1994). It was found recently that this
zymes, namely, NADP-malic enzyme (NADP-ME), NAD- could also be the case for genes of C4 enzymes located in the
malic enzyme, and phosphoenolpyruvate carboxykinase (PEP- bundle-sheath cells of C4 plants (see Miyao 2003). These find-
CK). The enzymes involved in the C4 pathway are also present ings have important implications that C3 plants possess the
in C3 plants, probably mostly in the photosynthetic mesophyll regulatory factors necessary for high-level expression of C4
cells, although their activities in C3 plants are very low. genes, and, more importantly, that the introduction of the in-
In the leaves of C3 plants, photosynthesis and subsequent tact C4-specific genes would lead to high-level expression of
carbon and nitrogen metabolism proceed mainly in the meso- the C4 enzymes in the leaves of C3 plants.
phyll cells. To have a significant effect on metabolism, any
genes for C4 enzymes (C4 photosynthetic genes) introduced Enzymes located in the mesophyll cells of C4 plants
into C3 plants will need to be expressed at high levels in these The introduction of the intact maize C4-specific gene, which
cells. A strategy based on the evolutionary scenario of C4 genes contained all exons and introns and its own promoter and ter-
has enabled C4 enzymes to be expressed at high levels and in minator sequences, was effective in overproducing PEPC and
desired locations in the leaves of C3 plants (Matsuoka et al PPDK in rice leaves (Ku et al 1999, Fukayama et al 2001).
2001, Miyao 2003). The activities of PEPC and PPDK in transgenic rice leaves
increased up to 110-fold and 40-fold, respectively, that of
nontransformants (Table 1). The level of the maize PEPC and
How to overproduce C4 enzymes PPDK proteins accounted for 12% and 35%, respectively, of
in the mesophyll cells of C3 plants total leaf soluble protein at most. In these transgenic rice plants,
Recent comparative studies have revealed that C3 plants have the levels of transcripts and protein and the activity of the in-
at least two different types of genes homologous to C4 genes, troduced C4 enzyme in the leaves all correlated well with the
one encoding enzymes of “housekeeping” function (C3-spe- copy number of the introduced gene. In addition, the levels of
cific genes) and the other similar to the C4 genes of C4 plants transcripts per copy of the maize C4-specific gene were com-
(C4-like genes), though expression of the latter is very low or parable in both maize and transgenic rice plants and the maize
even undetectable in C3 plants (Miyao 2003). It is now postu- gene was expressed in a similar organ-specific manner. These
lated that C4 genes evolved from a set of preexisting counter- observations suggest that the maize C4-specific genes behave
part genes in ancestral C3 plants, with modifications at the site in a qualitatively and quantitatively similar way in both maize
and in the level of expression in the leaves and kinetic proper- and transgenic rice plants.
aHighest
enzyme activities among the primary transgenic plants are listed. bHighest level of the enzyme protein relative to the level in Zoysia leaves is presented.
MC = mesophyll cells, BSC = bundle-sheath cells, prom = promoter, FL = full-length.
The transcriptional activity of the C4-specific promoter Miyao, unpublished results). This observation is consistent with
cannot be the prime reason for high-level expression, since the anaplerotic role of PEPC in replenishing the tricarboxylic
expression of the full-length cDNA encoding the maize PPDK acid (TCA) cycle with intermediates to meet the demand of
under the control of the promoter of the maize C4-specific carbon skeletons for synthesis of organic acids and amino ac-
PPDK gene or the rice Cab promoter increased the activity of ids. It was found that suppression of root elongation by alumi-
PPDK in rice leaves only up to several-fold that of num was mitigated by the overproduction of PEPC (Miyao-
nontransformants (Table 1; Fukayama et al 2001). It is quite Tokutomi et al 2001). This effect might be ascribable to the
possible that, in addition to the promoter region, the presence enhanced synthesis of organic acids, though exudation of or-
of introns or the terminator sequence, or a combination of both, ganic acids from roots remained unaffected.
is required for high-level expression.
PPDK
Enzymes located in the bundle-sheath cells Physiological impacts were minimal and no changes in photo-
of C4 plants synthetic characteristics were observed even in transgenic rice
Since the intact C4-specific genes for these enzymes would be plants with a 20-fold increase in activity (Fukayama et al 2001).
expressed specifically in the bundle-sheath cells of C3 plants, In general, the reaction of PPDK is reversible, depending on
they cannot be used for overproduction in photosynthetically concentrations of substrates, activators, and inactivators. This
active mesophyll cells of C3 plants (see Miyao 2003). More could be the reason why the overproduction of PPDK does
conventional techniques were applied and have proven suc- not result in significant effects on carbon metabolism in rice
cessful. The expression of the maize C4-specific NADP-ME leaves. Although preliminary, we found that overproduction
cDNA under the control of the rice Cab promoter increased of PPDK increased the grain number per plant and 1,000-grain
the activity of NADP-ME in rice leaves up to 30-fold that of weight of rice plants (not shown). It seems possible that the
nontransformants and the level of the NADP-ME protein up overproduction of PPDK might enhance the C4-like pathway,
to several % of total leaf soluble protein (Tsuchida et al 2001). which is operating in rice spikelets.
Such high-level expression was unique to the cDNA for the
C4-specific NADP-ME, and the expression of the cDNA for NADP-ME
the rice C3-specific isoform under the control of the same pro- Overproduction of the maize C4-specific NADP-ME led to an
moter increased the activity only some-fold (Tsuchida et al increase in the NADPH/NADP ratio in the chloroplast under
2001). The Zoysia C4-specific PEP-CK was also overproduced illumination and suppressed photorespiration, rendering pho-
by the introduction of a cDNA construct in rice leaves (M. tosynthesis more susceptible to photoinhibition (Tsuchida et
Miyao et al, unpublished results). al 2001). The C4-specific NADP-ME has a higher Vm value,
lower Km values for substrates, and higher optimum pH than
the C3-specific isoform. Such features are suitable for strict
Physiological implants of C4 enzymes overproduced regulation of the enzyme activity in the chloroplasts of bundle-
in rice plants sheath cells of C4 plants but they allow the enzyme to continue
PEPC operating in the leaves of C3 plants even when serious damage
Although PEPC is an enzyme involved in the initial fixation occurs.
of CO2 in the C4 pathway, its overproduction did not improve
photosynthetic efficiency of rice plants. Rather, it stimulated
respiration in the light and subsequent synthesis of organic
acids and amino acids (Fukayama et al 2003, T. Tamai and M.
The uptake and translocation of metal nutrients in plants are nongraminaceous plants (Mori and Nishizawa 1987, Shojima
essential for plant growth and, since plants are the primary et al 1990). In graminaceous plants, the amino group of NA is
source of food for humans, the nutritional value of plants is of transferred by nicotianamine aminotransferase (NAAT), and
central importance to human health. The most widespread hu- the resultant keto form is reduced to deoxymugineic acid
man nutritional problem in the world is iron (Fe) deficiency (DMA) and other mugineic acid family phytosiderophores
(WHO 2003). Increasing the ability of plants to provide higher (MAs). Rice secretes DMA, the first member of MAs. A re-
levels of minerals, such as Fe, will have a dramatic impact on cent study showed that transgenic rice containing NAAT genes
human health. Increasing the Fe uptake from soils is a prereq- from barley is more tolerant than nontransgenic individuals of
uisite to increase the amount of Fe in the edible parts of plants. low-Fe availability in calcareous soil since transgenic rice
This can be aided by identifying the transporters involved in plants secrete higher amounts of MAs than do wild-type plants
the translocation of Fe within plants. (Takahashi et al 2001).
Despite its abundance in soils, much Fe is present in the All other plant groups use a reduction strategy composed
insoluble Fe(III) form and not readily available to plants. There- of Fe(III) reductase and Fe(II) transporter. They do not pro-
fore, plants have evolved two distinct uptake strategies duce or release phytosiderophores. However, nongraminaceous
(Marschner et al 1986). Graminaceous plants use a chelation plants synthesize NA, which is the biosynthetic precursor of
strategy and release low-molecular-weight compounds, MAs and structurally similar. NA chelates metal cations, in-
phytosiderophores, from their roots to chelate Fe(III) in the cluding Fe(II) and Fe(III). Unlike MAs, NA is not secreted,
soil (Takagi 1976). Three molecules of S-adenosyl-methion- and is thought to have a role in metal homeostasis in
ine are combined with nicotianamine synthase (NAS) to form nongraminaceous plants. Since Higuchi et al (1999) isolated
nicotianamine (NA) in both graminaceous and the first NAS gene from Hordeum vulgare, NAS genes have
OsYSL10
OsYSL5 OsYSL11
OsYSL6
OsYSL13
OsYSL6
OsYSL13
OsYSL14
OsYSL12 OsYSL14
OsYSL7 OsYSL15
OsYSL17
rRNA
OsYSL3 OsYSL4
0.1
Fig. 1. (A) The unrooted phylogenic tree for the 18 OsYSL amino acid sequences. ZmYS1 is the Fe(III)-phytosiderophore
transporter from maize. (B) RNA gel-blot analysis of OsYSL transcripts in Fe-sufficient and Fe-deficient rice plants. Only the
expression of OsYSL2, OsYSL6, OsYSL13, OsYSL14, OsYSL15, and OsYSL16 was detected in these plants. The bottom panel
shows ethidium bromide-stained rRNA as a loading control. +Fe: Fe-sufficient plants, –Fe: Fe-deficient plants.
been isolated from barley again, and from tomato, Arabidopsis OsYSL2 is a transporter involved in the long-distance trans-
thaliana, rice, and maize. In graminaceous plants, the Fe(III)- port of Fe in rice, but is not a rice Fe(III)-phytosiderophore
phytosiderophore complex is taken up by a specific transporter for iron uptake from the soil. Based on the nucleotide se-
at the root surface. Maize yellow stripe 1 (YS1) is the gene quence, OsYSL2 was predicted to encode a polypeptide of 674
encoding the Fe(III)-phytosiderophore transporter (Curie et al amino acids containing 14 putative transmembrane domains.
2001). Subcellular localization of OsYSL2 and green fluorescent pro-
Our search for YS1 homologues in the Oryza sativa L. tein (GFP), together with the presence of the putative trans-
subsp. japonica (cv. Nipponbare) rice genomic database iden- membrane domains, showed that OsYSL2 is a transporter that
tified 18 putative YS1-like genes (OsYSL) that exhibited 36% is localized in the plasma membrane.
to 76% sequence similarity to YS1 (Koike et al 2004) (Fig. The cell-type specificity of OsYSL2 expression was in-
1A). Further, we examined whether Fe regulates the expres- vestigated using promoter (1.5 kb):β-glucuronidase (GUS)
sion of these genes in rice (Fig. 1B). Using Northern blot analy- analysis. Although Northern blot analysis did not detect OsYSL2
sis, the transcripts of six genes were detected in roots or leaves transcripts in the roots, GUS staining resulting from OsYSL2
of Fe-sufficient or Fe-deficient rice plants. OsYSL6 was con- promoter activity showed a dotted pattern in phloem compan-
stitutively expressed in both the roots and leaves of either Fe- ion cells in the central cylinder of Fe-sufficient roots. The ac-
sufficient or Fe-deficient plants, with Fe deficiency slightly tivity in companion cells increased in Fe-deficient roots, but
reducing its expression. OsYSL13 was preferentially expressed no GUS staining was observed in epidermal or cortical cells,
in the shoots, and Fe deficiency reduced its expression. even in Fe-deficient roots. This absence of OsYSL2 promoter
OsYSL14 was expressed in both roots and shoots, and Fe sta- activity in the root epidermis or cortex supports the contention
tus had no effect on its expression. OsYSL15 was expressed that OsYSL2 is not an Fe(III)-phytosiderophore transporter
only in roots and the expression was highly induced by Fe involved in the uptake of Fe from the soil. GUS staining was
deficiency. OsYSL16 was expressed in both roots and shoots, observed in phloem cells of the vascular bundles of leaves and
and the expression in roots was slightly increased by Fe defi- leaf sheaths of Fe-sufficient rice. The phloem-specific expres-
ciency. We could not detect the transcripts of other members sion of the OsYSL2 promoter suggested that OsYSL2 is involved
of OsYSLs in these plants by Northern blot analysis. in the phloem transport of Fe. In Fe-deficient leaves, the
The high expression of OsYSL2 in the leaves but not in OsYSL2 promoter was active in all tissues, with particularly
the roots of Fe-deficient plants led us to hypothesize that strong GUS activity evident in companion cells. Since the re-
Drought, high salinity, and freezing are environmental condi- independent pathways, a cis-acting element with nine base pairs
tions that have adverse effects on plant growth and crop pro- (TACCGACAT), named the dehydration-responsive element
ductivity. Plants respond and adapt to these stresses to survive (DRE), is involved in dehydration, and high-salt and low-tem-
under stress conditions at the molecular and cellular levels as perature–induced gene expression (Yamaguchi-Shinozaki and
well as at the physiological and biochemical levels. Expres- Shinozaki 1994).
sion of a variety of genes is induced by these stresses in di- Two cDNA clones that encode DRE-binding proteins,
verse plants. The products of these genes are thought to func- DREB1A/CBF3 and DREB2A, were isolated by using yeast
tion not only in stress tolerance but also in the regulation of one-hybrid screening (Liu et al 1998). cDNA clones encoding
gene expression and signal transduction in stress responses two DREB1A homologs (named DREB1B/CBF1 and
(Shinozaki et al 2003). DREB1C/CBF2) and one DREB2A homolog (DREB2B) were
Genetic engineering is useful for improving the stress also isolated (Liu et al 1998). Expression of the DREB1A gene
tolerance of plants. Several different approaches have been and its two homologs was induced by low-temperature stress,
attempted to improve the stress tolerance of plants by gene whereas expression of the DREB2A gene and its homolog was
transfer (Shinozaki et al 2003). The genes selected for trans- induced by dehydration. These results indicate that two inde-
formation were those involved in encoding enzymes required pendent families of DREB proteins, DREB1 and DREB2, func-
for the biosynthesis of various osmoprotectants. Other genes tion as trans-acting factors in two separate signal transduction
that have been selected for transformation include those that pathways under low-temperature and dehydration conditions,
encoded enzymes for modifying membrane lipids, LEA pro- respectively.
tein, and detoxification enzymes. In all these experiments, a Overexpression of the cDNA encoding DREB1A in
single gene for a protective protein or an enzyme was transgenic Arabidopsis plants activated the expression of many
overexpressed under the control of the constitutive 35S cauli- of these stress-tolerance genes under normal growing condi-
flower mosaic virus (CaMV) promoter in transgenic plants, tions and resulted in improved tolerance of drought, salt load-
although several genes have been shown to function in envi- ing, and freezing (Liu et al 1998, Kasuga et al 1999). To iden-
ronmental stress tolerance and response. The genes encoding tify target stress-inducible genes of DREB1A, we performed
protein factors that regulate gene expression and signal trans- microarray analysis. We searched for downstream genes in
duction and that function in stress responses may be useful for transgenic plants overexpressing DREB1A using the full-length
improving the tolerance of plants of stresses by gene transfer cDNA microarray and GeneChip array. We identified more
as they can regulate many stress-inducible genes involved in than 40 genes as the DREB1A downstream genes (Seki et al
stress tolerance. 2001, Maruyama et al 2004). The products of these genes were
not only proteins known to function against stress but also
protein factors involved in further regulation of signal trans-
DREB/DRE regulon in Arabidopsis
duction and gene expression in response to stress.
To understand the molecular process of signal transduction However, use of the strong constitutive CaMV 35S pro-
from the perception of the stresses to gene expression, we iso- moter to drive expression of DREB1A also resulted in severe
lated many environmental stress-inducible genes in Arabidopsis growth retardation under normal growing conditions. In con-
thaliana. The products of these genes are thought to function trast, expression of DREB1A from the stress-inducible rd29A
not only in stress tolerance but also in the regulation of gene promoter gave rise to minimal effects on plant growth while
expression and in signal transduction in the stress response. providing an even greater tolerance of stress conditions than
These gene products can be classified into two groups. The did expression of the gene from the CaMV 35S promoter
first group includes proteins that probably function in protect- (Kasuga et al 1999). We also analyzed the effect of
ing cells from dehydration. The second group of gene prod- overexpression of DREB1A on stress tolerance and growth
ucts contains protein factors that are involved in further regu- retardation in transgenic tobacco plants using the constitutive
lation of gene expression and signal transduction and that func- CaMV 35S promoter and the stress-inducible rd29A promoter
tion in stress response (Shinozaki et al 2003). (Kasuga et al 2004). Overexpression of DREB1A improved
We analyzed the expression of the drought-inducible drought- and low-temperature stress tolerance in tobacco. The
genes and identified at least four independent regulatory sys- stress-inducible rd29A promoter minimized the negative ef-
tems in drought-responsive gene expression: two are ABA- fects on plant growth in tobacco. These results indicate that
dependent and two are ABA-independent. In one of the ABA- the stress-inducible rd29A promoter and the DREB1A gene
are quite useful for improving drought, salt, and freezing stress On the other hand, we carried out microarray analysis
tolerance not only in Arabidopsis but also in other kinds of using rice cDNAs to identify stress-inducible genes in response
dicot plants. to dehydration, high salt, and low temperature in rice (Rabbani
et al 2003). We confirmed stress-inducible expression of the
candidate genes selected by microarray analysis using RNA
DREB/DRE regulon in rice gel-blot analysis, and finally identified a total of 73 genes as
Rice, an important crop plant, has now emerged as an ideal stress-inducible genes in rice (Fig. 1). Among them, 36, 62,
model species for the study of crop genes because of its com- 57, and 43 genes were induced by cold, drought, high salinity,
mercial value, relatively small genome size (approx. 430 Mb), and ABA, respectively. The products of the identified rice genes
diploid origin (2x=24), and close relationship with other im- are classified into functional proteins and regulatory proteins
portant cereal crops. It is important to analyze the DRE/DREB like those of Arabidopsis (Rabbani et al 2003). Comparative
(CRT/CBF) regulon in rice to understand the molecular mecha- analysis of stress-inducible genes in Arabidopsis and rice re-
nisms of stress tolerance and produce monocots with higher vealed a considerable level of similarity in stress responses
stress tolerance by gene transfer. We isolated five cDNAs for between the two genomes at a molecular level. Among 73 iden-
DREB homologs, OsDREB1A, 1B, 1C, 1D, and 2A, and ana- tified stress-inducible genes in rice, 51 have already been re-
lyzed their functions in rice (Dubouzet et al 2003). The ported in Arabidopsis with a similar function or gene name.
OsDREB proteins were specifically bound to the DRE se- These results indicate that rice has many stress-inducible genes
quence and activated the transcription of a GUS reporter gene in common with Arabidopsis, even though these two plants
driven by the DRE sequence in rice protoplasts. Expression of evolved separately a million years ago. Common stress-induc-
the OsDREB1A and 1B genes was induced by low-tempera- ible genes include LEA proteins, antifreeze proteins, detoxifi-
ture stress, whereas the expression of the OsDREB2A gene cation enzymes, and so on. All these genes are up-regulated in
was induced by dehydration. Recently, we searched the rice response to at least one of the abiotic stresses in rice and re-
genome database for amino acid sequences of the ERF/AP2 ported as stress-inducible genes in Arabidopsis. Transcriptome
domains of DREB and ERF, and found an additional five analysis of rice also revealed novel stress-inducible genes,
DREB1A homologs, OsDREB1E, OsDREB1F, OsDREB1G, suggesting some differences between Arabidopsis and rice in
OsDREB1H, and OsDREB1I, and three DREB2A homologs, their response to stress.
OsDREB2B, OsDREB2C, and OsDREB2D, in rice (Table 1). We analyzed expression of the stress-inducible genes
We generated transgenic Arabidopsis overexpressing the identified by the cDNA microarray and Northern blot analy-
OsDREB1A cDNA. The rice OsDREB1A protein exhibits a ses in the transgenic rice plants overexpressing the OsDREB1A
function similar to that of the Arabidopsis DREB1A protein in or DREB1C cDNAs and detected overexpression of some
transgenic Arabidopsis plants (Dubouzet et al 2003). We also stress-inducible genes. The promoter regions of these genes
generated transgenic rice plants overexpressing the rice contain the DRE sequences. These results indicate that there
OsDREB or Arabidopsis DREB cDNAs and analyzed their is a DRE/DREB regulon in the stress-responsible signal trans-
tolerance of drought, high-salt, and cold stress. These transgenic duction in rice similar to that in Arabidopsis. The DREB/DRE
plants showed high tolerance of these stresses. regulon should be useful for producing transgenic dicots and
Fig. 1. RNA gel-blot analysis of stress-inducible genes. Each lane was loaded with 10 µg of total RNA isolated from 2-wk-old rice seedlings that were exposed to H2O, dehydration,
250 mM NaCl, 100 µM ABA, and 4 °C cold treatment for 1, 2, 5, 10, and 24 h. RNA was analyzed by gel-blot hybridization with gene-specific probes of selected stress-inducible
clones by rice cDNA microarray. Stress-inducible clones were classified into various groups on the basis of their expression patterns in RNA gel-blot analysis under each stress
treatment.
monocots with higher tolerance of drought, high salinity, and/ Rabbani MA, Maruyama K, Abe H, Khan MA, Katsura K, Ito Y,
or cold stress. Yoshiwara K, Seki M, Shinozaki K, Yamaguchi-Shinozaki K.
2003. Monitoring expression profiles of rice (Oryza sativa
L.) genes under cold, drought and high-salinity stresses, and
References ABA application using both cDNA microarray and RNA gel
Dubouzet JG, Sakuma Y, Ito Y, Kasuga M, Dubouzet EG, Miura S, blot analyses. Plant Physiol. 133:1755-1767.
Seki M, Shinozaki K, Yamaguchi-Shinozaki K. 2003. Seki M, Narusaka M, Abe H, Kasuga M, Yamaguchi-Shinozaki K,
OsDREB genes in rice, Oryza sativa L., encode transcription Carninci P, Hayashizaki Y, Shinozaki K. 2001. Monitoring
activators that function in drought-, high-salt- and cold-re- the expression pattern of 1300 Arabidopsis genes under
sponsive gene expression. Plant J. 33:751-763. drought and cold stresses using a full-length cDNA microarray.
Kasuga M, Liu Q, Miura S, Yamaguchi-Shinozaki K, Shinozaki K. Plant Cell 13:61-72.
1999. Improving plant drought, salt, and freezing tolerance Shinozaki K, Yamaguchi-Shinozaki K, Seki M. 2003. Regulatory
by gene transfer of a single stress-inducible transcription fac- network of gene expression in the drought and cold stress
tor. Nature Biotechnol. 17:287-291. responses. Curr. Opin. Plant Biol. 6:410-417.
Kasuga M, Miura S, Shinozaki K, Yamaguchi-Shinozaki K. 2004. A Yamaguchi-Shinozaki K, Shinozaki K. 1994. A novel cis-acting el-
combination of the Arabidopsis DREB1A gene and stress-in- ement in an Arabidopsis gene is involved in responsiveness
ducible rd29A promoter improved drought- and low-tempera- to drought, low-temperature, or high-salt stress. Plant Cell
ture stress tolerance in tobacco by gene transfer. Plant Cell 6:251-264.
Physiol. 45:346-350.
Liu Q, Kasuga M, Sakuma Y, Abe H, Miura S, Yamaguchi-Shinozaki
Notes
K, Shinozaki K. 1998. Two transcription factors, DREB1 and
DREB2, with an EREBP/AP2 DNA binding domain separate Authors’ addresses: Kazuko Yamaguchi-Shinozaki, Biological Re-
two cellular signal transduction pathways in drought- and low- sources Division, Japan International Research Center for
temperature-responsive gene expression, respectively, in Agricultural Sciences (JIRCAS), Tsukuba, Ibaraki; Labora-
Arabidopsis. Plant Cell 10:1391-1406. tory of Plant Molecular Physiology, Graduate School of Agri-
Maruyama K, Sakuma Y, Kasuga M, Ito Y, Seki M, Goda H, Shimada cultural and Life Sciences, The University of Tokyo, Tokyo,
Y, Yoshida S, Shinozaki K, Yamaguchi-Shinozaki K. 2004. CREST, JST, Japan; Kazuo Shinozaki, Laboratory of Plant
Identification of cold-inducible downstream genes of the Molecular Biology, RIKEN Tsukuba Institute, Tsukuba,
Arabidopsis DREB/CBF3 transcriptional factor using two Ibaraki, CREST, JST, Japan, e-mail:
microarray systems. Plant. J 38:982-993. kazukoys@jircas.affrc.go.jp.
Environmental stresses exert a critical influence on rice yields tant. To develop new domestic gene technologies, we hit upon
and pathogen attacks are sometimes the most devastating bi- the idea of incorporating disease-resistance genes from veg-
otic stress. In a humid temperate climate, fungal rice blast etables, employing a selection technique that uses rice genes,
caused by Magnaporthe grisea is the most serious problem. and keeping the transgenes from functioning in the edible por-
Bacterial leaf blight caused by Xanthomonas oryzae is a seri- tions of transgenic rice. Based on this idea, we successfully
ous disease of rice in subtropical and tropical areas. The en- developed new transgenic rice strains with multiple disease
hancement of disease resistance in rice has made a great con- resistance by integrating new domestic gene technologies, in-
tribution toward increasing the productivity of rice and de- cluding new selection marker genes from rice, defensin genes
creasing the application of pesticides, which can negatively from vegetables, which are highly effective in conferring dis-
affect human health and the environment. Genetic engineering ease resistance, and new callus-specific and green-tissue-spe-
has given us opportunities and challenges for transgenic rice cific promoters from rice. We confirmed that the transgenes
for improving disease resistance, based on recent advances in are specifically expressed in the callus or green tissue, whereas
molecular biology. Cloning of several disease-resistance genes expression is suppressed in rice grains.
is providing insights into their function and how they protect
plants against pathogen attacks. Defensin genes
Defensins are antimicrobial proteins found in a wide variety
of organisms and that exhibit powerful antimicrobial activity
Developing domestic gene technologies against fungi and bacteria. In recent years, progress has been
Plant gene engineering meant to bring transgenic rice into prac- achieved in analyzing the characteristics of their antimicro-
tical use must address environmental safety and consumer con- bial activity against plant pathogens. Studies of the general
cerns, and thus the development of new technologies is impor- mechanism on antimicrobial activity have been reported and
mutated ALS
(mALS) Leu Ile
the detailed mechanisms of antimicrobial activity on defensin and these techniques suppress the accumulation of transgene
peptide have been studied (Thevissen et al 2004). Plant proteins in rice grains.
defensins have four disulfide bonds, distinguishing them struc-
turally from the defensins of insects and mammals, which have New selection markers
three disulfide bonds (Fant et al 1998). Acetolactate synthase (ALS) is one of the enzymes that is re-
Mature plant defensin proteins comprise 45 to 54 amino lated to metabolism, and it is widely found in plants. It is one
acid residues, and are small, basic proteins rich in cysteine. of the target enzymes for herbicides. Cells that expressed the
Their expression is observed in Cruciferae plants, as well as ALS gene suffer growth inhibition in the existence of bispyribac
in sorghum, dahlias, and many other plants. Although their sodium (BS), but a natural mutation with resistance to this
amino acid structures are diverse, cysteine residues are well herbicidal effect has been found among cultured rice plant cells.
preserved, and the alpha helix motif stabilized by cysteine is An investigation into the reason for this resistance revealed
common to these defensins. Cruciferae defensin proteins, that inhibition of enzyme activity changed because of the re-
which belong to the group exhibiting the strongest antimicro- placement of amino acid sequences at two locations in the ALS
bial activity, are expressed in places such as seed surfaces. protein (mALS) (Fig.1). Formerly, the marker genes used in
They exist as gene clusters, and their expression characteris- the development of transgenic plants involved mainly antibi-
tics and structures differ. One characteristic is that slight amino otic resistance genes obtained from microorganisms or other
acid sequential differences change the strength of antibacte- organisms. This new selection marker gene (mALS) can be used
rial activity and effectiveness against pathogens (De Samblanx in place of antibiotic resistance selection marker genes.
et al 1997). Defensin genes exist in commonly consumed
Cruciferae vegetables. They exhibit antibiotic activity against
Transgenic rice that integrates domestic gene technologies
a wide variety of pathogens such as fungi and bacteria, but
rice plants do not have them. We isolated and determined the Transgenic rice strains developed with multiple disease-resis-
structures of defensin gene clusters from eight kinds of veg- tance genes are characterized by the integration of original
etables, and found that adding defensin proteins suppressed new technologies, incorporating genes with high disease re-
the growth of rice blast hyphae in vitro (Kawata et al 2004). sistance from vegetables. In addition to mALS ligated down-
stream from the rice callus-specific promoter, and a defensin
New promoters gene ligated downstream from the rice green-leaf-specific pro-
Developing transgenic plants endowed with practical traits such moter, we constructed a binary vector incorporating an expres-
as disease resistance requires taking steps so that transgenes sion cassette consisting of the rice P10 terminator ligated down-
are expressed only in the necessary places, and not in the rice stream from each gene (Fig. 2). Transformation was accom-
grains. We isolated the green-leaf-specific promoter, which was plished by using seeds of rice variety Dontokoi. Estimation of
expressed in stems and leaves, and the callus-specific promoter. the rice blast resistance and bacterial leaf blight allowed us to
These promoters were not expressed in the rice grains, includ- obtain transgenic plants that had been endowed with resistance
ing the germ. Moreover, we modified the isolated promoters, to these pathogens. We confirmed that this resistance is stably
such as adding restriction enzyme recognition sites, thereby inherited (Kawata et al 2003). This positively confirmed that
making it possible to easily construct the vectors used for trans- the defensin gene makes it possible to develop transgenic rice
formation. These new technologies make it possible to induce strains with high resistance to fungi (the rice blast fungus) and
the functioning of “disease-resistance genes” and “marker genes bacteria (bacterial leaf blight). We also found that the resis-
that can select transformed cells” at the necessary time for each, tance conferred by this gene is not specific to a particular race
Fig. 2. Vector construction for the development of transgenic rice by making use of
domestic gene technologies. pCsp = callus-specific promoter from rice, mALS = mu-
tated acetolactate synthase gene as selection marker, P10 = rice 10K prolamine
terminator, pLsp = green-tissue-specific promoter from rice, Def = defensin gene as
disease-resistance gene.
of rice blast fungus. The transgenic plants did not have any References
significant difference with a nontransgenic variety in traits such
De Samblanx GW, Goderis IJ, Thevissen K, Raemaekers R, Fant F,
as leaf shape or fertility, and we were able to obtain progeny Borremans F, Acland DP, Osborn RW, Patel S, Broekaert WF.
seeds. We analyzed the expression specificity of the callus- 1997. Mutational analysis of a plant defensin from radish
specific promoter and the green-leaf-specific promoter, and (Raphanus sativus L.) reveals two adjacent sites important
confirmed that expression was suppressed in the rice grains. for antifungal activity. J. Biol. Chem. 272(2):1171-1179.
We investigated whether defensin has any significant homol- Fant F, Vranken W, Broekaert W, Borremans F. 1998. Determination
ogy between known allergens and toxic proteins. With refer- of the three dimensional solution structure of Raphanus sativus
ence to the Report of a Joint FAO/WHO Expert Consultation antifungal protein 1 by 1H NMR. J. Mol. Biol. 279:257-270.
(2001), homology was assessed as significant when either or Kawata M, Nakajima T, Yamamoto T, Mori K, Oikawa T, Fukumoto
both of the following were true: (1) overall homology is 35% F, Kuroda S. 2003. Genetic engineering for disease resistance
in rice (Oryza sativa L.) using antimicrobial peptides. JARQ
or more and (2) eight or more neighboring amino acid se-
37(2):71-76.
quences match. We confirmed that defensin peptides have no
Kawata M, Nakajima T, Mori K, Oikawa T, Kuroda S. 2004. Ge-
significant amino acid sequence homology either with known netic engineering for blast disease resistance in rice, using a
allergen proteins or with toxic proteins. plant defensin gene from Brassica species. Proceedings of
the 3rd IRBC Congress. (In press.)
Creating new demand Saido-Sakanaka H, Ishibashi J, Sagisaka A, Momotani E, Yamakawa
M. 1999. Synthesis and characterization of bactericidal
There is potential consumer demand for natural antimicrobial oligopeptides designed on the basis of an insect anti-bacterial
agents that offer eco-friendliness and excellent safety (Zasloff peptide. Biochem. J. 338:29-33.
2002). By focusing on the defensins, which exhibit broad an- Thevissen K, Warnecke DC, Francois IE, Leipelt M, Heinz E, Ott C,
timicrobial activity, as new natural antimicrobial agents, we Zahringer U, Thomma BP, Ferket KK, Cammue BP. 2004.
Defensins from insects and plants interact with fungal
expect that new antimicrobial agents will be developed, lead-
glucosylceramides. J. Biol. Chem. 279(6):3900-3905.
ing to the creation of new demand among consumers. It has
Zasloff M. 2002. Antimicrobial peptides of multicellular organisms.
been shown that animal and insect defensins inhibit the growth Nature 415:389-395.
of human disease-related bacteria (Saido-Sakanaka et al 1999).
This indicates the possibility that these defensins can be used
for the prevention and treatment of human diseases and food Notes
poisoning. Analytical research on plant defensins has also Author’s address: Hokuriku Department of Rice Research, National
broadened in scope from only targeting plant pathogens, and Agriculture Research Center, Inada, Joetsu, Niigata 943-0193,
there are hopes for the elucidation of pharmacological charac- Japan, and Graduate School of Natural Science and Technol-
teristics through antimicrobial activity spectrum analyses of ogy, Niigata University, Ikarashi, Niigata 950-2181, Japan.
diverse pathogenic organisms.
Deficiencies of vitamin A, iron, and zinc are widespread in year because of vitamin-A deficiency (VAD). Diet is the only
Asia and in developing countries, where the main diet is plant- source of vitamin A in mammals since they cannot manufac-
based and the staple food is rice. Rice provides 40–70% of the ture it on their own. Most of the dietary vitamin A is of plant-
total calories consumed in developing Asian countries. In food origin in the form of provitamin A that is converted to
Southeast Asia, an estimated 250,000 people go blind each vitamin A in the body (Sivakumar 1998). So far, none of the
0.040 0.040
0.035 0.035 lut
0.030 0.030
0.025 b-crt
0.025
0.020
0.020
0.015
0.010 0.015
0.005 0.010
0.000 0.005
0.005 0.000
0.010
10 20 30 40 50 60 70 5 10 15 20 25 30
0.040 0.040
b-crt b-crt
0.035 0.035
0.030 0.030
0.025 0.025
0.020 0.020
lut
0.015 0.015
0.010 0.010 lut
0.005 0.005
0.000 0.000
5 10 15 20 25 30 5 10 15 20 25 30
Minutes Minutes
Fig. 1. Carotenoid profile for selected rice cultivars under both polished and unpolished conditions.
Curves in red = polished seeds. Curves in blue = unpolished seeds. Curves in green = control seeds.
KDML 105 and Ase Pindjau (from Thailand and Indonesia) are nontransgenic seeds, whereas IR64 and
BR29 are genetically engineered Golden indica rice (lut = lutein, β -crt = β -carotene).
Higher plants are an attractive host for the production of re- scription and posttranscription, including translation, mRNA
combinant proteins for pharmaceutical and nutraceutical ap- stability, transport, protein stability, and folding, among oth-
plications. Plant production offers many advantages, the most ers. It has been found that the transcription level primarily
prominent of which are cost savings relative to animal determines the expression level. Therefore, screening for strong
transgenic systems, easy control of the production scale, and promoters from genes expressed in maturing seed has been
freedom from animal-derived pathogens. When recombinant carried out first. Promoter activities of more than 20 genes
proteins are expressed in seed under the control of a seed- were assayed in stable transgenic rice seed by fusing their pro-
specific promoter, they are highly and stably accumulated in a moters to the GUS reporter gene, and then introducing their
specific organelle designated as a protein body. Taking advan- chimeric genes into the rice genome by Agrobacterium-medi-
tage of these properties, the seed has been used as an ideal ated transformation. From the results of this promoter analy-
bio-reactor for the production of recombinant proteins. sis, rice seed promoters are classified into several types based
Rice is one of the most important crops and food re- on their expression patterns in seed. It was shown that glute-
sources in the world. Rice seed has several advantages over lin, 26-kDa globulin, and 10-kDa prolamin promoters exhib-
the other cereal crops, such as easier storage and processing, ited strong promoter activities in this order and different ex-
greater biomass, and lower production costs. In addition, a pression patterns in the endosperm (Qu and Takaiwa 2004).
transformation system of rice has been established and the Prolamin, glutelin, and globulin promoters directed expres-
whole genome sequence, including mitochondria and chloro- sion in the outer layer, subaleurone layer, and inner starchy
plasts, is now available. Therefore, rice seed can be used as a endosperm, respectively. Therefore, we have selected and used
good platform for producing recombinant proteins. the glutelin, 26-kDa globulin, and 10-kDa prolamin promot-
ers for the expression of foreign proteins in transgenic rice
seed dependent upon that purpose.
The rice seed production system Irrespective of high transcripts, it has been occasionally
For use as a bio-reactor for the production of recombinant pro- observed that final products were at a very low level. In this
teins or a direct delivery system of bioactive peptides or vac- case, the space for accumulation of recombinant proteins or
cines, establishment of a high expression system is required. subcellular localization as well as codon usage or mRNA sta-
Final accumulation levels are regulated at several steps of tran- bility is critical. Low-storage-protein mutants such as LGC-1
60
50
40
30
20
10
b-conglycinin
Ubiquitin
10 kDa Pro
13 kDa Pro
16 kDa Pro
GOGAT
GluB-2
Glb-1
AGPase
REG-2
Ole18
AlaAT
PPDK
SBE1
GluB-4
2.3 kB GluB-1
1.3 kB GluB-1
Fig. 1. GUS activities expressed by various promoters in maturing seeds at 17 days after flowering.
GUS activity is expressed as pmol (4 MU min–1 µg–1 protein).
will become good hosts, giving rise to a high accumulation of accumulated than A3B4 in transgenic rice. When both glycinins
foreign proteins, because they can provide ample space for were expressed together, the A3B4 level increased about three-
accumulation. It has been confirmed by crossing between fold more than the level obtained by the expression of A3B4
transgenic rice and a mutant that the transgene product was alone, thus indicating a requirement of assembly between dif-
enhanced more than twofold in a low storage background (Tada ferent subunits for the stable accumulation of A3B4. In the
et al 2003). LGC background, soybean glycinin was accumulated as a major
storage protein, accounting for 30% to 40% of total rice seed
protein. The glycinin level also increased more than twice com-
Transgenic rice accumulating soybean glycinin
pared with that of the normal background. Glycinin was ex-
Soybean protein has a function to lower the cholesterol level tracted as salt-soluble globulin and acid-soluble glutelin frac-
in human serum and is one of the best plant food proteins in tions, thus suggesting that hybrid glycinin-glutelin oligomers
terms of nutritional quality. The major storage proteins of soy- were also formed in transgenic rice seed. When the lysine and
bean are glycinin and β-conglycinin. Glycinin is superior to protein levels in the seed were examined and compared be-
β-conglycinin in nutritional value as well as in functional prop- tween a transgenic line expressing both glycinins and the low-
erties. Soybean proteins are deficient in sulfur-containing amino glutelin mutant background line, they were remarkably en-
acids and rich in lysine, and rice proteins are the opposite. hanced from 0.23 to 0.41 g per 100 g of seed and from 6.9 to
Therefore, if soybean glycinin were highly accumulated in rice 10.9 g per 100 g of seed, respectively.
seeds, an improvement of nutritional value and endowment of
functional properties such as gel-formation and emulsifying
abilities would be expected. We previously generated
Transgenic rice accumulating bioactive peptides
transgenic rice accumulating soybean glycinin A1aB1b It has been observed that some bioactive peptides derived from
(Katsube et al 1999). In this study, two representative types of digests of food proteins are active after oral administration
glycinin (A1aB1b and A3B4) belonging to different subclasses even though specific activities are not high compared with the
were expressed in the endosperm tissue under the control of endogenous bioactive peptides. Such peptides are expected to
glutelin GluB-1 and 26-kDa globulin promoters, respectively. be effective in preventing lifestyle-related diseases such as
Furthermore, to enhance their accumulation levels, they were hypertension, high cholesterol, and diabetes, among others.
transferred into the low-storage-protein background of LGC- Biological activities of peptides could be strengthened by re-
1 by crossing. It was shown that A1aB1b was much more highly placement of amino acid residues. Taking advantage of these
Western
B CBB staining analysis
NT TF NT TF
kDa kDa
75 75
50 50
Glutelin
37 (acidic) 37
25 Glutelin 25
(basic)
15 Prolamin 15
7 crp 7 Crp
10 10
Fig. 2. (A) Vector construct used to generate transgenic rice expressing the 7 crp gene in
endosperm tissue. (B) Accumulation of the 7 Crp peptide in seeds of transgenic rice plants.
Total protein was extracted from a seed of nontransformant (NT) or transgenic rice (TF:
pGPTV-HPT-GluB-1pro sig/7Crp+KDEL #10), separated by 15% SDS-PAGE, and then stained
with CBB or immunoblotted with anti-7 Crp antibody.
Chimeraplasts, which are also called chimeric RNA/DNA oli- gene in maize (Zhu et al 1999, 2000) and an acetolactate syn-
gonucleotides, have been reported to cause site-specific base thase (ALS) gene in tobacco (Beetham et al 1999). Both ALS
changes in chromosomal targets in mammalian cells (see Hohn and AHAS are the first common enzymes in the biosynthetic
and Puchta 1999, Oh and May 2001, Graham and Dickson pathway of the branched-chain amino acids leucine, isoleusine,
2002 for reviews). Chimeraplasts consist of 68 synthesized and valine. Mutations of certain amino acids in these proteins
oligonucleotides that have a DNA “mutator” region of five have been known to confer resistance to ALS-inhibiting herbi-
nucleotides complementary to the target site flanked by 2′-O- cides (Shimizu et al 2002). The efficiency of gene conversion
methyl RNA bridges of 8–12 nucleotides each. This DNA mu- mediated by chimeraplasts in maize was estimated to be 10–4.
tator region can cause a site-specific base change in the en- In the same manner, we demonstrated that chimeraplast-
dogenous gene. The first uses of chimeraplasts causing site- directed gene targeting is feasible in rice (Okuzaki and
specific base changes in plant cells were reported in maize Toriyama 2004). In our current study, we further confirmed
(Zhu et al 1999) and tobacco (Beetham et al 1999). In both the feasibility in rice using anther culture–derived haploid cells
cases, gene targeting has been directed toward generating base as well as scutellum-derived diploid cells. We also investigated
changes that result in a chemically selectable phenotype. The whether the haploid state promoted the efficiency of
targeted gene has been an acetohydroxyacid synthase (AHAS) chimeraplast-directed gene targeting.
Fig. 1. Chimeraplast Ch-W548L, which is designed to substitute TTG (a leucine codon) at codon 548 for
TGG (a tryptophan codon). Uppercase letters, DNA residues; lowercase letters, 2′′-0-methyl-RNA resi-
dues; bold letters, target codon.
Materials and methods matograms at the target nucleotide in codon 548, in the same
manner as we reported previously (Okuzaki and Toriyama
A japonica rice variety, Taichung 65 (Oryza sativa L.), was 2004). However, no base changes were observed in the whole
used in our study. A 6–8-day-old callus derived from scutel- coding region of the ALS gene, including another target codon,
lum of mature seeds and a 5-week-old callus derived from 627. Sequencing of the cloned PCR products demonstrated
anther culture were used as target cells. The calli were placed that both unconverted wild-type (TGG) and converted mutant
on a plate in a circle 2.5 cm in diameter for delivery of alleles (TTG) were present in codon 548 in all three plants.
chimeraplasts. The base change in ALS was transmitted to T1 seedlings. The
frequency of calli having predicted conversion at the target
Chimeraplast design and delivery site of W548L after delivery of Ch-W548L was 0.12 per plate.
Three chimeraplasts were designed based on the sequence of Out of 44 control plates, which included the delivery of Ch-
the rice ALS gene (accession numbers AB049822 and S627I, delivery of gold particles only, and no bombardment
AB049823 in the DDBJ databases). A Ch-W548L (Fig. 1) was treatment, 4 plants showed resistance to BS (Table 1). These
designed to substitute TTG (a leucine codon) at codon 548 for plants also contained the same base change in codon 548, which
TGG (a tryptophan codon). In chimeraplast S627I, AGT for seemed to be caused by spontaneous mutation. The frequency
serine was replaced by ATT for isoleucine. Ch-P171A was of spontaneous mutation of W548L was calculated to be 0.09
designed with the Pro-171 codon CCC altered to contain a per plate.
mismatch codon, GCC, which encodes alanine. Chimeraplasts Next, we delivered chimeraplasts to anther culture–de-
were delivered to callus cells by particle bombardment rived calli. Bombardment of a mixture of three chimeraplasts
(Okuzaki and Toriyama 2004). Calli were selected on N6 me- (Ch-W548L + Ch-S627I + Ch-P171A) and a mixture of two,
dium supplemented with 0.25–0.5 µM Bispyribac-sodium (BS, including Ch-W548L, yielded 22 BS-resistant calli, all of which
Kumiai Chemical Industry Co., Ltd., Shizuoka, Japan). A plas- contained the predicted conversion sequence, TTG, at Trp-548
mid containing cauliflower mosaic virus 35S promoter and codon (Table 1). Out of 23 control plates, in which Ch-W548L
green fluorescent protein (35S-GFP) was also delivered for was not delivered, 11 calli showed resistance to BS, having
the estimation of the number of cells receiving gold particles. the same base change, TTG, in codon 548, which was likely
caused by spontaneous mutation. The frequency of calli hav-
Molecular characterization of herbicide-resistant ing a predicted conversion at the target site of Ch-W548L af-
plants ter delivery of Ch-W548L was 0.96 per plate, and was twice
Genomic DNA was extracted from the leaves of regenerated as high as the frequency of spontaneous mutation, which was
plants. Target sequences were amplified from the DNA using 0.47 per plate. These results confirmed that the chimeraplast-
Ex Taq polymerase and the ALS primers (Okuzaki and directed gene targeting was feasible in rice.
Toriyama 2004). Amplified fragments were gel-purified and
were either directly sequenced or sequenced after cloning into Estimation of gene conversion frequency per cell
a pCR 2.1 TOPO vector. We estimated gene conversion efficiency using the following
calculation: Gene conversion frequency = (Number of calli
having predicted gene conversion per plate – Number of calli
Results and discussion having spontaneous mutation per plate) / Total cells receiving
Feasibility of chimeraplast-directed gene targeting chimeraplasts per plate. The total cells receiving chimeraplasts
At first, we delivered a mixture of Ch-W548L and Ch-S627I were estimated by transient expression of 35S-GFP. Gene con-
to scutellum-derived calli by particle bombardment. Three in- version frequency mediated by chimeraplasts was estimated
dependent calli resistant to BS were obtained out of 20 plates to be 3.3 × 10–5 and 3.2 × 10–4 for scutellum-derived calli and
of calli (Table 1). Direct sequencing of the PCR products of anther culture–derived calli, respectively. The frequency of
the regenerated plants showed two overlapping peaks of T and chimeraplast-directed ALS W548L change in anther-culture
G, and consequently the N nucleotide designation in the chro- calli was ten times higher than that of scutellum calli.
Scutellum calli
Ch-W548L + Ch-S627I 20 3 0 0
Ch-W548L 5 0 0 0
Total with Ch-W548L 25 3 (0.12)
Ch-S627I 6 2 0 0
Gold particles only 13 1 0 0
No bombardment 25 1 0 0
Total without Ch-W548L 44 4 (0.09)
Anther-culture calli
Ch-W548L + Ch-S627I + ChP171A 8 13 2 0
Ch-W548L + Ch-S627I 7 5 0 0
Ch-W548L + Ch-P171A 8 4 1 0
Total with Ch-W548L 23 22 (0.96)
Ch-S627I + Ch-P171A 8 5 0 0
Gold particles only 7 1 0 0
No bombardment 8 5 0 0
Total without Ch-W548L 23 11 (0.48)
Active oxygen species (AOS) such as superoxide radical conductivity of the solution (value 1) was measured with a
(O2–), hydrogen peroxide (H2O2), and hydroxyl radical (OH•) conductivity meter (Twin Cond B-173, Horiba, Japan). Next,
are generally produced during the biological process. The mi- the samples were boiled for 30 min and shaken again for 2.5 h,
tochondrial respiratory chain and the photosynthetic process and the conductivity (value 2) was measured. The electrolyte
are the major sites where AOS are produced. Indeed, approxi- leakage value (%) was expressed as (value 1)/(value 2) times
mately 2–3% of the oxygen used by the mitochondria can be 100.
converted into superoxide and H2O2 (Puntarulo et al 1988). Mitochondrial activity was measured by TTC reduction
Since plant growth depends on the balance between the pro- according to Steponkus and Lanphear (1967). Fatty acid
duction and detoxification of AOS, functional stability of AOS- peroxidation was measured by the TBA reaction according to
scavenging enzymes, such as superoxide dismutase, catalase, Heath and Packer (1968). Photosynthetic activity was exam-
and ascorbate peroxidase, is critical for plant survival under ined by measuring the value of Fv/Fm using a Plant Efficiency
stress environments. In fact, many transgenic plants express- Analyzer (Hansatech) according to Clarke and Campbell
ing such antioxidant enzymes were reported. We previously (1996). Respiratory activity was analyzed by measuring the
confirmed that the expression of wheat catalase in transgenic decrease in oxygen using Oxytherm (Hansatech).
rice plants shows tolerance for chilling injury (Matsumura et
al 2002). Although improved tolerance for stress circumstances
Results and discussion
was reported, the functional mechanism of the tolerance has
not been revealed. To make clear the reasons for improved Visible damage by chilling
tolerance in transgenic rice plants for low-temperature injury, After plants were exposed at 5 oC for 8 d, visible damage by
we examined the effect of chilling (5 oC up to 8 days) on elec- chilling was compared. As Figure 1 shows, leaves of
trolyte leakage, mitochondrial membrane integrity, fatty acid nontransgenic rice after 8 d of chilling were as severely dam-
peroxidation, and photosynthetic and respiratory activities. aged as withered. However, such severe damage did not occur
to the leaves of transgenic rice. These data clearly demonstrate
that transgenic rice plants are more cold-tolerant than
Materials and methods nontransgenic rice.
Transgenic rice plants (CT 2-6-4) in which wheat catalase
cDNA was expressed and nontransgenic control rice plants Chilling effect on catalase activity and H2O2 content
(cultivar Yuukara) (Matsumura et al 2002) were used in this While the catalase activity in leaves of nontransgenic rice at
study. All plants were grown in a growth chamber for 20 days 25 oC was around 100 (µmol mg–1protein min–1), the transgenic
under conditions of 12 h light (262 µmol m–2 s–1) at 25 oC and rice showed values of approximately 700. After 4 d of chill-
12 h dark at 20 oC. ing, the activities in both nontransgenic and transgenic rice
The effects of chilling (12 h light, 60 µmol m–2 s–1 at 5 were reduced to approximately 60 and 390, respectively (Fig.
oC and 12 h dark at 5 oC) on these plants were examined with 2). After 8 d, the transgenic rice showed 4 times higher re-
the method previously reported (Matsumura et al 2002). sidual activity than the control (Matsumura et al 2002).
Catalase (CAT; EC 1.11.1.6) activity in leaves was mea- In the transgenic plants, the content of H2O2 (about 0.36
sured according to Matsumura et al (2002). H2O2 concentra- µmol g–1 FW) was kept low during chilling. In contrast, the
tion in leaves was measured according to Patterson et al (1984). nontransgenic plants showed an increased content of H2O2 by
Electrolyte leakage was measured as follows. Leaf cut- chilling, and, at the end of the 8-d chilling period, the content
tings (3 cm) from plants after various treatments were added reached a level two times that of the control (about 0.68 µmol
to 3.0 mL of water in a test tube and shaken for 2.5 h. The g–1 FW). Thus, high catalase activities remaining in the
700
600
500
400
300
200
Fig. 1. Visible damage from chilling. Transgenic and nontransgenic
plants were grown in a growth chamber for 1 mo and then chilled
at 5 °C for 8 d. (A) Nontransgenic rice plant (control), (B) transgenic
rice plant. 100
0
transgenic plants seem to prevent an increase in H2O2 content 0 1 2 3 4
after chilling. Chilling period (day)
Wrap-up of Session 3
Rice is the most important food crop for half the world’s popula- Tolerance of unfavorable circumstances such as not using Fe-
tion. Because of the growing population in Asian countries, more deficient soil is important for expanding cultivation land to en-
rice will have to be produced from less land for cultivation. Ge- hance rice production worldwide. N. Nishizawa showed that
netic engineering of rice could solve this problem by permitting transgenic rice plants were tolerant of low Fe availability in cal-
higher productivity, including more resistance to biotic and abi- careous soil when barley genes encoding the enzymes involved
otic stresses as an alternative technology of conventional breed- in the biosynthesis of mugineic acid family phyotosiderophores
ing. Biotechnology using genetic engineering makes it possible (MAs) were introduced into the rice genome. This is because
to introduce valuable genes from all organisms into the rice ge- transgenic rice plants secrete higher amounts of deoxymugineic
nome. The application of genetic engineering is not limited to acid. Her group has characterized the transporter genes involved
agriculture but is relevant to industry as molecular farming. Rice in the uptake and translocation of iron and zinc from the soil.
can be used as a bio-reactor for producing high added value Plant productivity is affected severely by environmental
such as biopharmaceuticals (antibodies, vaccines). stresses such as cold, drought, and high salinity. K. Shinozaki’s
In Session 3, six leading researchers reviewed the present group has identified a cis-regulatory element designated as the
status of studies on transgenic rice. Transgenic rice for improved dehydration response element box (DREB), which is conserved
input traits showing resistance to biotic and abiotic stress and in promoters of many genes implicated in gene regulation in re-
output traits with improved quality of rice grain has been devel- sponse to the above stress. They demonstrated that the DREB
oped. transcription factor gene is quite useful for improving tolerance
To increase rice productivity, M. Tokutomi’s research group of environmental stresses in transgenic rice, when the DREB rec-
has introduced genes coding for maize C4-specific photosynthetic ognizing the conserved sequence was overexpressed under the
enzymes such as PEPC, PPDK, and NADP malic enzyme. Although control of a stress-inducible promoter.
transgenic rice overexpressing C4 enzymes exhibited a higher Yield is severely decreased by pathogen attacks such as
photosynthetic rate, these plants showed little increase in yield. fungal rice blast. Disease resistance can be enhanced by intro-
Sink and source function and their relationship are important sociated with complex traits, quantitative trait loci (QTLs).
aspects to understand the regulatory mechanism of crop pro- Ishimaru et al (2001a) applied QTL analysis to physiological
duction such as rice. In terms of carbon metabolism, the source traits, including sink and source functions. Using backcross
function in rice leaves includes photosynthetic carbon fixa- inbred lines of Nipponbare (japonica) × Kasalath (indica), a
tion, sucrose formation, and its loading into sieve elements. rice function map was constructed by collating the results on
The sink function in rice panicles includes sucrose unloading, QTLs for 23 important agronomic, physiological, and mor-
its breakdown, and starch biosynthesis. In addition, leaf sheaths phological traits. The source function-related traits such as
and culms show sink and source functions before and after chlorophyll content in leaves, photosynthetic ability, and δ13C
heading, respectively. In the latter stage, starch breakdown is value were mapped. These traits were first mapped in the QTL
a function of major carbon metabolism. analysis of rice in their study. One of three QTLs controlling
Rice genome research has been developing widely, pro- chlorophyll content partly overlapped with that controlling
viding a huge amount of gene information, DNA markers, and yield, although most of these source function-related traits did
high-throughput technologies such as DNA microarray, allow- not overlap with that controlling yield.
ing us to analyze these complex networks of carbon metabo- Among these source function-related QTLs, Ishimaru et
lism. At the same time, this makes it possible to elucidate the al (2001b) further analyzed Rubisco content, soluble protein
genetic basis of physiological traits such as sink and source content, and nitrogen content in flag leaves of rice. The kinet-
function using a genetic map and DNA markers. ics of Rubisco and the ratio of Rubisco to total leaf nitrogen
I introduce some examples of these new approaches on (N) are the two main factors determining N-use efficiency for
sink and source functions. photosynthetic carbon fixation. Since the kinetics of Rubisco
is very difficult to change, the ratio of Rubisco to total leaf N
content is the main target for the improvement of N-use effi-
Source function ciency. In their function map, QTLs controlling Rubisco con-
Recent progress in the generation of a molecular genetic map tent were not detected near QTLs for total leaf N content, in-
and markers has made it possible to map individual genes as- dicating that contents of Rubisco and total leaf N are controlled
Sunlight is the principal energy source for crops. It is also highly increased radiation conversion factor (RCF) (Sheehy et al
variable; differences in light levels caused by season, climate, 2000). It is essential to understand how such variation is ac-
and position can affect the photosynthetic rate and productiv- commodated efficiently within the plant such that biomass pro-
ity of crops. High yields are generally attained under high- duction rates are not affected, or can even be improved. This
irradiance conditions, in which photosynthesis is often satu- is largely carried out by processes termed acclimation. Accli-
rated and in full sunlight the number of photons may exceed mation is defined as a change in composition or morphology
photosynthetic requirements. The converse may be true in low- of the plant that is a specific response to an environmental
irradiance conditions such as the tropical wet season, in which change. There is ample evidence in plant science for signifi-
light interception is lower (Weerakoon et al 2000). Here, pho- cant variation in the capacity for acclimation to a number of
tosynthesis may often be light-limited and growing periods are environmental factors.
longer and low-yielding (IRRI 1998). These long-term pat- Additionally, it is vital to consider that the requirements
terns are punctuated by short-term variation in light. Ideally, a for light and N are strongly linked: the efficient use of high-
crop would rapidly maximize photosynthesis according to pre- irradiance levels (high photosynthetic rates) requires the as-
vailing local conditions. similation of large quantities of N. Leaf N content stored in
Understanding the processes that affect the efficiency of the lower leaves of the canopy is essential for high grain yield
transduction of light energy by crops is central to the improve- (Sheehy et al 2000) and any attempt to improve photosynthe-
ment of yield potential: this was highlighted by one recent study sis must take this into account.
that showed that the radiation-use efficiency (biomass produced This paper summarizes the responses of rice plants to
per unit of radiation energy intercepted) of rice is low, even excess irradiance. Genotype variation in light responses is
among C3 crops, and this is a major limitation to yield poten- described. Emphasis is placed on photosynthesis and the iden-
tial (Mitchell et al 1998). An increase in leaf-level photosyn- tification of limitations to biomass productivity.
thetic capacity at ambient CO2 levels is likely to lead to an
Poor filling of spikelets on the basal primary branches of the been identified as a regulator of senescence in detached rice
rice panicle limits genetic potential for yield and produces leaves (Lin et al 2002), but nothing is known about its role in
grains unsuitable for human consumption (Mohapatra et al the regulation of senescence in intact plants and grain filling.
1993). The problem is more acute in the newly developed su- The application of triazolic compounds reduces ethylene pro-
per rice with heavy panicles. Spikelet development is very slow duction in plants, and thereby delays senescence (Fletcher et
in the proximal part of the panicle; it results in poor partition- al 2000). In contrast, 2-chloroethylphosphonic acid (CEPA)
ing of dry matter for grain filling (Mohapatra et al 1993). As- application enhances ethylene production and may expedite
similates for grain filling are received from the flag leaf and senescence. We need to ascertain the role of ethylene in regu-
photo-assimilatory tissues located within the panicle structure lating the filling of rice spikelets. It is also desirable to distin-
(Yoshida 1981). Senescence of these organs is programmed guish its action from that of senescence-controlled reduction
genetically, and has a fixed time schedule. Therefore, it will in assimilate supply through depletion of source activity.
be of interest to know if the late-developed basal spikelets of
the rice panicle become source-limited for assimilates during
Materials and methods
their peak period of growth because of the coincidence of se-
nescence of the photosynthetic tissues. Similarly, the forma- A high-yielding rice plant, Oryza sativa cv. Mahanadi
tion of ABA or ethylene during senescence of these organs (IR19661-131-1-3 × Savitri), was cultivated in irrigated field
may make the spikelets sink-limited for growth. Ethylene has conditions of the Regional Research Station, Chiplima, during
H2 O 6.12 ± 0.95 3.76 ± 0.48 190.7 ± 31.7 151.7 ± 34.1 40.0 ± 8.2 19.70 ± 0.80 79.5 ± 5.3 21.0 ± 2.4
Uniconazole 6.85 ± 0.03 3.94 ± 0.15 201.0 ± 1.4 167.5 ± 6.4 33.5 ± 7.7 19.62 ± 0.85 83.3 ± 3.3 16.7 ± 2.8
Paclobutrazol 7.14 ± 0.79 4.10 ± 0.67 202.7 ± 4.6 176.7 ± 14.5 26.0 ± 1.1 20.24 ± 2.88 87.3 ± 5.0 12.8 ± 5.0
CEPA 5.54 ±0 .56 2.78 ± 0.40 204.3 ± 36.5 126.7 ± 13.8 77.7 ± 5.0 13.58 ± 0.46 62.0 ± 4.8 38.0 ± 4.8
the wet season of 2000. The experimental area was divided Ethylene concentration of panicle,
into three parts. In each part, uniformly developed plants were primary branches, flag leaf, boot, and spikelets
divided into four groups for growth regulator treatments. On Ethylene concentration of the plant organs was low at 5 d be-
each day, 0.5 mL of the chemicals paclobutrazol (100 µL L–1) fore anthesis; it reached a peak at anthesis and declined there-
and uniconazole (0.5 mg L–1), 2-CEPA (10–5 M), and H2O in after (Fig. 1). It was minimum at maturity. The concentration
aqueous solution were injected into the boot (the leaf sheath of ethylene inside the boot of the flag leaf also increased tem-
enclosure confining the young panicle before emergence) of porally during the pre-anthesis period. The ethylene concen-
the flag leaf. The treatments began 5 d before the day of anthe- tration of the apical primary branch was lower than that of the
sis and continued for 5 d. basal branch of the panicle. The treatment with triazolic com-
Plants were harvested from the day of treatment to ma- pounds decreased the concentration of ethylene of the boot
turity at 5–7-d intervals. Six spikelets from the uppermost pri- and the plant organs significantly, but CEPA increased the con-
mary branch (apical) and eight spikelets from the lowermost centration. The treatment with the chemicals was more effec-
primary branch (basal) were collected from the panicle on the tive on the basal branch than on the apical branch.
main shoot of four different plants in each treatment. The api-
cal spikelets emerged 5–6 d earlier than the basal spikelets. In Pigment concentration of glumes
each set, the lemma and palea (glumes) were dissected out, The concentrations of chlorophylls, carotenoids, soluble sug-
weighed fresh, and used for biochemical analyses. The four ars, proteins, and free amino acids of the glumes increased
sets of spikelets were used for estimation of pigments, soluble temporally from 5 d before to 1 wk after anthesis. It declined
carbohydrates, lipid peroxidation, and activity of peroxidase thereafter till maturity in the apical branch (data not included).
enzyme, respectively. At maturity, grain yield of the panicle In the basal branch, the concentration of these materials in-
was estimated. Separate samples were taken for measuring eth- creased for a longer time before declining. During the declin-
ylene concentration of the panicle, basal and apical primary ing phase, the glumes of basal spikelets possessed a higher
branches, flag leaf, and boot of the flag-leaf sheath. The cut concentration of the materials than those of the apical spike-
end of the plant material was dipped in 1 mL of water inside a lets. The treatment with the triazolic compounds improved the
test tube. The test tube was sealed and incubated for 2 h in concentration significantly in the glumes of basal spikelets,
darkness before 1 mL of head space gas was used for ethylene but CEPA application had the opposite effect. The chemical
estimation in a gas chromatograph (model 6890, Hewlett- treatments did not change the concentration of the materials in
Packard Company, USA, equipped with a flame ionization the apical spikelets.
detector) (Naik and Mohapatra 2000).
Peroxidase activity and peroxidation
of lipids in glumes
Results The activity of the enzyme peroxidase and the amount of lipid
Grain yield and yield components peroxidation in the glumes of apical spikelets increased from
Uniconazole and paclobutrazol treatments improved panicle 5 d before to 16 d past anthesis and declined thereafter till
weight, spikelet weight, and number of filled spikelets, and maturity (data not presented). The activity was higher in the
reduced the number of barren spikelets significantly (Table glumes of the basal spikelets than in the apical spikelets. Per-
1). In contrast, CEPA application had the opposite effect. The oxidase activity increased up to 23 d past anthesis in the basal
triazolic compounds did not improve the average weight of spikelets and declined thereafter. The treatment with triazolic
spikelets, but CEPA application depressed it significantly. compounds increased activity of the enzyme, whereas CEPA
decreased it significantly. The treatment with chemicals was
not effective on the apical spikelets.
3 3
2 2
1 1
0 0
10 12
(Flag leaf, 2000) (Panicle, 2000)
10
8
8
6
6
4
4
2
2
0 0
5 0 5 5 0 5
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0
5 4 3 2 1
Days from anthesis
Fig. 1. The effect of uniconazole, paclobutrazol, and 2-chloroethylphosphonic acid on the ethyl-
ene concentration of apical and basal primary branches, panicle, flag leaf, and boot of the flag
leaf of rice at anthesis during the wet season of 2000. Anthesis date is recorded when the apical
spikelets reach that stage. This time scale is used for all organs except the basal primary branch.
Grain yields of rice (Oryza sativa L.) as well as other major Data
cereals have increased markedly in the last century. Genetic
improvements have contributed largely to these increases. Yield Most of the rice responses to elevated [CO2] presented in this
potential can be increased by increasing biomass production paper are from the FACE experiment conducted in Shizukuishi,
and/or increasing allocation to harvested organs. Genetic im- Iwate, Japan (39o38′N, 140o57′E) from 1998 to 2000. The
provements in the yield potential of major cereals have largely design of the FACE system was reported by Okada et al (2001)
been through the latter. Photosynthesis plays a central role in and growth conditions were given by Kim et al (2003a). Briefly,
biomass production and recent varieties tend to have longer rice was grown in farmers’ paddies equipped with 12-m-diam-
photosynthetic duration than older varieties, but only limited eter FACE systems. Pure CO2 gas was released from periph-
evidence exists that the maximum photosynthetic capacity has eral emission tubes 0.5 m above the canopy, the sides of the
been improved. Even where improved photosynthetic capac- emission depending on the wind direction. The target [CO2] at
ity was observed, no direct evidence of biomass increase was the center of the FACE plots was 200 µmol mol–1 above the
shown (review: Evans 1993). Efforts are under way to improve ambient [CO2]. Each plot had some subplot treatments such as
the genetic potential of photosynthesis, but the question “Does N rates and varieties, but here we present only the results from
improving photosynthesis benefit crop production?” needs to cultivar Akitakomachi under standard N conditions (8–9 g
be answered experimentally. m–2 of N).
Photosynthetic capacity under current atmospheric CO2 For this paper, growth and yield data in Shizukuishi were
concentrations ([CO2]) is generally limited by the carboxyla- from Kim et al (2003a,b). Leaf photosynthesis data were from
tion process in the Calvin cycle. Increases in [CO2] accelerate Seneweera et al (2002) and Anten et al (2003). In the former
carboxylation velocity by increasing [CO2] supply from sto- study, [CO2] and [H2O] exchange rates were measured on the
mata and reducing photorespiration (review: Long et al 2004). top fully expanded leaves on two occasions during the season,
In general, CO2 enrichment studies aim to identify the magni- whereas in the latter study light-saturated CO2 assimilation
tude of changes in crops with increased [CO2], but these re- (Asat) and maximum apparent quantum yield of CO2 uptake
sults can also be used to infer the impact of increased photo- (QY) were measured at different positions of the canopy at
synthesis on crop yield. around flowering. By also measuring canopy light attenuation
Among various experimental facilities that have been and leaf area distribution, Anten et al (2003) estimated canopy
used in elevated CO2 studies, free-air CO2 enrichment (FACE) photosynthesis by integrating the photosynthesis of sunlit and
systems, which do not have enclosures and therefore a mini- shaded leaves.
mum of artifacts, are usually regarded as providing the best Other rice data presented in this paper come from the
estimates of yield responses to elevated [CO2] in the open field. rice-wheat FACE experiment conducted in Wuxi (31 o37′N,
Across all species, fewer FACE than chamber experiments have 120o28′E), Jiangsu, China, in 2001, 2002, and 2003. A FACE
been conducted, though the rice FACE system has been run- system similar to the one in Shizukuishi was used. Final bio-
ning at two sites (Japan and China) for more than three years. mass and grain yield data of cultivar Wuxiangjing 14 (Huang
Many data have been accumulated, leading to reviews of the et al 2004) are presented in this study.
yield responses of a range of plant species to FACE (Kimball To compare rice growth and yield responses to FACE
et al 2002, Long et al 2004). To expand on these reviews, this with those of other C3 species, we present data from a review
paper summarizes the FACE effects on photosynthesis, bio- by Kimball et al (2002) and a meta-analysis conducted by Long
mass, and yield of rice in order to examine the effect of in- et al (2004) as a reference; the latter reviewed results from
creased source capacity on grain yield. The rice results are large-scale (>8 m) FACE experiments only.
compared with those of other C3 species responses reviewed
by Kimball et al (2002) and Long et al (2004).
Jmax
SF S8
V,max/Jmax
SF S8
Narea
A
Asat
SF A S8
QY
A
Acanopy
CGR
Final biomass
Rice (Shizukuishi)
Rice (Wuxi)
Agricultural yield
Rice (Shizukuishi)
Rice (Wuxi)
Wheat (Maricopa)
Potato (Rapolano)
Cotton (Maricopa)
40 20 0 20 40 60
% increase with elevated [CO2]
Fig. 1. The percent change under elevated [CO2] of (1) maximum car-
boxylation rate (Vc,max), (2) maximum rate of electron transport (Jmax),
(3) the ratio of Vc,max:Jmax, (4) N content per unit leaf area (Narea), (5)
light-saturated CO2 assimilation (Asat), (6) maximum apparent quantum
yield of CO2 uptake (QY), (7) daily integral of leaf CO 2 uptake (A’), (8)
estimated daily net canopy CO2 assimilation at the heading stage
(Acanopy), (9) crop growth rate (CGR) around the heading stage, (10)
final biomass, and (11) seed yield. Meta-analysis results of large-scale
free-air CO2 enrichment experiments for C 3 species along with 95% con-
fidence intervals (Long et al 2004) are presented as a reference. Rice
leaf photosynthetic properties are from Seneweera et al (2002) and Anten
et al (2003); S8 and SF beside symbols stand for the 8th-leaf and flag-
leaf measurements, respectively, of Seneweera et al (2002), and A for
Anten et al’s measurements. Rice growth and yield data in Shizukuishi,
Japan, are from Kim et al (2003a,b) and in Wuxi, China, from Huang et al
(2004). Agricultural yield data of other C3 crops (seed for wheat, tuber
for potato, and boll for cotton) were from Kimball et al’s (2002) review;
only results from crops grown under ample water and nitrogen condi-
tions are presented.
We have a unique opportunity to solve all of the rice produc- The C4 syndrome evolved from the C3 syndrome and C4 plants
tion problems of Asia by redesigning one plant mechanism: have a photosynthetic system that concentrates CO2 around
photosynthesis. A lack of scientific understanding of the yield- Rubisco, and so nearly saturates the carboxylation reaction
shaping mechanisms, combined with a triumph of correlation and eliminates photorespiration. Consequently, C4 plants use
over common sense, have led to the commonly held belief water more effectively. In addition, C4 crops require less ni-
that leaf photosynthesis and yield are unrelated. To counter trogen fertilizer because they make better use of nitrogen in
that misconception, we argue the link between photosynthesis photosynthesis. The maximum rate of leaf photosynthesis for
and yield and explore what it would take to be a C4 rice plant. both C3 and C4 plants appears to vary linearly with nitrogen
Recent discoveries of single-cell C4 photosynthesis could pro- content per unit leaf area over a wide range of nitrogen con-
vide new impetus to convert rice from a C3 to a C4 photosyn- centrations (Evans and von Caemmerer 2000). Assuming the
thetic system without having to incorporate the Kranz anatomy data shown by Evans and von Caemmerer (2000) could be
(Sage 2002). described by linear relationships, then maize had a gradient of
Solar energy captured in photosynthesis gives plants the approximately 0.74 µmol CO2 s–1 mmol–1 N compared with
capacity to grow. Two aspects of energy capture set the limits that for C3 plants (rice and wheat) of 0.26 µmol CO2 s–1
for crop growth in a given thermal environment. The first is mmol–1 N: a threefold difference in leaf photosynthetic nitro-
the quantity of energy captured, and that depends on the avail- gen-use efficiency (Table 1).
able irradiance and the fraction absorbed by the crop canopy. Photosynthesis in C3 plants should benefit more from
The second is the efficiency with which the absorbed energy rising concentrations of atmospheric carbon dioxide than the
is used for the chain of synthetic processes that culminate in photosynthesis in C4 plants. However, work of Baker et al
harvestable yield. Modern cultivars with erect leaves absorb (1990) showed that canopy photosynthesis in rice did not re-
all the available photosynthetically active radiation, leaving spond to increases in atmospheric CO2 beyond 50 Pa. Average
little room for further improvements in energy capture. In high- atmospheric concentration of carbon dioxide increased by ap-
yielding rice, sink size is unlikely to limit yield; the potential proximately 4.5 Pa from 1968 to 1998. However, over that
sink size is almost double that actually realized. period, yield for the same cultivar (IR8) grown on the IRRI
In plants using the C3 photosynthetic pathway (rice, farm at nitrogen inputs of approximately 150 kg N ha–1 fell by
wheat, etc.), oxygen competes for the active sites on the key 2.6 t ha–1 (Peng et al 1999).
photosynthetic enzyme, Rubisco, leading to as much as a 30%
loss of captured CO2 in a process known as photorespiration.
0
The link between yield and photosynthesis 0 20 40 60 80 100 120
DAT
There are two nearly equal phases of growth in rice (Fig. 1), Fig. 1. The two phases of rice growth for IR72 growing in the dry
with about half of the first phase of vegetative growth preced- season of 2001 at IRRI with nonlimiting fertilizer applications. Open
ing the second phase of reproductive growth (largely the squares denote vegetative parts and filled squares denote panicles.
DAT = days after transplanting.
panicle). The growth of the panicle is largely heterotrophic
and in high-yielding rice has a maximum growth rate about
the same as that achieved by the autotrophic vegetative com-
ponent. How could the growth rates of both components be the canopy (Ln) were those that were close to the light com-
increased by 50%? There is a linear relationship between ac- pensation point and largely useful for the N they contributed
cumulated intercepted photosynthetically active solar radia- to the grain (Ln = L – Lo – Ls; where L is the total LAI). Sheehy
tion and accumulated biomass; the slope of that linear rela- (2000) went on to show that the loss through senescence of Ls
tionship is known as the radiation-use efficiency (ε). The rela- and Ln would reduce canopy photosynthesis by about 22–27%.
tionship between the growth rate of shoots (dWs/dt) and ε is The maximum rate of canopy photosynthesis (Pgmax) has an
asymptotic value given by
dWs/dt = ε Iint (1)
Pgmax = Lo Amo + Ls Ams as I → ∞ (3)
where Iint is the total amount of photosynthetically active ra-
diation (PAR) intercepted by the crop for the same day. To where I is the incident PAR, Amo is the maximum rate of sunlit
increase growth rates by 50%, the value of ε would have to leaf photosynthesis, and Ams is the maximum rate of shaded
increase by 50%, from the average value of ε for rice of 2.2 g leaf photosynthesis. The maximum quantum yield of the canopy
dry weight (DW) MJ–1 PAR to 3.3 g DW MJ–1 PAR, the aver- at full light interception is
age value for maize (Table 1, Kiniry et al 1989). The link be-
tween ε and photosynthesis on any given day can be written dPg/dI = α(1 + τ) as I → 0 (4)
as
where α is the leaf quantum yield in low irradiance and τ is
ε = (1 - β) [(Pg(t) – R(t) – D(t))] /Iint(t) (2) leaf transmittance. It can be seen that the limits for canopy
photosynthesis are set by the properties of the individual leaves.
where β is the root weight ratio, Pg is the rate of canopy gross At LAIs greater than approximately 6.0, improvements in
photosynthesis, R is the rate of shoot and root respiration, D is canopy photosynthesis can only result from improvements in
the loss of dry matter through detachment, and t is time in leaf quantum yield and/or leaf photosynthesis.
days. On a given day, to increase ε by 50%, photosynthesis At daytime temperatures typical of the tropics (>26 °C),
would have to increase by a little more than 50%, that is, to the quantum yield of C4 plants is greater than that of C3 plants
values comparable with those of C4 plants. (Ehleringer and Björkman 1977) and the optimum tempera-
ture for rice plants is close to 30 °C, whereas that for C4 plants
is closer to 36 °C. Climate change models predict that by 2100
Upper limits to canopy photosynthesis mean planet-wide surface temperatures will rise by 1.4 to 5.8
The simple model of Monteith (1965) shows that the maxi- °C, making C4-ness even more desirable for tropical and semi-
mum photosynthetic leaf area index of an erect canopy is ap- tropical environments.
proximately 6.6. The sunlit leaf area index (Lo) reached a maxi-
mum of 3.3 and photosynthetically active shaded leaf area in-
dex (Ls) reached its maximum at approximately 3.0. Sinclair
and Sheehy (1999) proposed that the third class of leaves in
There are four strategies for producing a C4 rice plant: (1) Baker JT, Allen LH Jr, Boote KJ, Jones P, Jones JW. 1990. Rice
achieve single-cell C4 photosynthesis by combining the C4 photosynthesis and evapotranspiration in subambient, ambi-
pathway and the Calvin cycle in one cell, (2) produce a C3-C4 ent and superambient carbon dioxide concentrations. Agron.
intermediate with reduced photorespiration, (3) produce a C4 J. 82:834-840.
Ehleringer J, Björkman O. 1977. Quantum yields for CO2 uptake in
rice with Kranz anatomy and specialization of chloroplast func-
C3 and C4 plants: dependence on temperature, [CO2], and [O2]
tion in mesophyll and bundle-sheath cells, and (4) produce
concentration. Plant Physiol. 59:86-90.
more Rubisco, or better Rubisco, or find ways to make Rubisco Evans JR, von Caemmerer S. 2000. Would C4 rice produce more
work harder. biomass than C3 rice? In: Sheehy JE, Mitchell PL, Hardy B,
All of the genes for the C4 syndrome are already present editors. Redesigning rice photosynthesis to increase yield.
in C3 plants and some plants switch between C3 and C4 modes Proceedings of the Workshop on The Quest to Reduce Hun-
of photosynthesis. In the sedge Eleocharis vivipara, the sub- ger: Redesigning Rice Photosynthesis, 30 Nov.-3 Dec. 1999,
merged culms are C3 but the emergent culms are C4 with the Los Baños, Philippines. Makati City (Philippines): Interna-
Kranz anatomy (Ueno et al 1988). Orcuttia has been shown to tional Rice Research Institute and Amsterdam (The Nether-
have two types of anatomy: the terrestrial leaves are C4 with lands): Elsevier Science B.V. p 53-71.
Kranz anatomy but submerged leaves have C4 biochemistry Hibberd JM, Quick WP. 2002. Characteristics of C4 photosynthesis
in stems and petioles of C3 flowering plants. Nature 415:451-
without Kranz anatomy (Keeley 1998). Hibberd and Quick
454.
(2002) showed that cells in stems and petioles of a typical C3
Keeley JE. 1998. C4 photosynthetic modifications in the evolution-
plant, tobacco, displayed C4 characteristics, opening up the ary transition from land to water in aquatic grasses. Oecologia
possibility that a form of C4-ness may already exist in some 116:85-97.
rice plants. Kiniry JR, Jones CA, O’Toole JC, Blanchet R, Cabelguenne M,
Recently, another plant has been reported to operate the Spanel DA. 1989. Radiation use efficiency in biomass accu-
C4 mechanism within single cells (Voznesenskaya et al 2002). mulation prior to grain-filling for five crop species. Field Crops
In the chenopod Borszczowia aralocaspica, the elongated Res. 20:51-64.
chlorenchymatous cells are closely packed, without intercel- Monteith JL. 1965. Light distribution and photosynthesis in field
lular air spaces, at the inner end where they have abundant, crops. Ann. Bot. 29:17-37.
large chloroplasts. The outer two-thirds of the cells have the Peng S, Cassman KG, Virmani SS, Sheehy JE, Khush GS. 1999.
Yield potential trends of tropical rice since the release of IR8
appearance of mesophyll cells, with normal air spaces.
and the challenge of increasing rice yield potential. Crop Sci.
39:1552-1559.
Work in progress Sage RF. 2002. C4 photosynthesis in terrestrial plants does not re-
quire Kranz anatomy. Trends Plant Sci. 7:283-285.
At IRRI, we have been screening the wild rice collection for Sheehy JE. 2000. Limits to yield for C3 and C4 rice: an agronomist’s
evidence of C4 traits. The δ13C values range from –31% to view. In: Sheehy JE, Mitchell PL, Hardy B, editors. Rede-
–25% and the interveinal spacing (at the middle of the blade) signing rice photosynthesis to increase yield. Proceedings of
ranges from 167 to 313 µm (the value for IR72 is 227 µm). the Workshop on The Quest to Reduce Hunger: Redesigning
The number of vascular bundles per unit width was not related Rice Photosynthesis, 30 Nov.-3 Dec. 1999, Los Baños, Phil-
to the overall width of the leaf. Currently, we are counting the ippines. Makati City (Philippines): International Rice Research
number of chloroplasts in the bundle-sheath cells and measur- Institute and Amsterdam (The Netherlands): Elsevier Science
B.V. p 39-52.
ing the mesophyll cell size. Work is under way to transfer maize
Sinclair TR, Sheehy JE. 1999. Erect leaves and photosynthesis in
genes for PEP carboxylase (PEPC) and pyruvate, orthophos-
rice. Science 283:1456-1457.
phate dikinase (PPDK), individually and in combination, into Ueno O, Samejima M, Muto S, Miyachi S. 1988. Photosynthetic
elite IRRI indica cultivars. Table 1 shows the advantage of characteristics of an amphibious plant, Eleocharis vivipara:
maize over rice in water-use efficiency, radiation-use efficiency, expression of C4 and C3 modes in contrasting environments.
and photosynthetic nitrogen-use efficiency. It is hard to imag- Proc. Natl. Acad. Sci. USA 85:6733-6737.
ine any trait other than C4 photosynthesis that could increase Voznesenskaya EV, Franceschi VR, Kiirats O, Artyusheva EG, Freitag
the value of rice by US$44 billion annually through increased H, Edwards GE. 2002. Proof of C4 photosynthesis without
yield and reduced water and nitrogen costs. Kranz anatomy in Binertia cycloptera (Chenopodiaceae). Plant
J. 31:649-662.
Notes
Authors’ address: Crop, Soil, and Water Sciences Division, Interna-
tional Rice Research Institute (IRRI), DAPO Box 7777, Metro
Manila, Philippines, e-mail: j.sheehy@cgiar.org.
Increasing yield is still the most important objective of rice Breeding of the second-generation new plant type
breeding programs in developing countries because of the
growing demand for food resulting from population growth In 1995, development of second-generation NPT lines began
and a reduction in area devoted to rice production. The Inter- by crossing first-generation tropical japonica NPT lines with
national Rice Research Institute (IRRI) began developing the elite indica parents. Multiple site-year comparisons of first-
new plant type (NPT) rice through ideotype breeding ap- generation NPT lines with the highest-yielding indica variet-
proaches in 1989 (Khush 1995). The goal was to develop an ies have shown that the original NPT design did not have suf-
NPT with a yield potential 20–25% higher than that of exist- ficient tillering capacity. An increase in tillering capacity is
ing semidwarf rice varieties under a tropical environment dur- needed to increase biomass production and to improve com-
ing the dry season. The NPT was designed based on the results pensation when tillers are lost because of insect damage or
of simulation modeling and the new traits were mostly mor- other causes during the vegetative stage. A slightly smaller
phological since these are easier to select than physiological panicle size without a change in panicle length also appeared
traits in breeding programs. The proposed NPT has a low to be advantageous to reduce the compact arrangement of spike-
tillering capacity (3 to 4 tillers when direct-seeded), few un- lets. Genes from indica parents have effectively reduced panicle
productive tillers, 200 to 250 grains per panicle, a plant height size and increased tillering capacity in the second-generation
of 90 to 100 cm, thick and sturdy stems, leaves that are thick, NPT lines. Indica germplasm also helped improve other NPT
dark green, and erect, a vigorous root system, 100 to 130 days’ attributes such as grain quality and disease and insect resis-
growth duration, and increased harvest index (Peng et al 1994). tance. Some second-generation NPT lines (F5 generation) with
In 1989, about 2,000 entries from the IRRI germplasm the above refinements were then selected and were planted in
bank were grown to identify donors for the desired traits (Khush a replicated observational trial for the first time in the 1998
1995). Donors for the low-tillering trait, large panicles, thick wet season. These second-generation NPT lines have been
stems, a vigorous root system, and short stature were identi- tested in breeders’ replicated yield trials since the 2001 dry
fied in the “bulu” or javanica germplasm mainly from Indone- season and in replicated agronomic trials since the 2002 dry
sia. This germplasm is now referred to as the tropical japonica season. However, only data from the replicated agronomic tri-
(Khush 1995). Hybridization began in the dry season of 1990. als in the dry seasons of 2003 and 2004 will be presented in
The first-generation NPT lines based on tropical japonicas were this paper.
developed in less than 5 years. They were grown in a repli-
cated observational trial for the first time in late 1993. As in-
Performance of the second-generation new plant type
tended, the NPT lines had large panicles, few unproductive
tillers, and lodging resistance. However, they did not yield well Field experiments were conducted under flooded irrigation at
because of limited biomass production and poor grain filling the IRRI farm in the dry season of 2003 and 2004. Five sec-
(Peng and Khush 2003). Low biomass production was caused ond-generation NPT lines and five indica inbred check variet-
by an excessive reduction in tillering capacity and low photo- ies were grown. Seedlings were transplanted at a hill spacing
synthetic rate. Poor grain filling was associated with panicle of 20 × 20 cm. The plants received a basal fertilizer supply of
morphology and source limitation. The poor grain filling of 30 kg P ha–1, 40 kg K ha–1, and 5 kg Zn ha–1 incorporated 1 d
NPT lines was probably caused by the lack of apical domi- before transplanting. Total N fertilizer applied was 200 kg
nance within a panicle (Yamagishi et al 1996), the compact ha–1 in four splits: basal (60 kg ha–1), midtillering (40 kg ha–1),
arrangement of spikelets on the panicle (Khush and Peng 1996), panicle initiation (60 kg ha–1), and heading (40 kg ha–1). Stan-
a limited number of large vascular bundles for assimilate trans- dard cultural management practices were followed. To avoid
port (S. Akita, personal communication), and source limita- yield loss, pests were intensively controlled using recom-
tion because of early leaf senescence (Ladha et al 1998). The mended pesticides.
first-generation NPT lines were also susceptible to diseases In the 2003 dry season, PSBRc52 produced the highest
and insects and had poor grain quality. Therefore, they could yield among indica inbred check varieties and IR72967-12-2-
not be released for planting in farmers’ fields. However, they 3 was the top yielder among the second-generation NPT lines
are valuable genetic materials used in rice breeding programs (Table 1). IR72967-12-2-3 produced 10.16 t ha–1, which was
worldwide. significantly higher than the yield of PSBRc52. The higher
yield was associated with the higher aboveground total bio-
mass production and greater grain weight in IR72967-12-2-3.
Indica checks
PSBRc52 117 9.51 1,762 49.8 483 94.4 86.9 22.2
IR72 117 9.31 1,731 49.2 468 84.5 89.4 24.1
PSBRc54 119 9.06 1,766 47.2 412 106.0 84.1 22.7
IR8 133 8.72 1,858 41.8 379 94.8 76.0 28.4
PSBRc82 111 8.54 1,515 50.8 452 82.2 87.6 23.7
Mean 119 9.03 1,726 47.8 439 92.4 84.8 24.2
Table 2. Crop growth duration, grain yield, and yield attributes of second-generation new plant type lines and
check varieties grown at the IRRI farm in the dry season of 2004.
Indica checks
IR72903-121-2-1-2 124 9.32 1,907 44.7 376 118.6 79.1 24.2
PSBRc18 116 8.37 1,775 42.6 418 103.0 78.6 22.4
PSBRc52 111 8.04 1,612 43.3 466 95.3 78.8 20.0
IR68440-36-2-2-3 116 7.96 1,758 42.1 447 107.9 79.1 19.4
IR72 116 7.25 1,720 41.4 481 81.8 82.1 22.0
Mean 117 8.19 1,754 42.8 438 101.3 79.5 21.6
Overall, the second-generation NPT lines did not outyield the 2003 dry season because of lower solar radiation and higher
indica inbred check varieties. There was a small difference in minimum temperature in the 2004 dry season. Overall, there
biomass production and harvest index between the two groups. was no significant difference in grain yield, biomass produc-
Grain filling of the second-generation NPT lines was poorer tion, and harvest index between the second-generation NPT
than that of the indica inbred check varieties. The spikelet num- lines and indica inbred check varieties. As observed in the 2003
ber per panicle of the second-generation NPT lines was less dry season, grain filling of the second-generation NPT lines
than 115, and was on average 15% greater than that of the was poorer than that of the indica inbred check varieties. On
indica inbred check varieties. average, spikelet number per panicle of the second-generation
In the 2004 dry season, an indica inbred line (IR72903- NPT lines was 11% greater than that of the indica inbred check
121-2-1-2) produced the highest yield (Table 2). IR71700-247- varieties, which was just enough to compensate for the reduc-
1-1-2 recorded the highest harvest index in both seasons. The tion in panicle number per m2.
grain yield in the 2004 dry season was lower than that in the
Rice is the main staple food in China, being planted on 23% of pends on 7% of the world’s arable land, will not encounter any
the world’s rice area and contributing 37% of global rice pro- food security problems in the next few years, it has hidden
duction. Rice production in China plays an important role for risks of food shortage as its rapid economic growth eats away
the Chinese people and also affects the world rice market. Al- at arable land.
though China, where one-fifth of the world’s population de-
35
30
Area of rough rice
25 Area of hybrid rice
20
15
10
0
1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000
Year
Fig. 1. Area of hybrid rice (million ha).
Current situation of breeding to increase rice yield in China Table 1. Yield standards (t ha–1) set for super rice in China.a
A look back at crop breeding for super high yield Hybrid rice
A series of high-yielding semidwarf varieties has made dra- Phase Yield
matic progress in rice cultivation since the 1950s. Hybrid rice Early-season Late-season Single-season increase
indica indica rice
at the end of the 1970s marked another breakthrough in China’s
rice production (Yang et al 1996, Zhou et al 1997, Huang et al Before 1996 7.50 7.50 8.25 0
2003). For the first time in 1977, hybrid rice was produced in 1996-2000 9.75 9.75 10.50 >20%
farmers’ fields. Subsequently, the area of hybrid rice expanded 2001-2005 11.25 11.25 12.00 >40%
rapidly to 54% of the total rice area in 1997, and has remained aItis required that grain yield be up to standards at two locations of an ecological
at more than 60% since then (Fig. 1). area with a planting scale of 6.7 ha at each location for two consecutive years.
Despite this progress, China’s grain output dipped from
a record high of 512 million tons in 1998 to 435 million tons
in 2003. In view of the rising population and falling number of brid has met the standard of super high-yielding rice, though
hectares of land across the world, the most effective and eco- only in a yield trial in a small area. Pei’ai64S/E32 is character-
nomical way to increase production is considered to be grow- ized by a set of yield components that lead to a theoretical
ing hybrid rice. Hybrid rice, which has been grown in China yield of 13.95 t ha–1 and actual yield of 12.87 t ha–1. Several
for a long period, will continue to play a significant role in hybrids, such as Chuanxiangyou2, P088S/0293, and
guaranteeing China’s food security. IIyouming86, increased their yield by more than 20%.
Germplasm enhancement
The historic discovery of the semidwarfing gene (sd1) in
Notes
Taichung (Taiwan, China) revolutionized rice production in Author’s address: Institute of Crop Sciences, Chinese Academy of
the world, extending varieties carrying this gene to almost all Agricultural Sciences, Beijing 100081, China, e-mail:
the rice-growing countries. Following this model, favorable wanjm@caas.net.cn.
QTLs have been identified from a set of wild rice based on
molecular analyses and field experiments. Near-isogenic lines
carrying the QTLs are created by marker-facilitated backcross-
ing and selection.
Hybrid rice
All the rice hybrids grown so far are indica hybrids, except for
japonica hybrids in the northern part of China. Since the in-
dica × tropical japonica hybrids are shown to outyield indica
It is estimated that the world’s population will increase 1.4- to cumulated in the aboveground part at harvest was not exam-
1.5-fold by 2025, and this projected increase will be mostly in ined.
Asia (IRRI 1995). More than 90% of the world’s production Herein, we report on the achievement of a high yield of
of rice is in Asia. It is therefore crucial to increase rice produc- over 12 t ha–1 of rough rice with Akita 63 and analyze the
tion within a relatively short period. With little scope for ex- factors responsible for its high yield compared with the yield
panding the area of land available for rice cultivation, this in- of common reference cultivars. Finally, the yield-limiting fac-
crease in rice production must be achieved by an increase in tors of recent japonica cultivars are discussed in terms of physi-
yield from the land currently used for such cultivation. On the ological NUE.
other hand, interest is increasing in the environmental impact
of nitrogen (N) management practices (Cassman et al 1998).
Materials and methods
Therefore, it is desirable to have high-yielding rice cultivars
with high N-use efficiency for grain production. However, se- Akita 63 was grown with different levels of N supply (0–160
lection or breeding of such rice cultivars has not been seri- kg N ha–1) in an experimental field of the Agricultural Experi-
ously attempted (Ladha et al 1998, Ying et al 1998a,b). mental Station of Akita Prefecture, Japan (39°34′N, 140°11′E,
A new large-grain japonica-type cultivar of rice, Akita 16 m altitude), for three years (2000-02). As reference culti-
63, has recently been released by the Akita Agricultural Ex- vars of japonica rice, a high-yielding local cultivar
perimental Station in northern Japan and its grain size is 30– (Yukigesyou), a common cultivar (Toyonishiki), and a mod-
31 mg, which is 25–50% larger than that of current japonica ern cultivar (Akitakomachi) were also grown with the same N
cultivars. In test cultivation with standard levels of fertiliza- treatments. The soil type was a Gley soil (Eutric Gleysols; FAO)
tion at different locations in Akita Prefecture and in different with pH 5.3, 31.6 g total C kg–1, 2.4 g total N kg–1, and 25.2
years, the average yield of Akita 63 was 18% higher than that cmol kg–1 cation exchange capacity. Phosphorus (43 kg P
of Akitakomachi, which is the current leading cultivar of Akita ha–1 as fused phosphate) was applied to all plots 30 days be-
Prefecture (Masaki et al, unpublished data), indicating that fore transplanting. Potassium was not supplied because of its
Akita 63 might be a new type of high-yielding cultivar with sufficient level in the soil. Two to four levels of N fertilization
high N-use efficiency. Although high productivity with large were applied by using ammonium sulfate and controlled-re-
grain cultivars of rice has been previously reported (Kobayashi lease fertilizers (urea LP100 type and LPS100 type polyolefin-
et al 1990, Wang et al 1995), their physiological (internal) N- coated, Chisso Co., Japan) (Table 1). A basal application was
use efficiency (NUE) defined as grain yield per unit of N ac- conducted a week before transplanting. Weeds, insects, and
Co., Japan). cControlled-release fertilizer (urea LSP100 type polyolefin-coated, Chisso Co.,
Japan).
1,200
900
Akita 63 in 2000
Akita 63 in 2001
600 Akita 63 in 2002
Yukigeshou in 2000
Toyonishiki in 2001
Akitakomachi in 2000
Akitakomachi in 2002
Akita 63: y = 48.32x + 276.7
300 R2 = 0.853
Reference y = 52.89x + 91.1
cultivars: R2 = 0.961
0
0 5 10 15 20 25
m2)
Plant nitrogen content (g
Fig. 1. Relationship between panicle dry weight and plant nitrogen content
per unit land area at harvest in Akita 63 and the reference cultivars.
diseases were controlled as required to avoid yield loss. All Results and discussion
the experimental plots were arranged in a randomized design
with three replicates, except in the case of Toyonishiki, for The grain yield of Akita 63 was 22–58% higher than that of
which there were only two replicates at the standard level and the reference cultivars for all N treatments throughout the three-
no application treatments in 2001. The area of each plot was year period. The yields were high in 2000 and 2001 and low in
26.25 m2 (3.5 m wide and 7.5 m long). About 30 five-day-old 2002, mainly because of the differences in climatic conditions
seedlings were transplanted with a rice transplanter on a day during the culture period among the three years. The highest
in the middle of May at a hill spacing of 0.3 × 0.14 m (24 hills yield was 12.8 t ha–1 of rough rice (9.83 t ha–1 of brown rice),
m–2) with 4–5 seedlings per hill. close to the highest yield of japonica rice (10.52 t ha–1 of brown
The sample plants were separated into leaf blades, culms rice) previously recorded in Japan (Honya 1989). The dry
plus leaf sheaths, panicles, and other parts (dead parts and non- weight of the aboveground part, the number of total spikelets,
reproductive tillers). Leaf area was measured with a leaf area and the amount of N accumulated in the aboveground part
meter (Type-AMM, Hayashi-Denko, Tokyo, Japan). All (plant N) per unit land area at harvest were almost the same
samples were oven-dried at 105 °C for several days, weighed, between Akita 63 and the reference cultivars in 2000 and larger
and powdered. At the time of harvest, plants from three addi- in Akita 63 than in the reference cultivars in 2001 and 2002
tional hills with the mean panicle number from each plot were for all the N treatments. However, the dry weight of the
collected and hand-threshed for measurement of the number aboveground part and the number of total spikelets for a given
of filled and unfilled spikelets, and grain weight. A survey of amount of plant N per unit land area at harvest did not differ
yield was carried out as follows: plants from 80 hills were col- between Akita 63 and the reference cultivars. Clear differences
lected from the center part of each plot. Winnowed unhulled in sink capacity (total spikelet number per unit land area × the
rice weight was measured after reaping, threshing, and wind weight of 1,000 kernels) and panicle dry weight for a given
selection. Unhulled rice of 80 hills was husked to obtain whole amount of plant N per unit land area were found between Akita
brown rice, and then put through a 1.8-mm sieve to remove 63 and the reference cultivars. The sink capacity was 30–40%
any immature kernels (brown rice). The weight of winnowed greater in Akita 63 than in the reference cultivars because the
unhulled rice and that of brown rice was adjusted to a mois- weight of 1,000 kernels was 29–36% greater in Akita 63. Fig-
ture content of 0.15 g H2O g–1 fresh weight. The harvest index ure 1 shows the relationship between panicle dry weight and
was defined as the ratio of panicle dry weight to the total dry plant N content per unit land area at harvest. The panicle dry
weight of the aboveground part. weight for a given amount of plant N per unit land area was
N content was determined with Nesler’s reagent after 12–43% greater in Akita 63 than in the reference cultivars.
Kjeldahl digestion of powdered samples with H2SO4 and H2O2. This was because the ratio of dry matter partitioning to the
References
Cassman KG, Peng S, Olk DC, Ladha JK, Reichardt W, Dobermann
A, Singh U. 1998. Opportunities for increased nitrogen-use
efficiency from improved resource management in irrigated
rice systems. Field Crops Res. 56:7-39.
The use of heterosis and hybrid rice has been one of the most eties can be used as restorers to produce hybrid F1. The possi-
important achievements in China. Hybrid rice occupies about bility to find good combinations is much greater than with the
15 million ha of fields, or half of China’s rice-planting area. three-line approach. Second-generation hybrids yielded 5–10%
The yield potential of hybrid rice is about 20% greater than more than three-line hybrids although their grain filling was
that of inbreds. Since its adoption in 1976, hybrid rice had poor.
been planted on 271 million ha of rice land until 2001 in China.
The increased grain from hybrid rice can feed 60 million people Super hybrid rice
annually. The Ministry of Agriculture of China established a mega project
China’s rice yield has experienced two quantum leaps on the development of super rice in 1996 (Min et al 2002).
since the 1950s (Peng et al 2004). The first one resulted from The objectives were
the extension of semidwarf rice varieties in the 1960s. Rice 1. To develop “super” rice varieties with a maximum
yield increased from 2.08 t ha–1 in 1961 to 3.14 t ha–1 in 1966, yield of 9–10.5 t ha–1 by 2000, 12 t ha–1 by 2005, and
with an annual increase of 0.21 t ha–1. The second leap was 13.5 t ha–1 by 2015 measured from a planting area of
brought about by the development of hybrid rice in 1976. Grain at least 6.7 ha.
yield increased from 3.50 t ha–1 in 1976 to 5.36 t ha–1 in 1984, 2. To develop “super” rice varieties with a yield poten-
with an annual increase of 0.23 t ha–1. To feed the increasing tial of 12 t ha–1 by 2000, 13.5 t ha–1 by 2005, and 15
population, China’s rice production must increase from 177 t ha–1 by 2015. These yields will be achieved in ex-
million tons in 2003 to 240 million t in 2030. Yield must in- perimental and demonstration plots.
crease from 6.27 to 7.66 t ha–1 during the same period. Hybrid 3. To raise the national average rice yield to 6.9 t ha–1
rice is expected to play an important role in the new increase by 2010 and 7.5 t ha–1 by 2030.
in rice yield potential. In addition, the super rice varieties should outyield the
locally wide-grown check varieties by 10%, or daily yield
achieve 100 kg ha–1, with acceptable grain quality and pest
Three generations of hybrid rice resistance. Significant progress has been achieved in China’s
Three-line hybrid rice super hybrid rice breeding in recent years.
The first-generation hybrids used intervarietal heterosis through
the three-line approach. Cytoplasmic genetic male sterility
Strategy for super hybrid rice breeding
(CMS) was used to facilitate hybrid seed production. Yuan
started his pioneering work on hybrid rice breeding in 1964. The strategy for super hybrid rice breeding was to combine
However, no significant progress was achieved until 1970, the ideotype approach with the use of intersubspecific hetero-
when the male sterile wild rice plant, “wild abortive,” was sis (Yuan 1997). This was different from the counterpart
found. In 1973, all three lines—male sterile lines (A lines), projects of other countries and/or organizations. According to
maintainer lines (B lines), and restorer lines (R lines)—were the principles of hybrid rice breeding, there are two ways to
made available. The heterosis of hybrid rice was shown by achieve super high yield. One is to make full use of the domi-
faster tillering, greater leaf area index, higher dry matter pro- nant complementary effects of the two parents to improve the
duction, and a stronger root system. The three-line hybrids morphological characteristics of the hybrid. Another is to ex-
yielded 20% more than the best inbred check varieties. Plant- tend the genetic diversity of the parents to increase the hetero-
ing area of hybrids was 0.135 million ha in 1976, which ex- sis level. The heterosis of intersubspecific hybrids is much
panded rapidly to 8.84 million ha in 1984 and 16.6 million ha stronger than that of intervarietal ones. The use of
in 1990. intersubspecific hybrids is thus the most feasible approach for
super high yield. The discovery of P/TGMS and wide compat-
Two-line hybrid rice ibility (WC) genes made it possible to directly use
The yield of hybrid rice remained stagnant until the mid-1980s. intersubspecific heterosis. Several WC lines were developed
To break the yield barrier, the two-line approach using as “bridges” to overcome the too great genetic difference in
intersubspecific heterosis was proposed (Yuan 1997). Envi- typical intersubspecific hybrids, which resulted in physiologi-
ronment-sensitive genic male sterility, including photoperiod- cal barriers. The WC lines could freely cross with both indica
and thermosensitive sterility (P/TGMS), was employed in this and japonica varieties without problems in grain filling (Yang
approach. In contrast with the three-line system, normal vari- et al 1996).
Morphological design of super hybrid rice Performance of selected super hybrid rice
The poor grain filling of second-generation hybrids and More than 10 hybrid varieties that meet the super rice criteria
progress in IRRI’s new plant type breeding led Chinese scien- have been released by the provincial or national seed board
tists to pay more attention to morphology. Different versions (Table 1). The cumulated planting area of super hybrid rice is
of morphological design were developed according to regional estimated at 6 million ha.
ecological conditions. Well adopted are the “panicle-under- Lianyoupeijiu is a two-line intersubspecific hybrid (Lu
leaves” type for the Yangtze River region (Yuan 2001), the and Zou 2003). The female parent Pei’ai64S is an intermedi-
“heavy-panicle” type for western China (Zhou et al 1997), and ate type between indica and japonica, whereas the restorer 9311
the “rapid-early-growth-and-deep-root” type for South China is a typical indica line. In the 1997-98 yield trial, Lianyoupeijiu
(Huang 2001). achieved 9.56 t ha–1 and outyielded check variety Shanyou 63
Yuan (2001) suggested the panicle-under-leaves type by 7.75%. An average yield of 12.15 t ha–1 was achieved in a
with the following morphological traits: 67.8-ha area in Jiangsu Province in 2001.
O Plant height about 100 cm, with culm length of 70 Xieyou 9308 is a three-line intersubspecific hybrid
cm. (Cheng and Zhai 2000). The female parent Xieqingzao-A is a
O Top three leaves: typical indica sterile line. The restorer line Zhonghui 9308 is
1. Long: flag leaf is 50 cm and -2nd and -3rd leaves an intermediate type. Grain yield ranged from 10.53 to 11.95 t
55 cm in length. The top two leaves are higher ha–1 in farmers’ fields of 6.7 ha in Zhejiang Province during
than the top of the panicle. 1999-2001 and a yield of 12.23 t ha–1 was recorded in a 697-
2. Erect: leaf angles of the flag, -2nd, and -3rd leaves m–2 field.
are around 5°, 10°, and 20°, respectively. Liaoyou 1518 is a japonica hybrid (Hua et al 2003). The
3. Narrow and V-shaped: the leaves look narrow but female parent 151A is a japonica line. The male parent C418
they are 2 cm wide when flattened. is an indica line with good compatibility with the japonica va-
4. Thick: 55 g m–2 of specific leaf weight for the riety. In field trials in 2001, Liaoyou 1518 yielded from 8.37
top three leaves. to 12.08 t ha–1 and outyielded check variety Liaojing 454 by
O Plant type: moderately compact type with moderate 6.37–30.2%.
tillering capacity; droopy panicles, the panicle top
about 60 cm from the ground after filled; erect leaf
Common characteristics of released super hybrid rice
canopy without appearance of panicles.
O Panicle weight and number: grain weight is 5 g per The super hybrids have several characteristics in common:
panicle; 270–300 panicles m–2. O Many super hybrids, such as Lianyoupeijiu, Xieyou
O Leaf area index (LAI): the LAI of the top three leaves 9308, Yueza 122, Liaoyou 1518, and P88S/0293, use
is about 6. indica-japonica intersubspecific heterosis.
O Harvest index: about 0.55. O Large sink. The panicle size is significantly enlarged,
with a small reduction in panicle number m–2. Spike-
lets reach 35,000–45,000 m–2.
In most rice-growing countries in Asia, achieving self-suffi- lose content, etc.). In 2000, MCI registered the first commer-
ciency in rice production and maintaining price stability are cial japonica hybrid rice varieties, Mitsuhikari 2003 and
important. Since the 1980s, many Asian countries, and espe- Mitsuhikari 2005, which were bred by combining a conven-
cially China, have developed breeding of hybrid rice to in- tional breeding method and modern biotechnology in Japan.
crease yield. MCI started selling these seeds in Japan in 2002.
Since 1986, Mitsui Chemicals, Inc. (MCI) has devel- Mitsuhikari 2003 and 2005 possess high yield and good
oped a breeding program for a japonica hybrid rice variety eating quality. To gain wide acceptance in Japan, it is neces-
acceptable to Japanese markets. To breed a new variety pos- sary to detect suitable districts to make full use of these variet-
sessing both high yield and good eating quality, we aimed at ies. On the other hand, Mitsuhikari varieties have some unde-
traits such as plant type (erect leaf), ear type (panicle weight), sirable traits such as long culms and a long growing period. To
strength against lodging (stiffness), and grain quality (amy- improve these traits, we have continuously developed a hy-
Mitsuhikari
2003 7.1 140 91 26.3 283 2.11 5.5 3.0 19.5 8.2 70
2005 6.7 138 90 26.2 262 2.06 5.6 3.0 18.8 7.4 77
MH 3001 7.1 126 82 26.0 305 2.25 5.7 3.0 19.2 7.5 74
Check 5.5 120 91 17.7 400 2.14 5.2 3.0 19.0 7.9 75
(Koshihikari)
aValues mean average of 2001-02. Input dose of fertilizer was 100 kg N ha-1. bEating quality means the analyzed value of the grain quality inspector
(Satake RCTA11A).
brid rice breeding program. The fruit of our efforts, a new lose content, protein content, and eating quality with a rice
japonica hybrid rice line, MH 3001, appeared in 2002. quality inspector (Satake RCTA11A).
This study presents the productivity and prevalence of
both Mitsuhikari varieties and MH 3001. Performance test of Mitsuhikari
Since 1995, performance tests of Mitsuhikari varieties have
been carried out at the prefectural agricultural research center
Materials and methods
in seven districts, Kanto, Tokai, Kinki, Chugoku, Shikoku,
Mitsuhikari varieties Kyusyu, and Okinawa. Cropping practices followed those of
Mitsuhikari 2003 and 2005 were selected after the evaluation each district. The input doses of fertilizer were 70–120 kg N
of 800 different F1s crossed between Japanese varieties as fe- ha–1. Productivity was expressed by the yield of brown rice
males and japonica-like restorer lines from China as males. after screening.
They have erect leaves, panicle weight type, long growing
periods, long culms, and strength to resist lodging. For the
Results and discussion
easy production of F1 seeds, the female lines were cytoplas-
mic male sterile (CMS) lines produced through asymmetric Table 1 shows the yield performance and grain quality of
fusion (Akagi et al 1989). We have also developed DNA mark- Mitsuhikari 2003 and 2005 and MH 3001. They yielded nearly
ers such as PCR, RAPD, and microsatellites that can identify 30% more than Koshihikari. Their growing period to heading
CMS lines, restorer lines, and F1s (Akagi et al 1995, Ichikawa date was 30–40 days longer than that of Koshihikari. The
et al 1997). These markers are powerful tools for breeding panicle length of these hybrids was also about 10 cm longer
parental lines and certifying F1 seed purity. than that of Koshihikari, and their period from heading to
maturation took more than twice as long as that of Koshihikari.
MH 3001 The culm length of Mitsuhikari varieties and Koshihikari was
Since 1993, we have developed some dwarf restorer lines. From about 90 cm, and that of MH 3001 was about 80 cm. This trait
1998 to 2002, performance tests of F1s produced between of MH 3001 will be an advantage for tolerance of lodging.
japonica CMS lines and these restorer lines were practiced. The grain weight of MH 3001 was the heaviest among
One combination, whose productivity and eating quality were the four varieties. The length of brown rice of all the hybrids
the same as those of Mitsuhikari 2005, but whose growing was longer than that of Koshihikari. For grain quality, both the
period and plant height were shorter than those of Mitsuhikari amylose and protein contents of these hybrids were much the
2005, was selected. The F1 of this combination was named same as those of Koshihikari. Also, eating quality of the hy-
MH 3001. brids was similar to that of Koshihikari.
Table 2 shows the results of the performance test. In all
Yield performance and grain quality of the districts, the productivity of Mitsuhikari 2003 was higher
We assessed the productivity and grain quality of Mitsuhikari than that of Mitsuhikari 2005. Mitsuhikari varieties cropped
varieties and MH 3001, as compared with those of Koshihikari in the southern districts of Japan have outyielded the checks
(check variety) at MCI’s research paddy field in Ibaraki Pre- by more than 20%. This result means that these southern dis-
fecture from 2001 to 2002. Their productivity was expressed tricts are suitable for growing Mitsuhikari varieties to exploit
by brown rice yields after screening with a net of 1.75-mm their real productivity. We estimated that Mitsuhikari 2003 and
mesh. Also, agronomic traits such as culm length, panicle 2005 were cropped on about 1,000 ha last year, mostly in south-
length, panicles m–2, and both the weight and size of brown ern districts. On the other hand, the new hybrid line, MH 3001,
rice were measured. We analyzed grain quality such as amy- is earlier in heading than Mitsuhikari varieties. We expect that
District
Variety
Kanto Tokai Kinki Chugoku Shikoku Kyusyu Okinawa
Mitsuhikari
2003 6.9 (117) 7.3 (124) 7.4 (123) 7.5 (132) 7.2 (126) 6.9 (128) 4.3 (130)
2005 6.7 (114) 6.6 (112) 7.3 (122) 6.5 (114) 6.1 (107) 6.9 (128) 4.1 (124)
Check 5.9 (100) 5.9 (100) 6.0 (100) 5.7 (100) 5.7 (100) 5.4 (100) 3.3 (100)
aValues mean average of each district (1995-2003). Input doses of fertilizer were 70-120 kg N ha-1. Numbers in parentheses
this line will be accepted where Mitsuhikari 2003 and 2005 Ichikawa N, Kishimoto N, Inagaki A, Nakamura A, Koshino Y,
have not prevailed. MH 3001 is now under registration. Yokozeki Y, Oka M, Samoto S, Akagi H, Higo H, Shinjyo C,
Fujimura T, Shimada H. 1997. A rapid PCR-aided selection
of a rice line containing the Rf-1 gene which is involved in
References restoration of the cytoplasmic male sterility. Mol. Breed.
3:195-202.
Akagi H, Sakamoto M, Negishi T, Fujimura T. 1989. Construction
of rice cybrid plants. Mol. Gen. Genet. 215:501-506.
Akagi H, Nakamura A, Sawada R, Oka M, Fujimura T. 1995. Ge- Notes
netic diagnosis of cytoplasmic male sterile cybrid plants of
rice. Theor. Appl. Genet. 90:948-951. Author’s address: Mitsui Chemicals Incorporation, Agrochemicals
Division, Bio-product Department, 2893, Kaminoshima,
Azumamachi, Inashiki-gun, Ibaraki-ken 300-0732, Japan, e-
mail: atsushi.nakamura@mitsui-chem.co.jp.
The functional relations between the source capacity of leaves rent parent. FS1 resembles Fujisaka 5 for most plant types, but
and the sink capacity of panicles affect dry-matter production has panicles with few ripened grains because of incomplete
and determine rice yield. The source-sink relation has so far embryo sac formation (Yokoo 1984).
been studied on rice plants from which their panicles were Germinated seeds were sown on 21 April in a plastic
artificially removed. In rice, however, the role of panicle pho- box containing rice nursery soil. Two seedlings at the 6-leaf
tosynthesis is not neglected and the role of hulls as a nutrient stage were transplanted to a 1/5,000 area of a Wagner pot.
pool is important. We planned to analyze the source-sink rela- Plants were placed under natural conditions in a vinyl house
tion and the accumulation of dry matter in intact plants. Under from April to September 2000 at the Agricultural and Forestry
the assumption that the sink capacity of panicles is negligible, Research Center, University of Tsukuba.
we examined the source-sink relation in a female-sterile line At 10 days before heading and during the period from
of rice (FS1), and characterized the effects of losing the sink heading to maturity (42 d after heading), 6 plants were har-
function of panicles on dry-matter production and nitrogen vested every 7 d and divided into 6 parts: roots, culms plus
distribution among plant organs. leaf sheaths (stems), leaf blades, dead leaves, panicles, and
late tillers that appeared after heading. All the samples were
oven-dried at 80 °C for 72 h and their dry weights were deter-
Materials and methods mined. The dried plant materials were ground in a mill, and
A female-sterile line (FS1) and a normal fertile line (Fujisaka their carbon and nitrogen contents were determined by a C-N
5) of rice (Oryza sativa L.) were used. FS1 is a progeny from Corder (MT600, Yanaco, Japan). The amounts of carbon and
the cross between Japanese variety Fujisaka 5 and Indonesian nitrogen of each organ were determined by multiplying dry
variety Tjina four times backcrossed by Fujisaka 5 as a recur- weights by each content.
60
50
40
30
20
10
0
10 0 7 14 21 28 35 42 10 0 7 14 21 28 35 42
Days after heading Days after heading
Fig. 1. Changes in dry weights of each organ in FS1 and Fujisaka 5 during the ripening stage.
0.4
0.3
0.2
0.1
0.0
10 0 7 14 21 28 35 42 10 0 7 14 21 28 35 42
Days after heading Days after heading
Fig. 2. Changes in nitrogen accumulation of each organ in FS1 and Fujisaka 5 during the ripening stage.
5. The partitioning ratio to stems also declined rapidly during amounts of carbon and nitrogen in stems than Fujisaka 5. This
the period from 7 to 21 d after heading. On the contrary, in suggests the possibility of using FS1 for biological production
FS1, the decline in nitrogen partitioning ratio to leaf blades as a forage crop. However, the sink capacity for nitrogen in
was slower, and the partitioning ratio to panicles was almost panicles was so strong that vegetative organs such as stems
steady. The nitrogen partitioning ratio to stems increased gradu- and late tillers could not fulfill it. The difference between the
ally until 21 d after heading, and then decreased rapidly to dry-weight and nitrogen distribution patterns in organs caused
21% at maturity with the apperance of late tillers. The parti- the change in carbon-nitrogen balance during the ripening stage.
tioning ratio to roots decreased with ripening in Fujisaka 5, The present results on dry-matter production and nitrogen ac-
but that of FS1 remained comparatively higher. While panicles cumulation were obtained from the limited conditions for root
accumulated 55% of the total nitrogen in Fujisaka 5 at matu- growth and nitrogen nutrition in pots. Further field experiments
rity, late tillers were the largest sink organ for nitrogen in FS1, under higher nitrogen nutrition levels are needed to evaluate
occupying 32% of the total nitrogen of a whole plant. the potential biomass productivity and nitrogen accumulation
of this female-sterile line of rice.
Conclusions
References
FS1 did not decrease dry-matter production irrespective of the
loss of sink function of panicles. The slower decrease in nitro- Kato M, Kobayashi K, Ogiso E, Yokoo M. 2004. Photosynthesis
gen concentration in leaf blades caused retarded leaf senes- and dry-matter production during ripening stage in a female-
cence and the maintenance of relatively higher photosynthetic sterile line of rice. Plant Prod. Sci. 7:184-188.
activity in FS1 (Kato et al 2004). Its vegetative organs such as Yokoo M. 1984. Female sterility in an Indica-Japonica cross of rice.
Jpn. J. Breed. 34:219-227.
culms, leaf sheaths, late tillers, and roots substituted for panicles
as photosynthate and nitrogen sinks. The results of the changes
in dry-matter and nitrogen distribution indicated that, in the Notes
final ripening stage, the carbohydrate and nitrogen accumu-
Authors’ address: Graduate School of Life and Environmental Sci-
lated in stems rapidly translocated to late tillers. This unique
ences, University of Tsukuba, e-mail:
pattern of dry-matter and nitrogen distribution retarded the katomo@sakura.cc.tsukuba.ac.jp.
senescence of leaves and the decline in leaf photosynthesis,
and then contributed to the increase in dry-matter production
during the ripening stage. FS1 is considered to have larger
Dry matter production and yield are higher in rice cultivar filled with soil on 27 May. In this study, Nipponbare plants
Akenohoshi than in cultivar Nipponbare, primarily because of were treated with additional nitrogen fertilizer and cytokinin.
a smaller decrease in the rate of photosynthesis during the rip- Ammonium sulfate was applied to some pots on 1 September
ening stage of Akenohoshi (Jiang et al 1988). The rate of at a dose of 10 g per pot as additional nitrogen fertilizer (NF).
photosynthesis and nitrogen content during senescence are both For the treatment with cytokinin, 30 mL of a 10-4 M solution
closely correlated with leaf levels of ribulose-1,5-bisphosphate of 6-benzylaminopurine (BA), containing 0.05% Tween 20 as
carboxylase/oxygenase (Rubisco). During leaf senescence, a surfactant, was sprayed on the aboveground parts of each
Akenohoshi shows a smaller decrease in levels of Rubisco than hill at 2-day intervals from 5 September. Levels of Rubisco
Nipponbare, and we have previously demonstrated that and nitrogen were determined in the same flag leaf of the main
Akenohoshi maintains larger amounts of nitrogen in leaves stem. Levels of Rubisco were determined by the single radial
during ripening. This may account for the differences in the immunodiffusion method. Nitrogen was quantitated using a
ability of these two cultivars to maintain high leaf levels of CN analyzer (MT-600, Yanaco Inc., Kyoto, Japan). For quan-
Rubisco (Ookawa et al 2003). The high leaf nitrogen content tification of mRNA, the flag leaves on the main culms were
in Akenohoshi results from both greater total accumulation of collected between 1000 and 1100 on a clear day and were im-
nitrogen and greater partitioning of nitrogen to leaves. mediately frozen in liquid nitrogen. Levels of rbcL and rbcS
It was also observed that larger amounts of cytokinin mRNA were determined by Northern blotting analysis.
were transported from the roots to the aboveground parts of
the plant during the ripening stage by Akenohoshi and
Nipponbare (Soejima et al 1995). Cytokinin can delay senes-
Comparison of Rubisco, nitrogen, and rbcL
cence and recent studies have illustrated the importance of
and rbcS mRNA levels in leaves
cytokinin in the control of senescence (Gan and Amasino 1995). Changes in Rubisco and rbcL and rbcS mRNA levels in leaves
Cytokinins might also contribute to the maintenance of high during ripening were compared between the two cultivars. At
Rubisco content. However, it remains to be determined whether the heading stage, there were no differences between the two
cytokinin suppresses a decline in Rubisco content. cultivars, but, after heading, levels of Rubisco and rbcL and
Several studies have reported that cytokinin can induce rbcS mRNA remained higher in Akenohoshi than in Nipponbare
the expression of photosynthetic genes and promote protein (Fig. 1A, C, and D).
synthesis. It can be assumed that suppressing the decrease in The nitrogen content in flag leaves was higher in
accumulation of Rubisco gene transcripts by cytokinin results Nipponbare and Akenohoshi at the heading stage. After head-
in the maintenance of high Rubisco levels during leaf senes- ing, the nitrogen content decreased more slowly in Akenohoshi
cence. Cytokinin also affects nitrogen partitioning in the whole than in Nipponbare (Fig. 1B).
plant (Jordi et al 2000). High nitrogen partitioning to leaves
by cytokinin probably results in the maintenance of high
Effects of treatment with 6-benzylaminopurine
Rubisco levels during senescence.
and nitrogen fertilizer on Rubisco, nitrogen, and rbcL
In the following experiments, our main aims were there-
and rbcS mRNA levels in leaves
fore (1) to compare the levels of Rubisco, nitrogen, and rbcL
and rbcS transcripts between Nipponbare and Akenohoshi, (2) In Nipponbare controls, Rubisco content in flag leaves de-
to analyze the effects of exogenous cytokinin on levels of creased with time after heading. Rubisco content remained
Rubisco, nitrogen, and rbcL and rbcS transcripts in leaves of substantially higher in plants treated with NF or BA in
rice during the ripening stage, and (3) to analyze the effects of Nipponbare than in the controls (Fig. 1A). Levels of rbcL and
exogenous cytokinin on the nitrogen content of leaves by de- rbcS mRNA decreased with time in the controls. However,
termining nitrogen absorption and partitioning to various or- levels remained high in plants treated with NF or BA during
gans in the whole plant. ripening (Fig. 1C, D).
In the controls, nitrogen content declined with time in
flag leaves. In contrast, nitrogen content decreased but re-
Methods mained relatively high during ripening when compared with
Rice seedlings (Oryza sativa L., cvs. Nipponbare and the controls (Fig. 1B).
Akenohoshi) were transplanted to Wagner pots (1/2,000 a)
Relative level of rbcL mRNA (%) Relative level of rbcS mRNA (%)
120 C 120 D
b c
100 b 100
c
80 b a 80
a c
60 a 60 bc
a b b
40 40
a a
20 20
0 0
31 Aug 10 Sep 20 Sep 30 Sep 31 Aug 10 Sep 20 Sep 30 Sep
Date Date
Fig. 1. Effects of 6-benzylaminopurine (BA) and additional nitrogen fertilizer (NF) application on the Rubisco
content (A), nitrogen content (B), and relative levels of rbcL mRNA (C) and rbcS mRNA (D) in flag leaves of
Nipponbare (NI). The plants of Akenohoshi were grown in the same conditions as the control for Nipponbare.
Vertical bars represent standard deviations (n = 3). Symbols with different letters are significantly different
at the 5% level (LSD).
Whole plant
Control 592.4 ± 35.1 635.7 ± 19.3 a 43.3 100 100
BA – 653.9 ± 15.8 a 61.5 – 100
NF – 1,024.6 ± 6.2 b 432.2 – 100
Leaves
Control 292.4 ± 12.1 142.9 ± 4.9 a –149.5 49.4 ± 1.0 22.6 ± 0.3 a
BA – 238.7 ± 3.2 b –53.7 – 36.5 ± 1.3 b
NF – 307.2 ± 3.5 c 14.8 – 30.0 ± 0.6 c
Culms + leaf sheaths
Control 202.0 ± 18.2 114.1 ± 3.1 a –87.9 34.0 ± 1.0 17.9 ± 0.5 a
BA – 115.2 ± 0.3 a –86.8 – 17.6 ± 0.4 a
NF – 202.3 ± 12.2 b 0.3 – 19.7 ± 1.0 b
Panicles
Control 58.2 ± 1.6 328.7 ± 8.9 a 270.5 9.8 ± 0.4 51.5 ± 0.7 a
BA – 259.7 ± 18.8 b 201.5 – 39.8 ± 2.0 b
NF – 459.4 ± 3.1 c 401.2 – 44.8 ± 0.5 c
Roots
Control 39.8 ± 4.2 50.1 ± 7.4 a 10.3 6.8 ± 0.4 8.0 ± 1.0 a
BA – 40.3 ± 1.6 b 0.5 – 6.2 ± 0.4 b
NF – 55.7 ± 0.6 a 15.9 – 5.5 ± 0.0 b
aData represent means ± standard deviations of results from three replicates. Nitrogen partitioning is expressed as the nitrogen content
of the indicated organs as a percentage of the total nitrogen content of the whole plant. Means followed by different letters are
significantly different at the 5% level of probability (LSD).
Factor 2
Materials and methods 2108.617
Fifteen rice varieties (three popular varieties, five high-yield- Factor 3 0.5
ing varieties, two hybrid rice varieties bred in Japan, four new Factor 4 0.1290323
South
Fig. 2. The calculation of leaf area receiving direct sunlight of Hitomebore about 20 days before
heading. The canopy consists of six plants and the target plant is in the center of the back row. The
light-receiving triangles and the shaded triangles are drawn by thick solid lines and thin dotted lines,
respectively.
by computer graphics. The characteristics of leaf distribution tigated whether each triangle into which a leaf was divided
of three rice varieties (Hitomebore, Milyang 23, and received direct sunlight or not. When the center of gravity of
Mitsuhikari 2005) were compared by this method. one triangle was in the inside of the other triangle, we judged
that they overlapped each other. In addition, the triangle near
Results the sun was recorded as sunny and the other was recorded as
The differences of leaf distribution among three varieties were shade. After this calculation was performed for the combina-
not clear before heading, but, as the developmental stage ad- tion of all triangles, the percentage of leaf area receiving di-
vanced, the characteristics of leaf distribution became more rect sunlight was determined.
prominent. At the ripening stage, Hitomebore had many slant-
ing leaves that protruded outward in the middle height. In Light-receiving leaf area of the rice canopy
Milyang 23, leaves were more erect at the upper height than in In our experimental paddy field, rice plants were planted at 24
Hitomebore. Mitsuhikari 2005 was taller than the others, and by 24 cm. The three-dimensional coordinates of the leaves of
its leaves stood straight at about every height and were densely five plants that were placed at 24-cm intervals (east, south-
distributed near the vertical center line of the plant (Oka et al east, south, southwest, and west of the target plant) were cal-
2003). culated. By examining the mutual shading among all triangles
of leaves of these five plants, the light-receiving efficiency of
the target plant was evaluated by the same method as for an
Calculation of light-receiving efficiency isolated single plant.
It is thought that the rice plant with ideal plant type has a larger The light-receiving efficiencies of three rice varieties
area of leaves that receive direct sunlight and shows a higher were compared. The result for Hitomebore before heading is
percentage of sunny leaf area to total leaf area. To evaluate the shown in Figure 2. When the angle of elevation of the sun was
light-receiving efficiency of the rice plant, we made computer 70 degrees and the azimuth angle was 15 degrees from the
software to calculate the area of light-receiving leaves when south to the east, the leaf area receiving direct sunlight was
the sun rose at any angle. 475 cm2 and the percentage to the total leaf area was 18%. At
the same angle of the sun, the percentages of Milyang 23 and
Light-receiving leaf area of an isolated single plant Mitsuhikari 2005 were 19% and 9%, respectively.
The three-dimensional structure of a rice plant was rotated in
the direction that was the viewpoint of the sun, and we inves-
Wrap-up of Session 4
Improving rice yield potential is a persistent research target for metabolism related to grain filling at the molecular level
food security. Its significance is magnifying with the increasing will help to manipulate sink strength (R. Ohsugi).
need for improving the efficiency in the use of land, water, and 2. It is clear that enhancing biomass production without
other natural resources, and labor for rice culture. With the reducing harvest index would be an effective way to im-
completion of the rice genome sequence, more tools are be- prove rice yield potential. The increase in biomass pro-
coming available for understanding the genetic control of plant duction will largely depend on improving radiation-use
functions. Considering this background, this session had two major efficiency and canopy photosynthetic rate. Improving
aims: Rubisco efficiency would have great importance as a
1. To elucidate the key physiological and morphological prospect for enhancing N-use efficiency (A. Makino).
traits to be targeted to improve yield potential by sum- The introduction of C4 photosynthesis is a challenging
marizing research carried out after the Green Revolu- option (J. Sheehy). The results of experiments under
tion. elevated CO2 conditions suggested that increased
2. To characterize recently developed high-yielding genetic canopy photosynthesis capacity increased biomass and
resources, including hybrid rice, and to evaluate the yield, but to a lesser extent than single-leaf photosyn-
possible contribution of molecular breeding to the im- thesis (T. Hasegawa). The relationship among leaf- and
provement of rice yield potential. canopy-level gas exchange and crop growth needs to
Six papers were presented on physiological processes that be studied. During the entire crop-growing season, en-
govern rice yield potential and five papers on breeding for in- hancement of biomass production during the ripening
creased rice yield potential. The following key points and future phase would be most effective in improving rice yield
prospects can be extracted from the papers presented: potential, followed by high biomass production during
1. A balanced increase in source and sink capacity is the the reproductive growth stage. Significant genotypic
key to enhancing rice yield potential since it is obvious variation has been found in biomass production at these
that source and sink capacity interact in determining stages. Further understanding is needed on the mecha-
yield (P.K. Mahapatra, S. Peng). Understanding and nisms through which sink strength affects source-re-
quantifying sink strength are a major challenge in ex- lated ability, for example, photosynthesis and stomatal
plaining the differences among genotypes in conductivity, and biomass production after the repro-
remobilization efficiency, grain filling, and harvest index. ductive stage.
High remobilization ability was shown to improve yield, 3. The interaction of yield determination with environment
harvest index, and N-use efficiency in new japonica va- needs further attention. The effectiveness of traits should
riety Akita-63, with large grain size (T. Mae). It is ex- be evaluated in adaptability to different environments.
pected that recent progress in understanding carbon For example, breeding for large panicle size with more
In the mid-1980s, upland rice varieties of tropical japonica long time. To create plant populations to be used in the subse-
extraction attained an area bigger than 4.5 million ha under quent steps of the program, recurrent selection schemes are
Brazilian savannas. Afterward, crop area decreased gradually widely used. Progenitor combination capacity is evaluated to
and markedly, attaining only 1.8 million ha in 2003. However, obtain more successful crosses. Pure lines are developed
production has not declined at the same level because average through population advances by mass selection within fami-
yield doubled in the period 1986-2003, rising from 1.1 to 1.9 lies. Some characteristics, such as plant type, grain appear-
t ha–1. This was due to both crop migration toward more fa- ance, growth cycle, and resistance to major diseases, are se-
vored areas in rain distribution, and the adoption of modern- lected for starting from early generations. Yield trials are car-
plant-type varieties, tropical japonica × indica derivatives. The ried out at three levels, encompassing observational trials, pre-
impact of improved-plant-type varieties and associated tech- liminary trials, and advanced yield trials under a multi-institu-
nologies in the favored localities of the savanna region put tional network composed of breeders and connected special-
forward a new concept of upland rice, leading to a new de- ists, located in various localities of the savanna region.
nomination—aerobic rice. The shift in breeding strategies over Table 1 presents relevant data on aerobic rice lines,
time and the genetic gain attained in the latter part of the 1990s pooled from the best advanced yield trials, 86 out of 198, of
are presented and discussed in this paper. the National Upland Breeding Network, conducted for three
years at representative sites of the savanna region.
The National Upland Rice Breeding Program
Main program strategies
Since its inception in 1974, the former National Research Cen-
ter for Rice and Beans (CNPAF) of Embrapa, located in Goiás, Breeding for blast resistance
has been developing and coordinating the National Upland Rice In spite of the difficulties in breeding for blast resistance, breed-
Breeding Program. The objectives and priorities of the pro- ers and pathologists were able to cooperate strongly since the
gram have changed from time to time to cope with the alter- initial steps of the program (Prabhu et al 1999). Donors pos-
ations in geographic distribution of the crop, cropping system, sessing ample spectra of resistance were selected for the Na-
and consumers’ taste for the product. According to Pinheiro tional Blast Nurseries conducted in nine locations throughout
(2003), it can be divided into three distinct periods: (1) phase the country, and widely used in crosses. Results of subsidiary
1, 1975-85, higher emphasis on drought tolerance, blast resis- tests for leaf blast performed by pathologists in field nurseries
tance, and yield stability, targeting exclusively the unfavored were provided to breeders before field selection, in which seg-
savanna conditions; (2) phase 2, 1985-90, breeding strategy regating lines are evaluated under high pressure for leaf and
expanded to include selection for high yield potential, target- panicle blast. Lines showing high scores in these environments
ing favorable savanna conditions as well; (3) phase 3, 1990 to are eliminated. Recurrent selection schemes have been used
the present, higher emphasis on blast resistance, grain appear- to generate resistant lines (Filippi et al 1994). The strong pri-
ance, and yield potential, targeting mainly favored conditions. ority for blast resistance still holds true in the aerobic rice breed-
Drought was a strong priority during phase 1 of the breed- ing program. But, although great care has been taken to pro-
ing program, which relied on progenitors of the tropical duce and release blast-resistant varieties, their average life is
japonica group, from both national and African origin, but will very short because of the high variability of the pathogen un-
not be covered in this paper. After the studies of Steinmetz et der the upland ecosystem. The severity of the disease decreases
al (1988a,b), the breeding strategy was expanded to obtain the sustainability of the exploitation.
genotypes to be grown under supplementary irrigation and in Two strategies are currently being used to cope with this
favored microregions for rainfall distribution during phase 2. problem: increasing the number of releases possessing con-
Progenitors of the indica group were involved in crosses, seg- trasting blast-tolerance genes and submitting the most promis-
regating generations were evaluated under supplementary ir- ing advanced lines to a side backcross scheme, to transfer dif-
rigation, and advanced lines tested exclusively in favorable ferent genes of blast resistance to a given line. When the resis-
locations. A noticeable move from the frontier land to more tance of the original line breaks down, various isolines would
favored areas for water distribution led to a decrease in the be available to replace it. This procedure is accompanied by
priority for drought tolerance in phase 3. Currently, drought monitoring of the prevailing blast races in the region.
evaluation results do not restrict the release of aerobic rice
varieties. Breeding for grain quality
Even with alterations in the program objectives and plant Grain quality was not a strong priority during phase 1 of the
ideotype, some basic strategies have been maintained for a breeding program because the long and bold grains of tradi-
tional upland varieties were considered the standard for qual- aerobic rice ideotype, that is, 90–110 cm of plant stature, 200–
ity. During the 1980s, a successful marketing strategy was pro- 250 tillers m–2, erect leaves, resistance to lodging, and yield
posed by the irrigated rice producers’ associations and indus- potential around 6 t ha–1 (Pinheiro 1999).
try in southern Brazil to change consumer preferences and in- A preference survey among rice millers indicated that
crease the national demand for long and slender grains. Visual BRS Primavera received preference ratings very close to those
selection for grain dimensions (length, width, thickness) and of BR IRGA 409, currently considered the most commercially
endosperm appearance is routinely made starting with the first competitive irrigated rice variety (Guimarães et al 2001). Most
segregating generations. To assist in the process, a chart was recent releases, such as BRS Talento and BRS Soberana, have
developed, based on standard varieties, which allows for quick grains that warrant ample acceptance by industry and consum-
field and laboratory scoring of grain dimensions. Starting with ers (Table 2). These varieties have cooking characteristics and
the F5 generation, individual plants are selected for intermedi- commercial qualities similar to those of American long-grain
ate values of both amylose content and gelatinization tempera- varieties.
ture. This routine methodology has been adapted to use
dehulled grains, thus reducing the sample size and decreasing Measuring advances of the breeding program
the factor of variation induced by the milling process. In the The estimation of genetic gain, considering the period 1995-
F6 generation, when grain availability is increased, grain trans- 2000, indicates a consistent decrease in growth cycle duration
lucency and milling recovery are also evaluated. The result of of 0.52 day year–1, which is desirable for the new profile of
this concentrated effort is a high proportion of lines possess- aerobic rice as a component of cropping systems. Besides the
ing long, slender, and translucent grains, with high to interme- reduced time of exposition to biotic and abiotic stresses, the
diate amylose content, low to intermediate gelatinization tem- short growth cycle opens up an opportunity to insert it as a
perature, and milling recovery higher than 55%. The complete succession crop after soybean, to benefit from the remaining
set of grain quality criteria is also applied during the yield part of the rainy season in the savanna.
testing program. Plant height decreased by 0.85 cm year–1. Till 1998, there
Although there have been releases of improved-plant- was a consistent decrease of 2.6 cm year–1, when the program
type varieties for the favored savanna conditions since the early was reoriented to include lines with an intermediate height,
1990s, the biggest impact was attained with varieties Maravilha which are more competitive with weeds. Lodging score has
and BRS Primavera, released in 1996. They were the first up- decreased at 4.16% per year because of both direct selection
land varieties that combine high grain quality and desirable and plant size reduction. Positive gains were also attained for
Session 5: Broadening the gene pool and exploiting heterosis in cultivated rice 155
Table 2. Grain quality data of representative aerobic rice varieties, com-
pared with those of irrigated1 and upland2 quality checks (int. = inter-
mediate).
panicle blast and grain discoloration, whose scores have de- Guimarães EP, Vieira NRA, Pinheiro BS. 2001. Breeding for spe-
creased at 1.59% and 1.81% per year, respectively. Although cialty rice in Latin America: status and perspectives. In:
the overall yield gain was only 0.37% per year, it was very Chaudhary RC, Tran VD, editors. Specialty rices of the world:
significant when only the favored region of the Brazilian sa- breeding, production and marketing. Rome: FAO. p 317-322.
Pinheiro BS. 1999. Características morfológicas da planta
vanna was considered, attaining a value as high as 1.85% per
relacionadas à produtividade. In: Vieira NRA, Santos AB,
year in the state of Mato Grosso.
Sant’ana EP, editors. A cultura do arroz no Brasil. Santo
The introduction of indica germplasm to the genetic base Antônio de Goiás (Brazil): Embrapa Arroz e Feijão. p 116-
of the traditional upland tropical japonica population, associ- 147.
ated with strong selection pressure for grain characteristics, Pinheiro BS. 2003. Integrating selection for drought tolerance into a
probably imposed a restriction on possible gains for drought breeding program: the Brazilian experience. In: Fischer KS,
tolerance. The first aerobic rice releases do not possess the Lafitte R, Fukai S, Atlin G, Hardy B, editors. Breeding rice
same level of drought tolerance as the traditional upland vari- for drought-prone environments. Los Baños (Philippines):
eties, derived from crosses within only the japonica group International Rice Research Institute. p 75-83.
(Pinheiro 2003). However, the program now holds a represen- Prabhu AS, Filippi MC, Ribeiro AS. 1999. Doenças e seu controle.
tative number of elite lines, japonica versus indica derivatives, In: Vieira NRA, Santos AB, Sant’ana EP, editors. A cultura
do arroz no Brasil. Santo Antônio de Goiás (Brazil): Embrapa
providing a sound basis for further improving yield, blast re-
Arroz e Feijão. p 262-307.
sistance, and drought tolerance.
Steinmetz S, Reyniers FN, Forest F. 1988a. Caracterização do re-
gime pluviométrico e do balanço hídrico do arroz de sequeiro
Concluding remarks em distintas regiões produtoras do Brasil: síntese e
interpretação dos resultados. EMBRAPA-CNPAF Documentos
The alteration in plant type and grain quality of upland rice, 23. Goiânia (Brazil): EMBRAPA-CNPAF. v. 1. 66 p.
toward its evolution to aerobic rice, led to an increase in yield Steinmetz S, Reyniers FN, Forest F. 1988b. Caracterização do re-
potential and national market acceptance. Its increased par- gime pluviométrico e do balanço hídrico do arroz de sequeiro
ticipation in grain cropping systems, under no-tillage or mini- em distintas regiões produtoras do Brasil: catálogo básico de
mum tillage, is foreseen. Moreover, aerobic rice could greatly dados. EMBRAPA-CNPAF Documentos 23. Goiânia (Bra-
contribute to environmental sustainability, avoiding new sa- zil): EMBRAPA-CNPAF. v. 2. 278 p.
vanna deforestation, when used in the rice-pasture association,
to renew the extensive areas of degraded pasture in the Brazil- Notes
ian savannas.
Author’s address: Embrapa Arroz e Feijão, Caixa Postal 179, 75375-
000 Santo Antônio de Goiás, GO, Brazil, e-mail:
References beatriz@cnpaf.embrapa.br.
Filippi MC, Prabhu, AS, Neves PCF, Notteghen JL. 1994. Eficiência
da seleção recorrente sobre a resistência parcial à brusone em
arroz de sequeiro. Fitopatol. Bras. (Brasília): 19:279.
World rice production has more than doubled, from 257 mil- Alien gene introgression
lion tons in 1966 to 600 million t in 2000. To meet the grow-
ing needs of the human population, rice production must in- The main objectives of our wide hybridization program are
crease by 25% during the next 20 years. Further, several biotic (1) to broaden the gene pool of rice by transferring useful genes
and abiotic stresses adversely affect rice productivity. Some for resistance to major diseases and insects and tolerance of
of the major diseases and pests affecting rice production are abiotic stresses and to enhance the grain yield of rice through
bacterial blight (BB), blast, sheath blight, tungro virus disease, the introgression of QTLs/yield-enhancing loci from wild spe-
and rice yellow mottle virus (RYMV), and insects such as cies, (2) to tag alien genes/QTLs introgressed from wild spe-
brown planthopper (BPH), stem borer, and Asian and African cies with molecular markers for use in marker-assisted selec-
gall midge. Similarly, abiotic stresses such as drought, cold, tion (MAS), (3) to characterize alien introgression using mo-
salinity, acidity, iron toxicity, and submergence reduce rice lecular cytogenetic techniques, and (4) to isolate agronomi-
production. Changes in insect biotypes and disease races are cally important genes/QTLs using BAC libraries of Oryza.
becoming a continuing threat to rice production. There is thus Some accomplishments made in collaboration with na-
an urgent need to broaden the rice gene pool through intro- tional agricultural research and extension systems (NARES)
gression of genes from new and diverse sources. Wild species partners and advanced research institutes (ARI) are discussed
are an important reservoir of useful genes for tolerance of bi- below.
otic and abiotic stresses.
Production of interspecific hybrids,
alien introgression lines, and mapping populations
Useful traits of Oryza species Direct crosses and embryo rescue techniques have been used
The genus Oryza has two cultivated and 22 wild species. Of to successfully produce hybrids between rice (AA) and all other
the two cultivated species, O. sativa (2n=24 AA), commonly wild species (except O. schlechteri). Backcrossing with the
referred to as Asian rice, is grown worldwide, whereas O. recurrent rice parent is used to produce fertile progenies. In-
glaberrima (2n=24 AA), “African rice,” is cultivated in a lim- trogression lines (2n=24) have been produced from crosses of
ited area in West Africa. The wild species have 2n=24 or 48 AA with various wild species representing AA, BBCC, CC,
chromosomes representing 10 genomic types (AA, BB, BBCC, CCDD, EE, FF, GG and HHJJ genomes except for O. sativa
CC, CCDD, EE, FF, GG, HHJJ, and HHKK). The wild species (AA) × O. coarctata (HHKK) (Brar and Khush 2002).
germplasm maintained in the IRRI genebank contains 2,500 Doubled haploids (DH) have been produced from O.
accessions of AA-genome types, 500 accessions of O. sativa × O. glaberrima. Similarly, recombinant inbred lines
officinalis complex, 35 accessions of O. meyeriana complex, (RILs) from crosses of rice with A-genome wild species and
21 accessions of O. ridleyi complex, 19 accessions of O. near-isogenic lines (NILs) derived through backcrossing car-
brachyantha, and 1 accession of O. schlechteri. rying small segments from distant genomes serve as mapping
Several incompatibility barriers, such as low crossabil- populations. Segregating populations derived from crosses of
ity, increased sterility, and limited recombination between chro- alien introgression lines × recurrent rice parents are also used
mosomes of wild and cultivated species, limit the transfer of in mapping genes/QTLs.
useful genes (Brar and Khush 1986, 2002). Recent advances
in tissue culture and genomics have enabled the production of Production of monosomic alien addition lines
wide hybrids among distantly related species and allowed re- (MAALs) and chromosome segmental substitution
searchers to precisely monitor the introgression of chromo- lines (CSSL)
some segments from wild into cultivated species. More recently, MAALs (2n=25) have been produced representing 6 to 10
BAC libraries of wild species representing 10 different ge- chromosomes from 7 wild species (CC, BBCC, CCDD, EE,
nomes have been developed by the Arizona Genomics Insti- FF, GG, and HHJJ genomes). CSSL are being developed from
tute in Tucson (Rod Wing, personal communication, 2004), O. rufipogon, O. longistaminata, and O. glaberrima in the
which offer new opportunities for map-based cloning of use- background of O. sativa using molecular markers. These CSSL
ful genes/QTLs. are important resources in fine mapping of genes and func-
tional genomics.
Session 5: Broadening the gene pool and exploiting heterosis in cultivated rice 157
Table 1. Progress in the transfer of agronomically important genes tiva × O. officinalis were released as BPH-resistant varieties
from wild Oryza species into cultivated rice at IRRI.a (MTL98, MTL103, MTL105, MTL110, MTL114) in Vietnam.
Donor Oryza species
Recently, variety Dhanrasi (IE15358) from the cross of B32-
sel-4 × O. rufipogon was released by the Directorate of Rice
Trait Wild species Genome Accession Research in Hyderabad, India (T. Ram, personal communica-
number tion, 2004).
Some of the indica alien introgression lines developed
Transferred to O. sativa
Grassy stunt resistance O. nivara AA 101508
at IRRI have shown a wide spectrum of resistance to BPH and
Bacterial blight resistance O. longistaminata AA – blast in Korea. These lines are being used by the Rural Devel-
O. officinalis CC 100896 opment Administration (RDA) to broaden the gene pool of
O. minuta BBCC 101141 japonica rice.
O. latifolia CCDD 100914 We are evaluating advanced progenies derived from dif-
O. australiensis EE 100882
O. brachyantha FF 101232
ferent wide crosses for resistance to stem borer and tolerance
Blast resistance O. minuta BBCC 101141 of aluminum toxicity, P deficiency, and iron toxicity. Some O.
Brown planthopper O. officinalis CC 100896 rufipogon accessions resistant to sheath blight have been iden-
resistance O. minuta BBCC 101141 tified and are being used to produce advanced introgression
O. latifolia CCDD 100914 lines. Progenies derived from O. sativa × O. glaberrima have
O. australiensis EE 100882
Whitebacked planthopper O. officinalis CC 100896
shown increased tolerance of iron toxicity under field condi-
resistance tions. Advanced introgression lines are under evaluation for
Cytoplasmic male sterility O. perennis AA 104823 the transfer of weed competitive ability from O. glaberrima
O. glumaepatula AA 100969 into O. sativa.
Tungro resistance O. rufipogon AA 105908
O. rufipogon AA 105909
O. rufipogon AA 106423
Enhancing the yield potential of rice
Wild species are phenotypically inferior but are valuable ge-
Introgression lines under evaluation netic resources for enhancing the yield potential of rice (Xiao
Yellow stem borer O. longistaminata AA 110404 et al 1998). Preliminary results at IRRI from the crosses of
O. rufipogon AA – new plant type (NPT) × O. longistaminata and IR64 × O.
Sheath blight resistance O. minuta BBCC 101141
O. rufipogon AA –
rufipogon indicate possibilities to improve rice grain yield by
Increased elongation ability O. rufipogon AA CB751 introgressing yield-enhancing loci/QTLs.
Tolerance of acidity and iron O. glaberrima AA Many
and aluminum toxicity O. rufipogon AA 106412 Molecular mapping of introgressed genes/QTLs
O. rufipogon 106423 and characterization of introgression
Resistance to nematodes O. glaberrima AA Many
The introgressed genes Bph10, Pi9(t), and Xa21 have been
aModified from Brar and Khush (2002). mapped. One of the genes, Xa21 for BB resistance, has been
used via MAS in gene pyramiding at IRRI and by NARES in
the Philippines, India, China, and Thailand. Genes introgressed
Alien gene transfer and varietal development for BPH and tungro tolerance are being mapped for use in
Genes for resistance to BPH, BB, blast, and grassy stunt and MAS. One of the major QTLs for tolerance of aluminum tox-
tungro virus, tolerance of acidity, and cytoplasmic male steril- icity introgressed from O. rufipogon has been mapped on chro-
ity have been introgressed from different A-genome wild spe- mosome 3, which is conserved across other cereal species
cies, including distantly related genomes (CC, BBCC, CCDD, (Nguyen et al 2003).
EE, FF), into elite breeding lines of rice (see Brar and Khush Molecular marker analysis has revealed limited intro-
2002, Multani et al 2003, Table 1). Some of these lines carry- gression of small chromosome segments from distant genomes
ing genes from wild species for resistance to BPH and tungro (FF, GG) of Oryza into rice. However, introgression among
and tolerance of acid sulfate soil conditions have been released cultivated rice and A-genome wild species is frequent for all
as commercial varieties. 12 chromosomes. We are developing genome-specific clones
During 2002, three varieties were released from wide using representational difference analysis (RDA) to charac-
crosses. One of the breeding lines, IR73678-6-9-B from IR64 terize alien introgression.
× O. rufipogon, was released as variety AS996 for commercial
cultivation in the Mekong Delta of Vietnam. This variety is Molecular characterization using GISH
grown on more than 100,000 ha. Another variety, Matatag 9 GISH techniques have been used successfully to characterize
(IR73385-1-4-3-2-1-6), has been released for cultivation in parental genomes, extra alien chromosomes, and
tungro-prone areas of the Philippines. Similarly, another high- homoeologous pairing between rice and genomes of several
yielding elite line (IR72102-4-159-1-3) from O. sativa × O. wild species, which were difficult to identify through conven-
longistaminata has been released as variety NSICRC112 in tional cytogenetic techniques.
the Philippines. Earlier, five IRRI breeding lines from O. sa-
Session 5: Broadening the gene pool and exploiting heterosis in cultivated rice 159
Panaud O, Vitte C, Hivert J, Muzlak S, Talag J, Brar DS, Sarr A. Notes
2002. Characterization of transposable elements in the genome
of rice (Oryza sativa L.) using representational difference Author’s address: International Rice Research Institute, DAPO Box
analysis (RDA). Mol. Genet. Genom. 268:113-121. 7777, Metro Manila Philippines.
Xiao J, Li J, Grandillo S, Ahn SN, Yuan L, Tanksley SD, McCouch Acknowledgment: The help provided by NARES partners, IRRI col-
SR. 1998. Identification of trait-improving quantitative trait leagues, and the wide hybridization team is duly acknowl-
loci alleles from a wild rice relative, Oryza rufipogon. Genet- edged.
ics 150:899-909.
The improvement of grain quality is one of the most important lopectin is synthesized by concerted reactions catalyzed by
subjects in rice breeding. For this, genetic resources of en- soluble starch synthase (SS), starch-branching enzyme (BE),
dosperm properties must be collected, characterized, and evalu- and starch-debranching enzyme (DBE), using ADP-glucose as
ated. In maize, various kinds of mutants for polysaccharides, a substrate. In addition, multiple isoforms were found in each
lipids, or proteins are known. These mutants greatly contrib- enzyme, that is, three isoforms of SS and BE, and two isoforms
uted to the improvement of the grain quality of maize, and of DBE in developing endosperm of rice, respectively. There-
thus expanded the use of maize not only as a food but also as fore, at least these eight genes are responsible for amylopectin
an important industrial material in food chemistry. In addi- biosynthesis in rice endosperm.
tion, they offered valuable information on the study of gene
action in the biological processes of metabolic regulation in
Mutants modifying amylose content
higher plants and led us to develop the novel genetic resources
improving these properties. To improve starch quality, novel genetic resources must be
Starch plays an important role in the eating quality of developed and evaluated. For this purpose, we need to eluci-
cooked rice and the processing quality of industrial uses. In date the genetic regulation mechanism of starch biosynthesis
rice, the mutants of starch properties had been little known and the interrelationships among the gene, enzyme, starch struc-
except for the waxy endosperm. We treated the fertilized egg ture, and rheological properties. A mutant is one of the most
cells of rice with MNU (Satoh and Omura 1979) and obtained helpful materials for this purpose.
several thousand rice mutants for embryo or endosperm prop- After the MNU treatment, many mutants with modified
erties (Satoh 1985), in which various kinds of mutants modi- amylose content were obtained in rice (Yano et al 1988). They
fying starch properties were found, as many as in maize (Satoh were classified into waxy type and dull type with lower amy-
and Omura 1981, Satoh et al 2003). lose content. Amylose content in endosperm starch varied from
0.3% to 12.8%, depending on the mutant line. Phenotypes of
dull mutants were intermediate between those of nonwaxy rice
Structure of starch and its biosynthesis and waxy rice, and there was good agreement between the phe-
Rice starch consists of two types of glucose polymers, 20% notype and amylose content. Thermal gelatinization of rice
amylose and 80% amylopectin. Amylose is an essentially lin- starch is influenced by amylose content. There is little differ-
ear molecule composed of α-(1,4)-linked glucosidic chains, ence in the initiation of gelatinization of endosperm starch
although recently another type of amylose possessing some among the waxy, dull, and wild type. However, the termina-
very short-branched chains was reported. Amylopectin is a tion temperature of gelatinization is remarkably different among
highly branched glucan with α-(1,6) glucosidic bonds that con- them. When the amylose content increases, the termination tem-
nect linear chains. Hizukuri (1986) proposed a cluster model perature of gelatinization of endosperm starch shifts to a high
for amylopectin. In this model, A and B1 chains form a single temperature. When starch is gelatinized in 4 M urea solution,
cluster, whereas B2 and B3 chains extend to two and three clus- granules with lower amylose content swell more. However,
ters, respectively. Amylopectin is much larger in molecular size these mutations did not alter amylopectin fine structure. These
than amylose, tenfold or more. The amylose to amylopectin facts indicate that amylose content affects the termination tem-
ratio and their structures greatly influence the rheological prop- perature of gelatinization of starch and the swelling power of
erties of starch. gel, but it has little influence on the initiation of gelatinization.
Amylose is synthesized by ADP glucose
pyrophosphorylase (AGPase) and granule-bound starch syn-
thase I (GBSSI), which is encoded by the Waxy gene. Amy-
Session 5: Broadening the gene pool and exploiting heterosis in cultivated rice 161
Jahan MS, Nishi A, Hamid A, Satoh H. 2002b. Variation in amy- Umemoto T, Yano M, Satoh H, Shomura A, Nakamura Y. 2002.
lopectin fine structure of Bangladesh rice cultivars. Rice Genet. Mapping of a gene responsible for the difference in amylopec-
Newsl. 19:72-74. tin structure between japonica-type and indica-type rice vari-
Nakamura Y, Kubo A, Shimamune T, Mastuda T, Harada K, Satoh eties. Theor. Appl. Genet. 104:1-8.
H. 1997. Correlation between activities of starch debranching Wong KS, Kubo A, Jane JL, Harada K, Satoh H, Nakamura Y. 2003.
enzyme and α-polyglucan structure in endosperms of sugary- Structures and properties of amylopectin and phytoglycogen
1 mutants of rice. Plant J. 12:143-153. in the endosperm of sugary-1 mutants of rice. J. Cereal Sci.
Nishi A, Nakamura Y, Tanaka N, Satoh H. 2001. Biochemical and 37:139-149.
genetic analyses of the effects of amylose-extender mutation Yano M, Okuno K, Satoh H, Omura T. 1988. Chromosomal location
in rice endosperm. Plant Physiol. 127:1-14. of genes conditioning low amylose content of endosperm
Satoh H. 1985. Genic mutations affecting endosperm properties in starches in rice, Oryza sativa L. Theor. Appl. Genet. 76:183-
rice. Gamma Field Symp. 24:17-37. 189.
Satoh H, Omura T. 1979. Induction of mutation by the treatment of
fertilized egg cell with N-methyl-N-nitrosourea in rice. J. Fac.
Agric. Kyushu Univ. 24:165-174. Notes
Satoh H, Omura T. 1981. New endosperm mutations induced by Authors’ addresses: Hikaru Satoh, Faculty of Agriculture, Kyushu
chemical mutagens in rice, Oryza sativa L. Jpn. J. Breed. University (Hakozaki, Fukuoka 812-8581, Japan); Ken-ichi
31:316-325. Ohtsubo, National Food Research Institute (Kan-nondai,
Satoh H, Nishi A, Fujita N, Kubo A, Nakamura Y, Kawasaki T, Okita Tsukuba, Ibaraki 305-8642; Yasunori Nakamura, Faculty of
WT. 2003. Isolation and characterization of starch mutants in Bioresource Science, Akita Prefectural University
rice. J. Appl. Glycosci. 50:225-230. (Shimoshinjo-Nakano, Akita-City 010-0195, Japan), e-mail:
Satoh H, Nishi A, Yamashita K, Takemoto Y, Tanaka Y, Hosaka Y, hdatoh@agr.kyushu-u.ac.jp.
Sakurai A, Fujita N, Nakamura Y. 2003. Starch branching
enzyme I-deficient mutation specifically affects the structure
and properties of starch in rice endosperm. Plant Physiol.
133:1111-1121.
Heterosis is known to be a major factor for increased produc- Extent of yield heterosis
tion in several crops, including rice. It has become the basis of
multibillion-dollar agribusiness worldwide. Hybrid varieties Data gathered at IRRI and in India, the Philippines, Bangladesh,
in maize, sorghum, sunflower, and rice added 90 million tons Indonesia, Vietnam, Myanmar, and Sri Lanka clearly provide
annually to global food production, which spared about 34 evidence for a 15–20% yield advantage of rice hybrids over
million ha of land from cultivation of these crops to meet glo- inbred high-yielding varieties (HYVs) (Virmani et al 2003).
bal demand (Duvick 1999). Hybrid rice technology, commer- Even in Korea and Egypt, where yield levels are considerably
cialized in China, has spread to about 50% of the 30 million higher, a yield advantage (10–15%, 1–1.5 t ha–1) of rice hy-
ha of rice area of that country, contributing 60% of its national brids over inbred HYVs has been reported (Virmani et al 2003).
paddy production. Inbred rice covers the other 50% of the area, Heterosis for yield in rice hybrids has been attributed to their
contributing only 40% of the national paddy production. In increased dry matter production caused by higher leaf area
China, rice hybrids (6.9 t ha–1) on average have outyielded index, higher crop growth rate, and increased harvest index
inbred rice (5.4 t ha–1) by 1.5 t ha–1 (Ma and Yuan 2003). Since resulting from increased spikelet number and increased grain
1979, IRRI has been exploring the prospects of exploiting weight. Parental lines developed at IRRI and shared freely with
heterosis in the tropics to increase rice yield potential in the national programs during the past 15–20 years have contrib-
light of increasing rice demand, decreasing land area and wa- uted significantly toward achieving these results. More than
ter resources for rice production, and the need for sparing rice 40 heterotic rice hybrids, developed by the public and private
land for crop diversification to increase farmers’ income. This sector, have been commercialized during the past ten years in
paper summarizes the progress made so far and discusses the some countries outside China. Those known to be derived from
future prospects of hybrid rice in the tropics. IRRI-bred parental lines are listed in Table 1. The yield ad-
vantage of rice hybrids over inbred HYVs was even higher
(20–30%) in farmers’ fields in the Philippines and India, per-
haps because of their homeostatic effects. During 2004, al-
aRecommended for release by the Rice Technical Working Group, Philippines, 25 November 2004.
most 1.4 million ha were covered with rice hybrids in India Breeding of parental lines to exploit heterosis in rice
(560,000 ha), the Philippines (200,000 ha), Vietnam (600,000
ha), Bangladesh (40,000 ha), Indonesia (5,000 ha), and Cytoplasmic male sterility and the fertility restoration system
Myanmar (5,000 ha). Rice hybrids have also been commer- have been primarily used to develop heterotic rice hybrids in
cialized in the United States (Walton 2003) and Japan (Takita and outside China (Virmani 1996). During the past decade,
2003). Several other countries (Sri Lanka, Thailand, Egypt, thermo- (TGMS) and photoperiod-sensitive genic male steril-
Brazil, Iran, Malaysia, and Pakistan) should be commercializ- ity (PGMS) systems have also been developed in China (Ma
ing this technology in the next 3–5 years. and Yuan 2003) and outside China (Virmani and Ilyas-Ahmed
Rice hybrids have also shown significant heterosis for 2001). Commercially usable parental lines to develop tropical
earliness. The higher yield coupled with slightly shorter dura- rice hybrids must have adaptability to tropical conditions. In
tion resulted in higher per-day productivity (68–78 kg d–1 addition, the male sterile lines used for this purpose must have
ha–1) than inbred HYVs (62–72 kg d–1 ha–1) (Virmani and complete and stable male sterility, and good outcrossing po-
Kumar 2004). The shorter duration of rice hybrids can result tential essential for economically viable hybrid seed produc-
in a savings of water and creation of niches where an addi- tion. Likewise, commercially usable restorer lines should re-
tional crop can be grown to allow crop diversification. Peng et store normal fertility in the derived hybrids and provide ad-
al (2003) observed higher N-use efficiency in an IRRI-rice equate pollen for a prolonged period for economically viable
hybrid—IR68284H (named as Mestizo 1 in the Philippines)— seed production. The TGMS system is the most effective for
than in a widely grown inbred HYV, IR72. Earlier studies at developing tropical rice hybrids; the PGMS system cannot be
IRRI had shown a significantly higher response of an IRRI used for lack of sufficient variation in daylength during the
hybrid (IR64616H, named as Magat in the Philippines) to the year (Virmani and Ilyas-Ahmed 2001). Tropical TGMS lines
application of N at booting than IR72. These results clearly possessing a low critical sterility point (CSP), complete and
show the increased production efficiency of rice hybrids com- stable male sterility in the wet and dry seasons, and good fer-
pared with inbred rice and illustrate how hybrid rice becomes tility reversion in available high-altitude locations are now
relevant in the current competitive global economic scenario. available at IRRI and in national programs.
Session 5: Broadening the gene pool and exploiting heterosis in cultivated rice 163
Among the first set of IRRI-bred CMS lines, IR58025A, lines (Cohen et al 2003). The most widely used WA-CMS
IR62829A, IR68888A, and IR68897A have been used to de- system was not found associated with susceptibility to blast,
velop several tropical commercial rice hybrids in India, the bacterial blight, brown planthopper, and whitebacked
Philippines, Vietnam, Indonesia, and Bangladesh (Table 1). planthopper. Nevertheless, the use of genetically diverse CMS
The most widely used CMS line, IR58025A, resulted in hy- systems is advocated to avoid potential genetic vulnerability
brids possessing aroma and somewhat sticky grains on cook- of rice hybrids in the tropics where disease/insect pressure is
ing. More recently, CMS lines possessing better grain quality high.
and higher outcrossing than IR58025A have been developed
(Virmani and Kumar 2004). Some of these also possess a dif-
ferent CMS system. These are being used extensively at IRRI
Hybrid rice seed production technology
and are shared with public and private research institutions in
and rural employment opportunities
national agricultural research and extension systems (NARES) Extensive research and seed production experiences in China,
to enable them to breed improved hybrids. at IRRI, and in other countries have helped in identifying guide-
Elite indica and indica/tropical japonica derivative lines lines and practices for hybrid rice seed production, which are
possess 30–40% restorer lines, which are usable as such to packaged in manuals (Virmani and Sharma 1993) and research
develop heterotic rice hybrids in tropical countries. Restorer papers (Mao 1988, Virmani 1996, Virmani et al 2002, Virmani
frequency in temperate and tropical japonica rice is negligible and Kumar 2004). A seed production manual is also available
and is very low in basmati rice. To develop rice hybrids in in CD form and can be accessed from the IRRI Knowledge
these genetic backgrounds, restorer lines have to be bred by Bank at www.knowledgebank.irri.org/hybridRiceSeed/
transferring Rf genes from indica rice. To increase the frequency hybridRiceSeed.htm. Using these practices, hybrid seed yields
of restorer and maintainer lines in the tropics, R × R and B × B ranging from 0.7 to 4 t ha–1 (average 1–1.5 t ha–1) have been
crosses are made routinely in IRRI’s hybrid rice breeding pro- obtained in the tropics. Seed yields are getting higher as seed
gram. The nuclear male sterility–facilitated recurrent selec- growers and their supervisors gain experience.
tion approach is also being used to develop maintainer and The production of hybrid rice seed is labor-intensive but
restorer composite populations from which B and R lines are economically viable enough to attract seed companies. More
extracted. These populations are also shared with national pro- than 60 seed companies in the public, private, and NGO sec-
grams to enable them to extract locally adapted B and R lines. tors are working in Asia to produce and market hybrid rice
Marker-aided selection, using STS markers found linked with seeds. Seed production operations (such as differential plant-
Rf3 and Rf4 genes, is being used at IRRI to increase the effi- ing of male and female rows, roguing, flag-leaf clipping, GA3
ciency of selecting fertility restorer lines for wild abortive application, separate harvesting and processing of pollen and
(WA), Dissi, and Gambiaca cytoplasm inducing male sterility seed parents, etc.) require at least 50 person-days of additional
in rice. labor compared with normal nonmechanized inbred rice culti-
Six IRRI-bred TGMS lines—IR68301S, IR73727-23S, vation. Increasing area under hybrid rice creates increasing
IR73834S, IR75589-31S, IR75589-41-13-17-15-3S, and demand for hybrid rice seeds, resulting in large areas of hy-
IR75589-41-13-17-15-22S—have been found to have a low brid seed production, which in turn creates additional rural
CSP and are being used to develop tropical two-line rice hy- employment opportunities in labor-surplus countries through
brids. Other tropical rice-growing countries (India, Vietnam, the seed industry. In labor-scarce countries, such as Japan and
and the Philippines) are also developing TGMS lines. Marker- the U.S., hybrid rice seed production technology is mecha-
aided selection has been used at IRRI to pyramid tms genes to nized (Walton 2003, Takita 2003).
develop more stable TGMS lines.
Future opportunities
Grain quality and biotic resistance of rice hybrids Opportunities exist to enhance heterosis through indica/tropi-
Acceptable grain quality of the commercial rice hybrids is es- cal japonica crosses since the level of heterosis of intervarietal
sential to ensure profitability to hybrid rice farmers. Hybridity groups is higher than in intravarietal groups (Virk et al 2003).
does not impair grain quality if the parents chosen to develop Recent results at IRRI have shown stronger heterosis for yield
heterotic hybrids possess acceptable grain quality (Khush et in indica/new plant type (NPT) crosses in which the NPT par-
al 1988). Therefore, a critical evaluation of parents for grain ent was derived from indica/tropical japonica crosses (Table
quality is necessary before using them in hybrid breeding. Simi- 2). The identification and use of heterotic groups and gene
larly, hybrids showing strong heterosis for yield should be blocks may also help to further enhance heterosis. Hybrid
evaluated critically for grain quality before their release and breeding efficiency is increased by using the TGMS system
commercialization. and nuclear male sterility–facilitated recurrent selection to
Biotic stress resistance of hybrids was determined by breed parental lines. Also, molecular marker-aided selection
the resistance of their parental lines and evaluating whether helps in identifying fertility restorers, TGMS lines, and het-
this resistance was dominant or recessive. Hybrid vigor does erotic gene blocks. Transgenic parental lines possessing resis-
not make rice hybrids more or less resistant than the parental tance to bacterial blight and stem borers can be developed by
using Xa21 and Bt genes to develop transgenic rice hybrids George, T, Magbanua R, Laza M, Atlin G, Virmani S. 2002. Magat,
possessing resistance to these stresses. a wetland semidwarf hybrid rice for high-yielding production
Early vegetative vigor and the better developed root sys- on irrigated dryland. Int. Rice Res. Notes 27(1):26-28.
tem of hybrid rice make it adaptable to low-temperature-prone Khush GS, Kumar I, Virmani SS. 1988. Grain quality of hybrid rice.
In: Hybrid rice. Proceedings of the International Symposium
(during the boro season) and salinity-prone irrigated rice ar-
on Hybrid Rice, Changsha, Hunan, China, 6-10 Oct. 1986.
eas. Recent results at IRRI showed better adaptation of certain
p 201-215.
hybrids under alternate wetting-and-drying conditions and in- Ma GH, Yuan LP. 2003. Hybrid rice achievements and development
dicated prospects for increasing water-use efficiency through in China. In: Virmani SS, Mao CX, Hardy B, editors. Hybrid
hybrids. Hybrids have also shown better adaptation to aerobic rice for food security, poverty alleviation, and environmental
conditions (George et al 2002) and certain rainfed lowland protection. Proceedings of the 4th International Symposium
conditions. Hence, the prospects for hybrid rice in these eco- on Hybrid Rice, Hanoi, Vietnam, 14-17 May 2002. Los Baños
systems should be explored. Brown rice of some rice hybrids (Philippines): International Rice Research Institute. p 247-
was found to show a higher iron content in their grain than 256.
popular inbred HYVs and there was evidence of heterosis for Mao CX. 1988. Hybrid rice seed production in China. In: Rice seed
this trait (Gregorio, pers. comm.); hence, commercial rice hy- health. Manila (Philippines): International Rice Research In-
stitute. p 277-282.
brids in and outside China should be critically evaluated for
Peng S, Yang J, Laza RC, Sanico AL, Visperas RM, Son TT. 2003.
iron and zinc content in their brown and polished rice to find
Physiological bases of heterosis and crop management strate-
out whether any of these happen to combine high yield with gies for hybrid rice in the tropics. In: Virmani SS, Mao CX,
high iron and zinc content. Continuous diversification of CMS Hardy B, editors. Hybrid rice for food security, poverty alle-
systems of commercial rice hybrids would be a good strategy viation, and environmental protection. Proceedings of the 4th
to avoid the risk of potential genetic vulnerability associated International Symposium on Hybrid Rice, Hanoi, Vietnam,
with some disease/insect problems. Apomixis is the ultimate 14-17 May 2002. Los Baños (Philippines): International Rice
genetic tool to fix heterosis in rice, which would make it widely Research Institute. p 153-172.
used even by resource-poor farmers. Attempts should there- Takita T. 2003. Hybrid rice research and development in Japan. In:
fore be intensified to develop this tool using modern genetic Virmani SS, Mao CX, Hardy B, editors. Hybrid rice for food
tools. security, poverty alleviation, and environmental protection.
Proceedings of the 4th International Symposium on Hybrid
Heterosis in rice is useful for increasing farmers’ income,
Rice, Hanoi, Vietnam, 14-17 May 2002. Los Baños (Philip-
contributing to national food security, and increasing produc-
pines): International Rice Research Institute. p 337-340.
tion efficiency. The associated labor-intensive hybrid seed pro- Virk PS, Khush GS, Virmani SS. 2003. Breeding strategies for en-
duction helps to create additional rural employment opportu- hancing heterosis in rice. In: Virmani, SS, Mao CX, Hardy B,
nities through the seed industry. editors. Hybrid rice for food security, poverty alleviation, and
environmental protection. Proceedings of the 4th International
Symposium on Hybrid Rice, Hanoi, Vietnam, 14-17 May 2002.
References Los Baños (Philippines): International Rice Research Insti-
Cohen MB, Bernal CC, Virmani SS. 2003. Do rice hybrids have tute. p 21-30.
heterosis for insect resistance? A study with Nilaparvata lugens Virmani SS. 1996. Hybrid rice. Adv. Agron. 57:377-462.
(Hemiptera: Delphacidae) and Marasmia patnalis (Lepi- Virmani SS, Ilyas-Ahmed M. 2001. Environment-sensitive genic
doptera: Pyralidae). J. Econ. Entomol. 96(6):1935-1941. male sterility (EGMS) in crops. Adv. Agron. 72:139-195.
Duvick DN. 1999. Heterosis: feeding people and protecting natural Virmani SS, Kumar I. 2004. Development and use of hybrid rice
resources. In: Coors JG, Pandey S, editors. The genetics and technology to increase rice productivity in the tropics. Int.
exploitation of heterosis in crops. Madison, Wis. (USA): Rice Res. Notes 29(1):10-20.
American Society of Agronomy, Crop Science Society of Virmani SS, Mao CX, Hardy B, editors. 2003. Hybrid rice for food
America, Soil Science Society of America. p 19-29. security, poverty alleviation, and environmental protection.
Proceedings of the 4th International Symposium on Hybrid
Rice, Hanoi, Vietnam, 14-17 May 2002. Los Baños (Philip-
pines): International Rice Research Institute. 407 p.
Session 5: Broadening the gene pool and exploiting heterosis in cultivated rice 165
Virmani SS, Mao CX, Toledo RS, Hossain M, Janaiah A. 2002. Walton M. 2003. Hybrid rice for mechanized agriculture. In: Virmani
Hybrid rice seed production technology and its impact on seed SS, Mao CX, Hardy B, editors. Hybrid rice for food security,
industries and rural employment opportunities in Asia. Tech- poverty alleviation, and environmental protection. Proceed-
nical Bulletin 156. Taiwan (China): Food & Fertilizer Tech- ings of the 4th International Symposium on Hybrid Rice,
nology Center. Hanoi, Vietnam, 14-17 May 2002. Los Baños (Philippines):
Virmani SS, Sharma HL. 1993. Manual for hybrid rice seed produc- International Rice Research Institute. p 97-104.
tion. Manila (Philippines): International Rice Research Insti-
tute.
Notes
Author’s address: Principal scientist (plant breeding), International
Rice Research Institute, DAPO Box 7777, Metro Manila,
Philippines, e-mail: s.virmani@cgiar.org.
In 2002, commercialization of hybrid rice technology became 1A) (Redoña et al 1998). More recent analysis using 77
Philippine agriculture’s banner program for attaining self-suf- microsatellite loci and 46 currently used CMS lines, including
ficiency and increasing productivity and profitability in rice, 10 from the previous study, revealed eight clusters of at least
and generating rural employment. Four public hybrids (Magat, 25% genetic similarity (Fig. 1B).
Mestizo, Mestiso 2, and Mestiso 3) and four proprietary hy- In another study, four maintainers (B) and three TGMS
brids—Magilas (Monsanto), SL8 (SL Agritech), Bigante lines of PhilRice were separated from the B and TGMS lines
(Bayer CropScience), and Rizalina 28 (HyRice)—were made of IRRI at a 58% and 52% level of similarity, respectively,
available to farmers. With intensified training efforts, 40 seed based on 64 microsatellite loci (Fig. 1C,D) (Perez 2004).
growers’ cooperatives were formed that produce 60–70% of Among restorers and CMS lines, genetic similarity coefficients
the program’s seed requirements. From 2001 to 2003, hybrids ranged from 0.22 to 0.71 based on another study using 65
yielded 6.03 t ha–1 versus the 4.44 t ha–1 average for certified microsatellite loci (Ordoñez 2003). These results suggest the
inbred seeds. Average seed production yields increased from existence of distinct nuclear genetic backgrounds in the
<300 kg ha–1 in 2001 to 754 kg ha–1 in 2003. germplasm used and conscious efforts are made to breed hy-
Philippine hybrids are developed using the three-line ap- brids with genetic backgrounds different from those already
proach involving cytoplasmic male sterile (CMS) lines, and commercialized.
the two-line system using temperature-sensitive genetic male
sterile (TGMS) cultivars as female parents. Biotechnology has
Parental genetic diversity and heterosis
been applied to fast-track breeding of improved CMS and
TGMS lines. Biotechnology tools used at PhilRice, the lead Genetic divergence of parents is thought to be related to het-
public institution for rice research and development in the coun- erosis or hybrid vigor. Thus, molecular marker-assayed ge-
try, have been molecular marker technology, genetic engineer- netic variation may potentially be useful for predicting hybrid
ing, and in vitro culture. This paper summarizes research un- performance, thus increasing breeding efficiency.
dertaken using molecular marker technology. Based on 43 microsatellite loci from the 12 rice chro-
mosomes, the nature and extent of correlation between
microsatellite marker heterozygosity and heterosis for some
Genetic characterization of hybrid parental lines quantitative traits were investigated in 48 three-line and 13
To widen the genetic diversity of Philippine hybrid rice two-line rice hybrids. Results revealed heterozygosity and het-
germplasm, thus avoiding genetic uniformity in commercial- erotic performance (measured as the superiority over the male
ized hybrids while increasing the likelihood of developing parent) to be significantly correlated at the 0.01 probability
heterotic combinations, the diversity of PhilRice CMS level for the number of productive tillers per plant (r = 0.41**)
germplasm has been assessed using molecular markers. The and leaf area index (r = 0.39**) while grain yield (r = –0.30*)
nuclear genome diversity of 22 CMS lines from PhilRice, the was correlated with heterosis at the 0.05 probability level in
International Rice Research Institute (IRRI), and China was the three-line hybrids. SSR heterozygosity in the two-line hy-
discriminated into nine groups based on 20 microsatellite loci, brids ranged from 0.43 to 0.66, suggesting a moderate extent
25 RAPDs, and 10 +3/+3 AFLP primer combinations (Fig. of genotypic divergence among the parental cultivars. Heterotic
D
C IR58025B IR68301s
IR78378B 98.1
100 IR78375B (76.5%)
85.1 IR73827-23s IRRI
IR78376B 100.0
IR79123B (96.4%)
IRRI (63.5%)
IR78369B IR73834s
94.9 IR68888B
(59.2%)
IR68897B
TGMS1
IR70369B
IR73328B
(69%)
PR3B TGMS4 PhilRice
97.2 (52.1%) 100.0
IR78367B (90.1%)
PR2B PhilRice
(~58%) TGMS6
PR4B
PR9B
0.52 0.63 0.74 0.85 0.96
0.57 0.68 0.78 0.89 1.00 Coefficient
Coefficient
Fig. 1. Diversity analysis of hybrid rice germplasm. (A) CMS line dendrogram using 20 microsatellite markers, 25 RAPDs, and 10 +3/+3
AFLP primer combinations (Redoña et al 1998). (B) Clustering of 46 diverse CMS lines based on 65 microsatellite loci. (C) Dendrogram of
15 maintainer (B) lines based on 64 microsatellite loci and 255 allele types (Perez 2004). (D) Six TGMS lines clustered using 64 microsatellite
loci and 129 allele types (Perez 2004).
performance of the hybrids was highest for grain yield and is prevalent in the wet season. The disease can cause tremen-
percent spikelet fertility at 62% and 25.6%, respectively. How- dous yield losses if severe infection occurs before flowering.
ever, the relationship between heterosis and genetic diversity To address this potential problem, single and/or pyramids of
was generally weak for most of the traits. Xa4, Xa7, and Xa21 BB resistance genes were introgressed
into five B lines of commercialized and promising hybrids—
IR58025B, IR62829B, 913B, LianB, and BoB (Borines et al
Marker-aided pyramiding of bacterial blight (BB) 2003). B lines with 2–3 gene pyramids showed much shorter
resistance genes lesions in response to diagnostic Xoo races, indicating an in-
Some commercial hybrids in the Philippines, including the creased and wider resistance spectrum against BB than B lines
high-yielding and popular Mestizo, are susceptible to the containing single genes. Xa4 and Xa7 genes together in differ-
Xanthomonas oryzae pv. oryzae (Xoo)-caused BB disease that ent maintainer backgrounds imparted a complementary gene
Session 5: Broadening the gene pool and exploiting heterosis in cultivated rice 167
action in conferring a higher level of resistance to race 3- Conclusions
PXO340, and in LianB background to race 10-PXO341.
In two-line hybrid breeding, three potential TGMS ma- DNA marker applications are finding relevant uses in address-
terials (F2 plants no. 450, 473, and 700) homozygous for Xa4, ing problems and challenges arising from commercializing
Xa7, and Xa21 genes were identified in a segregating popula- hybrid rice in the Philippines. Genetic diversity analyses us-
tion developed from the cross TGMS1(–Xa)/AR32-19-3- ing markers have been useful for ensuring a diverse germplasm
3(+Xa21)/DiR32(+Xa4/+Xa7) (Perez 2004). STS PCR mark- pool for hybrid breeding and avoiding the redundant use of
ers M5F/R and Xa21F/R were used to detect the presence of similar genetic materials. While the utility of randomly selected
Xa7 and Xa21, respectively (Fig. 2A). These lines exhibited markers for predicting heterosis cannot as yet be ascertained,
high degrees of resistance to Xoo race 1 (PXO61), race 2 the usefulness of markers in the tagging and pyramiding of
(PXO86), and race 6 (PXO99) in the greenhouse using an in- useful genes in parental materials, and in seed purity analysis,
duced screening method and they are potential genetic stocks has been demonstrated. As intellectual property protection
for the development of BB-resistant hybrids. becomes emphasized, with the recent passage of the Philip-
pine Plant Variety Protection Law, DNA profiling and identi-
fication of diagnostic alleles for specific hybrid germplasm
Molecular tagging of TGMS genes in rice will find useful applications. An exciting but yet unsuccessful
In hybrid rice breeding, screening for the TGMS trait is an application of molecular markers is in the development of het-
intensive process influenced by environment. Tagging TGMS erotic pools to enhance and sustain higher heterosis levels. As
genes from new donors with DNA markers would increase the the Philippines needs to produce 65% more rice relative to
speed and efficiency of developing two-line hybrids. Based current production by 2025, promising biotechnological tools
on a 301 microsatellite marker survey, four possible chromo- such as molecular marker technology will continue to be used
somal locations (i.e., chromosomes 1, 2, 6, and 12) of the tgms for increasing the speed, efficiency, and accuracy of hybrid
gene from a new donor were identified by single-marker analy- rice breeding procedures while finding other relevant applica-
sis and interval mapping procedures using One-Way ANOVA tions in other aspects of hybrid rice technology commercial-
and Qgene software, respectively. The highest LOD score was ization.
6.48 between RM24 and RM113 on chromosome 1 (Fig. 2B).
Most of the 11 microsatellite markers highly associated with
References
sterility were located on chromosome 6. Three peaks were iden-
tified in this region with LOD scores >3.0, indicating the pres- Borines LM, Redoña ED, Porter B, White F, Natural MP, Vera Cruz
ence of a possible gene/QTL for the tgms gene in chromosome CM, Leung H. 2003. Marker-assisted pyramiding of bacterial
6 (Fig. 2B). Saturating this region with more markers and iden- blight resistance in parental lines of hybrid rice. In: Proceed-
tifying tightly linked markers for use in TGMS line selection ings of the Arnel R. Hallauer International Symposium on Plant
Breeding, Mexico City, Mexico. p 168-169.
would expedite two-line hybrid development, thus further wid-
Ordoñez SA. 2003. Genetic variation of restorer lines in rice (Oryza
ening the genetic base of hybrids for the Philippine commer-
sativa L.). Unpublished MS thesis. University of the Philip-
cialization program. pines Los Baños, Laguna, Philippines. 83 p.
Perez LM. 2004. Development of bacterial blight-resistant
Hybrid seed purity analysis based on mitochondrial thermosensitive genetic male sterile (TGMS) lines for hybrid
genome analysis rice (Oryza sativa L.) production. Unpublished MS thesis.
University of the Philippines Los Baños, Laguna, Philippines.
Genetic purity in hybrid seeds is essential for full heterosis 92 p.
expression and is therefore a prerequisite for a successful com- Redoña ED, Hipolito LR, Ocampo TD, Sebastian LS. 1998. Mo-
mercialization program. Based on analysis of mitochondrial lecular polymorphism of rice cytoplasmic-genetic male ster-
DNA sequences of CMS and fertile rice, the orf61-atp6-orf79 ile (CMS) lines based on AFLP, RAPD and microsatellite
markers. Philipp. J. Crop Sci. 23(1):12-20.
region of the rice mitochondrial genome revealed two regions
polymorphic among IR58025A (CMS), its B line (IR58025B),
and hybrid (Mestizo). The a6P5 and MdF1 primers designed Notes
to amplify these regions produced 300-bp and 500-bp bands,
Authors’ addresses: E.D. Redoña, L.M. Perez, L.R. Hipolito, V.E.
respectively, that discriminated the B line from the CMS line
Elec, I.A. Pacada, R.O. Solis, S.A. Ordoñez, and J. Agarcio,
and the hybrid. Successful detection of B line off-types in mix- Philippine Rice Research Institute (PhilRice), Muñoz, Nueva
tures was demonstrated using the bulked DNA assay. This tech- Ecija, Philippines; L.M. Borines, Leyte State University
nique is starting to be used by the Philippine National Seed (LSU), Baybay, Leyte, Philippines, e-mail:
Quality Control Services for seed purity testing in the hybrid edredona@philrice.gov.ph.
commercialization program.
AR32-19-3-3
IRBB4/7
TGMS1
IRBB7
IRBB4
IR24
473
450
700
M M
1,170 bp
294 bp
1,000 bp
Chr 1 Chr 6
B RM128
RM170
RM259
RM275
RM220
RM237
RM276
RM1
RM5
RM111
RM140
RM113 RM3
RM248
RM24 RM162
RM102
RM50
RM243
6.48 3.0 0.0 3.0 0.0
LOD LOD
Fig. 2. Use of DNA markers in two-line hybrid breeding. (A) Genotyping of potential TGMS
materials (F2 plants no. 450, 473, and 700) showing the 294-bp allele for Xa7 (M5F/R primer)
and 1,000-bp allele for Xa21 (Xa21F/R primer) for bacterial blight resistance in rice. TGMS1 is
a susceptible parent, whereas IRBB4/7 (DiR32) and AR32-19-3-3 are donor parents for Xa4/
Xa7 and Xa21, respectively. IRBB7, IRBB4, and IR24 are check varieties (Perez 2004). (B)
Possible location of tgms gene (LOD = 6.48) on chromosome 1. Three peaks with LOD>3.0
found on chromosome 6 suggest the presence of a QTL.
Session 5: Broadening the gene pool and exploiting heterosis in cultivated rice 169
Genetic evolution of Rf1 locus for the fertility restorer gene
of BT-type CMS rice
Tomohiko Kazama and Kinya Toriyama
Session 5: Broadening the gene pool and exploiting heterosis in cultivated rice 171
A polygenic balance model in yield components as revealed
by QTL analysis in rice
Wilhelm E. Hagiwara, Kazumitsu Onishi, Itsuro Takamure, and Yoshio Sano
Under stabilizing selection, typical or intermediate phenotypes Genotyping with molecular markers
are favored, encouraging the development of hidden variation A total of 17 PCR-based markers were used to determine the
within and between populations (Mather and Jinks 1982). extent of the introgressed chromosome segments in the RILs.
Based on the polygenic balance theory, repulsion phase link- Nine of these markers (RM508, RM589, RM510, RM204,
age could be accumulated along a chromosome as demonstrated RM314, RM253, RM136, RM527, and RM3) were
in Drosophila (Thoday 1960). Such a linked group of poly- microsatellite markers (McCouch et al 2002), and one marker
genes might be reshuffled through recombination, resulting in (C214) was a cleaved amplified polymorphic sequence (CAPS)
the formation of newly derived extreme phenotypes or trans- marker (Rice Genome Research Program, http://
gression. Our study was conducted to gain insight into the ge- rgp.dna.affrc.go.jp). In addition, three PCR-based markers were
netic architecture of quantitative traits in grain characteristics designed from two RFLP (restriction fragment length poly-
between the two rice subspecies, the so-called indica and morphism) markers, R2291 and S1520, and one STS (se-
japonica types. The grain dimension is well differentiated be- quenced tagged site) marker, R2349. In addition, gene-spe-
tween these two rice subspecies in a quantitative manner, af- cific markers were designed from the Wx, Se1, RFT1, and alk
fecting grain quality (Redoña and Mackill 1998). To enhance genes.
the power to detect linked QTLs with small effects, a large
chromosome segment from an indica type was first introgressed Data analysis
by backcrosses into a japonica type of rice. Then, the A linkage map was constructed using Map Manager QTX soft-
introgressed segment was dissected by repeated self-pollina- ware (Manly et al 2001). To detect QTLs, we employed inter-
tions after hybridization with the recurrent parent. The result- val mapping and MQM mapping methods using MapQTL ver-
ing recombinant inbred lines (RILs) are expected to enhance sion 4 (Van Ooijen et al 2002). The empirical LOD thresholds
the efficiency to detect minor QTLs because of the reduction corresponding to the genome-wide significance at the 1% level
in the residual variation. We report here that a cluster of QTLs were estimated by permutation tests, for each trait. The addi-
affecting grain characteristics is present on a chromosome 6 tive genetic effect, the percentage of variance explained by
segment showing repulsion and coupling linkages, and that each QTL (PVE), and the total variance explained by all the
transgressive segregation appears through recombination within QTLs affecting a trait were obtained with MapQTL in the fi-
the chromosome segment. nal multiple-QTL model in which one cofactor marker was
fixed per QTL. Two-way ANOVA was used to detect digenic
interactions between QTLs. The markers with the highest LOD
Materials and methods score found by MQM mapping for each QTL were assumed to
Plant materials be at the possible position of the QTL.
T65wx was a near-isogenic line (NIL) of T65 (a japonica type
of Oryza sativa from Taiwan) carrying wx (waxy) from
Results and discussion
Kinoshitamochi (japonica). T65Wx-pat was a NIL of T65 car-
rying a chromosome 6 segment of Patpaku, an indica type from Segregation for seed dimension in the RILs
Taiwan (Dung et al 1998). An F1 plant between T65Wx-pat The seed and kernel were shorter and wider in T65wx than in
and T65wx was self-pollinated to give the F2 generation, which T65Wx-pat, suggesting that genes for the traits might be lo-
was advanced to the F7 generation using the single-seed de- cated on the introgressed segment of chromosome 6. T65wx
scent (SSD) method. The plants were expected to carry differ- showed higher values in the weights of seeds and kernels (SW
ent lengths of introgressed segments from Patpaku in the T65wx and KW) than did T65Wx-pat. Each of the six traits showed a
background and they were regarded as near-isogenic RILs. In continuous variation among the RILs.
total, 163 RILs were established and used in our experiment.
Grain characteristics were measured in seeds (unhulled grains) Detection of putative QTLs for grain characteristics
and kernels (hulled grains) of the RILs (F7 generation) and the The linkage map was constructed based on the segregation of
parents. For each of the 163 RILs, two plants were examined. 17 markers in all the RILs, and had a length of 72.5 cM. In
The average length and breadth (expressed in mm) from 6 seeds total, 10 QTLs responsible for the six traits were detected
per plant and the weight (in grams) of 10 grains per plant were (Table 1). The percentage of the phenotypic variance explained
calculated. (PVE) ranged from 6.8% for a QTL controlling seed breadth
(SB) to 43.7% for a QTL controlling kernel length (KL).
aNML indicates the marker locus nearest to the QTL or the marker with the highest LOD score peak.
PVE is the percentage of the phenotypic variance explained. a is the additive effect of the Patpaku
allele and the positive value indicates that the effect of the genotype of the indica parent (Patpaku) is
in the direction of increasing the measured value of the trait.
Three QTLs were detected for seed length (SL) and SB, Hidden genetic variation
at approximately the same locations, near the RFT1, S1520, Actually, the RILs with the highest or lowest values in SL and
and Hd1 markers (Table 1). Regarding the Patpaku-derived SB resulted from an accumulation of QTLs with positive or
QTLs near RFT1, S1520, and Hd1, it should be noted that the negative effects (Table 2). Thus, the transgressive segregation
QTL near RFT1 decreased both SL and SB, while that near revealed that a hidden genetic variation was concealed in the
S1520 increased both of them and the QTL near Hd1 increased introgressed segment even though the phenotypic difference
SL but decreased SB (Table 1). The detection of the effect of in grain characteristics was slight between the parental lines.
each QTL on SL depended on the genotypes of the other two Linked QTLs with different signs in effects are suggested to
QTLs (Table 2); however, it could be due to a small number of occur alternately along a chromosome under the polygenic
some genotypes because two-way ANOVA failed to detect any balance model (Mather 1943). Such linked QTLs on a chro-
significant interactions between QTLs for SL. For SB, two- mosome behave as an effective factor in segregating popula-
way ANOVA detected distinct epistatic effects (interactions) tions, although the polygenic balance model is an extreme
between the RFT1 and Hd1 markers (F = 14.64, P = 0.0002) metaphor (Lynch and Walsh 1998). This leads us to consider
as well as between the Hd1 and S1520 markers (F = 7.92, P = that adaptive traits might be created in part through the recon-
0.0055), suggesting that the effect of each QTL also depended struction of gene combinations in the flanking region using
highly on the genotypes of the other linked QTLs for SB (Table preexisting variation rather than newly derived mutations. Some
2). For seed weight, only one QTL was detected near the RM3 agriculturally valuable traits might also have been reconstructed
marker. in this way, as predicted from frequent transgressions.
In contrast to the characteristics of seeds, the length and
breadth of kernels were each controlled by a single QTL with
References
high LOD scores and PVE (Table 1). For KL, one QTL was
detected near the Hd1 marker, and that for KB was detected Dung LV, Inukai T, Sano Y. 1998. Dissection of a major QTL for
near RM3, showing their slightly different position (Table 1). photoperiod sensitivity in rice: its association with a gene
The Patpaku allele increased KL, whereas the T65wx allele expressed in an age-dependent manner. Theor. Appl. Genet.
increased KB. For KW, a single QTL was also detected near 97:714-720.
Lynch M, Walsh B. 1998. Genetics and analysis of quantitative traits.
RM3.
Sinauer Associates, Inc., Sunderland.
The detection of multiple QTLs of small effects in our
Manly KF, Cudmore Jr RH, Meer JM. 2001. Map Manager QTX,
experiments is possibly related to the elimination of effects of cross-platform software for genetic mapping. Mammalian
segregating genes on other chromosomes. The variance caused Genome 12:930-932.
by segregation on other chromosome regions increases the re- Mather K, Jinks JL. 1982. Biometrical genetics. 3rd ed. New York,
sidual variance for the interval under consideration, reducing NY (USA): Chapman and Hall.
the power for QTL detection (Lynch and Walsh 1998). Conse- Mather K. 1943. Polygenic inheritance and natural selection. Biol.
quently, multiple segregating QTLs tend to reduce the effi- Rev. 18:32-65.
ciency to detect significant associations between phenotype
and genotype in the whole-genome QTL mapping.
Session 5: Broadening the gene pool and exploiting heterosis in cultivated rice 173
Table 2. Phenotypic effects of each marker allele depending on the genotypes of the
other two markers (background) for seed length (SL) and breadth (SB).a
a(T)and (P) refer to the T65wx- and Patpaku-derived alleles, respectively. SL (seed length) and SB (seed
breadth) were measured in mm. The statistical difference was based on the t-test. ns = not significant. *
and ** show significance at 5% and 1% probability, respectively.
McCouch SR, Teytelman L, Xu Y, Lobos KB, Clare K, Walton M, Van Ooijen JW, Boer MP, Jansen RC, Maliepaard C. 2002.
Fu BY, Maghirang R, Li ZK, Xing YZ, Zhang QF, Kono I, MapQTL® 4.0, software for the calculation of QTL position
Yano M, Fjellstrom R, DeClerck G, Schneider D, Cartinhour on genetic maps. Wageningen (Netherlands): Plant Research
S, Ware D, Stein L. 2002. Development and mapping of 2240 International.
new SSR markers for rice (Oryza sativa L.). DNA Res. 9:199-
207.
Redoña ED, Mackill DJ. 1998. Quantitative trait locus analysis for Notes
rice panicle and grain characteristics. Theor. Appl. Genet. Authors’ address: Plant Breeding Laboratory, Graduate School of
96:957-963. Agriculture, Hokkaido University, Sapporo 060-8589, Japan,
Thoday JM. 1960. Effects of disruptive selection. III. Coupling and e-mail: rysano@abs.agr.hokudai.ac.jp.
repulsion. Heredity 14:35-49.
Wrap-up of Session 5
Three conveners, Y. Fukuta (Japan International Institute for Agri- In the mid-1980s, upland rice varieties of a tropical japonica
cultural Sciences, JIRCAS), D. Mackill (International Rice Research background were grown in the Brazilian savannas. Afterward, crop
Institute, IRRI), and R. Ikeda (JIRCAS), organized Session 5, area gradually and then markedly decreased. However, produc-
chaired by B. Mishra, Directorate of Rice Research, Indian Coun- tion did not decline at the same level because average yield
cil of Agricultural Research. Three scientists presented papers on doubled from 1986 to 2003. This was due to both crop migra-
broadening the gene pool and two on exploiting heterosis in cul- tion toward areas with more favorable rainfall distribution and
tivated rice. the adoption of modern plant-type varieties, tropical japonica ×
B.S. Pinheiro, of Embrapa Arroz e Feijão, discussed “Aero- indica derivatives. The impact of improved plant-type varieties
bic rice development in Brazil.” She summarized upland breeding and associated technologies in favored localities of the savanna
research under Brazilian savannas from the mid-1970s to 2003. region gave rise to a new concept of upland rice, leading to a
Session 5: Broadening the gene pool and exploiting heterosis in cultivated rice 175
fied inbred seeds, for a 34% advantage. Among the key factors In summary, the potential of wide hybridization (between
related to the progress made are strong support at the policy indica and japonica, or between cultivated and wild rice), and
level, engagement of the private sector, increased capacity en- mutants and germplasm for different endosperm traits, was dis-
hancement activities, front-line demonstrations, development of cussed for broadening of the gene pool and application to rice
production support, rewards and marketing assistance systems, breeding. In particular, the exploitation of heterosis through hy-
involvement of local government units, and a massive informa- brid rice production was considered a good example of broaden-
tion campaign. A successful communication and extension sys- ing the gene pool to achieve higher yield.
tem for hybrid rice breeding was also demonstrated.
Economic factors and recent changes in rice production tech- area in Asia (Table 1). This includes upland and submergence-
nology are the major drivers that have led to shifts from trans- prone ecosystems where direct seeding is the traditional
planting to the direct-seeding method of rice establishment in method. If these ecosystems are excluded, the direct-seeded
Asia. The rising cost of labor, the need to intensify rice pro- area is approximately 15 million ha. The shift toward direct
duction through double and triple cropping, the development seeding occurred during the late 1980s to mid-1990s, mainly
of high-yielding short-duration modern varieties, and the avail- in rapidly growing economies such as Malaysia and Thailand
ability of chemical weed control methods have jointly led to as well as in countries where rapid intensification of the rice
this switchover. As the rice production systems of Asia un- production system took place (for example, Vietnam). How-
dergo further adjustments in response to the rising scarcity of ever, the growth rate in direct-seeded rice area appears to have
land, water, and labor, pressure will increase for a shift toward slowed down considerably in recent years.
direct-seeding methods. This paper provides a brief overview
of the trends in crop establishment methods in Asia, their im-
Determinants of adoption of alternative
pact, and implications for research and technology develop-
crop establishment methods
ment.
The direct-seeded area in Asia is estimated to be about At a general level, the availability of water and the opportu-
29 million ha, which is approximately 21% of the total rice nity cost of labor can be considered the major determinants of
aHuke and Huke (1997). Sources of direct-seeded area: Bangladesh—Huke and Huke (1997) and personal communication with S. Bhuiyan. India—
Palaniappan and Purushothaman (1991). Sri Lanka—Pathinayake et al (1991). Cambodia—Helmers (1997). China—personal communication with Lu
Ping. Indonesia—Huke and Huke (1997) and personal communication with Dr. Hamdane Pane. Malaysia—Huke and Huke (1997) and own estimate.
Myanmar—Huke and Huke (1997) and own estimate. Philippines—PhilRice-BAS (1995) and own estimate. Thailand—personal communication with
Dr. Booribon Somrith and data from Agricultural Extension Office, Khon Kaen. Japan—personal communication with Yujiro Hayami. Vietnam—personal
communication with T.P. Tuong and Government Statistical Office (1997). Korea—Kim (1995).
($ ha–1)
Suphanburi, Thailanda
Cash costb 152 148 3
Gross returnsc 505 476 6
Gross margind 353 328 8
Net returnse 168 132 27
Pangasinan, Philippinesf
Cash cost 230 273 –16
Gross returns 608 666 –9
Gross margin 378 393 –4
Net returns 288 247 17
aSource: Isvilanonda (1990). All values converted to $ using the exchange rate US$1 = B20. bCost of
all purchased inputs. cGross value of output. dGross value of output minus the cost of purchased inputs.
eGross value of output minus the cost of purchased and family-owned inputs. fPandey et al (1995).
gPandey and Velasco (1998). The comparison is between two crops of wet-seeded rice versus only one
and crop management practices to relax these constraints is tween the “best-practice” farmer and the average farmer when
needed. rice is direct-seeded. A greater variability of technical efficiency
However, if crop intensification is the major reason for of direct-seeded rice could be partly due to the use of varieties
direct seeding, research to facilitate early establishment and that were originally developed for transplanted culture. Vari-
early harvest of the direct-seeded crop would have a higher eties that are specifically targeted for direct-seeded methods
priority, as this will permit timely planting of the subsequent could help reduce such yield gaps. Better crop management
crop. The development of short-duration varieties would be practices, especially those that facilitate early and more uni-
important in this case. Even though the cost of labor may be form establishment, can be similarly helpful.
initially low in these areas, intensification of land use may lead Precise water management is critical for high productiv-
to labor shortages because of peak labor demand during the ity of wet-seeded rice (De Datta and Nantasomsaran 1991). A
harvesting of the previous crop and establishment of the suc- high level of control of water flow on irrigated fields is hence
ceeding crop within a short period. Suitable mechanical de- desirable. However, most irrigation systems in Asia have been
vices for land preparation that can reduce the turnaround time designed to supply water to transplanted rice for which preci-
between crops could help achieve a higher and more stable sion in water management is not as critical. Suitable modifica-
yield of the second crop. tions of irrigation infrastructure could ensure a high yield of
The high costs of weed control could be a major factor direct-seeded rice and also improve water-use efficiency. In
constraining the widespread adoption of direct seeding. The addition, appropriate mechanical systems of field leveling that
key to the success of direct-seeded rice is the availability of ensure uniformity in field water level are needed.
efficient weed control techniques. Varieties with early seed-
ling vigor and crop management technologies that help reduce
References
the competitive effects of weeds on crops are needed. It is
essential, however, to evaluate the environmental and health De Datta SK, Nantasomsaran P. 1991. Status and prospects of direct
consequences of potential technologies that are based on chemi- seeded flooded rice in tropical Asia. In: Direct seeded flooded
cal means of weed control. rice in the tropics. Selected papers from the International Rice
Empirical analyses have indicated that the technical ef- Research Conference. Manila (Philippines): International Rice
Research Institute. p 1-16.
ficiency of rice production is lower and more variable for di-
Helmers K. 1997. Rice in the Cambodian economy: past and present.
rect-seeded rice than for transplanted rice (Pandey and Velasco
In: Nesbitt HJ, editor. Rice production in Cambodia. Manila
1999). This suggests the existence of a higher “yield gap” be- (Philippines): International Rice Research Institute. p 1-14.
In 1991, the Rural Development Administration (RDA) intro- In dry-seeded rice, on the other hand, annual grasses were the
duced dry drill seeding and water seeding of rice with water most predominant weeds, such as E. crus-galli (54%), Digitaria
seeding being either wet-drill-sown or water broadcast-seeded. adscendens (9%), Leptochloa chinensis (8%), Setaria viridis
Despite problems, the direct-seeding area has increased rap- (6%), and weedy rice (Oryza sativa subsp. spontanea, 5%). In
idly; however, only about 8–9% of the total rice area was un- water seeding, E. crus-galli (37%) was also the most impor-
der direct seeding in 2003 (RDA 2003). The main constraints tant weed species even though the degree of dominance was
to direct seeding are weed problems and rainfall fluctuation lower than in dry-seeded rice. The second most important weed
during seeding time. was Aneilema keisak (15%), followed by Ludwigia prostrata
(11%) and Leersia japonica (7%). Weed growth in direct-
seeded rice was dominated by E. crus-galli, which became the
Weed ecology in direct-seeded paddy most important weed and had a slightly lower value in water-
Weed growth is greatly influenced by cultivation method (Kim seeded rice than in dry-seeded rice. Most of the annual grass
et al 1992) and thus the adoption of direct-seeding technology weeds occurring in direct-seeded rice fields belonged to the
results in shifts in weed growth in terms of dry matter produc- C4 photosynthetic pathway (Table 2).
tion and composition of dominant species. The greatest weed Recently, weedy rice (including red rice) has become
growth is recorded in dry-seeded rice, followed by water- widely distributed in farmers’ fields. The possible origins of
seeded rice, mechanically transplanted rice, and manually trans- weedy rice were summarized by Kim (1995b) as follows:
planted rice (Table 1). O Shattered grains of the previous year
Early rice growth and water management are considered O Outcross between cultivated rice species
as the main factors accounting for the above differences. Yield O Outcross between cultivated rice and red rice (dor-
losses from weed competition are related closely to weed mant seeds)
growth, with complete crop loss occurring where weeds are O Outcross between wild-type rice and weedy rice (red
not managed. Changes in crop establishment method from rice)
transplanting to direct seeding changed not only weed dry O Outcross between red rice species
matter production but also the composition of the weed spe- O Dormant red rice or wild-type rice itself
cies. In transplanted rice fields, the most troublesome weeds The degree of contamination by weedy rice observed in
were Eleocharis kuroguwai (20%), Sagittaria trifolia (16%), fields was highly variable, ranging from 0.5% to 35.2% for
S. pygmaea (13%), Echinochloa crus-galli (12%), and the southern area. However, it was difficult to estimate the
Monochoria vaginalis (11%) (Kim et al 1992). Approximately correct degree of contamination and distribution pattern for
50% of weed biomass was accounted for by perennial weeds. all farmers’ fields because field observations were made and
The aging of farmers and falling rice prices are creating se- motes the seedling emergence of direct-seeded rice in flooded
vere problems for farmers in Japan, and this has led to demand soil. Experiments on the practical use of CaO2 in the 1970s
for a reduction in the costs and labor required for rice (Oryza contributed to the development of an oxygen-generating prod-
sativa L.) production in recent years. Direct seeding, which uct containing CaO2 and its commercial release (Calper Fine
does not require seedlings to be raised or transplanted, is re- Granule). Thereafter, pregerminated seeds coated with the
garded as the most effective method of reducing costs and la- oxygen-generating product (simply called “coated seeds” here-
bor. However, direct-seeded fields currently represent less than after) could be sown in soil at depths of 5–20 mm in submerged
1% of the total rice area, although dry-seeded fields increased conditions.
during the 1970s before the use of transplanting machines be-
came widespread. One disadvantage of direct seeding, hinder- Water management after seeding
ing its widespread use, is the instability of rice yields of di- Until the mid-1990s, fields were flooded after direct seeding
rect-seeded fields compared with those of transplanted fields. because it was believed that flooding was necessary to main-
The main factors responsible for the lower yields in wet-seeded tain a more constant and higher soil temperature during seed-
fields are poor seedling establishment and frequent plant lodg- ling establishment. At that time, to control weeds, herbicide
ing. Recently, improved methods that stabilize seedling estab- was usually applied under flooded conditions. However, seed-
lishment and promote lodging resistance have been developed ling establishment was unstable with this type of water man-
and the area of wet-seeded rice cultivation has doubled within agement, even when coated seeds were used. Oba (1997) sug-
the past 5 years. This paper discusses recent improvements gested that drainage after seeding enhanced seedling estab-
that stabilize seedling establishment and lodging resistance in lishment compared with flooding for rice coated and direct-
wet-seeded rice. seeded in the soil. Several reports (Yoshinaga et al 2000, Sato
and Maruyama 2002, Tsuchiya et al 2004) on the effects of
drainage after seeding show that oxidized soil conditions sur-
Stabilization of seedling establishment rounding the seeds improved seedling establishment and stimu-
In Japan, most popular rice cultivars with high taste quality do lated initial growth (Table 1). It has also been shown that the
not show sufficient lodging resistance under wet-seeded con- time taken for seedling emergence in drained conditions is simi-
ditions, as they are bred for transplanting. Seeding such culti- lar to that in flooded conditions, although the minimum soil
vars on the soil surface often causes severe lodging during the temperature during emergence may be lower in drained condi-
ripening period. This disadvantage can be minimized by seed- tions. Furthermore, improved herbicides that can control weeds
ing in the soil, but lower and slower rates of seedling emer- at a later growth stage were developed and their use has sup-
gence result from the anaerobic conditions around the seed. ported the widespread drainage after seeding because improved
Furthermore, direct seeding in Japan is usually conducted in herbicides can control weeds even if they are applied after
periods of low temperature of less than 20 °C daily mean tem- rice seedling emergence subsequent to drainage. For these rea-
perature, which may result in lower and slower rates of seed- sons, drainage subsequent to the seeding of coated seeds had
ling emergence. Solutions to these problems are important to become widespread by the end of the 1990s and it has contrib-
further develop wet-seeding cultivation and recent research uted to the recent increases in the area cultivated by wet seed-
has focused on stabilizing seedling establishment in the soil. ing.
Further experiments are required to decrease the dose of
Development of oxygen-generating products CaO2 or replace the use of CaO2 and reduce costs by develop-
Ota and Nakayama (1970) reported that coating rice seeds with ing cultivars with high emergence ability at low temperature.
calcium peroxide (CaO2), acting as a source of oxygen, pro-
aSeeds were coated with calcium peroxide and seeded in the soil at 10-mm depth. Data are
means of 2–3 years. Soil temperature is the average of 2 weeks after seeding at 10-mm depth.
Dry weight was measured at 2–3 weeks after seeding. Values followed by the same letter are
not significantly different at the 5% level according to LSD. (Yoshinaga et al 2000.)
Improvement of lodging resistance per hill, which is composed of several plants (Yoshinaga et al
2001).
Lodging is problematic for rice production as it makes har- Hill seeding has enabled popular cultivars with low lodg-
vesting by machine difficult. Furthermore, the lodging of rice ing resistance to be grown in wet-seeding areas. The area cul-
plants during the ripening period results in a yield reduction tivated by the hill-seeder has been increasing since 1998 and
owing to decreased canopy photosynthesis from self-shading overtook the area cultivated by broadcasting in 2002. Hill seed-
and also decreases grain quality because of the increased col- ing occupied 24% of the total area of wet-seeded rice in 2003
oring of brown rice and/or decreased taste (Matsue et al 1991, and it can be said that the hill-seeder has contributed to the
Setter et al 1997). Wet-seeded rice is more susceptible to lodg- recent widespread application of wet seeding.
ing during the ripening period than is transplanted rice. Popu- To further improve lodging resistance, the introduction
lar cultivars are used in direct seeding to achieve better pro- of cultivars with improved lodging resistance is important to-
ducer prices, although they do not generally show sufficient gether with studies on the interaction of nitrogen application
lodging resistance. Therefore, alternative means of improving with seeding methods.
the lodging resistance of direct-seeded rice are required such
as amending nitrogen application, seeding depth, and seeding
method. A decrease in nitrogen fertilization limits culm elon- Conclusions
gation and improves lodging resistance. The smaller amount The recent improvements in the stability of establishment un-
of nitrogen application, however, might reduce yield in direct- der direct seeding have contributed to the recent increase in
seeded rice. Deeper seeding depths have been tested in con- direct-seeded area. For the further stabilization and widespread
junction with the use of coated seeds, and this has reduced the adoption of direct seeding, developments in plant breeding are
root lodging of direct-seeded rice. The base of the rice hills, required to allow the introduction of cultivars with desirable
however, is still shallow in these conditions and lodging resis- traits for direct-seeding cultivation. Improved crop manage-
tance is insufficient. When comparing different seeding meth- ment that corresponds to seeding methods and cultivars is also
ods, lodging resistance is considered to be the highest in hill- required.
seeded rice and higher in row-seeded rice than in broadcast-
seeded rice (Ogata and Matsue 1998). However, practical hill-
seeders did not exist until the late 1990s, and previously direct References
seeding had been conducted by broadcast or row seeding. Matsue Y, Mizuta K, Furuno K, Yoshida T. 1991. Studies on palat-
A practical hill-seeder was developed in the late 1990s. ability of rice grown in northern Kyushu. Jpn. J. Crop Sci.
The hill-seeder, combined with a harrow, enabled seeding and 60:490-496.
puddling to be performed simultaneously (Shimotsubo and Oba S. 1997. Seedling emergence improving method by draining
Togashi 1996). The seeder effectively drives seeds into the (just after seeding) in wet seeding cultivation of rice. Nogyo
Gijutsu 52:33-34.
puddled soil intermittently by using a rotating disk, and the
Ogata T, Matsue Y. 1998. Studies on direct sowing culture of rice in
established seedlings form hills. The greater seeding depth
northern Kyushu. Jpn. J. Crop Sci. 67:485-491.
effectively prevents the floating and lodging of seedlings, while Ota Y, Nakayama M. 1970. Effects of seed coating with calcium
hill seeding effectively increases lodging resistance. Hill- peroxide on germination under submerged conditions in rice
seeded rice has a greater resistance to lodging than broadcast- plant. Proc. Crop Soc. Jpn. 39:535-536.
seeded rice across a range of plant densities (Fig. 1) or seed- Sato T, Maruyama S. 2002. Seedling emergence and establishment
ing depths (Yoshinaga et al 2001). The higher lodging resis- under drained conditions in rice direct-sown into puddled and
tance of hill-seeded rice is due to its larger number of panicles leveled soil. Plant Prod. Sci. 5:71-76.
80
60
40
1998 BS
20
1999 BS
1998 HS
1999 HS
0
0 50 100 150 200
Plant density (no. m2)
Fig. 1. Relationship between plant density and pushing resistance in
broadcast-seeded rice (BS) and hill-seeded rice (HS). Pushing resis-
tance was measured at 20 days after heading by bending a hill 45o
from the vertical at 15-cm height. Vertical bars indicate standard
error (n = 3). (Yoshinaga et al 2001.)
Setter TL, Laureles EV, Marazedo AM. 1997. Lodging reduces yield Yoshinaga S, Wakimoto K, Tasaka K, Matsushima K, Togashi T,
of rice by self-shading and reductions in canopy photosyn- Shimotsubo K. 2001. Improvement of lodging resistance in
thesis. Field Crops Res. 49:95-106. submerged direct seeding rice cultivation using a newly de-
Shimotsubo K, Togashi T. 1996. Study on submerged direct seeding veloped ’Shooting hill-seeder’. Jpn. J. Crop Sci. 70:186-193.
combined puddling with seeding operation. Jpn. J. Crop Sci.
65(Extra issue 1):12-13.
Tsuchiya M, Sato T, Maruyama S. 2004. Growth enhancement by Notes
drainage during seedling establishment in rice direct-sown into Author’s address: National Agricultural Research Center for Tohoku
puddled and leveled soil. Plant Prod. Sci. 7:324-328. Region, NARO, Japan, e-mail: yosinaga@affrc.go.jp.
Yoshinaga S, Nishida M, Wakimoto K, Tasaka K, Matsushima K,
Togashi T, Shimotsubo K. 2000. Effects of drainage after sub-
merged direct-seeding on the effectiveness of fertilized nitro-
gen and the growth and yield of rice plants. Jpn. J. Crop Sci.
69:481-486.
The food supply in China must increase by 50–60% over the of the country (Li and Duan 2002). Since rice is the most wa-
next 30 years for a population that is estimated to reach 1.6 ter-consuming crop, alternative rice cultivation strategies that
billion by 2030. Rice is vital to more than half of the popula- require less water and increase water productivity are urgent.
tion and it is the most important field crop in China. However,
lowland rice cultivation area dropped from 33.76 million ha
Variety improvement and water saving of aerobic rice
in 1981 to 29.96 million ha in 2000 mainly because of water
shortages. Current projections suggest that by 2030 there will Many traditional upland rice types have been grown in hilly
be a shortfall of 12.9 billion m3 of water to meet the demand rainfed regions in China for hundreds of years with very low
The Philippines is situated in Southeast Asia, with tropical cli- nium B.C. (www.riceweb.org/History.htm#Origin), later called
mate and wet and dry seasons. Rice is the staple food and has “tubigan” or transplanting (Camus 1921). When irrigation sys-
been cultivated since time immemorial. Before the introduc- tems were introduced in 1840, only 30% of the total area
tion of wetland rice, direct seeding in the “caiñgin” or slash planted to rice was irrigated until 1913 and only a few farms
and burn on mountain slopes (Fig. 1A) and secano or “dry had sufficient irrigation water for two rice crops a year (Alfonso
planting” on undulating lands to flatlands (Fig. 1B) were prac- and Catambay 1948).
ticed. Transplanting became common when irrigation pumps
and canals were constructed in the 1950s (Sta Iglesia and Lawas
1959) and the National Irrigation Administration was created
Shifting cultivation in 1963. Almost all flatlands (approx. 3.126 million ha) were
Migrants from mainland Asia carried the traditions of lowland planted to rice during the wet season but the majority are still
rice cultivation to the Philippines during the second millen- rainfed. When irrigated area increased to 1,387,855 ha (NIA
Fig. 1. Different methods of rice planting in the Philippines (A-D) and adoption of
direct seeding in average percentage of the regional area (E). A: caiñgin on moun-
tain slopes, B: secano or dry seeding on flatland, C: lipat-tanim or transplanted
rice, D: dalatan or sabog-tanim in wet paddies.
2004), there were more dry-season rice crops in areas with good, lack of expertise on direct-seeding technology, and the
enough water. The introduction of short-duration area is prone to golden apple snail (Pomacea spp.) attack.
nonphotosensitive rice varieties and effective fertilizer man-
agement by IRRI led to improved yields. Land preparation for DS
The shift from transplanting (Fig. 1C) to wet direct seed- A well-leveled field with less depression is essential for good
ing (WDS) or “sabog-tanim” (Fig. 1D) with improved man- water management and weed control. Emerging seedlings can
agement practices for weeds, pests, water, and fertilizer could be killed or greatly retarded in the presence of pond water on
be due to high production costs (De Datta 1986, De Datta and low portions because of a lack of oxygen and accumulation of
Flinn 1986), labor scarcity for transplanting during the peak toxic concentrations of applied herbicides or attacks by golden
season, the opportunity for early establishment for more crop- apple snails. Weeds can grow abundantly in the high portions.
ping and maximizing available soil moisture, avoidance of bad These conditions will result in uneven tiller growth and matu-
weather conditions during harvesting, and conserving more rity, and yield loss.
water. In Central Luzon, Philippines, for example, the adop- Dryland preparation. This requires a bigger tractor to
tion rate of WDS was from <2% of the rice area in 1979 to plow and harrow the dry to moist soil. It is done before the
approximately 16% in 1982 (Moody and Cordova 1985). The onset of the rainy season on farms with lighter soil texture. It
same trend was reported in 1989-90 from the same area (Khan requires less water during land preparation and the crop can
et al 1992). Adoption has been on the rise, as reflected in a hence establish early. Dry seeds can be sown before the onset
2002 regional survey (Fig. 1E). of the wet season and rain can then germinate the seeds. A
Farmers prefer the dry season for WDS because of poor drawback is the availability of good leveling equipment and
crop establishment and lodging problems in the wet season. effective weed control.
Field inundation within 1–2 wk after sowing (WAS) usually Wetland preparation. This is common on farms with
results in poor survival of seedlings, leading to nonuniform ample water and in any type of soil. Locally manufactured hand
growth. The average relative distribution of farmers from 1998 tractors can be used to puddle the soil. It has a greater advan-
to 2002 that reportedly practiced DS was 68% during the Janu- tage than dry seeding because land leveling is much easier.
ary to June cropping (dry season), and 32% during the July to Semidryland preparation. This is done by plowing and
December cropping (wet season) (Table 1). harrowing the field immediately after harvest while the soil is
Wet direct seeding is the most common type of direct still moist. The field is kept flooded during the rest of land
seeding in the Philippines. It is practical, has better weed con- preparation, usually for 3–4 wk, for the dry-season crop. For
trol than dry seeding, and has a shorter time for the seedlings the wet-season crop, after the initial rotavation, the field is
to emerge. It is commonly practiced in lowland rice areas with kept as is until the onset of the wet season, when the final land
good irrigation and drainage. Further, water control is very preparation proceeds. These methods can help conserve soil
important, especially during the crop establishment period up moisture and control weeds and golden apple snails.
to a few days from seeding (Khan et al 1992). Generally, yield
from direct seeding and transplanting is similar (Fig. 2A,B). Planting
Dry direct seeding is still being done in areas with higher Problems of poor seedling emergence emanate from uncon-
elevation and slope, on nonirrigated flatlands, and on irrigated trolled water and golden apple snail, bird, and rodent attacks
lowlands with medium- to light-textured soils. Seeds are grown during the first 2 wk after seeding. Aggravating circumstances
just before the onset of the wet season to establish the crop are uneven land preparation and poor drainage.
earlier. Seed preparation. Commonly used seed preparation in-
There are also common reasons for not adopting DS: cludes dry seeds, primed seeds, and pregerminated seeds. Prim-
high weed pressure, especially when land preparation is not ing the seeds means soaking them for 12–24 h before sowing,
3 3
2 2
1 1
0 0
1 2 3 4 5 6 7 8 9 10 11 12 13 1 2 3 4 5 6 7 8 9 10 11 12 13
ARMM
ARMM
Region Region
% %
Dry season Wet season
80 60
C D
60
40
40
20
20
0 0
Broadleaves Sedges Grasses Broadleaves Sedges Grasses
Weed type Weed type
mm %
60
800 Water used (mm) E
Water saved (%) 50
600 40
30
400
20
200
10
0 0
Continuously Irrigation when Shallow irrigation Irrigation every
flooded no floodwater (02 cm) 14 d
Water management
Fig. 2. Regional average yields of direct-seeded and transplanted rice in the Philippines (A and B), weed-type distribution for wet- and
dry-season wet direct-seeded rice (C and D), and water used and saved by using controlled irrigation (E).
whereas pregerminating means soaking the seeds for 12–24 h, Seeding rate. Most farmers used seeding rates of up to
followed by incubating for 24–35 h. Dry and primed seeds are 200 kg of seed ha-1 although the recommended rate is 60–80
usually sown in dry to moist soil, whereas pregerminated seeds kg ha–1 (Table 2). The wide variation among regions could be
are usually sown in saturated soil. attributed to their culture, skills, technical knowledge, and the
Sowing methods. Broadcast seeding is the most com- presence of adverse conditions for DS such as uncontrolled
mon method because of simplicity. Drilling or distributing seeds water and the presence of pests that affect crop establishment.
evenly in rows is done manually or by a seeder. Dibbling or Higher seeding rates would be beneficial if no weed control is
localized planting of seeds is done in small holes created by a planned (Guyer and Quadranti 1985), but it is not necessary to
pointed object thrust into the soil. This is common in caiñgin use high seeding rates to suppress weeds if an effective herbi-
(minimal or zero-tillage). cide is used (Castin and Moody 1989).
The High Barind Tract of northwest Bangladesh is drought- TPR. However, weeds are a major constraint to the adoption
prone, with the majority of the 1,200–1,400-mm mean annual of DSR as the inherent advantage of weed control afforded by
rainfall occurring in June to October. Limited irrigation po- transplanted rice in standing water is lost (Mazid et al 2002).
tential restricts cropping intensity to 175%, considerably less Labor shortages for many households prevent timely first weed-
than in districts where irrigation allows two or three rice crops ing of transplanted rice so that with current practices 34% of
each year (Nur-E-Elahi et al 1999). The majority of farmers farmers lose over 0.5 t ha–1 of the attainable rice yield because
produce a single crop of transplanted rainfed rice, grown in of weed competition (Mazid et al 2001). The additional weed
this monsoon season. Some 80% of the area then lies fallow in problems in DSR may be overcome, however, by applying a
the postrice rabi season. The challenge in the Barind is to si- preemergence herbicide (Mazid et al 2001). We report on the
multaneously improve the reliability and yield of rice while productivity of direct seeding of an early-maturing rice culti-
increasing total system productivity by increasing the area var with herbicide application, followed by chickpea in the
planted to postrice rabi crops, including chickpea, linseed, and postrice season.
mustard (Mazid et al 2003). Mazid et al (2002, 2003) have
proposed that the productivity of Barind soils can be increased
by switching from transplanted rice (TPR) to direct-seeded
Methods
rice (DSR) to allow more reliable establishment of rabi crops Systems trial
on residual moisture immediately after the rice harvest. The productivity of two rice cultivars when direct-seeded or
Chickpea, a drought-tolerant and high-value crop, can be grown transplanted was evaluated in the Barind from 2001 to 2003
successfully when seeded after rice in late October to mid- under differing nutrient regimes. Modern cultivar BRRI dhan
November. This can make significant contributions to higher 39 (BR39) (maturity 120–125 days) was compared with the
productivity and improved farm income. A late onset of the widely grown Swarna (maturity 145–150 days). The experi-
monsoon delays transplanting as a minimum of 600 mm of ment was conducted with a split-split plot design with main
cumulative rainfall is needed to complete ploughing, puddling, plots (3) as crop establishment and associated weed manage-
and transplanting. Direct seeding can be completed after ment, subplots (4) as nutrient management, and sub-subplots
ploughing, however, following only 150 mm of cumulative (2) as cultivars. Establishment treatments were (1) transplanted
rainfall (Saleh and Bhuiyan 1995). Earlier planted DSR ma- rice (TPR)—soil puddled prior to transplanting and plots hand-
tures 1–2 weeks before transplanted rice, thus reducing the weeded twice at 30 and 45 days after transplanting (DAT); (2)
risk of terminal drought and allowing earlier planting of a fol- direct-seeded rice (DSR)—soil ploughed prior to dry seeding
lowing nonrice crop (Saleh et al 2000). An earlier rice harvest (2001) or ploughed and puddled before direct seeding of
can also be achieved by planting early-maturing rice varieties. pregerminated seed (2002 and 2003) in rows by hand, with
Swarna, the most widely grown cultivar, matures after 140 to hand weeding at 21, 33, and 45 days after sowing (DAS); (3)
145 days and, when transplanted, may not be harvested until direct-seeded rice with chemical weed control (DSRH)—as
early to mid-November. In many years, soil dries rapidly at with DSR but with oxadiazon (375 g a.i. ha–1) applied 2–4
this time, reducing the likelihood of successful chickpea es- days after seeding, with one hand weeding at 33 DAS. Nutri-
tablishment. ent regimes were (kg ha–1) (1) single superphosphate, 40 P +
DSR reduces labor and draft power requirements for rice 40 K; (2) compound 60 N + 40 P + 40 K; (3) farmyard manure
establishment by 16% and 30%, respectively, compared with (FYM) + inorganic fertilizer totaling 60 N + 50 P + 50 K; and
Rice
2001 1.92 ± 0.10 2.79 ± 0.19 2.91 ± 0.12 2.85 ± 0.11
2002 2.80 ± 0.13 2.59 ± 0.11 2.96 ± 0.08 3.75 ± 0.15
2003 0.61 ± 0.08 0.51 ± 0.05 1.62 ± 0.21 2.67 ± 0.16
Chickpea
2001-02 –a – 1.01 ± 0.06 0.91 ± 0.05
2002-03 – – 0.76 ± 0.05 0.49 ± 0.04
2003-04 – – 0.38 ± 0.04 0.16 ± 0.02
(4) diammonium phosphate (DAP) (18% N) + Guti slow-re- planting in 2002 and under direct seeding in 2001. Yields of
lease urea (45%N) totaling 43 N + 40 P + 40 K. Rice was transplanted rice in 2003 were severely depressed by drought
harvested in a 5-m2 area. Biomass of individual weed species from early July to mid-August that led to late transplanting,
was recorded in two unweeded quadrats per plot at 28 DAS/ whereas, under direct seeding, yields were higher and Swarna
DAT and total weed biomass at 45 DAS/DAT. yielded 1 t ha–1 more than BR39.
Trials began at 16 sites during the 2003 monsoon sea- Nutrient application (compound and DAP/Guti) in-
son to verify the profitability of a DSR rice-chickpea system. creased tiller and panicle number, plant height, and grain yield
Chickpea (cv. Barisola 2) was sown after the harvest of Swarna in both cultivars. Tiller numbers were lower in the nutrient
or three shorter-duration BRRI dhan cultivars established by regimes single superphosphate and FYM + inorganic, espe-
either transplanting or direct seeding. Prior to dry direct seed- cially in transplanting. Under transplanting, panicle number in
ing in June, the land was ploughed (at least 3 times) with an BR39 decreased in comparison with Swarna. The main weed
animal-drawn country plough and leveled with a ladder. Seed species present at the site were Fimbristylis miliacea, Cyperus
was sown in lines by hand into furrows opened by a hand- iria, C. halpan, and Cynodon dactylon. In the first two sea-
pulled lithao (plough). Seedbeds were established at the same sons, greater weed biomass developed in rice established by
time and seedlings were transplanted at approximately 30 DAS direct seeding than under transplanting (P <0.05). Overall, the
following conventional ploughing and puddling operations. In most weed biomass was present where DAP/Guti was used.
direct-seeded rice, a single application of oxadiazon (375 g When herbicide was used for weed control in direct-seeded
a.i. ha–1) was made to control weeds, whereas, in transplanted plots, weed biomass was least under all fertilizer treatments
rice, pretilachlor (450 g a.i. ha–1) was applied. except DAP/Guti. In transplanted plots and direct-seeded plots
Seasonal variation in rainfall was considerable. The an- not treated with oxadiazon, the rank order of weed biomass at
nual rainfall in 2001, 2002, and 2003 was 1,475, 1,464, and 45 DAT was sedge > grass > broadleaf weeds. The use of
932 mm, respectively. In July 2003, rainfall was 2.5 times less oxadiazon changed the ranking: grass > sedge > broadleaf
than in the same month in previous years and in 2001 it was weeds.
highest in October. The yields of chickpea after direct-seeded rice were sig-
nificantly higher following BR39 than after Swarna in 2002
and 2003 (P <0.001) and elevated in 2001 (Table 1).
Results
Systems trial On-farm verification of the DSR-rabi system
There were significant effects of cropping system on the phe- Rice yields were considerably higher on-farm in 2003 (Table
nology of rice. Flowering was always later with cv. Swarna 2) as the drought in July and August was more prolonged at
and under transplanting (P <0.05). On average, grain fill dura- the site of the systems trial. On-farm, the yields of cvs. Swarna,
tion was similar over varieties, but was reduced by transplant- BR31, and BR32 were independent of crop establishment
ing, especially for BR39. Direct-seeded rice reached maturity method, whereas yields of transplanted BR39 were over 1.8 t
significantly earlier than transplanted rice (P <0.01), allowing ha–1 higher than when the crop was direct-seeded. Chickpea
chickpea to be planted 8–16 days earlier. Swarna significantly yields were not significantly affected by the preceding rice
outyielded BR39 in 2002 when direct-seeded, and also under cultivar.
transplanting in 2001 (P <0.05, Table 1). However, no signifi-
cant difference was observed between varieties under trans-
2
References
Mazid MA, Bhuiyan SI, Mannan MA, Wade LJ. 2002. Dry-seeded
0 rice for enhanced productivity of rainfed drought-prone lands:
BR31 BR32 BR39 Swarna lessons from Bangladesh and the Philippines. In: Pandey S,
Cultivar Mortimer M, Wade LJ, Tuong TP, Hardy B, editors. Direct
seeding in Asian rice systems: strategic research issues and
Fig. 1. Productivity (t ha–1) of rice and chickpea
grown in transplanted rice (open columns) and opportunities. Manila (Philippines): International Rice Re-
direct-seeded rice (solid columns) systems. search Institute. p 185-201.
Data are means of 16 on-farm sites in 2003 Mazid MA, Jabber MA, Riches CR, Robinson EJZ, Mortimer M,
(monsoon and rabi seasons). Weed control in Wade LJ. 2001. Weed management implications of introduc-
rice is by preemergence herbicide. Data above
each pair of bars are chickpea yields (t ha–1 ± ing dry-seeded rice in the Barind Tract of Bangladesh. Pro-
S.E.). ceedings of the BCPC Conference–Weeds 2001. 1:211-216.
Mazid MA, Jabber MA, Mortimer M, Wade LJ, Riches CR, Orr AW.
Conclusions 2003. Improving rice-based cropping systems in north-west
Bangladesh: diversification and weed management. The BCPC
The systems trial confirmed that replacing transplanted rice International Congress, Crop Production and Protection.
with direct-seeded rice could improve farm productivity in the p 1029-1034.
Nur-E-Elahi AH, Khan MR, Siddique A, Saha M, Nasim M,
Barind by allowing a greater opportunity to raise a high-value
Shahidullah SM. 1999. Existing cropping patterns of
rabi crop. The early-season weed flush associated with direct
Bangladesh: potential techniques and strategies for improv-
seeding can be successfully controlled by oxadiazon applied ing system productivity. In: Mandal MR, editor. Proceedings
preemergence. However, one subsequent manual weeding will of the Workshop on Modern Rice Cultivation in Bangladesh.
be essential for yield protection from weed competition and to Gazipur (Bangladesh): Bangladesh Rice Research Institute.
prevent a buildup of Alternanthera sessilis, Cyperus iria, and p 107-169.
Paspalum distichum in particular. While extensive cultivar Saleh AFM, Bhuiyan SI. 1995. Crop and rainwater management strat-
evaluation is under way, cvs. BR31 and BR32 represent prom- egies for increasing productivity of rainfed lowland rice sys-
ising lines for direct seeding. BR39, on the other hand, is not tems. Agric. Syst. 49:259-276.
suitable for direct seeding as sterility is a major problem when Saleh AFM, Mazid MA, Bhuiyan SI. 2000. Agrohydrologic and
it is planted early because it tends to flower in rains. Success- drought-risk analyses of rainfed cultivation in northwest
Bangladesh. In: Tuong TP, Kam SP, Wade LJ, Pandey S,
ful chickpea cropping after rice is contingent upon the pres-
Bouman BAM, Hardy B, editors. Characterizing and under-
ence of residual soil moisture and the time-window for suc-
standing rainfed environments. Manila (Philippines): Inter-
cessful chickpea establishment may be difficult to exploit. national Rice Research Institute. p 233-244.
Chickpea yields in the systems trial reported above were al-
ways higher because crops were established immediately after
the rice harvest. In on-farm trials in 2003, the potential advan- Notes
tage from direct seeding was not evident because widespread Authors’ addresses: M.A. Mazid, Bangladesh Rice Research Insti-
rain showers during the first two weeks of October favored tute (BRRI), Post Rangpur Cadet College, Rangpur 5404,
establishment regardless of the time of the rice harvest. High Bangladesh, e-mail: brrirang@bdonline.com; M.A. Mortimer,
yields were not achieved, however, because of late-season University of Liverpool, L69 3BX, UK; C.R. Riches and A.
drought. Our associated socioeconomic studies indicate that, Orr, Natural Resources Institute, University of Greenwich,
although farmers are keen to gain additional income from grow- ME4 4TB, UK; B. Karmaker and A. Ali, BRRI, Post
ing chickpea, many, particularly resource-poor share-croppers Shyampur, Rajshahi, Bangladesh; M.A. Jabbar, BRRI, Post
who pass 50% of their production to the landlord, need prac- BRRI, Gazipur, Bangladesh; L.J. Wade, The University of
Western Australia, Crawley, WA 6009, Australia.
tices that maximize rice yield. To be widely adopted for direct
Acknowledgments: This work was partially funded by the UK De-
seeding in place of Swarna, an early-maturing rice cultivar will
partment for International Development (Crop Protection
need to be high-yielding and sheath-blight-resistant. The re- Programme Project R8234). However, the views expressed
duction in input costs (no nursery and substitution of labor are not necessarily those of DFID. Support has also been re-
with a herbicide) that can be achieved with direct seeding and ceived from the Consortium for Unfavorable Rice Environ-
early planting of chickpea was evaluated favorably by an on- ments (ADB/IRRI).
farm trial with farmers in 2003. Farmers considered that even
Malaysia’s rice area was 692,600 ha for the combined dry- dominant weeds. In 1993, these were replaced by L. chinensis,
and wet-season crops in 2000. In Peninsular Malaysia, rice E. crus-galli, M. vaginalis, and E. colona in areas that were
double cropping in the eight granary areas accounted for a by then mostly direct-seeded. From 1981 to 1994, the areas
third of this total. Direct seeding (DS) as an alternative to the infested by L. chinensis grew by an estimated 4% per annum
traditional transplanting method in the main granaries in Ma- and all areas in Muda were infested by 1992. A similar trend
laysia began in the late 1970s and thereafter expanded rapidly was observed in Vietnam, where E. crus-galli, L. chinensis,
(Azmi and Abdullah 1998). By 2000, this method accounted and weedy rice were dominant weeds in the DS systems (Chin
for more than 90% of the total rice area planted. Wet seeding 2001). In Thailand, Sphenoclea zeylanica, Monochoria
rather than dry is favored but is only feasible where good wa- vaginalis, and Marsilea minuta are dominant in the transplanted
ter management is possible. DS is undertaken by hand broad- systems, whereas E. crus-galli, L. chinensis, Cyperus iria, and
casting or using a motorized blower. C. difformis are dominant in the wet-seeded areas (Vongsaroj
Rice is the most important crop in Vietnam and Thai- 1997). Some of the shifts in the importance of weed species as
land. The crop is grown in Vietnam on 4.2 million ha or a result of changes in crop establishment method from tradi-
74.1% of the arable land and produces 29 million tons every tional transplanting (1970s) to DS culture are shown in Table
year. The Mekong River Delta in the south and Red River Delta 1. Weedy rice is difficult to control because of its similarity to
in the north are the two main regions for rice production, oc- the rice plant. It was detected in Malaysia in 1988 (Azmi and
cupying 52% and 18% of the national rice area. In the Red Abdullah 1998) and Vietnam in 1994 (Chin 2001), and there
River Delta, rice is largely transplanted and double-cropped, has also been a rapid ingression of weedy rice in the DS area
whereas, in the Mekong Delta, about half the rice area is irri- of the Central Region of Thailand in recent years. Recent weed
gated, there are three growing seasons, and the crop is mainly surveys conducted in Malaysia found that weedy rice (Oryza
direct-seeded. Wet seeding is the most common practice (about sativa) had become a problem, and the weed is thought to be a
60% of the area), with water seeding, zero tillage, and dry serious threat to rice production (Azmi et al 2003).
seeding making up the balance. Rice cultivation in Thailand
covers about 10 million ha. At the end of the 1990s, the major-
Yield loss from weeds
ity of the rainfed rice area was dry direct-seeded and the re-
mainder transplanted, whereas irrigated areas in the Central Yield losses largely depend on season, weed species, weed
Plain were largely wet-seeded. Ten years earlier, however, most density, rice cultivar, and growth rate and density of weeds
of the rainfed areas had been transplanted, as were 30% of the and rice. Weedy rice at 35% infestation can cause about a 60%
irrigated areas of the Central Plain, with the remaining areas yield loss (Watanabe et al 1997) and, under a serious infesta-
being dry-seeded. This rapid change in establishment meth- tion, a yield loss of 74% was recorded in DS rice (Azmi and
ods was the result of labor shortages in rural areas, particu- Abdullah 1998).
larly during the peak season.
Impact of prolonged herbicide usage in direct-seeded rice
Weed succession in rice ecosystems Labor shortages have led to an increase in direct seeding and a
The widespread introduction of DS, the repeated use of herbi- rapid increase in the use of and reliance on herbicides. This
cides, and limited irrigation supplies in the 1990s are the fac- reliance has resulted in an undesirable shift in weed species
tors responsible for the shift in weed species populations in and concerns about environmental contamination. Phenoxy and
rice ecosystems. Grasses, such as Echinochloa crus-galli, along sulfonylurea compounds are widely used herbicides in Malay-
with other Echinochloa spp. (E. oryzicola, E. colona, E. sia, Vietnam, and Thailand to control broadleaf weeds and
stagnina, and E. picta), Leptochloa chinensis, and Ischaemum sedges in DS rice. Weeds resistant to these herbicides have
rugosum, which were not prevalent and dominant in Malay- evolved and there is evidence that weed species such as S.
sian rice fields in the 1970s (Azmi et al 1993), became wide- zeylanica, Marsilea minuta, and F. miliacea have developed
spread in the 1990s. Weed surveys were carried out in areas resistance to phenoxy herbicides (Watanabe et al 1997). Lately,
where direct seeding was being introduced to replace trans- acetolactate-synthase (ALS) inhibitor-resistant biotypes have
planting from 1989 to 1993. In 1989, while transplanting was been reported in several weed species in many countries, in-
the principal system, Monochoria vaginalis, Sphenoclea cluding Malaysia (Azmi and Baki 2003). ALS-inhibitor herbi-
guyanensis, Fimbristylis miliacea, and Leptochloa flava were cides, including sulfonylurea, are widely used in the world
Grasses
Isachne globosa Echinochloa crus-galli complex E. crus-galli
Leersia hexandra Leptochloa chinensis L. chinensis
Ischaemum rugosum I. rugosum
Oryza sativa (weedy rice) O. sativa (weedy rice)
Broadleaf weeds
Limnocharis flava L. flava L. flavaa
Monochoria vaginalis M. vaginalis S. guyanensisb
M. minuta L. hyssopifolia Sphenoclea zeylanicab
Sphenoclea zeylanica M. minuta M. crenatab
Limnophila erectaa
Sedges
Scirpus grossus Cyperus iria C. iria
Fimbristylis miliacea F. miliaceab
C. difformis
aBiotypes with herbicide resistance against 2,4-D and ALS-inhibitor herbicides. bSpecies/biotypes with her-
because of their low dosage, mammalian toxicity, and herbi- species. Further, water seeding (WS) can be practiced in lev-
cide injury in many crops. eled fields with levees where the water depth at seeding time
is 5–10 cm, sufficient to suppress weedy rice (Chin et al 2000).
Water seeding as opposed to wet seeding is effective in weedy
Weedy rice
rice suppression and can reduce populations by 70–76%. In
The term “weedy rice” refers to populations of annual Oryza Vietnam, WS can be practiced 10–20 days earlier than wet
species that diminish farmer income both quantitatively through seeding, allowing farmers to save water. Further, energy in-
yield reduction and qualitatively through lowered commodity puts can be reduced by nearly half because of the lack of drain-
value at harvest (Mortimer et al 2000). Weedy rice popula- age at the beginning of the cropping period and reduced irri-
tions are easy-shattering weedy types of rice, and are unwanted gation requirements from tillering to ripening. Weed control
plants among cultivated rice. These weedy rice populations input for WS is also reduced by one-half to a third that of wet
are found in Malaysia, Vietnam, and Thailand. In Malaysia, a and dry seeding, and a 69% reduction in the weedy rice popu-
close relationship between weedy rice and cultivated varieties lation was recorded after WS practices in Selangor, Malaysia
has been shown, giving a strong indication that evolutionary (Azmi 2004, unpublished data). WS practices for two con-
forces are still present in the rice ecosystems (Abdullah et al secutive seasons reduced weedy rice infestation to below 10%.
1997). According to Mortimer et al (2000), three factors that
determine the population of weedy rice are (1) seed remaining
Conclusions
dormant in the soil over crop seasons, (2) dissemination through
crop seed contamination, and (3) seeds returning from plants Effective weed management practices are an important pre-
in the previous crop. No single control measure will effec- requisite in DS rice culture with herbicide application seem-
tively control weedy rice. An integrated approach involving ingly indispensable. With the advent of herbicide-resistant weed
cultural, physical, and chemical interventions is expected to biotypes and unwarranted environmental contamination fol-
be effective in managing the weedy rice problem in a sustain- lowing the continuous and indiscriminate use of herbicides in
able manner. Unless the problem is addressed, weedy rice in Malaysia, Vietnam, and Thailand, a more sustainable integrated
many areas poses a major threat to sustainable DS rice pro- weed management (IWM) technology needs to be developed.
duction. This might be through the thorough understanding of the biol-
ogy, ecology, and socioeconomics of the major weeds in the
rice ecosystems. IWM, the promotion of biological and cul-
Water seeding tural control options, and minimizing the use of herbicides are
Besides tillage practices, good water management is an im- seen as keys to sustainable rice-farming systems in the region.
portant prerequisite for controlling weedy rice and other weed
Rice and wheat are the staple food crops of India, contributing transplanting but direct seeding is likely to increase in circum-
nearly 80% of the total food-grain production, and the rice- stances of labor scarcity and increasing wage rates. These lat-
wheat rotation is the principal cropping pattern of the Indo- ter conditions are now common in parts of India and farmers
Gangetic Plains, where the system occupies some 13.5 million are considering the alternatives offered by direct seeding.
ha. Demand for food grains in India is expected to grow and Balasubramanian and Hill (2002) commented that di-
the requirement for 2025 is estimated to be a 40% production rect seeding offers the advantages of reduced labor require-
increase compared with 2003-04. There has, however, been ments and drudgery, earlier crop maturity, more efficient wa-
stagnation in rice productivity in recent years, and long-term ter use and higher tolerance of water deficit, fewer methane
experiments show a declining rice yield trend (Padre and Ladha emissions as the system is more aerobic, and often higher profit
2004). Further, increasing costs are leading to demand for in areas with an assured water supply. Weeds, however, are the
change in rice production methods to improve returns to farm- main constraint for farmers practicing direct seeding since the
ers. Direct seeding in place of transplanting is one alternative inherent weed control from standing water at crop establish-
being considered. ment is lost. Differing levels of farmers’ resources, land devel-
In the Indo-Gangetic Plains, rice is traditionally trans- opment and infrastructure, or weather patterns may dictate dif-
planted at the end of the dry season (May/June) and, after the ferent direct-seeding practices in the Indo-Gangetic Plains.
rice harvest, wheat is sown (November/December). Direct Options available to farmers include wet seeding with sprouted
seeding has already replaced transplanting in many parts of seeds (broadcast or drum seeding) on puddled fields, dry drill-
Southeast Asia. The combined effects of changing water avail- ing of seed with or without prior cultivation, or broadcasting
ability and the opportunity cost of labor are driving reevalua- dry seed after dry-cultivating the field. This paper summarizes
tion of crop establishment methods. Pandey and Velasco (2002) results from an ongoing study of the options for direct-seeded
argued that low wage rates and adequate water supply favor rice in the Indo-Gangetic Plains.
CW HW TO CW TO CW HW TO CW TO
TP 7.8 7.2 5.9 4.9 3.4 6.1 5.5 4.6 5.8 3.6
WS 8.1 5.6 1.0 – – 6.8 5.4 0.9 5.9 2.2
DS 6.1 1.0 0.0 4.9 2.6 6.7 5.1 1.0 6.3 3.6
ZT 6.6 1.5 0.0 – – 5.7 5.4 0.2 – –
Materials and methods experimental details are given in Singh et al (2003). In farm-
ers’ fields, CW and TO treatments were imposed as described
Treatments compared rice establishment methods as main-plot above for TP, WS, and DS establishment methods.
treatments with subplots of weed management treatments in
rice, as follows:
Results
Main plots (60 m2)—rice establishment: In 2001, grain yield from transplanted rice was similar to that
TP = conventional transplanting of approximately 21- of wet-seeded rice when weeds were controlled (CW), whereas
day-old plants after soil puddling, 20 × 20-cm spac- yield from dry-seeded rice (DS, ZT) was about 20% less (Table
ing (seedling nursery established at the same time 1). A single hand weeding (HW) was insufficient to prevent
as direct seeding). major yield loss from weeds in either the DS or ZT rice and
WS = wet seeding (pregerminated, drum-seeded, 35 kg yield loss was complete without weed control (TO). In 2002,
ha–1) after soil puddling, 20-cm row spacing. when weeds were controlled (CW), yields under WS and DS
DS = dry seeding (50 kg ha–1) after conventional tillage, were similar or higher than under TP and ZT. Yield losses be-
20-cm row spacing. cause of weed competition, however, were less severe than in
ZT = dry seeding (50 kg ha–1), zero-tillage after flush 2001 and, with a single weeding (HW), yield in dry-seeded
irrigation (7 d before glyphosate application). areas decreased by an average of 16% versus 80% in 2001. In
farmers’ trials in 2001, equivalent rice yields were achieved
Subplots (20 m2)—intensity of weeding, postemergence weed from DS and TP rice, with potential yield losses from weeds
control in rice: (TO) being 30% in TP and almost 50% in DS. In 2002, DS
TO = unweeded checks. rice gave the highest grain yield, while potential losses from
CW = weed-free “best-bet herbicide treatment” followed weeds were the lowest in TP (37%), followed by DS (44%)
by two manual weedings. In TP plots, butachlor and WS (62%).
(1.5 kg a.i. ha–1) was applied 2 DAT; on the WS Weed species shifts in response to crop establishment
plots, anilofos (0.4 kg a.i. ha–1) was used 5 days are well known. At the start of the experiment in 2000, in the
after seeding (DAS); and, for the DS and ZT, absence of weeding, Echinochloa colona, Caesulia axillar-
pendimethalin (1.0 kg a.i. ha–1) was applied 1 DAS. ies, and other broadleaf weeds (other dicots) and Fimbristylis
HW = one manual weeding 30 DAS. miliacea were the most abundant weeds at 56 DAS in both the
From each subplot, weed counts and biomass (fresh transplanted and dry-seeded plots (Fig. 1). After three seasons
weight) by species were taken from two 0.25 m × 1 m quadrats under transplanting, however, E. colona, Commelina diffusa,
covering 5 crop rows. After the rice harvest, the experimental and Ischaemum rugosum were dominant, whereas, in dry-
area was sown to wheat after either conventional tillage or seeded rice, C. diffusa dominated. Furthermore, by 2002,
zero-tillage (as strips). The overall experimental design was a Leptochloa chinensis, an annual grass previously unrecorded
strip, split-plot design, with four complete randomized repli- at the site, was noted, and Cyperus rotundus, a perennial sedge,
cations as follows: wheat (cv. PBW-343 or PBW-154, 100 kg had also become proportionally more abundant in dry-seeded
ha–1, 20-cm row spacing) was sown in December into either areas over the three years up to 2002. The population dynam-
conventionally prepared plots (harrowed, rotavated, and lev- ics of individual weeds responded noticeably to establishment
eled) or zero-tilled plots (paraquat, 0.5 kg a.i. ha–1), 1 week method. E. colona, C. rotundus, C. diffusa, and I. rugosum all
before seeding. Yield estimates were taken from 5 m2. Full declined in abundance (density at 28 DAT/56 DAS) under trans-
Other 100
(dicot) C. axillaris
1.0 E. colona
I. rugosum
F. miliacea C. diffusa
10
10
0.1
2002 C. rotundus
E. colona
1.0 C. diffusa
0.01 Ischaemum rugosum Commelina diffusa
I. rugosum
C. axillaris
100
0.1 C. rotundus
F. miliacea
10
0.01
C. axillaris Other
(dicot) 100
F. miliacea
1.0 E. colona C. diffusa
I. rugosum
10
0.1 10 C. rotundus
2002
C. diffusa
E. colona
C. rotundus
0.01 1.0 L. chinensis Ischaemum rugosum Commelina diffusa
C. difformis F. miliacea
I. rugosum 100
0.1 C. axillaris
10
0.01
Fig. 1. Changes in abundance of weed species in relation to crop establishment method. Diagrams on the left are rank abundance curves
based on biomass at 28 DAT or 56 DAS comparing plots in 2000 and 2002. Diagrams on the right illustrate change in density (mean ±
standard error of mean) of selected species (at 28 DAT or 56 DAS) over five seasons. Log10 scales.
In rice cropping, water constitutes a powerful selective agent seeded rice may, however, extend over a longer period than in
for weed management (Mortimer and Hill 1999). Permanently wet-seeded rice, depending on the time of initiation and depth
flooded fields of transplanted rice that are unweeded by 45 of flooding.
days after transplanting (DAT) exhibit a weed flora different Intrinsically, seed size confers an advantage in seedling
from fields where the soil remains saturated. Noticeable dif- establishment (autotrophic net gain in biomass after germina-
ferences are also evident across toposequences where land tion). From a phyto-centric seedling point of view, the emer-
formation imposes different drainage and flooding regimes in gence of photosynthetic tissue above a water surface is a criti-
rainfed rice (Pane et al 2000). However, it is the timing, dura- cal life-history event in that there is a relatively abrupt increase
tion, and depth of flooding events in association with the in available photosynthetically active radiation (PAR) and ab-
method of crop establishment that govern the precise nature sence of hypoxic conditions. Successful water seeding of rice
of weed suppression by water. Wet seeding onto a saturated relies upon rapid mobilization of these reserves that is partly
puddled soil with subsequent flooding 10–15 days afterward dependent upon dissolved oxygen in floodwater and, similarly,
in comparison with transplanting into standing water alters the successful direct seeding of rice requires the absence of flood-
dominance hierarchy of weeds with the replacement of obli- ing immediately after sowing. Understanding seed and seed-
gate aquatic dicotyledonous species by semiaquatic species, ling biology at this stage in the life cycle is critical in under-
including competitive grasses. Weed species shifts may be rela- standing the role of water management in governing weed seed-
tively rapid. From observations of farmers’ fields, Azmi and ling recruitment. We briefly review the available knowledge.
Mashor (1995) recorded that seven seasons of wet seeding
resulted in the appearance of 21 new weed species, particu-
Weed biology: toward a definition of functional weed types
larly Echinochloa species and forms of “weedy” rice, and the
exclusion of 12 others. Germination
Understanding the processes underlying the selective Seed size in rice weeds varies on a logarithmic range from
nature of water in governing the recruitment and survival of approximately 10 µg (Sphenoclea zeylanica, Cyperus
weed species involves comparative autecological studies of difformis) to greater than 104µg (Rottboellia cocchinensis) and
seed germination and seedling and plant growth in relation to seed size in weed species is well known to be associated with
flooding regimes. These regimes may vary from permanently suites of dormancy/germination traits. In small-seeded species,
aerobic but moist soils, through continuously saturated anaero- the absence of oxygen (Monochoria vaginalis) or lowered
bic soils, to saturated soils subject to varying water depths oxygen concentration (Zinzania aquatica), fluctuating diurnal
arising from periodic flooding and drainage, either as a result temperatures (Najas graminea), and differential responses to
of rainfall variation or as a policy of water management. Po- spectral light ratios (Sphenoclea zeylanica) have all been shown
tentially, plants may experience all of these flooding regimes to be cues involved in response mechanisms that govern seed
at different stages in their life cycle and the diversity in rice germination rates underwater. Phytochrome-mediated switches
weed floras is in part because, in many semiaquatic species, in response to red/far-red light ratios provide another prob-
the rooting systems of adult plants are tolerant of, or adapted able mechanism governing germination although there has been
to, submergence in flooded soils. However, controlled studies little detailed work on the features of photocontrol of germi-
of weed community dynamics in transplanted and wet-seeded nation of many small-seeded rice weeds (Sanders 1994). Gap
rice emphasize the fact that weeds are recruited early in the detection mechanisms are moreover likely to confer fitness
life of both crops. In transplanted rice continuously flooded at advantages in a rice-weed community that will experience
shallow (< 50 mm) or deeper (approx. 100 mm) depth, re- largely uniform canopy closure, often within 30–40 days.
cruitment of plants to the weed community occurred only in Equally evident is the fact that small-seeded species (e.g.,
the first 10–20 DAT and subsequently losses continued until Cyperus difformis) exhibit polymorphisms in germination re-
the rice harvest. In wet-seeded rice initially sown on saturated sponse to flooding, providing “bet-hedging” tactics against
soil and subsequently flooded to these depths by 10–15 days unpredictable flooding events. Conversely, some species ex-
after seeding (DAS), recruitment of sedges and dicotyledon- hibit little or no innate or induced dormancy and germinate
ous species was restricted to the same early period in contrast rapidly on the surface of saturated soils (Fimbristylis miliacea,
to graminaceous species that increased in overall density up to Echinochloa colona, Kim and Moody 1989) but not underwa-
60 DAS (Hill et al 2001). Cohort recruitment of weeds in dry- ter. It is useful, therefore, to attempt to qualitatively group weed
6 Total biomass
(illuminated)
9
0.02
Total biomass
(dark)
6
3
Root and shoot
biomass (dark)
0.01
Structural seed
biomass 3
Root length (dark)
0 0.00 0
12
150
6
In dark Under full illumination
50
3
0 0
0 100 200 300 0 100 200 300
Thermal units (°C days) Thermal units (°C days)
Fig. 1. Seedling growth responses in Echinochloa crus-galli and Cyperus difformis in response to flooding. Germinated
seeds were placed in water columns containing nutrients to a depth of 50 mm and either illuminated (light/dark) on a 12-
h cycle or maintained in total darkness at a temperature of 30/20 °C day/night. PAR = 450 µE m–2 s–1. Repeated destructive
harvests were taken over the time course, enabling fit of the functions shown.
Kennedy RA, Fox TC, Dybiec LD, Rumpho ME. 1989. Biochemi- Pane H, Noor ES, Dizon M, Mortimer AM. 2000. Weed communi-
cal adaptations to anoxia in rice and Echinochloa seeds. In: ties of gogorancah rice and reflections on management. In:
Taylorson RB, editor. Recent advances in the development Tuong TP, Kam SP, Bouman B, Pandey S, Wade L, Hardy B,
and germination of seeds. New York (USA): Plenum Press. editors. Characterizing and understanding rainfed rice envi-
p 151-163. ronments. Manila (Philippines): International Rice Research
Kim SC, Moody K. 1989. Germination of two rice cultivars and Institute. p 269-287.
several weed species. Korean J. Weed Sci. 9:116-122. Sanders BA. 1994. The need for understanding the life cycle of the
Mortimer AM, Hill JE. 1999. Weed species shifts in response to rice weed Cyperus difformis. Aust. J. Exp. Agric. 34:1031-
broad-spectrum herbicides in sub-tropical and tropical crops. 1038.
Brighton Crop Protection Conference (1999) 2:425-437.
Farmers in South and Southeast Asia will have to increase rice Trends and impact on weed management techniques
production significantly to meet the needs of a growing popu-
lation. Competition between industry and agriculture for wa- Direct seeded wet-sown rice
ter, labor, and land will oblige rice farmers to intensify pro- Direct seeded wet-sown rice is widely practiced in Southeast
duction while using less labor and water. The risk of herbicide Asia, the southern provinces of China, and South Korea. The
overuse coming from those changes was highlighted in the technique has been widely known since the 1980s and, after a
1990s (Naylor 1994). Weed management practices must there- rapid initial development, further expansion has slowed down
fore be integrated and adapted to new cropping systems and because of the high constraints of this cropping system. Chemi-
associated weed problems (Labrada 2004). cal weed control is essential to avoid yield losses but high crop
In northeast Asia, the change in diet, opening of com- sensitivity at the early stages restricts the number of active
modity markets, aging farmers, and reduction in the cultivated ingredients that can be used. Accurate leveling and good wa-
area also compel rice farmers to increase productivity while ter management are also required to ensure adequate weed
maintaining higher product quality. These changes will have a control. High infestations of grasses such as Echinochloa crus-
significant impact on crop establishment and water manage- galli (L.) Beauv. and Leptochloa chinensis (L.) become domi-
ment practices. They will also affect weed populations and nant after a few cropping seasons. Weedy rice (Oryza sativa
herbicide use. L. fp. spontanea) is expanding in areas where wet-sown rice is
Until recently, the crop protection industry was focus- grown continuously, such as Vietnam, Malaysia, Sri Lanka,
ing on the discovery of new active ingredients. High develop- Thailand, and Korea (Baki et al 2000). Resistance of grass
ment costs and a stagnating and fragmented rice crop protec- weeds such as Echinochloa or Leptochloa spp. to propanil,
tion market are leading to a reduced investment in discovery quinclorac, and ACC-ase inhibitors is starting to appear in sev-
and development of new rice herbicides. With less chemical eral countries. Resistance of annual broadleafs and sedges to
innovation, the crop protection industry in the future will have ALS-inhibitors has been frequently detected in Korea (Park
to adapt use recommendations for current active ingredients 2004) and is starting to appear in China.
in order to solve new weed problems and ensure reliable per- During recent years, the crop protection industry has in-
formance under changing cropping practices and conditions. troduced more flexible and powerful postemergence herbicides
In the future, recommendations for herbicide use will increas- from the ACC-ase and ALS-inhibitor groups for grass or broad-
ingly take into account integrated weed management techniques spectrum weed control. Increasing reliance on such chemical
(Kudsk and Streibig 2003). classes will intensify the risk of developing weed resistance
This paper describes changes in weed problems and their (Valverde et al 2000). Syngenta is adopting an alternative ap-
management in three major Asian rice-cropping systems. Sus- proach. The 25-year-old herbicide/safener combination
tainable weed management solutions based on development pretilachlor/fenclorim is being rejuvenated by using an earlier
activities from Syngenta in Asia are used to illustrate how the application timing 3–7 days before seeding. The earlier appli-
crop protection industry responds to changes. cation eliminated the major constraint for farmers, which is
that it must be sprayed on mud within a few days after seeding
(Fig. 1). Direct application by dripping the concentrate or
slightly diluted product into water at the last leveling is easy 2,4-D or sulfonylureas will lead to an increase in Echinochloa
(Fig. 2). Crop tolerance is excellent. Performance is stable and spp. infestations. In areas with reduced water availability such
less affected by weather conditions or water management than as “tail areas” of irrigated perimeters or during El Niño years,
conventionally applied products. The method provides reli- or in areas with high cropping intensity, farmers will need tech-
able broad-spectrum weed control, including Leptochloa sp. niques to reduce water use and time for soil preparation.
and weedy rice. It also offers farmers a herbicide with a differ- The role of industry will be to register herbicides for
ent mode of action as a resistance management tool. This ex- controlling grasses and train farmers in the safe and effective
ample shows how exploring novel use recommendations for use of such products. New formulations for easier application
an old product can provide a solution to new weed problems. with a shaker bottle may be favored over products applied
through only foliar application. Syngenta is also promoting
Transplanted rice in South and Southeast Asia preplanting applications with nonselective herbicides such as
In areas with increasing labor costs, farmers will implement paraquat, allowing quick planting without tillage in several
techniques that reduce production costs and increase yield and Chinese provinces (Fig. 3) or reduced tillage in Indonesia and
cropping intensity. Techniques such as the use of younger seed- eastern India. The technique helps farmers to speed up soil
lings and reduction in hand weeding and herbicides such as preparation and planting by up to 1 month and reduce water
consumption by up to 15% (Kartaatmadja et al 2004). Appli- cides that can be applied during other operations such as lev-
cation of paraquat followed by no-tillage is currently the only eling or transplanting will also gain in importance.
technique allowing sustainable rice cultivation in Indonesian The crop protection industry is also working intensively
tidal rice. The development of reduced- or no-tillage together on new solutions to SU-R problems. Specialized active ingre-
with an adapted rice cropping system remains a possibility in dients such as carfentrazone, bromobutide, benzobicyclon, or
the fast-growing no-till wheat area in northwest India. others are being added into the already complex “one-shot”
mixtures, adding to production costs and restricting formula-
Transplanted rice in northeast Asia tion options. It is pertinent to compare the cost of a resistance
In northeast Asia, stringent pesticide laws will be enforced to prevention strategy with the cost of the cure. Calculations based
ensure the safety of users, consumers, and the environment. on Italian rice production systems have, for example, shown
Aging “weekend or part-time” farmers and also professional that rotation of different herbicide mode of actions can pre-
farmers managing a large number of fields will require broad- vent or delay the appearance of resistance at a much lower
spectrum herbicides to control annual and perennial weeds cost than using curative treatments when the resistant weed
(“one-shot herbicides”), which can be easily applied through population has developed (Orson 1999). In northeast Asia, it
dripping at planting or later from dikes or into water inlets. has been recognized that repeated use of “one-shot” products
Sulfonylurea herbicide-resistant (SU-R) annual broadleafs or in rice, relying only on sulfonylureas to control annual
sedges such as Monochoria vaginalis (Burm. f.) Presl var. broadleafs and sedges, is one of the main causes of the devel-
plantaginea (Roxb.) Solms-Laub. and Scirpus juncoides Roxb. opment of ALS-resistance (Itoh 2001). In line with the Herbi-
subsp. juncoides Roxb. are already widely distributed and will cide Resistance Action Committee (HRAC 2004) guidelines
continue to expand (Itoh and Uchino 2002). The control of for the prevention of resistance, Syngenta recommends an ap-
these weeds will require resistance prevention measures or the plication of pretilachlor applied preplanting or preemergence
addition of an antiresistance component into herbicide combi- in sequence with a pyriftalid-based “one-shot” to delay the
nations. development of SU-R weeds in farmers’ fields in both Korea
Maintenance and reinforcement of registration dossiers and Japan (Table 1).
for new and old herbicidal active ingredients will continue to
be a key priority for the crop protection industry. However,
Recommendations and conclusions
because of limited financial investment in a stagnating mar-
ket, cost and benefit will need to be balanced, which could In the coming years, scarcer land, labor, and water in conjunc-
lead to the divestment or phasing-out of older or economically tion with the need for intensification will drive significant
nonviable active ingredients. changes in cultivation techniques and weed management in
Responding to farmers’ demands, the industry will in- Asian rice production.
creasingly focus new developments on so-called “easy-to-ap- The number of farmers switching from transplanting to
ply” herbicides such as “Jumbos” or SC formulations. Herbi- direct seeding is expected to be minor and limited to large
aChemicals were applied at 3 DAT (3 days after transplanting of rice) at preemergence of SCPJO. bThe sequential application followed at 24
DAT.
The shift in rice establishment method from transplanting to mination process could be subdivided into imbibition, activa-
direct seeding has recently become prominent in Asia. Direct tion, and postgermination stages. We found that drying seeds
seeding constitutes either dry or wet seeding. Puddling the field at the activation stage could increase seed vigor by increasing
is characteristic of rice crop production and is generally con- the germination rate (Yamauchi 2002b). Dry seeds with high
sidered to increase soil nutrient supply, suppress weed infesta- vigor produced in this manner were found to have the poten-
tion, and increase water holding, thus contributing to tial to attain the same level of plant stand as pregerminated
sustainability. Puddling is conducted for transplanting and wet seeds.
seeding, but not for dry seeding. Therefore, poor water hold- We hypothesized that coating “high-vigor dry seeds” with
ing and heavy weed infestation are major problems in dry- iron could make wet seeding more adaptable and labor-saving
seeded rice fields. (Yamauchi 2003). Because dry seeds are storable, iron-coated
A variable rice plant stand is the major constraint to wet dry seeds can be prepared in advance of the busy farm season.
seeding. Puddling creates an anaerobic soil environment, thus
making seed germination and growth unstable, particularly
Preparation of iron-coated seeds
when seeds are sown beneath the soil surface. In the tropics,
farmers broadcast pregerminated seeds onto the puddled soil The procedure described here is designed for and being tested
after drainage. It was recently shown that rice cultivars toler- by rice farmers in Japan. Seeds with more than a 95% germi-
ant of anaerobic soil could be sown underground (Yamauchi nation percentage are surface-sterilized by soaking in sterilant
et al 2000). In Japan, farmers are recommended to seed “oxy- solution for 24 h, followed by ordinary soaking in tap water
gen-supplier” chemical (major ingredient CaO 2)-coated for 24–48 h at 15–20 °C. The soaking time depends on the
pregerminated seeds at a soil depth of 10–20 mm after drain- temperature and cultivar. Soaking is stopped just before co-
age. When the seeding depth is too deep, however, emergence leoptile emergence on the husk, then the seeds are dried. The
and plant stand are poor. In contrast, seeding on the surface or seeds thus prepared are vigorous.
in shallow soil may lead to floating seedlings when water is Seeds are coated with a mixture of iron powder and cal-
subsequently introduced. Drainage at seeding tends to stimu- cined gypsum (CaSO4 0.5 H2O, 5–25% of iron powder weight)
late weed growth. (Yamauchi 2004). The iron is in a reduced form. The amount
Water seeding is a form of wet seeding in which of iron used for coating is determined by the ratio of iron to
pregerminated seeds are sown in standing water. In the trop- seed weight. When the coating ratio is as high as 1, it is recom-
ics, it is practiced when drainage is difficult. In Japan, it was mended that an electric motor-driven granulator be used be-
practiced in a limited area before the materialization of “oxy- cause the weight of the mixture is too heavy to handle manu-
gen-supplier” coating technology. Although water seeding is ally. When the ratio is 0.1, granulation could be achieved in a
advantageous in water-saving and weed control, the occurrence container by mixing seeds and the mixture by hand. Portions
of floating or lodging of seedlings limits adoption. of the mixture are poured successively onto the seeds while
This paper describes the new technology developed to water is sprayed occasionally to help the granulation. When
overcome the floating and lodging problems of rice seedlings the granulation is over, calcined gypsum alone is dressed so
in standing water. that the surface of coated seeds becomes smooth and not chalky.
Because wetting the mixture of iron powder and gyp-
sum initiates rapid oxidation of iron, the temperature of coated
Hypothesis wet seeds increases. Therefore, the seeds must be spread in a
Seedlings in water are buoyant and tend to be poorly anchored, shallow layer and dried. Otherwise, there is a risk that seeds
leading to floating or lodged seedlings. We can improve an- will be injured and die because of the heat.
chorage by increasing the specific gravity of seeds by adher- The well-dried coated seeds can be stored at room tem-
ing heavy material on the husk. Among the materials commonly perature for several months without a significant decrease in
used in agriculture, iron has a high specific gravity of 7.9. We germination percentage. The specific gravity and single-seed
confirmed by use of pregerminated seeds that the occurrence weight of seeds thus produced are increased with the increase
of floating seedlings was reduced by iron coating (Yamauchi in iron-coating ratio (Figs. 1 and 2).
2002a).
Seed vigor is important to achieve a rapid and uniform
plant stand in fields (Yamauchi and Tun Winn 1996). The ger-
100
50
Specific gravity
4
3
Fig. 2. Intact (top) and iron-coated (bottom) rice seeds.
2
puddling. However, noncoated dry seeds failed to achieve an-
1 chorage and floated, even when seeded on the same day of
puddling.
In wet seeding, the soil surface tends to become hard
0 after drainage. When the seeds are broadcast-seeded onto such
0 1 2 3 4
Iron-coating ratio hard soil, the contact between seed and soil is poor and seeds
may become desiccated and die. Iron-coated seeds are advan-
Fig. 1. The relationship between iron-coat-
ing ratio and specific gravity and single- tageous because the seeds are heavy and become buried more
seed weight. Iron-coating ratio = iron pow- easily in soil. Thus, iron-coated dry seeds seem to allow greater
der weight/seed weight. flexibility in terms of seeding time in water and wet seeding.
Bird damage
Characteristics of direct seeding with iron-coated dry seeds Seeds and seedlings are vulnerable to bird attack in the field.
Performance in wet and water seeding When the field is drained as in wet seeding, they were often
We tested the seedling establishment and grain yield of rice destroyed by sparrows (Passer montanus). When the field is
from iron-coated seeds with coating ratios of 0.5, 1, and 2 in flooded for water seeding, the seeds are eaten by ducks (Anas
comparison with “oxygen-supplier” chemical-coated seeds in poecilorhyncha). Therefore, it is difficult to avoid bird dam-
wet and water seeding. “Oxygen-supplier” chemical-coated age, particularly when sparrows and ducks are in the vicinity.
seeds floated in water seeding, achieving a plant stand of 62%, It was noted, however, that iron-coated seeds are more resis-
whereas iron-coated seeds did not float, achieving a plant stand tant to sparrow attack at seeding. Sparrows cannot break the
of 82–94%. In wet seeding, the “oxygen-supplier” chemical- husk of iron-coated seeds and thus cannot eat the caryopsis.
coated seeds were damaged by sparrows, whereas the iron- Ducks are still able to eat the iron-coated seeds in flooded
coated seeds were affected only slightly, exhibiting 64% and fields, however, and, therefore, when ducks are expected to
81–94% plant stand, respectively. Thus, the plant stand of iron- attack, it is recommended to drain the field.
coated seeds is more consistent than that of “oxygen-supplier”
chemical-coated seeds under the experimental conditions used Environmental effects
here. There were little differences in grain yield (5,280–5,660 The use of iron-coated seeds for direct seeding should be dis-
kg ha–1) among the treatments. Our recent study showed that cussed and evaluated in terms of sustainability. While most
the coating ratio could be reduced as low as 0.1, although this water used for puddling has to be drained in ordinary wet seed-
depends on soil properties, crop management, and seeding ing, it is not necessary with iron-coated seeds because these
practices. seeds can be sown and establish plants successfully in the pres-
Seeding at the right interval after puddling is important ence of water. The introduction of iron-coated seed technol-
to the success of direct seeding and poor timing may end in a ogy may save water used for puddling. In addition, the flex-
poor plant stand. We tested the relationship between the time ibility of water control with iron-coated seeds would make
of seeding after puddling and plant stand in water seeding. herbicide application easier than in ordinary wet seeding.
The iron-coated seeds anchored well and established seedlings Water seeding is advantageous in the control of weeds,
successfully even when the time of seeding was 5–6 d after including weedy rice. Despite this advantage, water seeding
Rice production has been transformed over much of Asia, with and concerns about environmental contamination and human
total harvest increasing by 2.4% per annum from 1968 to 1999 health.
(IRRI 2004) through the introduction of improved germplasm, In addition to labor, there are increasing concerns over
agronomy, pest management, and, in many cases, mechaniza- supplies of irrigation water. Farmers in many rice-growing ar-
tion. At the same time, the declining availability of labor for eas are likely to have only limited availability of irrigation
agriculture and increasing labor costs have required farmers water and, in the future, most of the 22 million ha of dry-sea-
to seek alternatives to manual weeding, which has long pro- son rice in South and Southeast Asia will fall into an “eco-
vided farmers with the means to limit losses to weeds. Herbi- nomic water scarcity zone” (Bouman and Tuong 2003). Water
cide use in rice grew dramatically in Asia from 1980 to 1995, scarcity threatens the sustainability of irrigated rice ecosys-
with more than a threefold increase in sales, to more than $900 tems since it may no longer be feasible for farmers to under-
million per annum (Naylor 1996), allowing a massive release take wet cultivation and flood in fields to ensure good crop
of labor from agriculture that is expected to continue. The use establishment and control weeds. The development and adop-
of herbicides has been accompanied globally by the evolution tion of alternative irrigation strategies such as “alternate wet-
of herbicide resistance in weeds, weed species population shifts, ting and drying” and “aerobic rice” may enable good crop
Focus
Domains
Policy adviser Extension Farmer
Socioeconomic Implication of changes to estab- Cost-benefit analysis of compo- Choice of weed management
lishment or weed management nent technologies and system methods in relation to
practices: profitability O labor
O Herbicide use O knowledge
O Displacement of labor O resource availability
O Patterns of land and water O input costs
use
O Farm productivity
O Public/private partnerships
Output: policy brief Output: technology options/knowl- Output: leaflets, advice/decision
edge banks trees
Biophysical characteristics of pro- O Yield potential O Integration of weed man- Weed management options in
duction systems O Yield gaps because of weeds agement technologies, ag- relation to cropping systems,
O Stability of rice-based crop- ronomic practices, and cul- site location, water manage-
ping systems tivar selection ment, and other agronomic vari-
O Diversification of cropping O Prevention of herbicide re- ables
systems sistance
O Crop establishment meth- O Preventive methods to pro-
ods hibit weed species shifts
O Environmental protection
Output: policy brief Output: technology options/knowl- Output: leaflets, advice/decision
edge banks trees
Weed managementtechnologies Choice of mechanical/chemical Information on individual compo- Information on
control nent technologies O Target weed species
O Herbicide use
O Timing of application
Output: policy brief Output: technology options/knowl- Output: leaflets, advice/decision
edge banks trees
sensitivity analysis by partial budgeting. Pandey and Velasco farmer in the favorable rainfed environment, the question “What
(2002) have pointed to the likelihood of tradeoffs. For instance, are my options for rice establishment?” might initiate a tree
the introduction of direct seeding may have a negative impact involving several steps and covering the range of direct-seed-
on labor demand for the adoption of direct seeding in place of ing options (see Fig. 1 for an example).
transplanting or a positive effect where it allows for cropping
intensification, both of which may have policy implications
(Pandey and Velasco 2002). Likewise, improvements in wa-
Conclusions
ter-use efficiency may be achievable only by a switch in herbi- Traditional and modern rice systems alike are complex, the
cide product (e.g., butachlor to pretilachlor), which reduces management often sophisticated, and substantial information
the spectrum of control and necessitates more costs in later is required to enable farmers to judge objectively what the
manual weeding. best technology options are. This is certainly true in the transi-
As a third framework component, decision trees pro- tion to direct seeding and the management of intractable weed
vide farm-level information in the form of structured ques- problems. Gaining access to such information may present a
tions that enable answers in the form of options to be chosen. major obstacle for potential adopters. The challenge for re-
These trees specifically focus on technical issues related to searchers is to adequately address the variability of rice-farm-
the adoption of a particular system component and are most ing systems for which they are making recommendations and
effective when heuristics are employed. The question “Are to synthesize the results in ways that will make the conclu-
weeds that emerge early in the life of a rice crop the only ones sions available to those who will use them. Only in this way
to contribute to yield loss?” may initiate a tree that can lead to will farmers obtain benefit from advances in technology and
the recognition of the risk of intransigent weeds that in prac- be able to meet the challenges of a changing production envi-
tice require identification skills in scouting rice fields. For a ronment.
Yes No
Transplant
Can field be dry-cultivated?
Is Cynodon dactylon or Cyperus rotundus absent?
Yes No
Are soil conditions suitable for line Is there a need for interow cultivation
seeding by machinery? or substantial hand weeding
Yes No Yes No
Weed management
Are annual grasses absent?
Good water management possible?
Yes No
Apply herbicide
Apply herbicide + limited manual weeding or
+ limited manual weeding interflow cultivation
Fig. 1. Illustrative decision tree for adoption of direct seeding with respect to favorable rainfed lowland rice.
Weeds are the major biotic stress in rice cultivation in Sri Lanka, SAS computer package and the treatment means were com-
causing yield losses of as much as 30–40% (Herath Banda et pared by Duncan’s multiple range test.
al 1999). Leptochloa chinensis (L.) Nees, Red Sprangletop, is
becoming an important and serious threat to rice production in
Results and discussion
all rice ecosystems in Sri Lanka. It has the same outbreak level
as Echinochloa crus-galli (L.) Beauv. (Abeysekera 1999). Grasses were the dominant weed group found at the experi-
Chemical weed control is commonly practiced by rice farmers mental site, where the grass L. chinensis was dominant.
in Sri Lanka (90%) and propanil 36% EC was the most widely Echinochloa crus-galli, Isachne globosa, Paspalum distichum,
used grass-killer herbicide in rice fields before the introduc- and Ischaemum rugosum were also observed. None of the her-
tion of bispyribac sodium (Abeysekera 1999, Amarasinghe and bicides affected the germination and growth of rice seedlings.
Marambe 1998). In 1998, bispyribac sodium was the leading
herbicide and L. chinensis was selected as a serious weed in Phytotoxicity
wetland rice fields. After it germinates, it grows profusely in Cyhalofop-butyl, propanil, fentrazmide + propanil, quinclorac,
waterlogged spots in rice fields because of very poor land lev- and bispyribac sodium applied to rice plants did not show any
eling in farmers’ fields. At present, many herbicides have dif- visual phytotoxicity symptoms in rice plants. Clomozone +
ficulty in controlling this weed. Therefore, the objective of propanil–treated rice plants showed a slight whitening effect 2
this study was to evaluate the control efficacy of L. chinensis d after spraying. But these symptoms did not affect the grain
with postemergence herbicides in wet-seeded rice. yield of rice and were not visible after 7 d.
25
b
20
15
b
bc bc
10 bc
c
5 c
d d d
0
Cl + pro Cy Fe + pro Qui Pro Bis Nw Hw
Treatment
Fig. 1. Dry weight of L. chinensis as affected by different herbicides at 6 weeks
after sowing.
Yield (t ha2)
6
Minor
a Major
5 a
c
b b
b d
4 a b
a a
b b
3
c a
2 a
0
Cl + pro Cy Fe + pro Qui Pro Bis Nw Hw
Treatment
Fig. 2. Effect of different herbicides on rice grain yield.
chinensis. Fentrazmide + propanil 44% WP at 3 kg ha–1, Amarasinghe L, Marambe B. 1998. Trends in weed control of rice
quinclorac 250 g L–1 SC at 0.6 kg ha–1, and propanil 36% EC cultivation in Sri Lanka. Proceedings of a Multidisciplinary
at 8.5 L ha–1 applied at 7–10 DAS gave moderate control effi- International Conference, University of Peradeniya, Sri Lanka.
cacy (60%, 58%, and 40%, respectively) of L. chinensis. The p 272-274.
Bambaradeniya CNB, Gunatilaka CVS. 2002. Ecological aspects of
lowest control efficacy of L. chinensis was observed with
weed flora in an irrigated rice field ecosystem in Sri Lanka.
bispyribac sodium 10% SC at 0.3 kg ha–1 applied at 10–12
JNSF, Sri Lanka. 2 p.
DAS. No adverse effect was observed from all the tested her- Herath Banda RM, Dhanapala MP, Silva GAC, Hossain M. 1998.
bicides on rice grain yield. Constraints to increase rice production in Sri Lanka. Work-
shop on the prioritization of rice research, International Rice
References Research Institute, Los Baños, Philippines.
Fig. 3. Rice yield as affected by dry weight of Leptochloa chinensis. (A) minor season, (B) major season.
Direct seeding has increasingly become popular and adopted affordable practices and is closely monitored and supervised
by farmers in many rice-growing areas in the Philippines, es- by an interdisciplinary team of research specialists and exten-
pecially during the dry season. A recent survey of the Philip- sion experts from the Philippine Rice Research Institute. The
pine Rice Research Institute (PhilRice) revealed that more than project is piloted in the Agrarian Reform Community of
75% of the farmers in the major rice-producing areas are adopt- Maragol, Science City of Muñoz, Nueva Ecija, Philippines.
ing direct seeding. Iloilo has the largest area devoted to direct- The objectives of the study were to (1) enhance the ca-
seeded rice, followed by Occidental Mindoro and Negros Oc- pacity of rice farmers, (2) increase the yield and income of
cidental (PhilRice-BAS 1997). Wet seeding of pregerminated rice farmer-partners by promoting improved rice production
seed in puddled soil is more common than dry seeding and technologies, (3) link and network farm products to the main-
more farmers wet-seed during the dry season. stream and local market, and (4) assess the socioeconomic
The economic advantages of this mode of crop estab- impact of technologies and provide insights for researchers.
lishment include reducing the labor requirement as a result of
the elimination of the seedbed and transplanting operations,
Methodology
mitigating methane emissions, advancing harvesting by more
than 1 week, and other advantages such as using less water to Site selection was conducted based on the accessibility of the
mitigate drought and allowing dryland tillage in relation to area: large contiguous irrigated rice areas and farmers practic-
farm management techniques and practices. ing direct seeding.
A gamut of mature technologies exists for the direct- A baseline survey and participatory rural appraisal were
seeding rice ecosystem. A strategy identified for more effec- conducted to generate a location-specific database crucially
tive promotion of technologies is to develop a village into a important in the determination of rice farm households’ tech-
community of farmers for direct seeding, where indigenous nology base, needs, and problems. Site mapping was also con-
and modern technologies and practices merge. The village ducted using the global positioning system (GPS).
model serves as a springboard for developed technologies be- Spatial and temporal pest dynamics in the area were
fore these are eventually promoted. It is where technologies monitored. Rat population dynamics were monitored using the
are fine-tuned with the participation of farmers. The commu- trap barrier system + trap crop (TBS + TC) for four rice-crop-
nity is envisioned to use ecologically sound, sustainable, and ping seasons in 2002-04 (PhilRice 2000). The soil test kit
Monitoring and
evaluation
Fig. 1. Schematic diagram of the methodology in the direct-seeding rice village model.
(STK), minus one element test (MOET), nutrient omission plot, ing the wet season because of drainage problems as a result of
and laboratory analysis were used to assess the nutrient status the difficulty of water control.
of various rice fields at the study site (PhilRice 2002). The farmers’ association and organization are consid-
Researchers and farmers identified technologies and ered as an effective conduit in extending assistance to indi-
adaptable farmers’ practices for direct-seeded rice. Demon- vidual farmers. A farmers’ organization called Ugnayan ng
stration trials on the package of technologies, which include Magsasaka Makabagong Teknolohiya at Yamang Kalikasan
the use of 40 kg ha–1 of certified seeds, the drum seeder, and para sa Kaunlaran ng Maragol, Inc. (UMAYKAMI) was
carbonized rice hulls and organic fertilizer, were set up in the formed, and was officially registered at the Securities and Ex-
village. The performance of hand-tractor-drawn riding-type change Commission (SEC). Identified leaders were trained on
implements such as the harrow and leveler was also tested. basic leadership and community empowerment. Its members
The technologies introduced are being monitored through were taught about the technologies in direct seeding.
the farm record-keeping (FRK) approach. Farm record-keep- The population of insect pests was higher during the late
ing calendars were distributed to farmer-partners every crop- growth stages of the rice crop. Occurrence of major pests was
ping season for them to record their daily farm activities, ex- prevented by the 2-month fallow period and the use of resis-
penses, and revenues. Monitoring, documentation, and copy- tant varieties. Using the trap barrier system + trap crop (TBS
ing of data were done on a weekly basis by technical staff to + TC), it was observed that fewer rats were trapped as soon as
check whether farmers were religiously recording their daily the TBS +TC was established. This increased a week after the
farm activities, and problems encountered by the farmers were start of the field operations because of flooding and the de-
discussed. The gathered data were coded and encoded using struction of rat burrows in the levees and surrounding areas.
Microsoft Access. These were processed and analyzed using More rats were trapped up to the early maximum tillering pe-
frequencies, means, averages, and costs and returns analysis. riod than at the later rice growth stages. Community-wide physi-
The profitability and productivity between direct-seeded and cal rat control was recommended at the start of the cropping
transplanted rice were also compared. season. It was recommended that the time to use baits was
Figure 1 shows the schematic diagram of the methodol- during the early rice growth stages, not throughout the crop-
ogy for the direct-seeding rice village model. ping period as is now being practiced.
Relative grain yields of the nutrient omission plots ranged
from 50% to 98% compared with those of the complete plots.
Results and discussion Results showed that indigenous nitrogen (N) and potassium
The project focused on about 200 ha of contiguous rice fields (K) were limited in supply but not phosphorus (P). The rec-
in the Agrarian Reform Community of Maragol. It is one of ommended rate of 90-40-40 kg NPK ha–1 provided an average
the 37 barangays in the Science City of Muñoz, Nueva Ecija. grain yield of 5 t ha–1. However, by increasing the yield target
It has a population of 2,294 or around 480 households. It has a to 7 t ha–1, the soil nutrient supply would be limited for all
total land area of 520 ha, of which 490 ha is devoted to rice three nutrients (N>K>P) and a higher P application would be
production. Rice is planted twice a year with the National Irri- needed to attain that yield target.
gation System (NIS) as the main source of irrigation. Direct The indigenous nitrogen-supplying capacity (INS) and
seeding is practiced by 90% of the farmers during the dry sea- the agronomic efficiency of applied N (AEN) of the farmer-
son. However, 90% of the farmers practiced transplanting dur- cooperators’ fields ranged from 50 to 70 kg N ha–1 and 5 to 25
1.03
32,250
4.30
7.50
3,865
15,868
16,383
3.69
7,324
4,679
DSR
sults showed that the areas tested have a high N-supplying
Wet season
capacity and that a small amount of N is needed as basal fertil-
izer or during the early vegetative stage.
0.54
34,861
4.73
7.64
4,832
6,308
4.79
11,538
22,679
12,182
The minus one element test (MOET) showed that most
TPR
of the fields were deficient in macronutrients but not in micro-
2003
nutrients.
Farmers were encouraged and trained to optimize the
5,879
1.20
56,063
6.67
8.41
3.82
6,514
13,044
25,438
30,625
use of biomass residue from the farm, for example, carboniza-
DSR
Dry season
4,966
1.06
54,298
6.46
8.41
15,445
4.09
5,998
26,409
27,889
TPR
0.54
30,353
3.84
7.90
5,425
19,728
10,625
5.13
DSR
seeds not only reduced the input costs in rice production but
also helped save seeds for other farmers’ use. This also had a
7,908
10,944
0.05
25,478
3.23
7.90
24,340
11,138
7.55
5,488
1.01
58,673
6.90
8.50
29,208
29,464
4.23
6,977
DSR
Dry season
cated by its higher gross returns, net income, and net profit-
cost ratios and its lower cost of production. Table 1 shows the
comparative costs and returns analysis between direct-seeded
7,080
16,950
0.71
52,773
6.21
8.50
30,825
21,948
4.96
6,795
Conclusions
DSR
–
–
–
–
–
–
–
–
–
Wet season
22,462
12,570
6,306
4,216
0.56
4.30
8.15
5.23
TPR
0.86
11,997
44,435
5.49
8.10
23,939
20,496
4.36
6,535
tion trials based on the results of the MOET and NOPT; provi-
DSR
Dry season
0.70
14,173
39,682
4.90
8.10
23,319
16,363
4.76
7,081
TPR
References
Gross returns (pesos ha–1)
Labor
Wrap-up of Session 6
The change in rice crop establishment from transplanting to di- Sowing pregerminated seeds into standing water (water
rect seeding has occurred in many Asian countries because of seeding) has advantages over wet seeding in terms of water sav-
rapid economic growth, increasing labor costs, and shortages of ings and weed control. The water used for puddling is not drained
water. In addition, there is a need to reduce the costs of rice and conserved in the field, whereas the presence of water in the
production in order to maintain profitability despite the declining paddy field, particularly at the time of rice crop establishment,
market price, a trend that is expected to continue in the future. A efficiently suppresses weed growth. A major limitation of water
change in cultivation methods could have a substantial influence seeding, however, is the tendency of seedlings to float, which
on the environment; therefore, a change from transplanting to causes uneven rice populations.
direct seeding should be carried out with a focus on the sustain- Direct seeding in place of transplanting can have a posi-
able use of water and concern regarding agricultural chemical tive impact on the productivity of cropping systems, particularly
use. in rainfed systems where water availability is limited. In the rainfed
Direct seeding is practiced in three main forms: dry seed- and drought-prone areas of Bangladesh, advancement of the rice
ing, wet seeding, and water seeding. In dry seeding, crop harvest through direct seeding increases the chance of a
nongerminated seeds are sown beneath the soil surface. The second crop. Direct seeding also has advantages in the rice-
plant stand is more consistent than in wet and water seeding as wheat cropping system of northern India, where water availability
the soil is aerobic. Dry seeding is advantageous when the water and labor cost are the major determinants of successful farming.
supply is limited at the time of land preparation and crop estab- Weed infestation is a major threat to yield and the further expan-
lishment. The cultivation system of “aerobic rice” in China, which sion of direct seeding in Asia. Where direct seeding has been
combines dry seeding with a “nonflooding” water regime, is sug- introduced, there is commonly not only an increase in weed bio-
gested as a water-saving technology. As the land is not flooded mass but also a change in the weed species. Repeated use of
and puddled, heavy weed infestation and high rates of water in- the same herbicide or those with the same mode of action risks
filtration are problems in dry seeding. the occurrence of herbicide resistance in weeds, and biotypes
Wet seeding is a widely practiced form of direct seeding in with resistance to acetolactate synthase inhibitor have been re-
which the land is flooded and puddled in the same way as in ported. Furthermore, the control of weedy rice presents major
transplanting. Pregerminated seeds are sown on the puddled soil challenges.
surface just after drainage. Drainage is important to achieve reli- With the high development costs for new herbicides, the
able crop establishment. In Malaysia, Vietnam, Thailand, the Phil- agrochemical industry currently focuses on novel uses and mix-
ippines, and Korea, the repeated use of herbicides and limited tures of current active ingredients in order to solve new weed
irrigation water have seriously complicated weed problems with problems associated with direct-seeded rice. Water management
the rapid shift in weed populations to annual grasses (e.g., in terms of timing, duration, and depth of flooding at the crop
Ischaemum rugosum and Leptochloa chinensis) and weedy rice establishment stage is critical in determining the scope for the
(Oryza sativa) that are difficult to control. use of herbicides. Understanding the biology of weeds in relation
In Japan, wet seeding is practiced differently from the to water profile is important in developing ecologically sound weed
method used in other Asian countries. Pregerminated seeds are control methods and to enable the exploitation of flooding for
commonly coated with an “oxygen-supplier” based on calcium weed control. A new technology that coats rice seeds with iron
peroxide and seeded under the surface of puddled soil. As the powder to increase the specific gravity would overcome the prob-
coating alone is not sufficient to overcome anaerobic soil condi- lem of seedling float and make water seeding possible. This may
tions, the field must be drained at the crop establishment stage. provide a means to control weedy rice and water seeding is sug-
Seeding beneath the puddled soil surface has the advantage of gested as a potential control measure. Thus, the weed problems
reducing seedling float and bird damage. Further, “hill seeding” associated with direct-seeded rice could be alleviated to some
of coated seeds under the soil surface using a specialized seeder extent by greater emphasis on combined herbicide use and wa-
has been shown to reduce crop lodging. ter management.
Rice has become an extremely important crop being cultivated good potential because of its low production cost. An auto-
worldwide. In fact, the total production of rice in the world matic coating machine has already been developed and com-
has reached almost 500 million tons and 90% of the total pro- mercialized. Coating technology will be expanded for the pur-
duction comes from Asian countries. Taking into consideration pose of reducing some operations by the simultaneous appli-
the world population of 6.3 billion, food per capita per year cation of herbicide, pesticide, and fertilizer.
has been estimated at 400 kg. This rough estimate was based
on the total world production of rice in addition to that of other Combine harvester
cereals (such as wheat, sorghum, soybean, and maize), almost A special head-feeding-type self-propelled combine harvester
equal to 2 billion tons, plus 400 million tons from vegetables, was developed and is commercially available in Japan. This
fruit, and meat. However, this amount is just enough for now machine can harvest from two to six rows depending on the
and the Chinese government would like to increase produc- specified width of cut. I present new concepts of the machine’s
tion to satisfy a growing population. In this paper, basic data functions below.
related to rice and its cultivation are discussed prior to devel-
oping solutions to increase rice production and protect the en- Combine husker (named by the author)
vironment. Normally, cutting of straw, threshing, winnowing, and clean-
Rice is cultivated basically for food consumption; how- ing of grains are the main functions of a commercial rice com-
ever, it can also be used as energy in the form of an environ- bine harvester. However, husking can be done simultaneously
ment-friendly fuel. This is possible by using its biomass, for through an additional mechanism I designed. Using the com-
which production efficiency is the highest among agricultural bine husker, rice can be directly harvested as brown rice, throw-
crops. About 430 L of ethanol can be produced from 1 ton of ing out other foreign materials such as rice chaff, husks, empty
rice biomass. Reliability is based on the fact that almost 17 to grains, and weeds after husking and scattering them on the
19 tons of carbon dioxide per hectare are absorbed by the grow- paddy field. The following improvements can be made by in-
ing rice plants, which is nearly equal to the amount of carbon troducing this type of harvester: first, a savings of energy for
dioxide absorbed by some forest trees. In addition, the rice transportation because of the reduction in bulk density through
maturity stage takes only 4 months, whereas trees take more removal of rice husks and other foreign materials during har-
than 10 years to reach maturity. In this case, ethanol from rice vesting; second, a saving of drying space equal to almost half
can easily be obtained as the demand for energy consumption of the rice volume.
increases. This will also help minimize the world’s dependency
on fossil fuel. The most significant advantage of using ethanol The turntable combine harvester
produced from biogas is the clean emission of gas compared To improve operating efficiency, the total time for completing
with that of fossil fuel during combustion. the harvesting operation and the total distance traveled to com-
These facts illustrate the important role of rice in help- plete the harvesting operation are normally used for evaluat-
ing resolve four global issues (called tetralemma), particularly ing how much this could be improved. The turntable combine
concerning the increased consumption of fossil fuel, which harvester turns normally using the turntable mounted on the
jeopardizes the environment. This condition will sooner or later main chassis. The shape ratio of the paddy field expressed by
result in fossil fuel exhaustion; therefore, we must develop an the ratio of length to width, the pattern of promoting harvest-
alternative solution to combat this threat. ing operation, the operating speed, the operating width of the
cutter bar, and the specifications of the machine must be con-
sidered as the factors for improvement. For the turntable com-
New technology in rice mechanization bine harvester, the total distance required to complete harvest-
Direct sowing by use of coated rice ing could be reduced by 10% compared with that of the con-
Direct sowing of rice is popularly accepted in many develop- ventional commercial machine. The total distance under the
ing countries because of its low cost and energy-saving tech- same conditions could be reduced almost 25% at maximum.
nology in planting; however, the yield per unit area is almost The turntable combine harvester can be recommended for its
half that obtained in many developed countries. Only the United safety, automation, and energy savings.
States practiced direct sowing of rice successfully, combined
with low-cost production and high application of technology. The combine harvester equipped
Yield is almost 7.7 t ha–1, which is 1.1 t ha–1 higher than in with a pivot turn mechanism
Japan’s case using the transplanting method. A new type of The improvement in turnability of the tracked vehicle can be
direct sowing by using calcium peroxide-coated rice seed has achieved by applying the pivot turn mechanism. The ground
Conclusions Notes
The world faces many threats. This paper has presented some Author’s address: Laboratory of Energy Utilization Engineering,
suggestions to help overcome these threats. First, the world Department of Environmental Science and Technology, Fac-
requires enough food, particularly rice, as the world popula- ulty of Bio-resources, Mie University, e-mail: ito-n@bio.mie-
tion is increasing dramatically. Second, world dependency on u.ac.jp.
fossil fuel must be minimized. Third, emissions of carbon di-
The field agricultural production system in Korea has focused planters and harvesters were distributed in the 1980s to over-
on paddy rice and plot size is relatively small (less than 1 ha). come labor peaks and shortages.
Until the late 1950s, manual and animal power were used in Agricultural mechanization technologies were developed
rice production. With the execution of the Five-Year Economic and extended in the 1990s to improve the agricultural system
Development Plans started in the early 1960s, agricultural and productivity. The change from manual and animal power
machinery such as water pumps and pesticide applicators were to about 98% mechanization in rice production took place
disseminated for disaster-prevention purposes. In the 1970s, within only 30 years, which is regarded as a significant achieve-
as a large portion of rural labor moved to factories in urban ment by other developing countries.
areas, full-scale agricultural mechanization projects began. The
power tiller replaced animal power in the 1970s, and rice trans-
1985 70 23 68 17 2.1
1990 84 78 93 72 15
1995 97 97 98 95 32
2002 100 98 100 99 49
Population changes in the rural community In 2002, the mechanization rate of tillage and land preparation
The farm population was 58% of the national population in was 100%, rice transplanting 98%, pest control 100%, har-
1960, but it gradually decreased as labor demand for other vesting 99%, and paddy rice drying 49% (Table 1). Except for
industries increased. The number of farmers decreased, from drying, all operations seem to be fully mechanized. Average
5.0 million in 1975 to 4.4 million in 1980, 3.1 million in 1990, tractor power was in the 45-hp (33.8 kW) class. Major attach-
and 2.2 million in 2000. The annual reduction rate is above ments for tractors are the Irang-Janggi (plow), rotavator, trailer,
3%. The current labor shortage is caused by (1) an increasing and others, and the width of the rotavator for the 40-hp class
portion of old and female labor and (2) the higher labor wages tractor is 170 cm.
in the rural community have promoted farm mechanization Four-row walking-type and six-row riding-type trans-
more actively. planters are the main machines extended. Three- and four-row
Farmers from 20 to 59 years old gradually decreased head-feed paddy rice combine harvesters are now the main
and farmers more than 60 years old gradually increased. In rice combine harvester. The number of paddy rice dryers in-
2000, farmers older than 60 made up 33.1% of the farm popu- creases as the paddy rice combine harvester is being propa-
lation. This is a big problem in Korea. gated, but sun drying is still practiced because of good weather
during the harvest period. Drying in the rice-processing cen-
Expanding farm size ter, which reached 340 units in 2003, is another choice for rice
Farm size averaged 0.9 ha in 1965. It increased to 1.37 ha in drying.
2000. However, the farm size is not big enough to carry out
farm-restructuring effects. A small farm is defined as a field
Research on and development of agricultural machinery
less than 1.0 ha in size. Medium-sized farms from 1 to 2 ha
for paddy rice
have decreased rapidly.
On the other hand, the number of large farms (more than The following machinery has been developed:
2.0 ha) has been increasing. The ratio of small farms among O Partial-tillage direct rice seeder (1999)
total farms decreased slowly. Farmers gave up the profession Attachable to a tractor, 8 rows, operation width 2 m
during the industrialization period and were absorbed by the (tillage 0.6 m, no-tillage 1.3 m, drain ditch 0.1 m).
industrial sector. O Paddy-field leveling rotavator (1999)
For preparing paddy fields before transplanting or di-
Number of agricultural machinery rect seeding of rice, 3.4 m in width and less than ±35
Since the 1960s, water pumps, two-wheel tractors, and pesti- mm in unevenness.
cide applicators have been extended to farms. Two-wheel trac- O Unmanned tractor (2000)
tors were 11,884 in 1970, 289,779 in 1980, and 751,236 in Autonomous travel technology based on a GPS navi-
1990. In 2003, the number reached 857,809. gation system and remote-control technology with a
Four-wheel tractors have been propagated on a large scale field monitoring system for the unmanned tractor.
since 1980, and their number reached 2,664 in 1980, 41,203 O Paper-mulching rice transplanter (2003)
in 1990, and 211,576 in 2003. With the rapid propagation of Developed for sustainable agriculture and no herbi-
agricultural machinery, rice cultivation operations, including cide application.
tillage, land preparation, transplanting, and harvesting, became O Partial-tillage rice transplanter (2002)
mechanized almost completely (about 98%) as of 1998. This Developed after experiments considered important
was enabled by the government’s strong enthusiasm for farm factors in untilled paddy fields and the design param-
mechanization and concentrated financial support to attain self- eters of a rotary-type machine.
sufficiency in major crops. O A rice wet-seeding broadcaster (2003)
To mechanize wet seeding, a broadcaster was devel-
oped for wet-seeding operations. The prototype is
composed of a metering device, a blowing part, and a
seeding part, which can be attached to a tractor.
The planting area of rice occupies nearly 31,000,000 hectares machine to directly sow seed in the paddy without irrigation,
on the mainland of China, and field yields account for 39% of to a depth of 2 cm. This method is required for crop unifor-
the total grain production. Rice is the largest grain crop, with mity. China now increasingly uses direct sowing of rice.
the largest planting area, highest single yield, and most total The rice factory-bred seedling technique. The rice fac-
production. China has 34.5% of the world’s grain yield, with tory-bred seedling is one means of mechanizing rice transplant-
only 21.4% of the rice planting area in the world. So China ing. According to the Chinese situation, in recent years, Chi-
has played a major role in global rice production. nese farmers have welcomed farm machinery and techniques
such as bred seedlings, precision-sowing equipment, carpet-
type seedlings, and pan-type seedlings using a shallow-trans-
The status of rice production mechanization in China planting machine.
Upon entering the 21st century, China had made great progress Mechanized rice transplanting. There are two ways:
in major areas such as tillage, breeding and planting, field man- transplanting and throwing seedlings. The transplanter uses
agement, and harvest. By the end of 2003, China had doubled carpet-type seedlings. Throwing seedlings uses pan-type seed-
the area of rice under mechanization since 1995. Mechanized lings with soil having less-bruising roots, more rapid green-
planting was 5.08% in 1995 and has increased nearly 3% per ing, and higher yield. Moreover, the efficiency of throwing
year since then. The mechanized area is 8 times more than that seedlings is higher and costs less, so Chinese farmers welcome
in 1995, whose level of rice mechanization reached 24%, with it.
an increase of 21.6%. Transplanters mainly adopt the crank-swing-type cap-
The status of rice production mechanization in China is ture-seedling mechanism and entire supporting slide plates with
such that the total level is relatively low, except for the level of a single wheel. The row distance is generally 30 cm, with ma-
tillage mechanization, which is higher. Mechanized planting nipulation by three operators. The thrown seedling is broad-
and harvest are lower, and planting is the weakest operation cast or falls into a row. The broadcasting machine throws seed-
and has the most difficulties. There are many reasons why farm lings using a centrifugal disc mechanism.
mechanization does not combine well with agronomy. China Harvest technology. In the past few years, mechanized
has a vast territory with complex agricultural situations, and a rice harvesting in China has developed. Implements include
large population with limited arable lands. The weak agricul- the general-purpose wheel-type combine harvester for rice and
tural economy, the large labor force, and expensive farm ma- wheat, as well as a cutting machine and binder. Recently, some
chinery have all contributed to the slow adoption of rice mecha- new implements have been developed, such as the rubber-track
nization. crawler, full-fed and head-fed combine harvesters, and strip-
per combine harvester suitable for harvest operations, thus
basically meeting needs in various areas.
Technical analysis of rice mechanization in China
Dry farming technology. China has been slow in devel-
Tillage mechanization oping machines suitable for dry cultivation. Even today, the
For the different steps in rice production mechanization, the level of mechanization is low. Rice mechanization has now
level of tillage mechanization is highest. There are many kinds become more important, so suitable machines for dry farming
of tillage machines: the plow, furrower, drive-disk plow, paddy- technology have become important.
field harrow machine, and rotary machines that meet require-
ments for use in the main rice production areas.
Prospects
With the economic change, the most imminent technique
is the straw-chopper. There is still much room for the development of new tillage
technology, and also perfecting existing technology for sow-
Rice planting mechanization ing and applying fertilizer. This would add much value. Imple-
Mechanized direct sowing of rice. China has two kinds of di- ments for returning straw to the field are urgently needed but
rect-sowing techniques for rice: wet direct sowing and dry di- are still in the development stage. Laser cultivation implements
rect sowing. The wet direct sowing is mainly aperture sowing, would be useful.
and some implements can sow the rice seed after it has been Under controllable environmental conditions, factory-
soaked and pregerminated. Other machines can sow sprouted bred seedlings are being provided to mechanized, large-scale,
rice seeds with a sprout of 3 mm. In contrast, the dry direct centralized commercial farms according to a standardized flow-
sowing is mainly drilling, which usually uses a wheat-drilling
Harvesting and related operations are among the most labor- In 1980, the Los Baños-based International Rice Re-
intensive operations in rice production in Asia. In the Philip- search Institute (IRRI) developed with Chinese engineers a
pines, harvesting and threshing consume roughly 60% of the local version of this type of reaper for release to Asian coun-
total labor devoted to rice (Takahashi 1994). Harvesting alone tries. The IRRI reaper was a simplified version of the Japa-
costs around 6–8% of total produce, whereas losses (including nese and Chinese designs and was released to small manufac-
gathering) can reach as high as 5% of total yield (Andales 1998). turers in the Philippines and a few other countries (IRRI 1987).
Back-breaking manual reaping alone requires 10–16 person- This design, however, was not as successful as the Japanese
days ha–l (Juarez et al 1988). model released shortly afterward in the Philippines because of
Attempts have been made to modernize rice harvesting its durability problem as a result of poor blade tolerance
in Asia, the most successful of which was reported by Ezaki (Stickney et al 1986), less convenient operation, and high walk-
(1970) on walking-type reapers in Japan. Later, modified ver- ing speed. The imported model is more durable, lighter, and
sions of this design were locally produced in China, Thailand, more convenient to use, with a slower forward speed and re-
India, and Pakistan. This design made use of a cutterbar as- verse gear that allows easy headland turning.
sembly, composed of reciprocating triangular blades placed Except for its high price, this imported reaper was found
above a stationary ledger plate. This cutterbar mechanism re- to be suitable to the variable field conditions in the Philip-
ciprocates at low speed, requires low power, and is highly suit- pines. Eventually, it slowly replaced the IRRI models in most
able for windrowing operations since the plants have practi- areas. No cheaper version was attempted locally to compete
cally no horizontal movement after cutting prior to conveying with this imported design because of a lack of alternative com-
to one side of the machine. ponents such as light diecast components and a small trans-
1,200
Gasoline
engine
Converter lug
Fig. 1. The PhilRice-JICA rotary reaper prototype (A) and commercial model (B). (A) Schematic view of
the prototype, (B) a commercial model during field operation.
Rice transplanting was mechanized in the 1970s in Japan, and with soil weighs about 6 kg, (2) 2,000 nursery boxes are nec-
now more than 99% of paddy fields are cultivated by mecha- essary for paddy fields of 10 ha, and (3) it is necessary to carry
nized transplanting. Even large-scale transplanters with higher nursery boxes many times from sowing to transplanting. The
efficiency and precision are becoming common, and mechani- larger the farmer is, the harder it is to carry the nursery boxes.
cal transplanting is assumed to have reached its technical per- Our center developed the long-mat seedling culture system
fection. But a problem remains. (long-mat system: LMS) to save labor and facilitate work in
Usually, a plastic box (60 × 28 × 3 cm), called a “nurs- comparison with the CSS (Tasaka et al 1996, 1997), and has
ery box,” is used for raising rice seedlings. Soil is packed into improved it further.
it, and seeds are sown. Then, nursery boxes are arranged in In this paper, we introduce the general characteristics of
vinyl houses and the seedlings are raised. When the seedlings this long-mat system and discuss the recent development of
grow up, the nursery boxes are put on a truck and taken to the rolling-up equipment that enabled a one-person operation, and
paddy fields. The seedlings are then transplanted by a trans- the improvement of the higher-precision transplanter for long-
planter. mat seedlings.
This conventional soil-seedbed system (CSS) has major
problems for the following reasons: (1) a nursery box filled
Push down
roller
Running
Running
direction
direction
Core for
Iron weight rolling up
plate
Running
direction
Motor
Belt
Nursery tray
Fig. 1. The principle of rolling up (upper right), manual rolling up by two people (middle), and mechani-
cal one-person rolling up (below).
facilities, too. It is expected that the long-mat technology will Tasaka K, Ogura A, Karahashi M, Niiyama H, Namoto H, Kaneko T.
spread to stimulate agriculture in the region. 1997. Development of hydroponic raising and transplanting
technology for mat type rice seedlings. Part 2. Development
and field test of rice transplanters for long mat type hydro-
References ponic rice seedlings. J. Jpn. Soc. Agric. Mach. 59:87-98.
Tasaka K, Ogura A, Karahashi M. 1996. Development of hydro-
ponic raising and transplanting technology for mat type rice Notes
seedlings. Part 1. Raising test of seedlings. J. Jpn. Soc. Agric.
Mach. 58:89-99. Author’s address: National Agricultural Research Center, e-mail:
kitaga@narc.affrc.go.jp.
Planting finger
Guides to gather
seedlings
Fig. 2. The outline of the long-mat transplanter (left) and remodeling for higher precision transplanting
(right).
With farm size increasing and the number of farmers decreas- describe the vehicle control methods and how the desired op-
ing, more efficient agricultural practices are needed. However, erating path is chosen. The experimental results are discussed,
in Japan, fields are not as yet sufficiently consolidated. They and the final section gives conclusions.
are frequently dispersed and it is difficult to use large machines
in areas that would be unable to bear their weight. This effi-
Materials and methods
ciency problem would be solved, however, if one operator were
able to control several lightweight machines simultaneously. Vehicle, sensors, actuators, and controllers
The goal of our research is to develop an autonomous operat- We modified a commercial 6-row rice transplanter (PH-6, Iseki
ing system for paddy fields that would permit one operator to Co., Ehime, Japan). Figure 1 shows the modified rice trans-
control multiple machines. The objective of this study is to planter controlled by a computer with a 486-compatible 66
develop an autonomous navigation system for the rice trans- MHz central processing unit. Figure 2 is a schematic for an
planter. automated rice transplanting system. An RTKGPS (MS750,
Several pieces of research on autonomous agricultural Trimble Navigation Ltd., Sunnyvale, California) with 2-cm
machinery have been published (Reid et al 2000). Recently, precision at 10 Hz data output was used to locate the position
an autonomous tractor using an optical surveying device and a of the rice transplanter. For communication between the
terrestrial magnetism sensor for plowing was developed RTKGPS reference station and the rover station, 5-mW out-
(Matsuo et al 2002). These researchers obtained good results put wireless radio modems (YRM211, Vertex Standard Co.,
but found it difficult to control multiple machines because each Ltd., Japan) were used with baud rates set at 9,600 bps. A FOG
vehicle required its own optical surveying device. Some re- sensor (JG-35FD, Japan Aviation Electronics Ind., Ltd., To-
searchers used a differential global positioning system (DGPS) kyo) was used to measure the yaw angle, and an inclination-
(Cho and Lee 2002). Kalman filtering of DGPS can effectively measuring apparatus (JCS-7401, Japan Aviation Electronics
correct DGPS position error (Han et al 2002), but it does not Ind., Ltd., Tokyo) composed of three FOG sensors and three
provide enough precision for operation in paddy fields. accelerometers was used to measure the roll and pitch angles.
Noguchi et al (2002) developed a robot tractor and Nagasaka The GPS position data, yaw, and roll and pitch angle
et al (2000) developed an automated rice transplanter. Both of data were transferred to a computer through an RS232C inter-
them employed RTKGPS and FOG sensors that established face. The sampling rate was 10 Hz for the GPS, 20 Hz for the
precise operations. In this paper, the results of automated rice yaw angle, and 25 Hz for the roll and pitch angles. The main
transplanting operation experiments are reported. We first de- computer corrected the GPS position data, which were influ-
scribe the sensors, actuators, and controllers used for the auto- enced by the vehicle inclination, and calculated the control
mated rice transplanter and how data are processed. Then we parameters, then sent them to an industrial programmable logic
Data
Modem Modem
RS232C
Main computer
RS232C
R5232C Yaw
Encoder Roll
Limit switch Pitch
Steering
HST
Clutch PLC
Brake (L, R)
Engine throttle
Attachment up-down Motor
Controller
Path planning
Before starting operation, the computer must create a desired
Conclusions
path along which the rice transplanter travels. In this study, the An automated rice transplanter was developed and autonomous
paddy field was assumed to be rectangular. The four corners guidance operation was conducted. The maximum deviation
in the field were measured beforehand. The sowing width of a from the desired straight path is 12 cm and RMS deviation is
6-row rice transplanter is 1.8 m. Since a paddy field usually 5.5 cm at a 0.6 m s–1 operating speed. This is only accuracy
has only one entrance, the desired path depends on the width enough for rice transplanting operations, and not enough for
of the field. spraying and mechanical weeding operations after rice trans-
planting. Since it is necessary to drive an autonomous vehicle
Vehicle control between the crop row to do spraying and mechanical weeding
In paddy fields, it is thought that the paddy ground is flimsier operations, the maximum deviation must be reduced to less
than upland fields. Several papers discuss steering control in than 5 cm. It is also necessary to evaluate the control algo-
upland fields. O’Connor et al (1996) used a steering controller rithms.
based on a set of linear motion equations. Cho and Lee (2000)
used a fuzzy controller for autonomous operation for an or-
chard speed sprayer. Kise et al (2002) developed an optimal
References
steering controller and obtained good curved-path guidance Cho SI, Lee JH. 2000. Autonomous speed sprayer using differential
results. Inoue et al (1997) developed an adaptive steering con- global positioning system, genetic algorithm and fuzzy con-
troller that corrected steering system delay. Most of these re- trol. J. Agric. Eng. Res. 76:111-119.
searchers analyzed vehicle dynamics and made a dynamic Han S, Zhang Q, Noh H. 2002. Kalman filtering of DGPS position
for a parallel tracking application. Trans. ASAE 45(3):553-
model, then decided on the control parameters.
559.
In this study, we used a simple proportional controller
Inoue K, Otuka K, Sugimoto M, Murakami N. 1997. Estimation of
for steering control before evaluating the effectiveness of ad- place of tractor and adaptive control method of autonomous
vanced controllers in loose-ground-like paddy fields. tractor using INS and GPS. Preprints of the International
At the headland, the rice transplanter moves forward and Workshop on Robotics and Automated Machinery for Bio-
backward during a turn so as to minimize the headland space. Productions, 27-36, EurAgEng, Gandía, Spain.
When the rice transplanter reaches the edge of the field, it first Kise M, Noguchi N, Ishii K, Terao H. 2002. The development of the
stops transplanting and lifts the transplanting attachment, then autonomous tractor with steering controller applied by opti-
the HST lever is set to the forward position. While the rice mal control. Proceedings of Automation Technology for Off-
transplanter is turning and the yaw angle is less than 160 de- Road Equipment. Chicago, Ill. (USA): ASAE. p 367-373.
grees, only the yaw angle is obtained. Under these conditions, Matsuo Y, Yamamoto S, Yukumoto O. 2002. Development of tilling
robot and operation software. Proceedings of Automation
the steering angle is kept at 40 degrees and the brake is ap-
Technology for Off-Road Equipment. Chicago, Ill. (USA):
plied on one side. If the yaw angle exceeds 160 degrees, the
ASAE. p 184-189.
rice transplanter is moved as close as possible to the next de- Nagasaka Y, Otani R, Shigeta K, Taniwaki K. 2000. A study about
sired path. Then it moves backward. an automated rice transplanter with GPS and FOG. ASAE
Paper No.00-1045. Milwaukee, Wis. (USA): ASAE.
Results and discussion Noguchi N, Kise M, Ishii K, Terao H. 2002. Field automation using
robot tractor. Proceedings of Automation Technology for Off-
The experiment was conducted 4 days after puddling. The field Road Equipment. Chicago, Ill. (USA): ASAE. p 239-245.
size is a 20 m × 100 m square. When the operating program O’Connor M, Bell T, Elkaim G, Parkinson B. 1996. Automatic steer-
started, the transplanter moved into the paddy field and reached ing of farm vehicles using GPS. Paper presented at the 3rd
the starting point. Then it started operation and went forward International Conference on Precision Agriculture. Madison,
Wis. (USA): ASA, CSSA, and SSSA.
and backward in the field 4 times. Finally, it planted the seed-
Dry seeding is the most common direct-sowing method used a high level of skill for its operation, especially for calculating
in Japanese rice paddy fields. In 1996, before this study be- the appropriate doses of chemical powder and water. This pro-
gan, only 7,000 ha (< 0.5%) of the rice fields in Japan relied cess required constant observation to control the coating effi-
on direct-sowing methods, and dry seeding accounted for ap- ciency after adding rice seeds to the inclined-rotator drum.
proximately 70% of the total. However, considering the cli- High concentrations of chemical dust in the working environ-
matic conditions in Japan and the working system available ment also caused a significant problem and a potential health
for rice production, we believe that wet seeding could consid- risk. The drill seeder was found to have poor accuracy for the
erably improve rice productivity. Because wet seeding relies control of seeding depth affected by the hardness of the soil
on cultivating methods similar to rice transplanting, however, surface. Several mechanical problems also occurred related to
it is generally considered to be more expensive than dry seed- the stability of the furrow-opener and the covering plate.
ing, in terms of both labor and machinery requirements.
The major challenge related to implementing wet seed-
Development of a precision-drilling method
ing in Japan has been crop stabilization and minimizing the
for flooded paddy fields
occurrence of lodging. Most rice varieties that are currently
grown in Japan have been selectively bred for eating quality, Following the Uruguay Round agreement in 1994, it was de-
and, as a result, few varieties are specifically suited for direct- cided that there should be a gradual reduction in the tariff placed
sowing methods and are also resistant to low temperatures and on imported rice. This agreement encouraged the production
lodging. However, several studies have been carried out in- of low-cost rice that could be produced with minimum labor.
vestigating direct-sowing methods using several different rice Research carried out in our laboratory led to the development
varieties. Japanese rice farmers have expressed a desire to pro- of a direct-sowing machine, with financial assistance from an
duce high-quality rice (Koshihikari, etc.) that can attain the “Urgent Development of Agricultural Machinery and Its Com-
highest market prices. As a result, it is not broadcast seeding mercial Enterprise” project aimed at facilitating improvements
that places seeds on the soil surface but the sowing method in agricultural machinery. This research led to the develop-
that buries seeds in the soil absolutely that could be the best ment of an automatic coating machine for rice paddy seeds,
option. and a precision drill seeder.
Agitator
Cover
Inclined-rotator
drum
Screw feeder
O The machine is contained within a shatterproof cover the preparation work such as dewatering, preparation of chemi-
and was designed to be resistant to chemical dust cals, drying of coated seed, and packing of the coated product
particles. needed to be carried out either before or after coating using
O A vent on the side of the rotator drum enables the the old model. However, using the newly developed coating
easy removal of the coated seeds. machine, all of these tasks could be carried out simultaneously.
Overview of the coating machine. The coating machine,
as shown in Figure 1, is composed of an inclined-rotator drum, The precision drill seeder in flooded paddy fields
a screw-type chemical feeder, a constant-pressure water nozzle Design. The depth at which the seeds were sown was known
with associated pump system, and a program-sequence con- to be affected by the hardness of the soil surface in the field.
troller. Research was therefore carried out to improve the seeding
Combining a large-diameter agitator and a screw-type characteristics of the traditional drill seeder.
feeder, a stable chemical feeding performance was achieved. O The drill seeder mounts on a paddy vehicle had a
Several research experiments were carried out to test different working width of eight rows or more.
input patterns of water and coating chemicals. Based on the O A float is used to improve the accuracy of placing
results of these experiments, a prototype machine was pro- seeds at depth. A function was also provided to accu-
duced. This prototype was equipped with an automatic sequenc- rately control the angle of the covering plate.
ing system to control the input of water and chemicals into the O Seeding depth is adjusted depending on soil charac-
inclined-rotator drum. Several performance tests were then teristics and the rice variety being planted.
carried out. O The machine adds fertilizer to the planted seeds.
Performance of the coating machine. The efficiency of Overview of the seeding machine. The seeder, as shown
the newly developed coating machine was compared with that in Figure 2, consists of an 8-row seeder mounted on a paddy
of other more traditional models. Efficiency was evaluated vehicle. The seeder consists of a seeding part and a fertilizing
based on the output of coated seeds, the mass of one coated part. This seeding machine was designed to improve seeding
seed, the hardness of the coating layer, and germination. The accuracy. As a high-performance fertilizing part is already in
results of these tests showed that, in most respects, the newly commercial use, commercial items were employed. The basic
developed coating machine performed similarly to the earlier components of the furrow-opener and the covering plate are
models. Although differences were few between the models also commercially available, as is the traditional drill seeder.
for coating time, the newly developed coating machine had To overcome some of the problems associated with varia-
double the capacity and was therefore much faster. Much of tions in soil-surface hardness, it was decided that the seeder
should be equipped with a soil-surface hardness sensor. Sev- cision drill seeder also contributes to crop emergence and es-
eral experiments were carried out to investigate soil-surface tablishment by decreasing the number of surface seedlings and
hardness in relation to covering-plate angle. The results of these floating seedlings, and decreasing lodging. Moreover, these
experiments showed that control of the covering-plate angle newly developed machines have working properties similar to
with a sensor could be used to overcome some of the soil- those of rice transplanters and will undoubtedly lead to more
covering problems associated with differences in soil-surface efficient rice fields because of the application of direct-sow-
hardness. The drill seeder was found to be more accurate for ing techniques in flooded paddy fields.
seeding depth than other traditional drill seeders. A float was The automatic coating machine has been marketed since
also installed to increase the area of ground-surface contact, 1999 by HATSUTA Industrial Co., Ltd. and Yanmar Agricul-
which significantly improved efficiency. The precision drill tural Equipment Co., Ltd. A precision drill seeder for use in
seeder was therefore provided with a mechanism to control flooded paddy fields has been available and marketed since
the covering-plate angle, a sensor for soil-surface hardness, 2000 by Iseki & Co., Ltd.; Kubota Corporation; Mitsubishi
and a large float with a two-step adjustable furrow-opener. Agricultural Machinery Co., Ltd.; and Yanmar Agricultural
Performance of the seeding machine. Field tests using Equipment Co., Ltd. The use of these machines has increased
the newly developed seeder were conducted at 14 sites in Ja- dramatically in recent years. Some 250 coating machines and
pan. The results of these tests showed that the seeder was much 300 drill seeders are now being used. As a result of the in-
more efficient than traditional seeding machines and led to a creased use of these machines, the amount of land that can be
33% decrease in the number of surface seedlings and a 40% used for direct sowing in flooded paddy fields will likely in-
decrease in the number of lodged and floating seedlings. In crease to three times the area that was available in 1998 (3,600
addition, the working capacity was improved by 30% as a re- ha).
sult of the wider working width and the higher operation speed
compared with those of other traditional seeders.
Notes
Commercialization and diffusion Authors’ address: IAM, Bio-oriented Technology Research Advance-
Since the development of an improved automatic coating ma- ment Institution, e-mail: Y_nishimura@affrc.go.jp.
chine, its use has increased dramatically. This machine, which
assists with the direct sowing of rice, was found to improve
crop emergence and establishment. The development of a pre-
In Japan, direct-seeding methods of rice are now being intro- rice seeds are spread over the rods, reciprocating movement is
duced to lower rice production costs. However, in the north- given to the rods and seeds, and the rods are cut between cut-
east area, direct seeding did not spread easily. Under chilly ting boards after the seeds are attached to the rod surface. The
spring weather conditions, seed establishment in direct seed- pellets are coated in Calper Fine Granule (an oxygen-generat-
ing was unstable in wet-seeded fields. Also, growth of broad- ing product) using a coating machine. The round-shaped pel-
cast rice is difficult to control. Moreover, good-tasting variet- let of about 10 mm in diameter contains 5–7 seeds, made by
ies, such as Akitakomachi (Oryza sativa L.), suffer from lodg- the clay-rod cutting method. The working efficiency of the
ing. Recently, to solve these problems, a precise direct-seed- granulation instrument is about 60,000 pellets h–1 (0.3 ha h–1).
ing system was developed that stabilizes seedling establish- This work needs about 1–2 people. The seeds used for granu-
ment and lodging resistance in wet-seeded fields. We devel- lation of multiple seed pellets are pregerminated to about 30%
oped this new rice direct-seeding system with multiple seed in pigeon-breasted conditions. The coated multiple seed pel-
pellets as one type of hill-seeding cultivation. This paper dis- lets are put in a sealed plastic bag, and have been stored for
cusses the granulation method of multiple seed pellets, and about 1 month in refrigeration conditions (Ogiwara et al 2003).
paddy field cultivation tests using multiple seed pellets in wet- Before seeding work, 48 h of warm processing (25 °C) of
seeded rice. multiple seed pellets were effective for bud stabilization un-
der low-air-temperature conditions. Moreover, we have made
a small and simple experimental granulation instrument of the
Developing the granulation method of multiple seed pellets
clay-rod cutting system for general marketing by joint research
Developing a granulation method with a private corporation. This granulation instrument was
of a clay-rod cutting system sharply miniaturized by simplifying the clay feeding device.
We developed a granulation method of a clay-rod cutting sys- However, the clay feeding unit needs to be improved for con-
tem as a high-precision new granulation method (Fig. 1, Togashi tinuous operation. At present, no granulation instrument for
et al 2004b). The clay-rod cutting method follows these steps: multiple seed pellets is on the market.
clay rods are injected from nozzles by a compressing pump,
Sorting device
Ditch board conveyor The clay rod is cut and rolled to the ditch
Clay-rod cut device board conveyor and seeds are rounded
Cloth belt for conveyance
Fig. 1. The machine for manufacturing multiple seed pellets by the clay-rod cutting method (Togashi et al 2004a).
Seeding method Fertilizer application Yield Panicle no. Culm length Lodging index Pushing resistance
(g N m–2) (g m–2) (m–2) (cm) (0–4) (g × cm)
aMultiple-seed-pellet seeding, hill-seeding, and transplanting were used with about 15-cm interrow spacing and 30-cm row spacing. Drilling was done with about 30-cm row
spacing. Hill-seeding was done every five grains within 6 cm of the direction of a rider and 4-cm width. Seeding depth was about 10 mm. The multiple seed pellets of a clay-
rod cutting system were used for the test. Large-quantity fertilization added 2 g N m–2 using LP30 (coated urea-N). This was adjusted at 4 wk after seeding to seedling
establishment at 22 shares m–2, 5 seedlings per share, and 111 seedlings m–2. Data are means of 3 sections. Yield is the weight of brown rice graded by grain size (>1.85
mm in thickness) at 15% moisture. Pushing resistance was measured at 20-cm height at 18–20 d after heading. * indicates a significant difference as opposed to multiple-
seed-pellet seeding at 0.05 according to the Dunnet method (Morita et al 2004).
Developing the granulation method of multiple seed pellets was also large compared with single-
using clay-powder coated seeds. The granulation method needs to be improved
To granulate more simply, we developed an alternative pro- for the broadcasting method.
duction method for multiple seed pellets. With this method,
seeds and clay-powder were rounded together between a wet
Field experiments
cellulose spongy sheet and a sponge-plastic mat on which hemi-
spherical hollows were arranged. Seeds and clay-powder were In precision cultivation tests in which the number of seedlings
made to condense and a spherical pellet was fabricated (Sekiya to be established was arranged, the resistance to lodging of
et al 2004b). The granulation instrument made as an experi- multiple seed pellets was improved compared with broadcast
ment can be continuously granulating seed pellets by moving seeding or drilling (Table 1, Morita et al 2004). In a large num-
conveyor belts with a spongy sheet and a sponge plastic mat. ber of fertilization conditions, the difference between multiple-
Clay-powder for ceramic art is mixed with an oxygen-generat- seed-pellet seedling cultivation and drilling cultivation was
ing product (Calper) used for a binder by a weight ratio of 1:1. remarkable. For labor-saving fertilizing work, we experimen-
This granulation method is as efficient as the clay-rod cutting tally produced a sidedressing fertilizer for a planter and car-
method. The pellet was about 9 mm in diameter and contained ried out cultivation tests by a single application of the fertil-
4 seeds (CV 49%) on average. This method is expected to be izer (Sekiya et al 2004a). A sidedressing fertilizer application
less costly than the clay-rod cutting method. showed that sufficient yield was obtained with a single fertil-
izer application by this machine.
In 6 years of field cultivating, results of multiple-seed-
Developing the seeding method of multiple seed pellets pellet seeding in Ohta test fields in Akita Prefecture indicated
We developed a precision hill-seeding cultivation method us- a stable brown rice harvest of 4.58–6.84 t ha–1 and less lodg-
ing multiple seed pellets. Multiple seed pellets are sown by a ing using Akitakomachi (Sekiya et al 2004b). Since the amount
planter, which is an improved commercial inclined-belt-type of fertilization is decreasing now, there were small lodging
planter attached to a 4WD tractor or a transplanter (Togashi et resistance differences in seeding multiple seed pellets and drill-
al 2002). The inclined-belt-type planter drops multiple seed ing, and there were also small yield differences. When culm
pellets on puddled soil with about 15-cm interrow spacing and length was comparable in multiple-seed-pellet seeding, lodg-
30-cm row spacing. The multiple seed pellets were precisely ing was mitigated compared with drilling.
hill-seeded at about 5-mm depth. Planting efficiency of an eight- As the farmers introduced the multiple-seed-pellet di-
row planter is about 0.35 ha h–1. After seeding, water is drained rect-seeding system, labor expenses, seedling materials, and
for 1 to 2 wk depending on soil conditions. Then, water is fertilizer expenses could be reduced. However, the entire ma-
refilled and the usual cultural management is practiced. This terials expense had only a small difference in multiple-seed-
planter can be used for soybean and wheat seeding, and seed- pellet seeding and transplanting with the increase in the ex-
ing accuracy is good. pense of Calper or additional herbicide. The merit of the mul-
Multiple seed pellets could be broadcast by a duster or a tiple-seed-pellet seeding system for farmers was that it was as
radio-controlled helicopter (Togashi et al 2004a). However, easy to manage as transplanting. On the other hand, farmers
the broadcasting method sometimes made a block of seeds and seldom felt a difference in resistance to lodging between the
cracked pellets. In direct seeding by a duster, the unevenness multiple-seed-pellet seeding system and the drilling system,
Wrap-up of Session 7
This session covered specific concerns related to the improve- that rice has good potential to reduce tetralemma as an environ-
ment of efficiency through innovations in mechanization for boost- mentally friendly fuel.
ing rice production. Some 90% of the total rice production is W.P. Park reviewed the current status of agricultural mecha-
produced from Asian countries. Invited speakers came from Ko- nization, which progressed rapidly in paddy rice production. As a
rea, China, and the Philippines. result, Korean agricultural policy has contributed to self-sufficiency
Professor N. Ito discussed the important role of rice as a in food and improvement of welfare in rural society. But, Korean
key resource to save our planet. He said that the world faces farmers have some problems to be solved because of the small
various global issues, or tetralemma. The rice world requires a farming scale and aging society. To solve these problems, Korean
sufficient food supply, particularly including rice, because the world rice farming needs new strategic studies of advanced technology
population is increasing markedly. On the other hand, global- development to lower production costs, enhance the quality of
scale dependency on fossil fuel must be minimized to prevent products, and use precision agriculture and farm robots.
global warming. These problems can be solved only by improving Professor Y. Li analyzed the development of rice mechani-
rice production; hence, mechanization plays a vital role in this zation, and noted that mechanization of planting and harvesting
endeavor as a force multiplier compensating for the labor short- is relatively less since farm mechanization does not go well with
age. In addition, N. Ito introduced new, unique technologies: a agronomy because of vast territory with complex agricultural situ-
combine husker, combine harvester equipped with a pivot turn ations. China will concentrate on developing rice planting and
mechanism, brown rice dryer, and others. Finally, he stressed drying mechanization techniques, especially for new equipment,
Rice is among the top food grains in the world and is the most high-amylose, intermediate-GT varieties of South Asia had
important one as human food. It is grown widely on all conti- about 45%; all other varieties showed about 55% solubility.
nents and under all agroclimatic conditions. This wide adapta- The calculated insoluble amylose content of rice on this basis
tion has led to the evolution of thousands of varieties of rice (total minus soluble amylose) showed a striking correlation
having diverse cooking, eating, and product-making charac- with rice texture, particularly in the high-amylose group, which
teristics. Rice palatable to one ethnic group or suitable for one was thereby divided into three distinct subgroups showing dis-
product may not be so for others. One striking example of this tinct texture after cooking (Bhattacharya et al 1978).
diversity is the unique geographical shift in the quality of rice Taking the above secondary factors into consideration,
in Asia, the home of the world’s rice. Rice grown in northern it appeared by the mid-1980s that the Brabender paste break-
and eastern Asia cooks soft and sticky, while that in southern down, gel consistency, and insoluble amylose, along with amy-
Asia is completely different, for it cooks firm and fluffy, with lose content of rice, could characterize its end-use quality very
that in Southeast Asia coming in between. well (Sowbhagya et al 1987). But the chemical basis was un-
clear. Evidently, these approaches had reached the limit of their
potential and a new approach was necessary.
Research on divergence in rice quality
This wide varietal divergence in rice quality has attracted the
A new approach
attention of researchers for close to a century. In fact, this sub-
ject has been one of the most intensively investigated topics in The third phase started at this juncture in the mid-1980s with a
cereal chemistry. The work can be divided into three phases. complete shift in the prevailing paradigm. Chemists’ under-
The first phase, up to the end of the 1940s, was merely a pe- standing of the structure of starch, especially amylopectin, has
riod of initiation. The work, carried out mostly in India, was been all along going through a metamorphosis. The chain struc-
sporadic and confined largely to studying the absorption of ture of amylopectin came under serious reexamination from
water by rice during cooking. the end of the 1970s with the use of various starch-hydrolyz-
The second phase, the period of data accumulation, ing enzymes and of gel permeation chromatography (GPC).
started soon after World War II, when there was a sudden spurt Biliaderis et al (1981), for example, showed that the chain pro-
of research activity in the subject in the United States, Japan, file of legume starches varied significantly from one species
Germany, Italy, Spain, and France. This was followed shortly to another. Again, Professors Fuwa and Hizukuri in Japan
afterward by intensive and sustained research at the newly set showed significant variation in amylopectin chain structure
up International Rice Research Institute (IRRI) in the Philip- among various starches. Chinnaswamy in my laboratory at
pines as well as at the Central Food Technological Research CFTRI at this time observed with surprise that the GPC-sepa-
Institute (CFTRI) in Mysore, India. rated amylopectins of rice starch gave a varying shade of blue
This work led to a mass of data. The main conclusion color with iodine. What is more, the absorption maxima (λmax)
was that the amylose starch content of rice was the single larg- of these iodine complexes showed excellent correlation with
est determinant of its end-use quality. The greater the amy- the texture of the corresponding rice varieties after cooking
lose, the firmer was the rice texture when cooked and vice- (Fig. 1). These results showed that (1) unlike as so far be-
versa. However, it was also realized that amylose, though nec- lieved, amylopectin also bound iodine, (2) this binding was
essary, was not sufficient; other secondary factors were in- apparently caused by the presence of some very long chains in
volved. In fact, pinpointing these secondary factors was the it, (3) rice varieties therefore apparently differed in their amy-
main brunt of research during this phase. Many factors were lopectin chain structure, and (4) the apparent amylose content
suggested: hydration power, starch-iodine blue value, gelati- of starch as determined by reaction with iodine was always an
nization temperature (GT), alkali digestion score, Brabender overestimation, part of the value being contributed by amy-
viscograph pattern, mobility of an alkaline rice-flour gel (gel lopectin. Hizukuri in collaboration with IRRI scientists shortly
consistency), as well as hot-water-insoluble amylose. This last thereafter showed that the chain profile of rice amylopectin
item has an interesting history. While investigating the starch- indeed differed from variety to variety, with high-amylose rice
iodine blue value of rice varieties, my school at CFTRI real- having more very long chains than low-amylose rice, with in-
ized that it was nothing but the water-soluble amylose of rice. termediate-amylose rice coming in between (Hizukuri et at
Researchers then observed that the solubility of amylose of 1989). Radhika Reddy et al (1993) later confirmed these find-
rice varieties fell into three distinct groups. The Taiwanese ings and also showed that high insoluble-amylose rice amy-
high-amylose, low-GT, dwarfing-gene donors and many of their lopectin not only had more long chains but also that these chains
progenies had an amylose solubility of only about 35%; many seemed to be largely located externally in the molecule and
Jaya I
(I)
580
S317 III
560 (III)
540
Intan
(V) V
520
0 5 10 15 20 T65
Insoluble amylose in rice (% d.b.) (VII) VII
Cooked-rice hardness
Fig. 1. Variation in λ max of rice amylopectin- Fig. 2. Photomicrographs of starch granules in 12% rice-flour pastes
iodine complex. Thirteen rice varieties having of four varieties (types I, III, V, VII, having from very high to low
different insoluble amylose contents were frac- insoluble amylose). Flours were pasted by heating up to 95 °C. (A)
tionated over Sepharose 2B and the gel-ex- Light microscopy, (B) scanning electron microscopy. For each, left
cluded fraction 1 (amylopectin) was reacted column: just pasted (0 min); right column: maintained at 95 °C for
with iodine. The absorption maxima of the blue another 60 min. After 60 min, the granules are hardly affected in
complexes were read. d.b. = dry basis. (Adapted type I but are completely degraded in type VII. (Reproduced from
from Chinnaswamy and Bhattacharya 1986.) Sandhya Rani and Bhattacharya 1995.)
vice-versa. Work in many laboratories worldwide (Korea, UK, with the GT (Umemoto et at 2002). However, the implication
Japan, Taiwan-China) subsequently confirmed these findings. of this finding in terms of rice quality is uncertain. From all
accounts, rice GT does not seem to have a significant influ-
ence on its cooking-eating characteristics.
The influence of amylopectin A few other factors may affect rice-use value in a small
Clearly, the texture of cooked rice was closely correlated with way. Protein content of rice has long been thought to influence
the relative abundance of long chains in its amylopectin. The rice-use quality. This was first proposed by Spanish workers
more long chains, the firmer was the rice when cooked and in the 1950s and 1960s. Subsequently, Japanese and other
vice-versa. The rationale was explained by rheological and workers supported this contention. Recently, Ohtsubo and his
microscopical studies in my laboratory. It was shown that the group at the Food Research Institute in Japan have shown that
presence of abundant long chains in the amylopectin led, ap- the surface hardness of cooked rice especially is related to pro-
parently by mutual interaction, to strong and resilient starch tein (Okadome et al 1999). American scientists have shown
granules that resisted swelling and breakdown, whereas their that rice protein may have some relationship to its paste prop-
absence led to weak and fragile granules that broke down eas- erties.
ily during cooking (Fig. 2). In addition, the content of free amino acids and sugars
Thus, after more than 50 years of intensive research, the in rice has been suggested by Japanese scientists to affect the
texture of cooked rice and other rice end-use quality has been palatability and flavor of rice after cooking.
attributed primarily to the relative abundance of long chains
in its amylopectin. Interestingly, other work has shown that
References
the bulk of rice starch consists of amylopectin, true amylose
being not more than about 10% in it (Ramesh et al 1999). Bhattacharya KR, Sowbhagya CM, Indudhara Swamy YM. 1978.
Importance of insoluble amylose as a determinant of rice qual-
ity. J. Sci. Food Agric. 29:359-364.
Other factors Biliaderis CG, Grant DR, Vose JR. 1981. Structural characterization
of legume starches. 1. Studies on amylose, amylopectin and
Very recent work, primarily in Japan, has shown that the GT
beta-limit dextrins. Cereal Chem. 58:496-502.
of rice starch, too, is correlated with the chain structure of Chinnaswamy R, Bhattacharya KR. 1986. Characteristics of gel-chro-
amylopectin. The proportion of very short chains among the matographic fractions of starch in relation to rice and expanded
overall short chains in amylopectin is negatively correlated rice-product qualities. Staerke 38:51-57.
Rice, the most important crop in Thailand, was expected to this property is amylose content. The low-amylose type (less
alleviate the economic problem of jobless people in 1998. than 20% amylose) always has a soft texture and is sticky when
During 1998-2003, rice production obtained from main and cooked. The intermediate-amylose type (21–25%) produces
second crops was 22–26 million tons annually. About 6.8–7.3 rather soft cooked rice, whereas the high-amylose type has a
million t of milled rice were exported in 1999-2003. Approxi- hard texture. To obtain an optimum cooked-rice quality, high-
mately 50% of the commodity was used domestically, and the amylose milled rice requires more cooking water than that
surplus exported to foreign markets. In 2003, Thai rice ex- having a lower amylose content. To cook rice, the cooking
ports made up about 27% of the world market. Among this time depends on the gelatinization temperature of the starch
proportion of exports, good-quality rice represented 46%. This (gel. temp.) in milled grain. On the basis of gel temperature,
included Thai aromatic rice or Jasmine rice. Thus, Thailand’s rice varieties can be classified into three groups: low, interme-
rice program is projected to enhance production, sufficient for diate, and high. Thai rice has a great variation in cooked-rice
self-consumption and for export. Rice with good grain quality (Table 1). We have glutinous rice and various types of
must aim to meet the requirements of both domestic and for- nonglutinous rice, providing an advantage to satisfy the pref-
eign consumers. erences of different groups of rice consumers. Aroma is a spe-
cial characteristic and is considered to be a special quality.
Quality testing is therefore primarily a matter of determining
Improving quality whether the rice is suitable for a particular use and whether it
Rice quality is considered to have two general aspects: meets specific requirements of cleanliness and purity.
1. Milling, cooking, and processing qualities, which re- Furthermore, milling yields are based on the amount of
fer to suitability of the grain for a particular end-use. whole milled kernels and all sizes of broken kernels obtain-
2. Physical quality, which means cleanliness, soundness, able from a unit of rough rice. Many factors influence milling
and being free from foreign materials. Usually, physi- yield; the miller must also judge the quality of rice for milling
cal quality and milling, cooking, and processing quali- on the basis of variety, moisture content, the presence of red
ties are interrelated in that certain market grades of rice, peckiness (parboiled kernels with obviously black or dark
rice are more suitable than others for consumer prod- brown area) and other kernel damage, foreign seeds and other
ucts. material, chalky kernels, and mustiness or other undesirable
The quality of Thai rice has long been well known in odors (Webb and Stermer 1972).
international markets. The quality of rice is mainly defined as The world rice market classifies the commodity into six
long and slender. Breeding selection has incorporated grain basic types: (1) predominantly indica, high-quality, long-grain
quality into varietal improvement with other agronomic char- raw milled rice; (2) predominantly indica, medium-quality,
acters. Rice with long, slender, and translucent grain is prefer- long-grain raw milled rice; (3) japonica short- or medium-grain,
able. Different varieties produce a different quality of cooked raw milled rice; (4) parboiled rice; (5) aromatic (fragrant) rice;
rice; glutinous rice is very sticky, and steaming is always ap- and (6) waxy (glutinous) rice. Among these types, five have a
plied to prepare the cooked product. Nonglutinous rice may significant role in the international market. These groups can
be soft and slightly sticky before cooking, and rather soft or also be subdivided depending on consumer preferences. Each
hard in its texture when cooked. The main factor influencing type has different markets and, within each type, there are spe-
aAromatic rice.
moisture, heat, insects, or something else.
O Foreign matter: including underdeveloped grain, im-
mature kernels, excess kernels, other seeds, rice husk,
cial quality demands for specific uses. In general, consumers and bran.
view rice in an uncooked or unprocessed state and make judg- O Black kernels (whole kernels), partly black kernels
ments on its quality. To meet market standards, major rice ex- (more than 25% of the kernel becomes black), and
porters and importers have established official grades to iden- pecked kernels (less than 25% become pecked) are
tify the relative quality of the commodity. limited in the parboiled rice standard.
O Paddy: glutinous rice in nonglutinous rice or
nonglutinous rice in glutinous rice is also specified in
Thai rice standards all types of rice standards.
As the major rice exporter, Thailand established Thai Rice
Standards B.E. 2540. These standards were revised in 1997 Milling degree
through a notification by the Ministry of Commerce. These This is divided into four levels: extra well milled, well milled,
rice standards described the minimum specifications of rice in reasonably well milled, and ordinarily well milled.
each type and grade for domestic and international trade. They
include white rice (milled rice), cargo rice (Loonzain rice, Moisture content
brown rice, or husked rice), white glutinous rice (milled waxy The moisture content of rice of all types and grades is speci-
rice), and parboiled rice. The judgment is based on physical fied as not exceeding 14%.
examination of the grain considering the following categories.
The adulteration problem of Thai Jasmine rice
Rice classification
Classes of rice refer to the unbroken kernels of various length, The popularity of Thai aromatic or Jasmine rice increased re-
which are a mixture in accordance to the proportion. They can markably from 1992 to 1997. The annual export quantities of
be divided into four groups as follows: aromatic rice were as high as 1.06–1.45 million t. It repre-
Class 1 long grain more than 7.0 mm sented 20–27% of the total quantity exported. The commodity
Class 2 long grain 6.6–7.0 mm was produced from varieties KDML 105 and RD15. Produc-
Class 3 long grain 6.2–6.6 mm tion was only 3–4 million t of paddy, which was not enough
Short grain Less than 6.2 mm for market demand. This brought about an increase in the paddy
High-quality rice is composed of class 1 long grain in a price. The grains of these two varieties result in soft and sticky
greater proportion than that of the lower grades. The percent- cooked rice that is too sticky for some consumers. The high
age of this grain becomes less as the grade becomes lower. price and this unfavorable cooked-rice texture brought about
a blending of grain of aromatic rice with that of other varieties
having a similar grain appearance. In general, these practices
Starch is the major component of rice grain, and its structure To investigate the structure of the amorphous material,
contributes significantly to the “quality” of cooked rice. Starch flour (250 mg) from 16 waxy varieties of rice was mixed with
structure can be described on several levels—ranging from 25 g MilliQ water. The slurries were gelatinized with a Rapid
the organization of compound starch granules (Whistler and Visco Analyser (RVA, Newport Scientific) (AACC method 61-
BeMiller 1997) within the cells of the endosperm down to the 02). The mixture was centrifuged immediately (10,000 g, 10
connectivity of the chains in the individual molecules. min). The supernatant contained the hot water soluble (HWS)
Amylopectin and amylose are accumulated in the devel- starch. The HWS material was debranched, labeled with APTS,
oping grains of all cereals. Amylopectin is deposited in semi- and analyzed as described above.
crystalline growth rings, in alternating zones of crystalline and
amorphous regions (French 1984). It is unclear whether amy-
lose is interspersed within the amorphous lamella of the semi-
Results
crystalline rings, or if it is deposited within the amorphous Figure 1A shows the logarithmic plots for the range of spe-
growth rings. The amorphous growth rings are known to con- cies. All the higher plants studied exhibit similar logarithmic
tain carbohydrates linked by both alpha 1-4 and alpha 1-6 bonds MWD plots: an initial increasing nonlinear region (which we
(Atkin et al 1999), but the exact contents of those growth rings define as Region 1), followed by two distinct decreasing lin-
are unknown. ear regions (Regions 2 and 4) separated by a nonlinear region
The semicrystalline growth rings of starch are understood (Region 3). Seven metrics of the plots have been defined (Fig.
much more than are the amorphous growth rings. For amy- 1B): four slopes and three intersects of the slopes. The slopes
lopectin, crystalline lamellae alternate with amorphous lamel- and position of the regions on the logarithmic plots are inde-
lae. The synthesis of amylopectin occurs through the agency pendent of the normalization of the original data. Comparing
of a suite of enzymes, many of which have been identified, the various species, the slopes of the linear regions (2 and 4)
cloned, and sequenced, showing that the biosynthesis of starch clearly vary, but all the intersects are similar (Fig. 1A).
has been strongly conserved throughout the plant kingdom. Figure 2 shows the logarithmic plot of the debranched
However, the exact function of each enzyme is not so well HWS starch and full flour for one representative from the col-
understood. Attempts to understand function have been mostly lection of 16 varieties of waxy rice. All intersects differ and
undertaken by comparing the molecular weight distribution the slope of both Regions 2 and 4 differs. The logarithmic plots
(MWD) of the amylopectin chains of wild types with that of of the full flour and HWS starch for all 16 varieties showed
mutants or antisense lines that lack the activity of one enzyme the same differences, but these are not shown here.
in the pathway of starch synthesis (Myers et al 2000). Here we
introduce a new technique for understanding the full MWD of
debranched amylopectin, using methods that have been devel-
Discussion
oped for quantitatively interpreting MWDs in synthetic poly- Starch synthesis is a multistep process, containing both ran-
mers (Clay and Gilbert 1995). In brief, if the only events oc- dom and nonrandom elements (Thompson 2000, Caldwell and
curring during synthesis are random chain growth and stop- Matheson 2003). Plotting the MWDs of the debranched starch
page, then the number of chains of degree of polymerization as logarithmic plots (Fig. 1) allows the identification of these
N, P(N), has the form ln[P(N)] ∝N, where the constant of pro- random and nonrandom elements. Monotonically decreasing
portionality is the ratio of the stoppage and growth rates. This functions (Regions 2 and 4) indicate random elements (Clay
starting point suggests that mechanistic inferences can be made and Gilbert 1995). The monotonically increasing nonlinear
from a plot of lnP(N) against N. Castro et al (submitted) have regions of the lnP curves (Regions 1 and 3) are of particular
described the validation of the application of this model to interest—if all chains began at DP 1 and grew until a random
starch, and mathematical proof of the model. stopping event, then P(N), and hence lnP(N), must necessarily
be a decreasing function. Two decreasing functions added to-
gether still produce a decreasing function. This implies that
Methods some nonrandom event must be involved in generating Re-
Starch was isolated from rice, wheat, barley, maize, and po- gions 1 and 3. Region 1 chains are between DP 6 and 10 and
tato by established techniques. The starches were gelatinized, are probably remnants of branching enzyme events (chain stop-
debranched, and labeled with 8-amino-1,3,6-pyrenetrisulfonic page) rather than originating from controlled synthesis.
acid (APTS) and analyzed by capillary eElectrophoresis Regions 2 and 4 appear to define two populations of
(O’Shea et al 1998). chains. The longer chains (Region 4) must originate from the
shorter chains (Region 2), as there is no evidence for a mecha-
1 16
15
0
14
1
13
2
12
3
11
4
10
0 20 40 60 80
5 N
0 20 40 60 80 100
N Fig. 2. lnP(N) plots of the debranched hot water soluble (HWS)
and full flour starch of chains from DP 6 to DP <100 showing
differences between the slopes and intersects of each.
InP(N)
16 n1
B
talline polymer with a lamella repeat distance of 9 nm (Imberty
et al 1991). Hence, although three linked glucose units span
s1 s2
about 1.1 nm in the unit cell of crystalline starch (Imberty et al
1991), giving a maximum DP of 27, part of the 9 nm is com-
12 s3 n3 posed of noncrystalline spacer. Thus, it is reasonable to infer
s4 that the maximum size of a chain spanning a lamella repeat is
n2
DP 33.
Chains that escape crystallization-induced stoppage to
elongate beyond 33 units will then protrude from the newly
8
formed lamellae. These randomly elongated chains must then
0 10 20 30 40 50 60 go on to seed the next cluster, indicating that positioning of
N subsequent clusters in the molecule (tertiary architecture) is
random. These chains must add a minimum number of units to
Fig. 1. A: lnP(N) as a function of degree of polymerization N for
higher plants: potato ( ), maize (×), barley ( ), wheat (+), and support new clusters—most likely the size of a spacer, consis-
rice ( ). Dotted lines are the offset spacing. B: the description of tent with the minimum size of Region 4 chains being DP 47.
the seven metrics—4 slopes (s1–4) and 3 intersects (n1–3). Thus, this suggests that the architecture of the amylopectin
molecule is random within the capacity of the suite of enzymes
present; this has implications for the rheological properties of
nism for branches to start at DP 50 (the start of Region 4). the starch.
Region 3 must then relate to the mechanism that distinguishes It follows from this discussion that Region 3 chains do
chains in Region 2 from those in Region 4. Region 3 then indi- not belong in the semicrystalline growth ring. Hizukuri et al
cates that there is some nonrandom “event” preventing Re- (1989) have suggested the presence of a third starch in the hot
gion 2 chains from becoming Region 4 chains. Here we pro- water soluble fraction of rice. Since chains of Region 3 do not
pose that crystallization of amylopectin to form the lamellae fit theoretically into the semicrystalline model, the amorphous
in the starch granule is the key chain-stoppage event that de- material in a starch granule, which is reported to become
termines the upper limit of Region 2 chains. soluble in hot water (Chinnaswamy and Bhattacharya 1986),
Chains in Region 2 do not span more than a single lamella was then investigated. Figure 2 shows a representative example
repeat (one amorphous and one crystalline lamella). This is of the logarithmic plots of the starch in the HWS fraction and
consistent with the idea that starch is a side-chain liquid-crys- full flour of a waxy rice variety. All slopes and intersects dif-
References Notes
Atkin NJ, Cheng SL, Abeysekera RM, Robards AW. 1999.
Authors’ addresses: Melissa Fitzgerald and Rosa Paula Cuevas, In-
Localisation of amylose and amylopectin in starch granules
ternational Rice Research Institute, DAPO Box 7777, Metro
using enzyme-gold labeling. Starch-Starke 51:163-172.
Manila, Philippines; Melissa Fitzgerald and Jeffrey Castro,
Chinnaswamy R, Bhattacharya KR.1986. Characteristics of gel-chro-
CRC for Sustainable Rice Production, PMB, Yanco 2703,
matographic fractions of starch in relation to rice and expanded
NSW, Australia; Melissa Fitzgerald, NSW Department of Pri-
rice-product qualities. Starch-Starke 38:51-57.
mary Industries, PMB, Yanco 2703, NSW, Australia; Jeffrey
Clay PA, Gilbert RG. 1995. Molecular weight distributions in free-
Castro and Robert Gilbert, Key Centre for Polymer Colloids,
radical polymerizations. 1. Model development and implica-
School of Chemistry, The University of Sydney 2006, NSW,
tions for data interpretation. Macromolecules 28:552-569.
Australia, e-mail: m.fitzgerald@cgiar.org.
Indica rice varieties are popular worldwide. Cooked indica rice Materials and methods
is hard but not sticky. However, Japanese people are fond of
japonica cooked rice because of its moderate elasticity and Two cultivars of Oryza sativa L. indica (Khao Dawk Mali 105
stickiness. It is generally said that the amylose content con- and unknown high-amylose Thai rice) obtained from Thailand
tributes to the hardness and stickiness of cooked rice. Japonica in 2001 were used. They were stored at 4 °C. Japonica rice,
rice has lower amylose content and moderate elasticity and Oryza sativa L. japonica (Nipponbare), grown in Shiga Pre-
stickiness. fecture, Japan (2001), was used. Khao Dawk Mali 105 and
This study was carried out to confirm the textural differ- Nipponbare were 90% milled in a polishing machine before
ence between cooked indica and japonica rice and to know the use.
reasons for the difference. Among indica varieties, Khao Dawk Milled rice grain was ground with an electric miller, and
Mali is known to contain less amylose; thus, Khao Dawk Mali the resulting flour was passed through a 100-mesh sieve. Starch
and a higher amylose Thai rice and japonica Nipponbare were was isolated from milled rice flour by using cold 0.2% NaOH
used to compare their texture and some physicochemical prop- according to the alkali method of Yamamoto et al (1973) with
erties. modifications. The precipitate was washed with distilled wa-
ter until the pH decreased to 7.0. After rinsing with ethanol,
the starch was air-dried and passed through a 100-mesh sieve.
Milled rice grains were washed three times with distilled wa-
How cooked Textural property Nipponbare Khao Dawk Mali High-amylose rice
Cooked with 1.4 Hardness (kgf) 2.19 ± 0.22 a 2.45 ± 0.15 a 3.74 ± 0.38 b
times water Stickiness (kgf) 0.34 ± 0.04 a 0.10 ± 0.01 b 0.03 ± 0.04 c
Cooked with 1.9 Hardness (kgf) 1.98 ± 0.19 a 2.18 ± 0.30 a 3.17 ± 0.16 b
times water Stickiness (kgf) 0.41 ± 0.03 a 0.14 ± 0.06 b 0.04 ± 0.01 c
aMean ±SD of ten measurements. For letters a, b, and c, the same letter in the same row is not significantly
different (P<0.05).
ter. The washed rice was then soaked in 20 °C distilled water tained. The reducing-sugar content was measured by a modi-
for 1 hour. The weight ratio of rice grain to soaking water was fied Park Johnson method (Hizukuri et al 1981). The poly-
1:1.4 or 1:1.9. After soaking, the rice grains were cooked by merization level of each fraction was obtained by the follow-
an electric rice cooker (Toshiba Ltd. RC-74). Some 10 g of ing expression:
cooked rice, which was taken from the middle of the rice
cooker, was added to 50 mL of cold (5 °C) distilled water, Polymerization level = amount of total sugar/amount
stirred for 5 min, and the mixture strained by an ordinary of reducing sugar
strainer. The surface layer extract was then obtained. One part
of this surface layer extract was lyophilized and used for FT- To measure the absorption spectrum, 50 mL of 1% io-
IR spectrometry and gel permeation chromatography. The other dine-10% potassium iodine solution was added to each 5-mL
mixture was centrifuged at 16,000x g at 5 °C for 10 min and fraction. The maximum absorption spectrum was obtained with
the supernatant was obtained to determine sugar content. The a spectrophotometer (UV-2200 and Shimadzu). The fractions
precipitate was added to ethanol and acetone, was air-dried, eluted were divided into four (Fr. I: λ max ≥620 nm; interme-
and the weight measured. Cooked rice grain was also lyo- diate Fr: 620 nm > λ max ≥600 nm; Fr. II: 600 nm > λ max
philized for FT-IR and gel permeation chromatography. ≥525 nm; Fr. III: 525 nm > λ max).
The moisture, protein, fat, and ash of rice flour were
determined by AACC methods (2000). Protein content was
Results and discussion
determined by Lowry’s method (Lowry et al 1951) using bo-
vine serum albumin as a standard. Apparent amylose content Proximate analysis
was determined by the method of Juliano (1971). Total sugar The moisture contents of the three rice flours ranged from
and reducing-sugar contents of supernatant of extracts from 13.4% to 14.5% and crude protein from 5.34% to 7.62%. Ap-
the surface of cooked rice were determined by the phenol sul- parent amylose contents were almost the same (16%) in Khao
furic acid method and Somogyi-Nelson’s method. Glucose Dawk Mali and Nipponbare, which were different from high-
solution was used as a standard material for the measurement amylose Thai rice (27%). This confirmed that indica variety
of total sugar and reducing-sugar contents. Khao Dawk Mali and japonica variety Nipponbare have the
The β-amylase-prulanase method (Matsunaga and same level of amylose.
Kainuma 1981) was applied.
Hardness and stickiness were determined by the method Textural properties
of Okabe (1977) using three grains of cooked rice with a Table 1 shows the hardness and stickiness of rice cooked with
Texturometer (General Foods GTX-2). 1.4 times water or 1.9 times water. The hardness of Nipponbare
IR spectra were obtained using a Fourier Transform (FT) and Khao Dawk Mali was the same, but the stickiness was
Infrared Spectrometer (Japan Bunkou Ltd. JASCO FT/IR-420) different. And those of high-amylose Thai rice were quite dif-
equipped with a Diamond ATR accessory. Wave length was ferent from those of other rice varieties. Thus, the textural dif-
from 400 to 4,000 cm–1. Analysis was done in a room with 20 ference of cooked rice could not be explained by only the dif-
°C temperature. ference in amylose content.
Based on the method of Itoh et al (1994), starch from
raw rice, lyophilized cooked rice, and lyophilized extracts from Degree of gelatinization
the surface layer of cooked rice was analyzed. Before analy- The degree of gelatinization was almost 100% in Nipponbare
sis, these samples were under a debranching process by iso- and Khao Dawk Mali, both cooked with 1.4 and 1.9 times
amylase (Inouchi et al 1987). A column (2.5 id × 100 cm) water. In contrast, the value of the high-amylose Thai rice was
packed with Toyopearl HW-50SF (Tosoh Co., Ltd., Japan) 84% even when cooked with 1.9 times water. The degree of
eluted with 0.05 N NaOH–0.2% NaCl-NaN3 solution, at the gelatinization is not enough to explain the textural difference
flow rate of 0.44 mL min–1, was used. of cooked rice.
Carbohydrates in each fraction (5 mL) were estimated
by the phenol-H2SO4 method and the elution pattern was ob-
10 10 10
30 30 30
8 8 8
6 20 6 20 6 20
4 4 4
10 10 10
2 2 2
0 0 0 0 0 0
25 30 35 40 45 50 55 25 30 35 40 45 50 55 25 30 35 40 45 50 55
Fraction no. Fraction no. Fraction no.
Fig. 1. GFC patterns of extracts from three kinds of cooked rice surface after isoamylase treatment.
Weight of precipitates and sugar and protein and Khao Dawk Mali were similar and that of high-amylose
contents of supernatant of cooked-rice extracts Thai rice was different. The long-chain fraction was largest in
The weight of precipitates was the largest in Nipponbare, fol- high-amylose Thai rice.
lowed by Khao Dawk Mali, and the smallest in high-amylose When three kinds of rice were cooked, Nipponbare de-
Thai rice. The same order in the contents of total sugar, reduc- creased its amylose fraction, but its short-chain (about 15 dp)
ing sugar, and protein of supernatant was obtained. It seemed amylopectin fraction increased slightly. The pattern of the
that the large amount of sugar in the supernatant was an im- Nipponbare extracts was almost the same as with cooked rice.
portant factor, especially in the stickiness of cooked rice. In Khao Dawk Mali, the amylose fraction decreased to the
same level as Nipponbare when cooked. However, this amy-
FT-IR spectrometry lose fraction increased in the extracts. In high-amylose rice,
FT-IR is a useful tool to determine intermolecular interaction. the amylose fraction decreased largely after cooking, whereas
This time, the difference in intermolecular interaction of the the short-chain fraction increased. And, in the extracts, the long-
three kinds of rice starches (flour with 50% w/w of water) was chain fraction (amylose fraction) increased even much more
measured. A wave length of 1,000–1,100 cm–1 is known to than raw starch.
denote the interaction between C and O and C and C, and 1,020 From these findings, we assumed that the distribution of
cm–1 is also known to show the absorption of sugar. Thus, we amylose and amylopectin was almost equal in Nipponbare, but
used a wave length of around 1,000 cm–1. not in the indica varieties.
Three FT-IR patterns for each three kinds of rice were
obtained, that is, raw rice starch, cooked rice, and extracts of
Conclusions
the surface of cooked rice. Moreover, a difference in IR spec-
trum of the cooked rice and the extracts was found. From the It is suggested that the stickiness of cooked rice is related to
results, the difference in IR spectrum of this approximate wave the sugar content in the surface layer of cooked rice extracts.
length was negatively larger in Nipponbare, followed by Khao These extracts also contained a lower amount of amylose frac-
Dawk Mali. On the other hand, high-amylose Thai rice was tion and larger amount of short-chain amylopectin fraction.
positively larger. This negatively large difference in cooked These components are once eluted from the rice during cook-
rice and the extracts means that a larger amount of sugar was ing and finally absorbed into the surface layer, thus giving
in the surface layer than inside the cooked rice. This may be stickiness to rice.
one reason to explain the textural difference of the cooked
rice.
References
Gel permeation chromatography Hizukuri S, Takeda Y, Yasuda M, Suzuki A. 1981. Multi branched
Figure 1 shows the gel permeation patterns of rice starch after nature of amylose and the action of debranching enzymes.
iso-amylase treatment. For raw rice, the patterns of Nipponbare Carbohydr. Res. 147:342-347.
The World Health Organization estimates that 3.5 billion people nificantly different at the end of the 9-month trial for the
in the developing world are affected by iron-deficiency ane- nonanemic women who consumed the high-iron rice instead
mia. Given the need for a sustainable solution to reduce this of the control after controlling for baseline values, amount of
number, biofortified rice (variety IR68144) was used to test rice consumed, and convent. The greatest changes in ferritin
the feasibility of high-iron rice as a source of dietary iron. We and body iron were observed in the women who consumed the
conducted a double-blind longitudinal (9 months) interven- most biofortified rice and the calculation of the amount of iron
tion study with religious sisters from 10 convents in and around transferred from the diet to iron stores is consistent with the
Manila, Philippines. Some 317 religious sisters were randomly observed changes in body iron. Therefore, it is possible to
assigned to consume either an iron-enhanced (high-iron) rice conclude that there was a significantly positive effect of con-
or a local variety of rice (control). Rice preparation and distri- suming biofortified rice on the iron status of women with an
bution were standardized. The sisters consumed their assigned iron-poor diet.
rice at every meal and food intakes were weighed on 3 days
every 2 weeks for each woman. Blood samples were collected
at baseline, mid-point, and end-point and used to determine
Notes
serum ferritin, the transferring receptor, and hemoglobin, Authors’ addresses: J. Haas, J. Beard, and L.M. Kolb, Cornell Uni-
among other indicators of iron status. The efficacy trial indi- versity, USA; G. Gregorio, International Rice Research Insti-
cated that the high-iron rice provided an additional 1.41 mg of tute, Los Baños, Philippines; A. Del Mundo and A. Felix,
iron per day, representing a 17% increase in dietary iron in the University of the Philippines Los Baños, Los Baños, Philip-
pines.
diets of these women. Ferritin and body iron levels were sig-
Rice (Oryza sativa) is widely consumed in the world, and the the antioxidative and radical-scavenging properties of rice
most common type (>85%) has a white pericarp. Other types because of their potential to provide health protection against
have a colored pericarp, and the most common are green, black, reactive oxygen species and free radicals, which have been
and red. The black and red varieties are planted mainly in South implicated in more than 100 diseases (Halliwell 1992).
Asia and other countries, such as Italy, Greece, and the United On the other hand, there have also been some reports
States (Simmons and Williams 1997). Rice with a colored peri- concerning the antioxidative compounds found in rice. Gamma-
carp has long been consumed in Japan and China and is con- oryzanol, which is a mixture of ferulate esters of triterpene
sidered to be a healthy food. We are especially interested in alcohol, is well known as an antioxidant in rice bran (Xu et al
ND
ND
ND
ND
ND
0 0
1.5 2.0
Black rice Black rice
1.5
1.0
1.0
0.5
0.5
ND
ND
ND
ND
ND
ND
0 0
1.5 2.0
Red rice Red rice
1.5
1.0
1.0
0.5
0.5
ND
ND
ND
0 0
Deionized water
Deionized water
Ethyl acetate
Ethyl acetate
Diethylether
Diethylether
Acetone
Methanol
Acetone
Methanol
n-hexane
n-hexane
Fig. 1. DPPH• (A) and t-BuOO• (B) scavenging activities of extracts prepared from rice with red, black,
and white pericarp by using solvents with various polarities. ND = not detected.
proanthocyanidin in the extract from red rice was the Radical-scavenging components in black rice
procyanidin type, that is, (+)-catechin and/or (–)-epicatechin The black rice extracts were prepared from eight varieties pur-
derivatives. The gel permeation chromatography of the acety- chased at the local market in the Kyushu and Okinawa regions.
lated proanthocyanidin revealed that the average molecular DPPH• scavenging activity of the extracts ranged from 3.1 to
weight is about 5,000, in a range of about 500 to 18,000 after 10.8 µmol-Trolox equiv. mL–1 of the extract. The radical-scav-
proanthocyanidin was purified from the acetone extract by enging activity increased with the increase in anthocyanin con-
using a Sephadex LH-20 column chromatography. tent, and the correlation coefficient was 0.987. By RP-HPLC
We also examined DPPH• scavenging activity, polyphe- of the extracts, two major peaks were observed in all eight
nol content, and proanthocyanidin content in various rice cul- black rice varieties, and these peaks were identified as peonidin-
tivars with red pericarp. The DPPH• scavenging activity var- 3-glucoside (Pn-3-Glc) in addition to Cy-3-Glc. A slight com-
ied according to the cultivars (1.4 to 46.7 µmol-Trolox equiv. positional difference in anthocyanin was observed, however.
g–1), and Tanegashimazairai, a traditional wild type of red rice The sum of Cy-3-Glc and Pn-3-Glc constituted more than 85%
in Japan, showed very high radical-scavenging activity. The of the total anthocyanin. Therefore, it was revealed that the
DPPH• scavenging activity was highly correlated to polyphe- dominant radical-scavengers in black rice were Cy-3-Glc and
nol content and proanthocyanidin content, with correlation Pn-3-Glc.
coefficients of 0.982 and 0.944, respectively (Fig. 2). This re- We have already found that two anthocyanins (Cy-3-Glc
sult indicated that the major radical-scavenging component in and Pn-3-Glc) were detected in rat plasma after oral adminis-
red rice was proanthocyanidin. tration of the anthocyanin-rich concentrate prepared from black
rice (BR-ANT) (Oki et al 2002c). Furthermore, the BR-ANT
40
30
20
10
R = 0.982 R = 0.994
0
0 10 20 30 40 50 0 10 20 30 40 50
Polyphenol content Proanthocyanidin content
(mmol-gallic acid equiv. g1) (mmol-(+)-catechin equiv. g1)
Fig. 2. Correlation of polyphenol content (A) and proanthocyanidin content (B) with DPPH• scaveng-
ing activity in various rice cultivars with red pericarp.
suppressed hemolysis and the oxidation of LDL (Oki et al Oki T, Masuda M, Kobayashi M, Nishiba Y, Furuta S, Suda I, Sato
2002c). The BR-ANT also inhibited lipid peroxidation induced T. 2002a. Polymeric procyanidins as radical scavenging com-
by four different radical generation systems (Oki et al 2002c). ponents in red-hulled rice. J. Agric. Food Chem. 50:7524-
These results suggested that dietary consumption of black rice 7529.
Oki T, Masuda M, Furuta S, Nishiba Y, Suda I. 2002b. Involvement
might prevent many diseases related to lipid peroxidation
of anthocyanins and other phenolic compounds in radical-scav-
caused by free radicals.
enging activity of purple-fleshed sweet potato cultivars. J. Food
On the other hand, there is currently no satisfactory ex- Sci. 67:1752-1756.
planation for how the polymeric proanthocyanidin is absorbed Oki T, Masuda M, Ikuo S. 2002c. Antioxidative activity of black
into the bloodstream. Ling et al (2001) have reported that, when rice commercially available in Kyushu and Okinawa region.
a hypercholesterolemia rabbit had red rice as a part of its diet, In: Noguchi T, editor. The 56th Annual Scientific Meeting,
the total antioxidative properties of the serum and liver in- 19-21 July 2002. Hokkaido University (Japan): The Japanese
creased. They suggested that constituents with antioxidative Society of Nutrition and Food Science. p 230. (In Japanese.)
activity could be responsible for the increased antioxidative Osawa T. 1999. Protective role of rice polyphenols in oxidative stress.
properties in the body; however, they did not identify the anti- Anticancer Res. 19:3645-3650. (In Japanese.)
oxidant compounds in red rice. This report suggests that, after Simmons D, Williams R. 1997. Dietary practices among Europeans
and different South Asian groups in Coventry. Br. J. Nutr.
proanthocyanidin in red rice is consumed, it could be absorbed
78:5-14.
by the body directly and/or indirectly (i.e., structural change
Xu Z, Hua N, Godber JS. 2001. Antioxidant activity of tocopherols,
or degradation), which would elevate the antioxidative capac- tocotrienols, and γ-oryzanol components from rice bran against
ity. Now, we are investigating the physiological functions of cholesterol oxidation accelerated by 2,2′-azobis (2-
red rice in vivo and elucidating the mechanisms at work in methylpropionamidine) dihydrochloride. J. Agric. Food Chem.
these functions. 49:2077-2081.
References Notes
Halliwell B. 1992. The role of oxygen radicals in human disease, Authors’ addresses: Department of Crop and Food Science, National
with particular reference to the vascular system. Haemostasis Agricultural Research Center for Kyushu Okinawa Region
23(suppl. 1):118-126. (KONARC), National Agriculture and Bio-oriented Research
Ling WH, Cheng QX, Wang JMT. 2001. Red and black rice de- Organization (NARO), 2421 Suya, Nishigoshi, Kikuchi,
crease artherosclerotic plaque formation and increase antioxi- Kumamoto 861-1192, Japan, e-mail: tomooki@affrc.go.jp.
dant status in rabbits. J. Nutr. 131:1421-1426.
Nawa Y, Ohtani T. 1992. Property of pigments in rice hulls of vari-
ous colors. Food Industry 11:28-33. (In Japanese.)
Amylose content of endosperm starch is an important charac- Table 1. Effect of temperature on the accumulation of amylose dur-
teristic of rice in determining eating and cooking quality. Amy- ing seed development.a
lose content is genetically controlled by a major gene, the Waxy Strain 20 oC 26 oC Difference (20–26 oC)
(Wx) gene (Nelson and Pan 1995). In addition, amylose con-
tent is also affected by several modifying genes and environ- Snow Pearl 18.3 ± 0.8 5.2 ± 0.5 13.1
mental factors such as temperature. The Wx gene encodes the SP14 24.7 ± 0.3 11.7 ± 0.6 13.0
granule-bound starch synthase that is responsible for amylose Koshihikari 25.1 ± 0.8 17.0 ± 0.3 8.1
synthesis in the endosperm. Rice strains having the Wxb gene aSnow Pearl, SP14, and Koshihikari seeds matured at 20 or 26 oC. The amylose
contain 15–20% amylose, whereas rice strains having the Wxa content of the samples was analyzed based on an iodine colorimetric assay (Juliano
gene contain 20–25% amylose (Sano 1984). The du loci modify 1971) with slight modifications (Suzuki et al 2002). Data (amylose content, %) are
mean values ± SE from 5 trials. Twenty mg of white rice powder was used for each
amylose content by controlling expression of the Wxb gene trial.
and mutations at the du loci cause a reduction in amylose con-
tent (Okuno et al 1983, Hirano 1993). Although the amylose-
extender (ae) mutant shows increasing levels of apparent amy-
lose content, this increase is caused by the altered amylopec- seeded, and amylose content of SP14 (M3 seeds) was com-
tin structure, and the ae mutation is found in the gene for starch- pared with that of Snow Pearl and Koshihikari using a tem-
branching enzyme IIb (Nishi et al 2001). While there are many perature-controlled growth chamber. The amylose content in
mutants in starch composition, there is no report about a gene mature seeds of SP14 was about 12% at 26 oC and 25% at 20
whose mutation gives rise to an increase in amylose content in oC (Table 1), indicating that amylose content of SP14 is re-
the endosperm. In this paper, we describe the isolation and sponding to maturation temperatures and is 6–7% higher than
characterization of a rice mutant with enhanced amylose con- that of Snow Pearl. Furthermore, to clarify whether amylose
tent in the endosperm that does not affect the amylopectin struc- content in SP14 was increased by the ae locus or not, the amy-
ture. lopectin chain-length distribution in rice endosperm was ana-
To screen mutants with enhanced amylose content, we lyzed by high-performance anion exchange chromatography
used a low-amylose rice variety, Snow Pearl (Higashi et al with a pulsed amperometric detector, as previously described
1999). In this variety, amylose content was very low (about (Suzuki et al 2003). The chain-length distributions of amy-
5%) and the endosperm appeared chalky when seeds matured lopectin were almost the same between the two strains ma-
at a normal temperature (26 oC) (Table 1). On the other hand, tured at both temperatures (data not shown). In addition, while
when seeds matured at a cool temperature (20 oC), amylose endosperm appearance of EM10 (ae mutant) was floury, that
content was about 18% and the endosperm appeared of SP14 as well as that of Snow Pearl and Koshihikari was not
semitranslucent. Because amylose content in rice seed influ- floury (i.e., wild type = normal). These results indicated that
ences the degree of transparency (Suzuki et al 2002), a mutant the mutation in SP14 would not be the ae mutation, but a new
with enhanced amylose content in (semitranslucent) seeds mutation that enhances amylose content in rice endosperm.
matured at a normal temperature could be easily distinguished It should also be noted whether SP14 is derived from
from the original variety with low-amylose (chalky) seeds. Snow Pearl or not. As Snow Pearl has a single nucleotide poly-
Dry seeds of Snow Pearl rice were treated with 1 mM morphism (SNP) on the 3rd exon of the Wx gene (Nippon
sodium azide in 0.1 M sodium phosphate buffer (pH 3.0) for 6 Suisan Kaisha, Ltd. 2000), a part of the SP14 Wx gene (344
h at room temperature, following washing with tap water. The base pairs) containing the SNP was compared with that of Snow
first screening used 7,200 M1 plants that matured at an aver- Pearl and Koshihikari. SNP of SP14 was identical with that of
age temperature of 24–26 oC. Although the panicles of almost Snow Pearl, but not with that of Koshihikari, indicating that
all of these plants contained exclusively chalky seeds, one of SP14 is a mutant derived from Snow Pearl. Furthermore, the
these plants contained a mixture of chalky and semitranslucent nucleotide sequences of the wx locus (6,800 base pairs: Wx
seeds, 41 and 6 seeds, respectively. Next, semitranslucent M2 gene and promoter region) of SP14 and Snow Pearl were iden-
seeds were chosen as a candidate (designated as SP14) for a tical. These results suggested that the mutation from Wxb to
mutant with enhanced amylose content. The M2 seeds were Wxa in SP14 would not occur and that the mutation in SP14
would not be on the wx locus.
Seed number
Chi-square value
Strain Amylose content Total for 15:1 ratio Probability
< 8% ≥ 8%
aSnow Pearl, SP14, and F2 rice seeds (SP14/Snow Pearl) matured at 26 oC. The amylose content of white rice
seeds (15–25 mg) was analyzed based on an iodine colorimetric assay (Juliano 1971) with slight modifica-
tions (Suzuki et al 2002).
Next, the mode of inheritance of the character was ana- Nishi A, Nakamura Y, Tanaka N, Satoh H. 2001. Biochemical and
lyzed by crossing SP14 with Snow Pearl. The seeds of F1, Snow genetic analysis of the effects of amylose-extender mutation
Pearl, and SP14 plants matured at 26 oC. As segregation of in rice endosperm. Plant Physiol. 127:459-472.
amylose content in F2 seeds derived from the cross fit the ratio Okuno K, Fuwa H, Yano M. 1983. A new mutant gene lowering
amylose content in endosperm starch of rice, Oryza sativa L.
of 15 Snow Pearl-type seeds:1 SP14-type seed, the mutation
Jpn. J. Breed. 33:387-394.
is controlled by two recessive genes (Table 2). Thus, we have
Sano Y. 1984. Differential regulation of waxy gene expression in
isolated a new rice mutant (SP14) in which amylose content in rice endosperm. Theor. Appl. Genet. 68:467-473.
the endosperm was enhanced compared with the original vari- Suzuki Y, Sano Y, Hirano H-Y. 2002. Isolation and characterization
ety. This mutant should be useful in breeding programs de- of a rice mutant insensitive to cool temperatures on amylose
signed to produce rice of a high quality and in understanding synthesis. Euphytica 123:95-100.
genetic and molecular mechanisms for amylose synthesis. Suzuki Y, Sano Y, Ishikawa T, Sasaki T, Matsukura U, Hirano H-Y.
2003. Starch characteristics of the rice mutant du2-2 Taichung
65 highly affected by environmental temperatures during seed
References development. Cereal Chem. 80:184-187.
Higashi T, Saitoh S, Takita T, Yamaguchi M, Sunohara Y, Yokogami
N, Ikeda R, Tamura Y, Oyamada Z, Kowata H, Inoue M,
Notes
Matsumoto S, Kataoka T. 1999. Breeding of a new rice culti-
var with low amylose contents, “Snow pearl”. Bull. Tohoku Authors’ addresses: Yasuhiro Suzuki, Ushio Matsukura, Noriaki
Natl. Agric. Exp. Stn. 95:1-12. Aoki, and Hiroyuki Sato, National Institute of Crop Science,
Hirano H-Y. 1993. Genetic variation and gene regulation at the wx Kannondai, Tsukuba, Ibaraki 305-8518; Hiro-Yuki Hirano,
locus in rice. Gamma-Field Symp. 24:63-79. Graduate School of Sciences, The University of Tokyo, Yayoi,
Juliano BO. 1971. A simplified assay for milled-rice amylose. Ce- Bunkyo-ku, Tokyo 113-8657; Yoshio Sano, Graduate School
real Sci. Today 16:334-340. of Agriculture, Hokkaido University, Kita 9, Nishi 9, Kita-ku,
Nelson O, Pan D. 1995. Starch synthesis in maize endosperms. Annu. Sapporo 060-8589; Kazuo Ise, Japan International Research
Rev. Plant Physiol. Plant Mol. Biol. 46:475-496. Center for Agricultural Sciences, Ohwashi, Tsukuba, Ibaraki
Nippon Suisan Kaisha, Ltd. 2000. Method to distinguish a specific 305-8686, Japan, e-mail: suzuyasu@affrc.go.jp.
low-amylose-rice variety. Japan patent exhibition 2000-
201679.
Pasting behavior was widely used to estimate the physical visco analyzer (RVA) profiles in nonglutinous and glutinous
properties of various cereals or starch-based products, includ- rice varieties. Especially, the inhibition of amylase activity
ing rice products. The components of rice, such as amylose using copper sulfate or mercury chloride in Japanese gluti-
content, protein, and enzymes, especially amylase activity, nous rice varieties increased peak viscosity and could express
were known as the factors affecting pasting behavior. There the variety difference of glutinous rice varieties (Kikuchi 1998).
were several reports on the effect of amylase activity on rapid However, the effects of amylase inhibitor in Thai high-amy-
G1 with CuSO4
2,400
1,800
G4 with CuSO4
G3 with CuSO4
1,200
G4 with water
G1 with water
0
0 3 6 9 12
Time (min)
Fig. 1. RVA profile of rice flour (with and without water-soluble fraction) in water and in CuSO4 solution.
ter-soluble fraction (G3 and G4) showed no peak increasing Elution of water-soluble polysaccharide
even with an amylase inhibitor. Therefore, it is suspected that from broken rice
the water-soluble fraction contains the substance, which af- Even less physical force was applied on the rice kernel for the
fects pasting properties and is sensitive for inherent amylase preparation of broken rice. Water-soluble polysaccharide was
activity. eluted in GPC analysis. Therefore, water-soluble polysaccha-
ride was not categorized as “damaged starch,” which might be
Effect of amylase activities on the water-soluble produced during longer grinding (Chen et al 1988, 1999).
fraction Obtained DSC thermograms of water-soluble polysac-
Figure 2 shows an example of a GPC profile of the water- charide, water-insoluble powder, and original rice flour of
soluble fraction. The water-soluble fraction contained an ex- Hakuchoumochi indicated that the water-soluble fraction never
pected peak as water-soluble polysaccharide, when inherent gelatinized, and the thermal properties of rice may be repre-
amylase activities were inhibited by copper sulfate or heating. sented by water-insoluble powder. Water-insoluble powder is
The peak disappeared when suspensions were allowed to mix considered to represent “wet-milled” rice flour. Therefore, the
for 30 min by RVA. However, when an amylase inhibitor such wet-milling process might be a process to remove water-soluble
as copper sulfate was added after 10 min, the peak remained, polysaccharide, which might affect gelatinization (Chen et al
which corresponds to the water-soluble polysaccharide. These 1988, 1999).
results suggested that water-soluble polysaccharide was ex-
tremely sensitive to inherent rice amylase. Since the heating Constituent analysis of water-soluble
procedure inhibited inherent amylase activities, high-tempera- polysaccharides
ture drying might control the digestion of water-soluble All samples showed constituents as glucose only. Thus, water-
polysaccharide through denaturation of inherent amylase. soluble polysaccharide in rice was expected as a homoglycan
built from glucose.
With these properties, water-soluble polysaccharide was
easily digested by inherent rice amylase, then released oligosac-
charides, which show lower viscosity in RVA and “stickiness”
V0
Water
CuSO4, 10 min
CuSO4
Heated
on the surface of products. Therefore, the polysaccharide might Matsukura U, Suzuki Y, Iwai Y, Monma M, Kaneko N. 2004. Com-
work like “cement” in rice gel, which is considered as “con- parison of α-amylase activity in rice grains of nonglutinous
crete.” and glutinous varieties and effect of α-amylase activity on
These results might be a clue to understanding pasting rice pasting property. Nippon Shokuhin Kagaku Kougaku
of rice, and this is needed to understand rice gel, not only starch, Kaishi. 51(10):554-558. (In Japanese.)
Tungtrakul P, Yoshihashi T. 2001. Elucidation of amylase effect on
but also polysaccharides. However, further studies are needed
amylopectin chain distribution in various rice flour. Report
to elucidate more detailed properties of polysaccharide.
submitted to JIRCAS. 8 p.
References
Chen JJ, Lu S, Lii CY. 1988. Thermal characteristics and micro- Notes
structure changes in waxy rice flours by different milling
Authors’ addresses: Tadashi Yoshihashi and Eizo Tatsumi, Japan
methods. Food Sci. 25:314-330.
International Research Center for Agricultural Sciences
Chen JJ, Lu S, Lii CY. 1999. Effects of milling on the physicochemi-
(JIRCAS), Tsukuba 305-8686, Japan; Vipa Surojanametakul,
cal characteristics of waxy rice in Taiwan. Cereal Chem.
Patcharee Tungtrakul, and Warunee Varanyanond, Institute of
76:796-799.
Food Research and Product Development (IFRPD), Kasetsart
Horino T, Ondona A, Kusumoto K, Soegusa T, Mori Y. 1988. Novel
University, 50 Phahonyothin Road, Chatuchak, Bangkok
water-soluble polysaccharide from rice. Research highlights
10900, Thailand, e-mail: tadashi@jircas.affrc.go.jp.
on food research, NFRI. (In Japanese.)
Kikuchi H. 1998. Studies on chemical components in rice kernel for
breeding. Hokkaido Pref. Agric. Res Report. 1998:1-68.
Rice (Oryza sativa L.) and rice-based products derived from Starch properties
rice grain and rice flour include parboiled rice; quick-cooking
rice and ready-to-eat convenience foods; rice flours; rice starch; Thermal properties of starch of importance to product quality
cakes and puddings; baked breads and crackers; breakfast ce- include gelatinization temperature (GT) and degree of gelati-
reals and expanded rice products; extrusion-cooked and puffed- nization, amylose-lipid complex I (melting at temperatures
rice snacks; noodles, paper, and pasta; baby/weaning foods; <100 °C) and II (melting at temperatures >100°C), and staled-
fermented foods and beverages; pet foods; and bran products amylopectin melting (45–60 °C). Final GT represents the tem-
(Juliano 2003). The rice ingredient is preferably aged to have perature when at least 90% of the starch granules have swol-
predictable, stable functional properties, and freshly milled and len in hot water with loss of crystallinity. Residual ungelatinized
well-milled for minimal initial fat rancidity and long shelf life. starch may affect product properties. Parboiling studies showed
The rice ingredient in most of these products has the same that low-GT samples tended to have amylose-lipid complex I,
apparent amylose content (AC) type as that of the preferred and intermediate-high-GT rice tended to have amylose-lipid
table rice in the country. Milled rice AC (on a dry-weight ba- complex II in parboiled rice (Biliaderis et al 1993). The pres-
sis) is classified into waxy (0–2% dry basis) and nonwaxy: ence of amylose-lipid complex II in pressure-parboiled rice
low (12–20%), intermediate (20–25%), and high (25–33%). (>100 °C) may explain the hard texture of its cooked rice since
Waxy and low-AC rice are used for alcoholic beverages be- amylose-lipid complex II does not melt during rice cooking.
cause of the higher enzymic conversion of starch to glucose Rice noodles and corn starch noodles have a similar tex-
than high-AC rice, and for products requiring starch-gel sta- ture when freshly cooked. However, the cooked rice noodles
bility of wet products. High-amylose rice is preferred for prod- are more stable, even on reheating, and have a better shelf life,
ucts requiring an intact cooked product, resistant to disinte- whereas the corn starch noodles become mushy on reheating
gration, although it may be harder in texture. and readily spoil. Is this because of the greater stability of starch
Traditional rice products in developing countries are gels of high-AC rice relative to nonwaxy corn? Or is it be-
small-scale and poorly packaged, and have a short shelf life. cause of the presence of more protein in milled rice (7%) that
Improvement of value-added traditional rice products, selected is reported to contribute to the stability of rice starch gel (C.R.
based on marketability and profitability, will create jobs in rural Mitchell, California Natural Products, 1994, unpublished
areas and make a country’s products compete globally as im- data)? Corn starch has only 0.7–1.0% protein.
port substitutes or as exports to niche markets. The glycemic index is the relative increase in plasma
glucose within 3 h after a fasted subject ingests 50 g of carbo-
hydrate, with white bread or glucose taken as 100%. It is used
Semiwet-milled rice flour as a guide for diets of noninsulin-dependent diabetes mellitus
The availability of a semiwet (semidry)-milled rice flour of patients. The glycemic index tends to be higher for cooked
specified AC (waxy, intermediate-AC, and high-AC) is cru- waxy and low-AC rice than for cooked intermediate- and high-
cial to the development of the rice products industry in the AC rice (Table 1) (Panlasigui 1989, Foster-Powell and Brand-
Philippines, since a dry-milled flour cannot be used in many Miller 1995, Juliano 2003). Cooking processes that allow
baked and steamed products requiring a fine flour with little amylose staling such as parboiling and noodle extrusion fur-
starch damage. By contrast, wet-milled and semiwet-milled ther reduce the glycemic index of intermediate- to high-AC
rice flours can be used in all rice products and are whiter than rice. Processes that suppress amylose staling such as puffing
dry-milled flour. However, wet-milling is tedious and a source and precooking increase the glycemic index of all rice. How-
of environmental pollution. Yeh (2004) considered semidry-/ ever, varietal differences in glycemic index are shown in
semiwet-milled rice flour as that produced from steeped milled molded puffed brown rice cake.
rice (30% water) with a roller mill, a stamp mill, and a pin Filipino consumers (laborers) prefer rice that is more
mill, without water added during milling. Its starch is reported filling (Unnevehr et al 1992), for example, high-AC PSBRc10.
to be 40% gelatinized from heat generated during milling, Satiety is defined as the state of being fed or gratified to ca-
which may affect its functional properties. It has lower con- pacity. Boiled rice and molded puffed brown rice cake with a
tents of protein, lipids, ash, and reducing sugars than dry-milled higher glycemic index and lower AC tended to have less per-
flour. The development and production of semiwet rice flour ceived satiety and greater subsequent food intake than high-
of uniform properties (specified AC) to replace wet-milled flour AC rice (Holt and Brand Miller 1995, Tetens et al 2003). The
should improve the efficiency of small and medium rice enter- satiety situation for waxy and intermediate-AC rice should be
prises since the tedious wet-milling step can be dispensed with. determined. The glycemic index of intermediate-AC rice tended
aMean ± standard deviation. Numbers of samples in parentheses, if mean of more than one sample. The glycemic index based on
glucose is 0.70 of the glycemic index based on white bread.
Table 2. In vivo resistant starch in cooked rice and rice products of various amylose
(AC) and gelatinization temperature (GT) types as measured in five growing rats (Eggum
et al 1993) and in an ileostomate (Jenkins et al 1987).
aR= rat, H = human. Values followed by the same letter are not significantly different at the 5% level by
Duncan’s multiple range test. Mean ± standard deviation.
to overlap with that of high-AC rice, and waxy and low-AC Other nutritional concerns
rice have a similar glycemic index (Table 1).
The effects of food processing on protein quality, mycotoxin
The resistant starch (staled amylose) in boiled rice is
level, acrylamide content, and level of antioxidants have to be
low at 0–4% in milled rice and tends to increase with AC, GT,
monitored to ensure maintenance of the nutritive value of the
and cooking time (Table 2) (Jenkins et al 1987, Eggum et al
rice product (Juliano 2003). The amino acids lysine and cys-
1993). The total dietary fiber (TDF)/neutral detergent fiber is
teine are sensitive to heat, particularly during toasting of ex-
0.7–2.3% of milled rice and 2.9–4.0% of brown rice. Brown
panded rice. Hydrogen sulfide is produced during extrusion
rice has a lower TDF content than whole-wheat grain, with a
cooking of rice. Alkaline treatment may induce the formation
minimum TDF of 5.6%. The increase in TDF during cooking
of lysinoalanine, a carcinogen. Mycotoxin is a problem with
of nonwaxy rice is due to the formation of resistant starch. The
wet grain, including delayed drying after harvest and parboiled
IR36-based amylose extender mutant with 40% AC can still
rice. It is concentrated in the bran layer.
be cooked in boiling water and was not lower in glycemic in-
Acrylamide, a carcinogen, was analyzed in 2002 because
dex than the high-AC intermediate-GT parent. It contains both
of its presence in high levels in baked and fried cereals and
elongated (B type) and normal polyhedral starch granules. Its
potato, but absence in boiled cereals, including rice (NFA
in vivo resistant starch in rats was surprisingly lower than that
2002). Acrylamide is produced by the Maillard reaction of free
of high-AC intermediate-GT rice (Table 2). Its TDF was also
asparagine with free sugars. Developments in research on high
lower (7.5%) than that reported for high-amylose corn mu-
acrylamide levels in foods should be closely monitored as they
tants (30–40%) that are cooked only by autoclaving (>100 °C).
relate to baked, fried, and toasted rice products.
Recently, a lot of instant foods have been made from rice. The Materials and methods
main categories are aseptic packaged rice and rice cookies. In
this food process, the control of microorganisms, especially Electrolyzed water is produced by electrolyzing tap water with
heat-resistant spores from raw materials to products, is the most the addition of a small quantity of NaCl. Acidic electrolyzed
important point to confirm safety in product flow. If the heat- water (AcEW) created at the anode has been observed to have
resistant spores can be controlled by pretreatment before cook- sterilization effects on microorganisms, and alkaline electro-
ing, excess heat should be omitted to make long shelf-life rice lyzed water (AlEW) created at the cathode has been observed
products, thus improving the quality of the products. Several to have a rinsing effect on organic compounds. In Japan, AcEW
papers have demonstrated the sterilization effects of electro- already was approved as an indirect food additive in 2002.
lyzed water on food ingredients (Koseki et al 2004a,b). The principle of electrolyzed water is shown in Figure 1.
In this paper, we try to confirm the sterilization effect on We used a flow-type electrolyzed water producer (Rox-
rice using electrolyzed water and check the quality of rice dur- 20TA: Hoshizaki Electric Co., Japan). This apparatus gener-
ing pretreatment. ates electrolyzed water by the electrolysis of a dilute (0.1%)
saline solution in an electrolytic cell separated into an anode
and cathode region with a diaphragm. The current passing
4
HCl
NaOH
HClO
~Chemical reaction~
<+ part>:
2OH®H2O+1/2O2+ 3
2e
2Cl®Cl2+2e
+
Cl2 H2 Cl2+H2®HClO+HCl 2
H +
OH (low pH condition)
HClO®H++ClO
(high pH condition)
Cl Na+ <part>: 1
2H++2e®H2
NaCl NaCl
0
A B C D
Fig. 2. Sterilization effects of combination of
Separation electrolyzed water on rice preparation. (A) Con-
membrane trol rice, (B) washing with AlEW (5 min), (C)
Tap water soaking with AcEW (30 min) after B treatment,
with NaCl (D) soaking with D/W (30 min) after C treat-
ment.
Fig. 1. Principle of acidic electrolyzed water. Left: process
flow of apparatus, right: chemical reaction during electro-
lyzing process.
through the electrolysis apparatus and voltage between the elec- Results and discussion
trodes were set at 14A and 18V, respectively. AcEW was pre-
pared within the anode region of the electrolytic cell, and AlEW We examined several tests to prove the influence on the mi-
was prepared within the cathode region. The physicochemical croorganism and the change in the raw rice.
properties of electrolyzed water are as follows: AcEW and AlEW changed the color and pH of rice rap-
O AcEW: pH 2.7, oxidation reduction potential (ORP) idly. But, when the rice was treated with a combination of elec-
1,481 mV, available chlorine concentration 51.5 ppm. trolyzed water (washing with AlEW, soaking with AcEW, and
O AlEW: pH 11.6, ORP –576 mV. soaking with distilled water), pH and color of the treated rice
We used distilled water as a control processing solution. were no different from conventional washing and soaking with
Polished rice (Kinuhikari) was purchased and used as a distilled water.
sample. Prepared Bacillus subtillis PCI219 was used as a heat- AcEW showed a strong effect on microorganism con-
resistant spore sample. trol, even on heat-resistant spores in vitro. However, in the
Some 1 mL of spore solution (109 cfu mL) was added to rice preparation test, AcEW could not kill B. subtillis spores
9 mL of AcEW and mixed. The sample was collected from the on the rice completely. But, after washing by AlEW, AcEW
mixed solution every 1 min and the survival number counted could kill completely. So, finally, we set the optimum method
in each sample after cultivation. of combination of electrolyzed water as follows. To remove
Some 100 g of polished rice were prepared for each rice bran powder and other foreign materials that coat the sur-
preparation test. We devised a preparation process with 3 stages face of the raw rice, at first we washed the rice with AlEW for
(washing: 5 min, soaking 1 ½ min, soaking 2 ½ min). We used 5 min and soaked it with AcEW to sterilize microorganisms
AcEW, AlEW, and distilled water, respectively, as each stage for 30 min. Then, to remove the odor of chlorine and to revert
solution. pH and color of rice were measured by a conven- the pH level, we resoaked the rice with distilled water for 30
tional pH meter and color meter. min. Figure 2 shows the effects against microorganisms using
To confirm the microorganism control effect, we added the combination process of each electrolyzed water.
the B. subtillis spore into the rice and we counted the survival In this experiment, we could confirm the effect of the
number in each sample of the preparation test. combination of electrolyzed water on microorganism control.
This effect can be considered the result of available chlorine
concentration and the rapid change in pH and it raises the sen-
Korea has maintained self-sufficiency in rice production since Achievement of varietal improvement in specialty rice
1975, except during severely cold years. However, self-suffi-
ciency in total cereal food supply, including other cereals, is The target of rice breeding in Korea is to develop high-quality
still limited to about 54% and to about 30% for total consump- and high-yielding rice cultivars suitable for labor-saving, low-
tion, including livestock feeds. Rice products, the staple food cost, and safe grain production under different environmental
in Korea, cover 23% of farmers’ income and 41% of total ag- conditions. The main direction of the country’s rice breeding
ricultural resources, and provide 35% of caloric intake and program was on how to increase yield potential and safe grain
21% of protein intake per capita at present. production and how to improve marketability and palatability.
Historically, the continuous improvement of Tongil-type Also, efforts to develop various specialty rice varieties started
high-yielding rice cultivars greatly contributed to self-suffi- in the mid-1980s to support an increase in rice consumption
ciency in rice production through the so-called Green Revolu- and use through food processing (Choi 2001).
tion during the 1970-1980s. Moreover, the development of Since 1990, 23 specialty rice cultivars, including gluti-
high-yielding and high-quality japonica rice cultivars during nous rice, have been developed. The first specialty rice culti-
the 1980-1990s played a major role both in ensuring self-suf- var was a large-kernel variety, Daeribbyeo 1, developed in
ficiency of rice production and in enhancing the competitive- 1993. It has about 1.7 times heavier 1,000-grain weight of 34.8
ness of rice products under free trade. We believe that we have g than ordinary ones. It showed a good usability for popping
the most advanced techniques in rice breeding and maintain and brewing. In the same year, high-yielding and semidwarf
the highest level of rice yield potential per hectare while pro- indica scented rice, Hyangmibyeo 1, was developed. It has a
ducing the best-quality rice in the world. However, Korea is considerably strong aroma and a good flavor in cooked rice or
still lagging behind in the international competitiveness of rice processed rice foods such as traditional rice cakes and saccha-
goods primarily because the size of rice farm land is so small rified rice beverage. A high-quality japonica scented rice,
and the production cost is still very high (Choi 2001). It should Hyangnambyeo, was developed in 1995. Another high-yield-
also be noted that per capita rice consumption in Korea has ing Tongil-type scented rice, Hyangmibyeo 2, was developed
decreased dramatically from 120 kg in 1990 to 83 kg in 2003. in 1996. Since then, two scented glutinous rice varieties,
The development of high-quality and specialty rice varieties Aranghyangchalbyeo and Seolhyangbyeo, a premium scented
should continue to be an important element in rice research. rice, Mihyangbyeo, and a blackish purple-scented rice,
By developing a variety of rice products other than ordinary Heughyangbyeo, were developed successively (Table 1).
cooked rice, we can increase rice consumption and interna- In 1994, a medium-sized, chalky-kernel rice variety hav-
tional competitiveness. ing suitability for brewing, Yangjobyeo, was developed. An
opaque nonglutinous rice, Seolgaeng, a 9% low-amylose dull
aL/W ratio = length/width ratio of brown rice, YW = yellowish white, BP = blackish purple, RB = reddish brown, HL = high lysine, HF = high fiber.
Source: Choi (2004).
mutant rice, Baegjinju, and a high-fiber and high-amylose rice, ity of cooked noodles might be due to some factors indepen-
Goami 2, were developed through mutation breeding by MNU dent of each other in rice. The potassium and magnesium con-
(N-methyl-N-nitrosourea) treatment on a premium-quality tents of milled rice were negatively related to the gross score
japonica rice variety, Ilpumbyeo. Another semiglutinous rice, of noodle making when mixed with wheat flour half and half
Manmi, which has 12.9% amylose content, was developed by (Choi 2002).
a conventional breeding method using a dull mutant as a par- Batters for making brown rice bread had more volume
ent. expansion, better loaf formation, and more springiness in rice
Some colored rice varieties having reddish brown or bread. The higher the protein content in rice, the more moist
blackish purple pericarp were also developed (Table 1). In milled rice bread resulted. The springiness of rice bread was
2001, a premium-quality rice variety with high lysine content, significantly correlated with high amylose content and hard
Yeongan, was developed. Some specialty rice varieties, such gel consistency. A considerably different tendency of accept-
as blackish purple or reddish brown glutinous rice and giant- ability for rice bread baking among tested rice varieties was
embryo glutinous rice, were bred by mutation or from weedy revealed between brown and polished rice (Kang et al 1997).
rice collections (Choi 2004). There was a large varietal difference in suitability for koji
The cultivation area of specialty rice, including gluti- (malted rice) fermentation by either Aspergillus oryzae or
nous rice, is only about 1.5% of the total rice cultivation area Monascus anka. The nonglutinous rice having opaque en-
since its usability and processing have not been fully devel- dosperm, Seolgaeng, showed better rooting density of myce-
oped. lia and higher saccharogenic power in A. oryzae-fermented
rice, and higher pigment concentration in Monascus anka-fer-
mented rice when compared with dull or ordinary nonglutinous
Use of specialty rice rice. The chalky and large-grain rice showed better suitability
The large-kernel and low-amylose rice varieties revealed both for fermentation and brewing than less chalky and smaller rice
improved popping and brewing characteristics. Their softer (Choi 2002).
gel consistency and lower amylose content were closely re- Colored rice can be used as natural functional pigments.
lated to the higher ratio of popped rice grains and also to the The anthocyanin pigment of blackish purple rice showed sig-
lower bulk density of popped grain in either brown or milled nificantly high antioxidant and antimutagenic reaction (Kang
rice (Choi 2002). 2002). The pigments of colored rice are very useful for tradi-
Of the high-amylose and high-protein rice varieties, some tional colored rice cakes or for brewing colored rice wines.
showed a desirable quality of processed rice noodles or rice The extracted anthocyanin pigment can be used for diverse
bread. The acceptability for noodle formation and eating qual- processed rice foods or for beauty products (Choi 2002).
Rice is one of the most important cereals in the world in addi- of fish using natural fermentation (narezushi). In this case,
tion to wheat and maize. Rice production is more than 500 cooked rice was used as a substrate with salt for Lactobacillus
million tons (as paddy per year) in the world and about 90% of bacteria. Nowadays, nigirizushi was invented as an easily pre-
it is produced and consumed as a major staple food in the pared rice dish by pressing cooked rice balls manually and
densely populated Asian countries. topping them with sliced raw fish in Tokyo in the 1800s.
(Cooked rice becomes tastier by adding salt, sugar, and vin-
egar after cooking.) Soy sauce and grated horseradish (wasabi)
Processing of rice grains in Japan are necessary to enjoy the original taste of sushi.
Rice is processed and used as various kinds of foodstuffs be- Onigiri or omusubi is another traditional rice-based con-
sides direct food use, such as parboiled rice, fermented rice venience food in Japan. These cooked rice balls wrapped with
wine, rice noodles, rice crackers, rice cakes, rice snacks, rice bamboo leaves were used as a box lunch for many years.
flour, and other fermented rice products.
Breakfast cereals from rice
Traditional rice-based convenience foods in Japan Breakfast rice cereals are made from rice grains, milled rice
One of the most famous traditional rice-based convenience flour, or cooked rice dough. These rice materials are precooked,
foods in Japan is sushi. Sushi originated as a preservative food dried, flaked, then expanded or puffed and toasted. Examples
The use of pressure (P) in addition to temperature (T) (Hayashi High pressure also has an effect on starches: a thick sus-
1987, Hayashi et al 1987) has been accepted in the field of pension of rice starch forms a ball by standing against its own
food science and technology over the past 15 years, and re- weight after pressurization at 700 MPa, indicating that starches
search and development are now under way in the food indus- are gelatinized by the pressure treatment.
try, universities, and government institutes (Hayashi 2002).
Several commercial products are now on the market that are
Versatile utility of high pressure
prepared using high-pressure techniques.
This paper describes the principle of high-pressure treat- A high-pressure treatment generally coagulates protein, thereby
ment and the effects of high pressure on foods, with emphasis inactivating the enzymes, gelatinizing the starches, and killing
on the high-pressure effects on starches. Finally, recent suc- microorganisms. Thus, the use of high pressure promises to be
cesses of the Echigo Seika Company in the application of high- a versatile process in food science and technology. Pressure
pressure techniques to rice and rice products are discussed. produces a new texture in meats and starch-based foods, while
keeping the original nutrients, flavor, and color.
Principle and method
High-pressure effects on pure starches
High pressure means high hydrostatic pressure generated by
the compression of water. A pressure of 100 MPa or higher High-pressure-induced gelatinization of starch
(usually lower than 1,000 MPa) is used under temperatures Effect of pressure on amylase digestibility of starches. The
below 100 °C. A unit of pressure is expressed in kg cm–2, bar, starches of potato, maize, and wheat are gelatinized by pres-
or pounds in–2, but now the international Pascal unit is com- sure treatment at warm conditions of 45–50 °C. The pressur-
monly used. Therefore, 1,000 bar or 1,000 kg cm–2 almost ization produces unique properties that are different from those
equals 100 MPa. of heat-gelatinization: heat-treatment destroys starch granules,
Food, which is contained in a plastic bag and sealed by resulting in a transparent solution, but a pressure-treatment
carefully removing air, is placed in a pressure vessel filled with swells the granules while maintaining the granular structure.
water and high pressure is then generated in the vessel by a Nevertheless, amylolytic enzymes such as α-, β-, and glc-amy-
pressure pump. lases digest the pressurized starches well, being similar to the
phenomenon in which the pressure-treatment of proteins in-
creases protease digestibility (Hayashi and Hayashida 1989).
Pressurization of food The pressure-induced gelatinization of starches exhibits
An egg is not crushed by compression at 600 MPa, but the egg a sigmoid curve, suggesting that a two-state transition is in-
white and yolk are coagulated. The color of egg yolk of a pres- volved, as in heat-induced gelatinization (see below).
surized egg is naturally yellow, whereas a boiled egg changes Effect of pressurization time on amylase digestibility of
to a faded yellow. The color difference is attributed to changes starches. To attain full amylase digestibility of starches, pres-
associated with pressurization: pressure affects only surization under warm conditions for 2 to 6 h is necessary.
noncovalent bonding and coagulates proteins without splitting Interestingly, pressurization of starches for a longer time
covalent bonds, thus keeping the color and smell intact. makes amylase digestion difficult: the amylase digestibility of
High pressure induces protein denaturation in the same starches decreased by 20–50% after pressurization for 17 h
way as high temperature. compared with the maximum digestibility obtained after pres-
Meat protein is also denatured by pressure treatment at surization for 2 to 6 h.
400 MPa, preserving the properties of raw meats. An example These observations suggest that pressure induces gelati-
of prawns or shrimp is interesting: although a boiled prawn nization of starches, similar to heating, but prolonged pressur-
turns red and the meat coagulates, the appearance of a pres- ization produces a new stable structure of starches, which is
surized prawn is the same as that of raw shrimp, but the meat not susceptible to attack by amylase (Ezaki and Hayashi 1992).
coagulates after pressurization at 400 MPa for 10 min. Birefringence of starches after pressurization. The bire-
fringence of starches is lost as increasing pressure is applied.
An expanding global rice market and diet diversification have ery goods, breakfast cereals, candy, desserts, beverages and
presented the rice industry in the United States and other de- dairy items, side dishes, package mixes, entrees (frozen, shelf-
veloped countries challenges and opportunities to develop new stable, chilled), meats and meat substitutes, sauces and spreads,
markets for rice. Global trade has expanded with advances in snacks, soups, and pet foods.
technologies leading to record yields of high-quality rice. Coun- Rice usage in entrees has increased 10-fold in the last 20
tries that at one time were importers or had only a small share years. Interest in ethnic foods has boosted this growth. Frozen
of the export market are now major exporters. The United ethnic entrees represented a $2.2 billion industry in 2000
States, once the lead exporter of high-quality rice, has been hit (Brandt 2001). A popular new addition is bowl meals—a fro-
hard by this shrinking export market. Outside of the U.S., coun- zen entree in a microwaveable bowl usually containing indi-
tries such as Brazil, China, Japan, Taiwan, and India have been vidually quick-frozen rice, meat or poultry, and vegetables.
faced with declining per-capita consumption because of diet They have contributed largely to the growth of the entree mar-
diversification. Value-added rice products that capture the ket.
unique nutritional and functional attributes of rice offer a means Package rice mixes to serve as side dishes combine rice
to keep rice as a mainstay of the diet and to expand domestic with seasonings, dehydrated vegetables, and other ingredients.
markets. This paper will highlight new product introductions The use of rice in packaged mixes doubled during the 1990s
in the U.S. and technologies recently developed at the USDA- (USA Rice Federation 2000). Mixes for Louisiana Cajun and
ARS Southern Regional Research Center for value-added rice Creole foods, such as jambalaya, gumbo, and etouffee, and
products. difficult-to-make dishes, such as risotto, are popular. Rice for
mixes is either regular or parboiled, with a low moisture con-
tent around 6.5%. For a product in which distinct grains are
New product introductions important, such as a pilaf, or when a sauce will add moisture,
Domestic usage of rice in the United States has tripled in the parboiled rice often performs better than regular rice.
last ten years in response to consumer demand for more health- Snack bars using rice include granola, breakfast, and
ful foods, more convenient products, and a growing interest in energy bars. Energy bars, introduced in the 1990s, account for
ethnic foods. The market has been flooded with new product 80 to 100 new product launches a year and represented about
introductions. Foods that use rice now include baby food, bak- $500–700 million in sales in 2000 (Kreuzer 2001, Mintel In-
Dehydrin is a group of late embryogenesis-abundant (LEA) supplier’s instructions. Protein fractions were mixed with
proteins, which accumulate in the maturing seed of many plants. sample buffer (3-10NL) and IPG strips (pH3-10NL, 7 cm) were
Dehydrins are reported to protect protein or membranes in plant soaked in the sample solution for 15 h. The IPG strips were set
tissue under abiotic stress such as desiccation or low tempera- on the system, and electrophoresis was carried out at 200–
ture (Close 1997). Besides the biological function, applica- 3,500 V for 90 min and at 3,500 V for 65 min. IPG strips were
tion for preservation of food quality is expected, similar to then equilibrated in 50 mM Tris-HCl, pH 8.8, containing 6 M
antifreeze proteins in fish (Breton et al 2000). Since rice bran urea, 30% (w/v) glycerol, 2% SDS, 10 mg mL–1 dithiothreitol,
could be a good source of dehydrin for nonfood usage, we and a trace of bromophenol blue for 20 min. Second-dimen-
tried to characterize rice dehydrin in this study. Rice bran pro- sional electrophoresis (SDS-PAGE) was carried out on a 5–
tein was analyzed by two-dimensional (2D) electrophoresis, 20% polyacrylamide gradient gel. Gels were stained with CBB
and dehydrins in the protein fraction were detected by R-250.
immunoblotting using antibody against the conserved motif-
structure in order. A major rice dehydrin was partially purified Immunoblot analysis
from rice bran and its cryoprotective activity was estimated to Electrophoresis gels were incubated in a blotting buffer (25
assess the function of the protein. mM Tris-HCl, pH 9.5, containing 40 mM ε-aminocaproic acid,
20% methanol, 0.05% SDS) for 20 min, and then protein on
the gel was blotted onto a PVDF membrane (Immun-Blot,
Materials and methods Biorad) at 2.5 mA for 30 min. The PVDF membrane was kept
Fractionation and partial purification of proteins overnight in 3% BSA-TBS (Tris-buffered saline, 20 mM Tris-
Rice bran was obtained in the process of refining 1 kg of rice HCl, pH 7.5, 500 mM NaCl, containing 3% bovine serum al-
(cv. Koshihikari) to 90%. Fifty grams of the rice bran was de- bumin). Dehydrin protein was detected by polyclonal antise-
fatted with 500 mL of hexane three times and kept at –20 oC rum raised against conserved lysine-motif sequence
before use. The defatted rice bran (20 g) was mixed with 20 (DQNEKKGIMDKIKEKLPGGH) conjugated with ovalbu-
mM HEPES-NaOH, pH 7.0, for 3 min by a hiscotron mixer min. The PVDF membrane was incubated for 60 min in the
(NS-50, Nichi-On), centrifuged twice at 8,000x g for 20 min. antiserum of rabbits diluted to 1:200 with 1% BSA-TBS, then
Supernatant was collected as a water-soluble fraction of rice washed with TBS-t (Tris-buffered saline containing 0.05%
bran. The heat-tolerable protein fraction was prepared by heat- tween 20) for 20 min three times, and with TBS for 20 min.
ing the water-soluble fraction in boiling water for 10 min and The washed membrane was incubated with peroxydase-con-
removing precipitate by centrifugation at 12,000x g for 20 min. jugated antirabbit goat antibody diluted to 1:2,000 with 1%
Rice dehydrin was partially purified by modifying the BSA-TBS, washed with TBS-t for 20 min three times, and
method of purification of soybean dehydrin (Momma et al washed with TBS for 20 min. Reacted polypeptides were de-
2003). From 20 g of defatted rice bran, a 10–50% saturated tected by a detection kit (Immunostain HRP1000, Konica).
ammonium sulfate fraction was isolated, dialyzed, heated in
boiling water, and then centrifuged. Supernatant was put onto Cryoprotective activity on lactate dehydrogenase
an S Sepharose Fast Flow column (2 × 35 cm), and eluted with Cryoprotective activities of dehydrins in comparison with bo-
a linear gradient of 0–300 mM NaCl. The dehydrin fraction vine serum albumin (BSA, A4378, Sigma) on freeze/thaw in-
was collected based on detection by SDS-PAGE and activation of lactate dehydrogenase were assayed following
immunoblotting with antiserum against the conserved lysine- the method of Lin (1992). Lactate dehydrogenase from rabbit
motif of dehydrin. The water-soluble fraction, heat-tolerable muscle (L5132, Sigma) was dissolved in 10 mM sodium phos-
fraction, and partially purified dehydrin were dialyzed against phate buffer, pH 7.5, at a concentration of 2.5 µg mL–1. The
distilled water and freeze-dried and used to assess cryoprotec- enzyme solution (0.1 mL) was mixed with equal amounts of
tive activity. test compounds dissolved in the same buffer at a concentra-
tion of 2 × 102 – 2 × 10–4 µM. The solutions were frozen at
Electrophoresis of rice bran proteins –20 oC for 24 h and thawed at room temperature for 5 min.
SDS-PAGE was carried out on 5–20% polyacrylamide gradi- Enzyme activity was measured with the change of absorbance
ent gel (PAGEL, NPG-520L, ATTO) by the method of Laemmli at 340 nm at 25 oC using an assay kit (DG1340K, Sigma). All
(1970). Gels were stained with Coomassie Brilliant Blue (CBB) samples were assayed twice in triplicate. Residual activity was
R-250. Two-dimensional electrophoresis of protein fractions shown as the percentage of the control activity assayed imme-
isolated from rice bran was conducted with the Multiphore II diately after mixing of enzyme and protein solution.
System (Amersham Pharmacia Biotech) following the
31.0 31.0
32.5
27.5 21.5
21.5
14.4
14.4 18.5
kDa
kDa kDa
Fig. 1. Two-dimensional electrophoresis and immunoblotting. (A) 2D electrophoresis of water-soluble fraction, (B) immunoblot of water-
soluble fraction, (C) 2D electrophoresis of heat-tolerable fraction. = spots reacted with antibody.
were prepared in a grinder equipped with a screen in 0.25-mm HMR increased continuously during the heating until 95 °C
apertures. (Table 1).
The breakdown value, which shows the levels of col-
Gelatinization properties of rice flour lapse for swelled starch granules, of HMR and HDR was
Dynamic oscillatory measurement of 10.0% (w/w) rice flour smaller than that of UTR. The starch granules in UTR easily
suspensions during gelatinization was done using a stress-con- became swollen at a lower temperature. However, the rigidity
trolled rheometer equipped with 35-mm-diameter parallel plate of the swollen granules was much weaker than that of HDR
geometry. The gap size was set to 1.0 mm. The sample was and HMR. These phenomena suggested that starch granule
heated at 5 °C min–1 from 30 to 150 °C at 6.28 rad s–1. swelling in HDR and HMR was restricted by the heat treat-
The 30% (w/w) rice flour suspensions were hermetically ments. It was known that heat treatments suppressed granule
sealed into a silver pan. The scanning temperatures and heat- swelling and retarded gelatinization (Ahmad and Williams
ing rate were 30–150 °C and 5 °C min–1, respectively. Differ- 1999). Tan δ of all samples decreased rapidly at TG2 , and then
ential scanning calorimetry (DSC) was performed using a DSC the curves had a flat part. UTR with low tan δ showed gelling
calorimeter with distilled water as a reference. behavior at lower temperature than with HMR and HDR, but
the value of UTR increased when the heating temperature (>85
X-ray diffractometry °C) increased slightly. HMR paste exhibited weak gel because
Crystallinity of the starch in rice flour was measured by an X- the tan δ of HMR was greater than that of UTR and HDR. It is
ray diffractometer. The sample was scanned through the 2θ recognized that a rigid starch granule formed with the interac-
range of 5–35°. Diffraction pattern was obtained using Cu-Kα tion of both amorphous and crystallite regions during heat-
radiation and a scintillation counter at 40 kV voltage, 300 mA moisture treatment. The rheological behavior of rice flour paste
current, and 5° s-1 scanning rate. was primarily attributed to interaction in the system, such as
contact between the granules directly and between the gran-
Cooking and processing quality ules and glucose chains (Hoover and Vasanthan 2000).
Rice snacks were prepared substituting 50% rice flour for wheat Table 2 shows the DSC characteristics of 30% (w/w)
flour with a twin-screw extruder. Crispiness of the rice snacks rice flour suspensions. The To of HMR was 8 °C higher than
was evaluated from discrete wavelet transformation. Rice pud- that of UTR and HDR (Table 2). Other endothermic peaks
dings were prepared with 10% (w/w) rice flour pastes and other (Tp2 and Tp3) in the DSC corresponding to amylose-lipid com-
ingredients (granular sugar and agar), and their cooking prop- plexes were observed at temperatures of 100–120 °C (Table
erties were determined by rheometry and sensory evaluation 2). The peak transition for HMR was at a higher temperature
using a 5-point hedonic scale. (112 °C) than for UTR and HDR (102 °C). Lai (2000) re-
ported that increases in gelatinization (To) and the peak of
amylose-lipid complex (Tp2) temperature were observed in DSC
Results and discussion curves of parboiled rice flour. It was also reported that the
Effect of heat treatments in rice flours peak of the amylose-lipid complexes shifted to a higher tem-
on gelatinization property perature with heat-moisture treatment. Changes in DSC pa-
The storage modulus (G´) of rice flour suspensions increased rameters of HMR indicate that rearrangement of starch chains
drastically at a certain temperature, which is hereinafter desig- occurs during heat-moisture treatment, and leads to the forma-
nated as TG´ (Table 1), during the heating period. Eliasson tion of crystallites of different thermal stabilities. Therefore,
(1986) reported that the initial increase in G´ was due to the the modification of G´ p and To of HMR is attributed to more
starch granules swelling progressively and becoming closely thermally stable crystallites in starch after heat-moisture treat-
packed in the suspension. The results indicated that HMR and ment.
HDR retarded the starch granules’ swelling in comparison with The intensities of the major peaks of HMR were weaker
UTR. By continuous heating, G´ of UTR and HDR rose to a than that of UTR and HDR measured by X-ray diffractometry.
maximum and then dropped down. On the other hand, G´ of However, the diffraction peak of HMR at 19.5° (2θ) became
much sharper and another feature peak at 13° (2θ) was also
UTR 60.2 ± 0.3b 68.1 ± 0.1 b 78.4 ± 0.3 b 8.6 ± 0.4 b 101.4 ± 0.6 a 0.5 ± 0.0 a – –
HDR 58.1 ± 0.3 a 66.5 ± 0.2 a 76.8 ± 0.2 a 9.6 ± 0.2 c 102.4 ± 1.9 a 0.8 ± 0.1 b – –
HMR 68.3 ± 0.0 c 76.2 ± 0.2 c 86.7 ± 0.5 c 5.2 ± 0.1 a – – 111.9 ± 0.0 0.7 ± 0.0
aT , onset temperature; T , peak temperature; T , conclusion temperature; ∆H, gelatinization enthalpy. Values followed by the same letter in the same column are not significantly
o p c
different (P<0.05). Mean ± standard deviation (n = 3).
observed. An increase in the intensities of two peaks (13° and compared to UTR and HDR. Moreover, HMR paste behaved
19.5°, 2θ) was attributed to the amylose-lipid complexes as a liquid-like paste. This indicates that the internal structures
(Miyoshi 2002). The weaker intensities indicate that the starch of starch granules in HMR became more thermally stable dur-
granules in HMR flour gelatinized partly during heat-mois- ing heat-moisture treatment. Dehydration and decreases in the
ture treatment, which supports a decrease in the gelatinization damaged starch content retarded starch granule swelling by
enthalpy from the DSC results. On the other hand, To and TG2 heat-dry treatment. Rice puddings made from HMR and HDR
of HMR flour shifted to a higher temperature during the treat- were harder, more brittle, and less pasty than those of UTR.
ment (Tables 1 and 2). These results indicate that heat-mois- Modified rice flours were effective in altering the expansion
ture treatment of starch in the HMR caused “thermal steady and crispiness of rice snacks. Therefore, heat-moisture and
amorphous,” which did not occur during the dry heating of heat-dry treatments indicated feasibility for improving the cook-
rice in HDR. Furthermore, heat treatment of rice caused the ing and processing qualities of rice-based products.
swelling properties of HDR and HMR to change as indicated
by G´ p in the dynamic oscillatory measurement.
References
Cooking and processing qualities Ahmad FB, Williams PA. 1999. Effect of salts on the gelatinization
of modified rice flour and rheological properties of sago starch. J. Agric. Food Chem.
Although the expansion index of the rice-flour-based snacks 47:3359-3366.
substituted at 50% (w/w) rice flour for wheat flour (rice flour Eliasson AC. 1986. Viscoelastic behavior during the gelatinization
of starch. J. Texture Stud. 17:253-265.
systems) was smaller than that of wheat flour snack with about
Hoover R, Vasanthan T. 2000. Effect of heat-moisture treatment on
9–25% levels, HMR snack has less effect than others. The
the structure and physicochemical properties of cereal, legume,
crispiness of HDR and HMR snacks was greater than that of and tuber starches. Carbohydr. Res. 72:455-463.
UTR snacks when analyzing by discrete wavelet transforma- Juliano BO. 1971. A simplified assay for milled-rice amylose. Ce-
tion. The food texture of HDR and HMR snacks expressed real Sci. Today 16:334-340.
“hard and crispy” and “soft and crispy,” respectively. There- Lai HM. 2001. Effects of hydrothermal treatment on the physico-
fore, it is possible that HDR and HMR are more suitable for chemical properties of pregelatinized rice flour. Food Chem.
manufacturing rice-flour-based snacks rather than UTR, which 72:455-463.
improved the expansion index and food texture of the snacks. Miyoshi E. 2002. Effects of heat-moisture treatment and lipids on
The application of HMR to rice pudding could improve the the gelatinization and retrogradation of maize and potato
handling and yield of products because the HMR paste was starches. Cereal Chem. 79:72-77.
fluid and less adhesive on cooking utensils. HDR and HMR
puddings were harder, more elastic, and/or more brittle than Notes
UTR pudding evaluated for breaking characteristics. More-
Authors’ addresses: T. Takahashi and N. Ohisa, Akita Research In-
over, HDR and HMR puddings were harder, more brittle, and
stitute of Food and Brewing, Akita, Japan; M. Miura and S.
less pasty than UTR pudding tested by the sensory evaluation.
Kobayashi, Iwate University, Morioka, Japan; K. Mori, Na-
Therefore, heat-moisture and heat-dry treatments for rice flour tional Food Research Institute, Tsukuba, Japan, e-mail:
could improve the food texture of rice-flour-based products. toru@arif.pref.akita.jp.
Conclusions
The gelatinization temperature and endothermic peaks corre-
sponding to amylose-lipid complexes of HMR shifted to a
higher temperature, while endothermic enthalpy decreased
In advanced countries, the aging of society is becoming a seri- Interaction of α′ and β with rice seed storage proteins
ous problem. Especially in Japan, elderly people over age 65
account for 14.5% of the population, and it is estimated that To investigate whether α′ and β interacted with rice seed stor-
they will increase to 24.7% by 2025 and 32.3% by 2050. In age proteins glutelin and prolamin, sequential extraction of
such an ultra-aged society, lifestyle-related diseases such as proteins from transgenic rice seeds and western blotting using
hypertension, hyperlipemia, and diabetes could spread. Fur- anti-α′ and β sera were conducted (Fig. 1-I). β contains no
ther, our immune system weakens against various viruses as Cys residue, whereas α′ contains four Cys residues in the pro
we age. Therefore, demand for food that can prevent lifestyle- region and one near the N-terminus of the mature region. Most
related disease and stimulate the immune system will become of the β was extracted with buffer without 2-mercaptoethernol,
larger and larger in advanced countries. 2-ME (fraction 1), but not by buffer with 2-ME (fraction 5)
Soybean β-conglycinin, composed of α, α′, and β sub- (Fig. 1-IB). On the other hand, α′ was extracted by buffer with-
units, is reported to have many physiological functions such out and with 2-ME (Fig. 1-IA). In this case, fraction 5 con-
as cholesterol-lowering activity in human serum (Sirtori et al tained acidic polypeptide of glutelin (data not shown), sug-
1995), serum triglyceride-lowering activity (Aoyama et al gesting that α′ interacted with glutelin by disulfide bonds. Frac-
2001), LDL-cholesterol-lowering activity (Manzoni et al 1998), tions 9–16 containing glutelin extracted by lactic acid gave
and phagocytosis-stimulating activity in the α′ subunit (Tsuruki bands of α′ and β, corresponding to 19% and 20% of their
et al 2001). On the other hand, rice seed storage proteins, glu- total amounts, respectively, and there was no difference in their
telin, prolamin, and globulin, do not have any significant physi- distributions in each fraction. Therefore, α′ and β in fractions
ological functions. Furthermore, they are deficient in lysine, 9–16 probably interacted with glutelin by hydrophobic inter-
whereas β-conglycinin, especially the α′ subunit, is rich in action. Fraction17 containing prolamin extracted by SDS buffer
lysine. Therefore, in this study, we attempted to develop also gave the bands of α′ and β, corresponding to about 10%
transgenic rice containing the α′ subunit for conferring high of their total amounts. This indicates that α′ and β interacted
nutritional value, the modified β subunit having improved ph- with prolamin by disulfide bonds or hydrophobic interaction.
agocytosis-stimulating activity (Maruyama et al 2003) for
physiological property, and the β subunit for control of ph-
agocytosis-stimulating activity.
Subcellular localization of β-conglycinin
in mature transgenic seeds
Transformation of rice with chimeric α′ and β genes Subcellular localizations of α′ and β in transgenic rice en-
β-conglycinin α′ and β cDNAs were connected with GluB-1 dosperm were analyzed by immunocytochemical analysis us-
and GluB-2 promoter regions in the pGTV-HPT vector (Becker ing immunoelectron microscopy (IEM). Rice seeds have two
et al 1992), respectively. The resultant pGTV-HPT/α′ and types of protein bodies (Tanaka et al 1980). Protein body I
pGTV-HPT/β were introduced into rice calli by Agrobacterium (PB-I) is derived from ER, mainly accumulating prolamin, and
tumefaciens-mediated transformation. Eleven and 22 lines protein body II (PB-II) is derived from vacuole, mainly accu-
expressing α′ and β were regenerated, respectively. Total pro- mulating glutelin and globulin. A morphological difference was
teins extracted by SDS buffer from T1 seeds were spotted onto not observed between nontransgenic and α′-transgenic rice
nitrocellulose membranes and the accumulation levels of α′ seeds. When ultra-thin sections of α′-transgenic rice seed were
and β were estimated immunologically. The accumulation level labeled with anti-α′ antiserum conjugated with gold particles,
of α′ was basically higher than that of β. The average and high- these particles were mainly detected in the peripheral region
est accumulation levels of α′ and β were 3.9% and 1.9%, and of PB-II (Fig. 1-IIA), and nonspecific adsorption was not ob-
8.5% and 4.3%, respectively. The lines with the highest accu- served in the nontransgenic rice seeds under the same condi-
mulation underwent subsequent analyses. To investigate the tions. On the other hand, there was an obvious morphological
reason for this difference, we compared the transcription level difference between nontransgenic and β-transgenic rice seeds
of α′ and β in rice seeds. The transcription level of α′ was (Fig. 1-IIB). The electron density of PB-II was uniform in
close to that of β. This suggests that the difference in accumu- nontransgenic rice seed, but PB-II in β-transgenic rice seed
lation level between α′ and β did not depend on the transcrip- contained a slightly low electron density region. When this
tional step, but depended on the posttranscriptional step. seed was reacted with anti-β serum, the gold particles were
M 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
–2-ME +2-ME Lactic acid
M 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
II A C E
B D F
Fig. 1. Accumulation of α′ and β -conglycinin in rice seeds. (I) Sequential extraction of α′ (A) and β (B) from individual transgenic rice
seeds. At first, proteins were extracted from ground rice seeds four times by buffer without 2-ME (lanes 1 to 4), then four times by
buffer with 2-ME (lanes 5 to 8), subsequently eight times by lactic acid (lanes 9 to 16), finally by SDS buffer (lane 17). (II) Electron
micrographs of α′- (A, B, C) and β - (B, D, F) transgenic rice seeds. A and B, mature seeds; C to F, developing seeds.
observed only in the low electron density regions, and no non- these vesicles (Fig. 1-IIE). This suggests that the α′ subunit
specific adsorption was observed in the nontransgenic rice was transported to the vacuoles via the Golgi apparatus.
seeds. Anti-glutelin gold particles were not observed in this In the β-transgenic rice seeds, some PB-II had two re-
region. These results indicate that β and glutelin were sepa- gions, high and low electron densities (Fig. 1-IID). The low
rately accumulated in a different region of PB-II. Around 10% electron density region of PB-II was not seen in immature α′-
of α′ and β expressed in the rice seeds was accumulated in transgenic and nontransgenic rice seeds. Anti-β gold particles
PB-I. mainly existed in low electron density regions of PB-II. On
the other hand, anti-glutelin gold particles did not exist in the
low electron density regions of PB-II. Moreover, morphologi-
Subcellular localization of β-conglycinin in developing seeds cally different vesicles were observed (Fig. 1-IIF, arrow).
Rice storage proteins are basically synthesized on the ER, then Though the size of these vesicles was analogous with that of
assembled in the ER lumen, transported to the Golgi appara- conventional vesicles (Fig. 1-IIE, arrowhead), with a diameter
tus, and sorted into the vacuoles (Krishnan et al 1986). To of 100 to 500 nm and uniform smooth density surface (Oparka
investigate the trafficking pathway of α′ in rice seeds, the de- and Harris 1982), the morphologically different vesicles were
veloping seeds of α′-transgenic rice were observed immuno- surrounded by ribosome membrane-like precursor-accumulat-
logically by EM. There was no morphological difference be- ing (PAC) vesicles, observed in pumpkin seeds (Hara-
tween α′-transgenic and nontransgenic rice seeds. In the early Nishimura et al 1998). It has been reported that PAC vesicles
stage of PB-II formation, α′ was already localized in the pe- serve to carry storage proteins from the ER to the vacuole,
ripheral region (Fig. 1-IIC). Vesicles budding from the Golgi bypassing the Golgi apparatus. So, rice seeds may possess the
apparatus were observed, and anti-α′ gold particles recognized same trafficking pathway. To investigate whether the PAC-like
B C
1 2 3 Phagocytotic index
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0
0 0.01 0.1 1.0 10
Concentration (mM)
Fig. 2. Development of rice to stimulate human phagocytotic activity. (A) Substitution of amino
acids of β subunit to confer the phagocytotic activity. I, β subunit; II, I122M/K124W. (B) SDS-
PAGE of proteins extracted by buffer from β and I122M/K124W transgenic rice seeds. Lane 1,
nontransgenic rice; lane 2, β transgenic rice; lane 3, I122M/K124W-transgenic rice. (C) Phago-
cytotic assay of I122M/K124W and β. Solid line, I122M/K124W; dotted line, β .
Wrap-up of Session 9
B.O. Juliano of the Philippine Rice Research Institute gave an High-fiber and low-digestible rice was revealed to have a consid-
“Overview of rice and rice-based products.” He pointed out the erable dietetic effect and to lower the blood-sugar level. He con-
importance of the amylose content of rice as table rice, for ex- cluded that breeding to improve the utility and nutritional func-
ample, in relation to palatability and glycemic index. For materi- tion in rice for the future should focus on diversification in the
als for baked and steamed products, the milling method was morphological, physicochemical, and nutritional characteristics
shown to be important. He indicated that the effect of food pro- of rice grain suitable for processing various value-added rice foods.
cessing on protein quality, level of antioxidants, mycotoxin level, K. Ohtsubo of the National Food Research Institute, Ja-
and acrylamide content has to be monitored. Finally, he con- pan, discussed “Processed novel foodstuffs from pregerminated
cluded that the high level of antioxidants in rice bran, oryzanol, brown rice by a twin-screw extruder.” He mentioned various kinds
tocopherols, and tocotrienols, as reflected in high unsaponifiable of Japanese processed rice products, such as retort-pouched
matter in bran oil, contributes, in addition to phytic acid, to the rice, dried cooked rice, aseptic cooked rice, and pregerminated
hypocholesterolemic effect and other health benefits of full-fat brown rice. He developed a novel foodstuff from pregerminated
bran. brown rice by a twin-screw extruder. The pregerminated brown
S. Isobe of the National Food Research Institute, Japan, rice was rich in gamma-aminobutyric acid (GABA). The extruded
discussed the “Sterilization effect of electrolyzed water on rice pregerminated brown rice contained more oryzanol, inositol, ferulic
food.” He pointed out that controlling microorganisms from raw acid, and dietary fibers than polished rice. Wheat bread prepared
materials to products and confirming safety in product distribu- with 30% of the extruded pregerminated brown rice contained
tion are most important for enhancing rice consumption in the more GABA, free sugars, and amino acids than ordinary bread.
development of new rice products, such as aseptic packaged These results showed that a novel foodstuff from extruded
rice and rice cookies. He found that the combination of acidic- pregerminated brown rice would be acceptable to consumers or
electrolyzed water and alkaline-electrolyzed water is very effec- the food industry as a promising foodstuff that contains more
tive for controlling microorganisms without a change in color and bio-functional components than ordinary rice products.
pH. He reported the effect of electrolyzed water on preventing Professor R. Hayashi of Nihon University covered “High-
raw rice from microorganism infestation by an in vitro test using pressure food processing of rice and starchy foods.” He pointed
Bacillus subtillis spores. out that high pressure is used for food processing at high tem-
H.C. Choi of the National Institute of Crop Science, Korea, perature. When starch is pressurized at 200 MPa or higher, it
covered the “Current status of varietal improvement and use of changes its physical properties, including loss of birefringence
specialty rice in Korea.” He mentioned various kinds of newly and increases in amylase digestibility and crystalline properties.
developed specialty rice varieties, such as low-amylose dull, Thus, high-pressure food processing is now applied in rice and
opaque endosperm, or high-fiber mutants, and high-amylose and its related food industry in Japan. He explained the high-pres-
high-lysine rice in Korea. He explained the relationship between sure effects on starches and mentioned the recent development
the physicochemical properties and processing suitabilities for of the high-pressure technique in the food industry.
rice snacks, rice noodles, rice bread, and fermented products.
India produces about 93 million tons of rice, which is about Table 1. Mean average percentages of additional outturns of total
one-fifth of world production. India is the second-largest rice- rice and head rice obtained in modern mills versus conventional
mills for raw and parboiled paddy.
producing country in the world. More than 200 commercial
rice varieties of indica subspecies are grown on an area of Mean average percentage of additional outturns
about 45 million hectares, with 25% irrigated area. Yield aver- of rice in modern mills versus conventional mills
ages 2,086 kg ha–1 (FAI 2003). Rice provides an occupation Type of paddy
for more laborers than any other industry in India. About two- milled Total rice Head rice
thirds of the production is retained by producers for their own Sheller Huller Sheller Huller
consumption and for seed. Of the total quantity of rice avail-
able, about 10% is processed into products such as beaten rice, Raw 2.5 6.6 6.1 15.1
puffed rice, puffed paddy, and deep-fried products such as (0.8–4.2)a (1.8–12.5) (1.9–12.9) (6.9–24.7)
papads and chakli, and fermented products such as idli, dosa, Parboiled 0.8 1.6 1.3 4.1
(0.0–1.3) (0.3–2.5) (0.8–2.5) (1.0–8.5)
and uthapam. India exports rice valued at about US$1.4 bil-
lion, of which Basmati is $420 million and non-Basmati is aNumbers in parentheses represent the variations of actual additional outturns from
$980 million (MFPI 2003). different varieties of paddy and over different mills. Source: MA (1971).
Development of a modernization program single hullers were left out for the pending development and
In 1955, the government of India set up a committee to exam- supply of modern equipment of the same capacity (250 to 300
ine the problems concerning the development of a rice-milling kg h–1) in the country.
industry. It recommended that preference be given to sheller- Apart from the seven modern rice mills initially set up
type mills over the existing metallic huller. On the basis of in the pilot study, the Food Corporation of India established
these recommendations, the government promulgated the Rice 25 modern rice mills without including silo storage. In addi-
Milling Industry (Regulation and Licensing) Act and the Rules tion, the cooperative and private sector together established
thereof for regulating the industry. In 1963, the government, 55 modern rice mills and modernized 51 existing mills.
under the Intensive Agricultural Development Program, along
with the Ford Foundation, undertook a study on the rice-mill- Modernization of huller rice mills
ing industry. Seven modern rice mills of rubber-roll sheller Over the years, several designs of low-cost mini rice mills be-
type were imported from Japan and Germany and were set up came available in the country. In July 1976, by an amendment
in seven rice-growing states. The remaining components, to the Rice Mill Industry Rules, provision was made for the
namely, parboiling plants, mechanical dryers, silos, and me- gradual modernization of single-huller mills by adding rub-
chanical handling and conveying equipment, were constructed ber-roll shellers or centrifugal dehuskers, paddy cleaners, and
and fabricated in the country. A report submitted in 1968 (MA separators. A huller subsidy scheme with 50% of the cost of
1971) showed that modern mills gave a substantial increase in modernization provided to each beneficiary was being imple-
total rice outturn and head-rice yield for both raw and par- mented in six states to encourage the modernization of huller
boiled paddy over existing sheller mills and a much higher mills.
increase over existing huller mills (Table 1). Further, modern
mills yielded by-products (husk and bran) separately for bet- Modernization of paddy parboiling
ter end uses. Mechanical drying increased yield by 1–2%, and During the 1950s, the Indian Council of Medical Research
the improved technique of parboiling in conjunction with me- sponsored research at the Central Food Technological Research
chanical drying increased yield by 2%. With silo storage, esti- Institute (CFTRI), Mysore, and Jadavpur University, Calcutta,
mated additional yield ranged from 2% to 4%. for improving the parboiling and drying processes as well as
The government of India recognized that there was vast nutritional and cooking quality of rice. In 1957-58, the Minis-
potential for improving the entire postharvest technology of try of Food and Agriculture recommended the new technique
the paddy/rice system. Accordingly, the Rice Milling Industry of parboiling developed by CFTRI for popularization.
(Regulation) Act, 1958, and Rice Milling Industry (Regula-
tion and Licensing) Rules, 1959, were amended in 1968 and Modernization of paddy/rice storage
1970, respectively. In the initial phase, a battery of hullers, In 1965-66, the “Save Grain Campaign” was organized by the
huller-cum-sheller combinations, and shellers in the organized government, covering 19 states/union territories to popularize
sector was brought under the purview of modernization. The an effective method of grain storage among farmers, traders,
25,000
etc. The Indian Grain Storage Institute was established at Hapur, undertake testing of mill machinery according to standards
with two substations at Ludhiana and Bapatla for conducting prescribed by the Bureau of Indian Standards.
applied research, development, and training in grain storage
with the assistance of the United Nations Development
Present position of the industry
Programme. Storage structures made of steel, concrete, and
metal-plastic combinations at the farmers’ level were being Rice postharvest technology in India has come a long way over
popularized, initiated through assistance received from the the past three decades. Now, more than 50% of the overall rice
United Kingdom and Netherlands Freedom From Hunger Cam- production is processed by modern mills, with a steady growth
paign, and USAID. in numbers (Fig. 1), about 40% by traditional mills, and about
10% by hand pounding. There are now 35,088 modern/mod-
Training, research, and extension ernized mills (0.5–4 t h–1), 4,538 under-runner disc shellers,
The government established in 1970 a full-fledged Rice Pro- 8,385 hullers-cum-shellers, and 91,287 metallic hullers (MFPI
cess Engineering Centre (presently, Post Harvest Technology 2003).
Centre) at the Indian Institute of Technology, Kharagpur, to Fifty percent of the total paddy production is parboiled.
provide training facilities at the Bachelor, Masters, and Ph.D. The CFTRI method of parboiling is used mostly by modern
levels, and short-term training courses for rice mill engineers, mills to drastically reduce soaking time and improve rice qual-
managers, and operators. The Ministry of Food Processing ity. However, the age-old practices of premilling treatments at
Industries (MFPI) has been sponsoring various research and the village level and traditional parboiling methods are still
development activities through this Centre as well as the Paddy being followed in some commercial mills.
Processing Research Centre, Thanjavur. MFPI has also set up Drying of paddy is carried out by sun drying in open
a technical cell to assist the industry by providing technical drying yards and/or by mechanical drying in L.S.U.-type dry-
assistance and consultancy service. In addition, extension and ers in commercial mills.
training work was undertaken through a network of ten regional The estimated losses in storage and handling are about
extension service centers. Grant-in-aid is being provided to 10%. About 70% of the produce is stored by farmers in small
various institutions to promote research and development, ex- indigenous storage structures with some modifications and
tension, and training work on postharvest technology for rice. made of locally available materials. The commercial-scale use
of bag storage of paddy/rice in godowns is common.
Development of modern equipment Out of a potential availability of 1 million t of bran oil,
Realizing that the country would require modern rice-milling only 0.5 million t is produced from 3.4 million t of bran, two-
equipment on a large scale, three manufacturers were initially thirds of which is of edible grade and the rest is of industrial
given licenses to manufacture machinery with suitable modi- grade. A part of the edible quality is consumed via root blend-
fications in collaboration with the companies from where seven ing or hydrogenation. About 70,000–80,000 t of the oil are
modern mills were imported for the pilot study. used for direct cooking purposes. De-oiled bran is mainly used
To encourage the production of good-quality rice mills for cattle feed.
and allied machinery by various manufacturers, four testing Rice husks are primarily used as fuel in husk-fired fur-
centers were set up in four different regions of the country to naces to produce steam through a boiler for parboiling, run-
Session 10: Postharvest technology for efficient processing and distribution of rice 295
ning the mill using a steam engine and hot air for drying. Pro- Even though sufficient systematic development has been
ducer gas is being produced from husks as fuel for a diesel made and experience has been gained, much remains to be
engine to run the mill and to generate electricity. Husks are done for rice postharvest technology to play an important role
also being used commercially to produce furfural. in the proper use of valuable resources and in contributing
At present in India, a large number of trained technical substantially toward Indian economic growth.
personnel are available. Many research workers are actively
involved in rice postharvest technology research. India manu-
References
factures and also exports all types of rice-processing machin-
ery and equipment. India also provides training to foreign per- FAI (Fertilizer Association of India). 2003. Fertilizer statistics. New
sonnel. Delhi (India): FAI.
MA (Ministry of Agriculture, Government of India). 1971. Modern
versus traditional rice mills—a performance study. Delhi (In-
Constraints and concerns dia): MA.
MFPI (Ministry of Food Processing Industries, Government of In-
There have been several problems with the modern large-ca-
dia). 2003. Annual report. New Delhi (India): MFPI.
pacity rice mills: the inadequate supply of paddy for continu-
ous operation throughout the year, controls and restrictions
imposed by the government with regard to price, a levy, and Notes
movement for paddy and rice. The high capital cost for silos Author’s address: Professor, Post Harvest Technology Centre, Agri-
prohibits the use of integrated units. Because of the unavail- cultural and Food Engineering Department, Indian Institute
ability of good-quality rice bran and its inadequate timely col- of Technology, Kharagpur 721302, India.
lection from the mills, the use of bran is very poor. A genuine
need exists for efficient low-cost mini rice mills to replace the
existing metallic hullers.
Drying, especially for rice, is an important postharvest pro- dryer structure and modify it to use an FIR electric heater, so
cess in monsoon climate areas such as Japan, where there is that it was difficult to develop commercial models.
much precipitation during the harvest period. The Japanese We decided to use a design including the FIR body with
Ministry of Agriculture, Forestry, and Fisheries has taught farm- a kerosene burner in the recirculating batch dryer because it is
ers to dry crops within 8 hours after harvest. widely used in Japan.
Currently in Japan, heated-air (HA) drying by burning
kerosene is regularly used for artificial grain drying. However,
Schematic of the FIR dryer
to maintain grain quality, the combustion heat is mixed with
ambient air and its temperature is lowered to about 40 to 50 The concept of an FIR dryer and an HA dryer is shown in
°C. Heat energy cannot be used effectively because of the need Figure 1. The HA dryer heats the air to 40 to 50 °C by mixing
to maintain grain quality. Although the HA dryer has evolved, the flame and the ambient air, and passes the heated air through
its capability is limited from the energy viewpoint. the grain layer. The FIR dryer converted thermal energy by
The applicability of far-infrared radiation (FIR) to grain heating an FIR body with a kerosene burner. Moreover, ex-
drying was investigated earlier by Harry and David (1959) and haust heat was reused for drying to increase drying efficiency.
then by Faulkner and Wratten (1969). FIR radiation can effi- The FIR body was made from a stainless-steel pipe coated
ciently transfer energy to an object. Moreover, the wavelength with silicone resin. FIR body surface areas were 1.6 to 3.6 m2.
of FIR is 2.5 to 1,000 µm, and grain has an infrared absorption The rate of radiation was 0.9, and the FIR body surface tem-
wavelength in this range. For this reason, FIR can be used to perature was 300 to 500 °C. The center of the wavelength was
heat grain efficiently. Heat emission causes very little energy 4.3 to 5.2 µm as determined by Wien’s displacement law. An
loss in transmitting drying energy to the grain. electromagnetic wave containing far-infrared radiation is gen-
In Japan, the possibility of FIR drying for rough rice erated. These spectra are the absorption band of water. More-
was also evaluated by Bekki (1991) and Matusoka (1990). Itoh over, 30–50% of the input energy was converted into FIR.
et al (1994) and Mouri and Huoqing (1996) also explored the The FIR body was useful wherever installed. Two types
possible application of FIR to other agricultural products. In of dryers are used, one with the FIR body installed in the dry-
their studies, they acquired basic drying data for rough rice ing chamber and the other with it installed in the space be-
using an electric heater. However, they needed to change the
Session 10: Postharvest technology for efficient processing and distribution of rice 297
Table 1. Results of drying tests.
Although the mechanism is not yet perfected, we demonstrated Harry WS, David WR. 1959. Drying rough rice with infra-red radia-
the potential of high-quality drying using an FIR dryer. tion. Rice J. June, p 16-38.
Itoh K, Chung SH. 1994. Drying of agricultural products using long
wave infrared radiation. Part 1. Fundamental heating charac-
Conclusions teristics of long wave infrared radiation. SASJ 25(1):39-45.
Matsuoka T. 1990. Drying characteristics of rough rice by far-infra-
The FIR dryer developed has satisfactory drying, energy effi-
red radiation heating. SASJ 21(2):85-93.
ciency, and energy conservation compared with the HA dryer. Mouri K, Huoqing L. 1996. Study on the drying of agricultural prod-
The reduced noise level will be comfortable to work with. In ucts by using far-infrared ray. Part 1. The performance of the
addition, palatability tended to improve. Thus, the potential drying device and the heating properties of agricultural prod-
for high-quality drying was demonstrated. ucts. SASJ 27(2):57-63.
These dryers have been produced with commercial mod-
els from cooperating research companies since 1998, and the
34,000 FIR dryers so far produced were distributed among
Notes
individual farmers in Japan. We now have a series of FIR dry- Authors’ address: Bio-oriented Technology Research Advancement
ers with a capacity of 1 to 30 t for a drying facility. Institution (BRAIN), Institute of Agricultural Machinery
(IAM), 1-40-2 Nisshin, Kita-ku, Saitama, 331-8537 Japan, e-
mail: yhidaka@affrc.go.jp.
References Acknowledgments: We are very grateful for the cooperation of the
Bekki E. 1991. Rough rice drying with a far-infrared panel heater. research companies Iseki Co., Ltd.; Kaneko Agricultural Ma-
JSAM 53(1):55-63. chinery Co., Ltd.; Satake Corporation; Shizuoka Seiki Co.,
Faulkner MD, Wratten FT. 1969. The Louisiana State University Ltd.; and Yamamoto Co., Ltd.
infrared preheat rice drier. 61st Annual Progress Report, Rice
Research Station, Louisiana State University Agricultural
Center, Crowley, Louisiana.
In addition to energy, materials can be recovered from bio- The use industry and new rice products
mass, and it is clear that the world must, from now on, sub-
stantially increase the use of biomass as a source of both en- Figure 1 presents uses of rice and by-products in the rice in-
ergy and materials by optimizing biomass production, conver- dustry. In Japan, 1,200,000 tons of rice are used for process-
sion, and use technologies. ing raw materials. This is about 12% of the total consumption.
Rice milling and processing need to be part of this ef- The main application is for sake (600,000 t), rice confectionary
fort, given that the rice-milling process generates by-products, (200,000 t), miso (180,000 t), and other purposes (200,000 t).
such as husks, bran, and broken kernels. Here, we report on The product of processed rice consists mainly of whole-milled
several aspects of rice milling and processing: (1) the use in- rice grain and its directly processed materials. Sake, miso,
dustry and new rice products, (2) the application of the dosa packed rice-cake, and arare are produced directly by koji-
fermentation process to indica-type rice use, (3) a finishing making and rice-cake making from whole-milled rice grain.
system for manufacturing prewashed rice, and (4) an overview Flour and starch-making rarely occur by milling and wet-mill-
of rice milling and processing in developing countries. ing such as are done with wheat and maize. A very small amount
of rice flour and starch has been distributed and used.
Prewashed rice
White rice Enriched rice-vitamin, mineral
Steamed rice
Flour milling For bread, dosa
Cooking Freezing Frozen food
Drying
Instant rice
Solidifying
Sushi/rice, ball/rice, burger
Session 10: Postharvest technology for efficient processing and distribution of rice 299
Rice grits and their processing by wet-milling knives, on which are plated fine nylon fiber 0.275 mm across,
Rice grits or broken rice kernels are used for brewing. Gener- containing 800-mesh carbon silicate particles at 15% weight.
ally, wet-milled rice flour is superior to dry-milled flour for Rice is fed into the lower portion of the machine and is pushed
making baked products. Adjuncts (usually 20–40% of the to- upward to the roller, where rice is polished by the brushes
tal ingredients) are used to obtain fermentable sugars at a lower embedded on the roller. Rifure comes in two capacities, 1 and
cost, and play a very important and specific role in giving a 1.5 t h–1, the former in two lines and the latter in three lines.
characteristic quality to beer to attract consumers. All-malt beer
is richer in flavor and has more fullness of body and a darker The wet-milling and finishing system
color. On the other hand, beer with adjuncts has a lighter color The Super jiff rice (SJR) consists of two stages of processing,
and taste, so that consumers drink more because of its less the first being wet-milled and the second being finishing. The
satiating taste. Also, beer with adjuncts has a higher colloidal milled rice is fed from the top of the wet-milling section, or
stability because of its lower nitrogen content (Yoshizawa and the first stage, and rice flows downward to the milling roller,
Kishi 1985). where the rice is wet-milled. The retention time of the rice in
the first stage is about 3 seconds. Water is added at about 50%
Application of the dosa fermentation process by weight to the feed rice, and heated to 20 to 30 °C by the
to indica-type rice use heat exchanger on the water tank. Subsequently, rice under-
Dosa is a spongy pancake that has a desirable flavor, texture, goes dehydration by centrifugation, and then is fed into the
and sourness. The Japanese are receptive to the characteristics finishing process, where the rice is finished by air heated to
of dosa. A new food product is expected to have a desirable about 45 °C. The water discharged from the wet-milling sec-
flavor, texture, and function through fermentation. We exam- tion is sent to a dryer, where water is evaporated by the heated
ined the lactobacillus identification and physical-biochemical air from the boiler, which then leaves flakes of suspended sol-
parameters of dosa and tofu-by-product-additive dosa batter. ids.
Dosa and tofu-by-product-additive dosa were measured using
the creep compliance test, and antioxidant activity was mea- The special processing and finishing system
sured by the DHHP method. The fermentation characteristics The Neo Tasty White Process (NTWP), a new special pro-
and texture of tofu-by-product-additive dosa (additive ratio cessing and finishing system, has been adopted by large rice
less than 60%) were similar to those of dosa produced by con- milling plants since 2000. This system consists of three sec-
ventional fermentation. Tofu-by-product-additive dosa had tions: the moist pressurized milling section, the heated adhe-
improved antioxidant activity, although the tofu-by-product- sion low-pressure mixing section, and the drying and separa-
additive ratio was more than 60%, yielding a poor texture tion system. The moist pressurized milling section mixes the
(Shimizu et al 2003). grain after adding water at a rate of 5% of the milled rice. The
retention time of milled rice in this section is about 10 sec-
onds. The subsequent secondary process in the heated adhe-
The finishing system sion low-pressure mixing section adds tapioca starch, heated
Facility for manufacturing prewashed rice to about 85 °C, to the grain at a rate of 50% of the milled rice.
A new manufacturing facility aims to polish conventional milled Here, tapioca starch acts as a heated adhesive agent. The tapi-
rice to a condition good enough to be cooked without prior oca starch removes the remaining bran layer from the milled
washing. Today, facilities for manufacturing prewashed rice rice while being mixed with the latter. The retention time of
can be broken down broadly into three types: (1) a dry-polish- the milled rice in this section is about 10 seconds. The tertiary
ing and finishing system, (2) a wet-milling and finishing sys- process separates the tapioca starch agent from the bran by the
tem, and (3) a special processing and finishing system. Facili- bag filter and the separator installed separately, and tapioca
ties of all these types finish the milled rice to a prewash rice starch is recycled back to the system after being heated by a
state by removing the surface portion of rice by 1–2% by weight boiler.
of the rice. These facilities are designed to process japonica
rice.
Overview of rice milling and processing in the Philippines
The dry-polishing and finishing system The following factors have been highlighted by many reports
Kapika and Rifure are major ones of this type. Kapika has a as the main causes of loss through the various stages from har-
long roller, 1.7 times longer than the roller of an ordinary fric- vest to consumption.
tion rice-milling machine, with three extrusions on the mixing 1. Characteristics of the rice variety. High-yielding va-
roller. The screen has a dodecagon shape. The milled rice flows rieties, especially IR64, are the main rice crop these
downward from the top of the machine and rice grains are days. These varieties are characterized by a high shat-
milled by grain-to-grain friction. The machine comes in four tering loss because of the delayed harvesting and
models, with capacities of 0.3, 1, 1.8, and 3 t h–1. threshing time.
Rifure was developed by converting a rice-milling ma-
chine into a dry-polishing machine. The machine has brush
The production of brown rice in the world is about 560 mil- ferulic acid has become a very important chemical material.
lion tons a year. Japan produced about ten million tons. Brown The carbon resource of ferulic acid is carbon dioxide in the
rice is usually polished before eating, and 10% of its weight is air. Further, ferulic acid will be obtained forever as long as
discharged as rice bran. Thus, 1 million tons of rice bran are humankind keeps on producing rice. Therefore, ferulic acid
discharged in Japan. About 40% of the rice bran is used to can become a renewable resource of the future chemical in-
produce rice edible oil in Japan, by extraction. In the course of dustry. This development is an environmentally friendly one.
production of rice edible oil, rice bran pitch is discharged. This Figure 1 illustrates the concept of renewable resource usage
is a dark and viscous oil and it is a waste material. We pre- with a system flowchart of the organic chemical industry us-
pared ferulic acid from rice bran pitch (Taniguchi et al 1994). ing ferulic acid obtained from rice bran.
Ferulic acid was produced from petroleum before. Petroleum,
however, is predicted to be used up in the future, after about
Preparation of ferulic acid
40 years. So far, the cost of ferulic acid has been about
US$2,500 per kg. No one could use this as a chemical mate- Ferulic acid and rice bran pitch
rial because of its expensiveness. In the method we have de- It is well known that ferulic acid can be prepared by the con-
veloped, the cost of ferulic acid is about $100 per kg. Thus, densation reaction of vanillin with malonic acid (Adams 1942).
Session 10: Postharvest technology for efficient processing and distribution of rice 301
Sun Renewable resource
Light
Synthetic methods Evaluation studies
— Organic chemistry
— Biochemistry Cancer prevention
Rice salad oil Antioxidant activity
Rice plant Brown rice Rice bran
Fossil resource
(Petroleum-coal)
Consumers
Fig. 1. A system flowchart for the organic chemical industry using ferulic acid obtained from rice bran.
This produces ferulic acid at a high yield, but takes as long as solvent of ethyl alcohol and water. Ferulic acid was obtained
3 weeks. In addition, the product is a mixture of trans- and in a 100% yield based on γ-oryzanol that existed in rice bran
cis-isomers. A vigorous search has therefore been made for a pitch (Taniguchi et al 1999).
commercial method for manufacturing high-purity trans-feru-
lic acid.
EGMP and ferulic acid, which act as anticarcinogens
In the manufacture of rice edible oil, a blackish brown
waste oil having a high viscosity, called rice bran pitch, an The synthesis of EGMP
alkaline oil cake rich in oil components, called soap stock, It is known that many kinds of chemopreventive agents are
and a by-product rich in crude fatty acids, called dark oil, are present in a diverse range of edible plants and herbs. It is dif-
discharged. These waste materials are known to contain useful ficult, however, to isolate or extract chemically pure substances
components but they have generally been disposed of as use- from plants because these substances are present in very small
less industrial by-products, for example, by burning, because amounts in plants.
of the lack of techniques for effective use. However, ferulic acid can now be obtained easily in a
We have investigated the components of rice bran pitch large amount from rice bran. This development prompted us
and found by chromatography that about 30% by weight is γ- to synthesize potential cancer chemopreventive agents. On
oryzanol itself. Because other components are contained in searching for an appropriate target compound that should be
large amounts, it is very difficult to recover γ-oryzanol alone, synthesized by the reaction of ferulic acid with other reagents,
leading to a high manufacturing cost. Thus, it is economically we found that aurapten, a citrus coumarin, is an excellent can-
not feasible to recover γ-oryzanol from rice bran pitch. How- cer chemopreventive agent. Recently, Koshimizu et al (1997)
ever, hydrolysis under certain conditions must make it pos- have isolated aurapten from citrus natudaidai (natsumikan in
sible for us to produce ferulic acid, as illustrated in Figure 2. Japanese) and showed that aurapten inhibited 12-O-
Yet, because γ-oryzanol was included within various oil tetradecanoylphorbol-13-acetate (TPA)-induced skin tumor
components of rice bran pitch, hydrolysis under usual condi- promotion in mice.
tions did not proceed at all. It should be noted that rice bran Aurapten consists of 7-hydroxycoumarin and a geranyl
pitch is quite immiscible with water so that a suitable alcohol group. 7-hydroxycoumarin has one phenolic hydroxyl group,
must be applied as a solvent. With a simple alcohol having a one carbon-carbon double bond, and one carbonyl group in its
small number of carbon atoms, which can be infinitely dis- molecule. Ferulic acid also has these functional groups in its
solved in water, the alkaline compound and rice bran pitch can molecule. Therefore, we synthesized many compounds in which
also be dissolved without difficulty. ethyl 3-(4-geranyloxy-3-methoxyphenyl)-2-propenpate
In our procedure, hydrolysis was carried out at 90– (EGMP) was contained, resembling the chemical structure of
100 °C for 8 hours under an atmospheric pressure, resulting in aurapten using ferulic acid. Especially, for EGMP, the geranyl
crude ferulic acid with a purity of about 70–90%. The crude group is attached to the phenolic hydroxyl group of ethyl
ferulic acid was purified by the recrystallization from a co- ferulate.
Brown rice
Rice bran
COOH
H3CO MOH
COO
HO ROH
Ca. 30% OCH3
OH
Session 10: Postharvest technology for efficient processing and distribution of rice 303
is an excellent UV absorbent, it is not stable at high tempera- tion in vitro. Bioorg. Med. Chem. 11:3807-3813.
tures. Thus, we have developed a facile preparation method of Taniguchi H, Hosoda A, Tsuno T, Maruta Y, Nomua E. 1999. Prepa-
the ferulic acid dimer, that is, bis-4,4′-(2-ethoxycarbonyl-1- ration of ferulic acid and its application for the synthesis of
ethenyl)-2,2′-methoxyphenoxymethane, which can become a cancer chemopreventive agents. Anticancer Res. 19:3757-
3762.
better UV absorbent than available UV absorbents such as
Taniguchi H, Nomura E, Hosoda A. 2003. Useful application of feru-
benzophenone and salicylic acid. The compound absorbs light
lic acid produced from rice bran. Yukigouseikagakukyoukaishi
at about 260–370 nm and exhibits two absorption maxima 61:310-321.
(λmax) at 321 nm (εmax: 3.28 × 104) and 292 nm (εmax: 3.21 × Taniguchi H, Nomura E, Tsuno T, Mimami S, Kato K, Hayashi C.
104). The decomposition temperature is 359 °C. This tempera- 1994. Method of manufacturing ferulic acid. US Patent No.
ture is far higher than that of ferulic acid, which decomposes 5,288,902. Japanese Patent No. 2095088.
a t Tsuda H, Park CE, Takasuka N, Baba-Toriyama H, Sekine K, Moore
176 °C. Based on these results, the compound is able to be- MA, Nomura E, Taniguchi H. 1999. Influence of ethyl 3-(4-
come a thermally stable absorbent of UV light for plastics. geranyloxy-3-methoxyphenyl)-2-propenpate (EGMP) on early
stage colon cacinogenesis in rats treated with azoxymethane
(AOM). Anticancer Res. 19:3779-3782.
Other applications Tsuno T, Maruta Y, Hosoda A, Nomura E, Taniguchi H. 2001. Prepa-
ration of a novel thermally stable UV absorbent from natural
Ferulic acid is used as the starting material for the preparation
resources. ITE Letters on Batteries, New Technologies &
of reagents having antibacterial activity, fragrances, food ad- Medicine 2C6:808-811.
ditives, control of germination, and cosmetics (Taniguchi at al
2003).
Notes
References Authors’ address: Industrial Technology Center of Wakayama Pre-
fecture, 60 Ogura, Wakayama 649-6261, Japan, e-
Nomura E, Kashiwada A, Hosoda A, Nakamura K, Morishita H, mail:taniguti@wakayama-kg.go.jp.
Tsuno T, Taniguchi H. 2003. Synthesis of amide compounds
of ferulic acid, and their stimulatory effects on insulin secre-
Recently, there have been new movements in the Japanese rice introduced to all of Fukuoka Prefecture in Japan starting in
market to inspect residual agrochemicals before harvesting, 2004, under the guidelines of JA Central Fukuoka.
that is, the crops are checked for residual agrochemicals for
each farm or each harvesting lot. If agrochemicals are not de-
Background of the new movements
tected, the crop will be harvested and sent to a large-scale dry-
ing facility, but, when agrochemicals are detected, harvesting I will explain the background of these new movements. For
will be postponed until the sun and rain dissolve the chemi- each agrochemical, the maximum permissible level is deter-
cals, and the crop will be harvested later, after another residual mined by a law, which is set based on studies of the medical
agrochemical inspection. effect of agrochemicals on the human body. However, in the
The reason for this preharvest inspection is that, when market, it is not so simple to judge that residual agrochemicals
crops are sent to a large-scale drying facility, they are mixed under the standard are acceptable and that residual agrochemi-
with other crops within a storage silo. If a part of the crop has cals over the standard are not.
residual agrochemicals, the whole grains will be contaminated Suppose that there are two rice products, one of which
by the agrochemicals and it will be impossible to separate sound is almost certain not to exceed the legal level of residual agro-
grains and contaminated grains afterward. chemicals from past experience and the way of cultivation but
Raw paddy is collected from standing plants and threshed not confirmed with measuring equipment, and the other is
by simple threshing equipment to make a 130-g sample, and checked for confirmation. Which rice will customers buy if
then taken to a large-scale drying facility to be processed with other commercial factors such as price, taste, and appearance
a test dryer and test husker to make brown rice, 25 g of which are the same? At least, in a culturally mature society, the latter
are used for checking residual agrochemicals. This method was will be chosen. Or, suppose two samples are checked by an
Measurement 15 min
Acetylcholine
Measuring principal enzyme inhibition method
Add substrate
H 2O
2H2O (Oxidant) (Reductant)
POD M
ChE Agrochemical e–
(Red.) Current detection
M: electron-transfer
Choline ChOD H 2 O2 POD M mediator
Electrode
(Red.) (Ox.) POD: peroxidase
H 2O + O 2
agrochemical inspector, one of which has near but under the placed among rice-producing areas. Therefore, a low-cost de-
regulatory level of residual agrochemical, and the other con- tecting system is indispensable.
tains no residual agrochemicals. As before, if other factors are Moreover, it is also requested that the system be able to
the same, the latter rice will be chosen. Even when the level of process a large amount of samples in a short period of time,
residual agrochemicals is under the standard, the product’s and that it produce test results rapidly, since the test is con-
value still differs widely in human society because of actually ducted right before harvesting.
checking the residues or confirming whether the level of resi- Also, the system must be able to check not only that resi-
due is near the standard or almost zero. Therefore, the issue of dues are under the regulatory level, but it must have a high
residual agrochemical in the recent market is not that of a le- sensitivity to confirm that residual agrochemicals do not re-
gal or a medical problem, but more of a cultural or human ally exist. To meet these demands, we have developed a mea-
nature matter. suring system using the enzyme inhibition method.
In recent rice trading in the market, this cultural or hu- Many agrochemicals have the effect of inhibiting the
man nature aspect is claimed in competitions. This is clear neural transmission of insects, so this system measures the
because agrochemical-free production is promoted in the mar- degree to which the components of agrochemicals inhibit cho-
ket. linesterase, an important factor for neural transmission func-
Another reason that rice producers are becoming sensi- tion.
tive to residual agrochemicals is that, if their rice products are Acetylcholinesterase (ACHE) is known as a hydrolytic
found to contain residues over the standard, because of the enzyme of excitatory neurotransmitter acetylcholine that ex-
abovementioned cultural and human nature aspects, the brand ists widely from invertebrates to Pisces and mammals, and
name could suffer critical damage through news reports. activity is inhibited by organophosphorus compounds and car-
The new movements regarding residual agrochemical in- bamate compounds.
spection started from this background, and are attempting to Figure 1 shows the measuring procedures. First is a pre-
achieve a higher level of residue-free rice production. processing to collect the possibly contained residual agrochemi-
cals, and then the toxicity of agrochemicals to the enzyme is
activated before adding the enzyme cholinesterase. After add-
A new residual agrochemical detector to study the trend ing the enzyme, acetylcholine is added as a substrate. This
Because of this new trend, we need to check residual agro- acetylcholine is transformed to choline by the effect of cho-
chemicals for each farm or harvesting lot in a short period of linesterase. The amount of the generated choline is measured
time; thus, a large number of agrochemical detectors must be by transforming the choline into hydrogen peroxide and check-
Session 10: Postharvest technology for efficient processing and distribution of rice 305
Measuring principle A method to measure the amount of agrochemicals from the enzyme inactivities by
agrochemicals = enzyme inhibition method
Acetylcholine
Reference
The generation of H2O2 is fast since ChE
cholinesterase is intact
Acetylcholine
Rapid increase of electric current Choline
Current/nA
ChE Agrochemical
300
ChOD
Without agrochemical Choline
250 H 2 O2
ChOD
200 H 2 O2
VR
150
VS Sample
100 The generation of H2O2 is slow since
cholinesterase is affected
0
2 3 4 5 6 7 8
Time (min)
ing by a hydrogen peroxide sensor. The resulting output is in be checked by liquid chromatography and mass spectrometer,
the form of an electrical current. During this process, the etc. However, this legal method is high in cost and the measur-
amount of generated choline is influenced by the inhibition of ing efficiency, the number of samples that can be checked per
the activity of cholinesterase by agrochemicals, which in turn day, is low. What if screening tests are done by using our simple,
affect the amount of hydrogen peroxide and change the elec- inexpensive, high-efficiency and low-measuring-cost method,
trical current. The amount of electrical current is determined and the legal method is performed only on the samples that are
by the degree to which residual agrochemicals inhibit the ef- picked up by this screening test? If the legal method detects
fect of cholinesterase. Therefore, the electrical current is af- one out of 300 samples, our high-accuracy screening will de-
fected by the amount of agrochemicals, so the residual agro- tect about two samples that contain residual agrochemicals.
chemicals can be measured by using a hydrogen peroxide sen- Therefore, while the conventional method needed the ineffi-
sor. cient legal method to be performed on all 300 samples, this
To be more specific, as shown in Figure 2, if there are high-accuracy screening allows you to use the legal method
no agrochemicals, the generation of hydrogen peroxide is fast, on only two samples. The measuring efficiency and the cost
so the increase in electrical current is faster. On the other hand, for measuring residual agrochemicals are significantly im-
if there are residual agrochemicals, the generation of hydro- proved in this total system with the new technology and legal
gen peroxide is slow, thus the increase in electrical current is method combined.
slower. Therefore, the amount of residual agrochemicals can The idea of screening before the legal method is not new;
be measured by the differential VR or VS. This method cannot however, the point is how far you can reduce the number of
specify the type of agrochemical component, but it can detect samples using the screening method. If the degree of narrow-
the agrochemicals that inhibit the cholinesterase as a whole. ing down by a screening is small, and the dependence on the
In the case of DDVP (dichlorovinyl dimethyl phosphate), legal method is still high, the screening will make the mea-
which is an organophosphorus insecticide, sensitivity is high surement inefficient. However, because of its high measuring
enough compared with the legal standard. sensitivity, our new screening method can significantly reduce
Therefore, this method can be considered as a high-ac- the number of samples to be checked by the legal method, and
curacy screening method. According to data of several years can well be called a high-accuracy screening method.
ago, residual agrochemicals are actually detected in one out of The theory of agrochemical measurement by the enzyme
300 samples in Japan, but they contain only a minute amount inhibition method existed from before, although this may be
of agrochemicals and the residues barely exceed the legal stan- the first practical system that allows for a high-accuracy mea-
dard. In the conventional legal method, all 300 samples must surement.
One of the constraints to the growth of Philippine rice produc- mestic resource cost (DRC), and the seemingly unrelated re-
tion is the lack of appropriate and adequate infrastructure. This gression (SUR) analysis.
includes marketing and distribution infrastructure such as farm- The rice-based farm households across ecosystems, sea-
to-market roads; limited access to credit, processing, and stor- sons, and infrastructure with yield as the major indicator were
age facilities; and the lack of effective irrigation systems. Sev- regressed to measure the impact of infrastructure indicators
eral policy studies had pointed out the importance of these on total productivity, output prices, input prices, technology
facilities, but studies that look into the impact of infrastructure adoption, and net farm income. The models were estimated
on profitability and global competitiveness are still lacking. It using the SUR analysis. The models were as follows:
is therefore important to evaluate and measure the impact of 1. Output = f (production inputs, infra indicators)
irrigation and road infrastructure on rice profitability. 2. Output price = f (yield, infra indicators)
Our study evaluated the impact of infrastructure (roads 3. Input price = f (inputs, infra indicators)
and irrigation) development on farm productivity, rural income, 4. Technology adoption = f (inputs, infra indicators)
technology adoption, and transaction costs in rice farming, and 5. Farm income = f (inputs, infra indicators)
compared the profitability and global competitiveness of rice The explanatory variables for the regression model were
farmers who have access to good infrastructure and those who the inputs of production that include farm size (ha), seed use
do not. (kg ha–1), fertilizers—nitrogen, phosphorus, and potassium (kg
ha–l), pesticides (active ingredients in kg), and preharvest la-
bor (person-days), as well as infrastructure indicators such as
Methodology road classification, distance and travel time to major whole-
The study used the 1996 wet season (WS) and 1997 dry sea- sale markets, transport cost of rough rice from farm to major
son (DS) farm-level and barangay-level data of the Regular wholesale market, availability of transportation vehicles, num-
Monitoring of Rice-Based Farm Household Survey (RBFHS) ber of input/output dealers, number of business establishments,
project of the Philippine Rice Research Institute’s (PhilRice) and number of communication facilities.
Socioeconomic Division (SED) and the Bureau of Agricul-
tural Statistics (BAS). The farm-level data included produc-
Results and discussion
tion costs and returns and socioeconomic and technological
profiles of 2,500 rice farmers from 281 barangays (villages) Yield and profitability
of 30 selected provinces in the Philippines. Sample barangays Irrigated versus nonirrigated farms. Table 1 shows the yield
were classified according to their state of road infrastructure performance and relative profitability of farmers from irrigated
development. A barangay with good infrastructure was char- and nonirrigated environments. Data in both seasons showed
acterized by its year-round accessibility owing to asphalt or a substantial yield advantage of irrigated over nonirrigated
concrete roads and sufficient transport facilities. On the other farms. The average yield for irrigated farms during the WS
hand, a barangay with poor road infrastructure was character- was 3.54 t ha–l, whereas the yield for nonirrigated rice was
ized by a lack of all-weather roads, making it barely acces- 2.55 t ha–l. Similarly, in the DS, yield of irrigated rice farms
sible, and having insufficient transportation vehicles. averaged 3.66 t ha–l, while yield of nonirrigated farms aver-
The methods employed in the analysis were the costs aged 2.34 t ha–l, or around a 1 t ha–l (56%) yield difference.
and returns analyses, competitive advantage analysis using do-
Session 10: Postharvest technology for efficient processing and distribution of rice 307
Table 1. Average costs and returns of rice production in the Philippines, 1996
WS and 1997 DS by agroecological system.
aIncludes herbicides, insecticides, fungicides, molluscicides, and rodenticides. bIncludes food costs,
interest expenses, and land tax. Official exchange rate: P26.30 = US$1.
Source: Department of Agriculture-PhilRice Rice-Based Farm Household Survey, 1997.
Costs and returns analysis showed that irrigated farms Under the import trade scenario, there is a competitive
were more profitable than nonirrigated farms in both seasons. advantage in domestically producing rice only on irrigated
The higher production cost incurred by irrigated rice farms farms. For infrastructure development, farmers with access to
was offset by their higher yield gains. Net farm income real- good infrastructure had higher levels of competitive advan-
ized from irrigated farms was higher at US$375 ha–l for the tage.
WS and $443 ha–l for the DS versus $182 ha–l for the WS and
$180 ha–l for the DS obtained from nonirrigated farms. More- Seemingly unrelated regression analysis
over, net profit-to-cost ratios were better for irrigated farms. Results of the SUR analysis showed that nitrogen fertilizer,
Good versus poor infrastructure. Table 2 shows the com- pesticides, and irrigation were positively correlated with yield
parative productivity and profitability of farmers with access and showed a very high level of significance. These indicate
to good and poor road infrastructure. Rice yield of farmers on that an increase in one of these explanatory variables will posi-
irrigated farms with access to good road infrastructure was tively and significantly affect rice yield. Road structure affected
higher at 3.61 t ha–l for the WS and 3.77 t ha–l for the DS than the price of urea fertilizer. Urea fertilizer was more expensive
that of farms with poor infrastructure at 3.31 t ha–l for the WS in areas with poor road conditions owing to higher transporta-
and 3.24 t ha–l for the DS. Results of the analysis showed that, tion costs incurred in the purchase of this input.
except for the nonirrigated farms during the DS, farms with Another major finding of the study is that road structure
good infrastructure appeared to be more profitable as indi- and vehicle density positively and significantly affected the
cated by their higher gross returns, net income, and net profit- quantity of nitrogen fertilizer used in rice production. Farmers
to-cost ratios. Net incomes from farms with good road infra- would apply more nitrogen fertilizer to their crops because of
structure were higher by 14–40% for all seasons. its lower price brought about by lower transaction costs, good
road conditions, and the availability of sufficient transport fa-
Competitive advantage analysis cilities in the area.
Results of the competitive advantage analysis showed that the The major determinants of farm income were the major
Philippines has no competitive advantage under the export trade inputs of production such as nitrogen fertilizer, pesticides,
scenario. The country has no competitive advantage in domes- preharvest labor, as well as infrastructure indicators such as
tically producing rice for export as indicated by the resource irrigation, road structure, and transportation costs.
cost ratios greater than 1. A major reason for this export
noncompetitiveness was the high per unit cost of rice produc-
tion shown in the costs and returns analyses.
Yield (t ha–1) 3.61 3.31 2.61 2.39 3.77 3.24 2.26 2.32
Price of output ($ t–1) 290 290 290 290 300 300 300 300
Gross returns ($ ha–1) 1,047 960 757 693 1,131 972 678 696
Production costs ($ ha–1)
Seeds 49 42 46 40 48 43 43 38
Fertilizer 87 79 53 52 90 80 58 47
Pesticidesa 52 45 48 38 49 43 45 40
Irrigation fee 34 0 – – 37 40 – –
Transportation 3 3 3 3 3 3 3 3
Fuel and oil 18 19 10 13 21 19 10 14
Labor costs 339 353 289 285 346 336 296 297
Depreciation 24 24 20 17 24 23 17 17
Land (rental value) 27 27 14 14 27 27 15 15
Other costsb 38 35 32 27 34 34 28 23
Total production costs ($ ha–1) 670 629 515 490 679 648 514 493
Net farm income ($ ha–1) 377 331 242 203 452 324 164 203
Net profit-cost ratio 0.56 0.53 0.47 0.41 0.67 0.50 0.32 0.41
Cost per metric ton ($ t–1) 186 190 197 205 180 200 227 213
aIncludes herbicides, insecticides, fungicides, molluscicides, and rodenticides. bIncludes food costs, interest ex-
Policy implications izer and certified seeds. The restriction in the use of these yield-
enhancing inputs results in lower farm productivity and con-
Results of the study generally imply that good infrastructure sequently lowers farm income.
improves farm profitability and productivity. Therefore, there The dismal condition of rural roads has been cited as the
is a need to evolve an infrastructure development strategy in major reason (1) traders are able to dictate farm-gate prices,
irrigation and farm-to-market roads for the rice sector. (2) transportation costs are exorbitant, and (3) agricultural
growth did not lead to growth of the nonagricultural sector as
Public investment policy for irrigation it did in the newly industrializing economies (AGRICOM
Irrigation increases the cropping intensity of farmers. The 1997). Given the long gestation period in investment in public
Department of Agriculture (DA) should fast-track the construc- infrastructure, strategic alliances among agricultural stakehold-
tion of additional irrigation systems because of their high in- ers should always advocate the inclusion of public infrastruc-
teraction effect in terms of efficiency in the use of water and ture investment in the budgetary process.
other inputs such as fertilizer and seeds to sustain productivity The open trade regime in agricultural products has made
and reduce risks in the rice sector. Irrigation also provides a the Philippines vulnerable to globalization. Given the state of
safety net from distortion caused by natural agro-climatic phe- noncompetitiveness of the rice sector, there is a need to focus
nomena such as drought, as indicated by the higher net farm on key result areas, some of which are proposed by Gonzales
income of farmers in irrigated areas. (2000a,b), following the agribusiness framework to cushion
the impact of globalization.
Public investment policy for farm-to-market roads
Inadequate farm-to-market roads generally increase the trans- Accelerate farmers’ access to good-quality seeds
action costs of inputs and outputs. According to Gonzales Results of this study showed that only a very small percentage
(2000a,b), inadequate infrastructure in marketing and distri- of Philippine rice farmers use good-quality seeds. Since good-
bution accounts for 35–40% of the total costs of production quality seeds enhance yield by 10–15% over ordinary seeds,
from the farm to the wholesale market in the grain sector. It full adoption of good-quality seeds can partly enhance pro-
also restricts the adoption of technology because of the rela- ductivity in the rice sector.
tively higher cost on-farm of yield-shifter inputs such as fertil-
Session 10: Postharvest technology for efficient processing and distribution of rice 309
Improvements in postharvest facilities Gonzales LA. 2000a. Philippine agriculture in the next millennium:
Postharvest losses account for a substantial portion in the pro- some strategic issues and directions. Proceedings of the Aventis
duction-distribution chain and should therefore be minimized, Launch Conference, 4 February 2000, Makati City, Philip-
if not totally eradicated, to enhance productivity. pines.
Gonzales LA. 2000b. The Philippine rice industry: strategic issues
and directions. STRIVE Foundation, Putho, Tuntungin, Los
Promote a higher rate of fertilizer use
Baños, Laguna, Philippines.
The complementarity of fertilizer use, availability of irriga-
tion water, and adoption of high-yielding varieties should be
considered in promoting higher rates of fertilizer use. Provi- Notes
sion of other services such as credit and marketing services as Authors’ addresses: Rowena G. Manalili, Philippine Rice Research
well as improvement in infrastructure such as irrigation and Institute, Maligaya, Science City of Muñoz, 3119 Nueva Ecija,
transportation will encourage farmers to apply the right amount Philippines, e-mail: rgmanalili@philrice.gov.ph; Leonardo A.
of fertilizer. Gonzales, SIKAP/STRIVE Foundation, One Tepeyac Place,
Gov. San Luis Road, Putho, Tuntungin, Los Baños, Laguna,
Philippines, e-mail: lag@strivefoundation.com.
References
AGRICOM (Congressional Commission on Agricultural Modern-
ization). 1997. Modernizing agriculture. Manila (Philippines):
Congress of the Philippines.
There are two commercial brown-rice storage systems in Ja- Materials and methods
pan: an environment-temperature storage system, in which the
temperature during storage is not controlled, and a low-tem- Storage structure
perature storage system, in which the temperature is maintained Two of 12 silos of a grain elevator constructed in Hokkaido,
below 15 °C during storage. The low-temperature storage sys- Japan, were used for the on-farm experiment. Each silo had a
tem minimizes insect activities and mold growth, and fumi- diameter of 7.4 m and a height of 23.2 m. Each silo was made
gants are therefore not required during storage. However, this of steel with a 75-mm insulation layer on the outside of the
system requires electricity to cool the stored brown rice. wall. An automatic system was installed for aeration from the
A previous basic study (Kawamura et al 1997) revealed bottom through to the top of the silo, and an automatic system
that the quality of rice stored at a temperature below ice point was installed for ventilation in the upper vacant space of the
is comparable with that of newly harvested rice. Kawamura et silo.
al (2004) reported that rice with moisture content of less than
17.8% does not freeze at a temperature of –80 °C and that a Rice samples
temperature below ice point minimizes the physiological ac- Kirara397 and Hoshinoyume varieties were used for the stor-
tivities in rice and hence minimizes the deterioration of rice age experiment. The moisture content of each rough rice sample
quality. Rice storage at a temperature below ice point was was 15.4%.
named “super-low-temperature storage” by Kawamura because
the storage temperature was much lower than that of low-tem- Storage conditions
perature storage. Five hundred tons of Kirara397 rough rice was stored in one
Based on the results of the studies mentioned above, we silo and 494 t of Hoshinoyume rough rice was stored in the
conducted an on-farm experiment to develop a new rice stor- other silo. The storage period was from the end of November
age technique using ambient fresh chilly air in winter and to 1999 until the end of July 2000. The rough rice in the two silos
preserve the quality of rice. was simultaneously aerated in January 2000. Aeration was
automatically carried out when the temperature of the fresh air
was below –7 °C and it was continued until the cooling front
had moved through all of the rough rice in the silos. Total aer-
ating time (fan time) was 91 h.
20
Silo storage
Aeration Low-temperature
storage (7.6 ºC)
10
Silo storage in
0 center of silo (1.5 ºC)
10
10 Nov 10 Dec 10 Jan 9 Feb 11 Mar 10 Apr 11 May 10 Jun 11 Jul 10 Aug
Period (1999-2000)
Fig. 1. Rice grain temperatures during on-farm silo storage and control storage (Kirara397).
Three control storage experiments were also carried out Results and discussion
at the same time: a room-temperature storage, in which rice
samples were stored in a laboratory room; a low-temperature Grain temperature during storage
storage, in which rice samples were stored in a commercial The range of grain temperatures in the vertical direction in
rice warehouse and kept at a temperature below 15 °C; and a each silo was less than 3 °C, and there was no tendency in the
storage experiment at –5 °C, in which rice samples were stored grain temperature distribution. The temperatures recorded in
in a refrigerator and kept at –5 °C. About 15 kg each of rough the center and near the wall of each silo were averaged, and
rice and brown rice were stored in polyethylene containers in the average values were used as indicators of changes in grain
each control experiment. temperature during on-farm storage (Fig. 1).
The grain temperature was 10 °C at the beginning of
Temperature measurement and sampling storage. The temperature of rice grains near the wall gradually
The temperatures of rough rice in the center of the silo and at decreased as the ambient temperature fell. The minimum grain
10 cm from the silo wall were measured by thermocouples set temperature near the wall in the middle of February was
at 2.2-m intervals from the bottom to the top of the silo. The –2 °C. From the end of March until the end of the storage
temperatures of rough rice and brown rice in each container in period (at the end of July), the grain temperature near the wall
the control storage experiments were also measured. gradually increased. The maximum grain temperature near the
Rice was sampled and tested for quality before, during, wall in the middle of July was 21 °C. The grain temperature in
and after storage. Rough rice samples were taken from the the center of the silo remained constant (10 °C) at the begin-
center of the silo and at four points 15 cm from the silo wall ning of storage and fell to –2 °C when the silo was aerated at
(north, west, south, and east in the silo) at the end of the stor- the end of January. The grain temperature in the center of the
age period. Sampling depths below the surface of the rough silo remained below ice point until the end of July. After aera-
rice were 0.1, 0.5, 1.0, 2.0, and 4.0 m. tion, the grain temperature throughout the silo remained be-
low ice point until the end of March. These results indicate
Quality assessment that super-low-temperature storage of rice in a farm-scale silo
Moisture content, germination rate, free-fat acidity, and can be achieved by using aeration and chilly fresh air in win-
texturogram property were determined to assess rice quality. ter.
Session 10: Postharvest technology for efficient processing and distribution of rice 311
The thermal conductivity of rough rice (about 0.09 W Storage capacity of rough rice (000 t)
m–1 K–1) is nearly equal to that of lumber (0.15 W m–1 K–1) 120
and glass wool (0.04 W m–1 K–1). This means that rough rice Number of newly 104
115
is a thermal insulating material. On the other hand, the spe- 100 constructed facilities 111
cific heat of rough rice (about 1.7 J K–1 g–1) is larger than that Each year
Total
of lumber (1.3 J K–1 g–1) and concrete (0.8 J K–1 g–1). This 79
80
means that rough rice is a refrigerant material. Because of these
64
physical properties of rough rice, the grain temperature in the
60
center of the silo remained below ice point until the end of 47
July despite the increase in outside temperature in summer.
40
(3) (6)
(4)
Quality of rice sampled from different parts 19 28
20 (2) (5) (4)
of the silo 8
25
17 16 (1)
Rice samples were collected from different parts of the silo at 0 11 (2)
6 5
the end of the storage period and were examined for quality. 0
1995 1996 1997 1998 1999 2000 2001 2002 2003
The germination rates of all samples were more than 97%.
Year of construction
Free-fat acidities of all samples ranged from 12 to 15 mg. A
germination rate of rice grains of more than 90% and free-fat Fig. 2. Storage capacity of rough rice in Hokkaido, Japan.
acidity of rice grains of less than 20 mg mean that there has
been no deterioration in the quality of the rice. The tempera- use of fresh chilly air enables the quality of rice to be pre-
ture of rice grains near the silo wall increased to about 20 °C served at a high level without requiring a cooling unit or elec-
in July. However, the temperature of rice grains near the wall tricity. The use of the super-low-temperature storage technique
was below ice point during winter, and the mean temperature has been increasing in cold regions of Japan in recent years. In
of rice grains near the silo wall during storage was 5.8 °C (Fig. Hokkaido, the northernmost island in Japan, 26 grain-eleva-
1). The results of quality measurements showed that there was tors have been constructed since 1996. The storage capacity
no deterioration in the quality of rice grains near the silo wall. of rough rice in Hokkaido was 115,000 t at the end of 2003
(Fig. 2).
Quality assessment of rice before and after storage
The germination rates of rice subjected to the silo, –5 °C, and
low-temperature storage were more than 98% after storage, as
References
high as that of samples taken before storage. On the other hand, Kawamura S, Natsuga M, Kouno S, Itoh K. 1997. Super-low tem-
the germination rate of rice subjected to room-temperature stor- perature storage for preserving rice quality. Proceedings of
age decreased to about 50%, indicating that the rice had lost the Joint International Conference on Agricultural Engineer-
vitality during room-temperature storage. The free-fat acidi- ing & Technology, Volume III, 15-18 December 1997, Dhaka,
Bangladesh. p 820-824.
ties of rice increased in all of the storage experiments. How-
Kawamura S, Takekura K, Itoh K. 2004. Rice quality preservation
ever, there were differences in the rates of increase in free-fat
during on-farm storage using fresh chilly air. Proceedings of
acidity: the rate of increase in rough rice during storage was the 2004 International Quality Grains Conference, 19-22 July
less than that of brown rice. Moreover, the rate of increase in 2004, Indianapolis, Ind. (USA). p 1-17.
free-fat acidity was highest for rice stored at a room tempera-
ture, next highest for rice stored at a low temperature and in
the silo, and lowest for rice stored at –5 °C. These results indi- Notes
cate that the decomposition process of fat is slower in rough Authors’ addresses: Shuso Kawamura, Graduate School of Agricul-
rice during storage and in rice during storage at lower tem- tural Science, Hokkaido University, Kita 9, Nishi 9, Sapporo
peratures. Changes in the texture of cooked rice were mini- 060-8589, e-mail: shuso@bpe.agr.hokudai.ac.jp; Kazuhiro
mized by silo storage and –5 °C storage and by storage of Takekura, National Agricultural Research Center, 3-1-1
rough rice. Kannondai, Tsukuba, Ibaraki 305-8666, e-mail:
The results of quality assessment indicate that storage takekura@affrc.go.jp; Kazuhiko Itoh, Graduate School of
of rice at temperatures below ice point (silo storage and –5 °C Agricultural Science, Hokkaido University, Kita 9, Nishi 9,
Sapporo 060-8589, Japan, e-mail:
storage in this study), which is called super-low-temperature
kazu@bpe.agr.hokudai.ac.jp.
storage, preserves the quality of rice at a much higher level
than that of rice stored at higher temperatures.
Parboiled rice is the staple food in some developing countries, Goal definition and scoping
including Bangladesh. In Bangladesh, per capita consumption This stage defines the purpose of the study, the expected prod-
is reported to be 168 kg year–1 (FAO STAT 2001). Kar et al uct of the study, the boundary conditions, and the assumption.
(1999) reported that one-fourth of the world’s paddy is par- Furthermore, a reference unit (functional unit, FU), to which
boiled. Parboiled rice has been produced by both traditional all the environmental impacts are related, has to be defined.
and modern methods. Modern methods are energy- and capi- The goal of this study was to investigate the life cycle of rice
tal-intensive, and are not suitable for small-scale operation at produced by the vessel, small-boiler, medium-boiler, and un-
the village level. In rural areas, various methods are being used treated process. Figure 1 shows the life cycle of rice under
to produce rice and these consume different amounts of en- different processing methods and the boundary of this study,
ergy. The commonly used local parboiling processes are ves- which is encircled by a broken line.
sel (0.5–1.2 t batch–1), small-boiler (2–4 t batch–1), and me- It has been reported that agricultural LCAs often exclude
dium-boiler (5–10 t batch–1). Biomass, that is, rice husk, is the production processes of medicine and insecticides, machines,
main source of energy for parboiling. In Bangladesh, 63% of buildings, and roads because of a lack of data (Cederberg and
the total energy consumption is met by biomass fuel and 37% Mattsson 2000). In this study, environmental releases related
is commercial fuel. The household sector consumes 80% of to the cultivation of paddy and construction of the parboiling
total biomass energy and rural households use it almost exclu- facilities were not considered because of the unavailability of
sively for cooking (Bari et al 1998). It is reported that biomass data. Usually, the paddy and rice are marketed at the nearby
combustion contributes as much as 20–50% of global green- market or at the mill-gate in local areas. Manually operated
house gas emissions, of which one-third may come from house- three-wheeled rickshaw-vans are the main transportation used
holds. This has an adverse effect on human health and the en- in local parboiling processes. Therefore, energy consumption
vironment (Smith 1999). With the growing concern about en- and environmental releases from transportation were also not
vironmental pollution and health risks, it is very important to considered. The FU has been defined as the mass of the prod-
find a sustainable rice-processing method. Therefore, this study uct, that is, 1 ton of head rice.
attempts to evaluate the environmental effects of different rice-
processing methods using the life-cycle assessment (LCA) Inventory analysis
methodology. The life-cycle inventory (LCI) analysis quantifies resource use,
energy consumption, and environmental releases associated
with the system being evaluated. The energy consumption in
Methodology the parboiling processes, dehusking, milling, and cooking, was
LCA is a methodology to assess all the environmental impacts taken from our own study (Roy 2003). Energy consumption
associated with a product, process, or activity throughout its during drying of parboiled paddy was derived from the litera-
life cycle by accounting for and evaluating resource consump- ture (Palipane et al 1988). Energy consumption in dehusking,
tion and emissions. This concept consists of four major steps: milling, and cooking was considered to be the same for par-
goal definition and scoping, inventory analysis, impact assess- boiled rice produced by different processing methods.
ment, and improvement assessment (SETAC 1993).
Seedling
Fresh
Harvesting Paddy Untreated Dehusking Milling Rice Cooking Consumption
Parboiling Parboiled
Presteaming Soaking Steaming Drying Dehusking Milling
Fig. 1. Life cycle of rice and the system boundary of this study (marked by the broken line).
Session 10: Postharvest technology for efficient processing and distribution of rice 313
Table 1. Inventory results (inputs and outputs) of the rice life cycle produced by differ-
ent processes.a
aThehead-rice yield of the vessel, small-boiler, medium-boiler, and untreated process was 65.9%, 69.2%,
66.6%, and 60%, respectively.
It was assumed that the energy requirement in the life process. Fresh rice consumes less energy than parboiled rice,
cycle of rice was met by biomass energy and that biomass (rice but it consumes more resource (paddy) because of low head-
husk) was the source of primary energy for all types of energy rice yield (Table 1).
consumed in the rice life cycle, except diesel energy. The bio-
mass used in parboiling and electricity generation is consid- Atmospheric emissions
ered to be the industrial use of biomass and an improved do- The environmental load was found to be lower for untreated
mestic cook-stove was used for cooking. It has been reported rice than for parboiled rice. Parboiled rice produced by the
that the electricity efficiency of IGCC (integrated gasification medium-boiler process had a lower environmental load than
combined cycles, biomass option) is 43% (Gustavsson 1997). that produced by the vessel and small-boiler. The atmospheric
The efficiency of an improved cook-stove (ASTRA) is reported emission inventory indicates a need to switch methods to re-
to be 30% (Bhattacharya et al 1999). Based on these factors, duce air emissions.
the primary energy, atmospheric emissions (CO2, CO, TSP,
CH4, NOx, SOx, and VOC), water emissions (amino nitrogen, Water emissions
phenol, BOD, and COD), and production of final solid waste The excess soak-water drains after the soaking treatment in
(ash) were determined. the parboiling process and is the main source of water emis-
sions in the life cycle of parboiled rice. There is no soaking
treatment in the untreated process and hence no water emis-
Results and discussion sions were reported for fresh rice.
Although the inventory results consist of an exhaustive list of
parameters, the only parameters discussed in this study are Solid waste (ash)
resource use, energy use, environmental emissions, and solid Untreated rice produces the lowest amount of solid waste com-
waste (ash) production. The inventory results (inputs and out- pared with the other processes. For parboiled rice, solid waste
puts) of the rice life cycle are reported in Table 1. was lower for the medium-boiler process (Table 1).
LCI analysis of rice reveals that all the processes have a
Energy consumption negative effect on the environment and the environmental load
The rice-processing industry consumes some energy and at the depends on the production process. The results showed a
same time produces some energy in the form of a by-product gradual decrease in inventory (energy consumption, atmo-
(rice husk). Based on this assumption, an energy inventory was spheric emissions, and solid waste) from the small-boiler to
worked out and expressed in terms of primary energy (rice the untreated process (small-boiler>vessel>medium-
husk). In the case of parboiled rice, the energy inventory was boiler>untreated) and there are no waterborne emissions in
lower for the medium-boiler than for the vessel and small-boiler the case of the untreated process. The environmental load is
A great deal of improvement in rice grain end-use quality has change. On the contrary, changes made using postharvest meth-
occurred over the past decades using postharvest technology. odology must be made to each rice harvest.
For example, a method has been developed to fabricate rice Advancement in genetics methods and biochemical sepa-
flour into grain-like shapes possessing superior mineral con- ration and identification tools have lagged behind the techno-
tent. Preharvest technology, however, has been used to a lesser logical advancements underpinning postharvest methods used
degree to affect rice end-use quality. To date, only traditional to modify rice end-use quality. But, the gap between the tools
breeding methods combined with molecular markers and rela- of pre- and postharvest technology is closing. Consequently,
tively simple physical and biochemical tools have been used some research programs across the world are using these tools
to change the end-use quality of rice cultivars. The few traits to evaluate the potential of developing cultivars with modified
that have been modified using these tools are amylose con- traits such as storage, cooking, processing, and health-benefi-
tent, gelatinization temperature, kernel morphology, bran color, cial properties. These programs consist of collaboration among
chalkiness, and milling quality. Changing rice end-use quality geneticists, breeders, food scientists, nutritionists, and chem-
characteristics using preharvest techniques has an advantage ists. Below is a review of research efforts using the tools of
over postharvest methods because they create permanent
Session 10: Postharvest technology for efficient processing and distribution of rice 315
genetics to modify rice end-use quality, including traditional, Located primarily in the aleurone layer, phytic acid, when
mutation, and molecular breeding techniques. consumed by humans, is thought to reduce mineral
bioavailability. Also, when fed to animals, the phosphorus in
this molecule is excreted in their feces and rain takes it to wa-
Traditional breeding
terways, where the algae growth that occurs as a result de-
To use traditional breeding to modify rice traits, genetic diver- pletes the oxygen and other aquatic life dies. Thus, bran low in
sity for the trait must exist. Nearly 420,500 Oryza samples are phytic acid is desired, but little genetic variation for this trait
maintained in germplasm collections across the world. Most has been found in rice germplasm.
of this genetic material has yet to be explored for its variation A low-phytic-acid mutant has been reported by Larson
in any aspects of end-use quality. There is an increasing inter- et al (2000). Gamma radiation mutagenesis was used to create
est by some breeding programs around the world in using these the nonlethal low-phytic-acid mutant. The mutation was des-
genetic resources to develop rice cultivars with modified ker- ignated low phytic acid 1-1 (lpa1-1). Homozygosity for rice
nel characteristics. As an example of this, work by my col- lpa-1 changed the wild-type portion of seed phosphorus in
leagues and me at the U.S. Department of Agriculture is dis- phytic acid from 71% to 39%. The inorganic portion of phos-
cussed below. phorus increased from 5% to 32%, whereas the total phospho-
A germplasm collection was chosen to contain as much rus content of the seeds was unchanged.
genetic diversity as possible. More than 200 rice cultivars and Cool temperatures during grain maturation can interact
exotic accessions from over 25 countries were grown using with some genotypes to cause greater amylose content and thus
the same cultural management practices during two growing firmer rice may ensue. Variability in textural properties is a
seasons. The collection has been used to evaluate the potential problem for the food-processing industry. Therefore, cultivars
for developing cultivars with a reduced need for postharvest that are insensitive to low growing temperatures would be ad-
bran stabilization and enhanced levels or ratios of bran vantageous. Suzuki et al (2002) set out to create such a geno-
phytochemicals using genetic diversity and cross-pollination type by mutating a temperature-sensitive low-amylose type.
techniques. This genotype had a mutation at the du2 locus and, as a result,
Oil content (twofold) and fatty acid profiles in the col- the seeds could be visibly distinguished depending on whether
lection varied a great deal (Goffman et al 2003). Significant they matured at cool or warmer temperatures. A mutant that
genotypic differences in bran E vitamer (i.e., tocopherols and was insensitive to cool temperature was identified based on its
tocotrienols) content (twofold) and gamma-oryzanol (three- semitranslucent phenotype. Amylose synthesis in the mutant
fold) content were also found (Chen and Bergman, unreported and wild type was not significantly different. The authors are
data). Total gamma-oryzanol content and the three primary now confident that the effect of the mutation is not limited to
gamma-oryzanol compounds were not correlated with the level the du2 mutant background. They are now incorporating the
of E vitamers across the germplasm collection. The three mutation into locally adapted germplasm (personal communi-
gamma-oryzanol compounds were not correlated with each cation).
other. Nor was the tocopherol content correlated with the As with other cereal grains, rice is lower in the essential
tocotrienol content. The lines with white or light brown bran amino acid lysine than other sources of protein consumed by
had low phenolic contents, whereas those with darker bran humans. Elevated lysine levels in rice could help reduce hu-
(dark brown, red, and black) showed a much greater range in mans’ reliance on animal protein, which is in shorter supply
phenolic concentration (23-fold) and antiradical efficiency than grain protein and its production is considered to be less
(Goffman and Bergman 2003). The rate of hydrolytic rancid- sustainable. Kim et al (2004) irradiated calli of a cultivar with
ity in this collection was well correlated with esterase activity, gamma-rays and mutant cell lines resistant to growth inhibi-
which varied approximately fourfold (Goffman and Bergman tion by S-(2-aminoethyl)-cysteine (AEC) were selected. The
2003). These results suggest that it will be possible to use tra- AEC-resistant M3 lines had lysine levels from 1.58 to 2.66
ditional breeding methods to increase the oil and various anti- times higher than the wild type.
oxidant fractions of rice bran, as well as to modify the bran’s
fatty acid profile and reduce its susceptibility to becoming ran-
cid.
Molecular plant breeding
Some characteristics that are desired in rice kernels will have
to be created using genetic transformation techniques. These
Mutation breeding are generally traits that do not exist, vary too little or too much
When an end-use quality trait is desired, but does not exist in in rice germplasm, or are unlikely to be found using mutation
rice germplasm or it varies little, then mutation or molecular methods. The creation of enhanced levels of iron, zinc, and
breeding must be used. Mutation breeding becomes the method beta-carotene in rice using molecular breeding techniques has
of choice when the genes controlling the trait have not been been well described (Beyer et al 2002, Vasconcelos et al 2003).
cloned, or if transgenic technology is not desired. Mutation Described below are examples of other efforts to use transfor-
breeding has recently been used successfully to create new mation technology to change rice end-use quality.
rice kernel traits. A few examples are described below.
Session 10: Postharvest technology for efficient processing and distribution of rice 317
The role of proteins in textural changes in aged rice
Toshihisa Ohno, Takahiro Kaneko, and Naganori Ohisa
The texture of cooked rice from aged grains is generally hard were as follows: plunger and stage, aluminum; bite speed, 2
and nonsticky. Japanese people prefer cooked rice from newly mm sec–1; and sample temperature at the time of measurement,
harvested grains because of its highly sticky and soft texture. 25 °C.
Many factors have been identified as the cause of these changes The influence of the addition of an oxidizing or reduc-
in cooking properties, including the increase in free fatty acids ing agent on the S/H ratio was also investigated. Instead of
during storage and their inhibitory effect on the gelatinization distilled water, an aqueous solution of 5 mM potassium iodate
of rice starch, changes in the physicochemical properties of (an oxidizing agent) or 8 mM sodium sulfite or 10 mM cys-
rice starch itself, and changes in the structure-maintaining com- teine (reducing agents) was used as the cooking water. The
ponents (Shibuya and Iwasaki 1982). method of cooking and measuring S/H ratio was the same as
Some investigators have specifically studied the proteins described above.
associated with these changes. Moritaka and Yasumatsu (1972)
suggested that disulfide bond formation is responsible for the Removal of external layer of polished rice
textural changes in cooked rice during storage, and Arai and After the rice samples were further polished in a mill (Grain
Watanabe (1993) reported that denatured proteins were asso- Testing Mill, Satake or HS-4 Mill, Tiyoda Engineering), they
ciated with these changes. Furthermore, Hamaker and Griffin were cooked. The rate of removal of the external layer of the
(1990) reported that reducing agents influence the texture of polished rice was 7–20%. The method of cooking and mea-
cooked rice. suring S/H ratio was the same as described above.
We therefore aimed to clarify the role of proteins in the
textural changes in cooked rice seen after storage. Electrophoresis
The rice flour derived from the external layer of the rice grains
was mixed with distilled water or an aqueous solution of 5
Materials and methods mM potassium iodate for 30 min. After the mixture had been
Experimental materials centrifuged, the pellet was rinsed in distilled water and centri-
A japonica-type rice (Oryza sativa L. japonica, cultivar fuged again. The precipitate was mixed with Solution A (10
Akitakomachi) harvested in Akita Prefecture was used in the mM sodium hydroxide, 1% SDS) and incubated for 2 h.
experiments. After brown rice grains harvested in 2001 were The mixtures were then centrifuged and the resulting su-
polished to wash-free grade, they were used as New Rice A. pernatants subjected to electrophoresis. The native-PAGE us-
New Rice A grains, stored for 2 months at 30 °C in a closed ing 0.1% SDS was performed on 5–20% gradient polyacryla-
aluminum pouch, were used as Aged Rice A. Brown rice grains mide gels. Coomassie brilliant blue was used as the staining
harvested in 2001 were normally polished and used as New agent.
Rice B. After the brown rice grains harvested in 1999 had been
stored for 3 years at 15 °C in a paper bag, they were normally
Results and discussion
polished and used as Aged Rice B. Thereafter, all samples were
stored at below –5 °C until use in the experiments. We evaluated the textures of cooked rice. Figure 1 shows the
S/H ratios of the various types of cooked rice. The S/H ratio of
Cooking and texture measurement aged rice cooked in water was low. Since Japanese people gen-
Ten-gram samples of polished rice were not washed and then erally prefer cooked rice with a high S/H ratio (Fuke et al 1991),
soaked in 16 mL of distilled water for 1 h in an aluminum cup Aged Rice A and B were inferior to New Rice A and B. These
covered with aluminum foil. Next, they were cooked in an elec- changes have been observed in numerous previous reports (e.g.,
tric rice cooker (National, SR-W100, with 75 mL of water Fuke et al 1991).
added to the outer pan) for about 12 min. After steaming for Next, the influence of the addition of an oxidizing or
30 min, the cooked rice was transferred to a closed dish and reducing agent on the S/H ratio was investigated (Fig. 1). The
kept at 25 °C for 90 min. addition of 5 mM potassium iodate to the cooking water low-
The hardness (H) and stickiness (S) of the individual ered the S/H ratio. On the other hand, the addition of a reduc-
cooked rice grains were then measured by 90% deformation ing agent to the cooking water raised it. The S/H ratio of aged
with a compression tester (Taketomo, Tensipresser, TTP- rice was low as described above, and the addition of the oxi-
50BX2). The most important parameter, stickiness/hardness dizing agent brought the S/H ratio of New Rice down. We there-
(S/H) ratio, was worked out. More than 40 cooked rice grains fore estimated that the textural changes in the aged rice are
of every type were measured. Conditions for the measurement attributable to oxidation.
0.3
0.2
0.1
0
New Rice A Aged Rice A New Rice B Aged Rice B
Session 10: Postharvest technology for efficient processing and distribution of rice 319
References Shibuya N, Iwasaki T. 1982. Effect of the enzymatic removal of en-
dosperm cell wall on the gelatinization properties of aged and
Arai E, Watanabe M. 1993. Enzymatic improvement of the cooking unaged rice flours. Starch/Staerke 34:300-303.
quality of aged rice: a main mode of protease action. Biosci.
Biotech. Biochem. 57:911-914.
Hamaker BR, Griffin VK. 1990. Changing the viscoelastic proper- Notes
ties of cooked rice through protein disruption. Cereal Chem.
67:261-264. Authors’ address: Akita Research Institute of Food and Brewing, 4-
Moritaka S, Yasumatsu K. 1972. The effect of sulfhydryl groups on 26, Aza-sanuki, Araya-machi, Akita 010-1623, Japan, e-mail:
storage deterioration of milled rice: studies on cereals (part ohno@arif.pref.akita.jp.
10). Eiyo Syokuryo 25(2):59-62. Acknowledgments: This work was financially supported by the Iijima
Food Science Foundation.
India has increased rice production markedly during the past in storing paddy grains with different storage practices and
50 years by enhancing production from 22 to 93 million t. On also to find out their needs for effective storing of grains.
the other hand, because of severe and unprecedented drought
conditions during 2002-03, the total annual rice production
Materials and methods
had a shortfall of 17 million t. However, rice demand in 2010
and 2025 will be 100 and 140 million t, respectively (Mishra The study was conducted in Vellore District of Tamil Nadu,
2004). Hence, the shortfall in unexpected years and the pro- India, where rice is grown on about 72,000 ha and there are 20
jected demand to feed the growing population have to be met blocks in this district. Among these blocks, Tirupattur block
by proper handling of produced rice besides increasing pro- was randomly selected for collection of data. This block con-
duction. The importance of storage is often ignored in rice sists of 32 revenue villages wherein agriculture is the main
production ventures. Inadequate storage facilities and improper source of livelihood and rice is grown in three distinct sea-
storage can cause considerable losses to rice farmers. Total sons: the dry (Jul-Nov), wet (Dec-Mar), and summer season
postharvest losses of food grains account for 9.3% and the (Feb-May). Employing a proportionate sampling method, 100
losses during storage alone account for 6.5% (Gandhi 1983). respondents (farm women) representing small farms (0.1 to 1
Unless losses of food grains in the postharvest phase are satis- ha) and big farms (>1 ha) were selected, out of which 50 were
factorily minimized, the problem of feeding hungry millions small-farm women and another 50 big-farm women. Thus, a
could continue even with a substantial increase in production. sample of 100 women farmers (20 from each village) was se-
Thus, farmers must learn the proper way of storing rice for a lected randomly from five randomly selected villages. Most
transitory period on the farm and the local market and for a of the women farmers cultivate relatively small farms (0.1 to 1
seasonal/national reserve or buffer stock, besides protecting ha) and some who have more land (>1 ha) live in clusters.
grains against weather and insects. Preventing grain discol- Through a well-structured interview schedule, the data on adop-
oration and a bad odor to maintain rice quality for feeding and tion of women farmers for storing rice grains in different stor-
marketing is essential. age methods and feedback on suggestions to overcome prob-
The majority of the world’s agricultural producers are lems in adoption were collected. The relevant data gathered
women: they produce more than 50% of the food that is grown were tabulated, processed, and subjected to statistical tests of
worldwide. Women are usually responsible for food process- percentage analysis.
ing and make a major contribution to food storage. Women
farmers play a distinct and well-accepted role in all activities
Results and discussion
of rice cultivation and 93% of the farm women are actively
involved themselves in storing rice grains (Sumathi and Budhar Adoption level of storage methods
2003). Women farmers receive less than 5% of extension ser- Among 12 practices, four (Table 1)—drying grains before stor-
vices worldwide and the priorities of women farmers are rarely age, keeping bags horizontally, keeping bags on planks and
reflected in agricultural research or national policies, and the leaving passages, and using metal traps for rodent control—
role of women as agricultural producers is still largely unrec- were found to have been adopted by a majority of the farm
ognized and has not been addressed (LEISA 2002). Hence, we women. Upadhay and Gupta (1987) also stated that the major-
conducted a study among women farmers to find out their role ity of rural women possessed sufficient knowledge of grain
storage. All these practices required less cost and skill, which supply of materials (64%). Half of the farm women needed
might be the reason for their better adoption. Nearly 50% of timely technical guidance on postharvest technology of paddy
the farm women followed the use of polythene-lined bags and (52%) since these women farmers are economically stable and
prebaiting and poison baiting with zinc phosphide against rats. hence need a timely supply of materials as well as technical
This practice is dangerous to domestic animals, has a high cost guidance.
of technology, and has no technical guidance. The rest of the In general, most of the farm women suggested financial
practices had a very low level of adoption because they were assistance on a credit basis (54%) and a timely supply of ma-
capital-intensive and demanded skill in practicing them. A terials (52%), followed by providing a subsidized facility
considerably higher percentage of nonadoption was found in (48%), an adequate supply of materials (47%), and timely tech-
other storage practices for paddy grains because these prac- nical guidance (44%). Less than 30% of the farm women
tices had a high cost, were complex, a local substitute was wanted the research system to develop low-cost storage struc-
available, and knowledge on storage structures and technical tures to prevent grain losses from infestation and arrange cam-
guidance were lacking. paigns on scientific storage structures.
The big-farm women had a higher percentage of adop- The results imply that, in many parts of the world, there
tion of storage practices such as impregnated bags, pretreat- is an increasing trend in the involvement of farm women in
ment of storage structures, tin-cone plates for storage struc- agriculture and other allied sectors and these women farmers
tures, and anticoagulants for rat control since the big-farm have more responsibilities, especially in the Indian agricul-
women can afford the costs of these storage methods. How- tural system. Rural women who are contributing to the rural
ever, the percentage of usage by small-farm women of metal economy should be supported by the government (Narayana
storage bins was high since these structures were given to the 2002). Hence, for a balanced and sustainable growth of rural
small-farm women only at the subsidized rate. India, women’s role in the developmental process and activi-
ties should be recognized and they should be adequately sup-
Suggestions for increasing the efficiency ported as were the men farmers to achieve gender democracy
of storing grains and guarantee food security.
The majority of the small-farm women (76%) suggested pro-
viding financial assistance on a credit basis for the purchase of
References
better storage structures such as metal bins. Also, 72% of the
farm women suggested having subsidized facilities for the same Gandhi NK. 1983. Stepping up rural warehousing. Kurukshetra
purpose and 34% of the respondents wanted special campaigns 31(22):18-20.
on storage structures (Table 2) since the percentage of sugges- LEISA. 2002. Women in agriculture. Magazine on Low External
tions was more than with big-farm women. On the other hand, Input Sustainable Agriculture (LEISA) India 4(4):4-5.
Mishra B. 2004. Exploring new opportunities: survey of Indian ag-
the big-farm women suggested an adequate (70%) and timely
riculture. The Hindu. p 29-31.
Session 10: Postharvest technology for efficient processing and distribution of rice 321
Table 2. Distribution of respondents according to their suggestions to over-
come problems in adoption.
aMultiple responses.
Wrap-up of Session 10
Session 10 included a wide range of interests from conditioning 2. Postharvest technology can improve production effi-
or primary processing of rice to safety and environmental mat- ciency more easily than other preharvest measures such
ters. as the development of new arable land or a new variety.
Professor P. Chattopadhyay from India reported on the 3. Therefore, more emphasis should be given to postharvest
changing scenario in postharvest technology in India, whose rice technology and related matters to achieve well-balanced
production and economy are growing rapidly. From Japan, the production after examining productivity, quality, safety,
latest technological information on drying, milling, and storage economic, and environmental issues.
facilities was introduced. Relatively new topics such as the use Two outstanding posters were chosen: “Postharvest of rice:
of by-products, residual agrochemical inspection, life-cycle in- the role of farm women in storing grains by different storage prac-
ventory analysis, and so forth were also presented. To meet these tices” by P. Sumathi from India and “The role of proteins in tex-
global needs, the importance of new concepts for rice breeding tural changes of cooked rice during rice storage” by T. Ohno from
was pointed out from the U.S. R.G. Manalili from the Philippines Japan. The first dealt with women’s labor in postharvest prac-
mentioned that the infrastructure surrounding rice production was tices during the storage of rice in rural areas of India. It was
essential for improving international competitiveness in rice pro- understood that the role of women’s labor would strongly affect
duction as well as in technological renovations. the future improvement of postharvest practices in developing
Session participants reached the following conclusions: countries.
1. The covering area of postharvest or postproduction tech-
nology is becoming wider. This includes safety, envi-
ronmental, and socioeconomic issues.
Approximately 54% of the world’s population is thought to Characteristics of paddy irrigation in monsoon Asia
live in the regions known as humid Asia or the Asian monsoon The principal grain crop in monsoon Asia is rice. In fact, nearly
region, which constitute only approximately 14% of the world’s 90% of the world’s rice is produced in the countries of this
land area. The majority of Asia’s massive population is sup- region. Monsoon Asia could be called a virtually homogeneous
ported by intensive paddy rice cultivation, which originates region in that paddy rice cultivation extends over almost the
from the warm and humid environment and offers high land whole of its area. Generally speaking, regions with annual pre-
productivity because of irrigation. cipitation of less than 400 mm are classified as perennial irri-
Paddy rice cultivation in the Asian monsoon region is gation zones, those with precipitation from 400 to 1,000 mm
not only an excellent form of agriculture offering high land as unstable irrigation zones, and those with more than 1,000
productivity; it is also seen as a sustainable and environmen- mm as replenishment irrigation zones. Of these, the whole of
tally friendly economic activity that suits the climatic and to- monsoon Asia belongs to the last classification. What is more,
pographical conditions of this region. This economic activity the Asian monsoon region is also characterized by large sea-
has continued to evolve for thousands of years across various sonal and short-term fluctuations in the supply of water re-
regions, as witnessed by the archaeological traces of 7,000- sources, as evident in the distinct dry and rainy seasons.
year-old rice cultivation obtained in China. Even today, paddy
rice cultivation is a unique way of life supported by the en-
deavors of people living in harmony with water. Moreover,
Paddy irrigation in arid and semiarid regions
paddies also make life more convenient for city dwellers by Overview
reducing floods, fostering groundwater, etc. These “products” As characteristics of paddy irrigation in the world’s arid and
cannot be sold in any market, and their functions have been semiarid regions, cases from the United States and Australia
referred to as “the multifunctional roles of paddy irrigation” are discussed. In the U.S., paddy zones extend along the Mis-
in this study. sissippi River and in five southern states as well as California.
In this paper, the multifunctional roles of paddy irriga- The rice cultivated area changes from year to year, from
tion are illustrated by comparing the hydrological environ- 300,000 to 1,500,000 ha. In Australia, about 100,000 ha of
ments, forms of irrigation practiced, and characteristics of pad- paddy fields are concentrated in the Murray and Murrumbidgee
dies in humid regions with those in the arid and semiarid re- River basins. With annual precipitation of about 400 mm, paddy
gions. Further, the paper examines the quantitative or economic irrigation is practiced there.
methods for evaluating these multifunctional roles and formu-
lates policy proposals for Japan and the rest of the world with Examples
a view toward solving water-related problems. The state of California in the U.S. is a semiarid region with
rainfall in winter but little in summer. The average annual pre-
cipitation is 575 mm, and agriculture there would be impos-
Characteristics of paddy irrigation in monsoon Asia sible without irrigation facilities. Paddies and water-supply
Definition of Monsoon Asia and characteristics channels are used as retarding basins and diversion channels
of climatologic and hydrologic environment for floodwater. Paddies also provide sanctuaries for migrating
Musiake (2001) asserts that regions such as monsoon Asia birds as well as for duck hunting. Although some features of
should be characterized by the region’s uniqueness (including paddy rice culture in California are similar to those of mon-
the nature of human activity), in addition to the climatic con- soon Asia, the California paddies and water distribution and
ditions that have been emphasized in the past. Thus, this paper the drainage system do not have the all-encompassing multi-
follows his new classification. The Asia monsoon region em- functional roles of paddy irrigation in monsoon Asia. In the
braces the Indian Ocean to the south, the expansive region of rice cultivation zones along the Murray River in Australia,
Tibet, the Himalayan mountain mass and continental China to meanwhile, some floodwater is stored during high flows. How-
the north, and the Pacific Ocean to the east. Most of it consists ever, paddies are not seen as having a function for relief from
of high-precipitation warm regions that have annual rainfall in flooding. Furthermore, groundwater percolation from paddies
excess of 1,500 mm, influenced by low pressure and monsoons causes an accumulation of soil salinity in surrounding nonrice
accompanied by westerly winds. Meanwhile, the water bal- arable land because the shallow groundwater is saline. In terms
ance (calculated by subtracting annual potential evapotranspi- of the sustainability of agriculture, this could be seen as an
ration from annual precipitation) generally exceeds 500 mm. example of a negative impact being exerted on a region’s wa-
ter resources.
Catherine typhoon
S0 Irrigation period
Status quo
200 Fukuoka headwork
Category
L B1 Block by gates
and drops
150
1/100-y return period
U1 D 3
Characteristics
1/50-y RP
Main irrigation
100 of channel unit
U1 I3 Smv A
channel 1
1/10-y RP
Ecosystem
Smv A evaluation by
50 hydrologic
5 4 3 2 1 0 indices
Evaluated points
0 G
0 2,000 4,000 D0 C
Gs A Ecosystem in
Drainage (conveying) capacity, D (m3 s1) plateaus
St.2
Fig. 1. Drainage capacity (urban area) and
storage capacity of paddies (Kinu River),
(Masumoto 2002). G C U5 I3
C A
U2 D3
Methods of quantifying the multifunctional roles Kokai
River Branch
of paddy irrigation Smv A channel
Main irrigation
channel 2
In 2001, the Science Council of Japan delivered a report on A A
Branch
(Nakadori River)
the multifunctional roles of agriculture and forests to the Min-
Main irrigation
irrigation
Drainage river
channel 3
ister of Agriculture, Forestry, and Fisheries. The report included channel
The benefits brought about by the external economy intrinsic Introduction of external economic evaluation
to the economic activity of agriculture are public properties to (basin management) and sharing of water
be enjoyed by all the people of a nation. Methods applied to management costs
evaluate the scale of these benefits in economic terms are the Studies should be made from the perspective of the basin as a
substitution method, CVM (contingent valuation method), the whole, rather than just being aimed at the direct beneficiaries
hedonic method, the travel cost method, and the direct method. to agriculture. If this, a case of external economic evaluation,
When evaluation is difficult to achieve by these methods, the were studied in terms of the basin as a whole, it could also be
tendency has been for the CVM and other methods to be considered from the perspective of the internal economy.
adopted. Numerous facilities have been developed in agricultural
So far, the multifunctional roles of paddies, upland fields, and rural development projects to date. In the past, facilities
and rural areas throughout Japan were evaluated and their re- for water management in regions and basins mainly concerned
sults produced values ranging from 4.1 to 11.9 trillion yen for with agriculture used to be developed upon applications from
functions that can be quantified. The following functions, not agricultural concerns. The development was mainly undertaken
included in these calculation results, are calculated through by farmers and land improvement districts, which also bore
economic evaluation. the maintenance and management costs. However, urbaniza-
As a method of economic evaluation for flood preven- tion has opened its way to all parts of the country in recent
tion functions, attempts have been made to evaluate the flood years. In many cases, farmers also bear the costs for operating
storage capability of a paddy area by substituting it with the and maintaining district drainage and drainage channels in
volume of a retarding basin. Using a method based on ques- suburban agricultural zones. These costs should ordinarily be
tionnaire surveys (stated choice method), meanwhile, meth- borne by urban residents. In other words, there should be a
ods of analysis are moving toward greater refinement by in- sharing of management costs commensurate with the manifes-
creasing the targeted range (functions), and there is a progres- tation of multifunctional roles of paddy irrigation, partly as a
sive shift from CVM toward choice experiments (choice-based consideration for providing comfort to local residents in ur-
conjoint analysis: contingent ranking method). With choice banized areas. Legal support systems also need to be devel-
experiments, however, several situations are evaluated for com- oped to address this.
parison.
Basin management during excessive flooding
In paddy-dominant basins where pump drainage is practiced
Policy proposals for Japan and the international community in low-lying areas, flood plains for its connecting main river
Development, economic activity (e.g., agriculture), are designed for the capacities on a 100-year probability scale.
and environmental preservation The maximum drainage volume for agricultural facilities is
Agriculture, of course, is a type of economic activity, but the normally assumed on a 10-year probability scale (in some re-
water supply and drainage systems that exist in rural regions cent cases, 20–30 years). Thus, runoff that exceeds the drain-
could also help protect a region’s ecosystems, for example, by age capacity on a scale of 10 years cannot drain outside the
providing nursery grounds for fish in connecting rivers. Many basin, and is forcibly stored in drainage channels and/or pad-
examples are demonstrating that agricultural development is dies. In other words, farmlands that include agricultural drain-
not necessarily in opposition to environmental preservation. age channels and paddies have the function of storing flood-
Rather, we would assert that several points still require im- water for drainage rivers, and, as a result, could be said to help
provement in terms of protecting ecosystems. The reason for reduce the risk of flooding in main rivers downstream where
this is that drainage channels fed by urban drainage continue urban areas extend.
to flow throughout the year, while water supply channels have
no flow of water outside the irrigation period.
References
Sustainable water-recycling agriculture Masumoto T. 1998. Paradigm shift in the evaluation of water stor-
The difference between paddy field agriculture in arid regions age function of paddies and in watershed management. J. Jpn.
and that in monsoon Asia needs to be clarified. Namely, there Soc. Hydrol. Water Res. 11(7):711-722. (In Japanese with
are huge differences in paddy irrigation between those regions. English abstract.)
Masumoto T. 2003a. Flood prevention function of paddies in mon-
For paddy fields in the arid and/or semiarid regions, excessive
soon Asia. Proceedings of Japan/OECD Expert Meeting on
irrigation to cope with arid climates causes salt damage due to
Agriculture and Land Conservation: Developing Indicators
a rise in groundwater levels and this in turns leads to external for Policy Analysis, Kyoto, Japan, May 2003. p 91-98.
diseconomy. In Japan and other highly humid zones, on the
other hand, the supply of groundwater from paddy fields can
Paddy cultivation forms the basis of traditional Southeast Asian other cases of externally imposed agricultural strategies, is that
societies and the livelihoods of the people who comprise those traditional farming systems are often highly evolved socio-
societies. Historically speaking, paddy cultivation has always technical systems, and the introduction of any improvements
(at least for several millennia) been multifunctional—provid- must be attempted very carefully, taking into account the po-
ing not only the raw material for subsistence and trade but also tential cascade of multiplier effects from seemingly small
serving as the central focus for family and community life as changes.
well as spiritual and religious expression. While times have Respect for the technical sophistication of traditional
certainly changed, this paper suggests that the multifunctional farming systems has now become an accepted tenet within the
nature of paddy cultivation continues to be important, and that current paradigm of agricultural development. Indigenous tech-
our concept of rural “livelihood” should incorporate these cul- nical knowledge even has its own acronym—ITK. And the
tural dimensions. concept of rural livelihoods goes even further in capturing the
complex interactions of the farming system with the larger
economic context of handicrafts, wage labor, remittances from
Traditional versus modern paddy cultivation abroad, etc. The legitimacy that the development profession
Twentieth-century economic developments changed the well- has grudgingly accorded to the economic strategies of rural
integrated world of paddy farmers and imposed new pressures farm families, however, has not been extended to the overall
while offering new options from modern agriculture. In Bali, cultural “lifeways” of the people themselves. Traditional be-
for example, farmers no longer had to follow the traditional liefs in river spirits and rice gods are dismissed as supersti-
cropping calendar imposed by the temple priests. Thanks to tions. Local customs about diet, nutrition, and health are viewed
new crop varieties from the International Rice Research Insti- with suspicion, if not ridicule. And traditional values referring
tute (IRRI) and new infrastructure from the Asian Develop- to time and money are met with exasperation by Western ad-
ment Bank (ADB)-financed Bali Irrigation Project, farmers visers preaching the virtues of market-based capitalism.
were encouraged to adopt modern methods that treat farming
more as agri-business rather than agri-culture. The conventional model
The Balinese case has become a symbol of the dangers According to the conventional model of development, tradi-
of ignorantly tinkering with a carefully evolved agricultural tional family and community-based modes of paddy cultiva-
system. The disruption of coordinated planting times allowed tion are doomed to an eventual demise in the face of competi-
insect populations to spike, causing severe crop losses (Lan- tion from industrial-style producers. While many observers of
sing 1991). The project-imposed diversion weirs were incom- economic development—myself included—are uncomfortable
patible with customary water rights, resulting in conflicts among with the social and cultural costs of economic change, the pro-
farmers (ADB 1998). The lesson from Bali, and from many cess is generally accepted as both necessary and inevitable.
Rice is the most important staple crop for many people in Asia, of agriculture, especially irrigated rice culture, from a rather
especially in East and Southeast Asia. It has a long history of different point of view.
cultivation, probably for more than seven millennia. Its culti- Irrigated rice culture getting water from an irrigation
vation plays a significant role in Asia’s customs and religions. system can be viewed to possess several interrelated elements.
In the year 2000, the world rice harvest area was around 152.0 Sustainable irrigated rice culture should be perceived here as
million hectares. Of this, Asia contributed nearly 90%. Global a prolonged existence and functioning of several important
rice production was 597.2 million tons of rough rice and Asia interrelated elements of irrigated rice culture. In other words,
as a whole produced more than 91% (Kim 2001). sustainability of irrigated rice culture should encompass the
For many countries, rice is not merely an economic good; sustainability of the following: (1) the irrigators’ association
it is also a social, cultural, religious, and even political good. or institution (institutional sustainability), (2) irrigation net-
A rice shortage could create national instability. Moreover, the works (technical sustainability), (3) agricultural production
world rice market is highly volatile and relying upon such an (economic sustainability), (4) irrigated land ecosystem (eco-
unstable rice market could be quite risky for national food se- logical sustainability), and (5) socio-cultural values linked to
curity. No wonder that many countries still insist on adopting rice cultivation (socio-cultural sustainability).
a rice self-sufficiency policy. This implies that preserving rice However, the sustainability of those five elements of ir-
culture seems to be of primary importance. rigated rice culture also largely depends on the local natural
This paper tries to (1) propose a concept of sustainable environment, especially the upstream watershed and the qual-
irrigated rice culture, (2) raise arguments for preserving irri- ity of river water upstream (environmental sustainability),
gated rice culture, and (3) recommend relevant policy mea- which are external to the irrigated rice culture concerned. An
sures to preserve irrigated rice culture. individual irrigation system for rice cultivation is also an ele-
ment or subsystem of the entire system along a river course.
This implies that the performance of the individual irrigation
Proposed concept of sustainable irrigated rice culture system or irrigated paddy culture getting water from a lower
There are a lot of definitions or concepts of sustainable agri- stream may also be affected by the system getting water from
culture. Pakpahan (1995) has cited several definitions of sus- the upper stream. The sustainability of the local environment
tainable agriculture but those will not be repeated here. Suf- of the system along the water course as well as the individual
fice it to say that most of the definitions seem to merely em- irrigation system is also affected by various external forces
phasize more the environmental and agro-technological as- such as demographic, social, economic, cultural, political, in-
pects. Since land and irrigation water are indispensable for dustrial development, tourism, government policy, etc.
irrigated agriculture, it seems necessary to see sustainability
As a developing country, China has made outstanding progress Although huge changes in agriculture with the development of
in economic growth during the past 55 years. Especially in the a social economy and a population peak of 1.6 billion are ex-
last two decades, China has emerged as one of the fastest-grow- pected in the next two decades, China faces increasing food
ing economies in the world (Anderson and Peng 1998). From demand and limited arable land and water resources (Liu et al
1949 to 2003, grain production in China increased from 113.2 2001). From 1996 to 2003, Chinese total cereal production
million to 430.7 million tons and peaked at 504.5 million tons declined by 14.6%. The main reasons influencing this decline
in 1996. Meanwhile, the Chinese population has increased from in Chinese cereal production are attributed to a shift in peas-
0.4 billion to 1.3 billion. China feeds 25% of the world’s people ants’ willingness to cultivate grain crops instead of other eco-
with 7% of the total cultivated land in the world. nomic crops, a price fluctuation in the international grain mar-
However, since 1998, Chinese total rice production has ket, and the limited land and water. China affects the interna-
declined continuously, from 200 million to 170 million tons. tional grain market because of its huge population and
food crop rice are to enlarge the planted area and increase Gate
yield by improving arable land with affluent water resources.
This environment provides the water-loving rice plant with the Channel
fundamental conditions necessary to achieve a high yield when Fishpond
combined with crossbreeding and gene-mapping technologies.
More attractive and efficient ways to use water are now cru-
cial for farmers in order to encourage them to plant rice be-
cause rapid economic growth in China has led to high compe- Gate
tition among water users and requires the introduction of ef-
fective systems for water charging. Farmers need to wisely
develop multifunctional uses of paddy fields and water to har-
vest more rice based on hydrologic conditions. Economic ex-
ternalities generated by the multifunctional uses of paddy fields
and water should be incorporated when the efficiency and
sustainability of water use is addressed in rice paddy agricul- Rice field
ture in the Asian monsoon region (Yamaoka et al 2004).
This paper emphasizes the multifunctional roles of paddy
fields and irrigation water, and introduces the multifunctional
social roles of wise water management in the Dujiangyan irri-
gation area.
Fig. 1. Fishery beside the channel.
Rice is the staple food and main agricultural commodity of Sri from Pakistan, Vietnam, or Thailand, and tended to abandon
Lanka, where 17.7% of the gross domestic product comes from rice farming. Table 1 shows rice imports by quantity and as a
the agricultural sector. Highly water-intensive rice cultivation percentage of total national consumption.
consumes more than 70% of the total water allocated for food The inevitable consequence of this pressure is that most
production in the country. of the traditional Wet Zone farmers are pushed toward selling
Paddy lands in Sri Lanka can be categorized into three their paddy lands in lucrative deals for real estate develop-
major types depending on the water usage: ment. This has significantly decreased the total area of rainfed
O Major and medium irrigated systems—264,602 ha in paddy cultivation schemes during recent times (see Fig. 1).
1999-2000, 50% of the total area sown From the statistics, it is apparent that the percentage of
O Minor irrigated systems—129,604 ha in 1999-2000, rainfed paddy fields, which was as high as 45% of the total
23% of the total area sown cultivated area in the early 1960s, had dropped to about 28%
O Rainfed systems—155,040 ha in 1999-2000, 28% of by 2000.
the total area sown
Out of this, the rainfed systems are mainly found in the
The impact of the loss of rainfed paddy land
Wet Zone of the country. The Wet Zone is a climatic subdivi-
on national food security
sion in the southwestern part of the country, which gets up to
2,000–2,500 mm of annual rainfall. It constitutes about one- A study of rice imports in the past few years showed a signifi-
third of the country’s total area and accommodates about 63% cant growth of rice imports in 1996-2000. This could be mainly
of the population. The major part of heavily urbanized areas, attributed to the severe shortfall of rain that the country had to
including the cities of Colombo, Kandy, and Galle, is situated face intermittently during this period. Importing water-inten-
in the Wet Zone. Because of the high population density and sive goods to a country where the water stress is high from
high level of urbanization, demand is overwhelming for land another country where the water stress is comparatively less is
in the Wet Zone. The population density in Colombo District a form of virtual water trade, and it is a practice encouraged
is 3,305 persons per km2 versus 957 persons per km2 in the by most of the modern water ideologues since the Dublin-Rio
Wet Zone. In the Dry Zone, this decreases to 163 persons per conference onward.
km2. Competition for land is therefore strong and, because of But, unfortunately in the Sri Lankan situation, most paddy
the availability of other employment, labor costs are increas- lands affected by a shortage of rain are in the irrigated systems
ingly high. in the Dry Zone, whereas water stress in the Wet Zone is much
In recent years, rice or paddy farmers (as they are termed less even during drought; therefore, the rainfed systems in the
in Sri Lanka) were either attracted by more lucrative alterna- Wet Zone could be cultivated to a certain extent even during
tives or demotivated by the competition from rice imported relatively dry years. Because of the loss of paddy cultivation
Year
Item
1994 1995 1996 1997 1998 1999 2000 2001 2002
Rice imports in million SLR 655 122 5,118 4,331 2,621 3,290 288 969 1,732
Imports as a percentage 2 0.3 16.5 13.7 6.2 7.5 0.5 2.0 3.3
of total local production
Source: www.statistics.gov.lk.
150,000
100,000
50,000
1993-94
1983-84
1989-90
1981-82
1985-86
1987-88
1991-92
1999-2000
1963-64
1969-70
1971-72
1973-74
1975-76
1977-78
1979-80
1995-96
1997-98
1965-66
1967-68
1961-62
Years
Fig. 1. Variation in total sown area in the rainfed system from 1961 to 2002.
2
4
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
Year
Fig. 2. External debt and trade balance.
in the Wet Zone, Sri Lanka is pushed toward dependency on tempted by high land prices to decommission their paddy land,
imported rice, hence the import of virtual water, placing a heavy thus creating a vicious circle (Fig. 2).
burden on the already precarious balance of payments while a
significant amount of real water is being lost as runoff because The impact of the loss of rainfed paddy land
of the limitation of cultivable land in the Wet Zone. Therefore, on the water sector
this is more akin to a trade of virtual land than virtual water. In From the above section, it is clear how the decommissioning
this process, not only is there a loss of water, but also food of rainfed paddy lands contributes to wastage of a significant
security and the trade balance of the country are at stake. Be- volume of water that could have been used for cultivation. In
cause of the competition from cheap rice imports, small-scale all the recent environmental resolutions from Dublin-Rio to
farmers in the rainfed system are further discouraged and are Rio+10 conferences, it has been stipulated that uses of water
As one of the multifunctions of irrigated rice cultivation, bio- In this presentation, some typical studies, which we or
logical diversity in rice paddies has been a critical issue in other pioneers have conducted, will be reviewed from the view-
Japan and some Asian countries under free marketing of the point of agroecological principles.
WTO. Especially, in Japan, conservation of biodiversity and
ecosystems in agriculture has been an important issue in agri-
Case studies
cultural and environmental policy since 1999, when a changed
agricultural foundation law stimulated progress by the Japa- Examining natural history
nese government and EU. Agro-biodiversity has now become More than 200 plant species, 20 species of amphibians and
a popular password in many nations, and the Internet now has reptiles, 30 species of fish and birds, and more than 1,000 spe-
30,000 articles on it. cies of arthropods can be reviewed as apparent agro-
According to the literature before, about 2,000 species biodiversity in rice paddies (Hidaka 1998). A few collecting
of plants and animals associated with rice paddy fields have studies were conducted with scientifically appropriate meth-
been recorded in Japan (Hidaka 1998). This biodiversity in ods of sampling. Kobayashi et al (1973) examined in detail
species richness can be regarded as a general characteristic in the biodiversity of arthropods using sweeping methods in rice
the rice paddy ecosystem. Actually, 645 animal species were paddies of several sites of Tokushima Prefecture in the late
recorded in the Philippines (Cohen et al 1994) and 765 spe- 1950s. More than 450 species of insects, spiders, and mites
cies of arthropods were examined in Indonesia (Settle et al could be recorded in spite of the methodological faults of
1996), for example. For rice, 212 species were recorded in the sweeping methods that can mainly collect arthropods only on
Philippines (Heong et al 1991). rice plants. Nowadays, if we had a chance to examine arthro-
We must not ignore, however, that there are actually two pod fauna in the same fields, it might be difficult to collect a
types of agro-biodiversity in rice paddies: one is “apparent closed number of species in rice paddies such as in the 1950s.
agro-biodiversity” and the other is “true agro-biodiversity.” An example of this examination for species number must be
Usually, we have been evaluating agro-biodiversity as a value significant and important to evaluate the agro-biodiversity level
for the number of species collected in arable land, but not all as a first step. But many species are stranger species in rice
species of plants and animals truly depend on the agricultural paddies, which the authors pointed out. We have to examine
ecosystem. In the case of rice paddy fields, some species may the fauna in other wetland ecosystems as well as rice paddies
be stranger species or may depend on other wetland habitats. to evaluate true agro-biodiversity. Recently, Saijo (2001, 2002)
There are many kinds in natural or managed ecosystems with tried to examine aquatic insect fauna in both rice paddies and
the exception of rice paddy fields: ponds, lakes, rivers, streams, irrigated ponds in Shimane Prefecture to make a clear repro-
and reservoirs in the limnological viewpoint. Furthermore, duction of each species site. These studies are significant in
species components of agro-biodiversity in rice paddies may the ecological sciences to reveal the dependency of each spe-
vary depending on the preserved situation of natural wetlands. cies on rice paddy ecosystems and to evaluate true agro-
biodiversity.
For botanists, Kasahara (1947) examined a list of rice
Approaches toward evaluating true agro-biodiversity paddy weed specimens in 25 prefectures over Japan and re-
To discriminate the true species depending on rice cultivation corded 174 plant species. All botanical species could be clas-
as true agro-biodiversity, there are three ways to examine this: sified as serious harmful weeds (96 species) and semiwild
natural history, review work, and field work in ecology and weeds (78 species) in rice paddy ecosystems.
systematics:
1. Review work in the literature (classical natural his- Habitat records of specimens
tory). Recording habitat when collecting specimens in natural his-
2. Collection of specimens attached to a living habitat tory work is important information for evaluating the agricul-
(new natural history) (Mineta, Hidaka, and Enomoto, tural dependency of each species as true-biodiversity. Although
in preparation). there are very few cases that recorded the habitat of collected
3. Ecological field work of each species population as- specimens, the same specimen wetland plants had details on
sociated with agricultural practices (agroecological the collecting place such as rice paddy, pond, irrigation way,
field work). or natural wetland. Some botanists had been recording habi-
tats, such as Kasahara’s collection and Fujii (2001). So many
A. humidum
rice paddies but also in ditches, ponds, or streams and rivers
L. indica subsp. and other natural wetlands should be examined in detail over
incnopnyiia several years. Many species may use several kinds of ecosys-
N. coreana
75 25 tems to round out their life cycles as their indispensable habi-
M. quadrifolia tats.
H. oliganthum N. indica
s
Irrig
fi
R. extorris G. japonica deyrollei (Hemiptera) and the diving beetle Cybister japonicus
atio
dy
R. cantoniensis A. japonica
pad
np
50 50 (Coleoptera), representative endangered aquatic insects, have
ond
M.korsakowii
e
V. undulata A. imbricata
s
R. nipponicus ricultural dependency on rice paddy fields since 2000. These
local populations of each species depend on irrigated rice fields
25
and agricultural practices for their reproduction, according to
75
results from our field work (Fig. 2). Furthermore, L. deyrollei
S. natans
populations depend on frogs, which strongly depend on the
P. chinense Blyxa spp.
U. minor rice paddy ecosystem as a reproductive habitat (Hirai and
Isoetes spp. Hidaka 2002). Their dependency on irrigated rice cultivation,
0 100
100 75 50 25 0 however, might not be permanent. It is necessary for their ag-
ricultural dependency to conduct similar field examinations in
Food plains in rivers
different places and different climatic conditions. Often, the
Fig. 1. Frequency of habitat records among endangered aquatic
weeds in Okayama Prefecture according to botanical specimen habitat use pattern of a species population has changed in a
collections (Mineta, Hidaka, and Enomoto, in preparation). fluctuating environment. For more adequate conservation man-
agement, it is necessary to conduct ecological field work that
data were accumulated so that dependency on rice paddies can evaluate the habitat use of each local population quantita-
could be demonstrated in weed species in irrigated rice culti- tively, such as ornithologists’ work (Elphick 2000, Fujioka et
vation (Fig. 1; Mineta, Hidaka, and Enomoto, in preparation). al 2001, Maeda 2001).
Constructing a database of specimen habitat records using a
GPS system is a key methodology for evaluating true agro-
Integrating conservation and agricultural production
biodiversity.
For endangered species in rural areas where rice has been cul-
Agroecological study of local populations tivated, some conservation practices are necessary immedi-
and communities ately to recover or sustain the population level so that it won’t
The most evidential and practical methods among these three become threatened. For other conservation practices, we have
approaches are agroecological field work at the population or to sustain agricultural production. Especially in the case of
community level for each species. For the community level, conserving agriculturally dependent species such as the giant
generally, examinations of an assembly of biological species water bug, maintaining agricultural activity is necessary in some
on grown rice plants such as Kobayashi et al (1973), Heong et hotspots of true agro-biodiversity. Therefore, integration of the
al (1991), and Une et al (1989) are correctly regarded as agro- conservation of biodiversity and agricultural production is
biodiversity. There is another assembly of aquatic animals fol- important and necessary. Recently, a new concept, IBM (inte-
lowing the irrigation system in rice paddies for agro- grated biodiversity management), has been proposed in Japan
biodiversity such as Katano (2001) in fish or Saijo (2002) in (Kiritani 2002). Practically, to manage biodiversity well in
insects. agroecosystems, finding and making devices that can buffer a
At the population level, it is difficult for all species, one conflict between biodiversity and food production are desir-
by one or gene by gene, to reveal their dependency on rice able for both hardware and software.
cultivation by analyzing basic data, which used to require tasks In our research field, some IBM trials have started to
such as genetic analysis of one local population. If we could manage an endangered population of the giant water bug asso-
make their dependency on paddy fields clear for a particular ciated with biodiversity and sustainable agriculture in several
species population, we could use the most practical tactics for practices of agricultural, educational, and scientific research.
the population in agricultural policy. Therefore, it is reason- For our scientific researchers in agroecology, quantitative
able to start from an emergent case such as endangered spe- analyses of population dynamics can produce optimal results
cies, which need to conserve their population soon in an envi- for the management of endangered species populations. Ana-
ronmental policy (Hidaka 1998). Recently, some ecological lyzing an endangered population by genetic methods at the
field workers have started detailed studies of population biol- molecular level, furthermore, must result in not only manage-
ogy for RDB animals and plants to demonstrate their habitat ment of the population but mitigation for its appropriate con-
use quantitatively. Some species, including endangered popu- servation. In addition to studies of an intrapopulation, impor-
MAR APR MAY JUN JUL AUG SEP OCT NOV DEC JAN
Larval development
Other wetland habitats
Emigration with drainage (ditches, ponds, rivers)
New adults
Irrigation
Water level of irrigation
start Dispersal
Emigration
Higher
Fig. 2. A schematic description of the life cycle of the giant water bug, Lethocerus deyrollei, rice
cultivation practices, and irrigation water level in a hotspot field of agro-biodiversity.
Paddy farming in the Asian monsoon area is closely connected shortage (Roongrueng et al 1996) and flooding (Pramote 1999)
to the rural social system. Paddy farming has important roles have to be examined in parallel.
and functions. Recently, energetic efforts have been made to Water management is mainly carried out under the au-
spread such recognition worldwide and to evaluate it quantita- thority of the Royal Irrigation Department (RID). To prevent
tively (Japanese National Committee of ICID 2003, MAFF water conflict, water allocation in the dry season from January
2003, JIID 2003). In this paper, we introduce the to June is very important. Hereafter, the period from January
multifunctionality of floating rice farming in the Chao Phraya to June is referred to as “WM (water management) dry sea-
delta of Thailand and its possibility for enhancement by water son,” while the period from July to December is referred to as
management. “WM wet season.”
Farming and water management in the Chao Phraya delta Multifunctionality of floating-rice farming
The Chao Phraya River basin has a catchment area of 162,000 Floating-rice farming in the delta has played important roles.
km2, including 1.34 million ha of low-lying delta area. The It can be summarized as having low input and low yield but
Greater Chao Phraya Project drastically changed farming and sustainable farming. Cultivated floating-rice area decreased
water management in the delta. Only one crop per year was from 228,000 ha in 1987 to 114,000 ha in 1997 (CTI et al
possible there before the project. The original purpose of the 1999). Floating rice can grow flexibly according to irregular
project was supplementary irrigation in the rainy season. Its increases in water level. It grows in a dry field with weeds at
purpose has now expanded to irrigation in the dry season and the beginning stage. Then, the stem increases in length from 2
domestic water supply, among others. The delta area has be- to 5 m according to water conditions. The cultivation calendar
come one of the largest rice granaries in the Asian monsoon matches the natural water conditions. It takes 7–9 months from
area. planting to harvest. The harvest starts in December or January
The present irrigable area in the delta is 1.08 million ha. after the stored water in the field is drained.
Rice cultivation for more than five times per two years is pos- In the past, people in the delta had to follow natural water
sible where water accessibility is good. New trials of rice cul- conditions. Recently, the water level and flow conditions of
tivation in the dry season started only about 30 years ago. The the main rivers and canals have been controlled to some de-
rice cultivation area increased rapidly, but soon reached its gree by the operation of diversion dams and regulators along
limit because of a shortage of water resources and insufficient both the main irrigation and drainage canals. A unit of floating
capacity of existing irrigation facilities. Farming and land use rice area is from a few thousand to 10,000 ha. Each unit has a
changed rapidly again in the last 15 years. Water management so-called “drainage box” that has a drainage regulator to con-
became more complicated than before because water demand trol the water level upstream. Newly constructed big roads
increased and competition among water users became more function as embankments. Therefore, the location of flooding
serious (Francois et al 1999, 2001). The problems of water areas is also partially regulated by water operation. Scattered
Method (how) Preliminary release of standing water Release of standing water considering the Introduction of new farming technology
from floating-rice area timing of water use downstream for dry-season cultivation
Conveyance of floodwater into floating- Change of drainage point to use water Drainage improvement
rice area resources more effectively Arrangement of irrigation facility
Insurance contract between government Arrangement of new facility such as New water allocation
and farmer for receiving surplus flood- pump, regulator, and regulating pond
water Modification of cultivation pattern in
Heightening of embankment downstream fields
Advanced operation of drainage regula- Development of a decision support sys-
tor tem
Development of decision support system
of infrastructure for information and
communication technology
Induced What Peak period of flood discharge Provision of new water resources New farm products
positive When Flooding period (October, November) From middle of December to end of Janu- From February to May
impact ary
Where Downstream delta, especially Bangkok Downstream fields Floating-rice field
Whole basin
For whom People in downstream delta Farmers of downstream fields Farmers
People in the basin Consumers
Induced negative im- Increase in risk in farming and harvest Increase in salinity in the Chao Phraya Deterioration of water quality and soil
pact or remarks yield River Increase in water demand
Decrease in water level in the down- New water conflict might occur
stream Chao Phraya River (it is some- Change in ecological environment
times unsuitable for navigation and in-
take for tap water)
The principal grain crop in the Lower Mekong River Basin derground water quality, and the relationships of various fac-
(LMB) is rice. In fact, nearly 90% of the world’s rice is pro- tors in paddy growing. Results at the farm level will be ex-
duced in the Asian countries including this region. It is com- trapolated to the subbasin (or catchments) and, ultimately, to
monly recognized that “paddy fields” have many functions and the whole basin. The expected outputs will provide useful in-
advantages or disadvantages but consume a large amount of formation for the optimum use of water to obtain higher paddy
water. The value of irrigation water encompasses more than yields under appropriate soil, crop, and water management.
the net economic returns of the rice produced. The multiple
values of irrigation are particularly evident in paddy cultiva-
Summary study of the DMPF
tion where water supports a sustainable and ecological water
cycle control, which is the basis for the characteristic socio- The total timeframe of the DMPF is proposed for 5 years. In
cultural system of Southeast Asia. the first and second year, activities will concentrate on data
Northeast Thailand has an area within the LMB of collection in the subbasin and experimental plot. A demon-
513,113 km2 and runoff showing as much fluctuation as 80– stration model for the multifunctionality of paddy fields will
85% in the wet season but very scarce in the dry season. Ac- be established in the third year and will be improved and ap-
cordingly, water control functions (flood prevention, ground- plied to a wider area in the fourth year. Finally, analysis and
water recharge, prevention of soil erosion) are essentially re- demonstration will be done.
quired to control flood and drought problems. Demonstration will be done through simulation of a logi-
In this connection, we need to know more precise infor- cal model and a small experimental study. Through the study
mation about the multifunctional roles and multipurpose uses for the demonstration of paddy fields in northeastern Thai-
of paddy-field irrigation. For this region, the Demonstration land, two types of models will be developed, at the field level
Program on Multifunctionality of Paddy Fields over the and at the subbasin level.
Mekong River Basin (DMPF) is developed by the Mekong
River Committee (MRC) member countries to collect data and Field-level model
information on the area that help to demonstrate the The field-level model shows mainly the inflow/outflow of water
multifunctionality of paddy fields in the LMB. in the field for agricultural land (paddy and upland). In addi-
Thailand has expected that the outputs of DMPF will tion, if possible, it is better to reflect some other aspects, such
provide useful information for the effective use of water to as field conditions (topography, soil, meteorology), farming
achieve high production in paddy fields under appropriate practices, and infrastructure (irrigation/drainage, roads).
water, soil, and crop management. The preliminary field-level model will be studied, based
The overall concept of the DMPF aims at promoting on the schematic model shown in Figure 1, and on other fac-
agriculture and irrigation development within the MRB in a tors. To evaluate the multifunctionality of paddy fields at the
sustainable and environmentally sound manner, and with good sub-/entire basin levels, it is necessary to consider the impact
participation and cooperation of all concerned. This will serve of upland crops. The model can be applied for both paddy and
as a practical and integrated approach for agricultural devel- upland by changing the coefficients of several factors.
opment at the farm, subbasin, and basin levels. Achievement
of this program will benefit local people, scientists, and man- Subbasin level
agers of relevant agencies. Local farmers would learn how to In this study, the subbasin is defined as one of the areas assum-
cultivate rice with an optimum or minimum use of water and ing a unit block from the viewpoints of conditions of hydrol-
how to manage the consequent impact from rice growing. Sci- ogy, topography, land use, water use, flooding, and environ-
entists would gain further insight into such aspects as the rice ment. It may include one or more rivers/tributaries of the
cultivation system and the possible impact of actual water use Mekong. The model at the subbasin level (Fig. 2) will be in
for paddy growing, the balance of water use, surface and un- conflict because of factors such as other water use (urban,
Irrigation
Bund Flood Soil erosion water
mitigation control
domestic, industrial), environmental impact on the area be- These data are available for designing the basic framework of
yond the field, flood damage, etc. In the case of including res- the subbasin model and for clarifying the present conditions
ervoirs in a subbasin, it is necessary to consider the operation of the subbasin. The entire LMB levels will be clarified by
of the reservoirs and conditions for hydropower generation. compiling the information on the subbasin.
To accommodate the model subbasin-wide, three options Natural conditions. The data on climate, topography,
are proposed for data collection. The subbasin consisting of geology, hydrogeology, soil, hydrology, drainage, and flood
rice ecosystems as designed will be selected. These ecosys- are selected for the natural conditions in the subbasin. These
tems are rainfed and the irrigated wet and dry seasons. Within data are basic calculation data of the entire LMB and the
the selected subbasin, a farmer-field experiment will be con- subbasin model and actual observation data will be used for
ducted to collect data as needed. calibration and improvement of both models.
Rice ecosystem. The data on rice ecosystem, paddy area,
soil moisture, rice crop, and agro-economy are selected for
Proposed data to be collected the rice ecosystem of the subbasin. These data will be used
Subbasin level mainly for analysis of direct impacts of the multifunctionality
Subbasin conditions. The data on subbasin boundary, admin- of paddy fields.
istrative conditions, infrastructure, irrigation, and land use are Other conditions. The data related to socioeconomics,
selected for basic items on the respective subbasin conditions. water use, community, and environment are selected for the
To meet the demand for high-quality green tea products, Japa- Table 1. Water quality in rice terraces farmed at Kurasawa,
nese tea farmers have applied excess nitrogen fertilizer (over Shizuoka.
1 t N ha–1 y–1) for a long time. As a result, heavy groundwater
Sampling point pH Nitrate Na Al Mg
pollution by nitrate occurred in tea areas. Nitrate pollution is
associated with strong acidity because of severe soil acidifica- Rice terraces A (in) 4.44 33.5 2.2 17.5
tion in tea fields. Therefore, we evaluated the bioremediation Rice terraces A (out) 6.09 23.1 0.4 15.8
effect of traditional small-scale rice terraces on effluent water Rice terraces B (in) 5.44 29.3 0.5 16.9
Rice terraces B (out) 6.85 13.8 0.5 14.8
from tea fields. Many small rice terraces have been used by
Original groundwater 4.28 27.2 5.7 17.4
farmers for a long time to cultivate rice for their home con-
sumption near tea plantation areas in the Makinohara plateau aNitrate,aluminum, and magnesium concentrations are in ppm. Data were
of Shizuoka, Japan. The water remediation effect of those rice gathered in mid-summer 2000. Rice terrace A is about 2,500 m2 with 7
terraces, including slopes, and rice terrace B is about 5,000 m2, with 13
terraces on strong acidic and nitrate-rich effluent water from terraces. The rice terraces were continuously irrigated. The original ground-
tea areas was investigated by a field survey and a field experi- water was measured at a water spring on the upper slope of Makinohara
ment. plateau.
RDB plants
extinction. There are 191 species of 43 families, including
natural wetland plants with origin in paddy fields, in the entire Creation of positional information
region (Kasahara 1954). Now, 49 species of 25 families in (Low accuracy) (High accuracy)
paddy fields are registered with the Red Data Book (RDB) in
the Kanto region as endangered aquatic plants. These plants Meshing Pointing
fields to ensure a sustainable rice-cropping system in the world. Extraction of hot spots
Fig. 2. Land use of areas around which RDB plants appeared. *Among plants in each type,
χ 2 test, χ 2 = 161.74, P<0.01).
significant differences were found as to their habitat (χ
sin, the Kinu River Basin, and the foothills of Mt. Torinoko. in Tochigi Prefecture in 1997 decreased more than 10% com-
Three or more species of threatened plants were collected on pared with that of 1991, the ratio of land use on grids with
25 grids, which were regarded as hot spots of the RDB plants RDB plants did not change greatly over 20 years, except for
in a rural environment. Such areas are regarded as preserva- rivers or a lake environment. On grids considered to be hot
tion areas and they should be carefully monitored when agri- spots with more than three endangered species, forests have
cultural or rural development projects are carried out. decreased more than 30% during a decade. Therefore, although
development of rivers, lakes, and forests has occurred on a lot
of grids, which confirmed RDB plants, it was reasoned that
Characteristics of the habitat of RDB plants endangered species that had lost endemic habitat used stable
To study the characteristics of the habitat of each RDB plant, rice paddy fields as refugia.
based on Digital National Land Information, we analyzed the From calculations of the land-use ratio of each RDB
land use on a standard grid square where a threatened plant plant, the environment of areas with confirmed RDB plants is
was found. Although the ratio of paddy fields on the total grids categorized into three types (Fig. 2):
Wrap-up of Session 11
Session 11 featured discussion on the various values and roles is most important to increase real water-use efficiency.)
in our lives as a whole fulfilled by rice cultivation with paddy fields 4. Ample agricultural water in humid regions generates vast
and irrigation under the designated theme. A welcome was given externalities that cannot be priced through market mechanisms.
by H. Satoh, president of the National Institute for Rural Engi- 5. Significance of the externalities has been maintained
neering (NIRE), K. Miyamaoto, executive director of NIRE, and Y. but has shifted its weight from benefits to farmers to those for
Tsutsui, convener. The session, with about 200 participants from the public (town dwellers) according to the socioeconomic devel-
various countries, discussed the subject of inquiry raised by Pro- opment of the country.
fessor R. Nakamura in his earlier keynote speech. K. Yamaoka, 6. Policy implications for publicly facilitating sustainable
co-convener, proposed recalling the following arguments summed irrigated rice paddy agriculture in restoring a natural balance should
up by Professor Nakamura. be further promoted in various development stages of countries
1. Water is indispensable to rice cultivation. Sustainable in the Asian monsoon region.
irrigated rice paddy agriculture is crucial for developing humans’ K. Yamaoka also presented the background and purpose
future. of the session. It was expected to feature discussions to endorse
2. Increasing water-use efficiency, namely, water produc- the arguments and to sum up the main points of the arguments
tivity, is in fashion and seems to be our universal mission. to be contributed to discussions at the OECD, the Fourth World
3. However, we have to note that in humid regions it is Water Forum, and relevant international conferences. Eight con-
rational to use ample water to substitute for farmers’ labor, tributors made presentations from various interdisciplinary as-
whereas in arid regions constantly valuable water cannot be re- pects.
placed by other resources. (In humid regions, developing appro- The first contributor, T. Masumoto from Japan, stated that
priate water resource allocation systems for occasional dry spells the geographical, climatological, and hydrological characteristics
The world’s annual rice production reached 589 million tons Table 1. Importance of alluvial land (in 106 ha).
in 2003 (FAO 2004). It exceeded that of wheat and amounted
Land area Alluvial soil area
to 28.4% of the world’s total cereal output. Maize production Region
has recently surpassed that of rice, but its increase is mainly to Total Potentially Total Potentially
meet the demand for animal feed and not for immediate hu- arable arable
man consumption. Thus, rice as the staple for the human diet
is becoming increasingly important in Asia and in some parts World 13,000 3,152 588 316
Asiaa 2,704 620 – 192
of Africa and the Americas, feeding nearly half of the world’s Tropics 4,893 1,652 365 172
population today. Tropical Asia 987 344 168 114
According to IRRI (2002), some 11% of the total area
aExcluding
the former USSR.
cultivated for rice worldwide is in uplands, the rest being in
Source: White House (1967).
wetlands, either as rainfed, deepwater, or irrigated paddies.
As upland rice is generally much less productive, rice produc-
tion in paddy soils should definitely exceed 90% or even 95% Characteristics of paddy soils
of the total global output. This justifies my confining this dis-
cussion to only paddy soils. Any soil that is used for growing aquatic rice can be called a
paddy soil. Paddy soil in this definition is related directly to
land use, but not to any particular type of soil in a pedological
Natural settings of paddy lands
sense. Paddy soils occur basically in lands with an aquic (note:
There are some 135 million ha of paddy lands in the world, of this is a technical term used in U.S. Soil Taxonomy and widely
which 126 million ha, or 93%, are in monsoon Asia, or in hu- recognized as a term to denote hydromorphic conditions) mois-
mid East, Southeast, and South Asia (IRRI 2002). Climate and ture regime, in floodplains, deltaic plains, fans, and terraces.
landforms are the two determinants for this strongly biased Fluvisols, Gleysols, and Cambisols are the most common ped-
distribution of paddy lands (Kyuma 2004). Monsoon climate ological members of paddy soils, but some other members,
is characterized by a yearly inversion of wind direction and such as Acrisols and Luvisols, occur among paddy soils on
this accompanies alternating rainy and dry seasons. Therefore, older land surfaces. Eswaran et al (2001) also listed all the soil
rainfall is concentrated during the rainy season, often bringing orders of soil taxonomy except Gelisols as soils used for aquatic
more than 1,000 mm of rain in less than half a year, thus en- rice cultivation, but they also indicated that Inceptisols and
abling cultivation of a rice crop. In terms of landforms, mon- Ultisols are the most frequently used.
soon Asia features an exceptional abundance of lowlands (Table Irrespective of their pedological nature, all paddy soils
1). Being in the region of active orogeny and volcanism, in are submerged for at least a few months a year, either naturally
combination with high rainfall, monsoon Asia undergoes se- or artificially. Managing paddy soils under water is entirely
vere erosion in its high mountainous lands and deposition in different from managing other soils used for upland crops, and
its riparian and coastal lowlands, producing extensive flood- this produces important differences, particularly in chemical
plains and deltas in the middle and lower reaches of gigantic and biochemical or microbiological processes. However, paddy
rivers such as the Yangtze, Mekong, Brahmaputra, and Ganges. soils are also kept drained for the rest of the year, again natu-
These lowlands are naturally inundated with monsoon rains rally or artificially. This cyclic change in micro-environmen-
during the rainy season. tal conditions exhibits properties not encountered in other soils
As depicted above, monsoon Asia, with its high seasonal and differentiates paddy soils from most other soil systems.
concentration of rainfall and its exceptionally extensive area The most important change upon submerging a soil is a
of lowlands, provides Oryza sativa, a plant species native to decrease in oxygen partial pressure, which accompanies a low-
the region, with the most adapted natural habitat. Thus, rice ering of redox potential (Eh), reduction of chemical species
culture originated as an adaptation to the given natural set- [O2→H 2O, NO 3–→N 2, Mn(IV)→Mn(II), Fe(III)→Fe(II),
tings of monsoon Asia. Later expansion of paddy lands has SO42–→S2–, CO2→CH4], elevation of soil pH, and changes in
been made possible mainly by the provision of irrigation. biotic composition (aerobic to anaerobic) (Table 2). All these
A small share of indigenous rice culture with O. changes further induce enhanced nitrogen availability and its
glaberrima is known to exist in West Africa, but it has been fixation, solubilization of soil phosphates, and so forth.
narrowly confined to its place of origin throughout history, for Draining a paddy soil reverses most of the above pro-
the climate and landforms of the region are not favorable for cesses: Eh rises, chemical species are oxidized, soil pH de-
its propagation. creases, and aerobic soil biota recover. However, some of the
Weakly reduced
Redox state
every cycle of submergence and drainage, impoverishing soil
Moderately
basic cations and leading to Al-interlayering and eventual de-
reduced
Oxidized
struction of smectitic clay lattices (Kyuma 2004).
organic matter decomposition Intrinsic merits and demerits of the paddy soil/rice system
Aerobic and semianaerobic
Hypothetical pattern of
Anaerobic decomposition
decomposition process
In comparison with the soils used for growing upland crops,
paddy soils cropped to aquatic rice have many intrinsic merits
and some demerits. Among the merits, the following can be
enumerated: (1) a higher natural supply of nitrogen, bases, and
silica; (2) higher availability of soil phosphorus; (3) relative
process
indifference to soil physical properties; (4) detoxification of
excessive nutrients; (5) detoxification of many agrochemicals;
(6) perfect resistance to soil erosion; (7) relative ease of weed-
ing; (8) tolerance for monoculture; and (9) carbon sequestra-
Advanced stage: rapid decrease
Early stage: rapid accumulation
without fresh organic matter
air. In countries with a high fertilizer use, paddy fields are con-
sidered to be the largest potential nonpoint source of pollut-
ants, particularly nitrogen and phosphorus, for water bodies.
Globally, paddy soils are said to be responsible for 11% of
Rapid process
Slow process
NH4-N
Conclusions
Initial Eh7 Expected pattern of
in soil, V energy metabolism
Aerobic respiration
(MnIII, IV reduction)
Sulfate reduction
Nitrate reduction
(FeIII reduction)
0.2
0
0.4
0.5
Formation of MnII
Formation of FeII
Second Formation of S–II
Chemical
First
Session 12: Conservation of soil, water, and environment in rice cultures 353
References Kyuma K. 2004. Paddy soil science. Kyoto (Japan): Kyoto Univer-
sity Press. 280 p.
Eswaran H, Moncharoen P, Reich P, Padmanaban E. 2001. Rice, Takai Y. 1978. Redox processes in the soil under submergence. In:
land and people: the faltering nexus in Asia. Proceedings of Kawaguchi K, editor. Suiden-dojyogaku (Paddy soil science).
the 5th Conference of the East and Southeast Asia Federation Tokyo (Japan): Kodan-sha. p 23-55. (In Japanese.)
of Soil Science Societies, Krabi, Thailand. Bangkok (Thai- White House. 1967. World food problem: a report of the president’s
land): Department of Agriculture. p 38-66. Science Advisory Committee. Vol. II. Report of the Panel on
FAO (Food and Agriculture Organization). 2004. FAOSTAT. Rome the World Food Supply. Washington, D.C. (USA): White
(Italy): FAO. House.
IPCC (Intergovernmental Panel on Climate Change). 1995. Climate
change 1994, radiative forcing of climate change and an evalu-
ation of the IPCC 1992 IS92 emission scenarios. New York, Notes
N.Y. (USA): Cambridge University Press. Author’s address: Professor Emeritus, Kyoto University and the
IRRI (International Rice Research Institute). 2002.World rice statis- University of Shiga Prefecture, 1-79 Nagatanicho, Iwakura,
tics, 2002. Los Baños (Philippines): IRRI. Sakyoku, Kyoto 606-0026, Japan.
Although the Lower Mekong Delta (LMD) covers only 12% The experiment was a randomized complete block de-
of Vietnam’s total land area, it is the main source of agricul- sign with four replications. All crops received 60 kg P2O5
tural products of the country, supplying more than 50% of the ha–1 as single superphosphate. The application of N and K2O
staple food. Among the diverse range of agricultural products fertilizer for each crop was 100-30 for rice, 30-60 for soy-
in the Delta, rice is the main crop. It is planted in irrigated bean, 100-60 for maize, 50-90 for sweet potato, and 60-30 for
lowlands with three major cropping seasons: winter-spring sesame. Nitrogen was applied as urea in three equal splits: 1/3
(November to February), spring-summer (March to June), and at 10 days after planting (DAP), 1/3 at 20 DAP, and 1/3 at 45
summer-autumn (July to October), which produce about 56% DAP. Potassium was applied as KCl. Rice variety IR66 was
of the nation’s total production. Some soils in the Mekong Delta used. Soil properties in the top 25 cm were pH = 4.2, organic
receive high amounts of fertilizer, either because of the pro- C = 2.8%, total N = 0.25%, CEC = 16 cmolc kg–1, clay = 42%,
duction of three crops per year or because soils such as acid and silt = 56%.
sulfate soils require much fertilizer to support production. A rotation of rice with other crops in nearly all cases
Yield of summer-autumn rice in the LMD is usually low significantly improved rice yield (Table 1). Yield and the ap-
(3–4 t ha–1). Therefore, the government and farmers are inter- parent recovery of N fertilizer by summer-autumn rice were
ested in diversifying continuous lowland rice culture to mul- the greatest following sweet potato, and they were greater for
tiple cropping. Urea is the main N fertilizer applied to lowland rice following soybean and maize than following rice. Nitro-
rice in the Mekong Delta. Phosphorus fertilizers have been gen fertilizer-use efficiency was higher following sweet po-
used on some soils in Mekong Delta agriculture for more than tato than rice. The decomposition of sweet potato residues in
30 years. Especially, along the main rivers, soils have received raised beds after harvesting provided N and potentially im-
P fertilizer for more than 20 years. proved soil physical, biological, and chemical processes. In
The objectives of this research are to (1) determine the addition, a symbiotic combination of N2-fixing microorgan-
effects of the previous crop on rice yield and the economic isms of Azospirillum brasilense with sweet potato roots was
merits of crop rotation, (2) determine phytoplankton activity found (Hill et al 1983).
in nitrogen fertilizer-use efficiency, and (3) assess the effects The profit for each cropping system was determined from
of long-term P fertilization in intensive rice production on cad- grain yield and market price. The profit (US$ ha–1 year–1) of
mium contamination of soils. the cropping systems was rice-soybean-rice = 644, rice-sweet
potato-rice = 612, rice-sesame-rice = 586, rice-maize-rice =
503, and rice-rice-rice = 421 (Ngo Ngoc Hung and Nguyen
Effects of crop rotations on rice Bao Ve 2004). Additional research conducted at four locations
Field experiments to assess crop rotations with three consecu- during summer-autumn 2001 and 2002 showed that, in all the
tive crops per year were conducted on alluvial soils in 2002 cases, rice yield was higher after a 3-month fallow (average:
and 2003 at Can Tho. Five crops (rice, soybean, maize, sweet 4.1 t ha-1) than with continuous rice cropping (average: 3.8 t
potato, and sesame) were grown in spring-summer and rice ha–1) (Nguyen Bao Ve et al 2002).
was then grown in winter-spring and summer-autumn seasons.
aNumbers followed by the same letter in a column are significant at the 5% level. bApparent fertilizer N
Phytoplankton activity and nitrogen losses in rice fields lower Cd content (1.73 mg kg–1). Cadmium content in LMD
rice soils ranged from 0.01 to 0.56 mg kg–1. Generally, the
The design was a randomized complete block with four repli- maximum value of Cd (0.56 mg kg–1) in rice soils was lower
cations. Floodwater pH, urea, ammonium, and biomass of phy- than the standard limit of Vietnam (2.0 mg kg–1), the EC Di-
toplankton were monitored after N fertilizer application. Fer- rective (1–3 mg kg–1), and Canada (1.4 mg kg–1) for Cd in
tilizer was applied to rice as urea or ammonium sulfate, super- agricultural soils.
phosphate, and potassium chloride. The correlation between duration of P fertilizer use and
The biomass of phytoplankton (dominated by Diatom, Cd content in soil is well recorded (Fig. 1). Cadmium content
Chlorophyta, Cyanophyta, and Euglenophyta) reached a maxi- was highest in alluvial and acid sulfate soils and relatively lower
mum within 2–4 days after urea application. During summer- in saline soils. This can be attributed to the more intensive
autumn, at 3 days after N fertilizer application, the biomass crop cultivation on alluvial soils (3–4 crops year–1) than on
reached 8.0 g m–2 in urea plots versus 5.0 g m–2 in ammonium saline soils (1–2 crops year–1). Although crop cultivation is
sulfate plots. Because of daytime consumption by algae of CO2 not as great on acid sulfate soils (usually 2 crops year–1) as on
dissolved in the floodwater, the floodwater pH rose to > 8 on alluvial soils, the rate of P fertilizer application is higher on
some days at the first and second N application. However, there acid sulfate soils (60–90 kg P2O5 ha–1) than on alluvial soils
was no significant difference in biomass of phytoplankton or (50–60 kg P2O5 ha–1).
pH of floodwater between these two treatments after the sec- In serious cases, increasing soil Cd content could de-
ond fertilizer application (Watanabe et al 2003). This was con- crease soil health, reduce crop yield (Steve 1994), and affect
ducive to a rapid conversion of NH4+ to gaseous NH3. humans by Cd contamination in food. Recent research has
15Nitrogen balance research during the dry season of shown that the Cd uptake in crops depends largely on soil and
1990 at Can Tho showed that the fraction of added urea-N not climatic factors, plant genotype, and agronomic management.
recovered in the plant and soil, and presumably lost, depended While plant breeding and agronomic management can mini-
on the timing of urea application. The estimated N loss for N mize soil-plant transfer of Cd, and maximize concentrations
applied at 10 DAS was very high (49% of applied N). It de- of antagonists to Cd assimilation in humans, implications of
creased to 25% of the applied N at 20 DAS and 11% of the the studies are that inputs of this metal to soil can be mini-
applied N at 44 DAS. The distribution of 15N in the soil plus mized (McLaughlin et al 1999).
roots suggested that leaching was not an important loss mecha-
nism, and therefore the N loss in this case was presumed to be
Conclusions
due to NH3 volatilization and denitrification (Ngo Ngoc Hung
et al 1995). The rotation of rice with other crops in nearly all cases signifi-
cantly improved rice yield. Rice yield was highest after sweet
potato, and the N fertilizer-use efficiency of rice was higher
Cadmium content in flooded rice soils following sweet potato (29%) than rice (19%). The rotation of
Soil samples from the top 20 cm, which covered three major rice with soybean attained the highest profit because soybean
soil groups (alluvial soils, acid sulfate soils, and saline soils) had a high market price.
in the LMD, were collected. In addition, P fertilizers used in Floodwater pH is a major factor affecting NH3 volatil-
the LMD were collected. ization, and photosynthetic aquatic biomass has a key role in
Cadmium content in P fertilizers ranged from 0.02 to increasing floodwater pH. The biomass of phytoplankton,
2.76 mg kg–1. The highest content was in imported NPK fertil- dominated by Diatom, Chlorophyta, Cyanophyta, and
izer (2.76 mg kg–1), whereas some domestic P fertilizers had a Euglenophyta, reached a maximum within 2–4 days after urea
Session 12: Conservation of soil, water, and environment in rice cultures 355
Soil Cd content (mg kg1)
0.6
0.5
y = 0.0066x + 0.0306, R2 = 0.32*
0.4
0.3
0.2
0.1
0.0
0 10 20 30 40
Duration of P application (y)
Fig. 1. The correlation between years of P fertilizer application and soil Cd content.
application. Management practices to reduce floodwater N Nguyen Bao Ve, Ngo Ngoc Hung, Nguyen Thanh Hoi, Pham Duc
concentration can help minimize N losses. Tri, Nguyen Van Nhieu Em. 2002. Effect of soil fertility and
farming practices on rice growth and yield of summer-autumn
Some soils in the Mekong Delta received high amounts
rice in the Mekong Delta. Vietnam Soil Sci. J. 16(2002):76-
of P fertilizer, which can contain varying degrees of contami-
83.
nants, including cadmium. The concentrations were lower than Steve PM. 1994. Case study 6: effects of heavy metals from sewage
critical values considered by Vietnamese and some interna- sludge on soil microbes in agricultural ecosystems. In: Ross
tional levels, but the results indicate an anthropogenic impact SM, editor. Toxic metals in soil-plant systems. John Wiley
on soil cadmium levels. Concerns about food-chain contami- and Sons. 247 p.
nation should be evaluated. Watanabe T, Ngo Ngoc Hung, Nguyen Do Chau Giang, Vo Thi Yen
Phi, Tran Chan Bac. 2003. Changes of nitrogen in floodwater
after fertilizer application. In: Development of new technolo-
References gies and their practice for sustainable farming systems in the
Hill WA, Bacon-Hill P, Crosman SM, Steven C. 1983. Characteriza- Mekong Delta. Proceedings of the Final Workshop of JIRCAS
tion of N2-fixing bacteria associated with sweet potato roots. Mekong Delta Project, 25-26 November 2003. Can Tho (Viet-
Can. J. Microbiol. 29:860-862. nam): College of Agriculture, Can Tho University. p 373-382.
McLaughlin MJ, Parker DR, Clarke JM. 1999. Metals and micronu-
trients: food safety issues. Field Crops Res. 60:143-163.
Notes
Ngo Ngoc Hung, Singh U, Vo Tong Xuan, Buresh RJ, Padilla JL,
Tran Thanh Lap, Truong Thi Nga. 1995. Improving nitrogen- Authors’ addresses: Ngo Ngoc Hung and Nguyen Bao Ve, Can Tho
use efficiency of direct-seeded rice on alluvial soils of the University, Vietnam; Roland J. Buresh, International Rice
Mekong River Delta. In: Denning GL, Vo Tong Xuan, edi- Research Institute (IRRI), Philippines; Mark Bayley, Univer-
tors. Vietnam and IRRI: a partnership in rice research. Los sity of Aarhus, Denmark; Takeshi Watanabe, Japan Interna-
Baños (Philippines): International Rice Research Institute. tional Research Center for Agricultural Sciences (JIRCAS),
p 137-149. Japan, e-mail: ngochung@ctu.edu.vn.
Ngo Ngoc Hung, Nguyen Bao Ve. 2004. Effect of crop rotation on Acknowledgments: This research was supported in part by grants
nitrogen use efficiency and yield of summer-autumn rice. Viet- from the projects of CAULES, funded by DANIDA (Danish
nam Sci. Technol. J. Agric. Rural Dev. 5(2004):634-636. International Development Agency), and JIRCAS (Japan In-
ternational Research Center for Agricultural Sciences).
Thailand is divided geographically into four main regions: The fertilizer treatments were nil, FYM (cattle dung at 6.25 t
northern, northeast (NE), Central Plain, and southern. North- ha–1), PK, NPK, CR-NPK (controlled-release N), all (NPK +
ern Thailand, with about 22% of the total rice area, contrib- micronutrients), N, and FYM + NPK. The rates of N, P2O5,
utes about 26% of the rice production. The NE has 57% of the and K2O were all 50 kg ha–1. Rice seedlings were transplanted
total area and 46% of the total production. In the Central Plain, at a spacing of 25 × 25 cm at almost all sites except in Tung
the rice-growing area is 17% and rice production is 24%. In Kula Ronghai 1996, where dry seed was broadcast at 62.5 kg
southern Thailand, the corresponding percentages are only 4% ha–1.
and 4%. Soil properties indicate that soil productivity was con-
Rainfed rice is grown on approximately 4.66 million trolled by many factors, including soil fertility, environment,
hectares or about 92% of the total rice-growing area in north- and cultural management. The decline in grain yield under
east Thailand. The average paddy yield in the 2003 crop sea- rainfed loamy sand soil at Ubon was lower than under irri-
son was 1.81 t ha–1, versus 3.21 t ha–1 for irrigated lowland gated conditions. Late dry-seeded sowing and late transplant-
rice in the Central Plain. Among the four regions, rice yield in ing reduced the growth and yield of rice (Naklang et al 1998).
the northeast is the lowest because of at least two major con- The results indicated that the most effective fertilizer was N,
straints, poor soil and the unpredictable amount and distribu- which made the average grain yield increase over all sites by
tion of rainfall. Low soil fertility recurs every year in the rainfed 39%. This result reflected that N was the most limiting nutri-
lowlands in this region, but the problem of drought does not ent in NE Thailand. The application of PK was effective only
occur in the same manner every year (Fukai et al 1995, Ragland when it was applied with N. The application of NPK gave a
1997). higher grain yield than N applied alone: but only 8%. In loamy
The quantity of organic matter in the soil is a major indi- sand soil, the addition of some micronutrients together with
cator of its quality. To achieve higher grain yield, we should NPK (all) increased grain yield from 2% to 21% compared
apply chemical and organic fertilizer. For sustainability, it is with applied NPK. The application of FYM increased grain
important to incorporate organic matter into the soil. Koopmans yield in many experiments. Average grain yield of FYM was
and Goldstein (2001) stated that to improve soil quality we only 7.5% less than the yield of NPK (mean yield of NPK was
should treat our organic matter like a bank account. A bank 2,425 kg ha–1). The additional role of FYM in contributing
account lets us deposit, save, and withdraw something we value. NPK was found in some sandy soils. The grain yield of rice
For sustainability, it is important to deposit in the account of was higher in clay soil than in loamy sand soil when chemical
active organic matter in the soil on a regular basis, thereby fertilizer was applied (Table 1).
building cultural fertility. Poulton (1995) summarized that long-
term experiments are essential in determining the factors af- Long-term effects of rice straw compost
fecting soil fertility and sustainable production. and chemical fertilizer on rice yield
and their residual effects
An experiment has been conducted at the Surin Rice Research
Experiments Center since 1976. Rice straw compost (RSC) at 0, 3.125, 6.25,
Effects of chemical fertilizer and farmyard manure 9.375, and 12.5 t ha–1 with and without chemical fertilizer (CF)
on grain yield of rice in NE Thailand at the rate of 50-50-50 kg of N-P2O5-K2O ha–1 was applied for
Twenty-six experiments were conducted at eight sites in NE 25 consecutive years (1976-2000). During the first 25 years,
Thailand from 1995 to 1997 to determine the effects of farm- the rice variety used was RD 7, short, nonglutinous, and
yard manure (FYM) and chemical fertilizers on rice yield in nonphotoperiod-sensitive. From 2001 to 2003, residual effects
rainfed areas. Out of eight sites, the experiments were sepa- of 25 years’ application of RSC and CF were also investigated
rated into two experimental plots: one was rainfed and the other without any addition of RSC or CF. The rice variety used was
under irrigated conditions. The soil texture was loamy sand at KDML 105.
three sites, sandy loam at three sites, and clayey loam and clay The results showed that average rice grain yield over 25
at one site each. The soil’s pH varied from 4.0 to 5.7. Organic years of application of RSC with CF was higher than grain
matter varied from 0.39% to 1.57% and Bray-II P ranged from yield of the application of RSC without CF. Rice yield increased
7 to 38 mg kg–1. Nonglutinous, photosensitive, and tall culti- when the rate of RSC increased. The residual effects also
vars RD 15, harvested in mid-November, and KDML 105, showed that the higher the RSC rate applied, the higher the
harvested in late November, were used in these experiments. rice grain yield obtained (Table 2).
Session 12: Conservation of soil, water, and environment in rice cultures 357
358
Table 1. Rice grain yield average for 3 years (kg ha–1) as affected by the application of chemical fertilizer and farmyard manure in different locations in northeast
Thailand.a
Treatment Ubon Ubon Tung Kula Sakon Khon Udon Surin Chumpae Phimai Treatment
irrigated rainfed Ronghai Nakorn Kaen mean
pH 4.2, pH 4.1, pH 4.6, pH 7.1, pH 4.5, pH 5.7, pH 4.0, pH 4.2, pH 4.6,
OM 0.52% OM 0.60% OM 0.77% OM 0.96% OM 0.39% OM 1.32% OM 0.59% OM 1.18% OM 1.57%
Nil 1,316 b 857 cd 1,154 b 1,632 de 1,310 c 2,674 2,271 c 1,993 d 1,426 c 1,634 d
FYM 2,234 a 928 cd 2,016 a 2,488 ab 1,669 bc 2,636 3,192 ab 2,876 c 2,371 b 2,264 c
PK 1,423 b 689 d 1,272 b 1,428 e 1,312 c 2,605 2,314 c 2,206 d 1,639 c 1,655 d
NPK 2,408 a 1,318 abc 1,602 ab 1,981 c 2,224 ab 3,005 2,896 ab 3,473 a 3,168 a 2,425 bc
CR-NPK 2,129 a 1,458 ab 1,638 ab 1,907 cd 2,286 a 3,020 3,045 ab 3,012 bc 3,262 a 2,385 bc
Location mean 2,099 1,186 1,645 2,058 1,952 2,810 2,859 2,936 2,681 2,230
aAteach site, means followed by a common letter are not significantly different at the 5% level by Duncan’s multiple range test.
Source: Naklang et al (1998).
Table 2. Rice grain yield (kg ha–1) as affected by rice straw compost (RSC) with and without
chemical fertilizer (CF) and their residual effects (RSC and CF were applied annually from
1976 to 2000 and their residual effects were evaluated from 2001 to 2003).a
aIn
each column, means followed by a common letter are not significantly different at the 5% level by Duncan’s
multiple range test.
term effects of compost, green manure (sunnhemp: Crotalaria
bon. Crop residue, leaf litter, and green manure with slow break-
The annual application of 1.5 kg dry matter ha–1 of leaf
difference in grain yield and obtained a higher rice yield than
Session 12: Conservation of soil, water, and environment in rice cultures 359
Site-specific nutrient management and the sustainability
of phosphorus and potassium supply in irrigated rice soils
of Asia
C. Witt, A. Dobermann, R. Buresh, S. Abdulrachman, H.C. Gines, R. Nagarajan, S. Ramanathan, P.S. Tan, and G.H. Wang
The management of soil phosphorus (P) and potassium (K) is that occur when fertilizer is omitted for longer periods. In the
receiving greater attention in intensive irrigated lowland rice following, we will re-evaluate this strategy.
systems of Asia because of concerns that fertilizer P and K
rates are not optimally adjusted to long-term needs (Greenland
Long-term P and K management strategies
1997, Johnston and Syers 1998). Breeding offers limited op-
portunities to change plant nutrient requirements or uptake In the SSNM approach, fertilizer P and K rates are based on
efficiencies so that long-term management strategies must fo- the yield difference between treatments with optimal fertilizer
cus on overcoming immediate nutrient deficiencies and main- use of N, P, and K (NPK) and omission plots that receive all
taining adequate nutrient balances in the top-soil layer. Cur- nutrients except for the omitted one (0P, 0K). Recommended
rent fertilizer P and K strategies in Asia are still mostly based fertilizer rates would increase with an increasing yield gap
on soil tests that have shown little correlation with the effec- between NPK and 0P or 0K treatments to overcome greater
tive nutrient supply to the irrigated rice crop (Dobermann et al nutrient limitations. Maintenance fertilizer P and K rates are
2003b). Site-specific nutrient management (SSNM) allows an suggested where a yield response to P or K application is lack-
effective management of indigenous P and K supplies by esti- ing, and fertilizer rates would increase with increasing yield
mating fertilizer requirements based on yield level, yield re- level to balance greater nutrient removal with grain and straw
sponse to fertilizer P and K application, and a nutrient balance (Witt et al 2002, Witt and Dobermann 2004). The NPK, 0P,
model (Witt et al 2002, 2004). In that approach, fertilizer P and 0K treatments were also included in five long-term ex-
and K maintenance rates are commonly developed based on periments with two rice crops per year, and the cumulative
only two seasons of on-farm experiments, which may be in- yield differences between NPK and omission plots for a pe-
sufficient to develop long-term strategies. In this paper, we riod of up to 15 seasons are plotted in Figure 1. Differences
evaluate SSNM strategies to prevent soil nutrient depletion were substantial in short- and long-term yield responses to fer-
using data from five long-term experiments established from tilizer P and K application among sites. Yields were not af-
1968 to 1995 in China, India, Indonesia, the Philippines, and fected by K application in India or P application in China in
Vietnam. the first two seasons, and initial yield responses to fertilizer
application were generally small except for Vietnam (P) and
China (K). However, yield responses developed within a few
Indigenous nutrient supplies seasons at all sites except for K application at Omon in the
Nutrient supplies from indigenous sources can be estimated Mekong River Delta of Vietnam, where annual flooding sup-
by measuring plant nutrient uptake or grain yield in nutrient plies a large K load through sedimentation (Nguyen 2003).
omission plots, which receive all nutrients applied as fertilizer Yield responses developed linearly except for China, where
except for the omitted nutrient of interest (Dobermann et al yield responses to P but not K application developed expo-
2003a). In a large multinational research project, omission plots nentially within eight seasons or four years. Long-term fertil-
have been used for several seasons to estimate the spatial and izer requirements would generally be underestimated if fertil-
temporal variation in indigenous nutrient supplies in farmers’ izer requirements were based only on short-term yield re-
fields (Dobermann et al 2004). These on-farm experiments have sponses without considering nutrient removal with grain and
confirmed for a wide range of soils and growing environments straw and the overall input-output balance.
that farmers’ current fertilizer P and K management causes In the absence of a yield response to fertilizer P and K
slightly positive P balances at a number of sites, while fertil- application, SSNM recommends fertilizer P and K maintenance
izer K application is commonly insufficient to balance K re- rates that are calculated based on a nutrient input-output model
moval with grain and straw. Apart from unavoidable inaccura- (Witt and Dobermann 2004). Since initial yield responses ap-
cies in the estimation, nutrient balances probably supply lim- pear to be generally small based on the data presented from
ited information regarding the risk of nutrient depletion, un- long-term experiments, fertilizer requirements with SSNM
less combined with estimates of medium- or long-term changes could, to a larger extent, follow maintenance strategies. At is-
in indigenous nutrient supplies. In our past on-farm research, sue is whether SSNM recommendations would need to be ad-
however, the location of the omission plots was changed every justed if yield responses to fertilizer P and K application were
season in individual farmers’ fields to avoid residual effects evaluated for longer periods. Table 1 shows fertilizer require-
15
10
1 3 5 7 9 11 13 15 1 3 5 7 9 11 13 15 1 3 5 7 9 11 13 15
Season
20
India China
(TNRRI) (ARS)
15 (NPK) (OP)
(NPK) (OK)
10
1 3 5 7 9 11 13 15 1 3 5 7 9 11 13 15
Season Season
Fig. 1. Cumulative grain yield increase in fully fertilized plots (NPK) over plots without fertilizer P (0P) and K (0K) application
in long-term experiments at five sites in Asia, 1995-2002. PhilRice = Philippine Rice Research Institute, Muñoz, Philippines;
CLRRI = Cuu Long Rice Research Institute, Omon, Vietnam; IIRR = Indonesian Institute for Rice Research, Sukamandi,
Indonesia; TNRRI = Tamil Nadu Rice Research Institute, Tamil Nadu, India; ARS = Agricultural Research Station, Zhejiang,
China.
ments calculated with SSNM based on data from two and eight nutrient-supplying power, while a strong yield response de-
cropping seasons in the long-term experiments. Note that the veloped within a few seasons is a sign of a low soil buffering
actual fertilizer P and K rates were several times higher than capacity for nutrient depletion. For example, four years of data
the recommended rates with SSNM. High fertilizer P and K from Omon revealed greater P but lower K requirements than
rates in long-term experiments are usually chosen to ensure the rates calculated after one year. Despite these adjustments,
that these nutrients are not limiting. the initial estimates of fertilizer requirements showed a good
Considering two seasons of data, recommended fertil- congruence with the fertilizer rates that were based on several
izer rates ranged from 21 to 32 kg P2O5 ha–1 and from 20 to 48 seasons of data.
kg K2O ha–1. Suggested rates were highest where both yield
and yield responses to fertilizer application were high (Philip-
Conclusions
pines, India, China). Fertilizer rates changed at some but not
all sites when eight instead of two seasons of data were con- Two-season estimates of indigenous P and K supplies provide
sidered in the calculation of fertilizer rates. Fertilizer rates a first, sufficiently robust estimate of fertilizer P and K re-
needed little adjustment at sites where yield responses devel- quirements. However, on-farm monitoring of indigenous nu-
oped steadily and yields in NPK treatments remained more or trient supplies for longer time periods is essential to fine-tune
less constant (e.g., Philippines). Adjustments in fertilizer rates fertilizer P and K recommendations based on historical trends
were needed where yield levels in the first year were not rep- in soil fertility. Omission plots embedded in farmers’ fields
resentative for a longer time period, as in Indonesia, where should be continued at the same location for several seasons if
yields increased with time. Adjustments were also needed data from representative long-term experiments are not avail-
where average yield responses to fertilizer application were able to identify the response pattern that develops with time.
either lacking or very strong over the four-year period. A lack-
ing yield response over a long period indicates a strong soil
Session 12: Conservation of soil, water, and environment in rice cultures 361
Table 1. Average fertilizer rates and fertilizer requirements with SSNM (Witt et al
2002, 2004) based on the average yield increase (response) in N-, P-, and K-fertilized
plots (NPK) over plots without fertilizer P (0P) and K (0K) application during two and
eight cropping seasons in five long-term experiments in Asia.
Sites
Parameter Unit
Philippines Vietnam Indonesia India China
(PhilRice) (CLRRI) (IIRR) (TNRRI) (ARS)
Fertilizer rates
Starting year 1968 1995 1995 1995 1997
Fertilizer P2O5 kg ha–1 60 57 57 57 57
Fertilizer K2O kg ha–1 60 90 120 120 120
References Nguyen MH. 2003. Soil potassium dynamics under intensive rice
cropping: a case study in the Mekong Delta, Vietnam.
Dobermann A, Witt C, Abdulrachman S, Gines HC, Nagarajan R, Wageningen University, Netherlands. p 1-203.
Son TT, Tan PS, Wang GH, Chien NV, Thoa VTK, Phung CV, Witt C, Balasubramanian V, Dobermann A, Buresh RJ. 2002. Nutri-
Stalin P, Muthukrishnan P, Ravi V, Babu M, Simbahan GC, ent management. In: Fairhurst T, Witt C, editors. Rice: a prac-
Adviento MA. 2003a. Soil fertility and indigenous nutrient tical guide for nutrient management. Singapore and Los Baños:
supply in irrigated rice domains of Asia. Agron. J. 95:913- Potash and Phosphate Institute & Potash and Phosphate In-
923. stitute of Canada and International Rice Research Institute.
Dobermann A, Witt C, Abdulrachman S, Gines HC, Nagarajan R, p 1-45.
Son TT, Tan PS, Wang GH, Chien NV, Thoa VTK, Phung CV, Witt C, Buresh RJ, Balasubramanian V, Dawe D, Dobermann A.
Stalin P, Muthukrishnan P, Ravi V, Babu M, Simbahan GC, 2004. Principles and promotion of site-specific nutrient man-
Adviento MA, Bartolome V. 2003b. Estimating indigenous agement. In: Dobermann A, Witt C, Dawe D, editors. Increas-
nutrient supplies for site-specific nutrient management in ir- ing productivity of intensive rice systems through site-spe-
rigated rice. Agron. J. 95:924-935. cific nutrient management. Enfield, NH (USA) and Los Baños
Dobermann A, Witt C, Dawe D, editors. 2004. Increasing produc- (Philippines): Science Publishers, Inc., and International Rice
tivity of intensive rice systems through site-specific nutrient Research Institute. p 397-410.
management. Enfield, NH (USA) and Los Baños (Philippines): Witt C , Dobermann A. 2004. Toward a decision support system for
Science Publishers, Inc., and International Rice Research In- site-specific nutrient management. In: Dobermann A, Witt C,
stitute (IRRI). p 1-410. Dawe D, editors. Increasing productivity of intensive rice sys-
Greenland DJ. 1997. The sustainability of rice farming. Wallingford tems through site-specific nutrient management. Enfield, NH
(UK) and Manila (Philippines): CAB International and Inter- (USA) and Los Baños (Philippines): Science Publishers, Inc.,
national Rice Research Institute. p 1-273. and International Rice Research Institute. p 359-396.
Johnston AE, Syers JK, editors. 1998. Nutrient management for sus-
tainable crop production in Asia. Wallingford (UK): CAB In-
ternational. p 1-394.
During 1970-2000, although rice production increased from cannot be sustained. Hence, the Green Revolution cannot take
2.4 to 7.4 million tons in West Africa (WA), consumption per place.
capita increased from 15 to 30 kg and imports increased from
0.7 to 4.3 million tons. Yield, however, has been stagnant at
The sawah system and integrated watershed approach
1.3–1.7 t ha–1 during the past 30 years in this region. More
than 100 high-yielding varieties for various rice ecologies are The upper part of Figure 1 shows a concept of macroscale
available and another 6 NERICA varieties are in the pipeline ecological engineering, that is, watershed ecological engineer-
for upland areas. The lowland elite varieties under the sawah1 ing. The soils formed in uplands and the nutrients released
system with good management could yield up to 6 t ha–1. But during rock weathering and soil formation processes in up-
these potentials are never achieved in farmers’ fields because lands are accumulated in lowlands (geological fertilization).
of poor water control and bad soil management systems. This If the sawah system exists in the lowlands, it can store and
is why a renewed call for a concerted effort needs to be made effectively use the nutrient-rich water and fertile topsoils eroded
for better water and soil management systems, referred to in from the uplands. Watershed agroforestry through the integra-
this paper as the sawah system (Wakatsuki et al 1998), and tion of upland forestry and lowland sawah systems in a unit
correct policies for rice development. watershed as seen in the upper part of Figure 1 is a typical
Why has the Green Revolution not yet occurred in West model of watershed ecological engineering. Optimum land-
Africa in spite of its successes in Asia in the 1960s? The Green use patterns and landscape management practices optimize
Revolution laid the foundation for the rapidly growing econo- geological fertilization through optimum hydrology in a given
mies of Asia today. As our team revealed, the ecological envi- watershed. This is an eco-environmental basis for the long-
ronment of soil and water conditions in this region is very se- term intensive sustainability of sawah-based rice farming in
vere. Lowland soil fertility in WA may be the lowest among Asia.
the major tropical areas in the world (Issaka et al 1997, Buri et
al 2000, Hirose and Wakatsuki 2002). The main cause of the
present agricultural and environmental crises in WA, however,
may be the underdevelopment of lowland agriculture. Envi-
ronmentally creative technology, or ecological engineering
1The sawah system is the rice farmers’ basic infrastructure for intensive and
technology, such as lowland sawah farming, is not tradition-
ally practiced in WA. Sawah is a multifunctional constructed sustainable rice production. The term “sawah” refers to a leveled rice field
surrounded by bunds with inlet and outlet connections to irrigation and drain-
wetland, which is the prerequisite for realizing the Green Revo- age canals. The term originates from Malayo-Indonesian. The English term
lution as well as for preserving and even restoring ecological “paddy or paddi” also originates from the Malayo-Indonesian term “padi,”
environments. Irrigation and drainage without farmers’ sawah which means rice plant. However, the term “paddy” refers to rice grain with
farming technologies have proved inefficient or even damag- husk in West Africa as a whole. The paddy field is almost equivalent to the
ing because of accelerated erosion processes. Thus, the devel- sawah for Asian scientists. However, the term “paddy fields” refers to just a
rice field, including an upland rice field in West Africa. Therefore, to avoid
opment of irrigation has been slow. In the absence of water confusion among the terms “rice plant,” “paddy,” and the improved man-
control, fertilizers cannot be used efficiently. Consequently, made rice-growing environment through ecological engineering, the authors
the high-yielding varieties perform poorly and soil fertility propose to use the term “sawah.”
Session 12: Conservation of soil, water, and environment in rice cultures 363
(A) Optimum land-use patterns and landscape management practices optimize geological
fertilization through the control of optimum hydrology in the watershed
Rainfall Application of various types
of humified organic matter
Watershed
Intensive sustainability in uplands
agroforestry
= combination with stock raising
Uplands: 1,000–5,000 ha
Nutrient cycle
tion Village
iliza
fert
cal l
logi na
Geo Top ca
s oil on
Nu
trie g ati
nt- irri Small reservoir
rich all Supply of
Top
s wa
ter Sm Fish culture pond nutrients by
oil
ero animals
sio Lowlands = about 5%
n
Rock weathering, soil formation,
and nutrient release Sedimentation Sawah hypothesis: 1 ha of sawah fields
Stream discharge is equivalent to 10–15 ha of uplands
Geological O 2 + N2 CH4
fertilization CO2 emission Technology development for nitrogen fixation through
integrated management of soil, water, algae, and rice
N2 N2 O
N2 (N2O, NO)
biological
nitrogen
fixation Algae Submerged
Algae Nitrificaton Denitrificaton water
A few mm CH4→ CO2 NH 3→NO 2→NO 3
to 2 cm +400–500 mV Oxidized
layer
Organic N2
NO3 N2 (N2O, NO)
NO3← NO2← NH4-N Mn (IV) Mn (II) –100 mV
Plow layer
Fe (III) Fe (II)
10–20 cm Reduced
SO4 H2S
N2 NH4 layer
CO2 CH 4 CO2 CH4 –250 mV
O2 , N 2 CH3COOH CH4 (Eh)
CH4→CO2
Accumulating layer Fe2CO3 ⋅ nH2O
Fig. 1. Macro- and microscale ecological mechanisms of intensive sustainability of the low-
land sawah system: (A) geological fertilization through watershed ecological engineering and
(B) multifunctional constructed wetlands for enhanced supply of N, P, Si, and other nutrients.
Although the rate of soil formation, and thus also soil tive scientific evaluation of the geological fertilization pro-
erosion, is much higher in Asia than in Africa, suppose that the cesses in a watershed will be an important future research sub-
soil formation rate in uplands, which make up 95% of the total ject.
area in the example of a watershed shown in Figure 1, is 1 t The excessive decomposition of organic matter in tropi-
ha–1 y–1. In a stable ecosystem in a watershed, the rates of soil cal soils is another problem. Therefore, organic matter man-
formation and erosion should be well balanced; therefore, the agement through agroforestry and cover-crop systems is im-
topsoils formed in uplands—which account for 95% of the portant for sustaining soil fertility. Sustainability under inten-
area—and the nutrients produced in the process will be con- sive cultivation of uplands is possible through integration with
centrated in lowlands, which make up 5% of the area. Thus, stock raising. Animals can accumulate nutrients on upland
the soil formation rate in the lowlands equals 20 t ha–1 y–1 farms and lowlands. However, except for lowland sawah farm-
(Hirose and Wakatsuki 2002). Though it will be impossible to ing, these options are not successful enough to make tropical
use all of the rich soils and nutrient-rich water from the up- upland farming sustainable under intensive cultivation. The
lands effectively, the sawah farming system could artificially sawah system can also control such excessive organic matter
reinforce the geological fertilization processes. The quantita- decomposition (Kyuma 2003). Another possibility to restore
Session 12: Conservation of soil, water, and environment in rice cultures 365
Forest destruction induced
savannization
Upland-lowland Intensified rainfed to irrigated
continuum lowland sawah system
Evaporation
Bunding
Spring irrigation
Fragile:
no sustainable upland Leaching runoff
rice including NERICA Leveling
Stabilized
Humid zone Sahel zone
Agroecological zone: Guinea savanna to Guinea savanna Sudan savanna to Sudan savanna to Sahel to
Guinea savanna to humid forest zone to humid forest humid forest zone humid forest zone humid forest zone
humid forest zone zone
Fig. 2. Rice ecologies along a continuum of inland valley watersheds and floodplains in West Africa (excluding the ecologies of deepwater
and mangrove swamp rice).
Greenland DJ. 1997. The sustainability of rice farming. Wallingford Wakatsuki T, Otto E, Andah WEI, Cobbina J, Buri MM, Kubota D,
(UK): CAB International, in association with the International editors. 2001. Integrated watershed management of inland
Rice Research Institute. 273 p. valley in Ghana and West Africa: ecotechnology approach.
Hirose S, Wakatsuki T. 2002. Restoration of inland valley ecosys- Final Report on JICA/CRI joint study project, CRI, Kumashi,
tems in West Africa. Tokyo (Japan): Norin Tokei Kyokai. 600 Ghana, and JICA, Tokyo. 337 p.
p. Wakatsuki T, Matsui K, Shibata K, Matsuoka K, Masunaga T. 2003.
Hsieh S-C. 2001. Agricultural reform in Africa, with special focus A method for organic fertilizer production. Japanese Patent
on Taiwan-assisted rice production in Africa, past, present and Application No. 2002-338185.
future perspectives. Tropics 11:33-58. WARDA (West Africa Rice Development Association). 1988. Stra-
Issaka RF, Ishida F, Kubota D, Wakatsuki T. 1997. Geographic dis- tegic plan. Bouaké (Côte d’Ivoire): WARDA.
tribution of selected soil parameters of inland valleys in West WARDA (West Africa Rice Development Association). 1999. Afri-
Africa. Geoderma 75:99-116. can rice initiative report. Bouaké (Côte d’Ivoire): WARDA.
Kyuma K. 2003. Paddy soil science. Kyoto (Japan): Kyoto Univer-
sity Press. 305 p.
Wakatsuki T. 2004. Watershed ecological engineering for sustain- Notes
able increase for food production and resoration of degraded Authors’ addresses: Toshiyuki Wakatsuki, Faculty of Agriculture,
environment in West Africa. www.kindai-ecotech.jp. Kinki University, 3327-204 Nakamachi, Nara 631-8505, Ja-
Wakatsuki T, Shinmura Y, Otto E, Olanian G. 1998. African-based pan, e-mail: wakatuki@nara.kindai.ac.jp; Md. Moro Buri, Soil
sawah system for the integrated watershed management of Research Institute-CSIR, Kumasi, Ghana, e-mail:
small inland valley in West Africa. Water Reports 17. Institu- sawahksi@ghana.com; Oluwarotimi O. Fashola, Watershed
tional and technical opinion in the development and manage- Initiative of Nigeria, WIN2001, International Institute of Tropi-
ment of small-scale irrigation. Rome (Italy): FAO. p 56-79. cal Agriculture, Ibadan, Nigeria, e-mail: r.fashola@cgiar.org.
Nitrogen fertilizer efficiency in flooded rice fields is usually For U275 and U550 treatments, readily available urea
poor. Upland crops frequently use 40–60% of the applied N, was applied as basal, tillering, and booting fertilizer, with 40%,
whereas flooded rice crops typically use only 20–40% (Vlek 20%, and 40% of N applied in the whole rice season, respec-
and Byrens 1986). This means that about 60% of the applied tively.
nitrogen in the rice season is released from the soil to a water All treatments received 100 kg K2O ha–1 and 100 kg
body (surface water and groundwater) and to the atmosphere. P2O5 ha–1 as basal fertilizer. Fertilizers for basal application
For these reasons, sustainable forms of agricultural systems were scattered evenly on the wet soil surface and flooded and
that are more stable and friendly to the environment are now puddled for transplanting. The other two applications adopted
gradually recognized by the government as a long-term policy. the topdressing method.
Meanwhile, to develop corresponding technologies and evalu- For each irrigation, the soil was kept flooded at 5-cm
ate their environmental impact have become important for depth, and supplemental irrigation water was added according
agroenvironmental researchers. Much research has been done to local practices. Meanwhile, the percolation rate was artifi-
in the area of NO3– movement, N2O emission, and NH3 vola- cially controlled at 5 mm day–1 continuously with a medical
tilization, but generally only one or a few components are in- transfusion apparatus. Some 7.5 L of percolate was collected
volved in one experiment. As we know, the processes in N every 3 days with a 10-L pail that contained 10 mL of 2 M
cycling are correlated and they interact. Therefore, there is a H2SO4.
need to determine as much as possible measurable items for N The runoff samples were collected in the rice season.
cycling at one time in one experiment. The rain sample was collected with a rain gauge. All water
A field experiment on monolithic lysimeters was carried samples were stored in the dark at 4 °C until analyzed.
out in the 2001 rice season in the Tai-hu Lake region. The Plant samples were collected for straw and grain nutri-
objective of our studies was to investigate nitrogen cycling ent analysis when the crop was harvested.
data for urea applied in paddy fields managed according to the The ammonia volatilization rate was measured using a
prevailing practices of local farmers, including losses through continuous airflow enclosure method (Kissel et al 1977). The
surface runoff, percolation loss, and ammonia volatilization, rate of airflow, generated by a pump, was adjusted using valves
and to evaluate their impact on the environment. to 15–20 times the chamber volume per minute (Liao 1983).
Measurement continued every day (except for a heavy rain
day) until there were no significant differences between the
Materials and methods NH3 volatilization rates from the N treatments and the con-
The experiments were carried out at the Changshu Agro-eco- trol. NH4+, NO3–, and total N in water samples were measured
logical Experiment Station (31º32′45″N and 120º41′57″E). The colorimetrically by the indophenol (NH4+), ultraviolet spec-
soil was classified as a Wushan soil, a waterlogged soil devel- trophotometry (NO3–), and K2S2O8 oxidation ultraviolet spec-
oped from a parent material of lacustrine deposit, with the fol- trophotometry (total N) methods.
lowing characteristics in the surface 0.15-m layer: pH 7.36
(H2O); total N 1.79 g kg–1; organic carbon 20.0 g kg–1; texture
Results and discussion
silty sandy clay loam; bulk density 1.22 g cm–3; and CEC 20.20
cmol kg–1. Ammonia volatilization potential
Ten profile-undisturbed lysimeters each had an 800-mm The cumulative NH3 losses from U treatments were signifi-
inner diameter and 1,000-mm height. One percolate outlet was cantly higher than from the control. The average cumulative
drilled at the center of the bottom. Each lysimeter was installed NH3 losses in three rice seasons were 39.43 and 100.41 kg N
with two runoff outlets in the column wall, one at 60 mm above ha–1 from the U275 and U550 treatments, which corresponded
and one at 100 mm below the top surface. to 24.06% and 32.34% of the applied N, respectively (Table
The experiment included three treatments: U550 (urea 1).
at 300 and 250 kg N ha–1 for rice and wheat, respectively; con- The cumulative NH3 losses in the wheat season in 2002-
ventional level), U275 (urea at 150 and 125 kg N ha–1 for rice 03 were 20.99 and 40.10 kg N ha–1 from the U275 and U550
and wheat, respectively); and a control (no N). The treatments treatments, which corresponded to 11.94% and 13.61% of the
(3 replicates) and control (4 replicates) were arranged ran- applied N, respectively (data not shown).
domly. The ratio of ammonia volatilization to nitrogen applied
increased, along with the increase in application rate in urea
Session 12: Conservation of soil, water, and environment in rice cultures 367
Table 1. The amount and ratio of N loss by NH3 volatilization potential and runoff in the rice season.
NH3 volatilization
Control 0 ± 0 Ca 6.92 ± 0.36 c 3.10 ± 0.85 c
U275 24.11 ± 6.25 b 48.91 ± 4.40 b 45.27 ± 7.49 b 16.08b 28.00 28.11 24.06
U550 77.98 ± 11.20 a 110.67 ± 7.34 a 112.68 ± 13.34 a 25.99 34.50 36.53 32.34
Runoff
Control 4.1 ± 0.4 c 2.2 ± 0.4 c 4.4 ± 1.8 c
U275 20.4 ± 4.5 b 12.3 ± 3.5 b 25.4 ± 2.0 b 10.9 6.7 14.0 10.53
U550 44.7 ± 7.1 a 18.8 ± 3.7 a 42.7 ± 1.6 a 13.5 5.5 12.8 10.60
aAverage values in the same row without a common letter differ significantly at P = 0.01. bLoss ratio of N applied is N loss through NH emissions
3
minus the control value divided by N rate.
treatments. But, the ratio of NH3 loss to N applied in the wheat The mean value of NO3–-N concentration in percolate
season was much lower than that in the rice season. in fertilizer treatments was 0.21 mg N L–1 and it ranged from
0.0 to 0.83 mg N L–1. This was much lower than the norm for
Runoff loss NO3–-N concentration (WHO 1984). In the wheat season, the
Surface runoff occurred 5, 7, and 2 times for the 2001, 2002, highest TN concentration in percolate was 4.8 mg N L–1, while
and 2003 rice seasons, respectively. Nitrogen loss by runoff the highest NO3–-N concentration was 1.3 mg N L–1, which
varied significantly with different treatments at different stages. was much lower than the norm for NO3–-N concentration, too.
Significant variance in both runoff total N (TN) concentra-
tions and TN losses by runoff between different years occurred. Rainfall
Nitrogen loss by runoff was 20.4 and 44.7, 12.3 and 18.8, Rainfall occurred 123 times from June 2001 to May 2003.
and 25.4 and 42.7 kg N ha–1 for U275 and U550 in 2001, 2002, The average N concentration in rainfall was 1.4 mg L–1 (0.1–
and 2003, respectively. The ratios of nitrogen loss by runoff to 8.8 mg L–1), 1.7 mg L–1 (0.1–10.5 mg L–1), and 0.7 mg L–1
N applied were 5.5–14.0% for U treatments. There was a sig- (0.0–5.2 mg L–1) for NO3–-N, NH4+-N, and dissolved organic
nificant positive correlation between nitrogen runoff losses and N, respectively (data not shown). The concentration of TN in
applied N amount in urea treatments; the ratio of runoff N loss each rainfall was higher than 0.2 mg N L–1 (the threshold con-
to N applied was similar (insignificant difference at P<0.05) centration of N for inland water eutrophication). The nitrogen
between U275 and U550 (Table 1). input by rainfall should be considered in research on nitrogen
Runoff amount had been tested, having no significant balance in the agroecosystem and eutrophication.
difference, among different treatments at the same runoff event.
Therefore, the N concentration in surface water was the key Field nitrogen balance
factor for determining N loss by runoff when surface runoff Bio-uptake was the largest output of nitrogen for all treatments.
occurred. The N concentration in surface water increased rap- Ammonia volatilization had the most importance for urea treat-
idly to the maximum in 2 days after urea application, and then ments, followed by runoff and drainage. The total N loss mea-
decreased quickly afterward. So, the likelihood of N loss by sured under the two rice-wheat rotations in 2001-03, which
runoff is not great if there is no heavy rain within 5 days after was directly involved with pollution of the water body and
fertilizer application. Therefore, if N application were avoided atmosphere, was 176.7 and 483.6 kg N ha–1, which accounted
in a week before heavy rain, N loss by runoff would decrease for 27.9% and 41.9% of N applied for U275 and U550, re-
markedly. N loss by runoff will also decrease markedly if the spectively (Table 2).
runoff outlet is heightened properly. The control and U550 got the deficit, while U550 got
the surplus. This suggests that the suitable N fertilizer applica-
Nitrogen in percolate tion rate should be > 150 kg N ha–1 and < 300 kg N ha–1 for
There was no obvious difference between different treatments common urea under the rice-wheat rotation in this region.
for the amount of TN loss from percolation in the same sea-
son. Even for U550, the TN loss from percolation accounted
Conclusions
for only about 0.1% of the N applied. This means that the TN
loss from percolation has not been rising in parallel with the The following conclusions can be drawn:
increasing urea fertilizer application rate. This suggests that 1. Urea treatments for ammonia volatilization were re-
nitrate in groundwater is not directly related to N percolation markable in this region in both the rice and wheat
from N applied to the paddy field (Zhu et al 2003). seasons.
An innovative yet simple technology that increases nitrogen- and out of the floodwater, where it is more susceptible to loss
use efficiency, and which is being widely disseminated for as gaseous compounds or runoff (Mohanty et al 1999, Savant
managing urea in flooded rice systems in Bangladesh, involves and Stangel 1990).
the deep placement of urea supergranules or briquettes into The UDP technology as applied to flooded rice in
puddled soil shortly after transplanting rice. The briquettes are Bangladesh offers a compelling example of an effective ap-
made by compressing prilled or granular urea in small ma- proach to managing urea fertilizer that results in not only im-
chines with indented pocket rollers that, depending on the size proved efficiency but also greater yield with less urea fertil-
of the pocket, produce individual briquettes varying in weight izer. Farmer awareness of the technology has been increased
from 0.9 to 2.7 g. Within a week after transplanting rice, the through training and promotional activities, including on-farm
briquettes are inserted into the puddled soil by hand, being comparisons of UDP and broadcast urea in trials managed by
placed to a depth of 7–10 cm in the middle of alternating farmers. For this report, data from on-farm trials are used to
squares of four hills of rice. Often referred to as urea deep examine changes in grain yield, urea fertilizer use, and N-use
placement (UDP), this technology improves N-use efficiency efficiency attributed to the use of UDP technology.
by keeping most of the urea N in the soil close to plant roots
Session 12: Conservation of soil, water, and environment in rice cultures 369
Methods UDP grain yield (t ha1)
9 Boro
Rice is grown throughout the year in Bangladesh, with two Aman
principal cropping seasons: boro (irrigated dry season) and
8
aman (wet season). The boro crop is transplanted in January-
February and harvested in May-June, while the aman crop is
7
transplanted in July-August and harvested in November-De-
cember. Yields are usually greater during the boro season be-
6
cause of more favorable temperatures and solar radiation for
crop growth. 5
On-farm trials 4
During the 2000-04 boro and aman seasons, farmers in seven
districts (Bogra, Chandpur, Jessore, Kishoreganj, Mymensingh, 3
Pabna, and Tangail) conducted 531 on-farm trials to measure
rice yields obtained in side-by-side comparisons of UDP ver- 2
sus the current farmers’ practice (FP) of applying split appli- 1:1 line
1,500
kg grain per kg N applied
1,000 100
500 50
0 0
500 50
100
DGrain yield DApplied N DPFP-N
150
Fig. 2. Comparative distributions showing the changes in grain yield (∆ ∆ Grain yield), N
∆ Applied N), and partial factor productivity of applied N (∆
fertilizer use (∆ ∆ PFP-N) attrib-
uted to UDP technology versus the farmers’ practice (FP) for the boro and aman rice
crops. The box plots display the 10th, 25th, 50th, 75th, and 90th percentile observa-
tions along with the mean (solid marker).
kg grain ha–1 in 75% of the on-farm comparisons, increasing rice (4 million boro, 3 million aman) grown annually in the
to more than 1,400 kg grain ha–1 in as many as 25% of the country could potentially produce about 7 million t more rice,
comparisons. At the same time, 75% of the farmers realized a and do so with about 0.4 million t less applied N using UDP
reduction in applied N of more than 45 kg N ha–1, while as technology.
many as 25% of the farmers achieved a savings that exceeded The efforts to improve N-use efficiency using UDP tech-
90 kg N ha–1. Furthermore, in 75% of the comparisons, the nology in Bangladesh and other Asian countries have also
increased yield and savings in applied N translated into in- helped to identify situations where the effectiveness of UDP
creases in the PFP-N of more than 30 kg of grain per kg of may be less than expected and where adoption may also be
applied N. Similar distributions, but with somewhat smaller limited (Mohanty et al 1999). For example, UDP would not be
values, were obtained for all three parameters during the aman as effective on soils with high internal drainage. Additionally,
season. Except for a few cases where the use of applied N was it is important to recognize that UDP will be most effective
greater for UDP aman than for FP aman (i.e., ∆Applied N was when using high-yielding varieties and good management prac-
positive), all distributions showed that UDP provided greater tices.
benefits in both seasons, and there was essentially no risk of For flooded rice systems, few practical improvements
obtaining a less favorable outcome using UDP versus FP. in fertilizer management or products prove to be as cost-effec-
The average UDP yield benefit over the current practice tive as UDP in increasing fertilizer-use efficiency and address-
was 1,120 kg grain ha–1 (SE = 32.4) during the boro season ing food production and environmental concerns. As UDP tech-
and 890 kg grain ha–1 (SE = 32.5) during the aman. Impor- nology continues to be improved and adopted in Bangladesh,
tantly, the savings in applied N amounted to 70 kg N ha–1 (SE it promises to significantly reduce N losses to the environment
= 2.4) during the boro season and 35 kg N ha–1 (SE = 9.1) while also increasing grain yield and improving farmer income.
during the aman. As impressive as these numbers appear, they
are certainly realistic and within the range of values cited by
References
others for both on-farm and on-station experiments (Daftardar
et al 1997, Mohanty et al 1999, Pasandaran et al 1999, Savant Daftardar S, Wagle S, Savant N. 1997. Agronomic performance of
and Stangel 1990). urea briquettes containing diammonium phosphate in rainfed
transplanted rice on farmers’ fields. J. Agric. Sci. Cambridge
128:291-297.
Discussion and conclusions Mohanty SK, Singh U, Balasubramanian V, Jha KP. 1999. Nitrogen
deep-placement technologies for productivity, profitability, and
The current practice of broadcasting urea represents a tremen-
environmental quality of rainfed lowland rice systems. Nutr.
dous waste because much of the applied N is lost before it can Cycl. Agroecosyst. 53:43-57.
be effectively used by the plant. Greater adoption of UDP by Pasandaran E, Gultom B, Sri Adiningsih J, Apsari H, Rochayati S.
rice farmers could improve the use efficiency of applied N 1999. Government policy support for technology promotion
and successfully reduce losses to the environment. If the num- and adoption: a case study of urea tablet technology in Indo-
bers obtained from the on-farm studies are taken to be repre- nesia. Nutr. Cycl. Agroecosyst. 53:113-119.
sentative for all of Bangladesh, then the 7 million hectares of
Session 12: Conservation of soil, water, and environment in rice cultures 371
Savant N, Stangel P. 1990. Deep placement of urea supergranules in R.B. Diamond, senior scientist-retired, IFDC; U. Singh, se-
transplanted rice: principles and practices. Fert. Res. 25:1- nior scientist, Resource Development Division; T.P. Thomp-
83. son, senior scientist, Market Development Division, IFDC,
P.O. Box 2040, Muscle Shoals, Alabama 35662, USA.
Notes Acknowledgments: We wish to acknowledge the significant contri-
butions made by M.M. Islam, S.A.M. Hossain, and K.M. Elahi
Authors’ addresses: W.T. Bowen, resident representative, Interna- in carrying out the work that has been presented. Funding for
tional Fertilizer Development Center-Asia Division, Road the on-farm studies was provided by the International Fund
54A-House #2-Apt. #6, Gulshan 2, Dhaka 1212, Bangladesh; for Agricultural Development.
In paddy fields, leaching and denitrification of applied fertil- ents in the soil are caused by the intervention of soil between
izer nitrogen result in a low fertilizer-use efficiency. The aver- nutrients and plant roots as shown in Figure 1. However, the
age recovery rate of basally applied rapidly available fertil- invention of controlled-availability fertilizer has made possible
izer (RAF) nitrogen has been clarified to be about 30% in the the contact application (co-situs application) of fertilizer with
tillage transplanting system using 15N-labeled fertilizer. These seeds or roots. The term co-situs, which means the existence
losses of fertilizer nitrogen lead to increased nitrous oxide in of both fertilizer and seeds or roots at the same site (situs),
the troposphere, and contribute to both global warming and was proposed by Shoji and Gandeza (1992) to distinguish the
destruction of the stratospheric ozone layer. Compared with seed-placement or contact-placement of small amounts of RAF.
basally applied nitrogen, topdressed N showed a fairly high Therefore, co-situs application of CAF can directly supply
recovery rate (50–60%). Therefore, multiple split application fertilizer components to plant roots, and thus increase fertil-
of N fertilizer reduces N losses, but increases operating costs. izer-use efficiency to a great extent by minimizing the reac-
On the other hand, the invention of controlled-availability fer- tion and interaction of fertilizer components with the soil (Fig.
tilizer (CAF) whose nutrient release rate meets plant nutrient 1). Consequently, it is also highly possible to mitigate the en-
demand allowed contact application of fertilizer with plant vironmental loading of N2O and NO3– caused by fertilizer N
seeds or roots (co-situs application). Consequently, CAF con- applied.
tributes to the improvement of fertilizer-use efficiency to a great
extent, mitigation of environmental loading from fertilizer, and
No-tillage direct-seeding culture of rice with CAF
the development of innovative fertilizer application methods
such as a single basal application of total fertilizer, a single In Japan, rice cropping has been facing serious problems, in-
basal nursery application of nutrients, the application of an cluding overproduction, trade liberalization, the advanced av-
aimed form of nutrients directly to plant roots, etc., and the erage age of farmers, and a small number of successors, and
establishment of innovative farming systems such as no-till- thus a reduction in labor and production costs in rice cultiva-
age transplanting rice culture, no-tillage direct-seeding rice tion is required. The most desirable rice farming system is di-
culture, etc. (Shoji and Gandeza 1992). rect seeding by no-tillage and a single basal fertilization using
The objectives of this paper are to show the concept of CAF, which can also reduce the growth process of rice seed-
co-situs application of CAF and to summarize the innovative lings.
field work that has been done by our research group in the last Single basal co-situs application of total fertilizer nitro-
two decades using this technique. gen with CAF, polyolefin-coated urea of the 100-day type
(POCU-100), in no-tillage direct-seeding culture was con-
ducted in comparison with a conventional culture using am-
The concept of co-situs application monium-sulfate (AS) for both basal and topdressing. Nitrogen
Rapidly available fertilizers such as ammonium sulfate, urea, recoveries of basal POCU-100, basal-AS, and topdressed AS
etc., applied at the same site as seeds or in the intensive root- fertilizer were 63.2%, 8.5%, and 41.5%, respectively. Reflect-
ing zone often cause salt and/or ammonia injuries to plants ing the N-use efficiency, the brown rice yields of the POCU-
because of their high solubility. Therefore, farmers generally 100 plot were 33–55% superior to those in the conventional
have to adjust the soil intervening between fertilizer and seeds AS plot (Sato et al 1993).
or roots, or to apply only small amounts of RAF as a contact- This no-tillage direct-seeding system is being used basi-
placement to avoid these injuries. Phenomena such as nitrifi- cally in well-drained paddy soil. However, in Tohoku (north-
cation, denitrification, immobilization, and fixation of nutri- eastern) District, the main rice-producing zone in Japan, more
Denitrification
(paddy soil)
Rapidly available N2 O N2 Controlled
fertilizer (RAF) Plant root availability Plant root
fertilizer
N2 H 4 + (CAF) NO3
Paddy
NH4+
Soil NH4+
NO3
Upland Soil Soil
than half of the paddy fields are occupied by poorly drained conventional transplanting system (Saigusa et al 1996, 2004).
soil. Therefore, the development of no-tillage direct seeding The recoveries of POCUs-100 in the no-tillage system, basal-
of rice on Y-shaped furrows without covering soil using CAF AS, and the 1st and 2nd topdressed AS in the conventional
was studied in a heavy-textured Hachirogata soil and resulted tillage system were 61.4–82.7%, 31.1–43.1%, 40.0–76%, and
in the same level of brown rice yield as with the conventional 43–82.7%, respectively. The recoveries of fertilizer in each
transplanting system (Tashiro et al 2000). treatment seemed to be in the following order: light clay allu-
vial soil > clay loam Andisol > sandy loam alluvial soil, and
Single co-situs application of CAF to a nursery box also basal POCUs-100 > 2nd topdressed AS > 1st topdressed
In the last two decades, about one-third of the total area under AS > basal AS. These results indicate that the co-situs appli-
rice cultivation in Japan was converted to other crops such as cation of sigmoid CAF in a nursery box for the no-tillage trans-
soybean, wheat, and forage crops by political control of rice planting system increases the N-use efficiency of fertilizer com-
cultivation. Therefore, for farmers, the quality of rice is be- pared with AS in the conventional transplanting system, and
coming a greater issue than the quantity because of its much thus could increase both the growth and yield of rice plants in
higher price. At the same time, farmers are forced to reduce each soil and reduce environmental degradation caused by fer-
labor and production costs in rice cultivation. The develop- tilizer N (Saigusa et al 2004).
ment of a sigmoid type of CAF, which shows a delayed N re-
lease for a specified number of days, signified a single co-
situs application of CAF to nursery boxes (Sato and Shibuya
Supply of the aimed form of nitrogen
using co-situs application of CAF
1991). The controlled release of nitrogen from sigmoid
POCUs-100 held by rice roots, which takes place after trans- In paddy soil, nitrogen applied as nitrate is much more un-
planting, remarkably improved the N-use efficiency of fertil- stable than ammonium because of denitrification caused by its
izer (Saigusa et al 1996, 2004). reductive condition (Fig. 1). Therefore, it is almost impossible
to supply the nitrate form of nitrogen for rice plants using RAF.
No-tillage transplanting cultivation On the other hand, the single co-situs application of CAF may
with co-situs application of CAF to a nursery box directly supply the fertilizer component to plant roots (Fig. 1).
No-tillage transplanting rice cultivation is expected to main- Therefore, the supply of the nitrate form of nitrogen to rice
tain the grain quality of rice while saving both labor and pro- plants was examined by the co-situs application of polyolefin-
duction costs by eliminating plowing and puddling. However, coated calcium nitrate (POC-CN) (Saigusa et al 2001).
in this system, farmers had to apply RAF on the surface of the As Figure 2 shows, the recoveries of calcium nitrate (CN)
soil as basal and topdressing, and thus fertilizer-use efficiency at the harvesting stage were only about 2% in both clay loam
was extremely low (< 10%). Therefore, the technique of grow- Andisol and light clay alluvial soil because of denitrification,
ing seedlings by the single co-situs application of CAF to a whereas those of POC-CN were 21–26%, which were almost
nursery box was introduced in this system and both the recov- the same as those of AS (22–24%). These results may show
ery of nitrogen and brown rice yield were examined in three that rice roots can absorb nitrate-N from POC-CN particles
different types of paddy soils in comparison with those in the because they can come into contact with the CAF without any
Session 12: Conservation of soil, water, and environment in rice cultures 373
Recovery (%)
80 80
Clay loam Andisol Light clay alluvial soil
60 60
40 40
20 20
0 0
CN AS POC-CN POCU CN AS POC-CN POCU
Fig. 2. Recoveries of different forms of nitrogen fertilizer by the rice plant. CN = calcium nitrate,
AS = ammonium sulfate, POC-CN = polyolefin-coated calcium nitrate, POCU = polyolefin-coated
urea.
burning. The highest recoveries were obtained in POCU (49– Saigusa M, Ombodi A, Owashi T, Watanabe H. 2001. Supply of
50%). Reflecting the recovery of each fertilizer, the grain yields aimed form of nitrogen using controlled availability fertilizer.
of rice were in the following order: POCU > AS = POC-CN > Fertilization in the Third Millennium: Fertilizer, Food Secu-
CN. These results imply that the supply of the aimed form of rity and Environmental Protection, 12th World Fertilizer Con-
gress of CIEC. p 315-321.
nutrient directly from fertilizer to plant roots is highly feasible
Sato T, Shibuya K. 1991. One-time application of total nitrogen fer-
through the co-situs application of CAF. Similar results were
tilizer at nursery stage in rice culture. Rep. Tohoku Br. Crop
obtained in supplying NH4+-N in upland soil, Fe2+ in alkaline Sci. Soc. Jpn. 34:15-16.
paddy soil, Cu2+ in high-humus Andisol, and P in Andisol by Sato T, Shibuya K, Saigusa M, Abe T. 1993. Single basal applica-
the co-situs application of CAF containing these components. tion of total nitrogen fertilizer with controlled-release coated
From the foregoing, I conclude that co-situs application of CAF urea on non-tilled rice culture. Jpn. J. Crop Sci. 62:408-413.
with seeds or plant roots is a unique technology and is highly Shoji S, Gandeza AT. 1992. New concept of controlled release fer-
feasible in developing an innovative farming system. tilization. In: Shoji S, Gandeza AT, editors. Controlled release
fertilizer with polyolefin resin coating. Sendai (Japan): Konno
Printing Co. p 1-7.
References Tashiro T, Sugawara O, Naganoma H, Chiba K, Saigusa M. 2000.
Saigusa M, Hossain Md Z, Shibuya K. 2004. No-tillage transplant- Direct seeding of rice on Y-shaped furrow without tillage and
ing system of rice with controlled availability fertilizer in a covering soil in poorly drained paddy soil. Jpn. J. Crop Sci.
nursery box. J. Integ. Field Sci. 1:67-73. 69:547-553.
Saigusa M, Hossain Md Z, Tashiro T, Shibuya K. 1996. Maximizing
rice yield with controlled availability fertilizer and no-tillage Notes
culture in controlling environmental degradation. Proceedings
of the symposium on “Maximizing sustainable rice yields Author’s address: Field Science Center, Graduate School of Agri-
through improved soil and environmental management,” Khon cultural Science, Tohoku University, 232-3, Yomogida,
Kaen, Thailand. p 799-804. Oguchi, Naruko, Tamatukuri, Miyagi 989-6711, Japan, e-mail:
masa@bios.tohoku.ac.jp.
Soil organic matter (SOM) plays a central role in the storage in other regions and from other agrosystems in which crop
and release of essential soil nutrients into plant-available forms. residues also decompose anaerobically indicate recurring nu-
Its behavior in nutrient cycling is thought to be affected by its trient deficiencies, which can also reduce yield. The specific
chemical nature, although evidence is sparse. Recent work in nutrients that become deficient are all prone to chemical bind-
irrigated lowland rice (Oryza sativa L.) soils in the Philip- ing with SOM. Hence, the question arises whether anaerobic
pines suggests that the chemical nature of SOM depends in decomposition of crop residues inherently alters the forma-
part on whether its main parent material, crop residues, de- tion of new SOM and its chemical composition, which results
composes anaerobically or aerobically, raising the question in greater binding of new SOM with nutrients, inhibiting their
whether anaerobic decomposition contributed to a long-term release to available forms. This report will summarize exist-
yield decline. In addition, scattered reports from lowland rice ing knowledge on the occurrences of reduced nutrient avail-
ability in relation to the chemical nature of SOM in flooded Soil phenol accumulation and N mineralization were
soils, with emphasis on lowland rice cropping in tropical Asia measured in a 4-year study at IRRI that compared anaerobic
and the southern U.S. rice belt. decomposition of crop residues with aerobic decomposition
for double-cropped rice. In the third wet-season crop, urea fer-
tilizer labeled with 15N was applied at transplanting. Two weeks
Intensive lowland rice cropping in tropical Asia later, the amount of 15N that became immobilized in the MHA
A long-term decline in grain yield has been noted in some field was similar for both treatments. During the rest of the season,
trials under double- and triple-cropping of irrigated lowland 45% of the immobilized 15N was remineralized in the aerobic
rice (Cassman et al 1995), which involves regular anaerobic decomposition treatment, but only 8% was remineralized in
decomposition of crop residues and long-term submerged con- the anaerobic decomposition treatment (Olk and Cassman
ditions. The principal cause of the yield decline appears to be 2002). In-season mineralization of total N from the MHA was
decreased availability of soil organic nitrogen (N, Dobermann 22 kg N ha–1 less with anaerobic decomposition than with aero-
et al 2000), especially during late-season crop growth stages; bic decomposition. Phenols accumulated in the MHA under
the quantity of soil N did not decrease during the years of de- anaerobic decomposition, in tandem with the inhibition of N
clining yields and crop uptake efficiency of fertilizer N changed mineralization. The phenol accumulation was also associated
little. Hence, research efforts have focused on changes in the with inhibited mineralization of soil N across crop rotations
chemical nature of soil organic N. (rice-rice versus rice-maize, Zea mays L.), humic fraction, N
An accumulation of phenolic compounds has been asso- fertilizer rate, and year. A crop response to the N inhibition
ciated with this intensive cropping (Olk et al 1996). The phenols could not be demonstrated because N fertilizer was applied in
are found in incompletely decomposed lignin residues from synchrony with crop N demand.
the woody tissues of crop roots and stubble. Studying the young
mobile humic acid (MHA) fraction that was extracted from
both a triple-cropped rice soil and a nearby rice field main-
Continuous rice rotation in Arkansas, U.S.
tained under aerobic conditions at the International Rice Re- The conventional crop rotation in eastern Arkansas is rice-soy-
search Institute (IRRI), Schmidt-Rohr et al (2004) demon- bean (Glycine max (L.) Merr.), with one crop per year. Pend-
strated that the triple-cropped fraction had an agronomically ing market changes may promote continuous production of
significant excess (55 kg N ha–1) of organic N bound covalently rice in fields near water bodies. In a 4-year field trial, how-
in aromatic compounds, mainly with phenolic lignin residues ever, grain yield of a continuous rice rotation was 19% (1,350
in an anilide molecule. This form would cycle only slowly under kg ha–1) less than for rice yield following soybean (Anders et
flooded conditions. al 2004). Similar to the IRRI yield decline, agronomic symp-
In a double-cropped rice experiment conducted at three toms attribute this yield gap to a late-season N deficiency. In a
sites in the Philippines, phenols accumulated in the MHA frac- 15N-microplot study conducted in 2002, crop uptake at mid-
tion despite large differences among the sites in the degree of to late-season crop growth stages of unlabeled N, presumably
soil drying during two 2-month fallows (Olk et al 1998). Grain mineralized from SOM, was less for continuous rice than for
yields declined at the two sites with either the wettest or the rice following soybean (Table 1). The difference between crop
driest fallows, and decomposition of crop residues was anaero- rotations for uptake of labeled fertilizer N was less severe,
bic at all three sites. Hence, the key feature of continuous rice suggesting an inhibition of soil N mineralization in the con-
cropping that may promote phenol accumulation and decreased tinuous rice rotation. In-season degradation of soil phenols also
availability of soil N may be anaerobic decomposition of crop slowed under the continuous rice rotation, as their enrichment
residues, not the total duration of annual flooding. compared to rice following soybean was greater near harvest
than at an early growth stage (Fig. 1). This study is ongoing.
Session 12: Conservation of soil, water, and environment in rice cultures 375
Phenol content (g kg1 organic C)
18 18
Rice-soybean
Continuous rice
6 6
0 0
p-Hydroxyl Cinnamyl Total p-Hydroxyl Cinnamyl Total
Vanillyl Syringyl Vanillyl Syringyl
Phenol class Phenol class
Fig. 1. Soil contents of four phenol classes for rice-soybean and continuous rice rotations at two
sampling times in Stuttgart, Arkansas, 2002. Units are g phenol kg–1 soil organic carbon (C). DAE is
days after emergence. Standard errors are indicated above each bar.
Rice-wheat rotation in Japan crease in soil N uptake was associated with a loss of grain
yield.
A rice sterility problem has occurred in rice-wheat (Triticum
sp.) rotations in Osato, Saitama, Japan. In a greenhouse study,
Discussion
Noguchi et al (1997) associated the sterility with a copper (Cu)
deficiency that occurred with anaerobic decomposition of crop Evidence from multiple agrosystems is consistent with the
residues from either rice or wheat. Anaerobic soil conditions hypothesis that anaerobic decomposition of crop residues pro-
without incorporation of crop residues did not cause rice ste- motes an enrichment of phenolic lignin residues in newly form-
rility. Rice uptake of manganese, and to a lesser extent of other ing SOM, which can bind specific soil nutrients into less avail-
nutrients, was also less with anaerobic decomposition than with able forms. This hypothesis would account for the decreased
anaerobic conditions without crop residue incorporation. Cop- in-season N mineralization that was observed in rice soils of
per and manganese have a high affinity for binding with humic both the Philippines and Arkansas. Phenolic compounds are
molecules in laboratory conditions (Schnitzer and Hansen also capable of binding with Cu, which also becomes defi-
1970). Rice crop uptake of Cu and manganese was also dimin- cient with anaerobic decomposition.
ished by anaerobic decomposition of crop residues in the IRRI Demonstration of strong nutrient binding, at least for N,
field comparison of anaerobic decomposition with aerobic does not itself prove that anaerobic decomposition leads to
decomposition as described above (data not shown). Similarly, yield loss. Even strongly bound nutrient forms, most notably
Cu deficiencies in organic soils are well known; the Cu is the anilide N found by Schmidt-Rohr et al (2004), must miner-
thought to be bound to SOM by oxygenated functional groups, alize at some point; otherwise, total soil N would increase in-
such as phenolic hydroxyls. definitely with anaerobic decomposition, which is not the case.
Future research will explore the hypothesis that mineraliza-
tion of strongly bound nutrients is so slow as to become
No-tillage in temperate regions unsynchronized with crop nutrient demand, and as the strongly
In no-tillage systems of temperate regions, crop residues de- bound forms accumulate in soil, the soil nutrient supply also
compose to a larger extent during the springtime than under becomes increasingly unsynchronized, potentially diminish-
conventional tillage, which by contrast typically promotes au- ing the yield. Possible mitigation options include timely soil
tumn decomposition through postharvest tillage. In many tem- aeration, which would promote decomposition of the lignin
perate regions, abundant springtime precipitation saturates the residues and release of the bound nutrients. For example, rice
soils for extended periods of time. Consequently, decomposi- fields in Japan are commonly drained during later crop growth
tion of crop residues is more often anaerobic, or partially an- stages to promote soil N mineralization.
oxic, with no-tillage than with conventional tillage. In several How relevant is chemical stabilization of soil nutrients
North American studies involving 15N-enriched fertilizer, crop to yield trends for Asian rice farmers? Available data sets do
uptake of unlabeled N, presumably mineralized from SOM, not allow comment on whether on-farm yields or productivity
was significantly less with no-tillage than with conventional (ratio of yield to inputs) have declined since lowland rice crop-
tillage (Doran and Smith 1987). At some locations, the de- ping was intensified 30+ years ago. A definitive database would
Rotation cultivation of rice and upland crops in paddies has depends on soil texture, groundwater level, and cropping sys-
been adopted for the multiple use of land and has affected the tem. Nutrient contents depend on fertilization and the applica-
nutrient imbalance in the plant and soil. Multiple use of pad- tion of farming materials. Rotation cultivation affects these
dies has been required to adjust the balance between demand properties. Therefore, the influence of the paddy-upland rota-
and supply of rice in some countries. Crop growth primarily tion on soil physicochemical properties, and crop growth and
depends on soil moisture content, and secondarily on nutrient productivity, was studied in the Honam Plain area of the Re-
content. The distribution of soil moisture and nutrients greatly public of Korea.
Session 12: Conservation of soil, water, and environment in rice cultures 377
Table 1. Change in organic matter content by cropping system (g Table 2. Change in available phosphate content by cropping system
100 g–1). (mg kg–1).
Cropping system First Second Third Fourth Fifth Cropping system First Second Third Fourth Fifth
year year year year year year year year year year
Continuous rice 2.7 2.3 2.5 2.2 2.6 Continuous rice 188 170 148 137 139
Rice-barley 2.8 2.5 2.5 2.6 2.3 Rice-barley 105 145 161 134 131
Soybean-barley 2.6 2.2 2.4 2.3 2.3 Soybean-barley 127 175 168 176 274
Soybean-rice 2.3 2.2 2.6 2.6 2.4 Soybean-rice 154 153 145 144 189
Soybean 2 y-rice 3.9 2.1 2.5 2.2 2.3 Soybean 2 y-rice 171 117 168 151 206
Soybean 3 y-rice 3.0 2.3 2.1 2.0 2.2 Soybean 3 y-rice 157 148 159 171 192
Materials and methods creasing soil porosity was in the order of soybean for three
years-rice rotation, soybean-barley cropping system, soybean
The experimental field was installed with a sand-gravel drain- for two years-rice rotation, and soybean-rice rotation. Iwata et
age trench with 2-m intervals, 60-cm width, and 30-cm height al (1977) reported that soil porosity increased annually in the
at 1-m soil depth in 1982. After installation of the drainage paddy-upland rotation with the development of aggregate.
system, rice was cultivated in 1988. Treatments were continu- There was little difference in organic matter content with
ous rice cultivation, rice-barley cropping, soybean-barley crop- the continuous rice cultivation, rice-barley cropping system,
ping, soybean-rice rotation, soybean for two years-rice for one and rice-soybean cropping system. But organic matter content
year rotation, and soybean for three years-rice for one year decreased drastically with soybean cultivation for two years
cropping. Applied amounts of N-P2O5-K2O were 11-7-8 kg and three years (Table 1). Kitada et al (1992) mentioned that
ha–1 in rice, 4-7-68 kg ha–1 in soybean, and 15-10-88 kg ha–1 soil organic matter decreased gradually with easy mineraliza-
in barley. tion of organic matter in upland paddy conditions.
Soil available phosphate content decreased gradually
with continuous rice cultivation, but increased with rotation
Results cultivation of an upland crop in the paddy (Table 2). This re-
Rice yield increased after the rotation with soybean, and this sult was not related to fertilization amount but to phosphate
effect increased gradually with two- and three-year continu- dynamics. There was little difference in the amount of phos-
ous cultivation of soybean. This resulted from the increased phate fertilization among the crops. Phosphate is released in
number of panicles and number of grains per m2 by the high aerobic conditions and fixed in anaerobic conditions. This in-
content of NH4+-N in the soil. Meanwhile, rice yield decreased crement of phosphate in soybean cultivation resulted from
by 7% more in the barley-rice cropping system than in con- phosphate fixation due to aerobic conditions. Kitada et al
tinuous rice cultivation. (1992) also reported that availability of phosphate decreased
Soybean yield was higher in the annual rotation than in in upland conditions because of fixation to Fe and Al.
the biennial and three-year terms: 3.92 t ha–1 in annual, 3.56 t There was no change annually in Ca2+ of soil with con-
ha–1 in biennial, and 3.72 t ha –1 in the three-year interval. This tinuous rice cultivation, but the content increased after upland
effect resulted from the changing number of pods and grains cultivation. Meanwhile, soybean cultivation caused an increase
per plant as time passed. Park et al (1993) reported that soy- in Ca2+ content in double cropping and rotation cultivation.
bean yield decreased greatly in 4-year continuous cultivation Exchangeable K+ content showed a small decrease in continu-
in sandy loam paddy and 3-year continuous cultivation in silt ous rice cultivation, but an increase in rotation cultivation and
loam paddy. Motomatsu (1990) also reported that soybean yield a clearer increase in soybean-barley and soybean two years-
decreased by 10% in 2-year continuous cultivation and by 20% rice cultivation rotation. This stable content of Ca2+ and K+ in
in 3-year continuous cultivation because of the reduction in continuous rice cultivation implied that submerged rice culti-
soil N content and root activity and increment in disease and vation had a positive effect on the balance of Ca2+ and K+ in
pest injury with continuous cropping. the soil.
Barley yield increased by 4–7% in the soybean-barley The content of ammonium nitrogen was high in the or-
cropping system vis-à-vis the rice-barley cropping system. der of soybean-rice, soybean two years-rice cultivation, rice-
Therefore, this was a positive effect of soybean culture on bar- barley cropping system, soybean-barley cropping system, and
ley yield. Meanwhile, the annual yield of barley decreased continuous rice cultivation. The released amount of inorganic
gradually in both cropping systems as time passed. nitrogen in the rotational paddy was influenced by the soil tex-
Upland crop cultivation increased soil porosity as time ture, organic matter content, period of rotation, and cropping
passed. The effect was higher in 2-year and 3-year cultivation system (Watanabe 1985, Kogano 1987). Mineralized nitrogen
of upland crops than with the annual rotation of rice. The in- was higher in upland culture than in continuous rice culture.
In Japan, rice production has tended to exceed domestic de- region of Japan, where the rotation cycles are one year of up-
mand since around 1970, and the rice production adjustment land for one to several years of paddy. Although there is no
by the government has already continued for more than 30 years. clear evidence, this decrease in productivity could be attrib-
In 2004, the area of rice cultivation is restricted to about 60% uted to a decrease in soil fertility. Therefore, to promote paddy-
of all paddy fields (i.e., 1.6 million ha). The government has upland rotation farming, sustainability of crop production in
recommended several types of alternative paddy field uses with the successive paddy-upland rotation system should be inves-
subsidiary payments to farmers, especially alternative crop- tigated from a long-term view.
ping of soybean, wheat, or fodder crops to increase self-suffi-
ciency rates of these crops.
Materials and methods
It is known that the yields of soybean and/or wheat de-
crease when these crops are continuously cultivated in the up- The changes in soil fertility, soybean productivity, and paddy
land fields that have been converted from paddy fields for many rice productivity were investigated by long-term field experi-
years. However, this injury from continuous cropping can be ments under single cropping in a year on gray lowland soil in
avoided by a paddy-upland (irrigated paddy rice and upland the cool region of Japan (National Agricultural Research Cen-
crop) rotation and the rotation leads to high yields of both rice ter for Tohoku Region, NARCT, Omagari, Akita).
and upland crops. Therefore, the paddy-upland rotation has In experiment A, a field experiment in combination with
been encouraged. Several previous studies on the combination different paddy/upland periods and organic matter application
of upland period and paddy period showed that more than 2–3 has been conducted since 1990. In a short-term upland rota-
years of paddy period is required for 2–3 years of upland pe- tion treatment, soybean (Glycine max (L.) Merr.) and paddy
riod to keep crop productivity and soil nutrient dynamics suit- rice (Oryza sativa L.) were grown with a cycle of 1–2 years of
able (Takahashi 1983, Hanai 1987, Tsukuda 1990, Kitada et al upland and 1–3 years of paddy, respectively. A cycle of 3–4
1993). Recently, it has been pointed out that soybean produc- years of upland and 1–2 years of paddy was adopted in a me-
tivity decreased at some production sites, especially in the warm dium-term upland rotation treatment. As a control, continuous
Session 12: Conservation of soil, water, and environment in rice cultures 379
paddy rice cropping was done in a continuous paddy treat- Available soil nitrogen (index)
ment. Rice straw was applied at 0 or 6 t ha–1 in each treatment 140
after harvesting every autumn. 127
In experiment B, the experimental paddy field was con-
120
verted to an upland field in 1982 and soybean was cultivated
Starting year
during 18 years. After that, the field was returned to paddy and
paddy rice was cultivated for three years. Then, in 2003, it was 100
reconverted to an upland field for cultivating soybean. As a
100
control, a continuous paddy treatment was prepared and paddy
rice was cultivated there. Rice straw compost was applied at 0 80 92
9 years after starting
or 20 t ha–1 to both treatments before plowing every spring. (after rice-rice)
The average of total nitrogen (N) content of the applied rice
60 75
straw compost was 5.5 mg g–1 (fresh-weight basis) for the 20 13 years after starting
years, and it was 3.5 mg g–1 for the last several years. The C/N (after SB-SB-SB (or R)-SB) 71 72
ratio of the rice straw compost was about 20. 40
Medium-term U/+RS
Short-term U/+RS
RS
RS
RS
Continuous P/+RS
For soil fertility, soil samples were collected from the
plow layer in 1982 (only in experiment A), 1991, and 1999-
2003. All the soil samples were air-dried and sieved (< 2 mm).
Total carbon (C) and N content of the soil were determined by
using the CN-Analyzer (Yanaco MT-700). The available N in
the soil was measured by the soil incubation method (under
submerged conditions at 30 ºC for 4 wk). Fig. 1. Changes in available soil nitrogen in the successive paddy-
upland rotation system. Available soil nitrogen (index): relative %
of that in the continuous paddy field without rice straw application
Results and discussion in the same year. P = paddy, U = upland, RS = rice straw, SB =
soybean.
Changes in soil fertility
In the continuous paddy fields of experiments A and B, the
total C, total N, and available N of soil were maintained at the bean was planted after a sufficient paddy period (the yield trial
initial levels without organic matter application, and were in- by the Soybean Breeding Lab, NARCT). In contrast, the yield
creased by organic matter application. of paddy rice in both cases (paddy-upland rotation and long-
When the successive paddy-upland rotations were con- term upland conversion) was higher than that in the continu-
tinued for 10 years or more, whether the rotation was for the ous paddy field.
medium term or for the short term, a decrease in total C of the Rice straw application in the successive paddy-upland
soil was observed and it wasn’t made up for by rice straw ap- rotation produced no increase in soybean yield, whereas the
plication at 6 t ha–1. For N fertility, although the total N and rice straw compost application in long-term upland conver-
available N of the soil were maintained roughly at their initial sion increased soybean yield slightly. According to the experi-
levels by rice straw application for the short-term upland rota- ment on the topdressing of N fertilizer, the N topdressing at 70
tion, the decrease in total N and available N of the soil wasn’t kg ha–1 scarcely improved the yield of soybean planted in the
fully mitigated by the rice straw application in the medium- upland field where the decrease in soil N fertility occurred.
term upland rotation (Fig. 1). These results implied that, in response to a decrease in
Cultivating soybean for 18 years in the upland field con- soil N fertility, soybean yield in the successive paddy-upland
verted from paddy field (long-term upland conversion) de- rotation system decreased. In contrast, in spite of the consid-
creased the available N of the soil more than the paddy-upland erable decrease in soil N fertility, the yield of paddy rice in the
rotation. The available N of the soil in the long-term upland paddy field just returned from an upland field became higher
conversion, even with application of rice straw compost at 20 than that in the continuous paddy field.
t ha–1, was markedly lower than that in the continuous paddy
field where no organic matter was applied. By returning to a
References
paddy field from long-term upland conversion and keeping it
as a paddy field for three years, the available soil N recovered Hanai Y. 1987. Cropping systems in paddy field for multiple use.
up to about 25% of the decrease caused by the 18-year long- Res. J. Food Agric. 10(9):28-32. (In Japanese.)
term upland conversion. Kitada K, Shimoda H, Kamekawa K, Akiyama Y. 1993. Changes in
soil nutrients affected by rotation of upland and paddy crop in
gray lowland soil and search for the most suitable term of
Changes in crop production related to soil fertility
rotation. Jpn. J. Soil Sci. Plant Nutr. 64:154-160. (In Japa-
As Figure 2 shows, soybean yield in the successive paddy-
nese with English summary.)
upland rotation and long-term upland conversion decreased
by about 5–25% compared with that in the field where soy-
100
75
Cadmium (Cd) is principally dispersed in natural and agricul- being from 0.2 to 0.4 mg kg–1 the weight of polished rice.
tural environments through human activities. Japanese arable According to a survey of Cd contamination in rice by the Min-
lands, especially paddy fields, are contaminated to some ex- istry of Agriculture, Forestry, and Fisheries (MAFF) of Japan,
tent by Cd through irrigation with river water originating from 3.3% or 0.3% of rice cultivated in Japan exceeded a limit of
mines or through emissions from metal-smelters. Recently, the 0.2 mg kg–1 or 0.4 mg kg–1, respectively. Although our diet is
joint FAO/WHO Expert Committee on Food Additives has below the PTWI of Cd (4 µg kg–1), rice is the largest source of
established a Provisional Tolerable Weekly Intake (PTWI) of dietary intake of Cd for the Japanese people, so it is urgent to
Cd at 7 µg kg–1 body weight per week. Based on this risk as- reduce the Cd concentration of rice. In this paper, current ame-
sessment of Cd, the Codex Alimentarius Commission will pro- liorating techniques being applied in Cd-polluted paddy fields
pose new criteria of Cd concentration in foods, especially in and promising techniques for reducing Cd contamination in
staple crops. The standard of Cd for rice has been discussed as rice are discussed.
Session 12: Conservation of soil, water, and environment in rice cultures 381
Current techniques for reducing Cd contamination in rice Table 1. Estimation of Cd removal from polluted
soils using a high Cd-accumulating rice variety,
Milyang 23.
Soil dressing
In 1971, the Japanese government enforced the law that paddy Item Amount
fields that had produced brown rice beyond 1 mg kg–1 were
subject to preventive measures against soil pollution such as Cd concentration of polluted soil 0.7 mg kg–1
soil dressing. There are two types of soil dressing: (1) remov- Average of Cd concentration 0.3 mg kg–1
of nonpolluted soils
ing the Cd-polluted soil layer and replacing it with nonpolluted Quantity of Cd to be removed per ha 400 g ha–1
soils and (2) covering Cd-polluted soils with nonpolluted ones. Cd concentration of shoots 15 mg kg–1
According to a survey of rice Cd concentration after soil dress- Dry matter weight 800 kg ha–1
ing by which the Cd-polluted soil layer with more than 25 cm Cd uptake by shoots 120 g ha–1
was replaced, it did not exceed 0.03 mg kg–1 over more than Years for phytoremediation to 3–4 years
be reached in nonpolluted soils
20 years (Kuwana et al 2003). Thus, soil dressing is a perma-
nent technique for reducing Cd contamination in rice, but an
alternative approach is needed because of its enormous cost
($300,000–500,000 per ha). shoots and hypertolerance of Cd. However, it is questionable
whether hyperaccumulator plants are applicable for
Water management phytoremediation in Cd-polluted paddy fields in Japan. The
The solubility of Cd depends on the redox potential in the soil. primary reason is that appropriate agricultural practices such
Bioavailable Cd in the soil decreases to a great extent under as planting, fertilization, water management, and mechanical
submerged conditions because of the formation of less-soluble harvesting are not well established for hyperaccumulator plants.
cadmium sulfide (CdS). Pot and field experiments conducted In addition, the main targeted areas in Japan for
by many researchers showed that keeping a submerged condi- phytoremediation have slight and moderate contamination of
tion for 3 weeks before and after heading time is a very useful paddy fields, which might potentially produce paddy rice or
method for reducing the Cd concentration of rice. Even with upland crops contaminated with Cd. Under such lands, it is
such conditions, however, rice grains are always exposed to unknown whether hyperaccumulator plants can have an ex-
the threat of Cd contamination because high levels of Cd still traordinary ability for Cd uptake. Our research group found
remain in the soil. Besides, the effect of submergence varies that some indica-japonica cross varieties cultivated under up-
among years depending on weather conditions (e.g., rainfall). land conditions accumulated a substantially high concentra-
tion of Cd in grains and shoots. When the shoot Cd concentra-
Application of soil amendments tion was compared between several rice varieties and B. juncea
The application of alkaline amendments is one of the practi- grown in two types of Cd-polluted soils, its concentration was
cable methods for reducing the Cd concentration of rice. The 2 to 4 times higher in rice varieties than in B. juncea. More-
combined use of calcium-silicate and fused magnesium phos- over, Cd uptake in the shoots was approximately 6 times higher
phate for a basal dressing showed positive effects on suppress- in rice varieties than in B. juncea. A remarkable decrease in
ing Cd uptake by paddy rice (Takijima and Katsumi 1973). Cd content in the rhizosphere was found after harvesting rice.
However, no effects or negative effects of alkaline amendments There are many merits for using rice as a cleanup plant, in
on Cd uptake in rice were also reported. Currently, the combi- addition to its superior ability of Cd uptake. First, the cultiva-
nation of water management and application of soil amend- tion technique for rice is well established. Second, farmers
ments for suppressing Cd contamination of rice is widely prac- can use their own machines to plant or harvest high Cd-accu-
ticed in Cd-polluted paddy fields in Japan. mulating rice varieties without any new investment. Third, they
can plant a popular commercial variety immediately after rice-
based phytoremediation.
Promising techniques for reducing Cd contamination in rice Based on the results of a pot experiment, estimation of
Phytoremediation Cd removal from a low-level polluted soil (0.7 mg kg–1 of 0.1
Phytoremediation, which is the use of plant systems to M HCl extractable Cd in the soil) using Milyang 23, which is
remediate contaminated environments, has drawn attention as one of the high Cd-accumulating rice varieties and also a high-
an environmentally friendly and cost-effective technique. The yielding variety, is shown in Table 1. If the shoot biomass of
most important point for developing a successful Milyang 23 is estimated at 0.8 t ha–1, the Cd concentration of
phytoremediation (phytoextraction) technology in Cd-polluted the soil could be reached at close to 0.3 mg kg–1, which is the
soils is to select promising cleanup plants. Hyperaccumulator average Cd concentration of nonpolluted soils, by successive
plants such as Thlaspi caerulescens (Brown et al 1994) and planting for 3 to 4 years. A high removal efficiency of Cd by
Brassica juncea (Salt et al 1995) have been exploited as ideal Milyang 23 was also confirmed by a field experiment, in which
plants for phytoremediation because they possess several pref- approximately 190 g ha–1 of Cd was removed from a paddy
erable characteristics for phytoremediation: the ability to take field polluted with 2.5 mg Cd kg–1 soil. These results suggest
up and translocate an exceedingly large amount of Cd to their that high Cd-accumulating rice is an ideal plant for
4.0
3.5
3.0
SL204 SL215
2.5
SL218
Y = 7.62x + 1.07
SL219
SL205
Kasalath R2 = 0.129
SL225
2.0 SL233
SL238 SL212 SL232
SL202 SL230
SL206
SL239 SL235 SL220 SL231SL234
SL214 SL228
1.5 SL221
SL213Koshihikari SL226 SL227 SL222
SL229 SL203
SL209
SL224 SL211 SL236
1.0 SL223
SL207 SL208
0.5
0.0
0.00 0.05 0.10 0.15 0.20
Grain Cd concentration (mg kg1) in paddy conditions
Fig. 1. Cd concentration of grains in 39 chromosomal substitution lines carrying Kasalath (an
indica rice variety) chromosomal segments in a Koshihikari (japonica) background. These
lines were grown on Cd-polluted soils (1.8 mg Cd kg –1 soil) under upland and continuous
paddy conditions. The concentration of Cd in grains was analyzed by inductively coupled plasma-
mass spectrometry after digesting with an acid mixture.
phytoremediation of Cd-polluted paddy fields in Japan. For a ries of SLs has covered the entire genome with overlapping
successful decontamination by phytoremediation, it is impera- introgressed segments of each line, and it is a powerful tool to
tive to establish postharvest technology for handling cleanup simply detect the QTLs. The parents and 39 SLs were grown
plants, which contain a large amount of Cd. One way is to on Cd-polluted soils (1.8 mg Cd kg–1 soil) under upland or
incinerate the cleanup plants, turning them into ash, then re- continuous paddy conditions. Figure 1 shows the Cd concen-
cover the Cd from the ash. The other way is to use the plant tration of their grain. The average grain Cd concentration was
biomass for a bio-energy such as bio-methanol after recover- 20 times higher in upland conditions (1.67 mg kg–1 weight of
ing the Cd. The National Institute for Agro-Environmental brown rice) than in paddy conditions (0.08 mg kg–1). The grain
Sciences (NIAES) has organized a research consortium for a Cd concentration in Koshihikari was significantly lower than
rice-based phytoremediation project. that of Kasalath in both conditions. Although there is no cor-
relation between the grain Cd concentration in rice grown in
Introduction of low Cd-accumulating rice varieties paddy conditions and rice in upland conditions, several SLs
Grain Cd concentration varies greatly among rice varieties (SL 207, SL 224, and SL 223) showed significantly lower grain
(Arao and Ae 2003). This suggests the possibility of breeding Cd concentration than Koshihikari under both conditions.
varieties with less grain Cd concentration. In durum wheat, Based on graphical genotypes of SLs, the putative QTLs con-
Clarke et al (1997) reported that the low grain Cd trait is highly trolling low grain Cd were detected on chromosomes 3 and 8.
heritable and is controlled by a single dominant gene. Recently, As the detected SLs represent more than 90% of the Koshihikari
their work suggested that the Cd uptake locus resides on chro- genetic background, the development of a new Koshihikari
mosome 5B. A DNA marker, which was linked in repulsion with low grain Cd should be feasible in the near future.
with a low-Cd allele, has been developed for an efficient wheat
breeding program. Although the genetic analysis of rice is much
References
more developed than that of other crops, there is no useful
information on the low-Cd trait of rice. We tried to identify the Arao T, Ae N. 2003. Genotypic variations in cadmium levels of rice
quantitative trait loci (QTLs) controlling the Cd concentration grain. Soil Sci. Plant Nutr. 49:473-479.
of rice grain. Substitution lines (SLs) carrying Kasalath (an Brown S, Chaney R, Angle J, Baker AJM. 1994. Phytoremediation
indica rice variety) chromosomal segments in a Koshihikari potential of Thlaspi caerulescens and bladder campion for
zinc- and cadmium-contaminated soil. J. Environ. Qual.
(japonica) background were used to identify the QTLs. A se-
23:1151-1157.
Session 12: Conservation of soil, water, and environment in rice cultures 383
Clarke JM, Leisle D, Kopytko GL. 1997. Inheritance of cadmium Takijima Y, Katsumi F. 1973. Cadmium contamination of soils and
concentration in five durum wheat crosses. Crop Sci. 37:1722- rice plants caused by zinc mining. IV. Use of soil amendment
1726. materials for the control of Cd uptake by plants. Soil Sci. Plant
Kuwana T, Aoyama Y, Tsudaka H, Yoshikura J, Adachi I. 2003. A Nutr. 19:235-244.
survey of paddy field past to 20 years of soil dressing. Abstr.
Annu. Meet. Jpn. Soc. Soil Sci. Plant Nutr. 41:288. (In Japa-
nese.) Notes
Salt DE, Prince RC, Pickering IJ, Raskin I. 1995. Mechanisms of Author’s address: National Institute for Agro-Environmental Sci-
cadmium mobility and accumulation in Indian mustard. Plant ences, 3-1-3 Kannondai, Tsukuba, 305-8604, Japan, e-mail:
Physiol. 109:1427-1433. isatoru@niaes.affrc.go.jp.
The application of fresh plant residue, such as wheat straw, Distribution of tracer 15N at 24 h after application A
often inhibits the growth and N uptake of paddy rice (Oryza
sativa L.) plants in the establishment phase. This inhibition is Without straw
generally attributed to a deficiency of available N in soils with
straw added because of considerable N assimilation by the soil With straw
microbes. This phenomenon is called biological N immobili-
zation. In our previous study, however, we reported some cases 0 20 40 60 80 100
in which much mineral N existed in paddy soils with added Distribution (%)
wheat straw, even though the growth of rice plants was de- Root Top Mineral N in soil
layed (Watanabe et al 1988). Here, N uptake activities of both
rice plants and soil microbes were measured with 15N tracing
Nontracer N uptake of rice plants B
to validate that the widely accepted biological N immobiliza-
tion theory was precisely the cause of rice growth inhibition
Without straw
by fresh wheat-straw application. Furthermore, the inhibitory
substances were identified and their concentration determined
in soil solutions under various conditions. With straw
0 10 20 30 40 50
Validation of the biological N immobilization theory Nontracer N uptake (mg pot1)
Root Top
To analyze the behavior of soil mineral N that existed at a
given time, 15N tracer experiments were conducted (Tanaka
and Nishida 1996). Rice seedlings were transplanted into a Tracer and nontracer mineral N in soil C
pot with flooded soil, to which was applied wheat straw or no
wheat straw. At 17 days after transplanting (DAT), tracer 15N, Without straw
15NH Cl solution, was added to the soil using a syringe. The
4
soil and rice plants were sampled at 24 h after application of With straw
the tracer to avoid remineralization of the immobilized tracer
15N. The plants were divided into tops and roots. The N con- 0 20 40 60 80
tent and 15N abundance were analyzed on plant and soil Mineral N soil (mg pot1)
samples. The behavior of mineral N in the soil upon the appli- Nontracer Tracer
cation of the tracer can be estimated by using the ratio of dis-
Fig. 1. Distribution of tracer and nontracer N in the rice plant and
tribution of 15N in plants and soil. soil. Rice plants and soils were sampled at 17 days after trans-
Figure 1A shows that the uptake of 15N of rice in a plot planting. Tracer N represents the behavior of mineral N at the
receiving straw was about half of that in the control plot, even application of tracer (A), nontracer N represents that which was
derived from basal fertilizer and soil (B), tracer and nontracer min-
though the mineral 15N was larger. In this experiment, the total eral N in soil (C).
weights of rice plants, N uptake, and N content in the straw
plot were smaller than those in the control. A decrease in 15N
Session 12: Conservation of soil, water, and environment in rice cultures 385
air phase, and low temperature, were closely related to the Tanaka F. 2002. Formation of aromatic acids and growth inhibition
increase in the accumulation of aromatic acids. These trends of rice (Oryza sativa L.) plants in flooded soils with wheat
agreed with the trends in the appearance of growth inhibition straw added. Bull. Natl. Agric. Res. Cent. Kyushu Okinawa
of rice plants. Furthermore, it is noteworthy that the aromatic Reg. 40:33-78. (In Japanese.)
Tanaka F, Nishida M. 1996. Inhibition of nitrogen uptake by rice
acids described here were not decomposition fragments de-
after wheat straw application determined by tracer NH4+-15N.
rived from lignin of wheat straw but metabolites from soil
Soil Sci. Plant Nutr. 42:587-591.
microbes under anaerobic conditions. Tanaka F, Nishida M. 1998. Inhibitory effects of aromatic acids on
In conclusion, rice growth inhibition in the presence of nitrogen uptake and transport in rice (Oryza sativa L.) plants
wheat straw cannot be associated with biological N immobili- cultured on hydroponics. Soil Sci. Plant Nutr. 44:691-694.
zation but with N uptake inhibition. Aromatic acids detected Tanaka F, Ono S, Hayasaka T. 1990. Identification and evaluation of
from soil solution with wheat straw inhibited rice root elonga- toxicity of rice root elongation inhibitors in flooded soils with
tion and N uptake. Their concentration and accumulation pat- added wheat straw. Soil Sci. Plant Nutr. 36:97-103.
terns provide an important explanation for rice growth inhibi- Watanabe F, Ono S, Hayasaka T. 1988. Effect of wheat straw appli-
tion. cation on accumulation of organic acid in paddy soil and rice
growth in early stage. Kyushu Agric. Res. 50:87. (In Japa-
nese.)
References
Tanaka F. 2001. Growth inhibition of rice (Oryza sativa L.) plants in Notes
flooded soils with added fresh wheat straw: the delicate inter-
action between rice plant and soil microorganisms. Author’s address: National Agricultural Research Center, 3-1-1
Nougyougijutu 56:506-510. (In Japanese.) Kannondai, Tsukuba City, Ibaraki 305-8666, Japan, e-mail:
fukuyot@affrc.go.jp.
Evaporation and transpiration from a paddy field are very im- curs and traditional and complicated irrigation systems are
portant for regional water balance, and therefore irrigation widely developed. Such conditions may disturb the formation
management for rice cultivation, especially in monsoon Asia. process of the isotope composition of paddy water, but few
Stomatal transpiration is accompanied by CO2 uptake for pho- studies have been carried out to investigate the relationship
tosynthesis and is essential for plant growth, whereas evapo- between the isotope composition and evaporative loss and to
rative loss of ponded water would moderate diurnal variations confirm the possibility to separate evaporation and transpira-
of temperature and moisture around the paddy field. These tion from a paddy field in the moist region. In this study, we
two vapor fluxes into the atmosphere are often integrated and investigated stable isotope ratios of paddy water samples and
called evapotranspiration; nevertheless, they are essentially evaluated the effect of evaporation to examine the potential of
different processes. One of the reasons is the difficulty of de- the use of environmental isotope analysis.
termining these fluxes separately by micrometeorological meth-
ods such as eddy correlation, which has been widely used to
Methods
directly measure evapotranspiration from various terrestrial
ecosystems. Field observations were made in a rectangular lot of a paddy
As another approach, isotope hydrological techniques field (36°03′N, 140°01′E, 15 m above sea level; 100 m long
could divide evapotranspiration into evaporation and transpi- and 54 m wide) located in Mase, Tsukuba, Japan. Rice is usu-
ration because only evaporation can affect the isotope compo- ally cultivated from May to August in the paddy field; the rest
sition of ponded water. For example, Simpson et al (1992) of year has no cultivation. We collected paddy water samples—
successfully calculated the proportion of evaporation occu- irrigation water, drain water, and ponded water—three times
pied in total evapotranspiration using isotope analysis, at a during cultivation: in mid-May, early in June, and at the be-
rice paddy field located in the semiarid region of southeast ginning of August in 2002. The ponded water samples were
Australia. In their study, the water balance of the paddy field collected at six places distributed in the paddy field to detect
was considerably simple: a small amount of rainfall less than isotopic change during surface water movement. Stable iso-
10% of irrigation, no surface drainage, and a constant ponding tope ratios of hydrogen (δD) and oxygen (δ18O) of the paddy
depth. In the moist Asia region, in contrast, much rainfall oc- water samples were determined with a dual-inlet-type isotope
PW #2
PW #3
PW #4
PW #5
PW #6
PW #1
DR
IR
ferences between irrigation water (IR) and drain water (DR)
reached 6.1‰ and 5.3‰, respectively. In August, however,
the increase in δ18O was little and the difference between IR Fig. 1. Change in δ18O of paddy water samples along the flow di-
and DR was only 0.6‰. The pattern of δD change along the rection. IR = irrigation water, DR = drain water, PW = ponded
water, and #n = the order of PW samples arranged along the wa-
flow direction was similar to that of δ18O. ter movement.
Figure 2 is a δ18O vs δD plot for the paddy water samples.
A local meteoric water line (LMWL), which was determined
from a long-term event-basis water collection by Yabusaki et
al (2003) and indicates that the isotope composition of rain-
water around the study area will be plotted along the line, is
also shown. The paddy water samples were plotted almost lin- d18O ( )
early for each month. The result of linear regression analysis 8 6 4 2 0
suggested good correlation between δ18O and δD (the correla-
tion coefficients were larger than 0.98), and the slopes of the
regression lines were 4.3, 4.3, and 5.8 for May, June, and Au-
LMWL (dD = 7.4 d18O + 8.5)
gust, respectively, showing smaller values than the slope of
LMWL (7.4). 20
Discussion
The results of the stable isotope analysis illustrated in Figures
1 and 2 clearly indicate that the isotope composition of paddy 40
water was intensively affected by evaporation, especially in
May and June. When water vaporizes, water molecules that
consist of lighter isotopes (H and 16O) tend to become vapor
15 May
more than those including heavier isotopes (D and/or 18O). 7 June
This causes isotope fractionation between liquid and vapor 1-3 August 60
phases; therefore, ponded water became more and more en-
dD ( )
riched isotopically as the water moved downstream under the
influence of evaporation. In addition, the relationship between Fig. 2. A δ18O
vs δD plot for the paddy water samples. The local
δ18O and δD of the water affected by evaporation is usually meteoric water line (LMWL) in Tsukuba after Yabusaki et al (2003)
is also shown.
different from that of the water not affected. According to pio-
neering work by Craig (1961), most meteoric water samples
collected all over the world were plotted linearly in the δ18O
vs δD plot (the global meteoric water line, GMWL) and the an evaporation line (EL), and it results from the difference in
slope of GMWL was 8. But the samples collected from lakes isotope fractionation ratios (18O/16O vs D/H) between evapo-
and rivers in East Africa, where isotope compositions were ration of surface water and condensation of meteoric water.
strongly affected by evaporation, showed another linear rela- The slopes of ELs obtained in the paddy field (4.3–5.8) are
tionship and the slope was about 5. This line is generally called therefore evidence of the influence of evaporation.
Session 12: Conservation of soil, water, and environment in rice cultures 387
In August, the increase in δ18O was much smaller than ine the relationship among the isotope composition of paddy
in May and June, while the slope of EL was largest, although water, the components of water balance, and the stage of rice
evapotranspiration determined by eddy correlation was rela- growth more precisely, and to construct an isotope hydrologi-
tively similar (3–5 mm day–1 for all months). Comparing the cal model that can calculate the ratio of evaporation vs tran-
condition of plant growth in these months, the values of LAI spiration from a paddy field, considering the influences of rain-
in the paddy field were 0.05, 0.73, and 5.16 m2 m–2, respec- fall and relatively complex irrigation management.
tively. The much larger LAI in August than in the other months
implies that active transpiration and the depression of evapo-
rative loss by the grown rice canopy would occur. Under such
References
conditions, transpiration from the rice canopy must be pre- Craig H. 1961. Isotopic variations in meteoric waters. Science
dominant. By transpiration, indeed, isotope fractionation oc- 133:1702-1703.
curs between the water in a leaf and the transpired vapor, but Simpson HJ, Herczeg AL, Meyer WS. 1992. Stable isotope ratios in
the isotopically enriched leaf water is never brought back to irrigation water can estimate rice crop evaporation. Geophys.
Res. Lett. 19:377-380.
the ponded water. Consequently, the isotope composition of
Yabusaki S, Tase N, Kihou N, Yuita K. 2003. Characteristics of oxy-
paddy water in August changed little in contrast to May and
gen isotopic ratios of soil water and groundwater at paddy,
June, when significant evaporation considerably advanced the upland and grove fields in Tsukuba, Ibaraki Prefecture, Cen-
isotopic enrichment. tral Japan. J. Jpn. Assoc. Hydrol. Sci. 33:161-176. (In Japa-
nese with English abstract.)
Conclusions and future studies
As a result of this study, it was suggested that isotopic enrich- Notes
ment of paddy water is closely related to direct evaporation of Authors’addresses: Yohei Hamada, Terrestrial Environment Research
ponded water, rather than the total evapotranspiration from Center, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, 305-
the paddy field. This supports the possibility to separate evapo- 8577 Japan; Shiho Yabusaki and Norio Tase, Graduate School
ration and transpiration using the isotope hydrological ap- of Life and Environmental Sciences, University of Tsukuba,
proach, usually impossible by micrometeorological measure- 1-1-1 Tennodai, Tsukuba 305-8572, Japan; Ichiro Taniyama,
ments. For more detailed and quantitative analysis, we con- National Institute for Agro-Environmental Sciences, 3-1-3
Kannondai, Tsukuba 305-8604, Japan, e-mail:
ducted intensive field observations of water balance and the
hamada@suiri.tsukuba.ac.jp.
collection of water samples in the same paddy field during
rice cultivation in 2004. Based on the data, we intend to exam-
Sawah, also known as a paddy field, has been sustainably pro- the existence of sawah with its multifunctionality can be main-
ducing rice, the staple food for most Indonesians, for hundreds tained.
or perhaps even thousands of years. Its other functions, espe-
cially in producing environmental services, are indispensable,
Objective
yet have not been recognized or are ignored by most stake-
holders. The objectives of this study were to assess soil loss and water
Sawah areas are characterized by high population den- retention capacity (as an indicator of flood mitigation func-
sity and perfectly flat plots. The sawah plot size is usually small, tion) of sawah relative to that of other land-use systems.
especially on sloping land. As such, sawah farming has low
efficiency, resulting in low profitability. Low-incentive sawah
Materials and methods
farming results in accelerating conversion of sawah to nonag-
ricultural uses. Location
This study evaluated the environmental functions, espe- Erosion from sawah was measured in Ungaran (07o20′S;
cially soil loss and water-retaining capacity, of sawah relative 110oE), Central Java Province. The soil subgroup in the sawah
to other land uses in Java in 2001. The results are important area was classified as Typic Tropaquepts. A water retention
for advocating internalization of environmental services pro- study was conducted in the Upper Citarum Watershed
duced by sawah in national land-use-related policies, such that (6o40′30″–7o15′00″S; 107o30′00″–107o55′00″E), West Java
Session 12: Conservation of soil, water, and environment in rice cultures 389
WRC (m) References
0.16 Agus F, Vadari T, Watung RL, Sukristiyonubowo, Valentin C. 2003.
0.14 Interception capacity
0.12 Ponding capacity Effects of land use and management systems on water and
0.10 Pore absorption sediment yield: evaluation from several micro catchments in
0.08 Southeast Asia. In: Maglinao AR, Valentin C, Penning de Vries
0.06 F, editors. Proceedings of IWMI-ADB Project Annual Meet-
0.04 ing and 7th MSEC Assembly, Vientiane, Lao PDR, 2-7 Dec.
0.02
2002. Bangkok (Thailand): International Water Management
0.00
Forest Sawah Mixed Annual Housing Industry Institute, Southeast Asia Regional Office. p 135-149.
cropping upland Nishio M. 1999. Multifunction character of paddy farming. Paper
Fig. 1. Water retention capacity (WRC) of sawah and other land- presented at the Second Group Meeting on the Interchange of
use systems. Agricultural Technology Information between ASEAN Mem-
ber Countries and Japan, Jakarta, Indonesia, 16-18 February
1999.
Conclusions van Dijk AJM. 2002. Water and sediment dynamics in bench-ter-
raced agricultural steeplands in West Java, Indonesia. PhD
Sawah is superior in controlling erosion to other land uses. dissertation. Vrij University, Netherlands. 363 p.
Erosion mainly happens in sawah during and shortly after till-
age operations.
The capacity of sawah to retain water during and shortly Notes
after rainfall is comparable with that of the tree-based mixed Authors’ address: Soil Research Institute, Jln. Juanda 98, Bogor
cropping system and significantly higher than that of annual 16123, Indonesia, e-mail: f.agus@cgiar.org.
crop-based and housing and industrial areas. Policy measures
should be implemented to maintain the existence of sawah con-
sidering the significant environmental services it can offer, in
addition to its tangible role as a rice producer.
Wrap-up of Session 12
Although it is recognized that the cultivation of paddy rice sus- Revolution with paddy rice farming would take place in this area
tains soil fertility, there is a large area whose fertility of soil is still with the proper integration of forestry, upland farming, and low-
low. Therefore, it is necessary to increase both the sustainability land rice farming.
and productivity of these soils urgently. One of the main themes To optimize fertilizer management in such diverse paddy
of the conference is “the development of sustainable rice culti- soils, C. Witt (Singapore) proposed an effective and site-specific
vation based on environment and food security.” In this session, management of nutrients on the basis of his field trials with fer-
this theme was discussed from the viewpoint of soil and water tilizer application.
conservation. Recently in China, rice yield has increased dramatically with
Participants entered into “the world of paddy soils” with the increase in fertilizer application. However, the amount of ap-
the presentation of Prof. K. Kyuma (Japan). His keynote lecture plication for rice exceeds the requirement. As a result of this
was “Paddy soils around the world.” Based on his more than 40 imbalance, N pollution is becoming a serious environmental prob-
years’ experience in the field of paddy soil science, he empha- lem. J. Zhu (China) evaluated the environmental impact of N loss
sized the importance of diversity of paddy soil and its environ- under the wheat-rice system in the region around Taihu Lake. To
ment. solve this problem, it is necessary to increase the recovery of N
Diversity and sustainability of paddy soils were discussed from fertilizer urgently. M. Saigusa (Japan) reported that the ap-
by N. Ngoc Hung (Vietnam), K. Naklang (Thailand), and T. plication of polyolefin-coated urea, which was used as a con-
Wakatsuki (Japan). N. Ngoc Hung reported on the current status trolled-availability fertilizer (CAF), improved N recovery. The in-
of soil fertility in the Mekong Delta of Vietnam. He pointed out vention of CAF now enables a co-situs application (contact appli-
some problems in fertility such as nitrogen (N) loss through am- cation of fertilizer with plant roots) and that is improving the effi-
monium volatilization. In contrast to the relatively fertile Mekong ciency of fertilizer, environmental loading from fertilizer, and so
Delta, the paddy soils in northeastern Thailand are very infertile. on. W. Bowen (Bangladesh) also reported a good recovery of N
K. Naklang introduced her long-term field trials, and discussed from fertilizer with the deep placement of supergranule urea for
the effect of organic matter application on the improvement of lowland rice in Bangladesh.
soil fertility. T. Wakatsuki reported that the area with lowland rice D. Olk (USA) indicated that the yield decline caused by
farming has increased in West Africa. He predicted that a Green intensive rice cropping under continuously flooded conditions was
Session 12: Conservation of soil, water, and environment in rice cultures 391
SESSION 13
Farmers’ participatory approaches
to facilitate adoption of improved technology
CONVENER: T. Paris (IRRI)
CO-CONVENER: J.S. Caldwell (JIRCAS)
How does a farmer accept a new technology?
Fujihiko Tozawa
Ogata village, located in Akita Prefecture in northern Japan, sible for the water quality deterioration of Hachiroko Lake
was built on the reclamation of Hachirogata Lake, formerly because we use its water for agricultural irrigation and drain-
the second-largest lake in Japan, during 1957-77. This was the age. During the 40-year history of Ogata, we have been strug-
biggest agricultural land reclamation in Japanese history. The gling to improve the water quality of Hachiroko Lake by stop-
new reclaimed land had an area of about 15,600 ha, with re- ping the sale of synthetic detergents in a store of the agricul-
sidual canals and lakes of about 5,000 ha. When the project tural cooperative, banning pesticide spraying by airplanes, stop-
was completed in 1977, Ogata was expected to be a “model of ping the project of a golf resort planned within the village,
modern agriculture,” based on large-scale and productive farm- developing various kinds of sustainable agricultural technolo-
ing systems. Some 589 farmers were settled in Ogata from all gies from organic to no-till, etc. These efforts were partial and
over the country. Each of them had a farm of 15 ha, regarded fragmentary at first, but they were expanded to the whole com-
as large-scale in Japan, where the average farm size is about 1 munity and became systematic in the 1990s, when researchers
ha. of Akita Prefectural University and the Akita Agricultural Ex-
Like other farmers of Ogata village, I am a full-time and perimental Station started joining farmers to help them develop
large-scale rice farmer. In 1998, I established a company with new technologies and social systems for sustainable farming
some friends to sell our rice. As full-time farmers, we are al- on the reclaimed land. A series of collaborating programs be-
ways interested in technological innovations in farming and tween farmers and researchers began at that time, including
farm management because we always want to reduce costs and Ogata Low-Input Sustainable Agriculture (O-LISA 1991), the
sell our rice under better conditions. Unlike other Japanese Research Project for Sustainable Agriculture in Ogata (RPSAO
farmers, the majority of farmers of Ogata do not sell their rice 1998), and the Declaration of Environmentally Creative Agri-
to agricultural cooperatives, but we have multiple marketing culture in Ogata (ECA 2001). In 2001, we established a volun-
channels. For instance, we can sell rice by individual direct tary organization to promote ECA, and I have been president
marketing, or group direct marketing, or to rice wholesalers. It of this organization.
is common here that a single farmer has different sales chan- The above research has revealed that Ogata is the larg-
nels and grows rice in different ways on his farm to respond to est single area of sustainable rice production in Japan. Table 1
his different consumers. Farmers of Ogata are all independent shows examples of sustainable farming practices in Ogata.
farm owners who must decide what kind of rice they will grow, From the top, we see that the area of “Less chemical farming”
to whom they will sell, and how much. Therefore, it takes us a is 5,110 ha and accounts for 74.1% of the total farmland of the
long time to consider whether we should accept a new tech- village. “No chemical farming” (not the same as “organic”) is
nology because we should examine whether the new technol- 520.8 ha and the percentage of the total farmland is 7.5%.
ogy is compatible with other factors of our farm management. These lower input practices already cover 80% of the total
As for environmental problems, we have been suffering farmland of the village, whereas conventional farming covers
from a specific one: worsening of the water quality of the re- 20%. As you can imagine, these figures are extraordinarily
sidual lake of the commonly called Hachiroko Lake, whose high in Japan in terms of total area, area per capita, and per-
water quality (measured by COD) was listed as one of the worst centage of total farmland. We are also active in accepting “more
five lakes in the country in 2001. Ogata farmers are respon- friendly practices with Hachiroko,” namely, the “use of con-
West Bengal, the largest rice-growing state in India, has made Since the introduction of high-yielding varieties, farm-
respectable progress in rice production over the last two de- ers have been applying more fertilizer than the recommended
cades. Rice production increased from 11.2 million tons in dose. This reduces the agronomic and economic efficiency of
1980-81 to 20.9 million t in 2001-02, mostly because of tech- N fertilizer and has negative consequences for the environ-
nological progress that raised yield from 2.16 to 3.39 t ha–1. ment. The International Rice Research Institute (IRRI), through
But the price of rice has increased at a much slower rate than experiments on soil-plant development analysis, developed a
the unit cost of cultivation, showing the strains of diminishing simple decision-making tool called the leaf color chart (LCC)
returns to farmers. One way to increase the profitability of rice for the timely application of N fertilizer to the rice crop. The
cultivation is to reduce the unit cost by adopting efficient crop color shades of the LCC card are designed to match leaf col-
management technologies that can save modern inputs. ors of the rice plant from signs of nitrogen “hunger” to over-
feeding of the plant. By matching the color on the LCC with
Session 13: Farmers’ participatory approaches to facilitate adoption of improved technology 395
that of the rice plants, farmers can decide on the proper time the 14th day after transplanting (DAT) at 7-day intervals until
and amount of N application. LCC validation experiments in flowering. For the pre-wet-season and wet-season crop, the
Vietnam and other countries have shown that farmers could critical value was 3.5, and for the dry-season crop it was 4.
save a substantial amount of nitrogen without any reduction in Yield was estimated through the crop-cutting method.
yield (Balasubramanian et al 2000, Bijay Singh et al 2002). Field days were organized before harvesting, enabling farm-
To validate the technology with farmers, social scien- ers to evaluate the technology. During field days, the farmers
tists emphasized conducting farmer participatory research that discussed among themselves and evaluated the performance
would help farmers assess not only the suitability of the tech- by ranking the efficacy of the LCC using their own criteria.
nology to the farmers’ biophysical and agronomic environment,
but also its acceptability to farmers given their socioeconomic
environment and cultural background. Once the farming com-
Results and discussion
munity sees the benefit of the technology through its own ex- Farmers’ general perceptions were that the greener were the
perimentation, demand is expected to grow and the process of leaves, the higher yield would be. However, after testing, the
adoption would become faster. Community participatory, participating farmers were of the opinion that, by using the
farmer-managed trials were therefore conducted by farmers in LCC, the amount of nitrogenous fertilizer applied was reduced
West Bengal, India, during 2002-04 to validate the efficiency without affecting production. This saved farmers money and
of the LCC in managing nitrogen in rice cultivation. helped increase net returns. Nitrogen use decreased by 20.2,
The overall objectives were (1) to enable farmers to 30.2, and 36.0 kg ha–1 over the farmers’ traditional practice in
evaluate the effectiveness of the LCC in N management and the pre-wet-, wet-, and dry-season rice crops, respectively.
(2) to assess the economic and other benefits associated with Participating farmers thus saved a substantial amount as fertil-
the use of the LCC. Specific objectives were to observe the izer cost is an important part of direct costs.
effectiveness of a community participatory approach in tech- Farmers found out from their observations and from tri-
nology assessment and its adoption. This paper reports the re- als designed and conducted by some of them that using the
sults of experiments in the dry season of 2002-03 and trans- LCC could significantly reduce the use of pesticides, too. They
planted rice premonsoon and monsoon seasons in 2003 in West were of the opinion that higher doses of N fertilizer were re-
Bengal, India, by Bidhan Chandra Krishi Viswavidyalaya, sponsible for increased pest and disease infestation.
IRRI, and a local NGO. The real-time use of nitrogen affected the cost of culti-
vation and net returns of rice in different seasons. Farmers
earned more profit per hectare in the LCC plots than in the
Materials and methods
farmers’ own plots. These amounts were more in LCC plots
Out of six agroclimatic zones of West Bengal, two zones, the than in farmers’ plots by US$13.10 in pre-wet rice, $20.40 in
new alluvial and old alluvial, were chosen for the dry season. wet rice, and $19.40 in dry rice (Table 1).
Initially, six villages from six districts of the two zones were During the field days, participating and nonparticipat-
randomly selected for the study. As it was thought to be diffi- ing farmers evaluated the performance of LCC technology.
cult to monitor intensively and interact closely with the farm- They found out that, in the LCC plots, the number of unfilled
ers, a cluster of four neighboring villages in the premonsoon grains was lower than in farmer-managed plots and plants of
season and five villages in the monsoon season was identified LCC plots did not lodge despite strong winds, whereas plants
purposively, during the subsequent seasons, based on repre- in the farmers’ plots lodged. Some of the varieties also ma-
sentativeness of the flood-prone rice environment and the tar- tured earlier by about 5 to 7 days when N was managed by the
get group, and ease of monitoring and supervision. To do the LCC than in farmers’ plots.
work smoothly, partnership with a local NGO, the Nadia Zilla These positive effects have significantly changed the per-
Farmers’ Development Organization, was established. Partici- ception of farmers that very dark green leaves are associated
pating farmers were selected using a community participatory with higher grain yield. They now understand that N should be
approach. Meetings were organized in each village to discuss applied when the plant starts feeling “hungry” for nitrogen.
the LCC and its role in the judicious use of nitrogen in rice The farmers’ rankings showed that the reduction in N
cultivation. Farmers from each village selected 10 representa- use was most important and reduction in lodging was the least.
tives from the willing farmers in the dry season. Farmers were The rankings were more or less uniform across communities,
informed of technical details for using the LCC. Since valida- suggesting that farmers were equally concerned about the over-
tion should be under farmers’ management conditions, they use of N. Influenced by the success of the LCC experiments,
were informed that no material and financial inputs would be the number of participating farmers tripled by the third sea-
given to the participants so that their decision-making would son. The rate of adoption has increased each season and there
not be affected or distorted. To establish an ongoing monitor- have been no reports of discontinuation of LCC use.
ing system at the community level, small groups of farmers The LCC technology was found to be eco-friendly. The
were formed to share each LCC. Measurements were taken reduction in N use has potential to reduce NO3 pollution and
using the LCC matching the color of 10 uppermost fully ex- minimize the use of pesticides, which would have a beneficial
panded leaves from randomly selected plants per plot from effect on human health, biodiversity, and soil and water qual-
Early wet season 254.90 263.80 8.90 332.60 328.40 77.70 64.60 13.10
Wet season 328.20 343.10 14.90 413.40 407.80 85.20 64.80 20.40
Dry season 421.50 445.30 23.80 591.40 595.80 169.90 150.50 19.40
ity. The participating farmers believed that this simple tool Community action in technology testing facilitated bet-
empowers them to make decisions on the time and amount of ter and faster decision-making for N management. Small groups
N to be applied. established to share each LCC facilitated such evaluation. Fre-
It was thought that the community participatory approach quent interaction with the scientists and within the farming
used strengthened the partnership among key stakeholders, community improved the knowledge base of the farmers. This
which facilitated community action and technology adoption. helped to strengthen the social capital. However, the trial con-
The farmers believed that more and more frequent interaction firmed that a varietal interaction does exist and the communi-
with the scientists and stakeholders would help enhance their ties should be encouraged to evaluate this in partnership with
knowledge of improved farming practices. It appeared that an researchers.
action group has been formed in this locality. The farmers of
this group were very close to each other. Evidence from the
References
participatory approach used showed that farmer-to-farmer
transfer of technology was taking place, and the technology is Balasubramaniam V, Morales AC, Cruz RT, Thingarajan TM,
likely to be adopted by the large group of farmers. This ex- Nagarajan R, Babu M, Abulrachman S, Hai LH. 2000. Adop-
periment also had an impact on the scientific community. The tion of the chlorophyll meter (SPAD) technology for real-time
scientists who visited the experimental plots appreciated the nitrogen management in rice: a review. Int. Rice Res. Notes
25(1):4-8.
capabilities of farmers in testing new technologies under their
Bijay Singh et al. 2002. Chlorophyll meter- and leaf color chart-
own management.
based nitrogen management for rice and wheat in northwest-
ern India. Agron. J. 94.
Conclusions
Real-time management of nitrogen with the LCC helped farm- Notes
ers minimize N use, reduce the cost of cultivation, decrease Authors’ addresses: B. Bagchi and S.K.T. Nasar, Directorate of Re-
pest and disease incidence, minimize chaffy grains, and pro- search, Bidhan Chandra Krishi Viswavidyalaya, Nadia, West
tect the crop from lodging. Farmers’ perceptions that the use Bengal, India; M.Z. Abedin, International Rice Research In-
of more nitrogenous fertilizer brings more yield changed. stitute, Los Baños, Philippines, e-mail: drbbagchi@vsnl.net.
Water plays a major role in rice cultivation. On average, some The water supply is increasingly becoming scarce, how-
4,000 L of water are required to produce 1 kg of rice (Tuong ever, because of the world’s increasing population that brings
and Bouman 2003). Thus, an inadequate supply of water from along multiple competing demands, such as for agriculture,
crop establishment to the reproductive stage of the crop gen- industry, and domestic use, and climatic changes such as El
erally leads to a significant yield reduction (Wopereis et al Niño phenomenon (IRRI 1995, Bouman and Tuong, 2001).
1996, Bouman and Tuong 2001). The increasing water scarcity threatens not only general food
security in Asia but also the livelihood of most Asian farmers
Session 13: Farmers’ participatory approaches to facilitate adoption of improved technology 397
because of the increasing irrigation costs borne by them. To Table 1. Farmers’ perceptions about controlled irrigation (CI) as a
meet the food demand of Asian consumers and producers, ef- water-saving technology.
ficient water management practices are needed that maintain Perceptions about CI (n = 15) Frequency Percentage
rice yields and production at a high level.
To solve the problem of water scarcity, researchers have 1. Less time, less expensive, and saves labor 15 100
been developing water-saving irrigation technologies. In the 2. It saves a lot of water 15 100
Philippines and in 2001, IRRI began the project “Technology 3. No yield difference vis-à-vis farmers’
practice (FP) 14 93
Transfer for Water Savings” (TTWS) in collaboration with the 4. I learned a lot of new techniques in modern
National Irrigation Administration (NIA) and Philippine Rice rice production such as fertilizer application,
Research Institute (PhilRice) to facilitate farmers’ adaptation selection of pure seeds, when to spray, how
and adoption of controlled irrigation (CI) (Bouman et al 2002, to save water during irrigation, etc. 11 73
Lampayan et al 2003). CI entails an irrigation schedule in 5. If CI would be adopted by all members in
P-38 ISC, the water rotation interval would
which, contrary to the normal practice of continuous flooding, be improved in timeliness, reliability, and
water is applied to the field a number of days after the disap- equity of water distribution from one field
pearance of ponded water and yet high yields are maintained to another. 11 73
(Tabbal et al 2002, Belder et al 2004). The CI technology was 6. Grains in CI plots are rounded and heavier 5 33
piloted among farmers using irrigation water drawn from deep- 7. FP plots could not produce more tillers
because of continuous flooding 3 20
well systems in Tarlac Province, where these deep wells are 8. I observed that, in modern rice
maintained and operated by farmers’ irrigation service coop- production techniques, straight planting
eratives (ISC). requires more people 2 13
The successful adoption of a new technology such as CI 9. Others 3 20
depends on not only its suitability and profitability but also on
Others: Crop stand in both FP and CI are the same, CI plots have no diseases.
social and cultural factors such as farmers’ perceptions. Fur-
thermore, where irrigation water use is organized in a commu-
nal way, the organizational structure and functioning play a
key role as well. The objectives of this paper are (1) to deter- Farmers’ perceptions of CI
mine farmers’ perceptions of CI as an effective water-saving The majority of the farmer-cooperators gave positive feedback
technology in rice production and (2) to explore success fac- about the effectiveness of CI as a water-saving technology, as
tors for collective action in an organized irrigation manage- shown below and in Table 1.
ment system that would facilitate and ensure eventual farmers’ No yield difference compared with farmers’ practice
adoption. (FP). Most farmers perceived that yield in the FP plots was
similar to that in the CI plots, regardless of less water being
used in the CI plots. This perception agrees with the yields
Materials and methods obtained from crop-cut samples of 2 × 2.5-m2 areas as indi-
We focus our study on the case of the deep well P-38 of the cated in Table 2. In both 2002 and 2003, there was no signifi-
TTWS project, a reactivated system in Canarem, Victoria, cant yield difference under FP and CI.
Tarlac, about 70 km north of Manila. Eleven and 12 farmer Saves water. Farmers recognized that CI saves water.
volunteers in the 2002 and 2003 dry season (DS), respectively, Aside from alternate wetting and drying of the paddy field, CI
participated in the CI experiment. Each farmer participant con- maintains only a 2-cm depth of standing water in contrast to
tributed two (neighboring) plots with a size of 500–1,000 m2 the usual farmers’ practice of maintaining a 5–8-cm water depth.
each, one representing the current farmers’ practice (FP) and On average, the amount of water saved in CI versus FP was
one the CI technology. The water management of the CI plots 16% in 2002 and 24% in 2003 (Fig. 1). The largest water sav-
was a joint affair among all project members and aimed at ings were 24% in 2002 and 33% in 2003. The higher savings
site-specific optimization of the technology (in terms of irri- in 2003 reflect the effects of the learning process: at that time,
gation rotation, depth of application, number of days without farmers were already confident about the performance of CI
ponded water, etc.). All activities and farm inputs and output and willing to take more risk in saving more water.
were closely monitored in both seasons. An input-output sur- Saves time, labor, and expenses. All farmer-cooperators
vey of rice production for all 63 farmers was done in the 2003 acknowledged that CI saves time, labor, and expenses. It re-
DS using a semistructured questionnaire. Focus-group discus- duced costs by using 20–25% less fuel and oil (Table 2). It
sions (FGDs) and key informant interviews were conducted reduced labor as farmers spent fewer hours in irrigation, de-
among different stakeholders such as ISC officers and mem- pending on the distance. Moreover, the ISC could abolish water
bers, NIA staff, and IRRI water scientists. Field observations delivery during nighttime.
were likewise made and records of ISC meetings were docu- Better grain quality. Farmers observed that rice grains
mented throughout the dry-season rice cropping. from CI plots were heavier, bigger, and in better shape be-
cause of soil aeration, whereas, in FP plots, rice grains were
lighter and more slender, and sometimes unfilled, because of
the absence of soil aeration. In 2002, members of the TTWS vidual supply. All farmers need to bring their own fuel and oil
project took grain samples from CI and FP plots for some qual- to the pump for irrigation of their fields, and this resulted in a
ity analysis. The results showed no significant difference be- remarkable 40% decrease in the total amount of fuel and oil
tween CI and FP grains in the milling indicators, that is, head used. During the time when the cooperative was using fuel
and broken rice, brewers’ rice, bran, chalky and immature provided by NIA, the farmers were quite extravagant in their
grains, discoloration, and damaged grains. use, but they were very prudent in their use when fuel was
individually provided.
ISC functioning and impact of CI After two years, the experiences with CI also helped ease
At the start of the TTWS project, tensions about water use tensions in water use. Farmers were no longer apprehensive of
were rising because of the increase in membership of the ISC not having ponded water on their fields for some days as they
(from 45 in 1998 to 63 at present), and because of subsequent had experienced that this did not reduce yields. Cohesive in-
increases in service area (from 50 to 70 ha) and the lengthen- teraction among farmers also increased during the experimen-
ing of the interval of water delivery (from 7 d to 12–14 d). The tation because everyone was curious about the performance of
introduction of CI added to the tension because farmers were CI. And, when farmers saw the effectiveness and viability of
apprehensive of looking at their rice fields with dry land, es- CI, many were convinced that CI could reduce their water con-
pecially when they saw the cracking of the soil. Many farmers sumption. Since CI reduced water consumption, CI became a
doubted that CI would give them the same yields as or higher significant factor in solving operation problems and induced
yields than their usual practice and, thus, some farmers started farmers to cooperate, that is, with the fuel scheme of bringing
stealing water by placing holes underneath their paddy dikes, one’s own fuel and oil, and to be a continuing member of the
hidden from other farmers’ eyes. Conflicts were resolved deep-well system.
through the intercession of village officials, particularly the
village security officers. Factors in success of collective action
At the same time, many farmers were delinquent in pay- and adoption of CI
ing their dues, resulting in a temporary stoppage of irrigation Collective action is the basic foundation for facilitating adop-
operations. Within 2 days, most farmers settled their bills, per- tion of CI in a deep-well irrigation system. Below are the fac-
haps because water deliveries ceased when the crop was close tors identified as affecting the success of collective action.
to the flowering stage, when water is really much needed. Con- Group size. Groups should be small to minimize trans-
sequently, the ruling for having access to water was likewise action costs, but cannot be too small or else the ISC won’t be
changed from using NIA’s fuel and oil supply to using an indi- able to cover the costs of operating and maintaining the sys-
Session 13: Farmers’ participatory approaches to facilitate adoption of improved technology 399
% savings P 38, Canarem
35
30
25
20
15
10
0
Average
Average
Middle
Middle
Low
Low
High
High
2002 dry season 2003 dry season
Fig. 1. Average water savings (%) in controlled irrigation over farmers’ practice in Canarem,
Victoria, Tarlac, for 2002 and 2003 dry seasons, based on data from 11–12 farmers.
tem. Relatively small group sizes are generally associated with policies and the use of CI reduced total fuel consumption by
homogeneity among members; in addition, communication 40%.
among members takes less time and effort. In the experience Excludability. Excludability (i.e., the degree to which
of P-38, communication and management (particularly in terms entitled users can keep out free riders) is an important factor
of payment) were easier with less than 50 members. in the effective operation of an ISC. Excludability is high in P-
Service area size. The size of the service area should be 38, as witnessed by the temporary suspension of operations in
small, just large enough to maintain a water delivery schedule 2002 and by the rule of providing one’s own fuel and oil, with-
that would not create water stress to rice plants. At P-38, when out which one does not have access to water.
the service area increased from 45 to 78 ha, the interval be- Institutional linkages. Partnership with local government
tween water deliveries increased to accommodate all the fields. is necessary. It reinforces cooperation and discipline among
A long time interval is not acceptable to most members, and members where the village security officers intervened in re-
they prefer a 7-day interval rather than 12- to 14-day intervals. solving conflicts related to the stealing of water. Moreover,
Profitability. Relative profitability of adopting CI should through close monitoring and mentoring by NIA of the ISC’s
be high enough to make collective action profitable. Based on operation and maintenance of the system, farmers were able to
the experiments in 2002 and 2003, the amount of water used receive guidance and technical support, which helped the P-
in CI was lower than in the usual farmers’ practice (Fig. 1), 38 ISC to be a stronger cooperative.
and hence irrigation costs were lower as well. Water savings
averaged 16–24% and costs savings 20–25% (Table 2).
Conclusions
Strong leadership—strict enforcement of rules to solve
free-rider problem. In the first four years of operation, fuel Controlled irrigation was tested through farmer participatory
and oil were supplied by NIA as part of a loan. Records show research and development, and found to be a viable technol-
that ISC members used fuel excessively, probably because the ogy for farmers’ use, as demonstrated by P-38 ISC members.
payment was the same for each member and because payments Farmers perceived that CI saves significant amounts of water,
were not immediate but made only after harvest. The chair- time, labor, and cash, and reduces the costs of rice production.
man, imbibing the Filipino values of smooth interpersonal re- Furthermore, CI is perceived to give similar yield and produce
lationships and a sense of shame, was not so strict in enforcing more tillers, and bigger and heavier grains with good shape.
the policies and regulations, particularly those related to pay- Scientific evidence has corroborated the farmers’ perceptions
ing dues to the ISC. But, with the strict enforcement of rulings on water savings and yield performance. Though CI is suit-
on payment (that had resulted in halting the operation of P-38 able for adoption by farmers based on their positive percep-
for two days) and access to water (bringing one’s own fuel), tions, good functioning and cooperation among members of
everyone was frugal in the amount of fuel used. These new an ISC seem to be another requirement for the adoption of CI
It is widely admitted that poor coordination among stakehold- number of stakeholders using common resources at the rice
ers leads to inefficient resource use, economic and environ- agroecosystem level. Our hypothesis is that an understanding
mental damage, negative externalities, and social conflict. Di- and modeling of the diversity of stakeholders’ perceptions,
verse stakeholders use resources for different purposes, with associated with participatory simulation sessions, can be used
differing perceptions of their dynamics, and adopt various strat- to improve resource management through better coordination
egies to cope with problems. Consequently, the number of so- of people’s actions in any type of rice ecosystem.
cial conflicts is increasing and they are frequently reported in
the national and international media, for example, the cases of
The proposed companion modeling approach
water sharing at rice transplanting among farmers and villages
in Bhutan, and conflicts over land use between highlanders Usually, models are used to assemble scientific knowledge and
and lowlanders in northern Thailand uplands. To manage these to propose recommendations to a given decision-maker. In our
problems, new legislative frameworks decentralizing the man- case, we propose to use models to represent the different per-
agement of renewable resources are being introduced in many ceptions of various stakeholders in order to facilitate the coor-
countries. Their success depends on the quality of the local dination of their actions on a common resource. Some experi-
coordination among stakeholders, who often lack tools, meth- ments were conducted by using geographic information sys-
ods, and trained managers to achieve success. tems (GIS) (Abbot et al 1998, Gonzalez 2000), but very few
Thus, there is a demand for innovative approaches and of them dealt with the use of simulations (Costanza and Ruth
tools to improve coordination processes among an increasing 1998). The main lesson learned is that these tools, like many
Session 13: Farmers’ participatory approaches to facilitate adoption of improved technology 401
technological innovations, can both marginalize and empower Case studies
people and communities.
Companion modeling (ComMod) is a methodology for For the past three years, the ComMod approach has been used
the collective implementation and use of simulation models, to examine various resource management problems in differ-
and, more precisely, multi-agent simulation systems (MAS) ent Asian rice ecosystems (Bousquet et al, in press).
(Barreteau 2003). The ComMod approach proposes an itera-
tive and evolving modeling process: participatory modeling is Water sharing at rice transplanting
used to facilitate stakeholders’ interactions and to identify and in Lingmuteychu, Bhutan
scope resource management problems to be discussed and Lingmuteychu is a catchment covering 34 km2 in west-central
negotiated. Then, simulation models are used to collectively Bhutan, which is drained by a totally rainfed stream originat-
assess scenarios selected by stakeholders. This may lead to ing from a rock face at an altitude of 2,400 m. Five small irri-
new questions, new discussions, changes in the model, and so gation systems formed of 12 canals irrigate about 200 ha of
on. terraced paddies belonging to 121 households of six villages.
ComMod combines the use of different tools such as These villages share irrigation water within a broadly respected
agent-based modeling (ABM), GIS, participatory mapping, and customary regime evolved during a time when demands were
role-playing games (RPG). From a methodology development lower. Under the current processes of market integration, de-
point of view, the ComMod approach has been tested and used centralization, and resource conservation policies, and changes
in several places, leading to concrete policy recommendations in villagers’ social needs, this customary water-sharing set of
or collective actions by communities (Aquino (d’) et al 2002, rules is not adapted to current farming conditions anymore and
2003, Etienne et al 2003). The use of ComMod implies the causes repetitive social conflicts, particularly at rice transplant-
execution of the following set of activities in an iterative way: ing, as these conflicts remain unresolved. There is a definite
O Framing. Preliminary diagnostic analysis at the sys- contrast in perceptions of users on the water resource: some
tem level. Identification and understanding of the consider it an infinite resource, many consider this resource as
system’s ecological and social dynamics and of key an exclusively common pool with a free-access regime, whereas
issues, concerns, and intervention points with stake- the state considers it as its property. The Draft National Water
holders. Collection of relevant existing data. Identi- Policy pronounces water as a state property while it empha-
fication of knowledge gaps to be filled through spe- sizes integrated management. Under the new national commu-
cific surveys. nity-based NRM policy (CB-NRM), this watershed has been
O Prioritization and visioning. Participatory selection selected as a pilot site to improve coordination among water
of a key concrete problem to be examined. Establish- users. It is located near the Bajo Renewable Natural Resources
ment of a common vision shared by all key stake- Research Center (RNR-RC), which is leading the national ef-
holders through RPG and participatory simulation fort in the field of CB-NRM, and the Natural Resource Train-
workshops. Delimitation of initial areas of agreement, ing Institute (NRTI) at Lobeysa, the principal institution for
disagreement, uncertainty, and room for coordination higher education in agriculture and resource management in
and negotiation. Production of qualitative guidelines the kingdom. Participatory land-use planning and rural ap-
for monitoring and evaluation of the subsequent praisal activities were carried out by the RNR-RC Bajo team
ComMod activities. at this site. More recently, an analysis of existing water dy-
O Participatory field work and modeling. Implementa- namics, water-sharing arrangements, and farming practices has
tion of an iterative, integrated, flexible, and user- been implemented and a first participatory workshop was held
friendly modeling approach combining participatory in the two upper villages (where the conflict over water use is
workshops and laboratory work. Joint validation of acute) in May 2003 to test the proposed ComMod methodol-
the simulation tool with all concerned stakeholders, ogy for supporting farmers from the two villages to examine
followed by participatory identification of relevant collectively the problem of water exchange at rice transplant-
resource management scenarios to be simulated and ing (Gurung, in press). Based on the successful outputs, an-
assessed collectively, taking into account the com- other participatory workshop was held in December 2003 to
peting uses of resources by multiple users. modify water-sharing rules. Figure 1 shows the assessment of
O Collective exploration and discussion of trade-offs the ComMod process on the perceptions of the water-sharing
displayed by the simulated scenarios. Choice and de- issue by the local stakeholders who took part in both work-
sign of an action plan to be implemented to tackle shops.
the resource management problem under consider- By using a MAS model reproducing the RPG played
ation. and validated by the farmers, simulations were carried out later
O Assessment of local impact. Assessment of the local to examine the relative effects of social networks, communi-
impact of the approach for participatory and integrated cation protocols, and climate conditions on water-use effi-
renewable resource management (IRRM). ciency. Results indicate that communication protocols consti-
tute the most sensitive factor.
50
40
30
20
10
0
Farm Valuation Water Share Income Canal Control
activities of water shortage water management water
use
Lessons learned
Fig. 1. Lessons learned by the stakeholders after two participatory workshops.
Soil and water conservation in diversifying highlands spective influence of formal and informal credit on the wel-
of upper northern Thailand fare of different kinds of stakeholders (Barnaud 2004).
Mae Salaep is a highland village in Mae Fah Luang District of
Chiang Rai Province, where small-scale poor Akha farmers
Conclusions
are being rapidly integrated into the market economy. While
their former agrarian system based on swiddening is being re- The ComMod approach seems to be well received by both
placed by semipermanent agriculture on steep slopes, the risk scientists working in the field of IRRM and local stakehold-
of increased land degradation through soil erosion by concen- ers. It is important to distinguish between the use of this ap-
trated runoff is a major problem. The diversity of farmers (eco- proach in two specific contexts: (1) to produce knowledge on
nomic status, agricultural practices, etc.) is already extensive, complex rice systems and (2) to support evolving, iterative,
and their economic and institutional environment is becoming and continuous collective decision-making processes for
more complex (Trébuil et al 2002). An increasing number of IRRM. Where the land policy favors a decentralized manage-
individual or collective stakeholders with differing land- and ment of resources, such a MAS-based companion modeling
water-use strategies interact in the dynamics of diversifying approach has great potential to improve the collective man-
sloping-land agriculture. In collaboration with Chiang Mai agement of rice lands. It can be used to facilitate dialogue, to
University, the Department of Public Welfare (a government mitigate conflicts, and to establish coordination mechanisms
development agency looking after highland ethnic minorities), regarding multiple uses of the land by multiple stakeholders.
and a local NGO, a spatially explicit multi-agent model linked It is also a powerful tool to integrate knowledge from different
to a small GIS was built to represent the key interacting eco- disciplines, sources, and levels of organization. ComMod fa-
logical and agronomic dynamics (slope characteristics, rain- cilitates the collective assessment of desirable scenarios and
fall, main cropping systems and succession of practices, farmer identification of suitable innovations to overcome current
differentiation, etc.) to assess the risk of soil degradation. This IRRM problems in rice ecosystems.
MAS model was later simplified and translated into an RPG
to be used with farmers to validate the understanding of the
References
agricultural dynamics by the research team. In December 2002
and May 2004, two participatory modeling workshops, focus- Abbot J, Chambers R, et al. 1998. Participatory GIS: opportunity or
ing on land-use changes (particularly the transition from an- oxymoron? PLA Notes 33:27-34.
nual crops to perennial plantations) and the organization of Aquino (d’) P, Le Page C, et al. 2002. A novel mediating participa-
the credit system at the village level, respectively, were held tory modelling: the ‘self-design’ process to accompany col-
lective decision making. Int. J. Agric. Resources Governance
with representatives from all categories of farmers and devel-
Ecol. 2(1):59-74.
opment agencies. From the first to the second workshop, at
Aquino (d’) P, Le Page C, et al. 2003. Using self-designed role-
the request of the participants, the emphasis shifted from playing games and a multi-agent system to empower a local
achieving a shared representation of the IRRM problem and decision-making process for land use management: the
its causes to the use of ComMod for improving the allocation SelfCormas experiment in Senegal. J. Artif. Societ. Soc. Simul.
of new village funds to improve the current situation and espe- <http://jasss.soc.surrey.ac.uk/6/3/5.html> 6(3).
cially the access of small farmers to plantation crops to limit
erosion risk. MAS simulations were used to examine the re-
Session 13: Farmers’ participatory approaches to facilitate adoption of improved technology 403
Barnaud C. 2004. Erosion des sols et systèmes agraires dens les hautes Gurung T. In press. Companion modeling to examine water-sharing
terres de la Thailandes: une approche de la complexité par arrangements among rice-growing villages in west-central
une modélisation d’accompagnement. Master Thesis, Mémoire Bhutan: preliminary results. In: Bousquet F, Trébuil G, Hardy
de DEA, Géographie et pratique du développement. Paris, B, editors. Companion modeling and multi-agent systems for
Université de Paris X-Nanterrep. integrated natural resource management in Asia. Los Baños
Barreteau O. 2003. Our companion modelling. J. Artif. Societ. Soc. (Philippines): International Rice Research Institute.
Simul. 6(1). Trébuil G, Shinawatra-Ekasingh B, Bousquet F, Thong-Ngam C.
Bousquet F, Trébuil G, Hardy B, editors. In press. Companion mod- 2002. Multi-agent systems companion modeling for integrated
eling and multi-agent systems for integrated natural resource watershed management: a northern Thailand experience. In:
management in Asia. Los Baños (Philippines): International Jianchu X, Mikesell S, editors. Landscapes of diversity.
Rice Research Institute. Yunnan Science and Technology Press, China. Proceedings
Costanza R, Ruth M. 1998. Using dynamic modeling to scope envi- of the 3rd International Conference on Montane Mainland
ronmental problems and build consensus. Environ. Manage. Southeast Asia (MMSEA 3), Lijiang, Yunnan, China, 25-28
22(2):183-195. August 2002. p 349-358. At www.cbik.ac.cn/cbik/resource/
Etienne M, Le Page C, et al. 2003. A step-by-step approach to build- MMSEA_Index.asp, accessed 8 September 2003.
ing land management scenarios based on multiple viewpoints
on multi-agent system simulations. J. Artif. Societ. Soc. Simul.
6(2). Notes
Gonzalez R. 2000. Platforms and terraces: bridging participation and Authors’ addresses: F. Bousquet, IRRI-Cirad-DOA, Bangkok, Thai-
GIS in joint learning for watershed management with the land, e-mail: f.bousquet@cgiar.org; G. Trébuil, IRRI-Cirad-
Ifugaos of the Philippines. PhD thesis. ITC-Wageningen Uni- DOA Project, Bangkok, Thailand, e-mail: guy.trebuil@cirad.fr.
versity. 186 p.
Most conventional breeding programs have been set up and Methods used in participatory crop improvement
designed by breeders, neglecting the role of users: farmers and
farming communities. Therefore, the dissemination process of Witcombe and Joshi (1996) defined PPB as involving farmers
so-called “technology transfer” was very slow and costly for in selecting genotypes from genetically variable segregating
both breeders and farmers. materials and PVS as involving the selection by farmers of
The use of participatory approaches assures the involve- nonsegregating characterized products from plant breeding
ment of farmers at different levels in the whole process of crop programs. In practice, the use of PVS and/or PPB depends on
improvement to overcome the shortcomings of conventional farmers’ varietal needs and farmers’ breeding knowledge and
breeding approaches. technical skills. The PVS approach has been used to improve
Can Tho University, as the leading research institution local landraces (crop cultivars/varieties that adapted and have
for adapting the participatory approach in rice improvement, been kept for a long time in local areas) and to evaluate the
started on-farm breeding programs as early as 1975 by send- finished breeding materials from research institutions. When
ing out its staff and students to work closely with farmers for varietal options available to farmers through PVS are limited
local crop improvement (Xuan et al 1993). Later, with the in- or exhausted, PPB begins.
ception of the Community Biodiversity Development and Con- A successful PPB program should involve farmers as
servation (CBDC) project in 1994, and the global in situ con- much as possible in the whole process of plant breeding (De
servation of biodiversity project implemented by the Interna- 2000).
tional Plant Genetic Resources Institute (IPGRI) in 1998, par- Participatory plant breeding involves the following steps
ticipatory plant breeding (PPB) and participatory variety se- and activities.
lection (PVS) approaches have been introduced to develop crop
varieties specific to niche environments and farmers’ prefer- Needs assessment and community organization
ences. Community meetings are organized to identify farmers’ prob-
This paper reviews the achievements, problems, and les- lems and needs. By discussing with farmers and local offi-
sons learned from these practices using new approaches in the cials, cooperating farmers are divided into three groups de-
Mekong Delta of Vietnam. pending on their technical knowledge and skill.
Genebank Collaboration
* Characterization
* Evaluation
Conservation
**
Use
Coordinating Receiving
Selection Selection
coding as coding as
* Replicated observation
Fig. 1. The participatory rice varietal improvement program in the Farming Systems Re-
search and Development Institute, Can Tho University, Vietnam. MTL (Mien Tay Lua) =
crossed, selected, and released by Farming Systems Institute; CBDC = community
biodiversity development and conservation; IPGRI = International Plant Genetic Resources
Institute with on-farm conservation project; farmer participation level = from little (*) to full
(****) participation.
O Group 1 (2–3 farmers) will take responsibility for are high yield, short duration, resistance to major pests and
making crosses and selecting segregating breeding diseases, good eating quality, and so on. Based on the breed-
materials (PPB activities). ing objectives, breeders then assist farmers in searching for
O Group 2 (5–10 farmers) will field-test stable lines and suitable donor parents. These donors could be found among
promising varieties (PVS activities). the available genetic materials at the local level or from re-
O Group 3 (30–40 farmers) could be involved in seed search institutions.
multiplication and distribution of selected varieties The level of farmers’ participation in the rice breeding
for local use. process is described in Figure 1.
Session 13: Farmers’ participatory approaches to facilitate adoption of improved technology 405
breeders in making desirable crosses to meet lo- varieties using the participatory approach. Of these, some fa-
cal breeding objectives. mous rice varieties are IR36 (named as NN3A), MTL2
O Selecting from segregating materials. Early gen- (NN6A), MTL30 (NN7A), MTL9 (NN2B), IR42 (NN4B),
erations will be evaluated and selected together MTL58 (IR13240-108-2-2-3), and MTL87 (IR50404-57-2-2-
with segregating lines provided by research in- 3). These varieties have made a great contribution to the im-
stitutions. This process is repeated until stable provement of rice production in the Mekong Delta. Many farm-
lines are obtained. Farmers with adequate train- ers, such as Hai Huu (Long An Province), Hai Chung, Tu Tai,
ing can practice mass selection and pure-line and Ba Chuong (Tien Giang Province), Ba Cung and Hai Triem
selection. (An Giang Province), Muoi Tuoc and Muoi Than Nong (Vinh
Long Province), and many others are known as “rice selection
2. Participatory varietal selection (PVS) kings.” Farmers are not only selecting suitable finished variet-
O Genetic improvement of local adapted variet- ies but also improving the varieties formally released for bet-
ies. Farmers can improve their local landraces ter grain quality and adaptation to specific conditions in their
or common varieties for better genetic purity, areas by pure-line selection. This process increases both crop
higher yield, better quality, and resistance to productivity and local crop diversity.
major pests by pureline selection or mass selec- Depending on local resource availability and manage-
tion methods. Sometimes, the local varieties col- ment capacity, each community has received 7–30 varieties
lected from the community or other communi- for testing and 3–4 varieties have been selected annually for
ties are reintroduced to the community after seed seed multiplication and use (Table 1).
loss caused by a disaster. In addition, new vari-
eties from research institutions are provided for Participatory plant breeding
testing together with finished products from PPB In the 1996-97 dry season, Can Tho University started provid-
activities. ing 63 segregating F2-F3 populations of 12 crosses for four
O Varietal testing for local adaptation. Stable lines pilot communities.
(20–40) selected from the segregating materials In 1998-99, L246-7-3-B and L247-1-5-B, the two prom-
or provided by research institutions are planted ising farmer selections noted as SiC-1 (Soc Trang Selection,
in an observation test with common local vari- no. 1) and SiC-2 (Soc Trang Selection, no. 2), respectively,
eties as local checks. The promising varieties (5– were purified by Ke Sach community (Soc Trang Province)
10) selected from the observation test are then using the bulk selection method. Mr. Canh (a farmer group
tested in larger plot sizes for a yield trial. Farmer leader) led the selection activities. Similarly, L246-10-1-B, a
field days (activities to assemble local farmers promising line selected by farmers in My Thanh community
at demonstration plots for direct observation, (Ba Tri District, Ben Tre Province), is also undergoing a yield
evaluation, and interaction) are organized just test and seed multiplication. In 2003, Mr. Thanh in Cho Moi
before harvesting for a joint evaluation. District (An Giang Province) selected two stable lines, TH1
O Seed multiplication. Desirable varieties (usually and TH2, from segregating breeding material of Can Tho Uni-
2–3) from yield trials are then selected for seed versity in a similar way.
multiplication by a larger group of farmers to
use in the community.
Discussion
3. Monitoring and evaluation With technical training and field assistance, farmers can man-
Farmers closely monitor and keep records on field con- age their own breeding programs. The total number of rice
ditions and crop performance for later analysis to deter- varieties in the communities is increasing through the intro-
mine suitable varieties. A major data set consists of duction of new varieties and locally improved cultivars. As a
growth duration, plant height, tillering capacity, grain result, land productivity and farm income have been improv-
yield, grain quality, and tolerance ability for major in- ing.
sects and diseases. Field visits and farmer field days in-
volving breeders, extension workers, and farmers are Problems
the most appropriate tools for monitoring and evaluat- Some problems have occurred.
ing PPB/PVS activities. O Farmers are more willing to multiply promising vari-
eties (PVS) than to select from segregating materials
or make crosses (PPB) because of the time and re-
Results sources needed and the requirement of skills and in-
Participatory varietal selection tellectual input.
From 1975 to 2003, hundreds of promising rice varieties were O The low education level of farmers limits the adop-
tested in farmers’ fields under farmers’ management, and many tion of PPB; therefore, more training and field coach-
varieties were identified and released as national recognized ing are required.
O Morris (2003) also pointed out three challenges of which are under review and reform in Vietnam. However, a
participatory breeding: successful participatory approach in crop improvement requires
— There is a need to develop varietal evaluation a willingness to contribute time, land, labor, and other inputs,
methods capable of generating credible data for and to incur possible risk. It requires technical knowledge and
widespread acceptability. skill from farmers, and a decentralized breeding strategy and
— A level of participation that will ensure equitable technical assistance from breeders. It also requires favorable
compensation for participating farmers is needed. policies and support in organization and management.
— National and international regulatory frameworks
that govern the evaluation, approval, and release
References
of new plant varieties from PPB are necessary.
De Nguyen Ngoc. 2000. Linking the national genebank of Vietnam
Lessons learned and farmers. In: Friis-Hansen E, Sthapit B, editors. Participa-
The following lessons can be drawn from these activities: tory approaches to the conservation and use of plant genetic
O Farmers conserve and improve biodiversity to meet
resources. Rome (Italy): International Plant Genetic Resources
Institute. p 62-68.
their needs for home consumption, the market
Morris ML, Bellon MR. 2003. Participatory plant breeding research:
economy, and adaptation to local environments and opportunities and challenges for the international crop im-
farm resources. Therefore, farmers’ needs and per- provement system. Euphytica (2003):1-15.
ceptions should be taken into consideration to bring Witcombe J, Joshi A. 1996. Farmer participatory approaches for
breeders’ objectives closer to farmers’ objectives. varietal breeding and selection and linkages to the formal seed
O Support from local authorities and organizations in sector. In: Participatory plant breeding. Proceedings of a work-
terms of organization, management, additional funds, shop on participatory plant breeding, 26-29 July 1995.
and facilitation is very important to ensure a success- Wageningen, Netherlands. p 57-65.
ful participatory approach. Xuan Vo-Tong et al. 1993. Present status of agricultural extension in
O Cooperation with a group/community on PPB/PVS Vietnam. Paper presented at the first Southeast Asia work-
turns out better than that with individual farmers be- shop on formulation of project proposals on technology trans-
fer for major food crop production, FAO and UAF, Ho Chi
cause group/community members can help each other.
Minh City, Vietnam, 6-9 Dec. 1993.
O Farmer field schools and farmer field days are good
ways to stimulate farmers’ participation and to train
farmers with a lower education. Notes
Authors’ addresses: Nguyen Ngoc De, Mekong Delta Farming Sys-
Conclusions tems Research and Development Institute, Can Tho Univer-
sity, Vietnam; Kotaro Ohara, Department of Sustainable Re-
Participatory approaches could improve breeding for local source Sciences, Faculty of Bio-Resources, MIE University,
adaptation, local genetic diversity, efficiency, and the empow- Japan, e-mail: nnde@ctu.edu.vn, ohara@bio.mie-u.ac.jp.
erment of rural communities. The achievements in these ac-
tivities have had a strong impact on national seed policies,
Session 13: Farmers’ participatory approaches to facilitate adoption of improved technology 407
IRRI’s approach to participatory research for development:
advances and limitations
Thelma R. Paris and M. Zainul Abedin
Why do participatory research? While the traditional linear Participatory research has diverse approaches, perspec-
and top-down research-extension-farmer transfer of technol- tives, practices, and methods. At IRRI, participatory research
ogy (TOT) model was successful in disseminating rice tech- differs according to various interpretations of the mode of re-
nologies for favorable rice environments, this approach no search partnership or “participation” in research, the goal or
longer holds for rice research in unfavorable environments. In rationale for encouraging participation, gender and stages in
the traditional top-down approach, technologies developed the project, the level of “control” or “ownership” that local
were packaged for delivery to farmers and experiments with people have over the research process, the scale of participa-
component rice technologies were conducted on-station. De- tory activities and stakeholder involvement and the levels of
spite the many available technologies generated at the station, management involved, and the level of disaggregation and rep-
the traditional research approach has had limited success in resentation of different stakeholders required for the research.
unfavorable rice environments, which have high biophysical, Clearly, there is no formula for deciding which level of
socioeconomic, and cultural diversity, and poor access to mar- participation is the best. The level chosen will depend on the
kets and infrastructure facilities. This created a mismatch be- objectives of the activity. What is more important is to have
tween research recommendations and farmers’ needs and ex- well-defined research problems and well-defined technologies
pectations and thus low or no adoption of technologies. Crop for a vigorous process whether with farmers or other stake-
and natural resource management technologies require an holders. It is important, however, to distinguish which of these
analysis of how farmers make decisions, develop a corrective levels of farmer or community participation we refer to when
heuristic, frame it as a hypothesis, and become motivated to research is called “participatory.”
participate in an experiment to test it. Similarly, germplasm
improvement requires an understanding of farmers’ preferences
Advances in participatory research at IRRI
and criteria for rice varietal selection in fragile environments.
We need to understand what “enabling inputs” may be required Through the years, progress has been made in participatory
to accelerate and widen adoption. These requirements neces- research at IRRI. Biological scientists in close interaction with
sitate the use of participatory research approaches through social scientists have tested, refined, and adapted protocols
which researchers and farmers can work as active partners in for farmer/community participatory research in natural resource
developing suitable technologies suited to their livelihood and management (NRM) and participatory plant breeding (PPB).
ecosystems. Such active partnership creates synergistic effects Participatory research (PR) has fast-tracked the adoption of
as it integrates the knowledge, experience, and skills of farm- innovations and in several cases has contributed to adoption
ers and scientists. on a large scale. Working with national agricultural research
and extension systems (NARES) on projects using the PR ap-
proach has changed the attitudes, mind-sets, and research ori-
What is participatory research? entation of scientists and research managers. The specific ad-
“Participation” and “participatory” have become such fash- vances in PR at IRRI in natural resource management and rice
ionable terms recently that any kind of activity involving a germplasm development are described below.
group of people is termed “participatory.” As these terms em-
brace a multitude of meanings, and these meanings become
Natural resource management
correspondingly dilute, a serious threat is posed to the use of
the term “participatory research.” The risk is that a catch-all In integrated pest management (IPM) in Vietnam, farmers
definition of participatory research is destined to fall out of “learn by doing” and decision rules are modified on the basis
fashion and to be discarded as fashion changes, without even of direct experience. A radical change in the research sequence
receiving the serious scientific evaluation of its potential that of IPM was made by starting with the farmers’ perspective
a rigorous but less trendy use of the term would invite. The during the problem definition or initial problem diagnosis. This
term “participatory research” is a collection of approaches that approach provided a mechanism for scientists to learn about
enable participants to develop their own understanding of and farmers’ decision constraints, determine research needs, use
control over the processes and events being investigated. It is research information, and “distill” these into testable hypoth-
loosely used to describe various types and levels of local in- eses for farmers (Heong and Escalada 2003, Huan et al 2004).
volvement in and control of the research process (Ashby 1997). In Bangladesh and in West Bengal, India, the use of the com-
munity participatory approach to research (CPAR) to evaluate
Session 13: Farmers’ participatory approaches to facilitate adoption of improved technology 409
O At the institutional level. The integration of PR ap- they are increasingly becoming the de facto farm managers
proaches into the main research and training programs and potential users of new technologies.
at IRRI should be supported by management. IRRI’s While most PR activities are strong in initial diagnosis,
comparative advantage lies in the application of par- planning, design, and testing, very few have had impact. Ex-
ticipatory research to strategic and pre-adaptive re- perience in scaling up should be documented.
search such as participatory research methodologies There is a need to study the impact of participatory re-
for use by NARES, NGOs, and others and in the man- search and assess other enabling mechanisms such as match-
agement of natural resources, plant breeding for ing funds to support partnerships among projects as well as
rainfed environments, IPM, geographic information resolutions, regulations (such as for the varietal release sys-
systems, decision support tools for soil management tem), and official instructions given by departments and local
and land-use planning, validation of technologies for governments.
rainfed environments, etc. There is a need to include We also need to reflect on the following issues:
farmer participatory research methods in IRRI’s and O To what extent should international research organi-
NARES’ training programs on conventional research. zations and NARES be involved in scaling up to ac-
There should also be a mechanism to strengthen link- celerate impact? (Methodologies)
ages and partnerships with NGOs and extension in- O What approaches to scaling up work and which do
stitutions. Donors, which pressure for more farmer not? (Methodologies)
participation, should make a commitment to the is- O Do international research organizations have the
sue with a long-term perspective and multiyear fund- “warm bodies” to do this with a reduction in nation-
ing. ally recruited staff and budget reductions? (Institu-
tionalization)
O How can we unite/complement formal traditional/con-
Conclusions and issues to study ventional research and informal science (participa-
Not all kinds of research can be participatory and PR is not a tory research)? (Research)
substitute for traditional or conventional research. Participa- O How can international research organizations facili-
tory research will be more effective if backed by formal re- tate mainstreaming or institutionalizing participatory
search or conducted in parallel with on-station research. Farmer research by NARES? How can the practice of PR be
participatory research is increasingly necessary as one moves brought out of donor-driven project activities? (In-
toward more diverse and complex environments. As one moves stitutionalization)
in this direction, recommendations cannot be broadly applied
and adaptive research becomes increasingly important.
References
Farmer participatory research can be an effective tool to
motivate farmers to “experiment” with new concepts, innova- Abedin Z. 2004. The emerging face of rice production in Bangladesh:
tions, ideas, etc. It is by far the most powerful tool for chang- direct wet seeding of rice using a plastic drum seeder to make
ing attitudes and beliefs, especially deeply entrenched attitudes rice production more profitable. Paper presented at the Inter-
(“more seed and fertilizer will give higher yields; all insects national Symposium on Rainfed Rice Ecosystems: Perspec-
tives and Potential, held at Indira Gandhi Agricultural Uni-
are bad and must be killed”).
versity, Raipur, Chhattisgarh, India, 11-13 October 2004.
Farmer participatory research helps breeders consider
Ashby J. 1997. What do we mean by participatory research in agri-
farmers’ needs and preferences in setting their breeding goals culture? In: New frontiers in participatory research and gen-
for a target environment. PVS, as a complement to conven- der analysis for technology development: proceedings. Inter-
tional breeding, provides farmers with opportunities to evalu- national Seminar on Participatory Research and Gender Analy-
ate and select the genotypes they want in their own fields un- sis for Technology Development, 1996. Cali (Colombia):
der their level of management before they are formally released Centro Internacional de Agricultura Tropical. p 15-22.
in the national systems. Atlin G, Paris T, Courtois B. 2002a. Sources of variation in partici-
The appropriate scale and level of representation of dif- patory varietal selection trials with rainfed rice: implications
ferent interest groups, the methods chosen, and the extent of for the design of mother-baby trial networks. In: Bellon MR,
local participation in and control over the research process Reeves J, editors. 2002. Quantitative analysis of data from
participatory methods in plant breeding. Mexico, D.F.: Centro
will depend on the project goals, available resources, and scope
Internacional de Mejoramiento de Maíz y Trigo. p 36-43.
of the research as well as the rationale for using the participa-
Atlin G, Paris TR, Linquist B, Phengchang S, Chongyikangutor K,
tory approach. Singh A, Singh VN, Dwivedi JL, Pandey S, Cenas P, Laza M,
Strategies and methods for involving farmers, local Sinha PK, Mandal NP, Suwarno. 2002b. Integrating conven-
people, extension people, research administrators, etc., will tional and participatory crop improvement in rainfed rice. In:
have to be followed on a case-by-case basis and may differ for Whitcombe JR, Parr LB, Atlin GN, editors. Breeding rainfed
germplasm improvement and NRM. Moreover, more women rice for drought-prone environments: integrating conventional
farmers should be included in participatory research because and participatory plant breeding in South and Southeast Asia.
Proceedings of a DFID Plant Sciences Research Programme/
IRRI Conference, 12-15 March 2002, International Rice Re-
search Institute, Los Baños, Laguna, Philippines. p 36-39.
410 Rice is life: scientific perspectives for the 21st century
Gregorio GB, Salam MA, Karim NH, Seraj ZI. 2004. Evaluation Paris TR, Singh RK, Atlin GN, Sarkarung S, McLaren G, Courtois
report of sub-project on development of high-yielding variet- B, McAllister K, Piggin C, Pandey S, Sngh A, Singh BN,
ies for coastal wetlands of Bangladesh. Dhaka (Bangladesh): Singh ON, Singh S, Singh RK, Mandak NP, Prasad K, Sahu
International Rice Research Institute. RK, Sahu VN, Sharma MK, Singh RKP, Thakur R, Singh
Heong KL, Escalada M. 2003. Farmer participatory experiments in NK, Chaudhary D, Ram S. 2002. Farmer participatory breed-
pest management. In: Pound B, Snapp S, McDougall C, Braun ing and participatory varietal selection in eastern India: les-
A, editors. Managing natural resources for sustainable liveli- sons learned. In: Whitcombe JR, Parr LB, Atlin GN, editors.
hoods: uniting science and participation. London (UK): Breeding rainfed rice for drought-prone environments: inte-
Earthscan Publications Ltd., and Ottawa (Canada): Interna- grating conventional and participatory plant breeding in South
tional Development Research Centre. p 210-212. and Southeast Asia. Proceedings of a DFID Plant Sciences
Huan NH, Thien LV, Chien HV, Heong KL. 2004. Farmers’ partici- Research Programme/IRRI Conference, 12-15 March 2002,
patory evaluation of reducing pesticides, fertilizers and seed International Rice Research Institute, Los Baños, Laguna, Phil-
rates in rice farming in the Mekong Delta, Vietnam. Crop Prot. ippines. p 32-36.
(In press.) Singh RK, Paris T, Thakur R, Singh VP. 2002. Institutionalizing
Linquist B, Bounthanh K, Houemchitsavak S, Horne P. 2004. Up- participatory plant breeding within the national research sys-
land research in Laos: experiences with participatory research tems: issues and future challenges. Paper presented at the In-
approaches. In: Participatory research and development. Vol. ternational Symposium on Agronomy Conference in New
3. CIP/UPWARD, Philippines. (In press.) Delhi. Organized by the Department of Extension, Indian
Paris TR, Singh A, Luis J. 2001. Listening to male and female farm- Council of Agricultural Research (ICAR), New Delhi, India,
ers in rice varietal selection: a case in eastern India. In: CIAT November 2002.
2000: an exchange of experiences from South and Southeast
Asia. Proceedings of the International Symposium on Partici-
patory Plant Breeding and Participatory Plant Genetic Re- Notes
source Enhancement, Pokhara, Nepal, 1-5 May 2000. Cali Authors’ address: Social Sciences Division, International Rice Re-
(Colombia): Centro Internacional de Agricultura Tropical. search Institute, DAPO Box 7777, Metro Manila, Philippines,
e-mail: t.paris@cgiar.org.
We present here a case study using a participatory approach in fertilizer and chemical fertilizer. When farmers use chemical
the regional development of sustainable rice-farming systems fertilizer, they have to think only about the necessary quantity
in Ogata village in northeastern Japan. After Mr. Tozawa’s in- of fertilizer and just place it on the ground. The fertilizer works
teresting report as a farmer, we would like to present, from a and dissolves soon. But, when they use organic fertilizer, they
scientific point of view, several principles of participatory re- have to consider a wider variety of factors: how long will it
search derived from our experience with the development pro- take organic fertilizer to dissolve under the ground? how much
cess for two contrasting technologies, organic and no-till, in of it will accumulate? and when and how will it work on plants?
Ogata. This presentation is based on the on-farm research we The reason organic fertilizer requires farmers to think more is
have carried out over the past 7 years in collaboration with that it works on plants through much more complex interac-
farmers of Ogata. The theme of this research is rural develop- tions among soil, water, and microorganisms than chemical
ment based on sustainable agriculture. As Mr. Tozawa said, fertilizer. Organic fertilization is more difficult because it is
Ogata is the largest single area of sustainable rice production more site-specific than chemical fertilization. The effect of
in Japan, and Ogata farmers are very active in using new tech- fertilizer may differ from one site to another. Thus, organic
nologies of sustainable farming. These research experiences farmers have to have much knowledge of soil science, plant
have taught us that participation and interactiveness are the nutrition, ecology, etc. No-till farming is also knowledge-in-
keys to success in the development of new sustainable tech- tensive in a different way from organic farming. Pioneering
nologies, for three reasons. farmers of Ogata have developed new technologies as partial
First, technologies for sustainability are much more technologies of no-till farming. These technologies include a
knowledge-intensive than conventional chemical-based tech- special type of transplanting machine for no-till and controlled-
nologies that farmers need much more time to learn, experi- availability fertilizer for seedlings. Thus, no-till farmers have
ment, and adapt them before they can decide when and how to to know much about the relationship between water perme-
use them. This can be seen in the difference between organic ability and the nature of soil, the most effective use of con-
Session 13: Farmers’ participatory approaches to facilitate adoption of improved technology 411
trolled-availability fertilizer, and the cost-reducing effect of requires the use of herbicide in early spring and after trans-
no-till and others. planting, and the use of chemical fertilizer as controlled-avail-
Second, since each sustainable agricultural technology ability fertilizer. This means that rice grown by no-till farming
is partial (not a total change of the production system), farm- is not regarded as “environmentally friendly” in the present
ers need to make mutual adjustments to incorporate a new tech- market. So, farmers cannot expect a premium. As Mr. Tozawa
nology: sometimes modifying the technology to make it more said, this is the most important reason why no-till has not spread
compatible with their existing farm management, and some- in Ogata.
times modifying their management of other components to To avoid these uncertainties, farmers tend to take a “step-
make them more compatible with the new technology. Farm- by-step” approach based on communication with other agents,
ers will use new technologies only when they reach the con- within the community, in markets, and in the public sector. In
clusion after experimentation that these are useful for their the above example, as was reported by Mr. Tozawa, “estab-
overall farm management. For instance, labor and agricultural lishing a study group and experimenting with new technolo-
machinery are important factors for mutual adjustments. One gies with other farmers are the most common ways of solving
Ogata farmer doubled his farm size by purchasing extra farm- problems.” There are 40 or 50 groups of this kind, and a farmer
land. He said he would be able to manage the farm of 30 ha often joins several study groups to learn various technologies
with only labor of his family of four adults. Another Ogata and materials. It takes time for Ogata farmers to use a new
farmer has been hesitating to begin no-till farming, mainly technology, but, once they accept it, that technology spreads
because he could not decide whether to make a new invest- rapidly all over the village. The use of controlled-availability
ment by purchasing a special transplanting machine for no- fertilizer is a good example. In 1998, it covered 1,214 ha, but
till. He is afraid that he would not be able to repay the loan if the latest statistics indicate that it covered more than 3,000 ha
the price of rice continues to decrease. in 2003. It is also common for other agents to be involved in
Third, the innovation process of sustainable farming tech- this process. As Mr. Tozawa said, Ogata farmers sometimes
nologies is always accompanied by uncertainties: the effects invite researchers to learn about new technologies and materi-
of new technology are not well defined. Technical success does als. In some cases, retailers and consumers participate in tech-
not necessarily mean economic rewards or social acceptance; nological innovation by encouraging farmers and inspecting
short-term results may change existing long-term goals. These soil conditions and residual pesticide with rice. A participa-
three types of uncertainty can be seen in the example of both tory approach parallels this natural process of farmers, and
organic and no-till farming. In organic farming, organic fertil- enables research and extension support agents in the public
izer is a cause of uncertainties. Its effects may differ according sector to interact with this process.
to weather conditions, for example. Especially during the pe-
riod of seedlings, organic farmers are extremely sensitive as
to whether organic fertilizer works properly. Even experienced
Notes
farmers sometimes fail with seedlings using organic fertilizer. Authors’ addresses: Y. Taniguchi, Department of Bioresource Sci-
Then they have to rush to dispose of the failed soil and seeds, ences, Akita Prefectural University; S. Sato, Department of
and to try again from the beginning. With no-till farming, a Bioresource Sciences, Akita Prefectural University, 010-0195
cause of uncertainty lies in the use of herbicide and chemical Nakano, Shimoshinjo, Akita City, Akita, Japan, e-mail:
tani@akita-pu.ac.jp, ssatoru@akita-pu.ac.jp.
fertilizer. The present technological system of no-till farming
West Godavari District is the rice bowl of Andhra Pradesh rice yields frustrated farmers and made them switch to horti-
State and the kingpin of aquaculture in India. Rice farmers cultural crops and aquaculture. In view of the existing soil and
have enjoyed the fruits of the Green Revolution with the intro- climatic conditions, the majority were still forced to adjust to
duction of new high-yielding rice varieties starting with IR8. rice cultivation and its stagnated yields. The farmers’ entre-
They reached a peak and have attained the highest rice yields preneurial behavior motivated them to turn to a new technique
in the state for more than two decades. Later hybrid rice vari- such as the System of Rice Intensification (SRI). Yet, the ini-
eties did not increase yield. The regular new high-yielding rice tial field problems and laborious technological skills involved
varieties became a ray of hope but have been disappearing in it have slowed down its introduction. But, to the surprise of
because of one problem or another, and continued stagnated everyone concerned, including the scientists, the farmers’ ini-
tiative in doing research in association with extension scien- of 20–30% has encouraged farmers to adopt this system on a
tists with the new SRI technology in farmers’ own fields as large scale in their fields.
well as at extension center farms helped to modify existing Farmers locally manufactured small implements such as
field situations and make them suitable for the new SRI tech- markers and weeders to suit their field conditions. They have
nology. This has helped with the successful implementation of made several modifications and developed the best of them
SRI in India. with their own efforts, such as ropes with beads, ropes with
The SRI was first introduced in India during the 2003 knots, bamboo sticks with nails, fixed line markers, adjustable
rainy season and it was simultaneously investigated at a uni- markers, and, ultimately, the multipurpose liner, which is now
versity extension center such as the Krishi Vigyan Kendra mostly used by farmers because it easily marks lines, both ver-
(KVK), Undi, West Godavari District, and in farmers’ fields. tical and horizontal, at the same time with less labor. Simi-
Technologies were prescreened in association with selected larly, weeders were locally manufactured, such as cono weed-
farmer groups through discussion and interaction sessions. Par- ers and rotary weeders with bearings, and a weeder with a mesh,
ticipation of farmer groups was encouraged at different stages etc. (Fig. 1).
of crop growth. At the harvesting stage, yield profitability, the The farmers also struggled for effective water manage-
cost-benefit ratio, practicability, and likes and dislikes of farm- ment practices in the delta region, which is a crucial step of
ers were assessed. Field days were conducted in farmers’ fields SRI cultivation under delta situations. These management prac-
as well as at the extension center. Farmer conventions were tices included making a small pit in one corner of the main
organized, duly inviting resourceful farmers and those farm- rice plots and pumping out the water, making inner new and
ers who had already adopted the SRI technology. This helped small bunds and deep alleyways at a 2-m interval, etc. The
with constructive interaction among the farmers and motivated farmers also experimented with the SRI method using various
others to adopt the technology on a large scale in the next sea- varieties in order to find out the most suitable ones (Fig. 2).
son itself.
Farmers as scientists have undertaken several experi-
Farmer participatory research
ments based on their own experiences. They themselves have
evolved location-specific and cost-effective devices such as An innovative and enthusiastic farmer, Mr. B. Sudhakar Reddy
line markers and weeders. This has encouraged other farmers of Akividu village of West Godavari District, is known for his
to modify accordingly. The findings reveal that real farmers’ realistic approach to rice cultivation. He successfully imple-
participation right from the introduction of technologies has mented the newly introduced SRI technique on 10 ha during
made farmers modify technologies according to their situa- the kharif and rabi 2003-04 seasons with varied combinations
tions, thus inventing needed implements such as markers. This of different varieties and spacings, with careful water and weed
has made it easy to adopt different and wider spacings in rice management practices. He achieved higher yields with all va-
transplanting, and the weeders were easy to operate even in rieties. He has proved that the SRI technology can easily be
critical field situations. The collective approach of farmers implemented in any field situation. He has obtained record
made it easy to manage water properly. Though at the initial yields as well as formed rice plots as rice gardens by using
stages more laborers were needed for leveling and transplant- wider spacing at 50 × 50 cm. He has also explained the tech-
ing young seedlings, this gradually became standardized with nology to hundreds of farmers and correlated it to other farm-
less labor and speedy transplanting. The additional crop yield ers’ situations. He has convinced others to adopt it in their
Session 13: Farmers’ participatory approaches to facilitate adoption of improved technology 413
Fig. 2. Location-specific water management practices and different varieties tested by farmers themselves for successful implementation
of SRI technology.
own fields. He has become a role model for others to emulate his conditions and found a way to overcome his chronic weed-
to adopt the SRI technique easily. ing problem.
He himself designed an adjustable marker and used it Large farmer Sri N.V.R.K. Raju belongs to
for marking lines. This has helped him to follow different spac- Gummampadu village of West Godavari District. Motivated
ings (i.e., 25 × 25, 50 × 25, 50 × 50 cm). by the KVK’s SRI demonstrations during kharif 2003, he
Weeding with a specially made rotary weeder was done implemented SRI on 40 ha during rabi 2003-04 on what was
at 7-d intervals to check further weed growth and to incorpo- originally a fish pond. When experiencing difficulty of paddy
rate weeds in the soil. He could locally manufacture a weeder cultivation under an acute water shortage, he was forced to
in collaboration with other farmers. switch to SRI cultivation after learning about it. He modified
As there was a drainage problem in his field, he made a several weeders and line markers to suit his field conditions
small ditch at one corner of his field and drained the excess apart from a channelized water management system. He
water from that field into the ditch and pumped the water out- pumped out the excess water in the field by using an oil en-
side with the help of a small oil engine. Deep alleyways and gine. He was the first farmer to cultivate paddy under SRI in a
small bunds were made to drain out excess water. larger area. He reaped a good harvest and motivated several
Small farmer Sri Tirupathi Srinivas has 1 ha of land. He others to follow this new technique.
belongs to Ballipadu village in Attili Mandal of West Godavari Many of these research efforts of a group of farmers in
District. He has cultivated paddy under the SRI method on their own fields have motivated several other farmers of the
about 0.5 ha. KVK, Undi, provided paddy seed and technical same village and neighboring villages to join in sharing their
guidance to him at regular intervals. He attended all training innovative efforts. This type of farmers’ innovative research
classes conducted on SRI technology at KVK. He followed all has become an eye opener even to our own researchers. This
the water and weed management practices regularly. As rats clearly shows that farmers are the real initiators of their own
were the major problem in that area, he installed a polythene research efforts; hence, if farmers are made partners in doing
sheet all around the field, erecting it to 1-m height. The paddy research in agriculture, a better outcome will be possible.
plants produced a high number of tillers and he got higher
yields, with 3,000 kg ha–1 more yield with SRI technology than Notes
with the conventionally transplanted variety MTU 1071. Dur-
ing winter 2004, he followed the SRI technology on his entire Author’s address: Acharya N.G. Ranga Agricultural University,
farm of 1 ha by joining with his brothers and he motivated 8 Andhra Pradesh, India, training organizer and head, Krishi
other young farmers of his village to adopt this new technol- Vigyan Kendra, Undi, West Godavari District, Andhra
Pradesh, India, e-mail: Drdandu_jraju@yahoo.co.in.
ogy. He reaped the highest yields in the district.
Young and dynamic farmer M. Ramesh from Kaikaram
village of Unguturu Mandal of West Godavari District struggled
with weeding practices. He modified different weeders to suit
Session 13: Farmers’ participatory approaches to facilitate adoption of improved technology 415
Farmers Researchers
advantages advantages
Goal and
Goals hypothesis
l cost reduction
l premium quality Methods
l environment l broadening scope of ideas to test
l verifiable methodologies
Farmer conditions
Research must be interesting Fusion of
(for example, use role playing) l quantifiable goals
l farmer participation
in goal setting
Fig. 1. Advantages (boxes) of participatory research for farmers and researchers (ovals).
1. What methods can enable farmers to contribute as re- first summarized its understanding of the advantages of farmers
searchers in the field to experimentation for technology and researchers participating together in on-farm research. Farm-
development? ers bring their goals and conditions, while researchers bring meth-
O assessment and prioritization of needs and opportu- ods that can help farmers assess new ways to achieve those
nities goals. The results are quantifiable goals and hypotheses devel-
O matching appropriate interventions [innovations to oped together (Fig. 1), leading to a shared, common representa-
test] tion of the problem to be examined and solved through farmer-
O rapport building researcher collaboration.
O assessment of what farmers have done so far (giv- The group then considered two questions for broadening
ing recognition to farmers’ own efforts and building participation:
on what they are doing already) 1. How can nonfarmer stakeholders (wholesalers, input
O technology introduction, farmer volunteerism, and ac- suppliers, local government, NGOs, etc.) contribute to
tive farmer/researcher partnership (researchers may technology development?
supply critical inputs but risks should be shared) Figure 2 shows the advantages for each type of stake-
2. What mechanisms and types of farmer-researcher in- holder. Business stakeholders can obtain concrete ad-
teraction can enable farmers to contribute their own vantages that will improve their business, while partici-
ideas for scientific evaluation, rather than simply involv- patory research can be a social laboratory for govern-
ing farmers to provide their evaluation of ideas from ment to learn about the actual and potential effects of
scientists? policies. Consumers can have a better chance of ob-
O consensus taining the type of rice (or other agricultural product) of
O rapport building (cultural sensitivity) the quality they desire.
O encouraging feedback between partners 2. How can technology development go beyond a few sites
O using community communication structures of intensive farmer-researcher interaction, without los-
O undertaking direct observation (in the field) and par- ing participation?
ticipant observation [researchers joining with farm- Several types of processes were proposed (Fig. 3). One
ers in carrying out trials in the field] process is based on farmers becoming trainers to en-
3. How can farmer participation help biological scientists able other farmers to undertake participatory research
do better and more relevant science? on their own, and participatory research spreading as a
O scientists gain access to farmers’ knowledge method spontaneously among communities. Two espe-
O scientists learn from farmers (who have a broader cially useful tools for this process are (1) study tours by
perspective of the farming system) communities new to the process to communities that
O widens scientists’ horizons already have experience and results through participa-
O enables scientists to be practically oriented (praxis), tory research, and (2) farmer field schools to develop
thus refining technology site-specific field management techniques through joint
O ensures appropriate technology development observation by farmers and researchers. The latter
Group B’s topic was “Broadening participation socially and method is well known for its successes in integrated
spatially.” Before considering broadened participation, the group pest management (IPM), but its use should be broad-
Large Niche
Wholesalers retailers retailers
Manufacturers
Fig. 2. Types of stakeholders (ovals) and advantages (boxes) of participatory research for each.
Other communities
Input
Retail suppliers
Other
Pilot community
Knowledge
communities: Media
workshops, l soap opera
b
information
technologies*
l 3 reductions,
3 gains
Spontaneous interaction
Message should be simple
*GIS, generic model
tailor each stakeholder can access
Web-based e-participation and e-learning
Session 13: Farmers’ participatory approaches to facilitate adoption of improved technology 417
ened to integrated biological management (IBM). A sec- Finally, results from participatory research can be used
ond process can grow out of the first, in the formation by extension through improved media approaches. These
of information communities. These can use workshops should use formats that farmers enjoy, such as soap
and, where possible, information technology, to develop operas (successful in the Philippines) and simple mes-
models that can be tailored to the needs of many types sages, such as “3 reductions, 3 gains” (successful in
of stakeholders, both farmers and nonfarmers. These Vietnam).
can provide a basis for e-participation and e-learning.
Rapid economic and income growth, urbanization, and glo- tem, the rice-wheat system, and the rainfed uplands. They ac-
balization are leading to a dramatic shift of Asian diets away count for about 80% of the agricultural population and some
from staples and increasingly toward livestock and dairy prod- 50% of the total agricultural area in Asia. The tropical low-
ucts, vegetables and fruit, and fats and oils. The tendency for land and rice-wheat systems are the dominant sources of rice
per capita rice consumption to decline with income growth supply in Asia. These systems witnessed rapid productivity
and with urbanization has been widely documented in the lit- growth during the Green Revolution and their productivity
erature (Ito et al 1989, Huang and David 1993). FAO projec- continues to be high in the post-Green Revolution period. Yet,
tions indicate that the per capita consumption of rice will level pressure to diversify out of rice is also the greatest in these
off by 2015 and start to decline by 2030 (FAO 2003). systems, primarily because of low returns to rice relative to
The rice sector in Asia is facing the dual challenge of high-value alternatives such as vegetables (Pingali et al 1997).
sustaining high rates of rice productivity growth while trans- The feasibility and cost of substituting other crops vary
forming itself from a subsistence-oriented monoculture sys- across all three farming systems. The flexibility of farmers to
tem to a diversified market-oriented system. This paper exam- respond to the changing relative prices and relative profitabil-
ines the scope for the diversification of rice-based farming in ity in their crop choice decision-making can be described in
Asia. Economic, agro-climatic, and technological constraints terms of the level of investments (both physical and human
to the commercial transformation of subsistence rice systems capital) required in switching from rice to nonrice crops and
are identified. Priorities for research, and the primary compo- vice versa. Flexibility is low during the rainy season for the
nents of a policy agenda, are described. tropical lowlands and the rice-wheat zone because the drain-
age costs for growing nonrice crops can be prohibitive (Pingali
et al 1997). Upland areas, however, can oscillate between rice
The diversification of rice-based farming systems and nonrice crops with minimum additional investment.
A recent FAO/World Bank study on farming systems and pov- Access to markets and the relative prices of rice and
erty suggested that diversification is the single most important nonrice crops, especially horticulture, are additional determi-
source of poverty reduction for small farmers in South and nants of diversification. Although roads and market places are
Southeast Asia (FAO and World Bank 2001). The three most important, proximity to urban areas expands the range of
important systems described in the study are all rice-based nonrice diversification options, especially for fresh produce.
farming systems (see Table 1): the tropical lowland rice sys-
Rice-wheat 9 22
Agricultural population 416 Irrigated rice; wheat, vegetables, livestock
Cultivated area 93 including dairy, off-farm activities
Irrigated area 158
Rainfed uplands 30 26
Agricultural population 636 Cereals, legumes, fodder, livestock, horticul-
Cultivated area 189 ture, seasonal migration, and off-farm ac-
Irrigated area 38 tivities
Source: Tables 5.1 and 6.1 in FAO and World Bank (2001).
Note: Population figures are in millions, area figures in million hectares.
Session 14: Potential for diversification in rice-based systems to enhance rural livelihoods 421
teristics, and greater tolerance of pest stresses. System-level The process of agricultural diversification should not be
research includes land management and tillage systems that expected to be frictionless. Significant equity and environmen-
allow for shifts of cropping patterns in response to changing tal consequences can arise in the short to medium term unless
incentives and farm-level water management systems that can appropriate policies are followed. Long-term strategies to fa-
accommodate a variety of crops within a season. Also impor- cilitate a smooth transition to commercialization include in-
tant at the system level is research on the carryover effect of vestment in rural markets, transportation and communication
inputs and management practices across crops, for instance, infrastructure to facilitate integration of the rural economy, crop
high insecticide applications or the effects of intensification in improvement research to increase productivity, and crop man-
terms of prolonged water saturation, the buildup and carryover agement and extension to increase farmers’ flexibility and re-
across crops of pest populations, rapid depletion in soil mi- duce possible environmental problems from high input use;
cronutrients, and changes in soil organic matter that could lead and establishment of secure rights to land and water to reduce
to reduced productivity of rice monoculture systems over the risks to farmers and provide incentives for investment in sus-
long term. taining long-term productivity.
Given growing populations and income-induced demand
for increased cereal consumption, there continues to be a strong
need to seek higher productivity levels for the staple cereals.
References
The need to increase the productivity of cereals is higher the FAO. 2003. World agriculture: towards 2015/2030. London (UK):
greater the diversion of high-potential irrigated lands to Earthscan Publications Limited.
noncereal pursuits. FAO and World Bank. 2001. Farming systems and poverty: improv-
ing farmers’ livelihoods in a changing world. Rome and Wash-
ington, D.C.: FAO and World Bank.
Agenda for a food and agricultural policy
Huang J, David CC. 1993. Demand for cereal grains in Asia: the
Commercialization trends require a paradigm shift in agricul-
effect of urbanization. Agric. Econ. 8:107-124.
tural policy formulation and research priority setting. The rel- Ito S, Peterson W, Grant W. 1989. Rice in Asia: Is it becoming an
evant development paradigm for the 21st century is one of inferior good? Am. J. Agric. Econ. 71:32-42.
food self-reliance, in which countries import a part of their Pingali PL, Hossain M, Gerpacio RV. 1997. Asian rice bowls: the
food requirements in exchange for diverting resources out of returning crisis? Wallingford (UK): CAB International.
subsistence production. Agricultural policy should emphasize Rosegrant MW, Binswanger H. 1994. Markets in tradable water
maximizing farm household income rather than generating food rights: potential for efficiency gains in developing country
surpluses. irrigation. World Dev. 22:1613-1625.
Governments have a difficult task to perform. On the Timmer CP. 1992. Agriculture and economic development revisited.
one hand, continued food security needs to be assured for popu- Agric. Syst. 40:21-58.
lations that are growing in absolute terms; on the other hand,
research and infrastructural investments need to be made for Notes
diversification out of the primary staples. As Timmer (1992)
Author’s address: Director, Agricultural and Development Econom-
stated, “Ultimately, the process of rural diversification must
ics Division, FAO, United Nations, Rome, Italy.
be consistent with the longer-run patterns of structural trans-
formation.”
Rice-based ecosystems in East and Southeast Asia are being for and constraints to diversification, adoption of technologi-
challenged by the simultaneous requirements for more diver- cal innovations, and a productivity increase in such systems.
sified products (Hossain 1998), increased productivity, and The interrelated issues at stake in rice-based ecosystems are
reduced environmental impact (Dobermann et al 2004). Loca- complex and require multidisciplinary research methods in
tion-specific conditions, such as access to labor and product which knowledge and information from different disciplines
markets, and biophysical conditions determine the potential are integrated and synthesized. Farm household modeling
Session 14: Potential for diversification in rice-based systems to enhance rural livelihoods 423
Table 1. Results (in units per year) of the reference, technological innovation, and vegetables scenarios
for farm households HH1 and HH2.
Thailand has experienced steady economic growth and struc- in 1999. In more recent years, it has hovered around 50%
tural changes in its economy in the last four decades that en- (Mundlak et al 2002).
abled it to gain a position among the newly industrialized na- Although Thailand has been very successful in reducing
tions. Although a great deal of this growth comes from indus- poverty because of rapid and steady economic growth, rural
trial development, the agricultural sector has contributed sig- poverty, especially in certain regions, is a serious problem.
nificantly to the process through exports, a cheap food supply, Rural-urban disparities also increased after the financial crisis
and the release of labor for industrial development. The struc- because of the inability of the urban sector to absorb rural la-
tural changes associated with economic growth reflect the bor at a rapid rate, and the declining importance of agriculture
changing role of agriculture in the economy. The share of ag- in the total value-added. We underscore in this paper that Thai
riculture in GDP declined from 44% in the early 1960s to 10% agriculture still has a major role to play, and the major chal-
in recent years (Isvilanonda 1998). Its share in employment lenge is to switch to high value-added products, in other words,
has shown a similar trend, albeit at a slower pace. The share of agricultural diversification.
agriculture in total employment fell from 83% in 1957 to 57%
Session 14: Potential for diversification in rice-based systems to enhance rural livelihoods 425
Background and aims of the study growing areas in Thailand, respectively. Three production en-
vironments are selected in each province. These two regions
Agricultural diversification in Thailand means primarily were investigated in surveys at different times—1995, 1998,
deintensification of rice production. Thailand became a major and 2001-02. The number of households investigated was 234
rice-exporting nation in the second half of the 19th century in 1995, 240 in 1998, and 280 in 2001-02.
because of its abundant suitable land. The favorable land/per-
son ratio is one of the reasons behind Thailand’s late adoption
Results
of land-saving modern rice technology. However, the nation
started losing its monopoly position in the international mar- Land use and cropping patterns
ket in the late 1960s. Traditionally, squeezing rice farmers According to our household-level survey data (2001), total
through an export tax, a rice premium, and government con- sown area is much smaller (only 2.19 ha) in KK than in SB
trol of rice sales was the main feature of Thai policies (11.7 ha), with a decline since 1995 in KK but an increase in
(Siamwala 1975). Discrimination against rice farmers has SB. The larger farm size coupled with the increase in cropping
weakened since the 1970s partly because of competition in intensity because of irrigation facilities has resulted in a rise in
the international market. The government supported the adop- average sown area in SB.
tion of new technology through investment in irrigation/infra-
structure, although this constitutes a small proportion of total Rice intensification
rice area (Isvilanonda 1998). In spite of this, rice is facing the In spite of greater cropping intensity, the percentage of sown
problem of low profitability, mainly because of the declining area under rice in 2001 in SB is lower (86.67) than in KK
demand in both the international and domestic markets. Ide- (90.67). In SB, rice intensification has declined since 1995 in
ally, the declining profitability of rice should induce farmers contrast to KK, where it has increased. Differences are con-
to switch to other crops. But this has not happened uniformly— siderable among the villages in SB, where rice intensification
the pattern of deintensification of rice and diversification into is lower in irrigated areas. Rice intensification in the flood-
other crops and nonfarm activities differs from region to re- prone areas is extremely high and has influenced the region’s
gion. average. In KK, on the other hand, rice intensification is higher
The overall aim of this paper is to analyze the pattern of in irrigated areas than in rainfed or drought-prone areas. Look-
diversification at the farm level, its effects on farm income, ing at the diversification pattern, it appears that, in SB, it is
and the constraints faced by farmers in different regions and cash crops and water chestnuts, rather than fruits, that led to
under different production environments. diversification out of rice. In KK, the importance of upland
This paper addresses the following questions: crops has declined with a corresponding increase in fruit trees.
O Is diversification out of rice more difficult in poorer Although all households grow rice in both regions, the
regions than in prosperous regions? proportion of households growing only rice is lower in SB
O Is diversification into other agricultural activities than in KK. In KK, it is increasing, whereas in SB it is de-
easier in irrigated areas growing modern rice than in creasing. Differences are considerable between SB and KK in
rainfed areas growing traditional rice? rice productivity: 3.44 versus 2.47 t ha–1, the annual average
O What factors affect the ability of farmers to switch to for each region as a whole. In SB, rice productivity differs in
high-value crops? different production environments. It is very high in irrigated
O How is diversification related to income inequality and flood-prone areas (4.5–5 t ha–l) compared with rainfed
in rural areas and across regions? areas with less than 1 t ha–l.
In contrast to SB, there are no significant differences in
productivity in different production environments in KK since
Concepts and measures used in the study farmers prefer to grow local varieties in these villages, even in
Three types of diversification have been considered: irrigated villages.
1. Crop diversification—diversification out of rice in-
tensification measured in terms of both land and value Diversification in terms of gross value
of production. of agricultural production
2. Diversification of agricultural production measured In both Suphan Buri and Khon Kaen, the shares of rice in terms
in terms of value of production. of sown area are much greater than the shares in terms of gross
3. Intersectoral diversification into nonagricultural ac- value of crops. But the trend is moving in the opposite direc-
tivities measured in terms of household income. tion. Rice is gaining importance in terms of both value and
sown area in KK, and is declining in importance in SB. There
are significant differences within each region depending on
The study areas and data the production environments.
The study is based on household-level data from two regions, In Suphan Buri, diversification has increased except in
Suphan Buri (SB) in the Central Plain and Khon Kaen (KK) in flood-prone areas. Both rice and water chestnuts are a top crop,
the Northeastern Province, the commercial and traditional rice-
Session 14: Potential for diversification in rice-based systems to enhance rural livelihoods 427
fication index for each household as the dependent variable whereas inequality has declined in the more prosperous re-
explained by access to land (+), price of rice (–), irrigation, gions.
infrastructure and water control (+), agricultural credit (+), It should be noted that these two regions are at different
proximity to urban areas (+), age of the household head (–), stages of agricultural transformation (Timmer 1988) and the
demographic pressure (–), and the proportion of rental land in policy implications for agricultural development and diversi-
total land (–). fication are therefore different. In the Central Plain, policies
In 1995, the above factors explained 25% of the varia- should be directed to removing contraints in the production
tion in diversification among the sample households. Irriga- and marketing of high-value crops. On the other hand, in the
tion and the age of the household head have turned out to be Northeastern Province, the main problem is to increase the
significant at the 1% level with positive effects and negative productivity of rice cultivation as well as cultivation of other
effects, respectively. The model worked poorly for 1998. It crops. Policies should be directed to securing property rights,
explained only 8% of the variation in the dependent variable, irrigation, infrastructure, and credit facilities.
and irrigation was the only significant factor at the 1% level.
The model using 2001 data explained 16% of the variation,
and several factors were found to be significant—irrigation
References
and the person/land ratio were significant at the 1% level, Isvilanonda S. 1998. Rice production and consumption in Thailand:
whereas the rental ratio was significant at the 5% level. The the recent trend and future outlook. Thailand J. Agric. Econ.
person/land ratio had a negative effect as expected. 55.
Mundlak, Larson, Butzer. 2002. Determinants of agricultural pro-
ductivity in Thailand, Indonesia and the Philippines. World
Conclusions and policy implications Bank Web site.
Siamwalla A. 1975. A history of rice price policies in Thailand. In:
Diversification primarily refers to deintensification of rice pro- Puey Ungphakorn et al, editors. Finance, trade and economic
duction and a switch to other cash crops, fruits, livestock, and development in Thailand. Essays in honour of Kunying Suparb
aquaculture. It has been found that rice intensification is higher Yossundara.
and increasing in the northeast compared with the Central Plain. Timmer PC. 1988. The agricultural transformation. In: Chenery H,
Also, diversification in terms of land-use pattern and gross Srinivasan TN, editors. Handbook of development econom-
value of agricultural production is lower in the former region ics. Amsterdam (Netherlands): Elsevier Science Publishers
than in the latter. B.V.
The factors that constrain diversification in these two
regions are quite different because of differences in the nature Notes
of farming (commercial versus subsistence). The main con-
straints perceived by KK farmers are the lack of access to pro- Authors’ addresses: Alia Ahmad, Department of Economics, Lund
duction factors, whereas in SB farmers face marketing prob- University, Sweden; Sompom Isvilanonda, Department of
Agricultural and Resource Economics, Kasetsart University,
lems. The differences in the ability to diversify in the two re-
Bangkok, Thailand.
gions are reflected in the growing inequality between regions,
Poor farmers are less able to cope with shortfalls in crop pro- rice production, and provides recommendations for develop-
duction and as a consequence diversify their activities as a ment strategies to progressively improve the crop-animal pro-
precaution. This occurs more frequently in rainfed areas where duction system. It is hypothesized that animal raising helps
uncertainties are higher because of unpredictable environmental increase rice income and improves production efficiency.
elements (IRRI 1995). In Asia, close to 30% of the total pro-
ductive area is categorized as rainfed; the majority of this is
Integrated crop-animal systems on small rainfed rice farms
rice-based, with animal raising as a key component (Thornton
et al 2002). Crop-animal integration on rainfed farms comes in two broad
This paper presents part of the results of the socioeco- categories: systems combining crops (mainly rice),
nomic study on crop-animal interactions under the project Sus- nonruminants, ponds, and fish; and systems combining crops
tainable Food-Feed Systems and Improved Livelihoods of the (also mainly rice) and ruminants (Devendra 1995). The most
Poor in Rainfed Lowland Areas, which is funded by the common is the interaction between ruminants, particularly
Systemwide Livestock Program. It seeks to explain the crucial cattle/buffalo and rice. Cattle/buffalo provide animal traction
role of animal production in promoting greater efficiency in during land preparation and manure (directly or indirectly) to
Session 14: Potential for diversification in rice-based systems to enhance rural livelihoods 429
430
Table 1. Costs and returns in rice production across project sites (in US$ per ha).
Group 1a
Yield 3.84 4.25 4.25 4.32 3.48 4.22 3.82 2.05 2.74 1.83
Total returns 1,485.56 663.2 663.2 858.97 511.21 818.53 1031.3 571.91 751.29 233.76
Total costs 487.53 615.47 615.47 589.23 415.43 569.02 516.03 321.15 397.25 344.28
Imputed 203.88 370.71 370.71 223.05 181.51 218.22 68.85 106.73 91.93 229.63
Cash 283.65 244.76 244.76 366.18 233.92 350.8 447.18 214.42 305.32 114.64
Group 2a
Yield 3.76 4.36 5.15 4.25 3.21 3.93 3.78 3.88 2.05 3.66 2.64
Total returns 1,453.57 626.79 1,080.29 675.82 671.9 566.11 588.15 967.56 359.83 893.89 340.53
Total costs 746.34 498.94 501.05 499.17 601.81 332.5 388.6 445 174.54 412.21 307.71
Imputed 109.35 207.53 102.19 196.14 210.22 91.85 116.51 37.3 30.07 36.42 145.97
Cash 636.99 291.41 398.86 303.02 391.59 240.64 272.09 407.7 144.47 375.79 161.74
Net returns above cash cost 816.58 335.38 681.43 372.79 280.31 325.46 316.06 559.86 215.36 518.1 178.79
Net returns above total cost 707.23 127.85 579.24 176.65 70.09 233.61 199.54 522.56 185.29 481.69 32.82
Group 3a
Yield 2.70 4.84 4.39 4.25 3.08 3.08 2.96 1.91 2.35 2.76
Total returns 976.82 872.53 941.45 938.70 644.23 644.23 616.62 555.38 580.9 316.92
Total costs 493.85 494.42 648.80 642.63 637.32 637.32 389.29 302.23 338.51 302.77
Imputed 129.11 91.06 228.13 222.65 284.02 284.02 35.39 57.43 48.25 128.01
Cash 364.74 403.36 420.67 419.98 353.30 353.30 353.90 244.80 290.26 174.76
Net returns above cash cost 612.08 469.17 520.78 518.72 290.93 290.93 262.72 310.58 290.64 142.16
Net returns above total cost 482.97 378.10 292.65 296.07 6.91 6.91 227.33 253.15 242.39 14.15
Average all
Yield 3.73 4.36 4.46 4.25 4.18 3.84 4.06 3.78 2.04 2.91 2.65
Total returns 1,437.32 641.81 955.34 795.69 835.00 554.67 735.32 979.01 559.05 769.03 319.32
Total costs 508.46 518.43 634.03 575.16 591.23 349.78 505.38 480.12 311.56 395.84 307.33
Imputed 189.73 230.84 215.54 223.33 222.28 110.53 182.55 54.42 98.04 76.23 141.21
Cash 318.73 287.59 418.49 351.83 368.95 239.24 322.83 425.7 213.52 319.61 166.12
Net returns above cash cost 1,118.59 354.23 536.85 443.86 466.05 315.43 412.49 553.31 345.53 449.42 153.20
Net returns above total cost 928.86 123.38 321.31 220.53 243.77 204.89 229.95 498.89 247.49 373.19 11.99
aGroup 1 includes farm households that own cattle or buffalo and use those animals in their rice cultivation. Group 2 includes farm households that own cattle and/or buffalo but that are not used in rice cultivation.
Group 3 includes farm households that do not own any cattle or buffalo.
Table 2. Results of the stochastic production frontier regression, IRRI-ILRI
Project, 2004.
crop fails. Larger farmers have better access to credit, often tier analysis also indicate that there is still some room for im-
for reasons of provision of collateral and lower administrative proving yield and efficiency in rainfed rice production. The
costs of servicing larger loans. But their yields are often not recommendations indicated are nothing new: the provision of
higher than those of small farms. water and the development of high-yielding and pest-resistant
Also shown in Table 2 are the estimated technical effi- varieties should continue to be on the priority list of the re-
ciency coefficients across the three groups of rice farm house- search and development agenda for these production environ-
holds. The overall technical efficiency estimate for all the re- ments, and public budgets for training and extension should
spondents is 0.69. It is highest among rice farm households be enhanced to reach remote and marginal areas.
belonging to Group 1 and lowest among farm households in
Group 3. The differences, however, are not very large.
References
Aigner DJ, Lovell K, Schmidt P. 1977. Formulation and estimation
Conclusions of stochastic frontier production models. J. Econometrics
Crop-animal integration is beneficial from the standpoint of 6(1):21-37.
the savings on input expenditures, particularly on energy use Devendra C. 1995. Environmental characterization of crop-animal
systems in rainfed areas. In: Devendra C, Sevilla C, editors.
in rice cultivation. The advantage of closer integration in terms
Crop-animal interactions. IRRI Discussion Paper Series No.
of promoting production efficiency, however, was not so strong.
6. Manila (Philippines): International Rice Research Institute.
The results seem to indicate the need to stress to farmers the p 43-64.
proper techniques for the use of animals in land preparation to
enhance their contribution to promoting production efficiency.
Additional results gathered from the stochastic production fron-
Session 14: Potential for diversification in rice-based systems to enhance rural livelihoods 431
Husain AMM, Hossain M, Janaiah A. 2001. Hybrid rice adoption in Thornton PK, Kruska RL, Henninger N, Kristjanson PM, Reit RS,
Bangladesh: a socio-economic assessment of farmers’ experi- Atieno F, Odero AN, Ndegwa T. 2002. Mapping poverty and
ence. Research Monograph Series No. 18. Bangladesh Rural livestock in the developing world. Nairobi (Kenya): Interna-
Advancement Committee, Bangladesh. tional Livestock Research Institute.
IRRI (International Rice Research Institute). 1995. Fragile lives in
fragile ecosystems. Proceedings of the International Rice Re-
search Conference, 13-17 Feb. 1995. Manila (Philippines): Notes
IRRI. 976 p. Authors’ address: A.S. delos Reyes, assistant scientist II; A.J.B.
Malabayabas, contractual researcher; M.A. Sombilla, consult-
ant, International Rice Research Institute, Los Baños, Philip-
pines.
A central part of current economic development in Vietnam Land use and diversification characteristics
concerns the transformation of the agricultural sector and de- of Vietnam’s delta regions
velopment of the rural economy. For a rice economy, like
Vietnam’s, this process includes the challenge to diversify out Table 1 displays the significant difference in the size of land-
of rice (cf. Jirström and Rundquist 1999, Taylor 1994). This holdings in the two regions, with the RRD farmers operating
paper focuses on the farm-level processes of agricultural di- farms only a third of the size of the farms of the Mekong farm-
versification in Vietnam’s two rice-bowl areas: the Mekong ers. In both regions, however, land use is dominated by rice,
River Delta (MRD) and the Red River Delta (RRD). which is grown intensively, with a cropping index of 172% in
the north and 219% in the south. In the Mekong, a rice intensi-
fication process raised the cropping index from 168 in 1996 to
Data, methodology, and organization of this paper 219 six years later. The Mekong farmers form part of a more
This paper presents results from a research project1 within general development of the 1990s in which farmers in several
which household panel survey data were collected on three Mekong provinces have switched from double to triple rice
occasions in the MRD—1996, 1999, and 2002—and on two cropping as a result of improved irrigation and drainage infra-
occasions in the RRD—1999 and 2002. The selection of study structure (Choeng-Hoy Chung 1997).
areas and households was carried out through a multistage In the Red River Delta, a higher proportion of cropped
stratified random sampling method. In the MRD, data gath- land is allocated to upland, or dry, crops, a fact that is ex-
ered from 180 households cover the time-span 1996-2002. In plained by the climatic conditions in the north not allowing
the Red River Delta, 189 farm households were surveyed in rice cultivation in the winter season. In terms of crop diversi-
1999 and re-surveyed in 2002. Our analysis in this paper will fication, Mekong farmers have made efforts to grow new crops
be limited to what we call rice diversification changes mea- to increase income. During the 1990s, many diversified into
sured as changes in the dependence on income from rice in the fruit production, but a combination of fluctuating and low prod-
total household economy and agricultural production system. uct prices as well as pest- and water management-related prob-
lems seems to have discouraged farmers to the extent that the
share of land allocated to fruit plantations dropped to half by
2002.
1“Structural Transformation of Southeast Asia’s Agricultural Systems: Agricultural Diversification in Vietnam,” funded by the Swedish Research Council
(Vetenskapsrådet). The project is carried out in cooperation with the Cuu Long Delta Rice Research Institute (CLRRI) in Can Tho and the Institute for Environ-
ment and Sustainable Development, Hanoi.
MRD RRD
Variables
1996 2002 1999 2002
aAll income values are expressed as 000 VND (Vietnamese dong) in constant 2002 values; US$1 =
15,279 VND (2002). n.a. = not available. T-tests for paired samples show that the difference between the
two periods for the MRD is significant at the 1% level for cropped area, rice sown area, and fruit crop area.
Also, for the MRD, significant differences at the 5% level are found for total household income, noncrop
agricultural income, and total off-farm income. In the case of the RRD, significant differences at the 5%
level, or higher, are found for all income variables. For land use only, the changes in the upland crop sown
area are significant, but this time at the 1% level.
Source: Information is based on sample surveys comprising 180 farm households in the MRD and 189
households in the RRD from a larger research project by the authors titled “Structural Transformation of
Southeast Asia’s Agricultural System: Agricultural Diversification in Vietnam.”
Turning our attention to household income and its com- bandry and aquaculture) has grown in the Mekong and now
position, the aggregate numbers in Table 1 show that Mekong represents some 10% of total income. In the RRD, its share
farmers, in spite of access to much larger landholdings, on has fallen somewhat but remained relatively higher at about
average have approximately 20% lower income than farmers 22% in 2002.
in the north. This finding is in line with the more general ob- The process of intersectoral diversification explains most
servation that rural inhabitants in the Mekong region become of the growth in income in both regions. Off-farm income has
relatively poorer than in other parts of the country (Govern- grown in the MRD at a rate high enough for it to retain its
ment of Vietnam 2000). The mentioned increasing intensity of relative importance—38% of total income. In the RRD, off-
land use in the MRD has not been able to reduce the increas- farm income has almost doubled, and farm households there
ing income gap between the two regions. In terms of depen- now depend more on off-farm income—about 52% of total
dence on rice income for total income, the regional difference income—than on on-farm income. While the proximity to ur-
is significant. In 2002, 46% of an average Mekong farm ban areas and income sources are more important in the RRD,
household’s income was derived from rice production, while farm families in the Mekong are more dependent on agricul-
an average Red River Delta household depended only to some ture-related off-farm income—38% of total off-farm income
16% on rice. The share of rice in the total income was never- in the MRD vis-à-vis 11% in the RRD.
theless decreasing in both survey areas (54% in the MRD in
1996 and 26% in the RRD in 1999).
In spite of the rather dramatic area intensification of rice
Behind the averages
production in the Mekong, farmers earned approximately the The farming and income characteristics discussed so far are
same income from rice in 2002 as in 1996. The lower profit- based on mean sample values. The studied agricultural diver-
ability of rice production in the MRD is also typical for the sification process is not, however, a smooth process including
RRD, where income from rice cultivation from the same size all farm households in an equitable manner. In this second sec-
of cropped land fell by some 25% from 1999 to 2002. tion of the paper, we try to capture some of the dynamics of
Because of the mentioned difficulties in diversifying into the diversification process by analyzing and discussing four
fruit production, nonrice crop income has decreased by 4% in categories of rice farm households.
the MRD. On the other hand, noncrop agriculture (animal hus-
Session 14: Potential for diversification in rice-based systems to enhance rural livelihoods 433
Methodologically, changes are described by looking at level of income diversification, the following observations can
specific categories of farm households and their “movement” be made for the four categories.
in rice dependence between the two points in time. Rice de- Significant ascenders as a group started out in 1996 at
pendence has been defined as the proportion of income from an income level well below the average total sample (cf. Table
rice in the household total income. These proportions have 1) but were operating slightly above average landholdings. Dur-
been categorized into four quartiles (0–25%, 25–50%, etc.). ing the period, they were able to allocate more of their total
In the second stage, changes in category affiliation were ex- harvested area to nonrice crops and were successful in raising
amined, and several new categories were defined. First, those their income from this source. Also, the increasing noncrop
who stayed in the first quartile at both times were classified as agricultural income contributed to their overall success, which
Stable high diversifiers (households displaying a high level of primarily was explained by a dramatic increase in off-farm
diversification, i.e., low dependence [d”25%] on rice income income.
in their total income). The opposite group (those who showed Significant descenders were, in terms of total income,
a high dependence [>75%] on rice income in their total in- starting from a relatively higher level than the previous group.
come) was classified as Stable low diversifiers. Second, those However, their access to land was only two-thirds of that of
who changed by two or three quartiles were classified as Sig- the ascenders and, as their off-farm income dropped, they were,
nificant ascenders and Significant descenders, respectively, in spite of some increase in land, and a significant increase in
depending on the direction of their changes. Finally, those who cropping index from 166 to 255, not able to compensate for
changed only one quartile between the two points in time were the negative change. The area used for nonrice crops did not
classified as “Moderate changers.” In the presentation (Table change, but their net income from this source turned negative.
2) and discussion below, we focus on only four of these cat- This was due to the problems of pest attacks and water man-
egories—Stable high diversifiers, Stable low diversifiers, Sig- agement related to fruit production.
nificant ascenders, and Significant descenders representing Stable high diversifiers enjoy high income from all in-
the more “spectacular” groups of the “stayers” and “movers.” come sources but rice. During the period, these households
Our data confirm a picture of significantly more diversi- allocated somewhat less land to nonrice crops (from 0.54 to
fied farm households in the RRD. In 2002, only 10 out of 189 0.40 ha) but were instead able to increase their noncrop agri-
households depended on rice production for more than 50% cultural income. Over the six-year period during which the
of their total income. In the MRD, the situation is more dy- majority of households in the total sample intensified their
namic, with approximately a fifth of the sample being involved cultivation by growing a third (summer-spring) rice crop, the
in rapid diversification out of rice. Stable high diversifiers group only marginally increased the
Among the four categories, the Stable high diversifiers cropping index of its land. This then seems well in line with a
at both times and in both regions show the highest household strategy of diversifying both agricultural and nonagricultural
income. However, the group Significant ascenders is rapidly income.
catching up, with a doubling—MRD 104%, RRD 115%—of Stable low diversifiers seem to follow a rice specializa-
total household income. Significant descenders, on the other tion strategy. They have significantly increased their landhold-
hand, are characterized by falling total income. Finally, in the ings, and, with a cropping index of 266, they are the most in-
Mekong, it is possible to discern a group of households—Stable tensive land users. An important question is whether such a
low diversifiers—characterized by above average and increas- strategy is sustainable over time. The possibility to further in-
ing size of landholdings following a rice specialization path. crease the intensity of land use is very limited, and only a ris-
This group, which depends almost completely on rice cultiva- ing price of rice will keep this group on a par with an economy
tion, has greatly increased its total income to a level well above experiencing overall economic growth.
the sample average. Data in Table 2 do not reveal any strong links between
the variables education and household labor, on the one hand,
and the diversification changes, on the other. Apart from the
Determinants and patterns of diversification striking difference in the level of education between heads of
Searching for determinants of the observed differences in di- household in the RRD and MRD regions, this variable does
versification, several variables were selected—access to fam- not differ significantly between the diversification categories.
ily labor, education, farm size, and sown area. The composi-
tion of income may in itself not only be an effect of the diver-
Conclusions
sification process, but also may determine its direction and
speed for individual households. This section focuses on the A household’s diversification process is complex. Our data
MRD region, where the share of farm households experienc- tally with other data sources (e.g., Government of Vietnam
ing significant changes in the composition of income is rela- 1994, 2000) pointing at an ongoing agricultural diversifica-
tively greater than in the already more diversified RRD re- tion process in Vietnam. The change in the RRD especially
gion. Furthermore, only in the RRD is there a group of farmers seems more unidirectional. In the Mekong, where rice domi-
following a nondiversification path. Starting with the house- nates to a greater extent, different strategies—whether adopted
holds experiencing the most dramatic changes in terms of the by preference or not remains to be analyzed—seem to co-ex-
Stable high diversifiers Significant ascenders Significant descenders Stable low diversifiers
1996 2002 1999 2002 1996 2002 1999 2002 1996 2002 1999 2002 1996 2002 1999 2002
Number of households 16 16 70 70 34 34 12 12 7 7 3 3 20 20 0 0
Total household income 21,353 24,164 23,332 27882 10,298 20,986 9,713 20,883 14,101 10,040 11,804 4,647 11,928 20,093 – –
Rice production income 2,431 2,520 3,664 2,655 9,188 5,416 5,737 2,910 3,814 8,939 1,508 3,000 12,040 18,702 – –
Nonrice crop income 6,496 3,894 2,311 1,594 124 1,133 1,255 1,736 1,551 -699 836 366 404 –176 – –
Noncrop agricultural 3,117 4,862 6,513 7,674 –730 1,766 1,928 3,719 510 664 1,701 265 –1,208 461 – –
income
Off-farm income 9,309 12,888 10,844 15,959 1,716 12,671 793 12,518 8,226 1,136 7,759 1,016 692 1,106 – –
Education (no. of y) 6.1 5.0 10.1 10.1 4.2 4.9 9.3 9.3 4.1 5.2 8.0 8.0 6.2 5.3 – –
Labor, family 15–65 y 2.9 3.7 3.5 3.4 3.0 4.4 4.3 4.0 2.6 4.0 3.0 2.7 3.2 4.2 – –
Total sown area (ha) 1.08 1.38 0.49 0.47 1.86 1.92 0.69 0.60 1.11 1.89 0.46 0.62 2.66 4.79 – –
Rice sown area (ha) 0.56 0.98 0.42 0.41 1.77 1.76 0.60 0.52 0.79 1.59 0.41 0.58 2.51 4.72 – –
Land operated (ha) 0.72 0.85 0.26 0.24 1.05 0.86 0.33 0.30 0.67 0.74 0.28 0.31 1.38 1.80 – –
aAll income values are expressed as 000 VND (Vietnamese dong) in constant 2002 values; US$1 = 15,279 VND (2002). T-tests for paired samples show that the difference between the two periods for the MRD is significant at
the 1% level for rice area for Stable high diversifiers. Similarly, significant differences at the 1% level are found for Significant ascenders in the MRD for total household income, rice production income, noncrop agricultural income,
and labor. Finally, for the MRD, a 1% level of significance is found for Stable low diversifiers for rice production income, noncrop agricultural income, total sown area, and rice sown area. Significant differences at the 5% level for
the MRD are found for education (Stable high diversifiers); education (Significant descenders); total household income and labor (Stable low diversifiers). Significant differences at the 10% level in the MRD are found for labor and
total sown area (Stable high diversifiers); nonrice crop income (Significant ascenders); rice production income, nonrice crop income, and rice sown area (Significant descenders); and land (Stable low diversifiers). In the RRD,
significant differences at the 1% level are found for total household income, rice production income, and off-farm income for Significant ascenders. Also, significant differences at the 5% level for the RRD are found for off-farm
income (Stable high diversifiers) and for total sown area and rice sown area (Significant descenders). Finally, significant differences at the 10% level are found for the RRD for total household income, total sown area, and rice sown
area (Stable high diversifiers); rice area (Significant ascenders); and noncrop agricultural income (Significant descenders).
Source: Information is based on sample surveys comprising 180 farm households in the MRD and 189 households in the RRD from a larger research project by the authors titled “Structural Transformation of Southeast Asia’s
Agricultural Systems: Agricultural Diversification in Vietnam.”
Session 14: Potential for diversification in rice-based systems to enhance rural livelihoods
435
ist. The next step in the search for determinants of diversifica- Jirström M, Rundquist F-M. 1999. Diversify or perish! Structural
tion in this project will be to break down the material further transformation of agricultural systems: intensification and
in order to analyze whether differences in “geographical capi- diversification in Vietnam. In: Cederlund K, Friberg T, Wikhall
tal” reveal any spatial diversification patterns at the district M, editors. Geografi I Lund, Essäer tillägnade Gunnar
Törnqvist, Institutionen för kulturgeografi och ekonomisk
and village levels (Sen 2003). Differences in physical access
geografi, Lunds Universitet. p 74-87.
to product and labor markets as well as varying agroecological
Sen B. 2003. Drivers of escape and descent: changing household
conditions at the village level will be research issues in this fortunes in rural Bangladesh. World Dev. 31(3):513-534.
next step. Taylor DC. 1994. Agricultural diversification: an overview and chal-
lenges in ASEAN in the 1990s. ASEAN Econ. Bull. 10(3):264-
References 279.
The role of nonfarm activities in promoting growth of the ru- determine whether it is a positive development for poverty
ral economy and reducing poverty is well documented (Hymer reduction (Islam 1984, Hossain et al 1994).
and Resnic 1969, Shand 1986, Ranis and Stewart 1993, This paper uses primary data available from a recent
Rosegrant and Hazell 2000). Termed the rural nonfarm survey of a nationally representative sample of rural house-
economy (RNFE), this sector accounts for a large proportion holds to (1) study the nature of the RNFE in Bangladesh, (2)
of rural employment and income, and grows faster than agri- analyze the determinants of participation in the RNFE, and (3)
culture with the development of the overall economy. As assess the impact of participation on a reduction in rural pov-
Rosegrant and Hazell (2000; p 81) observe, “From relatively erty.
a minor sector, often largely part-time and subsistence-oriented
at the early stages of development, the rural nonfarm economy
Data and methodology
develops to become a major motor of economic growth in its
own right, not only for the countryside but for the economy as The data for the study are drawn from a repeat survey of a
a whole. Its growth also has important implications for the nationally representative sample. The benchmark survey was
welfare of women and poor households, sometimes helping to implemented in 1987-88 by the Bangladesh Institute of De-
offset inequities that can arise within the agricultural sector.” velopment Studies (BIDS) on 1,245 rural households from 62
Generating productive employment for the growing la- villages in 57 districts, drawn by using a multistage random
bor force remains a formidable challenge for the Bangladesh sampling method. In the first stage, 64 unions were randomly
economy. The capacity to absorb the incremental rural labor selected from the list of all unions in the country. In the second
force in agriculture is extremely limited because of (1) no scope stage, one village was selected from each of the unions that
for expansion of the land frontier, (2) the intensity of cropping best represented the union with regard to landholding size and
has almost reached the limit, and (3) the growth of crop pro- literacy rate. A census of all households in the selected vil-
duction now depends almost entirely on technological progress, lages was conducted to stratify the households with regard to
resulting in low employment elasticity of output. Recent cen- the size of landownership and land tenure. A random sample
suses and labor-force surveys show a dramatic structural change of 20 households was drawn from each village. The Interna-
in the composition of the rural labor force in favor of nonfarm tional Rice Research Institute (IRRI) studied the same villages
activities (Hossain et al 2002). Questions arise, however, about in 2000-01. A sample of 30 households from each village was
whether the expansion is due to “push” or “pull” factors that taken using the stratified random sampling method. The strati-
fication was based on the wealth ranking technique of the par-
Session 14: Potential for diversification in rice-based systems to enhance rural livelihoods 437
Table 1. Factors affecting participation in rural nonfarm activities: estimates of a TOBIT
model.a
aThe dependent variable is measured as the share (%) of the nonfarm activity in total household income.
Numbers within parentheses are asymptotic “t” values. * denotes statistical significance at the 1% prob-
ability of error.
Source: Estimated from household-level data from the BIDS-IRRI surveys.
tence pressure indicate a smaller family size in households INCM = 3.56 + 399 WRKR + 0.22 CPTL + 26 EDCN (1)
engaged in service-sector activities. (0.90) (20.64) (35.11) (6.30)
Major determinants for participation in Business enter- + 403 INFR + 319 MIGRN
prises are the accumulation of nonland assets, a larger number (7.14) (6.57)
of workers in the household, and access to developed infra- + 0.29 MIGRN*CPTL R-square = 0.66, F
structure. Participation is higher in younger households. It is (16.84)
interesting to note that the intensity of the adoption of high- = 607, N of cases = 1,887
yielding rice varieties and the level of education of the worker
do not exert a significant influence on participation in trade where INCM is the amount of income from nonfarm sources
and business activities. Even the less-educated involve them- (US dollars), WRKR is the number of family workers engaged
selves in business provided they have access to capital. in nonfarm activities, CPTL is the value (US$) of nonland fixed
assets, EDCN is the years of schooling for the worker, INFR is
Contribution of the RNFE to income a dummy variable for villages with access to paved roads and
and reduction of poverty electricity, MIGRN is a dummy variable for households with a
The average household income is estimated from the survey at migrant member, and MIGRN*CPTL is an interaction term
$1,232 for 2000, a growth of 2.2% per year over 1987-2000. with migration and nonland fixed assets. The numbers within
The growth in rural income was almost entirely on account of parentheses are the estimated t-values for the regression coef-
nonagricultural activities. The fastest-growing economic ac- ficient. The value of R-square indicates that about 66% of the
tivities were business and services, followed by nonrice farm variation in nonagricultural income across households is ex-
labor. The income from rice production and agricultural wage plained by the variables included in the model.
labor declined in absolute terms. The share of nonagriculture Since the equation is estimated in linear form, the value
in total household income grew from 42% in 1987 to 54% in of the regression coefficient shows the marginal returns from
2000. the factor. Thus, a worker employed in nonagriculture earns
The following regression equation best represents the on the margin $399 per year, and the marginal rate of return on
variation in income earned from nonagricultural activities investment in nonagricultural capital is 22%. A worker earns
across the sample households: an additional income of $26 for each year of schooling. House-
Land owned (ha) Head count indexa Poverty gap indexa Squared poverty gap indexa
holds with migrant members earn an additional $319. If the bility from farm to nonfarm activities. As a result, the rural
village had electricity connections and access to paved roads, nonfarm sector has been expanding and has already become a
the household earns an additional $403. The statistically sig- major component of the rural economy. The findings in this
nificant regression coefficient of the interaction variable of paper demonstrate that the RNFE in Bangladesh is a vibrant
migration with capital indicates that remittances help the ac- subsector contributing to productivity growth and reduction
cumulation of nonagricultural capital, and that the capital is of poverty. Greater public-sector investment in education and
invested in activities that give higher returns (29%). expansion of the physical infrastructure will accelerate the pro-
The contribution of different factors to income is esti- cess of expansion of the nonfarm sector, and contribute to the
mated by dividing marginal returns by the average returns for sharing of benefits with resource-poor households.
specific factors (elasticity). It is found that 51% of the nonag-
ricultural income is on account of nonagricultural labor, 19%
References
on account of capital, 15% on account of education, and 8%
on account of infrastructure. There was a strong association of Ahmed R, Hossain M. 1990. Development impact of rural infra-
nonagricultural employment with education and infrastructure. structure in Bangladesh. IFPRI Research Report No. 83. Wash-
Thus, investment for education and infrastructure development ington, D.C. (USA): International Food Policy Research In-
contributes significantly to the growth of nonfarm income. Does stitute.
Foster JE, Greer E, Thorbeck E. 1984. A class of decomposable pov-
engagement in the RNFE help reduce poverty? To answer this
erty measures. Econometrica 52(3):761-766.
question, we estimated various measures of poverty and com-
Hossain M, Rahman M, Bayes A. 1994. Rural non-farm economy in
pared them for households in which at least one member is Bangladesh: a dynamic sector or a sponge of absorbing sur-
employed in nonagricultural activities with those that depend plus labor? SAAT Working Paper, International Labor Orga-
entirely on agriculture for livelihoods. Since land is a domi- nization, New Delhi.
nant asset in the rural economy, we also made similar com- Hossain M, Bose ML, Chowdhury A. 2002. Changes in agrarian
parisons for different landownership groups, to dissociate the relations and livelihoods in rural Bangladesh: insights from
effect of differential endowment of land. Table 2 reports the repeat village studies. In: Ramachandran VK, Swaminathan
findings. For the entire sample, the households with income M, editors. Agrarian studies. New Delhi (India): Tulika Books.
below the poverty line were estimated as 43% (head count p 369-391.
index). The poverty ratio was 31% for households engaged in Hymer S, Resnick S. 1969. A model of an agrarian economy with
non-agricultural activities. Am. Econ. Rev. 59(4):493-506.
nonagricultural activities vis-à-vis 48% for households engaged
Islam R. 1984. Non-farm employment in rural Asia: dynamic growth
exclusively in agriculture. The difference is more pronounced
or proletarization? J. Contemp. Asia 14:306-324.
for measures representing intensity and severity of poverty. Ranis G, Stewart F. 1993. Rural non-agricultural activities in devel-
Thus, diversification into nonagriculture contributes to a sub- opment: theory and application. J. Dev. Econ. 40:175-201.
stantial reduction in poverty. The conclusion holds when the Rosegrant MW, Hazell PBR. 2000. Transforming the rural Asian
effect of the differential endowment of land is controlled. economy: the unfinished revolution. Hong Kong: Oxford
University Press.
Shand R. 1986. Off-farm employment in the development of rural
Conclusions Asia. Australian National University, Canberra.
Session 14: Potential for diversification in rice-based systems to enhance rural livelihoods 439
SESSION 15
Challenges to expanding rice production
in unfavorable environments
CONVENER: S. Tobita (JIRCAS)
CO-CONVENER: R. Lafitte (IRRI)
In vitro selection of somaclonal and gametoclonal variants
for salt tolerance in rice
Nguyen Thi Lang, Dang Minh Tam, Hiromi Kobayashi, and Bui Chi Buu
Tissue-cultured cell lines can be selected in vitro for resis- many other crops, rice requires a unique growth regulator com-
tance to various stresses. Selection is done by placing a stress- bination, sucrose concentration, and light regime for each cul-
causing agent in tissue cultures containing dividing cells. Tis- tivar. Culture vessels are either glass screw vials or jars. The
sue culture techniques have been widely used for breeding vials contain 10 mL of medium and are used for callus initia-
purposes, especially in selection for stress tolerance. Tissue tion. The jars containing 20 mL of medium are used after em-
culture is a source of genetic variability that gives rise through bryogenic callus, which is isolated and ready for plant regen-
genetic modifications during the process of in vitro culture to eration.
a phenomenon called somaclonal variation. The possible causes For callus induction, mature seeds are surface-sterilized
of somaclonal variation include chromosome aberrations, DNA and placed on the appropriate medium for a cultivar. At the
amplification, and the occurrence of transposable elements. end of 28 days—referred to as a “passage”—the original ex-
Salt tolerance is an important plant character in areas where plant and all of its associated calli are transferred to a fresh
seedling growth is a problem. This paper reports on the cur- medium. After the second passage, embryogenic calli are iso-
rent status of tissue culture technology at the Cuu Long Delta lated and transferred again to a fresh medium. Embryogenic
Rice Research Institute (CLRRI). Our primary objective is to calli are usually yellow or cream-colored and dense in appear-
demonstrate that useful germplasm can be obtained from tis- ance. For long-term maintenance, embryo calli must be care-
sue culture and another way found to produce salt-tolerant lines fully selected and all nonembryogenic and dead calli removed
using anther culture techniques in which anthers of F1 hydrids at each transfer.
having one or both parents with salt tolerance are cultured. A Three crosses were tested for callus induction and plant
significant feature of this work is to determine what in vitro regeneration. The F 1 seeds were grown in pots in the
treatment, including concentration of the stressing agent and screenhouse, and panicles were collected at the stage when
the period of time in culture, will produce the greatest prob- the auricle distance of the flag leaf to that of the subtending
ability of field-tolerant plants for a specific environment. This leaf was 5–10 cm. The anthers were plated in various callus-
study aimed to create genetic variability in rice cell culture to induction media when the pollen grains were at the mid-uni-
obtain salt-tolerant plants through in vitro selection. nucleate to early binucleate stage of development. The plated
anthers were kept in cold at 10 °C for 8 d and were then trans-
ferred under dim light at 25 °C until calli formed.
Materials and methods Calli of about 1 mm in diameter were plated in N6 + 2
Twelve rice cultivars were used to conduct in vitro selection mg L–1 2,4-D + 1 mg L–1 NAA medium and incubated under
for salt stress. The work on salt stress tolerance was done at 1,000 lux at 25 °C for 4 wk. The calli were then transferred to
concentrations of 0.5%, 1.0%, and 1.5% NaCl. MS + 1 mg L–1 kinetin + 0.5 mg L–1 NAA + 2 mg L–1 BA
Mature seeds were dehusked and surface-sterilized with medium, until the plantlets were ready for transfer to normal
70% ethyl alcohol for 30 sec and then for 30 min with 5% growth conditions.
sodium hypochlorite (commercial bleach is mainly hypochlo- The plantlets were grown in nutrient solution for 2 wk.
rite). Surface-sterilized seeds were rinsed several times with The plantlets from each callus were individualized, with each
sterile distilled water before inoculation on callus induction one considered as a line. These were then transferred to pots
medium, consisting of MS basal organic and inorganic com- and grown in the phytotron until maturity.
ponents (Murashige and Skoog 1962) supplemented with 3.0% Seeds were then surface-sterilized with 0.1% HgCl2 for
sucrose, 2,4-D at 1.0–2.0 mg L–1, and kinetin at 0.0–0.5 mg 2–3 min and rinsed thoroughly with distilled water. The steril-
L–1 (Nguyen Thi Lang 2002). ized seeds were soaked in water for 24 h and incubated for 48
The cultures were incubated under both dark and light h at 30 ºC. The pregerminated seeds were sown, two seeds per
conditions (16 h day/8 h night). The temperature was main- hole and 10 holes per variety, on a styrofoam sheet with 100
tained at 22–25 ºC. The MS culture medium was supplemented holes with a nylon net bottom. The sheet was floated in dis-
with 3.0% sucrose, 0.8% agar, and various concentrations of tilled water for 4 d in nutrient solution in a plastic tray and
growth regulators. The medium pH was adjusted to 5.8 prior salinized to EC = 6 dS m–1 by NaCl. After 4 d, the seedlings
to autoclaving at 15 psi for 20 min. were salinized to EC = 12 dS m–1 by adding NaCl to the nutri-
The basal medium for rice tissue culture used is similar ent solution. The culture solution was renewed weekly and its
to that developed by Murashige and Skoog (1962). Unlike pH was maintained daily at 5.0 by adding either 1 N NaOH or
Tequing/Soc Nau obtained from the medium MS + 1 mg L–1 sults indicated that Teqing/Doc Phung and Tequing/Pokkali
kinetin + 0.5 mg L–1 NAA + 2 mg L–1 BA. In screening of were induced on calli-producing green plants (3.1, 18.2, and
anther culture-derived lines in a salt-screening nursery, a total 0.9 reciprocity). Among 23 lines tested, three promising lines
of 23 lines were evaluated for salt tolerance. Among 23 lines from Tequing/Doc Phung were identified to be better than the
tested, 3 promising lines from Tequing/Doc Phung were iden- resistant check.
tified to be better than the resistant check. The performance test of the 23 selected lines in the field
was within 10 dS m–1. The resulting three lines developed from
Estimation of salt tolerance in saline soil this study have a potential for salt tolerance since their agro-
A total of 104 A4 lines, the original variety having been used nomic characters are much better than those of the original
as a control, were tested in normal and saline fields. Some 23 cultivar.
lines exhibited salt tolerance. Variations were considerable
among the lines. Nine lines were selected for further observa-
tion. Grain filling per panicle, 1,000-grain weight, spikelets
References
m–2, and yield were statistically different at the 5% level of Chu CC, Wang CC, Sun CS, Yin KC, Chu CY, Bi FY. 1975. Estab-
probability from the original variety. lishment of an efficient medium for anther culture of rice
through comparative experiments on nitrogen sources. Sci.
Sin. 18:659-668.
Conclusions and suggestions Murashige T, Skoog F. 1962. A revised medium for rapid growth
and bioassay with tobacco tissue cultures. Physiol. Plant.
MS + 2,4-D (2 mg L–1) is considered as a medium suitable for
15:473-497.
cultivating target cultivars. The genotypes Doc Do and Nep Nguyen Thi Lang. 2002. Protocol for basics of biotechnology. Ho
Ao Gia and others such as Pokkali, Soc Nau, Mong Chim Roi, Chi Minh City (Vietnam): Agricultural Publishing House.
and KDM 105 are also suitable for embryogenic callus forma-
tion.
The regeneration MS medium supplemented with 0.5 Notes
mg L–1 BAP and 1.0 mg L–1 NAA affected Trang Thai Lan (82 Authors’ addresses: Nguyen Thi Lang, Dang Minh Tam, and Bui
shoots per E callus from 0.5 to 1.0 cm, obtained 18% E calli Chi Buu, Cuu Long Delta Rice Research Institute, Omon, Can
from shoots) and Trang Diep (3 shoots per E callus, obtained Tho, Vietnam; Hiromi Kobayashi, National Agrobiological
18% E calli from shoots). The other cultivars have been con- Research Institute, Tsukuba, Ibaraki, Japan, e-mail:
tinuously observed. ntlang@hcm.vnn.vn.
In NaCl 0.5% medium, E calli of all cultivars developed
normally. Both treated and nontreated calli that survived in
1.0% and 1.5% NaCl were of considerable benefit for further
studies.
Variation in culturability in terms of callus induction and
plant regeneration among these crosses was observed. The re-
In West and Central Africa (WCA), deepwater area accounts submergence conditions for several days, especially for juve-
for 6% of the whole rice-growing area, mainly in the flood- nile plants.
plains of large rivers such as the Niger River; and rainfed low- Water has slower rates of gas exchange, less capacity to
land and mangrove swamp account for 20% and 7%, respec- hold gases inside, and a higher extinction coefficient for light
tively (Jones 1999). About one-third of all rice cultivation area than the air. Floodwater is usually turbid so that severe shad-
faces danger from waterlogging. In addition, submergence ing occurs in the water. Under complete submergence, there-
could occur even in irrigated lowland at the seedling stage fore, photosynthesis and respiration of rice plants are depressed.
because of poor water management. Low rates of gas diffusion cause the accumulation of physi-
In the floodplains, sudden flood can submerge the crop ologically active gases such as ethylene produced in situ in the
and waterlogging could continue during the growth of rice. To plant (Jackson and Ram 2003). Visible symptoms of rice plants
avoid submergence damage, farmers sometimes dibble rice under submergence are faster elongation of a few leaves, yel-
seeds on ridges (Singh et al 1997). In some locations such as lowing of old leaves, slow or negative dry-matter growth, and
Mopti in Mali and the Sokoto Rima River floodplains in Nige- shoot decay. After a reduction in water levels, most leaves or
ria, Oryza glaberrima, which is the other cultivated rice spe- the whole shoot may collapse. Some can die and the survivors
cies besides O. sativa, and domesticated in WAC more than also suffer from poor rates of new leaf emergence and lodging
3,500 years ago (Jones et al 1997), is still cultivated in spite of (Jackson and Ram 2003). When rice plants under submergence
its unfavorable traits in yield formation, that is, lodging and succeed in re-contacting with the air because of elongation and
grain shattering, perhaps because of its high adaptability to survive, they are highly likely to lodge after the water level
waterlogging. Traditional O. sativa cultivars are preferably recedes.
cultivated by farmers, too. One example is Gambiaka, which
is considered to be adaptable to medium-deepwater conditions Varietal development for tolerance
(50 to 100 cm deep). This cultivar has more vigorous seedling of complete submergence
growth than improved cultivars; Gambiaka’s height and dry Physiological mechanisms of submergence tolerance are well
weight of 21-day-old seedlings were 42.3 cm and 123 mg, while elucidated in rice compared with other abiotic constraints such
those of WITA 4, WARDA’s improved cultivar released in as drought. Jackson and Ram (2003) have widely reviewed
Nigeria for the rainfed lowland ecology, were 28.9 cm and 75 mechanisms of injury by and tolerance of complete submer-
mg, respectively. Gambiaka therefore has a higher possibility gence in rice referring to both current and pioneer studies. The
to escape from submergence than WITA 4. best source of submergence tolerance in rice was already iden-
In savannah and forest zones in WCA, inland valleys tified, which is Indian cultivar FR13A selected from a local
prevail and rice in valley bottoms experiences waterlogging landrace. Mackill et al (1993) introduced the tolerance into an
for several days after excessive rainfall. Farmers transplant agronomically useful cultivar by conventional breeding. Al-
large older-aged seedlings in fields to keep them away from though the tolerance was considered to be a quantitative trait
possible submergence. However, transplanting of such seed- controlled by several genes, studies with molecular markers
lings is associated with lower yield than appropriate-aged seed- revealed that a single locus, Sub 1, controlled most of the tol-
lings because larger seedlings receive more damage at trans- erance (Xu and Mackill 1996). This locus was mapped on rice
planting and have a shorter remaining growth duration after chromosome 9, and a fine-scale map around Sub 1 was also
transplanting, which cannot compensate for the damage. developed (Xu et al 2000). The markers mapped close to Sub
1 can reduce the need for time-consuming and high-cost field
screening.
Waterlogging and rice plants
Damage to rice plants under waterlogging
Rice is an aquatic plant and is able to grow well under flooded
Plant types adaptable to waterlogging in WCA
conditions. In deepwater conditions, rice plants promote the In the floodplains of large rivers, waterlogging could continue
elongation of the coleoptile, leaf, or stem, which can reduce for a long time. Since existing cultivars with tolerance of com-
the danger from submergence. However, rice receives severe, plete submergence can survive or receive less damage than
sometimes fatal, injury when it is under complete overhead normal cultivars under submergence of several days, their type
is not suitable. An elongation type that can escape from sub-
60
40
20
80
CG 14 CG 20
60
40
20
80
Norin 30 IR36
60
40
20
0
0 5 10 15 20 25 0 5 10 15 20 25
Days of submergence
Fig. 1. Changes in the rate of increased height to the initial height with the time of the
submergence treatment (cited from Futakuchi et al 2001). Vertical bars indicate stan-
dard deviation of three replications. Closed circles = submergence; open circles =
nonsubmergence (control).
mergence is desirable. To tolerate complete submergence ing in contact with the air by leaves or other organs. Since
caused by sudden flooding until plants re-contact with the at- sudden flooding could occur in the bottoms, the same type as
mosphere by elongation, it will also be desirable for them to with the floodplains mentioned above will be highly suitable.
have submergence tolerance to a certain extent. Vigorous ini- Setter and Laureles (1996) have reported that there is a
tial growth will be important, too, to reduce the danger from trade-off between stimulated elongation and tolerance in com-
submergence. Resistance to lodging or recovering ability from pletely submerged conditions, suggesting difficulty in devel-
lodging after a water level reduction will be required for prac- oping cultivars with both elongation ability and tolerance.
tical cultivars acceptable to farmers. Futakuchi et al (2001b) have tested Oryza glaberrima in rela-
In inland valley bottoms, submergence-tolerant types will tion to elongation ability and tolerance under submergence.
be useful. In existing cultivars such as FR13A, however, the Four O. glaberrima lines, out of which two were identified as
tolerance does not mean unharmed by submergence (Jackson submergence-adaptable types (TOG 5810 and TOG 6283) and
and Ram 2003). Although such cultivars relatively tolerate sub- the others were identified as upland-adaptable types (CG 14
mergence, they also have some damage and collapse more or and CG 20), were cultivated in pots and submerged at the end
less after submergence for several days. Rice has well-devel- of the stem elongation stage with two O. sativa checks (Norin
oped aerenchyma and it receives much less damage when keep- 30 and IR36). Figure 1 depicts changes in relative elongation
1,800
1,600
1,400
1,200
1,000
800
600
400
200 CG 14 CG 20
0
1,800
1,600
1,400
1,200
1,000
800
600
400
200 Norin 30 IR36
0
0 5 10 15 20 25 0 5 10 15 20 25
Days of submergence
Fig. 2. Changes in the rate of photosynthetic oxygen evolution in a leaf slice with the time of
the submergence treatment (cited from Futakuchi et al 2001). Vertical bars indicate stan-
dard deviation of three replications.
rate—the percentage ratio of increased height to the initial medium-deepwater: 50 cm deep (Futakuchi et al 2001a). This
height—after the beginning of the submergence treatment. treatment started at maximum tillering and continued for 4
Submergence promoted elongation for all lines/varieties tested weeks. CG 14 showed the best elongation at 2 wk after the
but the extent of elongation was much larger in O. glaberrima treatment and a few interspecific progenies rivaled CG 14
than in O. sativa. Judging from the elongation ability, O. though elongation was promoted for all interspecifics during
glaberrima should have much less tolerance of submergence the treatment. One week after the end of the treatment, how-
than the O. sativa checks. However, O. glaberrima maintained ever, the actual plant height of the interspecifics in the control
high photosynthetic rates, which were determined by oxygen (always in shallow-water conditions) caught up with that of
electrode, compared with Norin 30, and rates similar to those the interspecifics grown in the treatment. Yield and tiller num-
of IR36 (Fig. 2). The same result was obtained with chloro- bers were much depressed by the medium-deepwater treatment.
phyll content. O. glaberrima had better tolerance than expected However, lodging was observed only with CG 14.
from elongation ability under submergence. The growth of O. glaberrima at the seedling stage is
Out of the O. glaberrima lines tested, CG 14 showed the vigorous. For example, plant height and dry weight of CG 14
fastest elongation soon after submergence (Fig. 1) though the at 21 days after seeding were 46.4 cm and 54.3 mg, while those
line had been identified as an upland type. Thirty interspecific of Bouake 189, a leading O. sativa variety for lowlands in
progenies from the cross of CG 14 and an O. sativa cultivar Côte d’Ivoire, were 30.4 cm and 38.0 mg, respectively.
(WAB56-104) were tested in terms of elongation ability in
Many rice-growing areas in the world do not enjoy the avail- tion to drought. For example, deep-root characters that are
ability of full irrigation water, and a rice crop often experi- useful to minimize the adverse effects of drought in some up-
ences drought at some time during its growth. Responses of land conditions may not be expressed fully under rainfed low-
different genotypes to water stress have been studied for a long land conditions where anaerobic soil conditions prevent the
time, and several physiological and morphological characters development of deep root systems. Timing of drought is an-
have been suggested to be responsible for drought tolerance in other ecological factor; if drought develops during the vegeta-
rice. Grain yield of some genotypes is affected less than others tive stage, the genotype’s ability to recover from drought is
by drought, but genotypic adaptation to drought is not consis- important, whereas, if drought develops around flowering time,
tent across different drought conditions. The single factor de- the plant’s ability to tolerate drought becomes a key factor in
termining genotypic variation in yield under drought condi- producing higher yield than with other genotypes. This paper
tions is crop phenology, particularly flowering time. However, describes physiological and morphological characters that are
within the same phenology group, no particular physiological useful under different types of drought (Table 1), and how that
or morphological characters have been found to be consistently understanding can assist in the selection of drought-tolerant
responsible for genotypic variation in yield. This lack of uni- rice cultivars.
versally useful traits has resulted in rather slow progress in
developing drought-tolerant cultivars in rice (Fukai et al 1999).
Some of the inconsistency in genotypic yield perfor-
Terminal drought
mance is caused by different types of drought that the rice crop Among genotypes of a similar maturity type, genotypes that
may encounter. Ecological factors such as the presence of flood can maintain high leaf water potential are often advantageous
water before drought develops, as is often the case in rainfed in producing higher yield under terminal drought conditions
lowland rice, are now recognized to affect genotypic adapta- (Jongdee et al 2002, Pantuwan et al 2002). Higher water po-
aNumbers in parentheses are total number of QTL clusters for the corresponding two trait groups. bLength of marker intervals
in centi-Morgans. cChromosome numbers of the corresponding locations of QTL clusters. dThe letter “s” in parentheses after
trait names indicates that the corresponding traits were recorded under severe water stress environments. eClusters 1 and 2
are close, at 4.0 cM distance away. fWith 24.2 cM apart from RG345 in cluster 2.
Iron toxicity is a widespread nutrient disorder affecting the Iron toxicity is a complex nutrient disorder and the defi-
growing of wetland rice in the humid tropical regions of Asia, ciencies of other nutrients, especially phosphorus (P), potas-
Africa, and South America. Large areas of wetlands ideally sium (K), calcium (Ca), magnesium (Mg), and zinc (Zn), are
suited for rice production remain underused, especially in West considered in the occurrence of iron toxicity in rice (Ottow et
and Central Africa, because of iron toxicity as a constraint. al 1983). Other nutrients may play an important role not only
Iron toxicity has been reported to reduce rice yields by 12– in reducing the effect of iron toxicity but also in the expres-
100% depending on the intensity of the stress and tolerance of sion of iron tolerance by various rice cultivars (Sahrawat et al
the rice cultivars (Sahrawat et al 1996, Sahrawat 2004). Iron 1996, Sahrawat 2004). Deficiencies of P, K, Ca, Mg, and man-
toxicity of wetland rice is associated with a high concentration ganese (Mn) decrease the iron-excluding power of rice roots
of ferrous iron in soil solution (Ponnamperuma et al 1955). and can thus affect the rice plant’s tolerance of iron toxicity
The stress occurs in reduced soils when a toxic amount of fer- (e.g., see Yoshida 1981, Sahrawat 2004). Deficiencies of Ca,
rous iron is mobilized in soil solution in situ or when inflow Mg, and Mn are not commonly observed in lowland rice, ex-
brings in soluble iron from upper slopes (van Breemen and cept probably on acid sulfate soils; deficiencies of P, K, and
Moormann 1978). Zn therefore deserve special attention (Yoshida 1981).
Iron toxicity occurs in soils (mostly Ultisols, Oxisols, This paper critically reviews recent research on the role
and acid sulfate soils) high in active iron and potential acidity, of tolerant genotypes and plant nutrients in reducing iron tox-
irrespective of organic matter and texture. But texture, cation icity. The ultimate goal is to provide information that can be
exchange capacity, and organic matter content influence the used for increasing rice production and productivity on iron-
concentration of ferrous iron in soil solution, in which iron toxic wetlands on a sustainable basis.
toxicity occurs (van Breemen and Moormann 1978). Plant-
and growing-medium-related factors such as plant age, accu-
Tolerant genotypes for reducing iron toxicity
mulation of hydrogen sulfide, organic acids, and other reduc-
tion products also influence iron toxicity occurrence in rice Rice cultivars differ in their tolerance for iron toxicity and the
(Sahrawat 2004). selection of rice cultivars with superior iron tolerance is an
Iron toxicity symptoms vary with rice cultivars. They important component of research for reducing iron toxicity.
are characterized by a reddish brown, yellow, or purple-bronz- Genetic differences in adaptation to and tolerance for iron-
ing or orange discoloration of the lower leaves of the rice plants. toxic soil conditions have indeed been exploited for develop-
Typically, iron toxicity symptoms are manifested as tiny brown ing rice cultivars with tolerance for iron toxicity (Gunawardena
spots starting from the upper tips and spreading toward the et al 1982, DeDatta et al 1994). Breeding and screening ef-
bases of the lower leaves. With progress in iron toxicity, the forts at the International Rice Research Institute in the Philip-
brown spots coalesce on the interveins of the leaves. With in- pines and at WARDA (West Africa Rice Development Asso-
creased iron toxicity stress, the entire affected leaves look pur- ciation) in Côte d’Ivoire have identified a number of rice cul-
plish brown, followed by drying of the leaves, which gives the tivars for growing in iron-toxic soils (De Datta et al 1994,
rice plant a scorched appearance. Equally important, the roots Sahrawat 2004).
of rice plants affected by iron toxicity become scanty, coarse, Sahrawat et al (1996) evaluated 20 lowland rice culti-
short and blunted, and dark brown in color; with the allevia- vars for tolerance of iron toxicity at an iron-toxic site in
tion of the stress, the roots may slowly recover to the usual Korhogo, Côte d’Ivoire, under irrigated conditions. The culti-
white color. Iron toxicity symptoms on rice leaves and changes vars differed in tolerance of iron toxicity. Grain yields varied
in root color and morphology are useful for diagnosis of the from 0.10 to 5.04 t ha–1 and iron toxicity scores, based on the
stress. Toxicity symptoms commonly develop at the maximum extent of bronzing symptoms on foliage, ranged from 2 to 9 (1
tillering and heading growth stage, but may be observed at any indicates normal growth and 9 indicates that most plants are
growth stage of the rice crop. dead or dying). Further evaluation of rice cultivars during 1992-
Since the first report of its occurrence (Ponnamperuma 97 showed that, among three promising iron-tolerant cultivars,
et al 1955), iron toxicity in rice has been reported in several CK 4 was the top yielder (mean grain yield 5.33 t ha–1), fol-
countries in Asia, South America, and West and Central Africa lowed by WITA 1 (4.96 t ha–1) and WITA 3 (4.46 t ha–1), and
(van Breemen and Moormann 1978, Yoshida 1981, De Datta tolerant check Suakoko 8 (3.80 t ha–1) (Table 1). These and
et al 1994, Sahrawat 2004). other results suggest that high rice yields and iron toxicity tol-
aEach season, all cultivars received a uniform application of 100 kg N ha–1, 50 kg P ha–1,
and 10 kg Zn ha–1.
Source: Sahrawat et al (2000).
erance are physiologically compatible (Sahrawat et al 2000, Table 2. Effects of field applications of nutrients on grain yield of
Audebert and Sahrawat 2000). iron-tolerant (CK 4) and susceptible (Bouake 189 and TOX 3069-
66-2-1-6) rice cultivars on an iron-toxic soil at Korhogo, Côte d’Ivoire
Work done at WARDA in West Africa showed that some (1995-98).a
Oryza glaberrima cultivars, adapted to lowland rice-growing
conditions, possess a higher tolerance for iron toxicity than Grain yield (t ha–1)
their O. sativa counterparts. Sahrawat and Sika (2002) con- Treatment
ducted experiments at an iron-toxic site (Korhogo, Côte CK 4 Bouake 189 TOX 3069-66-2-1-6
d’Ivoire) during the 2000 wet and dry seasons to evaluate the No fertilizer 4.3 (3)b 3.4 (5) 2.9 (7)
performance of promising O. sativa (CK 4, tolerant check; N 4.4 (3) 4.1 (5) 3.3 (7)
Bouake 189, susceptible check) and O. glaberrima (CG 14) N+P 5.3 (2) 4.3 (4) 4.2 (5)
cultivars. While CK 4 and Bouake 189 showed typical iron N+K 4.8 (2) 4.4 (4) 3.8 (5)
toxicity symptoms in varying degrees, CG 14 plants did not N + Zn 4.8 (2) 4.6 (4) 4.6 (5)
N + P + Zn 5.0 (2) 4.4 (4) 4.2 (4)
show any iron toxicity symptoms at all as measured by iron N + K + Zn 5.2 (2) 4.6 (3) 4.6 (4)
toxicity scores. Although CG 14 did not give high grain yields N+P+K 5.4 (2) 4.5 (3) 4.5 (3)
because of its lower harvest index, lodging of the crop, espe- N + P + K + Zn 5.7 (2) 4.7 (3) 4.7 (3)
cially under the application of nutrients, and shattering of seeds LSD (0.05) 1.01 1.02 1.15
at maturity, it showed remarkable tolerance for iron toxicity. aThe data are an average of four years (1995-98). All cultivars received a uniform
Research shows that CG 14 has a high tolerance for iron toxic- application of N (100 kg ha–1), P (50 kg ha–1), K (80 kg ha–1), and Zn (10 kg ha–1).
ity and remains an obvious choice as a donor for iron toler- bIron toxicity scores are given in parentheses on a scale of 1 to 9, where 1 = normal
ance in breeding programs (Sahrawat and Sika 2002, Sahrawat growth and 9 = most plants are dead or drying.
Source: Sahrawat et al (2001).
2004).
Upland rice is grown on a total of 19.1 million ha in Asia (60%), this degree of acidity occurred only in the surface soil layer in
Latin America (30%), and Africa (10%). The humid to the middle stage of rice development. This enhanced acidity
subhumid climatic conditions prevailing in upland rice areas was created by the sequential split application of urea, which
have typically led to various degrees of soil acidity because of reduced soil pH because of nitrification and by potassium fer-
deep weathering and leaching of cations. In fact, the major tilizers that exchange Al at the exchange site of clay minerals
soils for upland rice are Ultisols and Alfisols in Asia and West and increase Al concentration in soil solution in the moder-
Africa, and Oxisols and Ultisols in Latin America. Oxisols and ately intensive upland rice cultivation system (Okada and
Ultisols are especially typical acid soils. Fischer 2001).
Soil constraints are less critical for irrigated rice because Rice plants are moderately tolerant of acid soils in gen-
flooded conditions increase the availability of nutrients and eral, but showed wide genotypic variation in tolerance of this
stabilize soil pH closer to neutral, even in acid soils. On up- severe soil acidity. Semidwarf indicas were susceptible but
land soils, however, soil acidity can cause yield losses of up to tropical/temperate japonicas were usually tolerant. The causal
50% (Sarkarung 1986). It is therefore necessary to manage factor inducing genotypic differences was investigated under
soil acidity and related soil constraints to increase the produc- these conditions and it was found to be low Ca availability
tivity of upland rice. rather than high Al (Okada and Fischer 2001). The relative
growth of susceptible varieties was more correlated to ex-
changeable Ca than to exchangeable Al and Al saturation (Fig.
Growth-limiting factors for upland rice in acid soils 1). The tolerant genotype had a higher Ca absorbing capacity
The problems of acid soils are complex and are often regarded of the roots, probably owing to higher preferential adsorption
as an “acid-soil syndrome.” The major growth-limiting factor of Ca over Al at the surface of the root cell wall (Okada et al
for upland rice differs depending on the degree of soil acidity. 2003).
Roughly speaking, where soil pH (in water) is lower than When soil pH is higher than 4.3, the concentration of Al
4.3, the Al concentration in the soil solution can be higher than in soil solution is not toxic to the elongation of rice roots. This
100 µM. This is potentially toxic, with negative effects on the is the case for the commonly found acidic soils in West Africa
elongation of rice roots under solution culture conditions. In (Ultisols and some Alfisols), where the major chemical soil
upland rice fields, such soil conditions are not the norm but constraint is low P content/availability rather than soil acidity
can be induced by inappropriate management practices. For itself.
example, in Oxisols in the subhumid savannas of Colombia,
100
90
80
70
4.4 4.6 4.8 1.5 2.0 2.5 3.0 0.2 0.3 0.4 75 80 85
pH Exch-Al (cmol kg1) Exch-Ca (cmol kg1) Al sat. (%)
Fig. 1. The response of susceptible varieties of upland rice ( Oryzica 1; Oryzica Llanos 5) to soil pH
(A), exchangeable Al (B), exchangeable Ca (C), and Al saturation (D) in the topsoil of savanna (0 to 20
cm). Results of 1993 to 1996 at two sites in savannas of Colombia.
Genetic improvement of upland rice ment of rice cultivars capable of using a higher portion of the
for tolerance of soil acidity fixed P already present in soils may be an attractive and cost-
effective approach to increasing rice yields where P deficiency
Strongly acid soils are most predominant in South America. is the major constraint.
Besides the Amazonian rainforest, subhumid tropical savan- One promising step toward developing more P-efficient
nas cover 240 million ha (Cerrados in Brazil and Llanos in cultivars was the identification of the Pup1 locus in a QTL
Colombia and Venezuela). Dominant soils are the highly weath- mapping population that was derived from rice cultivars
ered Oxisols and Ultisols. A typical example of the chemical Nipponbare (low P uptake) and Kasalath (high P uptake) and
characteristics are pH 4.8 (1:1 water), organic matter content that had been evaluated on a highly P-fixing volcanic ash soil
4.2%, available P (Bray-2) 2.5 ppm, and exchangeable cat- (Wissuwa et al 2002). This locus was transferred from the do-
ions at Ca 0.18, Mg 0.13, K 0.09, and Al 2.1 cmol kg–1 soil. nor variety Kasalath into a Nipponbare genetic background.
The aluminum saturation of the soil is 84%. Thus, these soils The resulting near-isogenic line, NIL-Pup1, was genetically
are characterized by low pH and low effective CEC (sum of 92% identical to Nipponbare, but, when grown on the highly
exchangeable cations), with most of the CEC occupied by ex- P-fixing soil, NIL-Pup1 had a threefold higher biomass than
changeable Al. Nipponbare (Fig. 2). This improvement in growth was due to
To develop rice varieties with high yield potential and the ability of the NIL to acquire more P from the soil. Higher
acid-soil tolerance for the purpose of introducing upland rice root growth rates and more efficient P uptake per unit root size
in the rice-pasture cropping system of the savanna region was (uptake efficiency) were the main differences between NIL-
one objective at CIAT/Colombia. More than 1,360 cultivars Pup1 and Nipponbare (Fig. 2). These effects were detected
of diverse genetic background were evaluated for acid-soil tol- only under P deficiency but not when additional P (50 kg ha–1)
erance in the typical Oxisols (with 84–86%Al saturation) of was supplied to the deficient soil. This suggested that Pup1
the Colombian savannas (Sarkarung 1986). The acid-tolerant did not affect root growth potential per se but that it helped
accessions were crossed with blast-resistant lines, further evalu- maintain root growth when P was limiting. Model simulations
ated, and finally a commercial cultivar, Oryzica Sabana 6, was have shown that an increase in P uptake efficiency in the range
released in Colombia in 1991. This variety frequently achieved observed for NIL-Pup1 would be sufficient to explain subse-
outstanding yield (3,500–4,000 kg ha–1) under upland acid quent increases in P uptake and plant growth, including root
savanna conditions, and has desired characteristics such as long growth (Wissuwa 2003). Higher P-uptake efficiency is there-
grain and high amylose content. In 1996, a succeeding variety, fore the most likely mechanism affected by the Pup1 locus.
Oryzica Sabana 10, was developed. Currently, efforts at IRRI are directed toward identify-
ing the gene(s) at the Pup1 locus.
Further fine-mapping of Pup1 has advanced consider-
Marker-assisted breeding for low-P tolerant varieties ably and Pup1 has now been mapped to a 240-kb interval con-
As mentioned above, phosphorus (P) deficiency is a major taining 31 putative genes (Wissuwa, unpublished data). Most
abiotic stress that limits rice productivity, particularly under of these are hypothetical genes without similarity to known
upland conditions in acid soils such as Ultisols and Alfisols genes and those few that were functionally annotated were not
(Kirk et al 1998). In addition to areas of low absolute soil-P associated with P metabolism, root growth, or P-uptake mecha-
content, P deficiency can also arise in soils where P is strongly nisms (based on the Nipponbare sequence data). This suggests
bound to soil particles. Dobermann et al (1998) estimated that that the Pup1 locus either represents a novel gene or that the
more than 90% of added fertilizer P may be rapidly transformed gene is simply absent or highly distorted in Nipponbare. It
to P forms that are not easily available to plants. The develop- could also indicate, however, that the Pup1 locus performs a
Fig. 2. (Left) Effect of the Pup1 locus on plant growth on a highly P-deficient, highly P-fixing soil
(Andosol). NIL-Pup1 (+Pup1) has about three times as much P uptake, root dry weight, and total
biomass as Nipponbare (–Pup1). (Right) Effect of the Pup1 locus on P uptake, root weight, and
P-uptake efficiency. P-uptake efficiency is estimated as mg P taken up per g of root weight.
regulatory function, only indirectly affecting the uptake mecha- Sarkarung S. 1986. Screening upland rice for aluminum tolerance
nism. and blast resistance. In: Progress in upland rice research. Pro-
ceedings of the Second International Upland Rice Conference,
4-8 March 1985. Manila (Philippines): International Rice
References Research Institute. p 271-281.
Wissuwa M. 2003. How do plants achieve tolerance to phosphorus
Dobermann A, Cassman KG, Mamaril CP, Sheehy JE. 1998. Man-
deficiency: small causes with big effects. Plant Physiol.
agement of phosphorus, potassium, and sulfur in intensive,
133:1947-1958.
irrigated lowland rice. Field Crops Res. 56:113-138.
Wissuwa M, Wegner J, Ae N, Yano M. 2002. Substitution mapping
Kirk GJD, George T, Courtois B, Senadhira D. 1998. Opportunities
of Pup1: a major QTL increasing phosphorus uptake of rice
to improve phosphorus efficiency and soil fertility in rainfed
from a phosphorus-deficient soil. Theor. Appl. Genet. 105:890-
lowland and upland rice ecosystems. Field Crops Res. 56:73-
897.
92.
Okada K, Fischer AJ. 2001. Adaptation mechanisms of upland rice
genotypes to highly weathered acid soils of South American Notes
savannas. In: Ae N et al, editors. Plant nutrient acquisition:
new perspectives. Tokyo (Japan): Springer-Verlag Tokyo. Authors’ addresses: K. Okada, National Agricultural Research Cen-
p 185-200. ter, 3-1-1 Kannondai, Tsukuba, Ibaraki 305-8666, Japan; M.
Okada K, Fischer AJ, Perez-Salasar FA, Cañon-Romero Y. 2003. Wissuwa, International Rice Research Institute, Los Baños,
Difference in the retention of Ca and Al as possible mecha- Philippines.
nisms of Al resistance in upland rice. Soil Sci. Plant Nutr.
49:889-895.
Plant sciences that address the different physiological, genetic, age integrated research in plant sciences. Henceforth, while
and environmental aspects of crop yield are becoming more the problem of crop production under environmental stress is
divergent as time progresses. This is a result of the explosion a problem of the whole organism and the whole cropping sys-
of knowledge in each discipline on the one hand and the hu- tems in relation to their environment, science is becoming less
man limitation in coping with such an amount of knowledge. capable of dealing with this whole while it becomes proficient
As expertise becomes more specific, the chances for interac- in dealing with increasingly smaller components of the whole.
tion between scientific disciplines are diminishing. This is ac- This is reflected in the young upcoming scientists just out of
centuated by funding agencies that do not consistently encour- universities who enter the field of abiotic stress resistance with
Pollen formation is one of the most important processes in ing temperature. Although the microscopic size of the rice an-
plant reproduction and is widely known to be very sensitive to ther has been a factor limiting progress in its biochemical and
various forms of environmental stress, including low tempera- physiological analyses, recent advances in analyses of the rice
ture. The most sensitive stage of pollen was shown to be the genome have allowed us to use new approaches to solve the
young microspore stage that was demonstrated using pot-grown problem, such as small material size. For example, changes in
rice under phytotron conditions (Satake and Nishiyama 1970). the expression of transcripts in rice anthers caused by low tem-
Male sterility induced by chilling temperature (12 to 18 °C) at perature could be monitored using cDNA microarray analysis
the booting stage causes a serious loss of grain yield in rice satisfactorily (Yamaguchi et al 2004). Our research objective
and this chilling injury has been one of the most serious prob- is to discover the genes that have a key role in chilling injury
lems for agriculture in the temperate region (Yoshida 1981). by proteome analysis. In this study, we screened proteins sepa-
To solve this problem, it is important to understand the physi- rated from normal and chilled anthers, and the selected glyco-
ological nature of anther development under normal and chill- protein, which was specifically down-regulated by chilling, for
Common wild rice as well as indica-type rice are distributed cultivated strains and to evaluate their genetic bases by QTL
in tropical and subtropical areas, while japonica type rice is analysis.
grown from tropical to temperate areas. Tolerances of chilling
or low temperature are considered to play a role in the wide
distribution of cultivated rice toward the northern region. The
Materials and methods
response to chilling temperature (nonfreezing temperature) is We examined responses to chilling temperature at the three
a complex phenomenon that includes primary and secondary stages of early development (germination, plumule, and seed-
injury by exposure to low temperature, depending on the stage ling stages), using 40 cultivated and 21 wild strains. These
of development (Kratsch and Wise 2000, Allen and Ort 2001). were selected from diverse origins distributed from tropical
The genetic variation in responses to low temperature has been areas to Hokkaido (43oN), Japan. Hokkaido is one of the north-
reported in cultivated rice to be associated with its taxonomic ernmost areas of rice cultivation. Chilling tolerance was evalu-
groups as well as geographic distribution (Oka 1958, Kotaka, ated as follows:
and Abe 1988, Nagamine and Nakagahra 1990). Among vari- 1. Germinability under chilling temperature was tested
ous responses to low temperature, acclimation has not yet been with hulled seeds placed on moistened filter papers
evaluated substantially in rice. Acclimation, which is a response in petri dishes which were incubated in the dark at 13
to cold temperature by switching to a more cold-tolerant physi- °C. The number of germinated seeds was counted
ological state, is known to play a role in freezing tolerance in daily until 20 d after sowing. The germination coeffi-
various plant species, including wheat and barley (Thomashow cient (percentage of germinated seeds/mean days to
1990). The objectives of this study are to investigate genetic germination) was used for comparison.
complexities in response to low temperature among wild and 2. Chilling tolerance at the plumule stage was tested by
incubating germinated seeds with plumule length of
30 30
20 20
10 10
r = 0.44** r = 0.50**
0 0
10 10
0 1 2 3 4 0 1 2 3 4
40 40
30 30
20 20
10 10
r = 0.45** r = 0.70**
0 0
10 10
0 1 2 3 4 0 1 2 3 4
Tolerance Tolerance
Fig. 1. The relations between latitudinal locations and chilling tolerance at the plumule stage (A) or the
ability of acclimation (B) among 40 cultivated and 21 wild rice strains. ** indicates significance at the 1%
level.
erances except between the tolerances of seedlings with and analysis. In the northernmost areas of cultivation, it is consid-
without acclimation, suggesting that their genetic bases are ered that rice plants have been improved by adaptive changes
largely independent. Based on QTL analysis, the three signifi- in response to low temperature. However, our study revealed
cant QTLs for plumule tolerance were detected on chromo- that no major QTLs are involved in the tolerances examined.
some 1 (LOD = 3.9, near the centromere), chromosome 11 These results demonstrate that the responses to chilling tem-
(LOD = 6.0, long arm), and chromosome 12 (LOD = 6.8, short perature are rather complex traits and that the combination of
arm). For the other three tolerances, however, only lower LOD- various genes acting on different stages of development or
peaks (suggestive QTLs) were observed on chromosomes 5, under different temperature regimes might be responsible for
11, and 12. their local adaptation.
Furthermore, the three chilling tolerances, excluding ger-
minability, were significantly correlated with genomic simi-
larity index (GSI). GSI was computed as a mean value of 109
References
codominant PCR-based markers by assigning the genotypes Allen DJ, Ort DR. 2001. Impacts of chilling temperatures on photo-
as 1 (A58), –1 (W107), and 0 (heterozygote), expressing the synthesis in warm-climate plants. Trends Plant Sci. 6:36-42.
genomic likeness of an RIL to the parental strains. The results Kratsch HA, Wise RR. 2000. The ultrastructure of chilling stress.
showed that chilling tolerance might be controlled by multiple Plant Cell Environ. 23:337-350.
QTLs distributed across the genome as supposed from the QTL
Lateral roots constitute a great proportion of the root system diam) fixed on the side wall of the container at 1-cm intervals,
of plants in quantity. Two distinct types of lateral roots have and germinated seeds of each of the DHLs were sown between
been identified in rice: the L type, which is generally long and the two neighboring sheets of filter paper so that the grown
thick, and capable of branching into higher-order laterals; and roots would be sandwiched between them. When harvesting,
the S type, which is short and nonbranching but normally nu- it was easy to remove the root samples from the filter papers
merous (Yamauchi et al 1996). These two lateral roots differ without any loss or damage. Water stress was induced by us-
in vascular structure in such a way that the vascular system of ing PEG-6000. Preliminary observation showed that, in the
the S type is much less developed (Kono et al 1972). The abil- range of 0 to 120 g L–1, the concentration of 80 g L–1 resulted
ity of a genotype to alter its phenotypic expression in response in the most contrasted root response versus 0 g L–1 (control),
to the environment is termed “phenotypic plasticity” (Bradshaw and thus the concentration was set at 0 and 80 g L –1, repre-
1965). Yamauchi et al (1996) and Bañoc et al (2000) suggested senting well-watered and water-stress conditions, respectively.
that plastic development of lateral roots under water-deficit In the harvested root systems, seminal, nodal, and lat-
conditions may be one of the key traits of some genotypes for eral roots were identified. Lateral roots were classified further
the adaptation to such stress. This study aimed to identify QTLs into three types: L-type laterals, those that were long and thick,
for the plasticity of lateral root development in rice plants in and branched into a higher order; M types, those that were
response to water stress using a population of doubled-hap- long and thick but without a branch yet; and S types, those that
loid lines (DHLs). We also attempted to test a hypothesis that were short, slender, and nonbranching. The length of seminal,
developmental plasticity in response to water stress may differ nodal, and L- and M-type lateral roots, number of nodal roots
among different types of lateral roots. per plant, and number of the three types of lateral roots per
unit length of seminal root axis were measured. Plasticity of a
given trait was evaluated as the difference between mean val-
Materials and methods ues of the trait in stressed and control plants for each DHL.
Ninety-six DHLs were derived from a cross between CT9993- The marker map with 315 markers, including 145 re-
5-10-1-M (japonica, upland adapted) and IR62266-42-6-2 (in- striction fragment length polymorphism, 153 amplified frag-
dica, lowland adapted) (henceforth abbreviated as CT9993 and ment length polymorphism, and 17 microsatellite markers, has
IR62266, respectively) (Kamoshita et al 2000). been constructed at Texas Tech University, Lubbock, Texas
Plants were grown hydroponically using a modified (Zhang et al 2001). Total map distance is 1,788 cM. Putative
growth pouch method in an environmentally controlled growth QTLs (main-effect QTLs assuming no epistasis) for the traits
chamber for 2 weeks, where the day/night temperature was set were identified based on QTL Mapper (version 1.0) (Wang et
at 30/25 ºC, relative humidity at 70%, and photoperiod at 16 al 1999a,b). Main-effect QTLs were declared significant at
hours. A plastic container (25 cm in length, 19 cm in width, 27 the threshold of 0.005 and with the log of odds (LOD) value
cm in depth) was filled with water of different PEG concentra- set higher than 2.0.
tions as described below. Filter papers (18.0 cm in width and
26.0 cm in height) were hung from plastic pipes (0.8 cm in
Shoot height (cm) 10.6 10.2 8.6 6.8 12.0 6.0–21.0 9.0 3.0–15.8
Seminal root length (cm) 20.2 8.6 23.8 24.7 26.1 6.7–53.0 19.6 5.7–40.5
Nodal root number 6.0 7.3 4.0 4.8 6.4 2.0–11.0 5.2 1.0–10.0
Nodal root length (cm) 23.3 48.2 40.8 40.9 62.7 13.6–167.1 40.8 3.0–94.8
L-type lateral root numbera 2.0 1.8 0.1 0.6 0.7 0–2.7 0.7 0–3.2
M-type lateral root numbera 1.2 0.6 0 0.6 0.5 0–2.2 0.6 0–2.7
S-type lateral root numbera 10.2 8.7 9.9 9.1 9.4 2.8–19.6 9.6 1.6–23.0
L-type lateral root lengthb 39.3 17.0 2.7 3.0 17.0 0–79.4 16.5 0–74.6
M-type lateral root lengthb 21.3 1.7 0 2.0 7.3 0–40.7 5.5 0–39.3
aOn the basis of cm–1 seminal root axis. bOn the basis of 0–10-cm seminal root axis base from shoot.
Wrap-up of Session 15
The session on abiotic stresses, formally titled “Challenges to land ecology. He emphasized the importance of interactions be-
expanding rice production in unfavorable environments,” was tween genotype and environment in crop performance under
convened by S. Tobita (JIRCAS) and R. Lafitte (IRRI). The conven- drought stress. Characterization of the type of drought at a proper
ers recognized that the session should deal with the linkages site is sometimes not regarded as important but it does affect
across different disciplines and across different levels of institu- the selection criteria for better drought tolerance, he explained.
tions and organizations for the challenge, as well as understand- K. Futakuchi, a Japanese scientist currently working for the Af-
ing the kinds of abiotic stresses in rice ecologies. So, we invited rica Rice Center, discussed his research on submergence toler-
five speakers with a background in rice physiology or soil sci- ance in rice, including African rice (Oryza glaberrima) and Nericas
ence, with each having specialized experience in salinity, drought, (interspecific hybrids of Asian and African rice). His achievements
submergence, iron toxicity, and soil acidity, respectively. The covered the physiological elucidation of responses to submer-
molecular biology of stress responses in the rice crop were to be gence stress and genotypic variation in submergence tolerance.
reviewed, taking into account some of the recent progress in this The audience understood that O. glaberrima rice would be a good
area. It was also good to have a lecture on the principles of genetic resource for achieving rice with higher tolerance of sub-
stress physiology; herein the uniqueness of the rice crop would mergence stress in this region through wide hybridization.
automatically stand out. Because the Crop Science Society of K.L. Sahrawat, working for the International Crops Research
Japan concurrently organized a session on cold/heat tolerance in Institute for the Semi-Arid Tropics (ICRISAT) in India, discussed
rice, we did not examine temperature stress in the session, but another problem also prevalent in West Africa, iron toxicity, espe-
called for poster presentations on this subject. cially in the irrigated and rainfed lowland ecologies. With his back-
Our session began with more than 100 participants. Be- ground in soil chemistry, he explained a detailed process of chemi-
cause R. Lafitte could not attend the symposium, the session cal development of toxic iron ions in the soil. He also showed us
was chaired by G.V. Subbarao (JIRCAS) and me. I briefly explained the importance of nutrient management and the use of varieties
the aims and outlined the structure of the session, followed by tolerant of iron toxicity.
the first speaker, Nguyen Thi Lang from the Cuu Long Delta Rice K. Okada (National Agriculture Research Organization, Ja-
Research Institute (Vietnam). She discussed tissue culture and pan), in a paper co-authored by M. Wissuwa (IRRI), spoke about
in vitro approaches for selecting salt-tolerant somaclones in rice, soil acidity problems in tropical upland areas. An investigation of
for which successful examples were reported. The second lec- typical Oxisols in South America suggested that the real cause of
turer was Prof. S. Fukai of the University of Queensland, Australia the genotypic difference did not result from tolerance of toxic Al
(whose paper is co-authored by A. Kamoshita, the University of but from higher Ca absorbing capacity of the roots, probably be-
Tokyo), who talked about experiences with drought studies at cause of the higher preferential adsorption of Ca over Al at the
several field sites in Southeast Asia, mostly in the rainfed low- cell wall. He also mentioned recent progress in future applica-
Kiritani (1975) stated that the central issue for agriculture in IPM perspective in relation to biodiversity
the future would be how to manage and optimize biodiversity,
stability, and productivity within agroecosystems. The paddy In the past, most studies on paddy ecosystems have focused
ecosystem is an integrated, water-dependent system, which can on productivity and its stability in terms of rice yields.
contain many kinds of living organisms: birds, fish, reptiles, Arthropods in paddy ecosystems can be classified into three
amphibia, arthropods, and plants. Paddy fields were originally main groups according to their ecological requirements: (1)
wetlands that are artificially constructed devices for rice pro- resident species adapted to the continuous cropping of rice in
duction. Nowadays, very few natural wetlands remain, and the same field, (2) migratory species adapted to exploit rice as
many aquatic organisms now depend partly or fully on paddy an annual crop, and (3) aquatic species originating from still-
fields. water habitats in wetlands. Concerning groups 1 and 2, inte-
To raise both land and labor productivity, the Japanese grated pest management (IPM) programs, which have a pri-
government has promoted the conversion of poorly drained mary objective of maximizing economic profit on the farm,
wet paddy fields into well-drained ones in association with a have been implemented with various degrees of success. Al-
policy to consolidate fragmented farmlands. U-shaped con- though IPM is becoming widespread, those insects (Tada-no-
crete ditches have replaced traditional earth ditches and irri- mushi = species of unknown or uncertain function that rou-
gation-supply canals have been separated from drainage ca- tinely occur in the habitat) that have no direct economic im-
nals, which effectively reduced the variety of habitats for pact on rice production have been mostly ignored as an impor-
aquatic organisms. tant element in the rice ecosystem. Consequently, some aquatic
insects are in danger of extinction, thus requiring conserva-
tion (Kiritani 1979, 2000, Kiritani and Naba 1994).
Biodiversity of arthropod fauna in paddy fields
The green rice leafhopper, Nephotettix cincticeps, is 80%
Infestations by rice borers, Chilo suppressalis and Scirpophaga of the diet of a lycosid spider, Pardosa pseudoannulata, in
incertulas, and the intermittent and sudden occurrences of out- paddy fields. No lycosid spiders, however, developed to adults
breaks of Nilaparvata lugens were the major causes of losses when fed only N. cincticeps. When lycosid females were al-
in rice yield in temperate Asia. These were the major pests lowed to prey upon mixed species of prey, their fecundity
until around 1965. Thereafter, leaf- and planthoppers and the greatly increased (Suzuki and Kiritani 1974). Those species
viral diseases RSLV and RDV transmitted by them were pre- such as chironomids and collembola, for example, that are
dominant for about 30 years. Meanwhile, the rice water wee- neither pests nor natural enemies, and yet are useful as alter-
vil, Lissorhoptrus oryzophilus, invaded Japan in 1976 from native food of generalist predators, can be referred to as mi-
the United States, inflicting serious damage to rice in the late nor, yet important, components of the community (Kiritani
1980s. Since 1995, the damage caused by various kinds of 2000).
stink bugs and mirids has become the most serious problem. Immigration of spiders to paddy fields occurs after the
Many species of arthropods with diverse types of life appearance of chironomids. Early insecticide applications to
cycles occupy different habitats within the paddy control rice stem borers often result in the resurgence of
agroecosystem. Sympetrum dragonflies emerge from paddy planthoppers and leafhoppers 1 month later because insecti-
fields and stay in coppiced woodlots to mature sexually be- cide treatments simultaneously kill spiders and chironomids
fore returning to paddy fields to oviposit. The eggs hatch in (Kobayashi 1961). In the tropics, prevention of outbreaks of
the following spring when irrigation water becomes available. planthoppers and leafhoppers depends on protection of early-
Newly emerged adults of the water scorpion, Ranatra chinensis, acting natural enemies by avoiding early insecticide spraying
move from paddy fields to irrigation ponds for overwintering. (Way and Heong 1994, Settle et al 1996).
Oviposition takes place in paddy fields in the next spring. The Levees are likely to act as refuges for various kinds of
migratory planthopper pests, N. lugens and Sogatella furcifera, natural enemies of arthropod pests that occur in upland crops
are annually replenished by a long-range immigration from grown close to paddy fields. A dwarf spider, Ummeliata
tropical endemic habitats. insecticeps, common in paddy fields dispersed from levees by
The biodiversity of the paddy agroecosystem therefore ballooning in late May to uplands remains there until the end
depends not only on the paddy fields themselves but also on of the rainy season. It behaved like a specific predator attack-
water channels, irrigation ponds, levees, surrounding fallow ing a newly hatched colony of larvae of Spodoptera litura in
fields, neighboring farmlands, secondary forests, wetlands, riv- taro fields (Nakasuji et al 1973). Another example is the
ers, and remote hibernating areas (Kiritani 2000). anthocorid bugs, Orius spp., that are effective natural enemies
of Thrips palmi, a serious invasive alien pest of eggplant. O.
IPM Conservation
IPM
Density
50
IBM
Conservation
Conservation Extinction threshold
Rice planthoppers, Nilaparvata lugens Stål and Sogatella emies caused by an unstable rice habitat (made unstable by
furcifera Horvath, are major insect pests in the Yangtze River harvesting, pesticide application, and other practices) and win-
Delta rice ecosystem. About 50–100 million tons of rice were ter interruption (Way and Heong 1994). Therefore, neighbor-
lost per year during 1973-2000 in China because of the inci- ing habitats are very important in conserving and enhancing
dence of rice planthoppers (Yu et al 1999). Usually, the injudi- the population of natural enemies in rice fields (Yu et al 1998).
cious use of insecticides in the rice ecosystem caused serious To evaluate the role of habitat diversity in promoting natural
problems, such as resurgence and resistance of insect pests, enemies in rice habitats, the movement of spiders and Anagrus
residue, and adverse effects on the environment. Predators and spp. between rice and neighboring habitats was monitored.
parasitoids were found to be an effective bio-agent to suppress Meanwhile, spraying experiments to determine the optimal
insect pests, but natural enemies per se usually could not effi- dosage of chemicals were conducted on a demonstration farm
ciently suppress insect pests below the damage threshold be- with various habitat manipulations.
cause of the low parasitism and predation rate of natural en-
Application Spiders Anagrus spp. Brown Leaffolder Stem borer Yield loss
No. of sprayings time (date) (no. per m2) (no. per m2) planthopper (no. per hill) (deadhead %) (%)
(no. per m2)
aThe transplanting date was 1 June. Data were analyzed with Duncan’s multiple range test between treatments. Means followed by a different letter in the same column are
out the season. The dispersal of these natural enemies from Way MJ, Heong KL. 1994. The role of biodiversity in the dynamics
neighboring habitats highly promoted their population in rice and management of insect pests of tropical irrigated rice: a
fields (Yu et al 2002, Zheng et al 1999). The field data showed review. Bull. Entomol. Res. 84:567-587.
that burning vegetation on field bunds or other uncultivated Xu JS, Chen ZF, Zhu RL. 1987. Taxonomy and application of spi-
ders in rice fields in Zhejiang province. Natural Enemy
habitats for sanitary purposes was destructive for spiders and
9(3):140-144.
egg parasitoids such as Anagrus spp. Mowing vegetation and
Yu XP, Hu C, Heong KL. 1998. Parasitization and preference char-
spraying herbicides did not significantly affect the population acteristics of egg parasitoids from various habitats to
buildup of spiders. homopterns. Acta Entomol. Sin. 40(4):431-437.
Favorable habitat manipulations based on this research Yu XP, Zheng XS, Chen JM, et al. 1999. A study on the relationship
were suggested as follows: (1) keeping a weedy bund or plant- between egg parasitoid, Anagrus nilaparvatae, and green slen-
ing soybean with long-term flowering for sheltering egg para- der planthopper, Saccharosydne procerus, a species of insect
sitoids, (2) not spraying within 30 days after rice transplant- pest of wild rice, Zizania caduciflora. Acta Entomol. Sin.
ing, (3) making a canal every 3–5 m to protect frogs, (4) re- 42(4):387-393.
constructing field bunds for spider refuge, and (5) not spray- Yu XP, Zheng XS, Xu HX, et al. 2002. A study on the dispersal of
ing insecticides in Zizania fields to protect egg parasitoids. lycosid spider, Pirata subpiraticus, between rice and Zizania
fields. Acta Entomol. Sin. 45(5):636-640.
These habitat manipulations were incorporated into integrated
Zheng XS, Yu XP, Chen JM, et al. 1999. Dispersal patterns of natu-
pest management, and 1–2 insecticide sprayings per season
ral enemies of rice planthoppers between Zizania and rice
were saved in rice fields. About 5,500 ha of rice fields using fields. Acta Agric. Zhejiang 11(6):339-343.
habitat manipulation have been demonstrated in Zhejiang Prov-
ince in the past 4 years. As a result, about 1–2 sprays, or a 20%
reduction in insecticide spraying, and a 20–25% increase in Notes
natural enemies were achieved in rice fields. No significant Authors’ address: Institute of Plant Protection, Zhejiang Academy
yield loss was recorded as a result of this modification of the of Agricultural Sciences, Hangzhou 310021, China, e-mail:
spraying program. luzx@mail.hz.zj.cn.
References
Berry W L, Stimmann MW, Wolf WW. 1972, Marking of native phy-
tophagous insects with rubidium: a proposed technique. Ann.
Entomol. Soc. Am. 65(1):236-238.
Perfect TJ, Cook AG, Padgham DE, et al. 1985, Interpretation of the
flight activity of Nilaparvata lungens Stål and Sogatella
furcifera Horváth (Hemiptera: Delphacidae) based on com-
parative trap catches and field marking with rubidium. Bull.
Entomol. Res. 75:93-106.
Augmentative releases of natural enemies are a widely used the release of C. lividipennis. In 2001, the effect of an addi-
method of controlling many pests, mainly in greenhouse veg- tional release of C. lividipennis at the start of the second gen-
etables (Van Driesche and Bellows 1996). In open rice fields, eration (early August) was also evaluated.
however, few attempts have been made to manage insect pests A routine population census for the two planthoppers
by releasing natural enemies. The mirid bug Cyrtorhinus (released N. lugens and naturally-occurring S. furcifera) and
lividipennis Reuter is the egg predator of the brown planthopper C. lividipennis was conducted using a FARMCOP suction sam-
Nilaparvata lugens (Stål) and the whitebacked planthopper pler (Cariño et al 1979). A plastic cylinder (30 cm in diameter
Sogatella furcifera (Horváth), the two major pests of rice and 70 cm in height) was used to cover a rice hill. All arthropods
throughout Asia (Chiu 1979). Although the potential growth inside the cylinder were collected with the sampler, stored in
rate of C. lividipennis is very high (Suzuki and Tanaka 1996, 70% alcohol, and counted in the laboratory under a binocular
Matsumura and Suzuki 1999), the initial population density of microscope. The number of rice plants sampled was 15 per
C. lividipennis in Japan is usually low (Teramoto et al 1996). experimental block per sampling date. Censuses were made at
This is because C. lividipennis is not able to overwinter suc- intervals of 5–10 days from late July to late September (in
cessfully in Japan, and colonization occurs annually following 1999) or early October (in 2000 and 2001).
long-distance migration from southern China. Thus, we evalu- Generation boundaries and mean population density of
ated the effectiveness of augmentative releases of C. the two planthoppers and C. lividipennis for each successive
lividipennis to suppress N. lugens in open paddy fields. generation were calculated based on Kuno’s (1968) method.
All analyses were conducted using a 3-year data set of mean
population densities. The effect of the natural enemy release
Materials and methods on population dynamics of N. lugens was determined by
Mass rearing of C. lividipennis Yamamura’s (1999) key-factor/key-stage analysis. All statisti-
As no mass-rearing technique for C. lividipennis using an arti- cal analyses were performed with the JMP version 5 Statisti-
ficial diet has been established, we developed a simple method cal Package.
for mass rearing of C. lividipennis on rice seedlings using N.
lugens eggs as food. To calculate the reproductive efficiency
Results and discussion
of C. lividipennis, 25 or 50 pairs of adult C. lividipennis were
introduced into plastic cages (30 × 25 × 28 cm) containing 5– Mass rearing of C. lividipennis
6-day-old rice seedlings infested with 50 gravid females of N. Five weeks after the introduction of C. lividipennis adults into
lugens. After 2 weeks, new rice seedlings were introduced into the plastic cage, 286.8 ± 15.7 and 436.4 ± 30.3 individual adult
the plastic cages and the old rice seedlings were removed after offspring were obtained from 25 and 50 C. lividipennis adults,
all the insects had moved to the new ones. Thereafter, the rice respectively. The population growth rate of C. lividipennis was
seedlings were renewed at 7-day intervals. After 5 weeks, the 5.7 and 4.4 per generation for the 25- and 50-adult treatments,
number of adult offspring was counted. The experiments were respectively.
conducted in the laboratory at a temperature of 25 ± 1 °C and The present method of mass rearing of C. lividipennis is
a photoperiod of 16L:8D. high enough for stock maintenance and use in release experi-
ments. For practical and commercial uses of C. lividipennis as
Augmentative release experiment a biological control agent, however, there is a need to estab-
In 1999-2001, C. lividipennis adults (1-week-old adults pro- lish a mass-rearing technique using an artificial diet.
duced in the laboratory) were augmentatively released at
predator:prey (C. lividipennis adult:N. lugens adult) ratios of Augmentative release experiment
0:1 (nonrelease control), 1:1, and 1:2 (only in 1999) into rep- The population density of the first-generation N. lugens (late
licated experimental blocks (7 × 7 m) in open paddy fields. July to early August) was significantly suppressed when C.
The release was done at the start of the immigrant generation lividipennis was released at predator:prey ratios of 1:1 (high-
of N. lugens (early July). Because the density of N. lugens ratio release) (ANOVA, F = 8.38, P<0.05) (Fig. 1). Yamamura’s
immigrants was quite low in those three years, N. lugens adults (1999) key-factor/key-stage analysis revealed that the effect
were also released at a rate of 0.2 individual per hill prior to of the natural enemy release at the start of the first generation
Rep. 1
Rep. 1 Rep. 1
0.0
3.0
2.0
1.0
Table 1. Key-factor/key-stage table for the population density of the third generation in
Nilaparvata lugens.
Generation
Factor df F P
G1/G0 G2/G1 G3/G2 Total
was the most significant factor suppressing the density of the population density of naturally-occurring first-generation S.
third-generation N. lugens (Table 1). One-time release of C. furcifera (P<0.05, r2 = 0.645). This negative effect of S.
lividipennis was insufficient to suppress the density of third- furcifera on the population growth of N. lugens is the first
generation N. lugens (Fig. 1C). However, an additional release field evidence of interspecific interactions between N. lugens
of C. lividipennis at the second generation (early August) suc- and S. furcifera, which had been suggested in laboratory ex-
cessfully regulated the mean population density of the third- periments (Matsumura and Suzuki 2003).
generation N. lugens to 13.6–36.1 individuals per hill (1.13– The population density of C. lividipennis increased at
1.56 in log scale), which is lower than the density at which the next generation (early August) of natural enemy release,
“hopper-burn” damage occurs (Fig. 1C). but it decreased thereafter (data not shown). The population
Stepwise multiregression analysis revealed that the most density of C. lividipennis increased again in mid-September,
important factor decreasing the population growth rate from when the density of N. lugens increased (data not shown). The
the initial to the first generation of N. lugens was the density number of C. lividipennis was positively correlated with the
of released C. lividipennis (P<0.001, r2 = 0.628). In contrast, sum of the number of S. furcifera and N. lugens (P<0.001, r2 =
the most important factor decreasing the population growth 0.967). Because S. furcifera density peaks earlier in the sea-
rate from the first to the second generation of N. lugens was son than does N. lugens, the present results suggest that the
not the density of the first generation of C. lividipennis but the
The development and cultivation of shorter-duration and ag- nying this intensification is the use of genetically uniform va-
ronomically well-adapted varieties has fed half of the world’s rieties that reduce the buffering capacity in the rice-cropping
total population of about six billion people, mostly in Asia. system. This situation creates some concerns regarding mod-
This not only allowed intensification of the rice crop in time ern rice production in relation to pest outbreaks. Below, we
and space, which reduced the demand for land to grow it to outline some of these concerns and propose that scientists
meet this population pressure, but it also accounts for 30% to working on pest management should consider sustainable rice
50% of agricultural production and 50–90% of the calories production systems as a means to manage pests and diseases
consumed by these people (Hossain and Fischer 1995). While as an important option for another doubly green revolution in
there is a need to achieve greater rice productivity, intensifica- agriculture.
tion also makes the rice crop more vulnerable to attack by rice The first concern is the deployment of a few high-yield-
pests and diseases (Mew et al 2004). ing varieties over large areas, leading to a decline in cultivar
Pests and diseases are a moving target. Genetic elastic- diversity. The general landscape of rice production in inten-
ity allows them to readjust their population to adapt to new sive rice ecosystems is characterized as a monoculture sys-
environments and on new resistant varieties bred by plant tem. A reduction in cultivar diversity is a concern of rice inten-
breeders. It is not uncommon that the scenario of pest and dis- sification because maintaining adequate diversity and resil-
ease profiles shifts as production systems change. Accompa- ience is important in the humid tropics of Asia, where pressure
15
10
10
5
5
0
0 10
25 30 35 40 45 50 55
Sim. S:R (1:0) in 2003
Days from 1 June Sim. S:R (1:1) in 2003
Fig. 1. Observed (Obs.) and simulated (Sim.) development of rice 8 Sim. S:R (1:3) in 2003
leaf blast in a mixture of susceptible (S) cv. Sasanishiki and its Obs. S:R (1:0) in 2003
resistant (R) near-isogenic line, Sasanishiki BL No. 4, at propor- Obs. S:R (1:1) in 2003
tions of 1:0 and 1:1 from 2000 to 2001. The field experiments Obs. S:R (1:3) in 2003
were performed at the Furukawa Agricultural Experiment Station 6
in Miyagi Prefecture.
4
Results
Many traits of significance in agriculture exhibit quantitative defense for analysis of mapping populations for improving
inheritance. To characterize these quantitative traits, the can- resistance to rice blast.
didate gene approach has been applied in several cereal crops With the sequencing of many genomes in recent years
such as maize, wheat, and rice because of the detection of along with the genetic and bioinformatic resources for rice,
multiple genes defining a complex trait, their partial effects on our ability to comprehensively characterize multigene fami-
phenotypic variation, and their less precise localization on lies has also been greatly enhanced. These resources include
genetic maps in comparison with traits conferring qualitative the physical, linkage, and expressed sequence tag (EST) maps,
inheritance (Pflieger et al 2001, Wilson et al 2004). a collection of >130,000 ESTs, gene prediction algorithms and
The availability of the whole rice genome sequence from automatic genome annotation systems, and online resources
public and private sequencing efforts has provided opportuni- for comparative grass genomics. Through these resources, we
ties to predict putative functions of a gene based on sequence examined DR genes retrieved from the Rice Genome Program
information, thus allowing the identification of candidate genes. database using in silico analysis.
Candidate genes are DNA sequences similar to known genes
or conserved motifs that make it possible to infer their bio-
Identifying candidate genes by association
logical functions. Through their association with disease re-
with blast quantitative resistance
sistance, they become candidate defense-response (DR) genes
for conferring particular phenotypes. Using the candidate gene A japonica cultivar, Moroberekan from Africa, known to ex-
approach, we studied the associations between candidate gene hibit durable resistance to rice blast in Asia was crossed to a
markers with disease and insect resistance in a doubled-hap- popular upland cultivar, Vandana, following the advanced back-
loid rice mapping population (Ramalingam et al 2002). This cross QTL approach. This approach allows QTL analysis of
frame map has been a useful reference for selecting candidate the mapping population while simultaneously developing elite
genes involved in both pathogen recognition and general plant germplasm for commercial use (Tanksley and Nelson 1996).
Eighty BC3F3 and BC3F4 lines derived from advanced back-
rm473d 113.2
125 rm426 122.3 100 100
rm80 103.7
rm168 122.8
rm186 127.4
rm411 127.9
rm148 140.7 rm5720 115.2.6 rm281 119.9
0 1 3 5 7 9 rm248 118.0 1 3 5 7 9 1 3 5 7 9
2 4 6 8 2 4 6 8 2 4 6 8
100
rm426
osglp10up
rm473 113.2 osglp10
osglp11up
rm426 122.3 osglp11
osglp12
125rm168 122.8
osglp13up
rm186 127.4 osglp13
rm168 1 2 3 4 5 6 7 8 9
rm411 127.9
rm148 140.7
1 2 3 4 5 6 7 8 9
Fig. 1. (A) Genome scan of Vandana/Moroberekan intercross progenies using SSRs, 1: Moroberekan; 2: Vandana;
3–7: more resistant lines; 8–9: less resistant lines. Examples of candidate genes associated with each of the
critical areas marked: putative oxalate oxidases in chr. 3, putative aspartyl protease in chr. 7, and oxalate oxidase-
like proteins in chr. 8. Heterozygous loci are colored green. Because of selection in BC3F3 prior to intermating, our
susceptible lines in fact carry some degree of quantitative resistance. Therefore, Vandana is still the susceptible
control in this analysis. (B) Detailed marker analysis of chr. 3 loci identified with 4 putative oxalate oxidases on 1:
Moroberekan; 2: Vandana; 3–7: more resistant lines; 8–9: less resistant lines. PCR primers used were designed
from the coding region of each gene. Primers were also designed in the 1,000-bp upstream region of each gene
(OsGLP10UP, OsGLP11UP, and OsGLP13UP). Heterozygous loci are colored green. Monomorphic markers are col-
ored light blue.
These oxalate oxidase gene sequences are located next to each A scan of the rice genome revealed at least 20 thaumatin-like
other, with similarities ranging from 90% to 98%. For each proteins dispersed in all chromosomes except chromosome 5.
gene, there was variation in the copy number of cis-elements Of these, thaumatin-like sequences in chromosomes 6 and 7
related to biotic stress responses, such as W box, WNPR1, have been mapped to IR64/Azucena doubled-haploid lines
and WRKY, indicating that these genes have potential asso- (Ramalingam et al 2003). Analysis of the 1,000-bp upstream
ciations with the response of rice to pathogen infection, such of each gene reveals that the number of copies of cis-elements
as the blast fungus. such as WRKY, WBOX, and GCC core varies from one gene
Phylogenetic analyses clustered chromosome 3 oxalate to the next. PR5 proteins in barley were found to bind to
oxidases with barley oxalate oxidase Y14203. Several GLP β-1,3-glucans from yeasts and fungi (Trudel et al 1998). In
sequences in chromosome 8 were similar in conserved domain Vandana/Moroberekan, thaumatin was significantly associated
structure to barley oxalate oxidase-like protein X93171. These with partial blast resistance (Wu et al 2004).
genes are thought to play a role in defense response by degra-
dation of oxalate produced by fungi, the production of active
Future prospects
oxygen species for oxidative cross-linking of cell wall com-
ponents, and stress signaling by increasing extracellular The availability of rice genetic and bioinformatic resources
[Ca+ion] (Dunwell et al 2000). has clearly overcome the limitations for relating QTLs to can-
Another class of DR genes is the thaumatin-like proteins didate genes and metabolic pathways. To improve efficiency
or PR5 proteins. At IRRI, overexpression of PR5 delays the in the selection and pyramiding of desirable plant traits, marker-
onset of disease symptoms of sheath blight in transgenic rice. assisted selection may now allow candidate defense genes with
References Notes
Dunwell JM, Khuri S, Gane PJ. 2000. Microbial relatives of the Authors’ addresses: G. Carrillo, J. Wu, B. Liu, N. Sugiyama, I. Oña,
seed storage proteins of higher plants: conservation of struc- H. Leung, and C.M. Vera Cruz, IRRI, DAPO Box 7777, Metro
ture and diversification of function during evolution of the Manila, Philippines; M. Variar, Central Rainfed Upland Rice
cupin superfamily. Microbiol. Mol. Biol. Rev. 64:153-179. Research Station, Hazaribag 825 301, Jharkand, India; B.
Liu B, Zhang S, Zhu X, Yang Q, Wu S, Mei M, Mauleon R, Leach J, Courtois, CIRAD-CA, Avenue Agropolis, Montpellier Cedex
Mew T, Leung H. 2004. Candidate defense genes as predic- 5, France; J.E. Leach, Colorado State University, Fort Collins,
tors of quantitative blast resistance in rice. Mol. Plant-Mi- Colo. 80523-1177, USA; P.H. Goodwin, University of Guelph,
crobe Interact. 17:1146-1152. Guelph N1G 2W1, Canada.
Pflieger S, Lefebvre V, Causse M. 2001. The candidate gene ap-
proach in plant genetics: a review. Mol. Breed. 7:275-291.
Ramalingam J, Vera Cruz CM, Kukrejah K, Chittoor JM, Wu J, Lee
SW, Baraoidan M, George ML, Cohen MB, Hulbert S, Leach
JE, Leung H. 2002. Candidate resistance genes from rice,
barley, and maize and their association with qualitative and
quantitative resistance in rice. Mol. Plant-Microbe Interact.
16(1):14-24.
The transfer of genes into cereal crops from related grasses two types of resistance (Thordal-Christensen 2003) suggest
has been an important tool for improving the disease and pest that R genes probably play a role. One limitation of effective
resistance of some cereals. The approach to date has been lim- transfer of resistance genes is the fact that many R genes do
ited to cereals and their close relatives because conventional not function in the heterologous genetic backgrounds of dis-
genetic approaches have been used. If resistance genes could tantly related species (Hulbert et al 2001).
be effectively transferred among distantly related grasses in
different tribes (e.g., rice, maize, and wheat), this could have a
Transfer of a maize resistance gene to wheat and barley
large effect on resistance breeding. Of particular relevance is
the observation that a given cereal species is typically resis- The first transfer of a disease resistance gene among distantly
tant to most of the pathogens that cause disease in other cereal related cereals was the maize Rp1-D gene, which was trans-
species. If the basis of some of these nonhost resistances is ferred to wheat and barley (Ayliffe et al 2004). Rp1-D confers
due to the same types of genes that cause host resistance, we resistance to common rust (Puccinia sorghi) in maize, but no
may be able to identify and transfer these genes to transfer the rust resistance was observed in wheat or barley lines express-
resistance. The contribution of typical host resistance genes to ing the gene after inoculation with rust isolates from three dif-
nonhost resistance is not clear, but similarities between the ferent rust species. Although the Rp1-D transformation con-
The objectives of this study are to discriminate allelopathy ronment and makes sure human beings have a sustainable ex-
from other competition in rice by a specific bioassay (plant istence, is needed all over the world. In paddy fields, prob-
box method), and to check weed suppression activity in paddy lems of weed control, which had been tedious and laborious
fields. The final goal is weed control without herbicides, or work for a long time, have been successfully conquered by the
less use of herbicides. introduction of specific and powerful herbicides, and few prob-
The main target of breeding rice cultivars in Japan has lems remain except for some perennial weeds in Japan. We
been good taste and high productivity. Recently, cultivar thus need to be able to use allelopathy to control paddy weeds
Koshihikari, famous for its good taste, has been sweeping over without using much or any herbicide and to accomplish safe
all of Japanese rice cultivation except for Hokkaido, a north- and labor-saving rice cultivation. Our group therefore started
ern district of Japan, and a need remains for still more good research on allelopathy in rice. Our target is to search for po-
rice cultivars with good taste. Now, many other new candi- tent allelopathic rice varieties and to use these strains for weed
dates are being prepared for the market. On the other hand, control in paddy fields without herbicides or with little use of
sustainable agriculture, which aims at less impact on the envi- them.
20
Barnyardgrass
Konagi
15
10
Inuhatarui
5
0
Awa-Akamai Houman jinjya Midorimai Soujya Nebarimochi Nihonbare
Akamai Akamai
Fig. 1. Weed suppression by allelopathic rice in paddy fields.
For allelopathy of rice, little research has been done so plants were grown in sand culture in pots for 4 to 8 weeks.
far. Though effects of rice residues on the following crop were Roots were washed with tap water and then with deionized
reported (Chou 1986), little research has been done on the ef- water to remove sand. Roots of three rice plants were placed
fect of root exudates of rice. Our group has been engaged in gently into a nylon-mesh-bounded cylinder (3.2 cm diameter
work to search for allelopathic plants and to examine allelopa- by 6.6 cm tall), which was placed upright in one corner of a 70
thy and ascertain its mechanism. In this research, we have de- × 70-cm-wide by 100-cm-deep clear plastic box. Plant stems
veloped a new system that can prove the possibility of allel- were taped to the corner of the box to stabilize the plants in the
opathy, called the “plant box method” (Fujii 1999). This method box. Hot liquid agar (0.75%; 250 mL per box) was cooled to
is a sort of mixed planting using agar (no nutrient) as a me- 40 °C and poured into the boxes, covering all root material,
dium. By using this method, we can visualize allelopathy by but not covering stems. Boxes containing agar and rice plants
root exudates, and we can observe differences between variet- were cooled on ice until the agar had solidified (about 30 min).
ies, even in the same species. Thirty-one lettuce seeds were inserted partially into the sur-
Some farmers suggested from their experience that red face of the agar in an equidistant grid pattern. Seeds were about
rice of Japanese old-type rice varieties has few accompanying halfway into the agar with the germ (pointed) end down using
weeds. Some rice cultivars used before the introduction of porcelain forceps. Boxes were sealed around the rice stems
herbicides are possibly resistant to paddy weeds. And, some and over the box surface with plastic wrap to minimize water
rice cultivars or ancestors of rice growing in tropical countries loss, and placed in a fluorescent-lighted incubator for 5 days
where weed growth in paddy fields is serious may be resistant (25 °C, 14-h days; 20 °C, 10-h nights). Lettuce seedlings were
to weeds. So, we started screening allelopathic activity of rice then removed and root and shoot lengths determined to the
cultivars by the plant box method. Our work is quite prelimi- nearest mm. Regression analysis was used to estimate lettuce
nary, and our findings might not apply directly to actual weed root length at zero distance from rice roots (y-intercept), and
control. But, we hope that these approaches to the biological slope of the inhibition curve.
control of paddy fields will awaken Asians to weed manage- Experiments were conducted with rice using an agar dif-
ment using allelopathy from their original rice cultivars. fusion plant box method. Apparent allelopathic activity was
A total of 500 rice cultivars, varieties, and related plants observed in the second year of a 5-year field trial in trans-
were tested. They were supplied from the Japanese Gene Bank planted water-seeded rice. In this study, the red Shrine rice,
(MAFF, Tsukuba), the core collection by Dr. M. Okuno called Awa-Akamai, reduced total weed biomass by about 80–
(HNAES), the Malaysian Gene Bank (MARDI), the red rice 90% compared with the commercial Japanese cultivar
collection by Dr. T. Itani (Hiroshima Prefectural University), Nipponbare (Fig. 1). The mechanisms responsible for weed
and the wild and traditional rice collection in Bhutan by Dr. Y. suppression in Awa-Akamai may be valuable for improving
Sato (Shizuoka University). weed suppression in commercial rice cultivars. Suppression
The plant box method was followed according to Fujii of barnyardgrass and broadleaf paddy weeds has been shown
(1999). The receiver plant was lettuce; Great Lakes 366 is in the field in Tsukuba with Awa-Akamai and other red rice
highly susceptible to bioactive substances. The plant box cultivars.
method was used for assaying the activity of diffusible phyto- In plant box tests, roots of several rice lines, including
toxic metabolites (potentially allelopathic compounds) in agar Awa-Akamai, were highly inhibitory to the elongation of let-
in the laboratory and is summarized briefly as follows. Rice tuce roots. Nipponbare (a nonsuppressive cultivar in field tests
in Tsukuba) was usually less inhibitory to lettuce root elonga- breed new varieties that have allelopathic ability to control
tion than was Awa-Akamai. In plant box experiments in which weeds. This project may take another 10 years or more, and it
lettuce seed planting was delayed 2 or 3 days, allowing might be impossible to suppress paddy weeds completely by
allelochemicals to diffuse greater distances, apparent allelo- only allelopathy, but, in combination with other biological
pathic activity was greater following longer delays. control, allelopathic rice strains will help us to achieve sus-
The results of the effect of rice cultivars are shown in tainable and safe rice production.
Table 1, in which % means the inhibitory percentage of lettuce
radicle length in the root surface of rice by extrapolation of
References
the radicle length of each point apart from the root zone of
rice to the control. These data show the activity and quantity Azmi M, Abdullah MZ, Fujii Y.2000. Exploratory study on allelo-
of root exudates. Results of screening 500 rice varieties showed pathic effect of selected Malaysian rice varieties and rice field
a wide variation in activity. Some varieties showed no activity, weed species. J. Trop. Agric. Food Sci. 28(1):39-54.
some showed strong activity, and, in some cases, the top of the Baker P. 1992. Weed resistance hidden in rice genes. Rice J. 8-9
March.
lettuce radicle became brown to black. The strongest activity
Chou C-H. 1986. The role of allelopathy in subtropical
was of Awa-Akamai, an old red rice reserved in the old shrine
agroecosystems in Taiwan. In: Putnam AR, Tang C-S, edi-
in Shikoku. Its inhibitory activity was equivalent to the activ- tors. The science of allelopathy. New York (USA): John Wiley
ity of Mucuna and Avena species that showed the strongest & Sons. p 57-73.
activity we ever tested. In Table 1, the letter R shows red rice Fujii Y, Shibuya T, Yasuda T. 1991. Discrimination of allelopathy of
and, generally speaking, the allelopathic activity of red rice is upland rice, taro, and oat by substitutive experiment and its
stronger than that of other rice cultivars. modified experiments. Jpn. J. Soil Sci. Plant Nutr. 62:357-
Though our research aimed to search for potent allelo- 362.
pathic rice varieties that inhibit the growth of accompanying Fujii Y. 1994. The allelopathic effect of some rice varieties. In: Inte-
weeds, the work reported here is only the first step because we grated management of paddy and aquatic weeds in Asia. FFTC
used lettuce as a model plant and did not yet estimate the al- Book Series, No. 45. p 160-165.
Fujii Y. 1999. Discrimination and proof methods for allelopathy by
lelopathic effect in general paddy fields. But it became clear
bioassay, greenhouse and field tests. In: Macias FA, Galindo
that there were distinct differences among rice varieties.
JCG, Molinillo JMG, Cutler HG, editors. Recent advances in
We would like to assess allelopathic activity by using allelopathy. Vol. 1. p 25-28.
paddy weeds as the acceptors, to identify allelopathic sub-
stances and specific genes in relation to allelopathy, and to
Rice blast caused by Pyricularia oryzae is the most serious Mapping and breeding NILs having each QTL region
disease of rice in the world, and cultivation of disease-resis-
tant cultivars is highly desirable for stable rice production. In Kato et al (2002) crossed upland rice cultivar Sensho with blast-
Japan, rice cultivars with a complete rice blast-resistance gene, susceptible japonica paddy rice variety Norin 29 and analyzed
such as Kusabue, suffered from a massive outbreak of blast the resistance to leaf blast of the progeny by QTL analysis.
disease in the 1960s. Thereafter, the importance of a program They found four QTL regions, which were detected in the vi-
for breeding varieties with field resistance to rice blast was cinity of RFLP markers G271 (LOD = 24.8) and G177 (5.7)
strongly recognized (Takita and Solis 2002). However, the of chromosome 4 (Targets 1 and 4), in the vicinity of C1172
mechanism of field resistance has not been fully analyzed be- (3.6) of chromosome 11 (Target 2), and in the vicinity of S826
cause of the complicated hereditary mode and difficulty of (3.4) of chromosome 12 (Target 3).
bioassay; thus, breeding did not proceed satisfactorily. In this Based on this finding, we backcrossed (B4F4–B3F6)
study, we produced recombinant inbred lines using DNA mark- Mine-asahi, a susceptible variety with good eating quality, with
ers for high field resistance of upland rice to simplify compli- Sensho to breed lines having each of the four QTL regions
cated phenomena, and to develop a method of breeding resis- independently, using DNA markers for selection. We bred NIL
tant cultivars. (near-isogenic line)-1, which has the Mine-asahi genome ex-
cept for the RFLP marker G271 region of Sensho (smaller than
22.3 cM), the Target 1 region; NIL-2, which has the Mine-
History of the breeding of rice blast resistance at AARC asahi genome except for the SSR marker RM229–RM287 re-
At the Aichi Agricultural Research Center (AARC), Moun- gion (smaller than 16.9 cM) of chromosome 11 of Sensho, the
tainous Region Agricultural Research Institute, we have been Target 2 region; and NIL-3, which has the Sensho genome
breeding blast-resistant lines of rice since 1933 in an area where except for the SSR marker region RM313–S13752 (smaller
an outbreak of rice blast occurred, by introducing the field than 14.7 cM) of chromosome 12 of Sensho and a part of chro-
resistance to blast of upland rice cultivar Sensho into paddy mosomes 1 and 9, where a QTL was not detected, the Target 3
rice. Until now, scores of rice-blast-resistant cultivars have been region. In addition, NILs that showed resistance, although the
bred. However, it has been difficult to eliminate the inferior chromosome with the resistance gene was not specified, were
characteristics of upland rice such as burning up and poor qual- bred (NIL-5). From the results of a race test, the complete
ity, and only paddy rice with a part of the upland rice genes resistance genes of these NILs were Pia and Pii, like their back-
introduced has been released (Saka et al, in preparation). Breed- crossed parent Mine-asahi.
ing for excellent eating quality satisfying the needs of Japa-
nese consumers along with field resistance to rice blast has
Effect of each QTL region on blast severity
not succeeded because these traits are controlled by polygenes.
Therefore, we tried to resolve the long-pending problems us- The leaf and panicle blast tests gave consistent results in the
ing genomic techniques. two years of testing, and the difference between the criterion
varieties having Pia and Pii and other cultivars was clear. Under
the same test conditions, we performed a leaf blast test for
each NIL. The degree of resistance varied with the NIL, and
was in the order of NIL-1 > NIL-2 >NIL-5 > NIL-3 (Fig. 1).
Fig. 1. Leaf blast resistance in NILs of Mine-asahi backcrossed to Sensho. 1 = Chubu 105 (r), 2 =
Mine-asahi (ms), 3 = Mine-hibiki (rm), 4 = Hitomebore (s), 5 = Sensho (rr), 6 = NIL-1 (Target 1: Chr. 4-
1), 7 = NIL-5 (the region undecided), 8 = NIL-2 (Target 2: Chr. 11), 9 = NIL-3 (Target 3: Chr. 12), 1–5 =
check varieties.
This result is in agreement with the results of QTL analysis blast than Mine-asahi, which has slightly low resistance, and
reported by Kato et al (2002). The result of the panicle blast was judged to have medium to slightly low resistance.
test was similar to that of the leaf blast test, confirming that
the QTL region for resistance to leaf blast also confers resis-
Elite NILs with a narrow Sensho region
tance to panicle blast.
The field resistance to rice blast of the four NILs was The above-described four NILs retained the inferior traits of
compared with that of the criterion varieties (Saka 1996) (Fig. Sensho such as poor eating quality and coarseness of stems
2). NIL-1, which had the highest resistance, developed severer and leaves compared with repeatedly backcrossed parent cul-
symptoms of leaf blast than Sensho, but was more resistant tivar Mine-asahi. To narrow the Sensho region, we selected a
than Chubu 105, which is the criterion variety with high resis- line with a narrower region for the resistance from 672 indi-
tance, and was judged to have very high resistance. The resis- viduals of Mine-asahi*3/Sensho (B2F2) having the Target 1
tance to panicle blast of NIL-1 was slightly lower than that of region, using eight PCR markers included in RFLP marker
Sensho but was higher than that of Tiyonishiki, which had G271–G89B (14.6 cM). As a result, we succeeded in selecting
slightly high resistance. Thus, resistance of NIL-1 to panicle a NIL that has the Mine-asahi genome except for the chromo-
blast was judged to be high to slightly high. These results sug- somal segment smaller than 0.5 cM having Target 1. This re-
gest that NIL-1, having the Target 1 region, has very high re- gion coincided with the estimated locus of rice blast-resistance
sistance to rice blast. In particular, resistance to leaf blast of gene pi21 (Fukuoka and Okuno 2001), and this NIL had field
NIL-1 was very high compared with that of the criterion vari- resistance similar to that of the NIL having Target 1. Thus, the
eties, and we consider that this line could be cultivated with- NIL is considered to be a useful breeding material. In the fu-
out the use of agrochemicals for control of this disease in ar- ture, we will perform linkage drag associated with introgressed
eas where rice blast occurs commonly in Japan. The resis- Sensho DNA to eliminate undesirable characteristics by
tance to leaf blast of NIL-2 having the Target 2 region was marker-aided selection, using various NILs having different
similar to or slightly higher than that of Mine-hibiki, a crite- sizes of Sensho region that were selected in this study.
rion variety, and the resistance to panicle blast of NIL-2 was
similar to that of Tiyonishiki, which has slightly high resis-
The next target
tance. Mine-hibiki and Tiyonishiki are the varieties having the
highest resistance to rice blast among the recommended vari- The level of field resistance to blast of NILs having QTL re-
eties in Japan, and they show high resistance in the area where gions from Sensho was clarified as mentioned above. Among
rice blast commonly occurs. Therefore, Target 2 was consid- them, the NIL with Target 1 had high resistance, which was
ered to contribute by itself to stable rice production. NIL-5, in not observed in known paddy rice varieties in Japan, and we
which the region for the resistance gene was not specified, may be able to grow varieties with this region introduced with-
had medium resistance to leaf and panicle blast. NIL-3, hav- out using agrochemicals for control of rice blast. Hereafter,
ing Target 3, had slightly higher resistance to leaf and panicle not only the simple use of this region but also examination of
8 6
3
4
7 m
mr
6
2
1
5
4
5
0
(excellent) 0 1 2 3 4 5 6 7 8 9 (poor)
Leaf blast severity
Fig. 2. Blast resistance in NILs of Mine-asahi backcrossed to Sensho. 1 = NIL-1
(Target 1: Chr. 4-1), 2 = NIL-2 (Target 2: Chr. 11), 3 = NIL-3 (Target 3: Chr. 12), 4
= NIL-5 (the region undecided), 5 = Sensho (rr), 6 = Mine-asahi (ms); blue dot-
ted line: leaf blast check varieties; Chubu 105 (r), Mine-hibiki (mr), Hitomebore
(s); green dotted line: panicle blast check varieties; Tiyonishiki (mr), Nipponbare
(m).
the cumulative effects and confirmation of the accompanying Saka N. 1996. Panicle blast resistance: test method using the area
inferior characteristics may be necessary. where outbreak of rice blast occurred. In: Yamamoto R,
Recently, we started a national project to introduce these Horisue N, Ikeda R, editors. Rice breeding manual. Tokyo
QTL regions into the leading Japanese variety of rice, (Japan): National Agricultural Research Center: Yokendo
Printing Co., Ltd. p 15-19.
Koshihikari, which occupies 32% of the paddy fields in Ja-
Saka N, Kudo S, Sugiura K, Okuda T, Terashima T. History for 70
pan. If field resistance to rice blast can be introduced into this
years, present situation and the future of the breeding of rice
variety while retaining its excellent eating quality, we will be blast resistance in the Mountainous Region Agricultural Re-
able to cultivate rice stably without disturbing the environ- search Institute, AARC. Plant Prod. Sci. (In preparation.)
ment. Along with the above project, we are now trying to clone Takita T, Solis O. 2002. Rice breeding at the National Agricultural
the pi 21 gene (Fukuoka and Okuno 2001). By clarifying the Research Center for the Tohoku Region (NARCT) and rice
function of this gene, we should be able to understand easily varietal recommendation process in Japan. Bull. Natl. Agric.
the mechanism of complicated field resistance to rice blast. Res. Cent. Tohoku Reg. 100:93-117.
References Notes
Fukuoka S, Okuno K. 2001. QTL analysis and mapping of pi21, a Authors’ addresses: Norikuni Saka, Aichi Agricultural Research
recessive gene for field resistance to rice blast in Japanese Center (AARC), Mountainous Region Agricultural Research
upland rice. Theor. Appl. Genet. 103:185-190. Institute. Inabu, Aichi 441-2513, Japan, e-mail:
Kato T, Endo I, Yano M, Sasaki T, Inoue M, Kudo S. 2002. Mapping norikuni_saka@pref.aichi.lg.jp; Shuichi Fukuoka, National
of quantitative trait loci for field resistance to rice blast in Institute of Agrobiological Sciences (NIAS), Tsukuba, Ibaraki
upland rice, Sensho. Breed. Res. 4:119-124. 305-8602, Japan, e-mail: fukusan@affrc.go.jp.
In the past several decades, the international rice market has Million t
undergone major changes, in particular a shift in the general 160
policy setting, a strong expansion in the volumes of trade, and 140 Wheat: Y = 48.147e0.0269t
a lingering tendency for world prices to decline in real terms
120
and relative to the other two most-traded cereals, wheat and
maize. Nonetheless, the world rice market continues to be re- 100
Maize: Y = 24.223e0.0351t
garded as distorted, thin, segmented, and volatile. This paper 80
discusses whether these attributes still portray the market in 60
the light of the trade liberalization drives prevailing in the
40
1990s. Rice: Y = 6.0645e0.0321t
On the policy front, interventions have diminished in the 20
wake of the market liberalization launched by several coun- 0
1961
1964
1967
1970
1973
1976
1979
1982
1985
1988
1991
1994
1997
2000
2003
tries since the late 1980s. The WTO agreement, in 1994, also
disciplined government policies and helped improve market
Year
access. Nonetheless, rice continues to be one of the most pro-
tected commodities in both developing and developed coun- Fig. 1. Trends in global cereal trade, 1961-2003.
tries, subject to high tariff and nontariff barriers, export re-
strictions or aids, state trading, and domestic market interven-
tions. changes in the relative positions of the major exporters, the
Since the early 1960s, trade in rice has expanded at about supply side of the rice international market is still highly con-
3% per annum, not much different from the pace of growth in centrated, with the top four exporting countries (Thailand,
wheat or maize trade (Fig. 1). However, growth has been far India, Vietnam, and the United States) supplying 66% of trade
from steady. The liberalization thrust of the 1990s coincided and the top ten more than 90% of the total.
with a period of dynamic expansion in the volume of rice trade, Rice is not a homogeneous commodity and there are now
which succeeded a decade of relatively lackluster growth in more than 50 different published international price quotations
the 1980s. The volume of rice exchanged rose from less than 7 for rice. In fact, there are distinct submarkets featured accord-
million tons in 1961 to 24 million t in 2000 and has continued ing to several criteria, the most important of which are variety,
to expand further in the early 2000s, surpassing 28 million t in quality (defined mainly by the percentage of brokens), and the
2001 and 2002. Nevertheless, the international rice market is degree of processing.
still small relative to the other major cereals, with an average The expansion of trade witnessed in the 1990s was ac-
of 27 million t in 2000-03, about one-quarter of the volume companied by small but significant changes in the structure of
traded in wheat and a little over one-third of the trade in maize. the world rice market and in the relative importance of each
Rising import demand by countries in Asia and Africa segment (Table 1). The bulk of global trade continues to be in
was the main force underpinning trade in rice in the 1990s and the form of milled, indica, and higher-quality rice (defined as
early 2000s. The increases in imports were often a reflection containing less than 20% of brokens). However, aromatic rice
of more open trade policies but were also prompted by several varieties, lower-quality rice, parboiled, and paddy have made
production setbacks, for instance, in 1997 in the wake of an El large in-roads and have increased market shares. Those gains
Niño weather anomaly. Despite the consolidation of countries were mainly at the expense of trade in japonica, higher-qual-
in Africa and the Near East as important and stable destina- ity, and milled rice.
tions of rice trade, the demand side of the rice international Most of the highlighted changes can be associated with
market remains highly dispersed geographically, with the top shifts in the geographical pattern of trade. The increasing im-
ten importers accounting for only 40% of the total. portance of Africa and several Asian countries as destinations
Most of the trade expansion witnessed in the past two of rice flows, in particular, has sustained a large increase in the
decades was met by traditional exporters. Thailand, in par- trade of lower-quality rice. The growing importance of aro-
ticular, has maintained its leadership as the top rice exporter matic rice varieties in global trade reflects brisk imports to the
since 1980. Major inroads were made by Vietnam, which be- European Union (mainly of Basmati rice, imported under pref-
came the second most important source of trade supplies in erential access conditions), the United States, Canada, and
the 1990s, a position from which it was eclipsed in the early Australia. However, it can also be associated with large deliv-
2000s, when India started granting export subsidies. Despite eries of Hom Mali rice (a fragrant variety from Thailand) to
1960
1963
1966
1969
1972
1975
1978
1981
1984
1987
1990
1993
1996
1999
2002
High 11,781 77 20,226 75
Low 3,482 23 6,592 25
Year
Degree of processing (forms)
Paddy 263 2 1,122 4 Fig. 2. International rice prices.
Husked 508 3 1,077 4
Milled 12,559 82 20,639 77
Parboiled 1,934 13 3,980 15
a measure of variability on the order of 8%, not very different
Total trade 15,263 100 26,818 100
from that prevailing on the wheat and maize markets. Trade
fluctuated within a much broader band of 12 million to 28
million t in the 1990s, which gave rise to a much higher CV of
countries in Africa, particularly Côte d’Ivoire, Ghana, and 26%. Thus, the strong tendency for international trade to grow
Senegal, albeit with a high percentage of brokens. On the other in the 1990s was associated with much greater volatility in
hand, high degrees of protection in japonica rice markets such volumes.
as Japan, the Republic of Korea, the European Union, and The international rice market has also been character-
Turkey have constrained the opportunities to expand japonica ized by a long-term tendency for world rice prices (represented
rice trade. Tariff escalation, whereby the more processed forms by the Thai 5% broken rice, f.o.b. Bangkok) to fall in real
of a commodity are assigned higher tariff rates, has favored a terms (deflated by the index of unit value of manufactured
strong expansion of trade in paddy, principally to Latin America goods) from 1961 to 2003 (Fig. 2). The decline in constant
and the Caribbean. US$ has been 3% per annum and, in 2003, rice was worth less
An international commodity market is considered “thin” than 40%, in real terms, of its 1961 value. Rice prices also
when it represents a relatively small proportion of global pro- declined relative to wheat and maize. If 1 t of rice could be
duction. The international rice market represented only 3% to exchanged for 2.5 t of wheat in 1961, it could be bartered for
5% of global production in the 1980s, but a strong expansion only 1.3 t in 2003. A similar loss in value was evidenced rela-
of world trade since the mid-1990s has made it “deepen,” as it tive to world maize prices.
has come to represent 7% of global production in recent years. Although variability in the volume of rice trade rose in
Nonetheless, the international rice market remains thin com- the 1990s compared with the 1980s, this did not cause the vari-
pared with wheat or maize, whose trade now accounts for some ability of world prices to follow suit. On the contrary, rice prices
18% and 13% of global production, respectively. have become more stable over time, to the point of achieving
Thin commodity markets are often subject to large swings levels of volatility similar to those exhibited by wheat and maize
in volumes since relatively small changes in supply or use in prices. Actually, on an annual frequency basis, prices in the
important producing countries may give rise to large increases 1990s were more stable for rice than for wheat or maize, in
or contractions in their exports or imports. In general, how- sharp contrast to the pattern prevailing in the 1960s and ’80s.
ever, such countries have preferred to balance their domestic Stabilization of world rice quotations was also evidenced on a
markets by building up or drawing supplies from stocks, with monthly frequency basis.
trade considered only a “residual” option. Thus, the rising variability of trade flows was not asso-
Global trade in rice has fluctuated widely over the past ciated with more volatile world prices, which have instead sta-
two decades, from a minimum of 10.6 million to 28.3 million bilized. Several explanations can be offered to explain this
t, and variability measured by the coefficient of variation (CV) paradox. First, the deepening of the international market has
was high at 37% compared with a 12% variability of global meant a greater dependability of supplies. The existence of
rice production. It was also much higher than the variability of large buffer stocks, the improved flows of information on mar-
wheat and maize trade, which had a CV of 6% and 8%, re- kets and prices, and the introduction of disciplines (restric-
spectively. Variability of rice trade measured decade by de- tions on policies because of trade agreements such as WTO)
cade pointed to a greater stability. In the 1980s, volumes ex- on national and international policies are also believed to have
changed on the international rice market fluctuated within a fostered price stability on the market, in spite of the wider
relatively small range of 11 million to 14 million t, resulting in fluctuations in the volumes of trade.
Rice has been consumed as a staple food more than any other and Taiwan (China) during the last three decades. It is still a
grain in the world. However, in Asia, where 90% of global huge problem involving governments, producers, and consum-
rice is consumed, per capita rice consumption (PCRC) has been ers. In Taiwan, as total consumption decreased to a third, pro-
declining over time. In Japan, PCRC decreased by almost half duction decreased to a third as well. The situation in Japan is
from 125 to 65 kg during the last 45 years. In Taiwan (China), the same. This type of decrease in rice production, which is
it decreased by more than two-thirds from 160 to 50 kg during the core of Asian agriculture, is a tremendous setback. The
the same period. As a result, domestic rice production in both drastic plunge in PCRC in Taiwan and Japan is not unique, but
countries has decreased substantially. The declining trend of any Asian country would have to face the same fate if it did
PCRC seems likely to continue even though the governments not take any effective measures.
of these nations have tried hard to reverse the trend (Fig. 1).
The same type of situation has been occurring in other
Asian countries such as South Korea, China, and India. The
Pressure on market prices
current situation in South Korea is similar to what happened in Rice production is quite sensitive to changes in market prices.
Japan in the 1960s when the decline began. In China, the pace The rice of Thai 5% broken was as low as US$237 per metric
of the decline currently appears to be small at 0.5 kg per year, ton (mt), on a milled basis, in 1993. It increased since then and
decreasing from 110 kg in 1991. However, the diet in large reached $338 in 1996 and remained relatively high until 1998.
areas of China is quite similar to the Taiwanese diet. There- Meanwhile, rice production in Asia increased continuously,
fore, it would not be surprising even if the PCRC in China breaking records every year through 1999, reaching 371 mil-
were to decline as fast as what has happened in Taiwan (China). lion mt. The world total production, reflecting the situation in
The declining trend of PCRC in the Asian countries was first Asia, reached a new record at slightly over 400 million mt in
reported by Ito et al (1989) but it has not hit bottom yet. It was 1999, the first time in history to reach the 400 million mt level.
big news in Japan in May 2004 that PCRC in Japan dropped However, the market prices continued to plunge and dropped
by as much as 2.4% during the preceding 12 months (Asahi way below $200 at $173 per mt by 2001. The price in 2001
Newspaper, June 2004). The countries where PCRC is decreas- was the lowest since 1972. Although market prices recovered
ing are not only economically developed countries. This has some since then, they are still drifting around $250.
spread all over Asia. Accounting for about 90% of world rice production,
Even in China, which is the world’s largest rice-produc- Asian rice producers are quite sensitive to changes in market
ing/-consuming country (holding 1.3 billion people), its PCRC prices, and changes in global production are a reflection of the
has been decreasing steadily during recent years. PCRC in changes in Asian production. Asian rice production decreased
China was increasing until the early 1990s; however, it peaked three years in a row to as little as 344 million mt in 2002, a
at 110 kg in 1991 and then started decreasing. It was as low as 7.3% decrease from the record level in 1999. Asian produc-
105 kg in 2004, about a 4% drop during the last decade. If the tion recovered some in 2003 and 2004. However, the histori-
PCRC dropped by 1 kg in China, this would mean a surplus of cally largest amount recorded in 1999 has not been achieved
1.3 million tons of rice generated in the nation. Population in even after five years. During the last five decades, each his-
China is estimated to peak in the 2030s; therefore, total rice torical production record has been renewed mostly every year
consumption in the country could decrease anytime soon if or after three years as the longest period after the preceding
the current situation continues. record. Nowadays, however, the preceding record has not been
A continuous decrease in PCRC in a country eventually surpassed even at the fifth year. The very low prices around
leads to a decrease in total consumption in the country. And 2000 have affected rice growers. In real terms, taking the con-
this can lead to a decrease in domestic production. This causes sumer price index (CPI) into consideration, the low prices in
a political problem just like the one that prevailed in Japan 2001 were practically the lowest in history. The market prices
Japan
Taiwan (China)
South Korea
160 China
120
80
40
2001
2003
2005
2007
1971
1973
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
1961
1963
1965
1967
1969
Year
Source: S. Ito, World Food Statistics & Graphics (http://worldfood.muses.tottori-u.ac.jp).
Original data sources: USDA: PS&D View, November 2004; USBC: International Data Base, July 2003.
Fig. 1. Per capita consumption of rice for Japan, Taiwan (China), South Korea, and China during 1961-2004.
400,000
300,000
200,000
100,000
0
2002
2032
2038
2005
2008
2011
2014
2017
2020
2023
2026
2029
2035
2041
2044
2047
2050
1963
1972
1975
1978
1981
1984
1987
1990
1993
1996
1999
1966
1969
1960
Year
Fig. 2. World rice total consumption, 1960-2050.
While rice is the major agricultural crop and the major FAPRI (Food and Agriculture Policy Research Institute). 2004.
source of income and calories for most of the Asian people, FAPRI world agricultural outlook 2004: world rice,
less production of rice with weak demand would mean that w w w. f a p r i . i a s t a t e . e d u / O u t l o o k 2 0 0 4 / P a g e M k e r /
Asian people would have a less competitive agriculture rela- 9_WldRice.pdf.
Ito S, Wesley E, Peterson F, Grant W. 1989. Rice in Asia: is it be-
tive to that of other regions in the world and more serious en-
coming an inferior good? Am. J. Agric. Econ. 71:32-42.
vironmental and poverty problems in Asian rural areas. This is
Ito S. World food statistics and graphics, September 2004: http://
a critical problem not only for Asian agriculture but also for worldfood.muses.tottori-u.ac.jp.
any rice producer in the world. United Nations. 2003. United Nations Expert Meeting on World
It is therefore important for Asian nations to cooperate Population in 2003, Department of Economic and Social Af-
among themselves and cope with the declining per capita rice fairs, Population Division New York, 9 December 2003,
consumption. While more human consumption as table food www.un.org/esa/population/publications/longrange2/
should be promoted, more diversified use of rice should be longrange2.htm.
explored even more. USDA (U.S. Department of Agriculture). 2004. PS&D Online, Sep-
tember 2004: www.fas.usda.gov/psd/complete_files/
Note: All the data for rice are on the milled basis provided originally default.asp.
from the U.S. Department of Agriculture. The World Food Statistics
& Graphics (by Ito, http://worldfood.muses.tottori-u.ac.jp) is de- Notes
signed using the USDA’s PS&D Online data set.
Author ’s address: Tottori University, Japan, e-mail:
sito@muses.tottori-u.ac.jp.
References
Asahi Newspaper: Rice consumption declining consecutively for 9
months, 9 June 2004.
Rice is the most important crop in China, with the highest level %
of production. China’s rice production accounts for 28% of 100
the world’s total production. China produced 161 million tons Export
of rice in 2003, but this figure is relatively low compared with 80 Loss
Processing
past years. The rice grown in China consists of indica and Feed
japonica varieties. In general, indica varieties are produced 60 Seed
mainly in the south. Landraces of japonica are produced in the Food
north on the lower reaches of the Chang Jiang River, and im- 40
proved varieties of japonica that originated from Japanese rice
20
are cultivated mainly in the northeast. In recent years, japonica
varieties have increased rapidly, accounting for 26% of the 0
total production, whereas indica varieties account for nearly
2000-01
2000-02
1999-2000
1998-99
70%.
In China, rice has been consumed since ancient times.
The main staple food in China has been either rice or wheat,
depending on the suitability of local environments for cultiva- Years
tion and the diversity of races and eating habits. Because of Fig. 1. Share of rice demand. Source: www.agri.org.cn/
migration between the southern and northern regions, rice con- analysis/include/4/dmjs.htm, 2002 China agricultural
development report.
sumption is now seen all over the country. It is currently con-
sidered that 840 million people (65% of the total population)
eat rice as a staple food. The nationwide per capita rice supply
in choice of other dietary options and increased consumption
has dropped slightly since 1984, when it peaked at 99 kg, to
of meat and other foods. According to statistics on rice pur-
91 kg in 2000.
chases in urban areas, direct rice consumption peaked in the
mid-1980s, and per capita rice consumption (not including con-
Trends of rice consumption sumption at restaurants) has decreased ever since. In 2000, it
was 46 kg, a decrease of 12% from 1995 figures. In rural ar-
Although per capita rice consumption in China is a statistical
eas, as estimated from the total consumption of rice and wheat,
matter and precise figures have not been obtained, recent docu-
direct rice consumption has decreased slightly (90 kg in 1995
ments (based on the food balance sheet) issued by the Chinese
to 89 kg in 2000). However, the quantitative difference in con-
Ministry of Agriculture indicate that polished rice consump-
sumption between urban and rural areas is still significant.
tion is in the 90-kg range. This figure is considerably higher
than that in Japan (65 kg in 1999). Based on a recent food
balance sheet for rice (by crop years), 85% of total demand Increased income and rice consumption
was for food consumption (Fig. 1). In addition, yearly seed
demand has remained at 1.5 to 1.6 million tons, whereas de- In general, changes in rice consumption are influenced by con-
mand for rice for industrial processing has increased slightly sumer income and price. However, price data on rice are in-
to 1.8 million tons (2001-02). There is a significant loss dur- sufficient, and rice prices have been under the control of the
ing distribution and storage, but the total loss has decreased to government for a long time because of the importance of rice
some extent for four years (from about 8 million to about 7 as a staple. In this section, we analyze the relationship between
million tons). Feed demand has increased for four years. Its income and consumption of rice and other foods.
share of total demand has increased from 4% to 6%, and con- We obtained details of income elasticities in the 1980s
sumption of rice as animal feed has reached nearly 8 million and 1990s from data published by the State Statistics Bureau.
tons. Under the current situation, future rice demand will be Income elasticity of rice in urban areas has had negative val-
determined by the rate of decrease in rice consumption and ues since the 1980s, indicating that rice is now seen as an infe-
the trend of population increase. rior commodity in these areas (Table 1). On the other hand, in
Direct consumption accounted for more than 90% of total the 1980s, the income elasticity of rice in rural areas was 0.15
rice demand at the beginning of the 1970s. Since the latter half (meaning that when income increases by 10%, rice consump-
of the 1970s, it has started to decrease because of the increase tion rises by 1.5%), demonstrating that rice was still a valued
Asia has long been dominant in rice production, consumption, and threats in the East around the mid-1800s. China lost Hong
and trade. More than 92% of world rice production, 90% of Kong to the British after a humiliating defeat in the Opium
rice consumption, and 75% of rice surplus for exports are in War. Hong Kong was supposedly on a long-term lease, but the
Asia. The abundant rice supply in Asia explains why 60% of British never did pay any rent to China over 155 years (from
the world’s six billion people are concentrated in Asia. Will 1842 to 1997). India also lost its independence in 1857 after
this picture change in the future? conflicts with East India Company. The U.S. naval officer
Matthew C. Perry forced Japan to trade with the West in 1854.
Next, the British annexed Burma (now Myanmar) in the
Determinants
west, and forced Thailand to open up for trade, particularly in
Rice supply is determined by both natural endowments and rice. The French controlled Vietnam and Cambodia to the east.
human factors. Water, fertile soil, sunlight, and suitable tem- The British invested heavily in irrigation and transportation in
perature are prerequisites for growing rice. Human factors Burma, while the French did the same in Indochina. Thailand
could include particular government policies on free trade ver- also invested a moderate amount in similar infrastructure. In-
sus government control, investments in infrastructure in irri- dians poured into Burma, while the Chinese immigrated into
gation and transportation, and emphasis on technology, for Vietnam and Thailand. Rice production in excess of the do-
example, genetic engineering or traditional biotechnology. Both mestic requirement was stimulated by new demand for inter-
natural and human factors can have a tremendous effect on national trade. Immigrant entrepreneurs brought in their trad-
producing rice surpluses or deficits. Population density or hu- ing and managerial skills together with their initial capital. The
man-land ratios can also become an overriding determinant. British and the Indians helped Burma to export rice to Europe,
India, and Sri Lanka, while the Chinese helped Thailand and
Vietnam to export to Hong Kong and Singapore, and the Viet-
Origins namese were expanding exports to Indonesia and the Philip-
Even though the definitive origins of rice have yet to be estab- pines.
lished, it is quite reasonable to believe that rice originated in The rice economy of Southeast Asia flourished despite
Asia. The shear sizes of the Asian population and rice produc- fears that foreigners would exploit their countries. Prior to the
tion today are good testimonies that rice must have started Second World War, Burma managed to export more than 3
somewhere in Asia. Remains of rice plants from 10,000 B.C. million tons annually. Thailand and Vietnam took turns being
were discovered in Spirit Cave on the Thailand-Myanmar bor- the second- or third-largest exporters at around 1 million tons.
der. Rice and farming implements dating back at least 8,000 Foreign exploitation did indeed turn out to be a blessing in
years have been found along the Yangtze River in China. Rice disguise.
production outside of Asia constitutes only one-ninth of that in
Asia. Because of the crop’s origins in Asia, rice production,
Economic systems
consumption, and trade are all concentrated in Asia.
Nationalism and independence gained after the Second World
War did not help Burma to continue producing huge rice sur-
Trade pluses for export as when the country was operating under a
International trade in rice, in the modern era, started to de- market economy during the long period of British rule.
velop about 150 years ago in Southeast Asia. The West, par- Rice exports reached 2.1 million tons as early as 1900,
ticularly the British, extended its colonial campaign with trade and peaked at 3.4 million tons in 1934. In 1948, the year of the
India
Pressure from population density India used to import a lot of rice: 2.5 million tons in 1915, 3.7
Rice exporters are countries that are generally endowed with million tons in 1919, a peak of 4.4 million tons in 1934, and a
favorable human-land ratios; they also must have a proper gradual drop to 1 million tons in 1941. India became self-suf-
economic system that provides enough incentives for rice pro- ficient in 1977. Exports increased from 0.09 million tons in
ducers. 1994 to 4.9 million tons in 1995 and 6.7 million tons in 2002.
Population density per km2 for the major rice exporters
discussed here are Myanmar (62), Thailand (121), and Viet-
Technology and the market economy are the key
nam (246). In contrast, traditional rice importers prior to the
Second World War with their respective population density The world was not comfortable with food security in the early
were Hong Kong (6,688), Singapore (6,430), Bangladesh 1960s as rice production could not catch up with population
(926), Taiwan (627), South Korea (491), Japan (336), India growth. Like a miracle, IRRI introduced high-yielding variet-
(318), Sri Lanka (298), the Philippines (282), China (134), ies, which were of short maturity, and two to three crops could
and Indonesia (121). be planted in a year. These varieties were also responsive to
The importers that have limited land to support a large fertilizer. That scientific breakthrough was called the Green
population must overcome their scarce natural endowments Revolution. A food crisis has been averted up until today. How-
with technology and management. Japan, China, and India are ever, the growth of rice production seems to have leveled off
good examples of success stories. and again there is concern that population may outgrow rice
production in the decades ahead. In the last few years, produc-
Japan tion has been running behind consumption. This situation must
Rice imports into Japan reached 1.4 million tons in 1919 and be corrected because it cannot be sustained.
2.3 million tons in 1938, and then declined to 1.1 million tons The Green Revolution helped farmers to grow more rice
in 1943 and less than 1 million tons since 1956 onward. Up and real prices therefore came down. Roughly, from 1953 to
until 1945, Japan promoted rice production and imported 1999, rice production increased threefold, but prices dropped
mostly from its two colonies, Korea and Taiwan. Rice con- to one-fourth of those of the beginning period.
sumption in Japan peaked in 1962 at 118 kg per capita, but is Under such circumstances, a few observations are worth
only 61 kg per capita now. Japan has always emphasized high pointing out. Asia and Oceania’s shares of rice imports de-
technology and high rice prices to their producers and at the creased from 69% in 1961 to 38% in 2002. This implies that
same time tightly controlled rice production, imports, and dis- those Asian importers became more self-sufficient and needed
tribution. This self-sufficiency policy has been so successful to import less.
that Japan today is in the process of deregulating the rice in- On the other hand, Asian exporters increased their shares
dustry and using about US$3 billion a year to reduce rice grow- from 70% to 75% during the same period. However, poor
Any surprises?
The Green Revolution got the world out of the food crisis of
the 1960s. Biotechnology, particularly genetic engineering, may
change the landscape that we can see today. It is entirely pos-
Numerous models of the global rice market have been devel- Overview of quantitative models
oped and are used by university and governmental organiza-
tions and agencies. This paper discusses several of the key This paper will discuss primarily the third longer-term cat-
model frameworks for which documentation and projections egory of partial equilibrium quantitative models. Models of
exist. An overview of the various models in terms of type and the third type, which are widely recognized, include the World
scope is presented. Comparisons are made in terms of model Food Model (WFM) of FAO, the IMPACT model of the Inter-
structure and differences in the baseline development process, national Food Policy Research Institute (IFPRI), the AGLINK
model output, and projections. Differences in country/regional model of the Organization for Economic Cooperation and
coverage, product disaggregation, time period of projections, Development (OECD), the Country-Commodity Linked Sys-
and the purpose of the various models are noted. The chal- tem (CCLS) of the United States Department of Agriculture
lenges and prospects for improving quantitative modeling of (USDA), and the Food and Agricultural Policy Research In-
the global rice economy are discussed. stitute (FAPRI) model of a consortium of U.S. universities.
The FAO WFM is no longer maintained. The WFM was
a large multicommodity and multicountry model used to ex-
The need for quantitative models
amine resource, policy, and market issues. Currently, within
Quantitative models of the world rice market are needed for FAO, two global modeling efforts are in a developmental pro-
three alternative analytical needs (Liu and Seeley 1987). First, cess, including the @2030 model, which is more aggregated
we need models to use for short-term impact analysis. These by country/region than the previous World Food Model (FAO
short-term models must be relatively flexible to address issue- 2003). The second modeling effort within FAO is known as
specific policy and market questions, since we never know in COSIMO, which will be a multicommodity, multicountry model
advance what the particular problem, policy, or market issue based on the OECD AGLINK framework but with much greater
will be. Second, we need models that are capable of interme- disaggregation by countries. The IMPACT model maintained
diate-term production, consumption, trade, and price forecast- by IFPRI is an analytical framework used to examine alterna-
ing. Ideally, these models should be multicommodity and ca- tive scenarios for global food demand, supply, and trade. IM-
pable of accounting for substitution effects. It is also extremely PACT covers 16 commodities, including rice, and covers 36
useful if there is a spatial dimension that can account for trade countries and regions, accounting for essentially all of the
flows. The third use we make of market trade models is for world’s food production and consumption. It is structured as a
longer-term policy analysis. The longer-term models typically set of country or regional submodels that are linked through
are used to generate baseline projections against which policy trade and world reference prices that clear international mar-
alternatives can be compared. Typically, the longer-term mod- kets.
els are multicommodity and have a time horizon of ten years Unlike the @2030 and IMPACT models, the AGLINK,
or longer for the projection period. They tend to be more com- CCLS, and FAPRI models are updated at least annually through
plex and useful for generating analyses that evaluate changes a process known as a reference scenario or baseline exercise,
in market structure and policy regimes. Most of these models which is then published annually by each of the respective in-
are partial equilibrium that focus on a broad set of commodi- stitutions as a 10-year global agricultural outlook. In this pa-
ties and countries but take as exogenous variables income, per, comparisons of the AGLINK, CCLS, and FAPRI rice
exchange rates, interest rates, employment, and other factor models are made first by providing a brief description of each
markets. Computable general equilibrium (CGE) models such model with respect to model type, scope, model structure, and
as the Global Trade Analysis Project (GTAP) have been used reference scenario development, including the use of extra-
extensively by the World Bank and the WTO Secretariat for model information. This is followed by some comparisons of
analyses of the Uruguay Round and more recently of the Mil- output and projections. Finally, challenges and prospects for
lennium (Doha) Round of WTO negotiations (Hertel 1997, improving quantitative global rice modeling frameworks will
World Bank 2003). The macro variables and factor markets be discussed.
are endogenized in the CGE model framework. However, be- The AGLINK model maintained by the OECD is a set
cause of the large data and parameter requirements of the GTAP of country modules of OECD member nations and selected
model and other CGE frameworks, there is considerable sac- nonmember nations. As such, the model links the country mod-
rifice in terms of reflecting current market and policy condi- ules through trade and international reference prices. The rice
tions. At present, the GTAP model is based on 1997 data and component of each country model therefore constitutes the basis
will be updated to 2001 data only by spring 2005. for representation of the global rice model. Country models
include major rice importers and exporters among the OECD
Year
Item
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
OECD
Production (million mt) 393 409 417 422 425 429 434 438 441 444 448
Consumption (million mt) 415 415 418 421 425 428 434 438 441 444 448
Prices (Thai milled 100% B, US$) 204 232 252 264 275 286 291 299 303 310 316
USDA
Production (million mt) 399 410 422 425 430 434 438 442 444 447 451
Consumption (million mt) 416 417 423 424 428 431 434 436 440 444 446
Prices (Thai milled 100% B, US$) 200 206 211 216 221 227 233 240 247 255 262
AGRM/FAPRI
Production (million mt) 391 410 425 430 434 436 438 442 445 449 453
Consumption (million mt) 413 415 421 426 431 435 438 441 444 448 451
Prices (Thai milled 100% B, US$) 207 233 241 246 256 267 281 294 308 322 335
References Notes
FAO. 2003. World agriculture: towards 2015/2030: an FAO perspec- Author’s address: 217 Agriculture Building, Department of Agri-
tive. Bruinsma J, editor. London (UK): Earthscan. cultural Economics and Agribusiness, Division of Agricul-
FAPRI. 2004. FAPRI 2004 U.S. and world agricultural outlook. Staff ture, University of Arkansas, Fayetteville, Arkansas 72701,
Report 1-04. Ames, Iowa (USA): Iowa State University. USA, e-mail: ewailes@uark.edu.
Hertel TW, editor. 1997. Global trade analysis: modeling and appli-
cations. Cambridge (UK): Cambridge University Press.
As food consumption has become diversified, rice consump- The age composition of family members, annual income, and
tion has gradually declined in Japan. Individual consumption monthly purchases of various products and services of each
of food products varies significantly by age (Ishibashi 1997), household are recorded in the data. We analyzed the panel data
indicating the importance of the age factor in food demand of approximately 96,000 households each year for 8 selected
analysis. years from 1987 to 2001 (Courtesy of the JBC).
Household income is highly associated with the age of
the household head under the seniority system in the Japanese
Methodology
labor market. Real income-consumption profiles are likely ob-
scured by the fact that Japanese older people (old in age and We first classified the panel by the age of the household head
generation as well) consume more rice on average than (cur- and household size, with the age of family members taken into
rently) younger ones, and that older children generally eat more consideration.
rice than smaller ones. First, the amount of rice purchased and average rice price
Based on the analysis of the panel data of household paid were examined by each household type listed below, then
expenditure surveys, distinct differences are detected in quan- household purchases were regressed against the household in-
tity and price paid for rice among major household types. We come as follows, using the aggregate data grouped by every
then tried to estimate the age factor-free income elasticities of 0.5 million yen in household annual income.
demand for rice cross-sectionally by major household types.
ln Q = a + b ln Y
Data where Q = household monthly purchases (= consumption), Y
The Japanese government Bureau of Statistics (JBC) conducts = household annual income, a = intercept, and b = coefficients
the Family Income and Expenditure Survey (FIES) monthly for income.
with approximately 8,000 households across the nation on a
continuous basis, with one-sixth of them renewed every month.
12
10
4
3,500 4,000 4,500 5,000 5,500 6,000
Yen
Fig. 1. Changes in consumption of rice with respect to quantity and price paid by se-
lected household types.
Results million yen per year for most years, whereas the latter cat-
egory has 2.0 to 4.0 million yen. Regarding the average prices
Major household types in Japan paid, those in their 50s paid higher prices (rightward horizon-
According to the JBC, the sample household is selected in pro- tally) than those in their 60s in every year.
portion to the size of the household types in Japan. The major Within the same household type, that is, when the age of
household types selected for our analysis were family members is controlled, clear tendencies are observed
O Households in which a couple are in their 30s with that the greater the household income, the smaller the con-
two children under 10 years of age, 5,663 in 1987 sumption and the higher the price paid. For households of a
and 3,979 in 2001 in total number, for example. couple in their 40s with two children from 10 to 20, for ex-
O Households in which a couple are in their 40s with ample, those with an annual income of 2.0 to 4.0 million yen
two children age 10 to 20, 4,584 in 1987 and 2,880 in purchased 16.8 kg of rice per month on average at the average
2001, respectively. price of 5,257 yen per 10 kg in 1987, whereas those with in-
O Households in which a couple are in their 50s with a come over 10 million yen purchased 12.3 kg at 5,897 yen in
child in his/her 20s, 2,840 in 1987 and 3,099 in 2001, the same year. The corresponding numbers for 1996, for ex-
respectively. ample, are 12.7 kg at 4,815 yen versus 8.9 kg at 5,099 yen,
O Households in which a couple are in their 60s and and for 2001 are 8.9 kg at 3,743 yen versus 7.5 kg at 4,430
have no dependents, 3,761 in 1987 and 7,449 in 2001, yen, respectively.
respectively.
Income elasticities for rice
Rice consumption by household types Most previous studies, using cross-sectional data, suggest that
Figure 1 clearly shows that changes in consumption of rice income elasticities for rice should be positive in sign (Appen-
with respect to quantity and price varied greatly among the dix table 4 in FIES annual report, various issues). To deter-
selected household types from 1987 to 2001. The average price mine the pure income effects on rice consumption, we ran a
paid declined consistently from 1987 to 2001 (downward slope regression by household types.
along the X-axis) in every household type, with the younger Table 1 shows estimates of income elasticities for con-
households showing greater declines. At the same time, a de- sumption of rice by household type, using panel data aggre-
crease in quantity purchased occurred for all household types gated by income groups in 0.5 million yen. Both the house-
(downward slope along the Y-axis). The households of a couple holds in which a couple are in their 30s with two children un-
in their 40s with two children from age 10 to 20, in particular, der 10 years of age and those of a couple in their 40s with two
demonstrated the largest drop from approximately 15 kg per children from 10 to 20 are estimated to carry negative income
month in 1987 to approximately 9 kg per month in 2001 (the elasticities, –0.1 to –0.2, with meaningful t-values for most of
topmost line in Fig. 1). our survey years. When these two types are grouped together,
Figure 1 also shows that the households in which a couple however, we get positive income elasticities, 0.2 to 0.3, with
were in their 50s with a child in his/her 20s and those of a significant t-values and much greater R2 s. This might suggest
couple in their 60s without dependents share strikingly similar that household purchases of rice tend to increase with the grow-
tendencies in consumption changes, with the two lines over- ing up of children from under age 10 to teens, whereas the
lapping over time, despite big differences in income between household income also tends to increase in accordance with
the two categories, that is, the former register more than 10 the aging of household heads from their 30s to their 40s. The
Household types
A couple in their 30s with two A couple in their 40s with two Two types aggregated
children under 10 years of age children from 10 to 20
aDouble log model used: ln Q = a + b ln Y, where Q = household monthly purchases and Y = household annual income. Glutinous and nonglutinous rice for
above results may strongly suggest a need for classifying house- studies, including the government FIES annual re-
holds by the age of family members if we wish to determine ports.
the pure income effects on rice consumption in Japan using Our findings, as shown in Figure 1, suggest that age ef-
cross-sectional data in particular. fects should overwhelm income effects in rice consumption
and prices paid are found to differ significantly by age of house-
hold head and household income. This should be taken to im-
Conclusions ply the importance of incorporating the age factor into analy-
Tentative conclusions from our investigations are as follows: sis to better comprehend Japanese rice consumption, includ-
1. The household quantities of rice purchased and prices ing future projections.
paid differ, and the patterns of change in consump-
tion over time also vary significantly by types of
References
household.
2. When the age of family members and the household Ishibashi K. 1997. Changes in Japanese dietary patterns by age. Is
type are controlled, the households with a larger in- the “Japanese style” disappearing? Natl. Agric. Res. Ctr. Ru-
come are found to purchase a smaller amount of rice, ral Econ. Res. 36:17-31.
and to pay higher prices.
3. The cross-sectional income elasticities for rice are es- Notes
timated to be negative in sign and significantly dif-
ferent from zero in Japan, opposed to most previous Authors’ address: National Agricultural Research Center, Kannondai
3-1-1, Tsukuba, Ibaraki, Japan.
Wrap-up of Session 17
This session dealt with the international rice market situation, tural research, especially the development of technologies, should
with emphasis on Asian countries. The main objectives were to be consistent with these structural shifts, so that research re-
obtain common understanding on past trends and future pros- sults will be relevant. This session therefore attracted a large
pects in rice supply and demand among researchers by exchang- audience consisting of both economists and natural scientists.
ing views, as well as to identify research priorities in this field. Regarding the demand side, we are encountering epoch-
These include the identification of emerging factors influencing marking phenomena such as diversification of the rice-based di-
supply and demand of rice and the improvement of analytical etary pattern in Asian countries, the increase in processed and
tools such as econometric models. nonfood use for rice, and quality differentiation in the rice mar-
In addition to day-to-day market fluctuations, the supply ket. For the supply side, skepticism on yield potential prevails, as
and demand situation structurally and fundamentally shifts ac- water use and chemical inputs are no longer unlimited. The cost
cording to economic and social environmental changes. Agricul- of production tends to increase in many countries. In addition,
Studies (e.g., Wailes, 2003) showed that, although the removal are uncorrelated to each other. CMT considered two other al-
of tariff and export subsidies in the highly distorted rice mar- ternatives of the expression of νit in replacing λt by (1) θi*t
ket would benefit the world, developing countries that pro- indicating that the growth rate of TFP (coefficient on time trend)
duce rice but are net rice importers (e.g., Bangladesh, Brazil, is different for each country and (2) θ*t indicating that the
Indonesia, Madagascar, Nigeria, and the Philippines) would TFP growth rate is the same for all countries. Taking the loga-
lose. For these countries, with mostly poor rice farmers and rithms of the production function and taking into account these
low productivity, the increase in the producer’s gain would be three expressions of the residual, production per worker is
too little to overcome the consumer’s loss as the market price
rises. Seemingly, liberalization yields little benefit unless pro- k–1 k–1
ductivity rises first. But, did the lack of price competitiveness, (1a) yit = γyit–1 + Σ αjxjit + Σ βjxjit–1 + δ0Χkit
caused in part by distorted rice policies that depressed rev- j=1 j=1
enues and profits, prevent poor farmers from adopting avail- + δ1Χkit–1 + θi *t + µi + εit
able technology?
The objective of this paper is to determine how distorted k–1 k–1
policies in the international rice market affect productivity and (1b) yit = γyit–1 + Σ αjxjit + Σ βjxjit–1 + δ0Χkit
competitiveness in the rice sector. Specific objectives are to j=1 j=1
(1) measure the level and growth of total factor productivity + δ1Χkit–1 + θi *t + µi + εit
(TFP) in the main rice-producing countries, (2) explain the
sources of productivity gaps among rice producers, and (3)
k–1 k–1
determine the interactions among trade liberalization, produc-
(1c) yit = γyit–1 + Σ αjxjit + Σ βjxjit–1 + δ0Χkit
tivity, and competitiveness. The focus is on 33 rice-producing
j=1 j=1
countries during the period 1961-2002. This study provides a
+ δ1Χkit–1 + λt + µi + εit
new framework for measuring and analyzing sources of the
productivity gap.
Variables y and x represent the log of output and input per
worker, and ε’s are error terms.
Rice markets and policies
Rice trade is thin and represented only 6.5% of global rice Competitiveness and total factor productivity
consumption during 1997-2002 (Childs 2002). Wailes (2003) We define a competitiveness index C as the ratio between value
reviewed current policies in the rice-producing countries in added estimated at world prices and input cost estimated at
the OECD (Organization for Economic Cooperation and De- shadow prices: C = pV/(ωL+rK) where ω and r are shadow
velopment), where production and exports are highly subsi- prices of labor L and capital K. Variable K combines all
dized and protected. These policies have caused distortions in nonlabor inputs such as human capital, land, fertilizer, and trac-
prices and resource allocation for developing countries. Dur- tor uses. p is the world price of the value added V; V = f (K, L,
ing 1961-2002, the price-adjusted inflation of rice in the world A) where A is the level of technology. As in Nishimizu and
market has tended downward despite the growing global rice Page (1986), price p can be called “terms of trade,” reflecting
demand, but the price of imported farm inputs has increased differences between output price and intermediate input costs.
sharply. The change in competitiveness index is
Framework dC dp dω dr dL
(2) = – sL – sK – (sL – aL)
Measuring productivity C p ω r L
We assume that the rice production function is Yit = F (X1it,...,
dK
Xkit, ev ), where index i represents country, t is the time dimen-
it
– (sK – aK) – aA .dA
sion, and superscripts 1 …k represent input type. Y represents K
output, X’s are inputs, and Xk is labor. The exponent of the
residual νit represents the level of technology. Using Cermeño
et al (2003) henceforth as CMT, the residual in a dynamic form
is νit = µi + λt + Φνit–1 + εit , where εit˜ iid (0,σε 2), µ, and λ
represent individual and time-specific effects, and µ, λ, and ε
aGroup 2 includes Argentina, China, Colombia, Egypt, Guyana, India, Myanmar, Pa-
Growth of Growth kistan, Suriname, Thailand, Vietnam, and Uruguay. bGroup 3 includes Bangladesh,
TFP gap + Disturbances = of Brazil, Cambodia, Guinea, Indonesia, Iran, Laos, Madagascar, Mali, Malaysia, Nepal,
Nigeria, Peru, Philippines, Sri Lanka, and Tanzania. cNumbers in parentheses and
(home/world) below the coefficients are t-values. The symbols *** , **, and * are levels of signifi-
(3) cance at 0.01, 0.05, and 0.1, respectively.
Other Change in
Output Intermediate
price
–
good’s price
– factor + production Results
costs efficiency
Measures of TFP level and growth rate
From Coe et al (1997), we add control variables such as Between-groups estimation. In all three models, (1a), (1b), and
openness (OPEN), human capital (HK), and access to agricul- (1c), µi’s, the group-specific effects, are statistically signifi-
tural machinery—a function of equipment price and PC—and cant for group 1 and group 2 (group 3 is the basis for compari-
propose the following model: son). As expected, the high-income countries in group 1 have
the highest level of TFP versus group 2 and group 3. Model
TFPw (1b) produces the best fit and indicates, through θ, that the
(4) log ( )t = α0 + α1 log(Pw)t + α2 log(q)t trend in world rice productivity grew at 0.4% on average per
TFPi
year during 1961-2002.
+ α3 log (HKw / HKi)t + α4 log(OPENw / OPENi)t + ηit Within-group estimation. For group 1, model (1b) yields
the best fit. The coefficient θ is significant, indicating that TFP
where Pw and q are the prices of rice and inputs and farm in group 1 grew at 1.6% per year on average. For group 2,
equipment. Subscripts w and i designate the world’s and indi- model (1a) gives the best fit. Country-specific effects on TFP
vidual country’s variables. are significant in eight out of 11 countries. Model (1b) pro-
vides similar results and shows that TFP in group 2 grew at
1% per year during 1961-2002.
Data For group 3, model (1b) best represents the data. All
We divide the 33 countries into three groups. Group 1 includes country-specific effects of TFP are significant. Also, TFP
OECD countries, rice producers: Australia, EU (Italy), Japan, growth is about 0.8% per year, meaning that the “catching-up”
Korea, and the United States. Group 1 has high rice yields, hypothesis does not hold; countries in group 3 started at lower
subsidizes heavily its rice production and exports, and has high TFP levels and their TFPs have grown slower than in group 2
import protection. Group 2 includes developing countries, rice and group 1.
exporters, with little or no subsidies and relatively high yields:
Argentina, China, Colombia, Egypt, Guyana, India, Myanmar, Explaining productivity gaps
Pakistan, Suriname, Thailand, Vietnam, and Uruguay. Group 3 From model (1a), we build a time series on TFP levels and
includes mostly low-income countries that produce and import growth for each group. Estimates of TFP levels are plotted in
rice from the other two groups: Bangladesh, Brazil, Cambo- Figure 1 showing the widening TFP gap, especially between
dia, Guinea, Indonesia, Iran, Laos, Madagascar, Mali, Malay- group 3 and group 1. We employ TFP measures to estimate the
sia, Nepal, Nigeria, Peru, the Philippines, Sri Lanka, and Tan- parameters of (4).
zania. Data are yearly from 1961 to 2002. Appendix A explains Results are in Table 1. Coefficients have the expected
the data sources. signs and are significant, except for the openness variable. The
2.0
1.5
1.0
0.5
0.0
1961 1965 1969 1973 1977 1981 1985 1989 1993 1997 2001
1963 1967 1971 1975 1979 1983 1987 1991 1995 1999
Year
Fig. 1. Estimated TFP index levels for the main rice-producing countries (1961-2002).
Group 1 includes Australia, EU (Italy), Japan, Korea, and the United States. Group 2
includes Argentina, China, Colombia, Egypt, Guyana, India, Myanmar, Pakistan,
Suriname, Thailand, Vietnam, and Uruguay. Group 3 includes Bangladesh, Brazil,
Cambodia, Guinea, Indonesia, Iran, Laos, Madagascar, Mali, Malaysia, Nepal, Nige-
ria, Peru, Philippines, Sri Lanka, and Tanzania.
TFP gap has grown partly because of the lack of price com- sources of this productivity gap. Limited access to equipment
petitiveness. A 10% decrease in the world price below its free has increased the TFP gap. Also, trade openness has a weak
trade level has increased the TFP gap ratios between group 1 positive impact on the TFP gap; this may partly reflect diffi-
and the other two groups by 0.5%. Coefficients on q are statis- culties in choosing proxies (Edwards 1997).
tically significant. Each percent increase in equipment price Results imply that removing distortions on output price
would widen the TFP ratios by about 0.11% and 0.15% for and input access can boost revenue and profit for rice produc-
groups 2 and 3. The rising equipment price reduced poor farm- ers in developing countries. That will allow investments for
ers’ access to the inputs necessary to adopt available technol- the use of available technology and reduce the TFP gap.
ogy. As a result, their TFP levels lag behind TFP in rich coun-
tries and the TFP gap widens.
The coefficients on OPEN are positive, and are signifi-
Appendix A: Variables and sources
cant at the 0.1 level for group 2, but not significant for group Inputs: Xf , fertilizer use, proxyed as the total fertilizer use
3. This supports early criticism (Rodrik 1992) that openness times the share of harvested rice fields over arable land (UN
does not necessarily guarantee a reduction in TFP gaps. The FAO); Xk , amount of labor in rice production proxyed as the
coefficients on HK ratios are significant, indicating that rapid rural labor force, that is, active workers in rural areas (World
growth of human capital in groups 2 and 3 could help reduce Bank); Xtr, physical capital, proxyed as the average number of
TFP gaps. tractors used per unit of agricultural land times rice-harvested
area; Xl , rice-harvested area (UN FAO).
HK: human capital, is net secondary enrollment ratio (World
Conclusions and implications
Bank).
This paper focused on how distorted trade policies affect pro- OPEN: openness, is trade value over GDP (World Bank).
ductivity and competitiveness in the rice sector. We estimated Pw: world price of rice; proxy is the f.o.b. price of 5% broken
TFP levels and growth rates and explained the sources of the milled rice from Bangkok, Thailand; prices are in 1982 US
TFP gap among rice-producing countries. We found that the dollars constant (World Bank).
TFP gap in the rice sector has widened. The catching-up hy- q: price of agricultural and farm equipment; proxy is the index
pothesis does not hold: TFP levels and growth rates have been price of agricultural and farm inputs and equipment times the
higher in high-income countries than in developing countries. exchange rate (UN FAO, U.S. Department of Labor, World
The lack of price competitiveness, partly because of heavy Bank).
production and trade subsidies in the OECD, was one of the
Japanese rice farming was influenced by the liberalization of Prefecture, and farmers obtained new jobs in nonagricultural
the domestic market and minimum access to rice trading. The industries. Under those conditions, most farmers in the town
price of rice decreased by 27% from 1992 to 2002. Because of became part-time farmers and took part in community-wide
this problem, the Japanese government promotes the improve- group farming to maintain their rice production through the
ment of the productivity of rice production. This policy con- joint use of farm machines, joint farm work, and community
siders large-scale individual farmers (10–20 ha) and coopera- land-use arrangements because their farm size was too small
tive management farms (35–50 ha) to be model farms. These to modernize rice production and make enough money.
are estimated to be only 45,000 of the total of 3 million Japa- As a result, 21 of the 32 rural communities in Johana
nese farmers. So, in regions where large-scale farms aren’t now have group-farming units. The farmland managed by these
found, “community-wide group farming” based on the rural units is two-thirds of the paddy field in Johana (1,367 ha). On
community is expected as a model of farming. the other hand, only 33 of the total of 649 town farmers are
In Japanese agricultural areas, traditionally, rural com- full-time farmers engaged in beef-cattle raising, fruit produc-
munities of 20 to 30 farmers have been the basic units not only tion, or vegetable cultivation in addition to rice growing.
of social life but also of rice-farming work, such as the man-
agement of irrigation water and joint farm work. Group farm-
ing is an activity based on the rural community, which aims to
Foundation of the Noguchi farming group
modernize and restructure rice production through joint use of In Noguchi community, 16 farmers started the joint use of rice-
farm machines, joint farm work, and community land-use ar- farming machines in 1975. In 1983, when restrictions on rice
rangements. Therefore, about 10,000 farming groups were es- growing had become serious, Noguchi community started the
tablished in Japan. joint production of soybean and barley as an alternative to rice.
The Noguchi farming group of Johana town, Toyama Then, in 1985, as a project financed by Toyama Prefecture,
Prefecture, is an example of community-wide group farming. the community introduced rice-farming machines and con-
The Noguchi farming group is incorporated and is now work- structed a joint rice drying and processing plant (Table 1),
ing on the organic cultivation of rice, too. This report analyzes thereby starting community-wide agriculture with the partici-
the group’s farm management activities for over three decades pation of all 23 farmers in the community. In general, commu-
and discusses its improvements. nity-wide farming groups adopted the joint use of machines in
individual fields, and members got the harvest from individual
paddy fields. But, in the group farming in Noguchi reorga-
Outline of the area nized in 1985, a new system was adopted: the members worked
Traditionally, most farmers in Johana have been engaged in together and the harvest was processed and sold jointly in the
rice growing on their land of 1 ha or so. But, in the 1970s, community. And the earnings were distributed to the member
factories of metals and electronics were constructed in Toyama farmers as land rent and wages for farm work.
production and selling. Second, by reducing land rent pay- productivity as high as that of large, full-time rice growers.
ments and raising wages for farmers, farmers are encouraged That is the reason the Japanese government considers com-
to take part in joint farm work. Third, the incorporation en- munity-wide group farming to be an exemplary model of rice
abled the group to have stocked earnings, which made it easier farming and promotes that instead of further globalization.
to renew farm machines. Finally, by introducing organic rice
cultivation, profitable farming was obtained. With these strat-
Notes
egies, group farming has overcome the overinvestment in ma-
chines and ineffective management, the situation that Japa- Author’s address: NARO, e-mail: kyoichi@affrc.go.jp.
nese part-time farmers usually experienced, and has obtained
Rice has traditionally been the main source of livelihood among ing those periods, rice farmers mostly enjoyed subsidies, irri-
Indonesian farmers. Rice has also remained a unique and stra- gation establishment, and market protection from the govern-
tegic commodity, closely related to food security and economic, ment.
social, and political concerns. Rice has been the staple food Because of the economic crisis, beginning in 1998, In-
for the majority of Indonesian people and still holds the high- donesia introduced changes in its rice policies to include rice
est share among consumption commodities. The rice sector trade liberalization. The government opened the rice economy
employs more than 18 million farming households, or about more to the global market. Prior New Order strategies for the
70.8% of the total farming households in 2003, of which the rice economy were abolished gradually, including some subsi-
farming household itself constitutes more than 48% of the to- dies. The rice sector has lost some protection since rice trade
tal households. was liberalized in 1998, and local farmers now have to face
Because rice has an important role in the Indonesian global competition, whereas the government strategy has
economy, the government needs to protect both rice farmers shifted to food security rather than self-sufficiency.
and consumers, and assure enough rice supply to meet demand. This development has necessitated more research to ex-
Under the New Order administration (1969-98), the primary amine the performance of the rice economy as affected by food
effort in agricultural development was devoted to increasing security strategies, mostly from a macroeconomic point of view.
rice production to the point of achieving self-sufficiency. There are only a few studies on farmers whose lives revolve
Through intensification programs, the so-called BIMAS (lit- mainly around rice as their main livelihood. Using the case
erally meaning mass guidance) and INMAS (literally meaning study approach, this paper attempts to investigate the role of
mass intensification) programs, Indonesia managed to boost the rice economy given some rice policy reforms made, from
rice production and achieve rice self-sufficiency in 1985. Dur- the perspective of Indonesian rice farmers in Java that histori-
Principal factor
Item
F1 F2 F3 F4 F5 F6 F7
Wetland rice performance had better irrigation networks. This was clarified by the higher
R1: Crop intensity per year (%) proportion of technical irrigated land in lowland regions (IN2).
R2: Yield (kg ha–1) G2 and G3 were both located in upland regions and had
high annual rainfall, though several other variables showed
differences between the two. G2 was relatively high in IN4 (%
Results of principal component analysis and grouping of villages with paved main road) but relatively low in IN5
Results of principal component analysis show that 75% of the (number of households to number of cars), IN7, IN8, HH3,
total variance could be explained by seven principal factors and R1. This means that G2 and G3 had a different level of
(Table 1). The first factor, highly correlated to variables X1, infrastructure, household landholding, and crop intensity for
IN1, IN2, IN3, HH1, and HH2, represents production condi- wetland rice, even though they had similar natural resource
tions, contributing 27% to the total variance. The second fac- endowments.
tor, related to IN4, IN5, HH3, and HH4, stands for social in- A similarity of natural resource endowments is also ob-
frastructure. The third factor, related to IN7 and R1, indicates served for G1 and G4. Both groups were located in lowland
technological level and productivity. The fourth factor, related regions and had relatively low annual rainfall. However, G1
to HH4, could be interpreted as social structure. The fifth fac- had a better transportation infrastructure (IN4) and larger av-
tor, related to X2, indicates rainfall. Those five factors, being erage farm size (HH1 and HH2) than G4. In contrast, G4 had
able to explain about 65% of the total variance, are used for a higher crop intensity for wetland rice than G1.
kecamatan grouping.
Table 2 shows the grouping of kecamatans based on five
Rainfall decrease, area decrease, and vulnerability
principal factors. The comparison of mean values of variables
between groups shows that several variables were highly cor- The rainfall decrease caused by El Niño varies by locality be-
related to one another. For instance, kecamatans with the most cause of the difference in natural resource endowments. How-
villages located in regions with altitude of <500 m or in low- ever, at a homogeneous level of rainfall decrease, production
land regions (G1 and G4) had lower rainfall than kecamatans loss might differ, depending on farmers’ capability to cope with
located in regions at altitude of >500 m or in upland regions the effects of a rainfall decrease on their agricultural produc-
(G2 and G3). Wetland rice yield in lowland kecamatans was tion. The farmers’ ability to adjust to a rainfall decrease is a
also higher than in upland ones because the former generally function of the characteristics of farmers’ households, infra-
Group of kecamatans
G1 G4 G2 G3
Characteristics of kecamatans
IN4 Percentage of villages with paved main road (%) High Low High Low
IN5 Ratio of number of households to number of cars Low Low Low High
IN6 Ratio of wetland area to number of water pumps Low Low Moderate High
IN7 Ratio of wetland area to number of agicultural input kiosks High High Low High
IN8 Percentage of villages accessible to mass transportation High High Low High
HH1 Ratio of wetland area to farmland High Moderate Low Low
HH2 Mean wetland size (ha per farm household) High Moderate Low Low
HH3 Percentage of farm households with wetland size
<0.5 ha per household (%) Low High Low High
structure availability such as irrigation networks, and an insti- were located in lowland regions and had similar annual rain-
tutional support system such as extension, cooperatives, local fall. This indicates that the vulnerability index of kecamatans
community, etc. for El Niño onset was determined not only by natural resource
As shown in Table 2, a high rainfall decrease could be endowment but also by other factors. The probable other fac-
seen for groups G2 and G3, at 45% and 42%, respectively, tors are infrastructure, household social characteristics, and
whereas, for groups G1 and G4, the decrease was 30% and farming practices, which might influence farmers’ ability to
29%, respectively. This shows that the rainfall decrease in the reduce El Niño consequences.
El Niño event was lower in lowland regions than in uplands. To understand the other factors that affect vulnerability
Therefore, it was quite reasonable that the area decrease was indices of kecamatans, groups G2 and G3, both located in up-
higher in upland regions (6% and 8%, respectively, for groups land regions having relatively homogeneous annual rainfall,
G2 and G3) than in lowland regions (4% for both G1 and G4). were compared. As indicated in Table 2, G3 was higher than
The number of kecamatans having an area decrease was also G2 in vulnerability index, scarcity of water pumps (IN6) and
higher in upland regions, 58% and 85%, respectively, for groups input kiosks (IN7), proportion of small farmers (HH3), and
G2 and G3, whereas the percentages for groups G1 and G4 crop intensity (R1), but lower in the availability of transporta-
were 55% and 51%, respectively. tion infrastructure (IN4 and IN5). This reveals that the vulner-
The average vulnerability index of all kecamatans was ability index was higher in isolated regions, which commonly
0.187, which means that a 10% decrease in rainfall will lead to have few agricultural input kiosks. The index will also be higher
an area decrease of around 1.9% for wetland rice. Vulnerabil- with increasing annual crop intensity for wetland rice because,
ity indices were significantly different between groups G2 and if annual intensity for wetland rice were high, then most of the
G3 (0.137 versus 0.219), although both had similar natural wetland area in the dry season would be allocated for rice cul-
resource endowments (altitude and rainfall). The comparison tivation. Rice cultivation in the dry season had a high prob-
between G1 and G4 indicates a similar result, where vulner- ability of failure since El Niño 1997 caused a higher rainfall
ability indices were significantly different, although both groups decrease in the dry season than in the wet season.
During the last five decades, rice area in India has increased Methodology
from 20 million to about 45 million hectares and rice produc-
tion from 22 million to 90 million tons. Productivity increased Policy matrix analysis
from 700 to 2,000 kg ha–1. Rice is a major cereal for more The trade competitiveness of Indian medium-quality rice was
than 70% of the world’s poor in Asia, where more than 90% of analyzed using the framework of the policy analysis matrix
rice production and consumption take place. Despite the rapid (PAM) as proposed by Monke and Pearson and presented by
expansion of rice trade globally, as argued by Gulati and Yao (1997). To construct the PAM, world prices (free on board,
Narayanan (2003), the world rice market is characterized by FOB, for exports, and cost, insurance, and freight, CIF, for
thinness, volatility, and segmentation. Rice markets are affected imports) were used. For domestic factors, which are not traded
by distortions caused by a plethora of controls imposed by internationally, social costs (Pxi) were calculated using the
both developed and developing countries because of various value of marginal product approach that uses factor shares (Si)
political, economic, and food security compulsions. In the long of various inputs (Xi) together with the mean value of inputs
run, trade liberalization is likely to benefit both consumers and and outputs (Y) and price (Py). Social cost was obtained as
producers in developing countries, in addition to the avail- Pxi = { (Si/Xi) × Y} × Py. The nontradable inputs considered
ability of cheaper grain in the importing countries (Gulati and were labor (human and bullock), farmyard manure, and rental
Naranayan 2003). value of land. The tradable inputs were seed, fertilizers, and
India is a traditional exporter of high-quality basmati pesticides.
rice. But the export of nonbasmati rice was rather small in Nominal protection coefficients (NPCs), effective pro-
earlier years, as it was not export-competitive (Chand 2002). tection coefficients (EPCs), and domestic resource cost (DRC)
However, the devaluation of the Indian rupee drastically al- were computed to reveal trade competitiveness. Trade com-
tered the export scenario of the country. This was followed by petitiveness was estimated using the above three measures for
a gradual relaxation of restrictions on the export of rice, which rice in India using the data from Karnataka on the importable
resulted in a gradual increase in exports of medium-quality hypothesis for the two periods, preliberalization (1985-86 to
rice. To assess the export potential and consequent welfare 1991-92) and postliberalization (1996-97 to 2000-01).
gain from rice, it is mandatory to assess the export competi- NPC was calculated as a ratio between the domestic price
tiveness of medium-quality rice. As Karnataka symbolizes and the international price of a comparable grade of rice, ad-
medium-quality rice production in the country, we have cho- justed for all the transfer costs such as freight, insurance, han-
sen this state for assessing the export competitiveness and con- dling costs, margins, losses, etc.
comitant welfare gain due to the export trade of rice.
NSLp NSLc NSL Estimated welfare gain Estimated welfare loss Net effect of
(Rs.) (Rs. 000) (Rs. 000) of producers, WGp of consumers, WGc Liberalization on
(Rs. million) (Rs. million) welfare in the state
Conclusions Notes
The study shows that liberalization will benefit the rice sector Authors’ address: Department of Agricultural Economics, Univer-
in terms of giving farmers a better deal. Consumers may have sity of Agricultural Sciences, GKVK, Bangalore 560 065,
to pay a higher price because of the limited domestic supply India, e-mail: batlahalli@yahoo.com.
and increase in prices. The positive impact on the farming com-
munity may lead to more efficient rice production and in the
process increase the export prospects of rice.
In 2002, the domestic supply of Japanese rice consisted of Rice crop area (000 ha) Set-aside area (000 ha)
Producers rice price (US$ t1) Demand for rice (10,000 t)
8.876 million tons of production and 0.77 million t of mini- Rice yield (kg ha1)
mum access import. Rice was cultivated on 1.688 million ha. 3,000 1,100
To maintain the crop’s taste, farmers have not been pursuing
high yield, but, thanks to good weather, the average rice yield Set-aside area
2,750 1,000
was 5.27 t ha–1 and almost constant since 1994. On the other
hand, domestic demand for rice decreased from 10.26 million Demand
t (1992) to 8.95 million t (2002). Japanese per capita rice con- 2,500 900
sumption decreased drastically from 1960 to 1980, and this
considerable decrease in rice consumption still continues. It is 2,250 800
also important to notice that, in that period, the producers’ price Rice crop area Rice price
of rice fell, corresponding to the decrease in demand (Fig. 1). 2.000 700
This shows that the consumption of rice itself is diminishing.
The decreasing demand together with the high average Rice yield
1,750 600
yield necessitate a reinforcement of rice supply adjustment. In
2002, the set-aside area of rice production reached 39% of
1,500 500
paddy fields. Therefore, to convert crops from rice, a lot of 1994 1996 1998 2000 2002
subsidy was paid to farmers. The government expended 1995 1997 1999 2001
US$2.467 billion for this rice supply adjustment program in Year
2002, and it occupied one-half of financial expenditures re- Fig. 1. Trend of Japanese rice supply, demand, and price. Source:
lated to rice policy. Consequently, a decrease in these expen- References for rice production Production Bureau, Ministry of Ag-
ditures became an urgent issue. riculture, Forestry, and Fisheries.
On the other hand, in 1992, the government designed a
structural change in arable farming, targeting 80,000 farms cul-
tivating 20 ha each, with a target of sharing 60% of paddy under market mechanisms by 2010. The government also aims
fields by 2020. Nevertheless, in 2000, the share of the rice to encourage large-scale farms. Therefore, the government is
crop area cultivated by large-scale farms (more than 5 ha of now considering a conversion from a price-support policy to
rice crop area) was only 14%. This means that the accumula- direct payment for targeted farmers.
tion of farmland for core farmers is stagnant and structural This paper discusses the reactions and strategies of Japa-
adjustment has not progressed as the government expected. nese rice producers responding to various economic and po-
These facts show that nowadays there are three main is- litical changes. Whether the structure of Japanese rice farming
sues in Japanese rice production. First, market-oriented rice will change as the government hopes depends highly on farm-
production is required. It is especially essential to grasp mar- ers’ responses. Therefore, it is essential to understand farmers’
ket needs and seek out new demand for rice. Second, the con- strategies in their specific socioeconomic context. Reactions
version from rice to other crops in paddy fields should be rein- to policy and market conditions differ according to farming
forced to achieve a rice supply adjustment. It is also important types. Therefore, first I describe the major farming types of
to increase the rate of food self-sufficiency. Third, the govern- rice producers. Next, I discuss their socioeconomic character-
ment needs to tackle reform of the production structure by en- istics and behavior. Finally, I conclude with the future per-
couraging large-scale farmers. spectives of Japanese rice production and the necessary con-
ditions for changes in agricultural structure.
Japanese rice policy reform and objectives of this paper
Major farming types of rice producers and their behavior
To solve these problems, the government launched a Rice
Policy Reform in April 2004. By implementing the policy, the Japan has various types of rice producers such as the part-time
government expected to cope with market globalization through family farm, large-scale family farm, agricultural corporate
a remarkable change from the existing rice supply adjustment firm, group-farming organization based on the rural commu-
program. The government abolished the set-aside program to nity, farm contractor, rural agricultural public corporation, and
achieve efficient land resource allocation for rice production agricultural cooperative. However, the following four types
54.5
45.5
27.3
Farm land size (10 ha)
18.2
Farm land size (30 ha but no subsidy)
9.1
Rice is the dominant staple food for about 220 million people less than 10% of the total rice area of Indonesia. Government
in Indonesia. Indonesia achieved self-sufficiency in rice in 1984 efforts to achieve self-sufficiency in rice during the 1980s fo-
following the intensive effort of the government through vari- cused on lowland irrigated rice. Upland rice has remained
ous public assistance programs for increasing rice production. somewhat neglected in the public-sector assistance programs
Unfortunately, rice self-sufficiency could not be maintained of Indonesia.
for long and the import volume slowly crept up over time. In In upland areas, where the opportunity costs of land and
2003, rice imports reached 2 million metric tons, which made labor are relatively low, Indonesia may still have a compara-
Indonesia the biggest rice-importing country in the region tive advantage in upland rice production. The objectives of
(Sidik 2004). this paper are (1) to evaluate the importance of upland rice
Land is the basic resource for agricultural activities. The and its relation with poverty level and (2) to discuss the likely
upland and rainfed lowland play an important role in the agri- impact of trade liberalization on upland rice production and
cultural sector of Indonesia, since a vast area of cultivated land the livelihoods of upland farmers.
(about 65.8%) is under rainfed conditions (BPS 2003). The
existence of upland is dominant on all the main islands of In-
Characteristics of upland rice production
donesia (Table 1). In Java, the proportion of upland area is
about 33.5%. However, the high incidence of rural poverty, Upland rice production is characterized by tiny percentages of
instability of production, and fragility of the environment are marketed output, low yields, and relatively low levels of use
among the important characteristics of rainfed upland areas of modern inputs. Upland rice production activities involve a
(Partohardjono 1993). large number of Indonesians. Although the contribution of
Upland rice is one important product of upland agricul- upland rice to the total national rice production is relatively
ture. In 2002, the total production of upland rice reached 2.6 low, it has an important role in supporting farmers’ livelihoods.
million tons, whereas lowland rice produced 45.2 million t. Most upland farmers cultivate upland rice mainly to meet their
Sumatra and Java islands account for the major share of up- family needs. Because of the small area of land ownership,
land rice area and production in Indonesia. Total upland rice low yield, and lack of government support, most upland rice
area of Indonesia is 1 million ha and upland rice accounts for farmers are poorer than their wetland counterparts.
Globalization makes the world smaller and affects people’s Thailand, the WTO, and IT
lives in many aspects. With the fast development of informa-
tion technology (IT), our way of life has changed drastically. Thailand and the WTO
IT has markedly changed our way of life, economics, society, As a member of the WTO, Thailand committed to the follow-
and policy worldwide. Every country has to prepare itself to ing:
cope with this change. Besides IT development, the World 1. Reduce average tariffs on 740 tariff lines by 24% in
Trade Organization (WTO), which aims to create efficient and 10 years (1995-2004), with a minimum reduction of
fair international trade, has been established. To promote agri- 10% on all tariff lines. The average tariff rate will
cultural trade, three measures, market access and tariff reduc- decrease from 49% in 1995 to 27–40% (37% on av-
tion, domestic support reduction, and export subsidy reduc- erage) in 2004 (DTN 2004).
tion, have been implemented since 1995. In general, it is be- Thailand opened its domestic rice market in two
lieved that the WTO and IT will boost the volume of agricul- ways. The first one was to set import tariff rate quo-
tural trade, which will create benefits for many people, espe- tas at 30%. The quota amounts were set at 237,863
cially farmers. However, in Thailand, this belief is questioned; and 249,757 tons in 1995 and 2004, respectively. The
therefore, it is crucial to discuss the impact of the WTO and IT second was to reduce tariffs for imported rice (more
on farmers. Because the number of rice farmers in Thailand is than quota) from 58% in 1995 to 52% in 2004. For
the most compared to other farmers in the country, this paper processed rice products, there is no tariff quota, but
will focus only on the impact of the WTO and IT on rice farm- there will be a tariff reduction and commitment to an
ers. import tariff rate.
7.35
1,846.84
6.43
2.16
4.21
9.11
1.53
251.41
117.68
6.65
19.63
2003
rice production support was US$333.03, dropping
from $354.53 in 1995 (DTN 2004).
5.62
7.33
1,629.41
20.90
2.26
4.17
9.25
1.35
222.16
108.98
5.88
2002
Thailand has been using IT for about 20 years. However, this
technology was not widely used until 1995, when the Thai gov-
ernment established its IT 2000 policy framework for IT de-
6.06
4.34
1.39
7.69
2.14
9.24
1,577.46
19.79
205.10
97.40
-4.27
velopment from 1995 to 2000, with three fundamental prereq-
2001
uisites that must necessarily be in place and function together:
(1) a national information infrastructure, (2) a well-educated
populace and adequate IT labor, and (3) good governance with
5.16
4.10
1.26
2.10
9.05
6.14
19.02
1,631.38
265.64
114.99
6.20
a dare to dream and a resolve to act. The termination of IT
2000
2000 was followed by a second, called IT 2010, which was
approved by the cabinet in March 2002. The IT 2010 policy
framework, implemented from 2001 to 2010, set a higher goal
4.34
4.20
1.03
18.66
2.07
9.00
6.84
1,950.59
285.22
143.33
27.16
of transforming the Thai economy into a knowledge-based
1999
economy (Krutkaew 2004). The Ministry of Information Tech-
nology and Communication has been established to take these
Year
responsibilities. An IT master plan has also been set up to pro-
4.79
4.14
1.16
18.79
2.06
9.11
6.54
2,098.29
320.83
159.63
61.12
1998
mote widespread IT use. Consequently, IT has been used widely
in the last 2–3 years.
4.55
4.42
1.03
17.78
1.94
9.17
5.57
2,074.85
372.68
166.32
40.87
The impact of the WTO and IT on Thai rice farmers
It was expected that Thailand should have been able to in- 1997
crease its amount of rice exports because its climate is suitable
4.29
4.51
0.95
5.46
17.73
1.93
9.19
365.96
197.05
1,998.22
46.59
for rice cultivation. Many millions of tons of rice are left over
1996
2.95
4.28
0.69
1,948.19
18.16
2.01
9.02
314.33
161.74
barriers. This is a good chance for Thailand to increase its 17.10
1995
1.97
3.96
0.50
16.48
1.83
8.98
320.06
149.62
15.54
amount of rice exports has obviously increased. Even though Cultivated area (million ha)
it fell in some years, it climbed to over 7 million tons since
Total product of paddy (t)
2001 and in 2003 the amount of exported rice was 7.343 mil-
Value (million US$)
Amount (million t)
lion tons. It can be seen that income from rice exports has
Farm price ($ t–1)
Milled rice export
Wet season
Wet season
Wet season
Dry season
Dry season
Dry season
China is the largest country that produces rice in the world and temporary shortage in a bad harvest year, since the end of the
also the largest country that consumes rice. China’s rice pro- 1990s. Thanks to the good rice harvest in the second half of
duction plays a very important role not only in China’s grain the 1990s, the surplus grain supply gave the Chinese govern-
economy but also in the world rice economy. ment a good chance to smoothly implement policies of adjust-
Table 1 shows that China produces more than 30% of ing the agricultural production structure, including the rice
the rice with around 20% of the world’s paddy area. This im- production structure, to meet consumers’ diversified demand.
plies that the yield of rice in China is higher than the average Figure 1 tells us that we can categorize rice production
in the world. However, these shares have been going down. and consumption into three stages. The first stage was the pe-
This downward trend mainly results from the Chinese riod before the mid-1980s, when production and consumption
government’s policy of structural adjustment of rice produc- rose rapidly and in the same phase. During that period, pro-
tion to enhance the production of high-quality rice while re- duction and consumption were almost balanced. The second
ducing low-quality rice in response to the change in consum- stage was the period from the mid-1980s to the end of the
ers’ behavior in rice consumption. Chinese demand for rice 1990s, when production rose with a large fluctuation and con-
consumption is shifting from quantity to quality as economic sumption went up smoothly. During that period, the rice sur-
development occurs. Owing to the progress of science and tech- plus and deficit switched several times, with a large surplus in
nology for agricultural production, China’s grain supply ver- the end. The third stage was the period starting at the begin-
sus demand has changed from chronic shortage to surplus in a ning of the 2000s, when both production and consumption went
good harvest year, keeping a balance in a normal year and a down, but production at a faster speed. Fortunately, the trend
(000 t)
150,000
Production
140,000 Consumption
130,000
120,000
110,000
100,000
90,000
2000
2002
1992
1994
1996
1998
1982
1984
1986
1988
1990
1980
Year
Table 3. Cultivated area and yield of early rice, late rice, and single-harvest rice.
Early Late Middle rice and Early Late Middle rice and
rice rice one-season late rice rice rice one-season late rice
Data source: China Rice Research Institute, 1993. Agricultural Yearbook of China, 1990-1995.
Wrap-up of Session 18
The theme of session 18 was the “Impact of globalization on rice tion from the perspective of farmers,” Roosiana, Indonesia), and
farmers.” We focused on farmers’ response to the progression of the classification of areas against climatic shocks (“El Niño sen-
globalization. Globalization is usually treated as a matter of WTO sitivity, resource endowment, and socioeconomic characteristics:
agricultural negotiations on agricultural economics. Globalization the case of wetland rice in Java, Indonesia,” S. Yokoyama, Ja-
is the liberalization and deregulation of the market system, but it pan).
also fosters the interactions among policy, the economy, infor- These analyses indicated that current problems in rice pro-
mation, and culture of each country. Especially, the information duction and the impacts of globalization on rice farmers really
technology (IT) revolution will strongly affect agricultural produc- differed among countries and areas. They also indicated that farm-
tion. Therefore, it is important to consider globalization from a ers were undertaking regionally specific behavior regarding their
wider perspective. socioeconomic circumstances. Questions and discussion for these
The session focused on three major topics: presentations mainly concentrated on how researchers grasped
1. What are the major farming types of rice farmers (rice that diversity. A statistical approach and field research were ap-
producers) and what are their socioeconomic charac- plied but a detailed farming survey will be indispensable for an
teristics in each country? empirical study on rice production and rice producers’ behavior.
2. How do rice farmers face changes in economic condi- Symposium 2 had four presentations by speakers from Japan,
tions brought about by globalization, and how do they Indonesia, Thailand, and China. They reported their countries’
react to those changes? main issues on rice production and rice farmers.
3. What are the future perspectives of rice farmers in each M. Umemoto (“Behaviors and strategies of Japanese rice
country? producers under globalization”) showed the main types of Japa-
From these viewpoints, we explored the conditions of rice nese rice producers and their behaviors. Next, he analyzed the
farmers among Asian countries. impact of the rice price decline on large-scale family farms and
With this concept, we held two symposia. Symposium 1 agricultural cooperative firms. Finally, he pointed out the need
had four presentations: a comparative analysis for total factor for an effective income compensation policy in Japan.
productivity in 33 rice-producing countries from 1961 to 2002 Y. Syaukat (“Future perspectives of upland rice farmers in
(“Impacts of trade liberalization on competitiveness and produc- Indonesia in the globalization era”) analyzed the impact of glo-
tivity in the rice sector,” M. Rakotoarisoa, USA), an empirical balization on Indonesian upland rice farmers. He explained that
study on Japanese group faming (“New strategy of group farming upland rice is an important product of upland agriculture, but it
in Japan,” K. Miyatake, Japan), research on Indonesian rice farm- has remained somewhat neglected in public-sector assistance
ers under globalization (“The role of the rice economy after the programs. Then, he concluded that only the poorest of the poor
implementation of agricultural policy reform and trade liberaliza- who are dependent on upland rice for their livelihoods are likely
It is predicted that the carbon dioxide concentration (CO2) in response (R) to [CO2] under optimum temperatures revealed
the atmosphere will double within this century the concentra- that the relative response with yield at 353 µmol mol–1 with
tion at the beginning of the last century and that the increases unity can be well approximated by the following equation:
in [CO2] and other greenhouse gases will cause global warm-
ing of 1.5 to 5.8 °C at the end of this century (IPCC 2001). R = a (Ca – Co)/{K + (Ca – Co)} (1)
Such global environmental change is considered to have an
enormous impact on Asian rice production, so that prediction where Ca is ambient [CO2], Co is [CO2] at which R becomes
of its impact by region is important not only for future food zero, and a and K are empirical constants. Indica rice geno-
security but also for the sustainable development of rice-farm- types showed a higher yield response to [CO2] than japonica
ing societies in Asia. Many experimental studies (Baker 2004, genotypes (Fig. 1). The values of a, K (µmol mol–1), and Co
Horie et al 2000) and simulation studies (Matthews et al 1995, (µmol mol–1) for indica genotypes were 1.89, 288, and 30.6,
Horie et al 1995) have been made to clarify elevated [CO2] respectively, and those for japonicas were 1.39, 124, and 30.6,
and increased temperature effects on rice. On the basis of these respectively. Thus, relative yield increases at 700 µmol mol–1
previous studies and a multilocational field experiment (Horie are estimated to be 32% for indicas and 17% for japonicas.
et al 2003) covering wide-ranging rice genotypes and envi- With the increase in daily maximum temperature aver-
ronments in Asia, we developed a new rice model for simulat- aged over flowering period above about 36 ºC, rice yield gen-
ing growth and yield of different genotypes under variously erally declined because of spikelet sterility induced by high
different environments (Yoshida et al 2004, Yoshida and Horie temperatures. Importantly, elevated [CO2] increased spikelet
2005, Horie 2005). susceptibility to high-temperature damage (Kim et al 1996).
This paper briefly reviews important findings from the Nearly doubled [CO2] decreased the threshold temperature for
previous studies on [CO2] and temperature effects on rice, then high-temperature damage of spikelets by 1–2 ºC more than
explains the structure of our rice model and results of model the ambient [CO2].
simulation on rice yield at different locations in Asia under
increased [CO2] and temperature conditions, and discusses the
Simulation of rice growth and yield under elevated [CO2]
implications of the simulation results for future Asian rice pro-
and increased temperatures
duction.
The model used for the simulation
Many attempts have been made to predict global climate change
Rice responses to [CO2] and temperature: a review effects on rice by rice growth simulation models (Matthews et
To understand rice responses to [CO2] and temperature, ex- al 1995, Horie et al 1995). However, the models employed for
tensive research has been done using experimental facilities prediction did not consider the important interacting effect of
of closed chambers (Allen et al 1995), open-top chambers [CO2] and temperature on high-temperature-induced spikelet
(Ziska et al 1997), and temperature gradient chambers (Horie sterility, nor genotypic differences in the responses to [CO2]
et al 2000) since 1980. Since results of these studies have been and temperature that are described above. Considering these
fully reviewed in Horie et al (2000), only important findings insufficiencies of existing rice models, we synthesized a physi-
from these studies for model synthesis are summarized here. ologically based rice growth and yield simulation model named
Those previous studies showed that elevated [CO2] acceler- GEMRICE (Genotype by Environment simulation Model for
ated rice development; nearly doubled [CO2] increased leaf RICE). Since important components of this model have been
photosynthesis by 30–70%, canopy photosynthesis by 30–40%, reported in Yoshida et al (2004 and 2005) and Horie et al
and crop biomass yield by 15–30%, depending on genotype (1995), we describe only briefly the structure of GEMRICE.
and environment; and that elevated [CO2] had a minor effect GEMRICE consists of six model components for simu-
on rice nitrogen (N) uptake, which appeared to be associated lating processes of phenological development, photosynthe-
with the relatively insensitive response of leaf area growth to sis, respiration, biomass growth, spikelet number determina-
[CO2]. Those rice responses to [CO2] resulted in a substantial tion, and grain filling. The phenological development process
grain yield increase under elevated [CO2] and nearly optimum component is the same as the previous model SIMRIW (Horie
temperature conditions. Analysis of reported data on rice yield et al 1995). The photosynthesis component simulates leaf pho-
Ueda et al (2000)
Kim et al (1996)
Kim et al (2003) Indica
Huang et al (2002) R2 = 0.85
1.5 Allen et al (1995)
Baker (2004)
Ziska et al (1997)
Ueda et al (2000)
1.0
Japonica
R2 = 0.77
0.5
Ca
0
0 200 400 600 800 1,000
Ca=CO2 concentration (mmol mol1)
Fig. 1. [CO2]-yield response curves of indica (open symbols) and japonica (closed symbols)
rice genotypes grown under nearly optimum temperatures. The figure was drawn with experi-
mental data from various reports shown in the figure.
tosynthetic rate as a function of radiation flux density, tem- spikelet number sterilized). Source size is the sum of the cur-
perature, and [CO2], in which stomatal conductance and leaf rent daily photosynthesis and amount of nonstructural carbo-
N content are given as genotype-dependent plant parameters. hydrates stored in the plant.
Canopy photosynthetic rate is given by integrating leaf photo- In the current version of GEMRICE, LAI and plant N
synthetic rate with respect to leaf area index (LAI), taking into content are given as external parameters by assuming that ef-
account solar radiation flux density in the canopy as a func- fects of [CO2] and temperature on those parameters are rela-
tion of LAI. Respiration consists of growth and maintenance tively small.
respirations, both of whose rates are given as functions of plant
N content. The maintenance respiration rate is also a function Estimation of parameter values and model validation
of temperature. The photosynthetic and respiration rates are The model GEMRICE contains 49 plant parameters to describe
both calculated on an hourly basis by inputting hourly tem- rice growth and yield formations, of which 37 are genotype-
perature and solar radiation, which were estimated from their independent and 12 genotype-dependent. To estimate values
daily values. of these parameters and validate the model, a cross-locational
Crop biomass growth rate (CGR) is calculated on a daily field experiment named the Asian Rice Network Experiment
basis by subtracting daily respiration from daily photosynthe- (ARICENET, Horie et al 2003) was conducted on growth and
sis. The number of spikelets per unit area is calculated as the yield of nine common genotypes. The experimental sites were
difference between the spikelet number differentiated and that eight locations: Iwate, Nagano, Kyoto, and Shimane in Japan;
degenerated. The differentiated spikelet number is determined Nanjing and Yunnan in China; and Chiang Mai and Ubon
by multiplying plant N content at 2 wk before full heading by Ratchathani in Thailand, covering 39N (Kitakami, Iwate) to
a genotype-dependent coefficient, and the degenerated spike- 15N (Ubon Ratchathani). The nine genotypes used included
let number by a function of CGR during the 2-wk period be- representative cultivars from indica, temperate and tropical
fore full heading. Spikelet sterilities caused by both cool tem- japonicas, and a glaberrima-sativa hybrid (NERICA). The
perature at the meiosis stage and high temperature during flow- experiment was conducted in 2001 and 2002 at the respective
ering are simulated by the respective functions of tempera- sites.
ture. The [CO2] effect on high-temperature-induced spikelet Approximately half of the ARICENET data were used
sterility is also considered. Grain-filling rate is calculated as a for parameter estimation and the remainder for model valida-
function of sink and source size and temperature. Sink size is tion. GEMRICE with parameter values thus estimated could
given by a logistic function of the current grain yield and the simulate spikelet number per unit area, biomass, and grain yield
maximum sink size, which in turn is a function of effective of ARICENET data with R2 values of 0.86, 0.80, and 0.87,
spikelet number per unit area (= spikelet number produced – respectively (Yoshida et al 2003).
Rice yield has increased dramatically with the development of Monitoring crop growth and development
improved nursery techniques, new varieties resistant to lodg- using a rice crop model
ing and diseases, and new yield-enhancing agrochemicals.
The approach involved using the development stage and spike-
Although rice farmers commonly attain high yield, yield fluc-
let sterility and yield models, and crop and geo-climatic data
tuations sometimes occur because of unfavorable weather con-
sets to develop a prediction system.
ditions during the rice-cropping season in particular areas. A
Based on the development stage (DVS) model designed
recent study shows that summer weather in northern Japan since
by Horie (1987), values of 0, 1, and 2 are assigned to emer-
1982 appears to exhibit a distinct five-year cycle, with a pres-
gence (seed germination), heading, and physiological matu-
sure difference between Wakkanai in Hokkaido and Sendai in
rity, respectively. The value of DVS at any point of crop de-
Tohoku. The temperature also follows similar five-year cycles
velopment is calculated by integrating the developmental rate
and the second year of each cycle is typical of the Yamase, a
(DVR), expressed as a function of daylength and daily mean
cold northeasterly wind, and is consistent with the cool sum-
temperature, with time.
mer in northern Japan (Kanno 2004). For instance, severe spike-
Generally, dry matter accumulation at any point in rice
let sterility because of cool weather that occurred during meio-
growth is proportional to the accumulated solar radiation in-
sis of the pollen mother cell to the flowering stage of the rice
tercepted by the canopy, except at the final stage of rice growth,
crop in some areas of Tohoku Region in 1980, 1988, 1993,
when DVS is close to 2. The model used to estimate daily dry
and 2003 greatly reduced yield.
matter increase was adopted from Monteith (1977).
This paper intends to describe the research approach and
The spikelet sterility model on the relationship between
initial results of wide-area monitoring of crop growth and de-
spikelet sterility and cool-temperature sensitivity of the rice
velopment in 1993 and 2003 using a rice crop model (Yajima
plant at the panicle development stage was proposed by Yajima
1996) to construct the initial version of a climatic early warn-
et al (1989), with the following equation:
ing system against cool summer damage of rice under chang-
ing climate.
3
(1:1)
2
r = 0.944
1
0 1 2 3 4 5 6 7
0 1 2 3 4 5 6 Observed yield (t ha1)
Fig. 1. Spatial distribution of estimated yield (left) and its comparison with observed yield in
2003 (right), Tohoku Region.
Table 1. Estimated sterility and yield resulting from shifting transplanting date in 2003.
tion of mean air temperature, and its deviation from climatic References
normals, as well as spikelet sterility caused by cool tempera-
Horie T. 1987. A model for evaluating climatic productivity and water
ture, were determined. A significant linear correlation was balance of irrigated rice and its application to Southwest Asia.
obtained between the estimated spikelet sterility and yield. The Southwest Asian Studies, Kyoto Univ. 25:62-71.
results suggest the importance of using a crop model to moni- Kanno H. 2004. Five-year cycle of north-south pressure difference
tor and forecast rice development stages and spikelet sterility as an index of summer weather in northern Japan from 1982
at the regional level or in areas affected by cool temperature onwards. J. Meteorol. Soc. Jpn. 82:711-724.
damage. With this method, extension staff could easily pro- Seino S. 1993. An estimation of distribution of meteorological ele-
vide information on the possible occurrence of spikelet steril- ments using GIS and AMeDAS data. Jpn. J. Agric. Meteorol.
ity in particular areas, which could enable rice growers to take 48:379-383.
necessary countermeasures to minimize the yield reduction Yajima M, Nitto A, Seino H. 1989. Estimation of percent sterility of
rice spikelets with DVS model. Abstract for 1989 Annual
from cool temperature.
Meeting of the Agricultural Meteorological Society, Japan.
p 58-59.
Yajima M. 1966. Monitoring regional rice development and cool-
summer damage. JARQ 30:139-143.
Climate variability is a major source of risk in rice production frequently in recent years. Historically, El Niño events are of-
since it greatly affects yield variation and often leads to yield ten accompanied by the occurrence of temperature extremes
losses (Pantastico and Cardenas 1980, Lansigan et al 2000). (i.e., high temperatures during unusually prolonged periods
Climate change is also expected to affect crop growth and de- that induce changes in potential evapotranspiration and soil
velopment, resulting in yield reduction (Peng et al 1995, moisture availability), resulting in severe droughts that cause
Matthews et al 1996). The stability of rice production systems reduced crop yields. Different areas across the archipelago ex-
in tropical agroenvironments depends on adaptation strategies perience varying degrees of drought.
and mitigation measures applied to cope with these events. The Philippines is often visited by an average of 20 ty-
This paper analyzes the effects of climate variability and phoons a year, which mostly occur in the eastern and northern
change on rice production in the Philippines. Climatic events portions of the archipelago. Typhoons, which are often accom-
that are important to rice growth and production are analyzed. panied by intense rainfall and strong winds, usually occur from
The paper distinguishes the effects associated with risks from June to October. These cause significant yield losses during
long-term weather aberrations and short-term weather fluctua- the wet season, the growing period for most crops.
tions. Coping strategies to adapt to or mitigate the effects of Current climate change studies (IPCC 2001) anticipate
climate variability and change are discussed. that future climate will experience an appreciable increase in
temperature, changes in the spatio-temporal distribution of pre-
cipitation, and increased frequency and intensity of extreme
Climate variability and change weather events. The frequencies of occurrence of successive
The Philippines has a tropical climate that is conducive to rice wet and dry periods such as El Niño and La Niña and tropical
production year-round. The Philippine climate is generally char- cyclones and typhoons with strong winds are expected to in-
acterized by four climatic types in terms of the relative dura- crease. In recent years, typhoons have begun to be experienced
tion and intensity of the wet and dry periods in different parts during periods when they were less expected (e.g., November
of the country. Type 1 climate has a pronounced wet period or December). Unfortunately, these events coincide with the
from May to November, and a dry period from December to period when rice and other crops are about to be harvested, or
April. Type 2 climate is characterized by no clear dry season, when the second crop has been planted.
and maximum rainfall is experienced from November to Janu-
ary. Type 3 climate is characterized by no distinct wet and dry
Effects of climate variability and change
seasons but is relatively dry from November to April. Type 4
on rice growth and productivity
climate has rainfall more or less evenly distributed throughout
the year. Climate variability affects rice growth and the various crop
Extreme climate variability conditions such as during the processes and activities in rice production. It determines plant-
occurrence of El Niño and La Niña events have greatly af- ing date, duration of crop growth, crop yield, and also man-
fected the spatial and temporal distribution of rainfall in tropi- agement practices in rice production. The occurrence of ex-
cal rice-growing areas such as the Philippines. Figure 1 shows treme climate variability such as El Niño or La Niña events
the percentile of the 12-month total rainfall in the Philippines characterized by a prolonged dry period or heavy rainfall spell
for the period April to March during El Niño years in the last coinciding with the critical stages of crop growth and devel-
four decades. Significant changes in rainfall patterns occurred opment may lead to significantly reduced crop yields and ex-
during different periods within the year. These extreme events, tensive crop losses. Statistics on rice production and produc-
with a return period of 7–10 years, have tended to occur more tivity in the Philippines (PhilRice-BAS 2000) show that the
Percentile*
≤10 Severe drought effects 41–60 Near normal to above normal >90 Severe flood damage
11–20 Drought effects 61–80 Way above normal conditions *Percentile is a way to present
variability with respect to time
21–40 Moderate drought effects 81–90 Potential flood damage
Fig. 1. Spatial and temporal variability of 12-month (April-March) rainfall during El Niño years in the Philippines. (Source: PAGASA 2000.)
observed declines in production and yield coincide with the tion and mitigation measures that involve the choice of rice
occurrence of El Niño events. Rice production has declined crop variety, adaptation of cultural management practices, and
during these extremely dry periods since crop productivity or understanding of climate science. Crop management measures
yield level has decreased, and the area planted to rice has been include a range of possible strategies such as the following
reduced to adapt to the anticipated drought period. The risk (Lansigan 2003): (1) adjusting the cropping calendar to syn-
associated with climate variability of rice production in gen- chronize crop planting and the growing period with soil mois-
eral depends mainly on the growth stage of the rice crop when ture availability, (2) changing the rice variety to plant (i.e.,
the weather aberration occurs. planting a drought-tolerant or early-maturing variety), (3) vary-
Several studies using temperature gradient tunnels, crop ing planting density, and (4) optimizing water-use efficiency
simulation models, and field experiments have been conducted by improving irrigation facilities and introducing water-sav-
to evaluate the effects of climate change on rice crop growth ing techniques.
and development, particularly temperature increase and in-
creased or double CO2 concentration (e.g., Horie 1993, Horie Crop improvement strategies
et al 1996, Matthews et al 1996). In general, simulation analy- Biological measures being pursued to cope with climate vari-
ses and field experiments have shown a reduction in rice yields ability and change include the breeding of drought-tolerant
because of the sterility of spikelets, which are very sensitive to varieties and lodging-resistant cultivars in national and inter-
a temperature increase. The same studies have also shown that national rice breeding programs. Drought-tolerant rice variet-
rice varieties respond nonlinearly with increased CO2 concen- ies can be planted that are more adapted to warmer or drier
tration (Peng et al 1995). conditions. This involves the screening and testing of rice
germplasm collections that are sources of tolerance of water
stress or that require less water for crop growth. Lodging-re-
Coping strategies for climate variability and change sistant cultivars will be useful in areas generally experiencing
Crop management to cope with climate variability more frequent and intense typhoons with strong winds and more
Inasmuch as climate variability and change are inevitable, rice precipitation.
production systems should be able to adapt to weather fluc-
tuations and climatic aberrations to minimize their negative Crop insurance for and risk from climate variability
effects. Coping with or managing climate variability and change Risks associated with reduced rice production and yield losses
in rice production systems requires a combination of adapta- caused by climate variability can be partly alleviated by pro-
Downscaling
climate forecasts
Cropping strategy
Simulating
seasonal crop
yields
Fig. 2. Schematic of a knowledge-based crop forecasting system using seasonal climate prediction, estimation of crop area, and simula-
tion of rice yields.
viding appropriate crop insurance coverage to rice farmers. conditions during the cropping period. More specifically,
Current crop insurance practices in the Philippines cover only knowledge-based crop forecasting that uses these seasonal cli-
the cost of land preparation and rice crop establishment. The mate forecasts will be useful in guiding rice farmers and con-
crop insurance policy at the moment often does not consider cerned government agencies and other stakeholders to act ac-
the varying extent of vulnerability of rice-growing areas in cordingly to minimize the negative or detrimental consequences
different locations with different climatic types as well as the of climate variability. Such a crop forecasting system currently
differences in risk from climate variability, which are site-spe- being piloted in the Philippines (Fig. 2) involves using sea-
cific and time-dependent. Although crop insurance has been sonal climate forecasts and a validated process-based rice crop
recognized as an effective coping strategy, it is less popular simulation model to predict rice crop performance under the
among rice farmers because of the usually unaffordable insur- predicted seasonal climate outlook during the growing period.
ance premium. Promotion of crop insurance to rice farmers Seasonal climate forecasts are often made at a regional
requires not only the provision of a more reasonable insurance level, and have to be downscaled to a specific location (e.g.,
premium but also attractive crop insurance coverage schemes province of a country) to produce forecasts of rice yield per-
or options such as insuring not only the cost of crop establish- formance in a given area. Crop forecasts provide useful infor-
ment but also (1) the expected crop yield or (2) the expected mation to stakeholders (e.g., farmers, extension workers, trad-
revenue based on harvestable crop. ers, and planners) in rice production systems at different lev-
els. The crop forecasting system also includes efficient and
Seasonal climate prediction and rice production effective tools and mechanisms to deliver the forecast infor-
In recent years, advances and developments in climate sci- mation to the intended users in the area.
ence, meteorology, and global data observation and monitor-
ing networks have enabled the reasonably accurate prediction
Conclusions
or real-time forecast of seasonal climate up to a 3-month lead
time. Reliable forecasts of seasonal climate serve as important The vulnerability of rice production systems to climate vari-
and crucial information to an early warning system to support ability in tropical regions depends on its time of occurrence
crop production to recommend appropriate management ac- relative to the growth stage of the crop. Advances in crop sci-
tions or measures to take considering the anticipated climatic ence, meteorology, and other sciences offer potential for knowl-
The China rice/wheat FACE (free-air carbon-dioxide enrich- The FACE treatment resulted in a significant increase in N-
ment) has been running since 2001. It is the first FACE ap- use efficiency for biomass production (NUEp), which was
proach to determine the effects of elevated CO2 on intensive measured at the 28th day after transplanting, at heading, and
agroecosystems with a rice/wheat rotation under field condi- at maturity. A significant increase in N-use efficiency for grain
tions. The FACE system followed the Japanese design with output (NUEg) and nitrogen harvest index (NHI) under the
three FACE rings, 12.5 m in diameter, and three ambient rings. FACE treatment was also observed. Nitrogen content (%) and
The experimental setup was a split-plot design with CO2 treat- N accumulation in rice plants increased significantly under
ments (two CO2 concentrations) as the main plots and N fer- high-N treatment, but N-use efficiency in rice plants decreased.
tilization as subplot treatments (two N levels). Different from Elevated CO2 significantly increased P and K uptake in
others, the China FACE system is run continuously year-round, aboveground tissues. P fertilization had no significant effect
24 h a day. The target CO2 concentration ([CO2]) at the ring on various tissue dry biomass. Leaf P concentration at joint-
center was 200 µmol mol–1 above ambient. ing, heading, and ripening increased, but no significant effect
Results showed that elevated CO2 affects nutrient up- was found on P concentration in the stem, ear, and root. P and
take and nutritional conditions of rice. Nitrogen content (%) K concentration in rice was not influenced in the same way as
in the rice plants sampled at different growth stages decreased N concentration by elevated CO2. The dilution effect is al-
significantly under FACE treatment, but N accumulation in- ways used to explain the decrease in N content led by elevated
creased slightly because of the significant enhancement in dry CO2; however, it could not be used to explain the increase in P
matter production. Ear N concentration increased at the head- content. The average soil exchangeable K in the FACE ring
ing stage but decreased at the ripening stage. No significant increased more under the NN treatment than under the LN
effect was found on root N concentration at tillering, but root treatment in the rice season. The possible increase in soil or-
N concentration at jointing, heading, and ripening decreased.
Production and emission of CH4 from rice fields, a gas about This study contains a series of experiments to character-
23 times more radiatively active than CO2, are an important ize CH4 production in three different rice topsoil and subsoil
feature of the global carbon cycle (Ramanathan et al 1985). samples and their blends in different ratios created to produce
Rice fields were identified as a major source of CH4 and the gradients in soil properties for assessing CH4 production po-
intensity of CH4 release to the atmosphere depends on several tentials influenced by these properties. Blending was done with
factors, including soil parameters (Denier van der Gon et al the assumption that, during long-term farming, topsoil and sub-
1992, Yao et al 1999, Mitra et al 1999). According to Yagi and soil get mixed with each other, creating a similar situation. To
Minami (1990), emission rates (flux) of CH4 vary according estimate production potential, a laboratory incubation study
to soil type and the organic matter applied; Bachelet and Neue was conducted estimating both entrapped and emitted CH4.
(1993) illustrated the potential importance of differentiating Production potential was tested under special circumstances
soil types when estimating CH4 emissions from rice fields. such as sterilization followed by inoculation, straw incorpora-
pH Active Fe Active Mn Available P Total N Total C Clay Silt Sand CH4 productiona (mean ± SD)(mg kg–1 soil)
Soils (1:1 H2O) (%) (%) (mg kg–1) (g kg–1) (g kg–1) (%) (%) (%)
Control Sonication Straw app. Sterilization
Topsoil
Subsoil
Luisiana 5.3 4.31 0.29 3.1 0.58 0.57 51 43 6 0.1 ± 0.1 a 0.2 ± 0.07 a 1,996.7 ± 282.8 c 0.3 ± 0.1 a
Maahas 6.8 3.13 0.29 3.5 0.28 2.60 43 54 3 0.3 ± 0.1 a 0.1 ± 0.06 a 1,363.3 ± 291.9 b 0.9 ± 0.7 a
Pila 7.1 0.50 0.10 19 1.85 1.94 44 47 9 0.9 ± 0.2 a 0.4 ± 0.3 a 734.5 ± 58.2 a 4.0 ± 0.6 a
aValues followed by the same letters are not significantly different from each other at the 5% level of significance according to Duncan’s multiple range test (DMRT). DMRT was performed separately for CH methane production in
4
different treatments.
CH4 (mg g1 soil d1)
Control Control
25 0.06
A E
20
0.04
15
10
0.02
5
0 0.00
0 20 40 60 80 0 20 40 60 80
Sonication Sonication
50 0.16
C G
40 0.12
30
0.08
20
0.04
10
0 0.00
0 20 40 60 80 0 20 40 60 80
Sterilization Sterilization
80 0.16
PI D H
60 LU 0.12
MA
40 0.08
Inoculated Inoculated
20 0.04
0 0.00
0 20 40 60 80 0 20 40 60 80
DOI
Fig. 1. Methane production in topsoils (A,B,C,D) and subsoils (E,F,G,H) as influenced by
sterilization and subsequent inoculation, sonication, and straw incorporation. DOI = days
of incubation; PI = Pila, LU = Luisiana, and MA = Maahas. Please note different scaling on
Y axis.
Since the early 1980s, water management of rice paddies in tilization, manure amendment, irrigation, weeding, and graz-
China has changed substantially, with midseason drainage ing) have been parameterized and linked to the various bio-
gradually replacing continuous flooding (Shen 1998). This has geochemical processes (e.g., crop growth, litter production,
provided an opportunity to estimate how management alterna- soil water infiltration, decomposition, nitrification, denitrifi-
tives affect greenhouse gas emissions at a large regional scale. cation, etc.) embedded in DNDC (Fig. 1). To enable DNDC to
Both methane (CH4) and nitrous oxide (N2O) fluxes from simulate C and N biogeochemical cycling in paddy rice eco-
agroecosystems are highly variable in space and time, affected systems, we modified the model by adding a series of anaero-
by ecological drivers (e.g., climate, vegetation, and anthropo- bic processes. The paddy-rice version of DNDC has been de-
genic activity), soil environmental factors (e.g., temperature, scribed and tested in recent manuscripts (Cai et al 2003, Li et
moisture, pH, redox potential, and substrate concentration gra- al 2004), and is summarized briefly here.
dients), and biochemical or geochemical reactions (Li 2000, The Nernst equation, a basic thermodynamic formula
Li et al, 2004). Process-based models are used to quantify trace defining soil redox potential (i.e., Eh) based on concentrations
gas fluxes driven by the local climate, soil, vegetation, and of the oxidants and reductants existing in the soil liquid phase
management conditions at the site scale. Geographic informa- (Stumm and Morgan 1981), and the Michaelis-Menten equa-
tion systems (GIS) databases provide spatially differentiated tion, a widely applied formula describing the kinetics of mi-
information on climate, soil, vegetation, and management to crobial growth with dual nutrients (Paul and Clark 1989), were
drive the model runs across the region. To quantify the effects linked to each other in DNDC to simulate variation of soil Eh
of water management change on carbon (C) sequestration and and its impact on rates of the redox reactions, which produce
CH4 and N2O emissions from approximately 30 million hect- or consume CH4 or N2O during the frequent changes between
ares of rice paddies in China, we integrated a process-based saturated and unsaturated conditions driven by water manage-
model, DNDC, with a GIS database of paddy area, soil prop- ment. CH4 or N2O is produced or consumed under certain Eh
erties, paddy management (fertilizer use, water management, conditions (–300 to –150 mV for CH4, and 200 to 500 mV for
crop residue management, planting and harvest dates), and daily N2O production), so the two gases are produced during differ-
weather data. ent stages of the varying soil redox potential. The major sub-
DNDC was originally developed for predicting C se- strates controlling production of CH4 or N2O are dissolved
questration and trace gas emissions for nonflooded agricul- organic carbon (DOC), ammonium, and nitrate. DNDC allo-
tural lands, simulating the fundamental processes controlling cates the substrates to drive nitrification, denitrification, and
the interactions among ecological drivers, soil environmental fermentation based on a simple kinetic scheme, the so-called
factors, and relevant biochemical or geochemical reactions, “anaerobic balloon.” An anaerobic balloon is defined as the
which collectively determine the rates of trace gas production effective anaerobic volumetric fraction of a soil, which is quan-
and consumption in agricultural ecosystems (Li et al 1992, tified by the Eh value as calculated with the Nernst equation.
1994). Details of management (e.g., crop rotation, tillage, fer- When the substrates are allocated within the balloon, they will
Soil
environmental Temperature Moisture pH Eh Substrates: NH4+, NO3, DOC
factors
participate in the reductive reactions (e.g., denitrification, database of crop-sown areas with a Landsat TM-derived land-
methanogenesis); when the substrates are allocated outside of cover map for all of mainland China (Frolking et al 2002).
the balloon, they will be involved in oxidative reactions (e.g., The database also includes soil texture, pH, bulk density, and
nitrification, methanotrophy). When a soil is flooded, oxygen organic carbon content, and daily weather data for 1990 from
diffusion into the soil is severely restricted, which reduces the 610 weather stations across China. General data on fertilizer
soil Eh and causes swelling of the anaerobic balloon. The en- use, tillage, planting and harvest dates, crop residue manage-
larged balloon will hold more DOC or nitrate to stimulate ment, and crop varieties were taken from various sources. Two
methanogenesis or denitrification. This mechanism has been water management scenarios, continuous flooding (CF) and
embedded in DNDC to link the soil water regime to trace gas midseason draining (MSD), were used in the model simula-
emissions for rice paddy ecosystems. The modified DNDC tions.
model has been tested against several methane flux data sets The modeled results indicated that the ranges in total
from wetland rice sites in the United States, Italy, China, Thai- CH4 emissions with CF and MSD management were 6,400–
land, and Japan (Li et al 2002, Cai et al 2003). The test results 12,000 and 1,700–7,800 Gg CH4-C y–1, respectively. Taking
indicate that DNDC is capable of estimating the seasonal pat- the mean values, changing water management from continu-
terns and magnitudes of CH4 and N2O fluxes from the sites ous flooding to midseason draining caused a reduction in ag-
although discrepancies exist for some of the tested cases. gregate methane emissions from rice paddies of about 40%, or
A GIS database was constructed to hold the 30 million 6 Tg CH4 y-1. This is 5–10% of total global rice paddy meth-
ha of rice paddies with different rotation systems at the county ane emissions (Prather et al 2001). The mitigating effect of
scale in China. The area occupied by each rotation in each midseason draining on CH4 flux was highly uneven across the
county was quantified by combining a county-scale statistical country. The highest flux reductions (>200 kg CH4-C ha–1 y–1)
Sichuanc 11,000 1,500 240 1,600 –380 1,800 1,300 1,300 200 49 34 72 47
Hunan 12,000 370 –1,700 280 –2,000 1,100 550 580 110 38 32 54 41
Jiangsu 9,000 1,800 1,400 1,900 1,300 1,200 770 1000 480 25 18 36 25
aProvinces sorted by paddy area. Tibet and Qinghai provinces excluded because of small paddy areas. bMean annual rough rice production for 1991-2000 (IRRI 2004); carbon fraction of dry weight
assumed to be 50%. cIncludes Chongqing Province, which was established from land in Sichuan Province in 1997. dProduction data (see note 2 above) not reported individually for some provinces; total
production for all nonreported provinces = 370 Gg C y–1; China total equals sum of all reported and nonreported provinces.
Impact of rising atmospheric CO2 on CH4 emissions
from rice paddies
Weiguo Cheng, Kazuyuki Yagi, Kazuyuki Inubushi, Kazuhiko Kobayashi, Hidemitsu Sakai, Han-Yong Kim, and Masumi Okada
Recent anthropogenic emissions of key atmospheric trace gases trolled to follow ambient air temperature. In the 1998 and 1999
(e.g., CO2, CH4, N2O, and CFC) that absorb infrared radiation rice growth seasons, rice straw was applied to the soil at 300
may lead to an increase in global mean surface temperature, and 150 g m–2, respectively. The CH4 emitted from the cham-
and thus have the potential to bring about changes in climate bers was determined by an automated monitoring system dur-
(IPCC 2001). Depending on population growth and energy- ing all rice growth seasons.
use scenarios, atmospheric CO2 concentration is expected to To understand how elevated CO2 affects CH4 emissions
rise from 370 ppm currently to about 485 to 1,000 ppm by by plant-mediated transport or ebullition diffusion through
2100 (IPCC 2001). CO2 and CH4 are two important green- flooding water, and how drainage affects CH4 emissions in
house gases and they are responsible for approximately 55% elevated CO2 conditions, a pot experiment in the Climatron
and 20%, respectively, of the anthropogenic global warming was carried out in the 2002 season. No rice straw was applied.
effect. Most atmospheric CH4 is produced by microbial ac- The flooding water was maintained at a 5-cm depth until 2
tivities in extremely anaerobic ecosystems, such as natural and September (3 weeks before harvest); then, all pots were drained,
cultivated wetlands, sediments, sewage, landfills, and the ru- and the water level was maintained at 5 cm below the soil
mens of herbivorous animals. The methane from rice paddies surface. A cylinder and a doughnut chamber were used to mea-
accounts for about 17% of the anthropogenic sources. Also, sure CH4 flux with and without plants.
plant-mediated transport is a very important pathway for CH4
emissions from cultivated rice soils. Many studies have dem- FACE experiment
onstrated a positive effect of elevated CO2 on rice biomass The FACE experiments with rice plants were conducted at
(above- and belowground) production and grain yield. The Shizukuishi, Iwate, Japan. The soil was an Andosol paddy soil.
direct effect of elevated CO2 on rice root biomass and tiller A randomized complete block design with two levels of target
number would be expected to indirectly affect CH4 produc- atmospheric CO2 concentration (ambient CO2 and FACE =
tion and emissions from rice fields, although such an effect ambient + 200 µmol mol–1) and four replicate blocks was used.
has not been identified previously. Therefore, we carried out a The rice (Oryza sativa) test cultivar was Akita Komachi. CH4
controlled-environment chamber experiment at Tsukuba and a flux was determined with a cylindrical, transparent, acrylic,
free-air CO2 enrichment (FACE) experiment at Shizukuishi, closed-top chamber. The samples were collected and measured
Iwate, Japan, for a series of rice growth seasons, in order to once every 2 weeks. The total seasonal emissions from each
understand the influence of elevated CO2 on CH4 emissions plot for both the 1999 and 2000 seasons were calculated and
from rice paddies. analyzed statistically.
+ +
+
O2 diffused to
rhizosphere CH4
CO2 + CH4 emissions
(by aerenchyma) oxidation
+ + +
The size of the arrows in the last steps indicated potential changes during all the rice growth period.
Fig. 1. A conceptual model depicting how elevated CO2 affected CH4 emissions from rice
paddies.
effect of elevated CO2 on increasing total CH4 emissions in In conclusion, elevated CO2 affects CH4 emissions from
the 1999 rice growth season was much greater than that in rice paddies both by increasing the net balance between CH4
1998 because of lesser amounts of rice straw applied in 1999 production and oxidation and by increasing the pass-way
(Yagi et al 2000). through the rice plant (Fig. 1). Several studies were made on
The results from a pot experiment under controlled-en- the effect of elevated atmospheric CO2 concentration on CH4
vironment chambers during the 2002 rice growth season emissions from wetland and paddy soils (Table 1). CH4 from
showed that the increase in CH4 emissions caused by elevated rice paddies and natural wetlands accounts for 32.7% of the
CO2 was significant after the grain-filling stage of rice; the total global sources to the atmosphere. If we consider that CH4
total emissions were enhanced 58% by elevated CO2. The CH4 emissions from natural wetlands and rice paddies would be
emitted by ebullition-diffusion accounted for 11.3% and 11.9% enhanced by elevated CO2, the global warming potential of
of total emissions under ambient and elevated CO2 conditions, CH4 could become more important as atmospheric CO2 in-
respectively. In contrast, no CH4 was emitted from plant-free creases, as was recently predicted (Lelieveld et al 1998).
pots, suggesting that the CH4 emitted from rice-plant pots was
most contributed by the rice plant through root exudates and
References
root autolysis products. The CH4 flux decreased when the
flooding water was drained under both ambient- and elevated- Cheng W, Yagi K, Sakai H, Kobayashi K. 2003. Effects of elevated
CO2 treatments, implying that drainage management will have CO2 on CH4 and N2O emissions from submerged rice soil: a
an important role in mitigating future CH4 emissions from pot experiment. 2003 ASA-CSSA-SSSA Annual Meetings, 2-
paddy fields, when more CH4 will likely be emitted from 6 November 2003, Denver, Colorado, USA. CD-ROM.
Dacey JWH, Drake BG, Klug MJ. 1994. Stimulation of methane
flooded rice paddy soils under increasing atmospheric CO2
emission by carbon dioxide enrichment of marsh vegetation.
concentration (Cheng et al 2003).
Nature 370:47-49.
When free-air CO2 was enriched to 550 µmol mol–1, the Hutchin PR, Press MC, Lee JA, Ashenden TW. 1995. Elevated con-
CH4 emissions from the rice paddy field increased significantly, centrations of CO2 may double methane emissions from mires.
by 38% in 1999 and 51% in the 2000 season. Net CH4 emis- Global Change Biol. 1:125-128.
sions are determined by the balance between methane produc- Inubushi K, Cheng W, Aonuma S, Hoque M, Kobayashi K, Miura S,
tion and methane oxidation in the rice plant and paddy soil Kim H, Okada M. 2003 Effects of free-air CO2 enrichment
ecosystems. The CH4 production potential in the FACE was (FACE) on CH4 emissions from a rice paddy field. Global
significantly larger than that in the ambient-CO2 treatments at Change Biol. 9:1456-1464.
the flowering stage in the middle rice growth period because IPCC. 2001. Climate change 2001: the scientific basis. Cambridge
elevated CO2 increased CH4 production by root exudates and (UK): Cambridge University Press.
Lelieveld J, Crutzen PJ, Dentener FJ. 1998. Changing concentra-
root autolysis products. However, elevated CO2 had no net
tion, lifetime and climate forcing of atmospheric methane.
effect on CH4 oxidation activity during the same period. Also,
Tellus 50B:128-150.
the elevated CO2 increased rice tiller numbers and CH4 trans-
port through the rice plant in FACE treatments (Inubushi et al
2003).
Wetlands
Tidal marsh, open-top chambers 80 Dacey et al (1994)
(690 ppm)
Ombrotrophic mire, open-top chambers 145 Hutchin et al (1995)
(550 ppm)
O. aquaticum grew in wetland, glasshouse 136 Megonigal and Schlesinger (1997)
(700 ppm)
Paddy soils
Rice (IR72), IRRI open-top chambers 49–60 Ziska et al (1998)
(650 ppm)
Rice (Nipponbare), Climatron (650 ppm) 23–145 This research (Yagi et al 2000)
Rice (Akitakomachi), FACE (550 ppm) 38–51 This research (Inubushi et al 2003)
Rice (Nipponbare), pot in Climatron 58 This research (Cheng et al 2003)
(650 ppm)
Megonigal JP, Schlesinger WH. 1997. Enhanced CH4 emissions from Notes
a wetland soil exposed elevated CO2. Biogeochemistry 37:77-
88. Authors’ addresses: Weiguo Cheng, Kazuyuki Yagi, and Hidemitsu
Yagi K, Li Z, Sakai H, Kobayashi K. 2000. Effect of elevated CO2 Sakai, National Institute for Agro-Environmental Sciences,
on methane emission from a Japanese rice paddy. FACE 2000 Tsukuba 305-8604, Japan; Kazuyuki Inubushi, Faculty of
Conference, 27-30 June 2000, Tsukuba, Japan. p 40. Horticulture, Chiba University, Matsudo, Chiba 271-8510,
Ziska LH, Moya TB, Wassmann R, et al. 1998. Long-term growth at Japan; Kazuhiko Kobayashi, Graduate School of Agricultural
elevated carbon dioxide stimulates methane emission in tropi- and Life Sciences, the University of Tokyo, Tokyo 113-8657,
cal paddy rice. Global Change Biol. 4:657-665. Japan; Han-Yong Kim, Department of Agronomy, Chonnam
National University, Gwangju 500-757, Korea; Masumi
Okada, National Agricultural Research Center for Tohoku
Region, Morioka 020-0198, Japan, e-mail:
cheng@niaes.affrc.go.jp.
The atmospheric carbon dioxide (CO2) concentration is in- plants. Thus, O3 and CO2 have opposite effects on production
creasing year by year because of human activities such as fos- processes of crop plants. However, investigations on the inter-
sil fuel combustion and deforestation. The elevated levels of action of these two gases in lowland rice are limited (Ishioh
CO2 promote photosynthesis of C3 plants. For example, in low- and Imai 2003, 2004, Kobori and Imai 2003, Olszyk and Wise
land rice, photosynthesis and biomass production are promoted 1997).
and plant development is accelerated by elevated CO2. More In our study, we exposed a japonica lowland rice culti-
than a 20% yield increment of rice is often reported under var to O3 and CO2 at the vegetative growth stage and at the
doubled CO2 conditions (Imai et al 1985, Kim et al 1996). On heading to early-maturity stage and examined plant responses
the other hand, atmospheric ozone (O3), a major component over short (3 h) and long (2 wk) terms to clarify the interaction
of photochemical air pollutants, is often generated around ur- of these two gases.
ban areas during the sunny summer season in Japan. Ozone
causes injury, with visible symptoms, growth inhibition, and
Materials and methods
yield decline in a range of field crops. From now on, under
further industrialization and global warming, the frequency of A lowland rice cultivar (Oryza sativa L. cv. Koshihikari) was
occurrence of O3 above the environmental quality standard grown outdoors in small plastic pots containing 2.5 kg of soil
(0.06 ppm in Japan) will increase and detrimentally affect crop and 4 g of compound chemical fertilizer (NPK=15:15:15, %)
Outdoors Outdoors
In long-term
1-1 Flowering to early maturing
1-2
1-3 Early maturing
35
30
25
20
15
10
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 4 7 11 14 18 21
Days of treatment
Fig. 2. Effects of O3 and CO2 concentrations on net photosynthetic rate and stomatal conduc-
tance at vegetative growth stage (experiment 1, long-term; 1-4).
= 0 ppm O3 + 400 ppm CO2 plot, = 0.1 ppm O3 + 400 ppm CO2 plot,
= 0.1 ppm O3 + 800 ppm CO2 plot, = 0 ppm O3 + 800 ppm CO2 plot.
High temperatures during the rice-ripening period significantly was 27.7 ºC, which was 5–6 ºC higher than the optimum tem-
diminish grain dry weight and increase grain damage (Nagato perature.
and Ebata 1960, Yoshida and Hara 1977, Tashiro and Wardlaw To determine yield, yield components, and grain qual-
1991, Morita et al 2004). It is thought that unusually high tem- ity, 48 hills (2.3-m2 area) from each replication were harvested
peratures in recent years have caused a reduction in first-grade at maturity (2 October). Yield and 1,000-grain weight were
quality rice in Japan (Terashima et al 2001). The major factors recorded as filled grain (with a thickness of 1.7 mm or more)
in grade reduction were an increase in immature grains with a weights of brown rice adjusted to 15% grain water content.
white portion, deep ditch on the grain surface, and thin shape. Grain-filling percentages were calculated as 100 × filled grain/
Recently, it has been suggested that these adverse effects of total spikelet number. Filled grains were graded by the inspec-
high temperature on rice ripening have been accelerated by tion system, and grain quality was measured by a grain quality
the reduced topdressing during the panicle formation stage inspector (RGQI 10A, Satake, Japan). Nitrogen contents of
designed to improve edibility (Terashima et al 2001). filled grains were measured by the Japan Grain Inspection
The objective of this study was to demonstrate the effect Association (Tokyo, Japan).
of topdressing at the panicle-formation stage on grain dry For grain shape analysis, eight grains (3rd, 4th, 5th, and
weight and the occurrence of immature grains with a white 6th grains from the top on the 3rd and 4th rachis-branch in
portion, deep ditch, and thin shape under high temperature. each panicle) of one panicle per plot were sampled. The
Recently, immature grains with a white portion have been iden- samples were dehulled with a forceps and cut in half along the
tified by grain quality inspectors during the inspection system. cross section. The grain images, viewing from the grain top or
However, immature grains with a deep ditch and thin shape bottom, were digitized through a digital microscope at 50 mag-
were not detected by machines. To enable a numerical com- nifications. After transforming the digitized 24-bit full color
parison, we developed a new method of determining grain shape images into binary ones, an image analysis was performed to
using digitized images. measure grain contour with rθ polar barycentric coordinates (r
indicates the distance, %, from the center point, half of X and
Y radius, respectively, in the grain image, and θ denotes the
Materials and methods angles of direction toward the vascular bundle). θ angles were
Experiments were performed in fields of the National Agricul- within 0–180 degrees (i.e., numbers and areas of the cells at
tural Research Center for Kyushu Okinawa Region at Chikugo 90 degrees were the averages of those at 270 degrees and 90
(33°12′N, 130°30′E), Japan. The seedlings of rice (Oryza degrees), since the top and bottom of a grain image showed no
sativa L. cv. Hinohikari) growing 4 weeks after sowing were distinction from one another.
transplanted to paddy fields (gray lowland soil) on 18 June
2003. Planting density was 20.8 hills m–2 (30 cm × 16 cm),
Results and discussion
with 3 seedlings per hill. An experiment was established in a
randomized block design with three replications (11 m2 per Yield, yield components, and grain quality
replication). A basal dressing of 6 g N m–2 (commercial NPK Grain yields were higher at higher levels of N topdressing be-
(16:16:16) fertilizer) was applied in all plots. Topdressing at cause of higher grain numbers per unit area and higher 1,000-
18 and 8 d before heading (DBH) with ammonium sulfate was ripened-grain weights (Table 1). Grades of grain quality were
in six patterns: 3–3 (g N m–2), 3–1.5, 3–0, 1.5–1.5, 1.5–0, and higher, and the percentages of the immature grain number were
0–0. The average temperature during the 20 d after heading lower, at higher N levels of topdressing, especially at 8 DBH
(Table 1). There were no significant differences between the
Treatmentsa Ripenedb Grain no. Ripened 1,000- Grade of Sum of Milky Basal Grain N
(g N m–2) grain yield per m2 grain ripened-grain qualityc immature white white contente
(g m–2) (×100) (%) weight (g) (1–3) grains (%)
3–3 620 304 90.0 22.7 1.7 39.4 8.4 6.2 1.24 *
3–1.5 617 ns 306 ns 90.1 ns 22.5 ns 1.8 ns 42.6 ns 9.4 ns 9.4 ns 1.18 *
3–0 593 ns 300 ns 89.4 ns 22.1 * 2.0 ns 46.7 * 10.0 ns 10.0 * 1.11 *
1.5–1.5 578 ns 282 ns 91.7 ns 22.3 * 1.8 ns 44.5 * 7.8 ns 11.4 * 1.13 *
1.5–0 561 *f 286 ns 90.7 ns 21.6 * 2.0 ns 51.5 * 8.2 ns 14.0 * 1.07 *
0–0 544 * 281 ns 90.3 ns 21.5 * 2.3 * 53.5 * 8.8 ns 13.4 * 1.06 *
aTopdressing at 18 and 8 days before heading with ammonium sulfate. bGrains with a thickness of 1.7 mm or more were regarded as ripened grains. cBy inspection system. 1st
high = 1, 1st low = 1.5, 2nd high = 2, 2nd low = 2.5, 3rd high = 3, 3rd low = 3.5. dMeasured by RGQI 10A (Satake). eMeasured by the Japan Grain Inspection Association.
f*Indicates a significant difference from the 3–3 treatment at the 5% level by Dunnett’s multiple range test.
Ratio between the distance from the center point to the contour
of grain cross section of 3-3 treatment and that of other treatments
1.02
3-3 treatment = 1
1.00
0.98
3-3
0.96 3-0
1.5-0
0.0
0.94
0 30 60 90 120 150 180
Angle of direction toward vascular bundle (º)
Fig. 1. Comparison between the grain contour of 3-3 treatment and that of other
treatments. Standard errors of the ratios between the distance from the center
point to the grain contour of the 3-3 treatment and that of other treatments
ranged from 0.007 (1.5-0 treatment) to 0.016 (3-0 treatment), on average, from
0 to 180º.
percentages of milky white grains of the 3–3 treatment and at 30 to 50 degree angles from the vascular bundle direction
those of other treatments. However, the percentages of basal from the center point of the cross section (Fig. 1), where the
white grains of the 3–3 treatment were significantly lower than photoassimilates were thought to be accumulated at a later stage
those of other treatments, except for the 3–1.5 treatment. These of ripening. These areas of angles contain a deep ditch made
results coincide with the report that immature grains with a from the joint line of the lemma and palea.
white portion increased at a higher grain number per unit area Image analysis of the longitudinal projection from the
(Terashima et al 2001, Kobata et al 2004) and at a lower N grain lateral side also revealed that lower N levels caused in-
level (Terashima et al 2001) under high temperature at ripen- ferior growth at ventral shoulder parts (from –120 to –150
ing. In this study, it was clear that the reactions to the increase degrees) and at the dorsal bottom square parts (from 30 to 60
in the nitrogen application change with the type of immature degrees) of grains (Fig. 1). The latter parts corresponded to
grain (i.e., milky white or basal white) under high tempera- the parts of the white portion in the basal white grain.
ture. But, the reason for this has been unclear. Further research Akiyama et al (1997) have classified rice cultivars by
into the physiological and morphological factors of white por- image analysis of the size of white cores in the grain. How-
tions in the endosperm will be required. ever, image analysis of rice grain shape had not been previ-
ously reported. In this study, the inferior growth position in
Grain shape the grains, revealed by image analysis of grain shape, may in-
Image analysis of grain projection from the grain top or bot- dicate the time and/or position in which a lower nitrogen ap-
tom revealed that lower N levels in the reproductive stage un- plication under high temperature has an immediate adverse
der high temperature caused inferior grain growth, especially effect. This suggests that image analysis of grain shape will be
Wrap-up of Session 19
The conclusions of the Third Assessment Report (TAR) of the models have to have parameters measured in “controlled envi-
Intergovernmental Panel on Climate Change (IPCC) leave no doubt ronments.” The first paper of the session by T. Horie and H. Yoshida
that Earth’s climate is changing in a manner unprecedented in described the development of such a model, with particular ref-
the past 400,000 years. The report forecasts that by 2100 mean erence to high-temperature stress and its use to predict future
planet-wide surface temperatures will rise by 1.4 to 5.8 °C, pre- yields in China, Thailand, and Japan.
cipitation will decrease in the subtropics, and extreme events will One curious consequence of climate change is that in some
become more frequent. However, changes in climate are already regions temperatures could get cooler because of changes in
being observed—the last 60 years were the warmest in the last currents and atmospheric pressure differentials. The second pa-
1,000 years and changes in precipitation patterns have brought per of the session by M. Yajima focused on the effects of cool
greater incidence of floods or drought globally. The TAR also con- weather in northern Japan on spikelet sterility and methods of
cluded that natural phenomena do not explain these observed protecting rice against cold damage.
changes—that there is now clear evidence of human influence. Coping with the effects on rice of weather variability asso-
Climate change is a development problem as well as an ciated with climate change in the form of increased frequency of
environmental problem. It is a development problem because wet and dry periods, tropical cyclones, and typhoons in the Phil-
those engaged in agriculture must learn how to adapt to the ippines was the topic of the third paper by F. Lansigan.
consequences of an altered atmosphere, higher temperatures, Climate change will have direct effects on rice but may
and greater variability and more extremes in weather. Coping si- also alter soil nutrient availability for rice. The fourth paper by J.
multaneously with the other adverse environmental consequences Zhu described how free-air CO2 enrichments can affect plant
of anthropogenic activity, such as air quality, will be no easy mat- nutrients in the soil solution.
ter. The combination of changes in air quality and composition, Rice ecosystems will be affected by climate change, but
acid rain, and climate will produce a new bioclimate for food there has also been great concern about the contribution of rice
production systems. The potentially beneficial effects of increases to climate change. Understanding the properties of rice soils in
in CO2 may be offset by the increases in ozone at ground level relation to methane emissions is crucial to efforts to minimize
(tropospheric concentrations) as well as high-temperature stress. those emissions and this topic was explored by S. Mitra and D.
It is arguable that most developing countries are not well informed Majumdar.
about the probable impact of bioclimatic change and are there- Professor Li and his colleagues investigated the alterna-
fore most vulnerable to its consequences. tives for farmers in managing their crops so as to reduce meth-
To understand the quantitative effects of climate change ane emissions. They developed a tool by integrating a process-
on rice production, it is necessary to have rice simulation models based model with a GIS database to scale up their experimental
that can predict the outcome of the new bioclimates on rice. The observations from the field to a regional level. They investigated
Research on precision agriculture has been attracting atten- grain inside the combine tank. This monitor is a load cell unit
tion with regard to the production of rice in Japan (Umeda that is fixed under the grain tank of the harvester and responds
1999) and other Asian countries (Sung et al 2002). It is thought to particles filling the tank. The output signal of this sensor is
that precision farming could solve the problem of variability also analog. As position information is acquired by GPS at the
within and between fields, leading to economic and environ- same time as yield information, the yield map is drawn after
mental benefits, as well as stable production. The methods of the harvest.
precision agriculture are based on information technology. In A gathering unit for analysis of the quality of samples
precision farming, farmers make optimal decisions depending (Chosa et al 2004b) that can be attached to a combine har-
on site-specific information, such as on fertility, growth, and vester was used to acquire quality information. This system
yield. consists of a leading tube in the form of a cylinder, receiving
In recent times, progress has been made in developing cups that gather the grain from the leading tube, and a turning
information acquisition technology (Shibata et al 2002, Chosa table that supplies the receiving cups. The leading tube is in-
et al 2004); progress has also been made with respect to appli- serted into the combine tank at an angle and leads the grain
cations for site-specific management (Chosa et al 2003). How- sample from inside the combine tank to the outside. As the end
ever, it is difficult to make recommendations for management of the tube inside the tank is cut obliquely, the angle of the
or applications based on the information acquired. Although tube controls the sampling weight and can be used to stop the
some successful case studies have been reported (Toriyama et sampling. When the sampling is stopped, normal operations
al 2003), no general method for producing recommendations can be conducted. Grain dropped through the leading tube is
has been determined. gathered in sampling cups that are supplied by the turning table.
Theoretically, if the fertility and variations in fertility of Grain samples corresponding to random harvest positions can
an area can be determined, the added nitrogen required to reach be collected through appropriate operation of the turning table.
a targeted nitrogen absorption level, or to decrease variability, After the harvest, the quality parameters of the samples col-
can be calculated. Similarly, if the absorption of nitrogen is lected can be analyzed, such as the moisture content, protein
determined through yield, a revised amount of nitrogen can be content, and outward quality. Since position information is
calculated for application the following year. However, nitro- acquired through the GPS at the time of sampling, the quality
gen absorption is not the same in each area, for each year, map is drawn after the analysis.
because of differences in soil conditions, climate, fertilizer
application method, and type of cultivation. For this and other
Farm work for data acquisition
reasons, it is difficult to generalize the formulation of a rec-
ommendation. To examine variability in yield and quality, a field survey was
This paper presents guidelines for making recommen- conducted during combine harvesting.
dations for the next year’s crop using yield and grain quality The combine harvester used was a modified commer-
maps of paddy fields. The information provided by mining the cial four-row operating combine harvester (head-feeding type,
data in these maps should be more useful than the information 25.7 kW, harvesting width 1.2 m). The combine was equipped
provided by the individual maps. Combining the two types of with the yield and quality acquisition systems described above.
maps for the purpose of site-specific management can reveal Commercial combine harvesters can easily be equipped with
useful information. both systems. Although the person who runs the combine must
operate the information acquisition systems, the harvesting op-
eration is otherwise the same as is usual for this farm work.
Yield and quality acquisition systems After the harvest and analysis of the yield, quality, and
To acquire yield information, a hybrid yield-monitoring method position information, the yield map and quality map were gen-
(Chosa et al 2004a) was used. This method uses a sequence- erated. Protein content was used as a measure of rice quality.
monitoring sensor simultaneously with a batch-monitoring sen- The protein content of each sample was measured using near-
sor. The sequence yield monitor, which is used to measure the infrared light.
grain flow rate, consists of an emitting unit and a receiving The yield and quality variability maps were overlapped
unit fixed to the top of the clean grain auger of the combine to determine a recommendation for cropping and fertilizing
harvester. The output signal is an analog signal that varies ac- strategies for the next year.
cording to the particle flow between the emitting and receiv-
ing units. A batch yield monitor is used to measure the mass of
Environmental data such as weather data and crop data such Block) and analog multipliers. By combining these elements,
as rice growth in paddy fields are necessary for crop manage- the FMS can be connected to various sensors and probes di-
ment and scientific studies. In addition, production history rectly.
systems and traceability systems for food security are becom- A typical specification of FMSs follows:
ing indispensable. A wireless sensor network could be one of 1. Casing: heat-resistant, waterproof, dustproof, and suit-
the best solutions for this need. able to the landscape.
So far, many kinds of sensor-network solutions, such as 2. Sensors: air temperature, humidity, PPFD (photosyn-
Mote (Khan et al 1999) and TINI (www.ibutton.com/TINI/), thesis photon flux density), soil temperature, soil
have been proposed. However, specification of the sensors is moisture, CO2 concentration, leaf wetness, ultravio-
poor for monitoring crops. Wireless broadband communica- let light, infrared light, etc.
tion, high-resolution image-monitoring technology, and vari- 3. Image data: 0.3–8 M pixels depending on cameras.
ous sensors are needed for monitoring rice in real-time in paddy 4. Wireless communication range: 300 m–10 km de-
fields. For example, a rice blast prediction system, pending on external antennas.
MetBLASTAM (http://cse.naro.affrc.go.jp/ketanaka/model/ 5. Hotspot service: within 100 m–1.5 km depending on
applet/Blastam.html), requires information on air temperature, antennas and circumstances.
humidity, and leaf wetness. Specific data such as images of 6. Communication speed: 54/11 Mbps.
emerging rice blast are indispensable to revise the prediction 7. Power supply: AC adapter, embedded solar cell, or
system. To estimate the photosynthetic rate of rice, the spatial external solar cells.
distribution of CO2 concentration should be measured in real- 8. Size: diameter is 20 cm and length is about 30 cm.
time in the paddy field. 9. Cost: less than $500 to $2,000 per unit depending on
We developed Field-Monitoring Servers for those re- embedded devices.
quirements, and constructed a wireless sensor network in paddy
fields. Their cost is extremely low and their functions are much
The full-wireless Field-Monitoring Server for paddy fields
advanced compared with conventional sensor networks or
weather stations. The observed data are freely available on Conventional FMSs require cables for the power supply, since
our Web site. Any users can use the data directory, and appli- power from the solar cell embedded on the rooftop of the FMSs
cations such as the rice blast prediction system, through is too low (0.5 W) to drive an FMS continuously, as it requires
MetBroker, which provides data in standard format for appli- more than 50 W for nights and the rainy season. The hardest
cations by linking these applications to conventional weather job for the installation of FMSs was the construction of wir-
databases such as AMeDAS and NOAA. ing. This will be much harder in paddy fields. So, the full-
wireless FMS was developed, especially for paddy fields with
several energy-saving technologies (Hirafuji et al 2004). Cli-
The Field-Monitoring Server ent connection FMSs can rest periodically to save power. The
A field server is a Web server installed in fields, and the Field- full-wireless FMSs connect other conventional standard FMSs
Monitoring Server (FMS) is a kind of field server for real- as clients at hotspots provided by the standard FMSs, which
time monitoring. The FMS can monitor fields in real-time and are driven by an AC adapter or larger solar cells and consume
use wireless local area network (LAN) hotspots around the more energy to maintain the backbone and the hotspot service.
FMS (Hirafuji 2000). The FMS consists of a field server en-
gine (micro Web server for control and data acquisition), a
Fieldserver-Agents and MetBroker
wireless LAN access-point, a network camera, and sensors
(Hirafuji and Fukatsu 2002, Fukatsu and Hirafuji 2003). Field servers can be controlled by users using a Web browser,
The field server engine is a main panel that functions also simultaneously controlled by a Fieldserver-Agent, which
like a Web server, analog I/O, digital I/O, DDS (Direct Digital is a kind of artificial intelligence (Fukatsu and Hirafuji 2004).
Synthesizer), FPAA (Field Programmable Analog Array), and The collected data are stored in public database servers such
analog multipliers. The DDS generates an alternative signal as a PC-cluster at NARC or a computer center. A Fieldserver-
up to 70 MHz. The FPAA serves to construct dynamic analog Agent controls almost all FMSs in the world and collects data
circuits, such as a low-pass filter and super-heterodyne radio automatically, such as the Google Web crawler. The agent is
receiver, using its embedded 20 CAB (Configurable Analog accessing FMSs every 1–10 min. Farmers and consumers can
view the data. The sorted images are shown as time-lapse FMS-2: air temperature (2 types), humidity, intensity of
movies to monitor plant growth and farms. This function can solar radiation, inside temperature, and CCD
be a production history system and a farm security system. camera (0.3 Megapixels).
Applications can use the data by using MetBroker FMS-3: air temperature, CO2 concentration, leaf wet-
(Laurenson et al 2002), which has transformed conventional ness, intensity of UV, and inside temperature.
weather databases and data of FMSs into standardized data FMS-4: air temperature, PPFD (2 layers), and inside tem-
automatically. MetBroker is also a network service. Applica- perature.
tions determined to be used by MetBroker can access an enor- FMS-5: air temperature, humidity, recharging current (in-
mous amount of data through MetBroker. If a new FMS is tensity of solar radiation), battery voltage, and
installed in a field, the measured data instantly become avail- inside temperature.
able applications. In addition, we tested a 3-Megapixel high-resolution
digital camera with a magnifying lens to FMS-2. We could
observe the surface of a rice leaf; the resolution is up to about
Experiments 10 micrometers (Fig. 2). This microscope camera can be in-
We installed five FMSs; three FMSs are in a paddy field and stalled in the paddy field and also controlled by the Fieldserver-
one FMS is on a rooftop at NARC’s Yawara experiment sta- Agent. In total, 23 sensors and two cameras were installed in
tion in Yawara village. The first FMS (FMS-1) on the rooftop the paddy field. All of the measured data were collected by the
is connected to the Internet by the Asymmetric Digital Sub- Fieldserver-Agent automatically.
scriber Line (ADSL), since the Yawara station is not connected We isolated the FMSs from the global IP network using
to the optical fiber network of MAFFIN (the Ministry of Agri- VPN for security using FieldServerGateway (Kiura et al 2002).
culture, Forestry, and Fisheries Network). The second FMS However, real-time data and collected data are always avail-
(FMS-2) is connected to FMS-1 by a repeating mode of wire- able on the Field Server Web site (http://model.job.affrc.go.jp/
less AP (a kind of ad hoc network), and the third FMS (FMS- FieldServer/monitor/).
3) is also connected to the FMS-2 by repeating. The fourth
FMS (FMS-4) and fifth FMS (FMS-5) are connected by an
Results and discussion
AP-Client mode as a client around FMS-2, which provides a
hotspot service. FMS-4 and FMS-5 were installed in the paddy We needed four visits to the Yawara station for installation of
field (Fig. 1). The power for FMS-4 was supplied by a cable, the five FMSs and ADSL router. That is troublesome for farm-
which was hard to use in a such a dirty field. So, we employed ers and researchers, so we improved the FMSs and now we
a full-wireless FMS for the fifth FMS. Power for the FMS-5 can finish such installation within a half day. The function of
was supplied by an embedded small solar cell. Sensors for the the hotspot and ad hoc network could be connected among the
FMSs are FMSs successfully, and FMSs now provide a ubiquitous net-
FMS-1: air temperature, humidity, intensity of solar ra- work service at the station. The CO2 concentration in the paddy
diation (power of solar cell), and inside tempera- field can be measured continuously and the relationship with
ture. other monitored data such as UV light is clear.
So far, the FMSs were installed in both the paddy field References
and at many experimental sites in several countries such as
Fukatsu T, Hirafuji M. 2003. Development of Field Servers for a
Japan, the United States, Thailand, Denmark, China, Korea, field monitoring system. Agric. Info. Res. 12:1-12.
Canada, and Taiwan (China). The FMSs in paddy fields be- Fukastu T, Hirafuji M. 2004. The agent system for Field Monitoring
came extremely dirty after one year, and the sensors were dam- Servers to construct smart sensor-network. Fifth International
aged. So, we are improving the FMSs; for example, an air fil- Workshop on Artificial Intelligence in Agriculture, 8-10 March
ter was appended to the air intake. 2004.
Hirafuji M. 2000. Creating comfortable, amazing, exciting and di-
verse lives with CYFARS (CYber FARmerS), an agricultural
Conclusions virtual corporation. Proceedings of the Second Asian Confer-
A wireless sensor network is constructed in paddy fields using ence for Information Technology in Agriculture. p 424-431.
Hirafuji M, Fukatsu T. 2002. Architecture of Field Monitoring Serv-
FMSs. We can access the Internet in the area ubiquitously by a
ers. Proceedings of the Third Asian Conference for Informa-
Wi-Fi wireless-LAN. The sensor network has collected data
tion Technology in Agriculture. p 405-409.
in real-time since July 2003. The collected data were used to Hirafuji M, Fukatsu T, Hu Haoming. 2004. Full-wireless Field Moni-
develop new sensors such as a leaf-wetness sensor. Relation- toring Server for advanced sensor network. Proceedings of
ships hidden in enormous amounts of measured data and AFITA/WCCA, 2004. p 686-691.
weather databases can be found by MetBroker and data min- Kahn JM, Katz RH, Pister KSJ. 1999. Mobile networking for smart
ing. So far, we lack detailed data in paddy fields where we dust. ACM/IEEE International Conference on Mobile Com-
could not install measurement equipment, and only some data puting and Networking (MobiCom 99), Seattle, Wash., USA.
and numerical estimations have been employed. At present, Kiura T, Fukatsu T, Hirafuji M. 2002. Field server gateway: gateway
we can now collect an enormous amount of data automatically box for Field Monitoring Servers. Proceedings of the Third
and monitor them in real-time by live cameras with the FMSs. Asian Conference for Information Technology in Agriculture.
p 410-413.
Any people, such as researchers and consumers, can inspect
Laurenson M, Kiura T, Ninomiya S. 2002. Providing agricultural
paddy fields anytime at any place using the collected data.
models with mediated access to heterogeneous weather data-
This can make rice production more scientific, more reliable, bases. Appl. Engin. Agric. 18:617-625.
and safer.
Notes
Authors’ addresses: National Agricultural Research Center, e-mail:
hirafuji@affrc.go.jp, fukatsu@affrc.go.jp, hhaoming@
affrc.go.jp, kiura@affrc.go.jp, tomonari.watanabe@
affrc.go.jp, snino@affrc.go.jp.
The whitebacked planthopper Sogatella furcifera and the Atomic Energy Research Institute (JAERI), which calculates
brown planthopper Nilaparvata lugens are major pests of rice atmospheric dispersion of radioactive particles in case of a
in eastern Asia. They migrate from southern China into Japan nuclear accident (Ishikawa and Chino 1991). GEARN was
mainly in the Bai-u rainy season in June to July every year modified to model the planthopper’s migration behavior. Fig-
(Seino et al 1987). Generally, the prediction of planthopper ure 1 shows a schematic model. This model calculates the po-
migration provides farmers and plant protection advisers with sition of many planthoppers, and doesn’t discriminate between
information about the migration source as well as area of mi- N. lugens and S. furcifera. Nearly 2,000 planthoppers for each
gration. It is especially important to know the migration source, takeoff area took off randomly within the area at 10 or 21 UTC
which helps to understand the immigrant’s characteristics such (Universal Time Coordinate), which locally corresponds to
as biotype, pesticide resistance, etc. Conventional prediction dusk or dawn, respectively. They then flew up at a speed of
methods used 6- or 12-hour two-dimensional wind data 0.2 m s–1 for 1 hour. Taking off ceased 1 hour later, at 11 or 22
(Rosenberg and Magor 1983, Watanabe et al 1990), but their UTC. During their flight over the sea, the planthoppers moved
prediction quality was limited. A three-dimensional simula- at the same speed as the wind because they are slow fliers.
tion method using a boundary layer model was developed Vertical diffusion was taken into account by a random walk
(Turner et al 1999). However, a prediction system based on model, but horizontal diffusion was not because preliminary
this model has not yet been developed. results showed too much horizontal diffusion over Japan. Since
To achieve high-precision migration prediction, a real- the planthoppers don’t like cooler air, a temperature ceiling of
time prediction system was developed using three-dimensional 16.5 ºC was set up, and they could not go beyond that ceiling.
simulation models. This paper presents the simulation method Wind, temperature, and vertical diffusion coefficient data were
and its evaluation results. given by numerical weather forecasts. Simulation duration was
48 h. The model calculated the relative aerial density of
planthoppers based on their position on the model grid. The
Materials and methods
horizontal resolution was 33 km. Areas of relative density larger
Planthopper behavior than zero in the lowest level, which was less than 100 m above
Sogatella furcifera and N. lugens are tiny insects about 3–4 the ground, were used for prediction. These areas were called
mm in size and 1–3 mg in weight (Ohkubo 1973). Many field “migration clouds” (Fig. 2).
observations have shown that the planthoppers take off at ei-
ther dusk or dawn (e.g., Ohkubo and Kisimoto 1971). A radar Prediction system
observation showed that S. furcifera and N. lugens flew to an First, the latest meteorological data were supplied online to an
altitude of several hundred to 1,000 m above the ground at an advanced numerical weather prediction model, MM5 (Grell
estimated upward speed of 0.2 m s–1 (Riley et al 1991). The et al 1994). The model forecast atmospheric fields for the next
species fly at about 1 m s–1, whereas the wind speed when 72 hours at 1-hour intervals. Second, these forecast fields were
migration occurs is typically more than 10 m s–1 (Seino et al supplied to the planthopper migration simulation model,
1987). Laboratory experiments of tethered N. lugens adults GEARN, which calculated displacement of a number of mod-
indicated that they have the ability to fly for up to 23 hours in eled planthoppers and predicted relative aerial density. The
air of high humidity (Ohkubo 1973). N. lugens ceases flying data were converted to PDF files at 3-hour intervals and sent
at cooler temperatures; half of them stop beating their wings to the project’s Web site. These processes were conducted au-
when the air temperature is below 16.5 °C (Ohkubo 1973). tomatically. The system predicted migrations over the next two
Their postmigration landing process is not yet fully understood. days. The maps of relative aerial density provided informa-
tion on the timing and area of arrivals. The system also gave
Planthopper migration simulation model possible migration sources.
Takeoff area. Because information on the planthopper’s den-
sity in source regions was not available, it was assumed that
Results and discussion
the population density in June to July was high enough for the
planthopper to take off from source regions. In the simulation, Evaluation
several takeoff areas of 50–100 km2 were set up in southern Figure 2 shows an example of predicted migration clouds. An
China and Taiwan. These takeoff areas were located in paddy evaluation was conducted using daily catch data obtained at
fields, and covered the major source regions. three sites in Kyushu, western Japan, in 2003 and 2004. The
Model. The migration simulation model originates from sites were Saga (33.17°N, 130.33°E), Kumamoto (32.95°N,
a particle dispersion model, GEARN, developed by the Japan 130.78°E), and Kagoshima (31.52°N, 130.50°E). The system
Wind + V-diffusion
100 m
50 km Sim. grid
Fig. 1. Schematic of the migration simulation model. Black circles represent modeled planthoppers. As
many as 1,852 planthoppers per one takeoff area were randomly generated at 10 or 21 UTC. The
simulation duration was 48 h.
E 110º 120º 130º tions were evaluated by a hitting ratio defined as:
N
40º 40º
Hitting ratio (%) = Number of days when the migration
was correctly predicted/Total number of days × 100
Discussion
This is the first real-time prediction system to use three-di-
20º 20º
mensional wind fields given numerically by weather forecasts.
The system enabled migration prediction that presented infor-
110º 120º 130º mation about the timing and area of migrations. Nonetheless,
several issues need to be discussed.
Fig. 2. An example of migration clouds. This figure shows the pre-
dicted distribution of relative density of planthoppers at 6 UTC, The takeoff assumption. The strong assumption made
25 June 2004, which started at 21 UTC from all the takeoff areas. for the prediction was that planthoppers were supposed to take
off from every predefined takeoff area at every takeoff time.
The regions where planthopper density is high in June and
predicted daily migrations by seeing whether a migration cloud July are located in southern China at latitudes less than 25°N
would cover the evaluation site or not. To judge predictions, (Zhou et al 1995). The takeoff areas in Fujian, Guangdong,
daily catches of S. furcifera captured at the sites in June to Guangxi, and Hainan provinces qualify as such regions. In some
mid-July were used. Catches of N. lugens were too small to be years, however, northern regions around 30°N have been in-
used for the evaluation. Traps used were net traps and a vaded by the beginning of June. Therefore, the other northern
Johnson-Taylor-type suction trap. The net trap was a 1-m tow takeoff areas were included. Since the theoretically designated
net mounted at the top of a pole 10 m high. More than one takeoff areas widely cover possible source regions, this as-
catch was interpreted as a migration event. The daily predic- sumption seems reasonable.
There is no doubt that the agricultural sector, no less than any applied in agriculture and continue on to examine some of the
other, is facing a range of old and new challenges as a result of lessons learned.
today’s economic and environmental pressures. Key among
these are population growth, increased market complexity,
Applications of digital technologies to agricultural learning
continuing economic inequality, and the need to raise produc-
tivity without adversely endangering the natural resource base eLearning and digital technologies are increasingly influenc-
(McCalla 2001). ing and enriching all forms of agricultural education. This is
A growing global population means that agriculture will most apparent in informal and formal education, but with ten-
need to produce enough food to feed an expected two billion tative pilot applications in nonformal education as well.
additional people by 2025 and this additional production must Traditional means of informal education and learning in
be achieved with less natural resources. Compounding these agriculture, based on knowledge and skills being passed be-
problems is the changing economic nature of agriculture, with tween generations and between community members, are be-
increased commercialization, sophistication, and globalization. coming much less effective. This is a result of several factors,
There is a growing consensus that learning will be a major such as fewer experts in rural areas, agricultural innovations
factor to help agriculture and agricultural producers success- are increasingly coming from outside the community, fewer
fully deal with these challenges. “Knowledge—and related traditional courses are being offered, and much information is
information, skills, technologies, and attitudes—will play a key time-sensitive and/or needed quickly (Agriculture and Agri-
role in the sustainable intensification of agriculture and suc- Food Canada 2003). Comprehensive and growing reposito-
cess of rural development investments,” stated Alex et al ries of online agricultural information provide a powerful means
(2002). of overcoming these constraints and allow the independent
Getting the essential knowledge to those who need it learner to delve deeply into a myriad of subjects with the click
most remains difficult and expensive, but much optimism has of a mouse.
been generated as a result of the increased growth and sophis- Formal education has long been limited by geography,
tication of new electronic information services—even in re- high cost, and lack of access by particular groups; as a result,
mote rural areas. Information and communication technolo- eLearning and associated technologies are increasingly being
gies (ICTs), and such specialized ICT applications as used to overcome these obstacles. This trend is most promi-
eLearning, are offering new options to deliver knowledge and nent in higher education in developed countries and recent fig-
information to farmers directly and indirectly through knowl- ures show that 80% of the top U.S. and European universities
edge intermediaries. will offer global courses in 2004, with many of these offerings
eLearning is one form of distance learning, a type of related to agriculture.
educational situation in which the instructor and students are While informal and formal education are certainly im-
separated by time, location, or both. eLearning typically in- portant for the future of agriculture, nonformal education is
volves the use of the Internet to access learning materials; in- arguably the most critical. There is now almost too much in-
teract with the content, instructor, and other learners; and ob- formation available online for informal learning and taking
tain support during the learning process in order to acquire full advantage of these resources requires special skills to lo-
knowledge, construct personal meaning, and grow from the cate and evaluate. Agricultural knowledge acquired through
learning experience. formal education is soon outdated and obsolete. Properly con-
Proponents make several convincing arguments about ceived and developed, nonformal eLearning can substantially
the power and potential of eLearning. eLearning provides learn- complement formal and informal efforts and provide up-to-
ing opportunities in subjects not offered locally or where local date and relevant agricultural knowledge.
offerings lack quality. It is ideally suited for individuals who
lack time for classroom courses. Perhaps most importantly,
Lessons
participating in an online class gives students the skills required
for lifelong learning. Although schools and other providers of education first began
Although eLearning is still in its infancy, particularly in experimenting with online education only just over a decade
developing countries, some experience has been gained. Be- ago, much has already been learned. Below, some key lessons
low, we will look briefly at where and how eLearning is being are listed and discussed.
The fast and effective transfer of research findings to farmers graphic slide–audio tape modules, instructional video, early
has always been one of the biggest challenges facing those in attempts at computer-aided instruction, video-conferencing,
agricultural development. All too often, new knowledge is suc- and information presented on CD-ROM have all supported
cessfully developed and validated, only to fail in reaching those face-to-face classroom instruction both at IRRI headquarters
who need it most—the farmers. Into this gap between research and in country training programs. Now, IRRI is developing
and impact has stepped the International Rice Research and adding another set of tools to better serve its education
Institute’s (IRRI) Rice Knowledge Bank (RKB). Not only is it mandate: those that are collectively known as information and
one of the world’s first digital extension services for those who communication technologies (ICT), which include distance-
provide information and support for farmers, it is also the first learning methodologies and use of the Internet.
comprehensive, digital rice-production library containing a IRRI has moved to harness the power of ICT by care-
wealth of information for rice-related training and extension. fully matching the new media’s capabilities to the needs of
More importantly, it provides this service using a format that rice-related training to bring relevant knowledge, in the most
sets a new standard for knowledge access within the agricul- useful form, to field officers when and where they need that
tural development community. Containing the most up-to-date knowledge. The RKB has received critical acclaim (for ex-
and validated knowledge, the RKB is providing government ample, BBC Earth Report, September 2004—the “Further read-
extension services, nongovernment organizations (NGOs), and ings” section provides more reviews of the RKB) as a tool to
universities with unprecedented access to training and support distill, store, and provide access to the vast array of IRRI’s
knowledge. training and support knowledge for rice science and exten-
First and foremost, IRRI is a rice science research insti- sion.
tute. However, IRRI’s founders knew that, to get research find-
ings out of the laboratory and into farmers’ fields, there had to
Rationale for the RKB
be an education component. Over the past 40 years, IRRI has
used a variety of instructional methodologies and technolo- Ask any farmers what they perceive as their major needs and
gies to satisfy its education mandate. Printed materials, photo- you will probably be told three things: access to credit, a good
International germplasm exchange was the engine of the Green Several CGIAR centers, national agricultural research
Revolution. In the past, however, much of the important infor- and extension systems, and advanced laboratories are collabo-
mation generated from this exchange was accessible only lo- rating to develop ICIS as a generic system that will accommo-
cally, in field books or researchers’ files. Although major in- date all data sources for any crop. The vision of ICIS is to
ternational initiatives for germplasm collection and conserva- integrate different data types into a single information system
tion followed the Green Revolution, much collected material and provide specialist views and applications that operate on
is still not used because it is difficult to access. As a result, the a single integrated data platform. ICIS must also seamlessly
potential impact on agriculture has not yet been realized. How- integrate private local data with public central information to
ever, the free exchange of information, through international give local researchers access to global crop information. After
crop information systems, should provide the foundation for a all phases of development are complete, ICIS will seamlessly
second revolution that adds value to germplasm by seamlessly support a range of activities including germplasm conserva-
uniting its conservation, evaluation, use, and exchange. tion, functional genomics, allele mining, breeding, cultivar test-
Furthermore, new technologies in molecular biology and ing, and release. Data will be accessible from CD-ROM or the
genomics mean that traditional phenotypic information must Internet, and users could either adopt the complete system or
be linked to large quantities of sequence and genetic informa- individual components complementary to their own systems.
tion so that functional genomics and allele-mining activities The driving force behind ICIS is accessing and sharing data
can speed up germplasm enhancement. rather than providing analytical and statistical tools. This is
The International Maize and Wheat Improvement Cen- because the major bottleneck to intelligent data integration and
ter (CIMMYT) devised an information strategy and developed use was not statistical software, but rather the drudgery of find-
software on a main-frame computer during the 1980s to facili- ing, extracting, preparing, and managing the data.
tate unambiguous identification of wheat germplasm, thereby
establishing links among information from different sources Functionality
(Fox et al 1996a). In 1995, CIMMYT and IRRI canvassed The ICIS system is fast, user-friendly, PC-based, and distrib-
other CGIAR centers to establish a project to develop an In- utable on CD-ROM or via the Internet. It contains
ternational Crop Information System applicable to a wide range O A genealogy management component to capture and
of crops. process historical genealogies as well as to maintain
evolving pedigrees, and to provide the basis for unique
identification and internationally accepted nomencla-
The International Crop Information System (ICIS) ture conventions for each crop.
Extensive communication among CGIAR centers highlighted O A data management system for genetic, phenotypic,
the economies to be gained by collaborating on the develop- and environmental data generated by evaluation and
ment of an information system that could be used for many testing, as well as for providing links to genomic maps.
crops. At a meeting held at CIMMYT in September 1994, a O Links to geographic information systems that can ma-
prototype CD-ROM was demonstrated and Open Database nipulate all data associated with latitude and longi-
Connectivity (ODBC) was chosen as a programming standard tude (e.g., international, regional, and national test-
for international collaboration. This ensures that the choice of ing programs).
database system for a particular crop is independent of the
decision to use ICIS.
Has progenitors
Derivative germplasm
Developed by
Methods
Called
Names
Named at
Named by
Developed at Locations
Developed by Users
Assigned at
Assigned by
With value of Attributes
Defined property
User-defined fields
dure by which the value is observed or derived, and the SCALE changes are part of the update process and sufficient informa-
or measurement units in which the value is expressed. tion and justification need to be recorded in the changes table
to allow the process to be completed.
Correcting data Verification and completion of changes in the central
Corrections and changes will inevitably be made in any data- database may take some time but local users would like to see
base. Only authorized users can make these changes, and all their changes reflected immediately. This is achieved by the
changes are logged so that the sequence of changes can be DLL, which always checks the local CHANGES table for cen-
traced and can be “undone” if required. tral changes and applies them at run time for the specific in-
One common occurrence of changes is when new infor- stallation where they are recorded.
mation about an existing germplasm record is entered into a
local database. If the existing record is also in the local data- Stand-alone software modules
base, then the local user can complete the changes. But, if it is Components of ICIS include a Genealogy Management Sys-
in the central database, changes cannot be completed until the tem (GMS); Set Generation Module (SETGEN), including the
central database is updated. Verifying and completing requested External Pedigree Input Tool (EPIT); Field Book Module
Applied by
Observed by
Property management
module
Application of Measured of
Factor Property
Expressed in Expressed by Variate
Applied by
Scale
Indexed by Measured by
Element
of
Value of Method Measurement
of
Effect Vector
Representation of
(FLDBK); Trait Management System (TMS); Data Manage- ICIS has also enjoyed collaboration with the USDA-
ment System (DMS); a Work Book (WRKBK) for data input sponsored GrainGenes and Gramene database initiatives. Link-
and query; and the Data Retriever for cross-study data queries ing these implementations to ICIS will provide access to se-
(RTV). The first four modules focus on germplasm and man- quence information and molecular maps, which will facilitate
agement of genealogy and nomenclature. The next three handle functional genomics and allele mining and integrate informa-
the management of evaluation and characterization data, and tion across crops at the genomic end of the spectrum.
the Retriever provides access to both raw and processed data.
The International Rice Information System (IRIS)
User-defined data
Users will be able to define new relationships among IRIS is the rice implementation of ICIS. The GMS of IRIS
germplasm by specifying new breeding methods. They can stores information on about one and a half million varieties,
specify attributes of the germplasm to be stored, types of names, breeding lines, and accessions of rice. This allows pedigree
location descriptors, and traits and variates for characteriza- analysis to trace germplasm flows and relationships between
tion and evaluation data. Data with any form of factorial struc- lines that can be used to improve evaluation estimates or plan
ture can be managed. improvement programs. The DMS of IRIS contains 5 million
data values from over 500 studies from breeding, screening,
Links to global bioinformatics resources and international testing trials. This allows integrative analy-
Other databases of agricultural information have close links sis over different environments. These data are available
with ICIS. The System-wide Information Network for Genetic through a web interface and are linked to rice data sources
Resources (SINGER) aims to provide global access to genetic throughout the world (Bruskiewich et al 2003).
resources data across all CGIAR-mandated crops and com-
modities. Germplasm records in ICIS that relate to accessions
Conclusions
in the CGIAR collections are linked with SINGER so that the
ICIS can provide information on the use and deployment of The technical challenge for plant scientists and software de-
those genetic resources. velopers is to implement the type of system outlined here. The
Each individual ICIS database manages data for a par- challenge for administrators is perhaps more difficult—to fa-
ticular crop and generally does not share data with other crops. cilitate the continued free exchange of information and to nur-
This integrating role of genetic resources information across ture a scientific culture in which users take full advantage of
crops is played by SINGER (Fig. 3). the data of others and, in turn, contribute to shared databases.
ICIS
data model and information system
In the mid-1990s, the Consultative Group on International pended on new levels of connectivity. The IVDN brought the
Agricultural Research (CGIAR) positioned itself well to take first digital Internet connections to several of the countries in
advantage of new opportunities provided by information tech- which the CGIAR operates. Since then, the world has changed
nology, with projects such as the Integrated Voice and Data a great deal, as evidenced by disconnections from the IVDN
Network (IVDN) and the System-Wide Information Network and connections to local Internet service providers that didn’t
for Genetic Resources (SINGER)—an application that de- exist seven or eight years ago. Unfortunately, the CGIAR is no
2 Mbps
IRRI (E1) 1.5 Mbps
APAN
Commercial
ISP
Internet
Fig. 1. IRRI’s APAN connectivity in 2003: a shared local loop from IRRI to a co-located
router at the premises of the Philippine Network Foundation (www.ph.net) in Manila for
a 1.5 Mpbs APAN link and a 1 Mbps commercial Internet connection (one of two in
operation). ISP = Internet service provider.
After
Los Baños
U.P.
Los Baños PHNET, Makati
2 Mbps
U.P. IRRI (E1) IRRI
Open
University
2 Mbps Ai3
IPB (E1)
6 Mbps
IRRI (3 × E1)
APAN
Fig. 2. IRRI’s APAN connectivity in 2004: the 6 Mbps local loop to the Advanced Science
and Technology Institute (ASTI) at the University of the Philippines Diliman
(www.upd.edu.ph) will be operational in 2005. Here, a CGIAR institute serves as a bridge
between both national and international partners.
to the Philippines increased from 1.5 Mb to 6 Mbps in Febru- retained as a backup and made available to the local scientific
ary 2004. In the Philippines, the Department of Science and community. Locally, the University of the Philippines Los
Technology has sponsored the creation of a national network, Baños (UPLB), the University of the Philippines Open Uni-
PREGINET, the Philippine Research, Education, and Govern- versity (UPOU), and the Institute of Plant Breeding (IPB) were
ment Information Network (http://preginet.asti.dost.gov.ph). interconnected with fiber optic cable that converged in an IRRI
This network now connects all 19 regions of the Philippines building. They were not connected to APAN, nor, more im-
with broadband links and more than 70 institutions are con- portantly, to PREGINET.
nected to it. Video-conferencing between nodes and broadcast- Apart from local loop communication costs and infra-
ing of meetings using video-streaming technologies are rou- structure costs, advanced research network bandwidth is free
tine. to the end user. When CGIAR scientists are connected to
Recent changes in IRRI’s APAN connectivity are sum- Internet2 or to APAN or similar networks, including connected
marized in Figures 1 and 2. national networks in a growing number of countries, they can
With the change in the national termination point of make use, at no marginal cost, of video-conferencing, video-
APAN from PHNET to ASTI, as shown in Figure 2, IRRI opted streaming, and other communication-intensive technologies
to install a faster local loop that would enable better exploita- (e.g., for large data transfers, and for hosting and synchroniz-
tion of the extra APAN bandwidth. The original circuit was ing mirror sites of genomic, climatic, and other data).
IRRI has begun construction of an Access Grid node well documented. What the world needs, and will surely have
(www.accessgrid.org). Once additional CGIAR centers con- soon, is an affordable multifunctional device for voice and data
nect to advanced research networks, the CGIAR network can communications. A new state-of-the-art device, such as an HP
play a bridging role in facilitating face-to-face interactions iPAQ 6300, costs about $500. For half as much, one eminent
between national partners at some locations and external part- researcher, Raj Reddy, now claims it is possible to make and
ners. With the Access Grid’s suitability for informal group sell a device that connects the world’s poor.2
meetings, we hope that the technology “catches on” and is used It seems likely that we are, at most, only a few genera-
for informal meetings by communities of interest. tions of cellular and wireless technology away from a $100
In 2004, four CGIAR centers acquired grid cluster com- device. This is still not affordable by all individuals but should
puter systems for bioinformatics and genomics research. It is be accessible to most communities in populated areas.
planned that these resources will be shareable on the CGIAR- What will the connection be between researchers using
wide area network and improvements in ARN connectivity to networks like APAN and farmers and extension workers with
address this should have spillover benefits for bandwidth for Internet access via a prepaid subscription to a wireless Internet
content sharing and delivery. service? The “CGIAR Knowledge Bank in your pocket” will
be attainable. Whether it will feed anyone, of course, will be
another challenge. Just delivering content will not be enough.
APAN and the CGIAR tomorrow Clearly, we are only at the beginning of an exciting era.
Despite the explosive growth in the number of cell phones in
the region in recent years, access to communication technol-
Notes
ogy is out of reach for many Asians, in particular the rural
poor. In Bangladesh, for example, it is estimated that only 5% Author’s address: Head, Information Technology Services, IRRI,
of the population has a cell phone despite annual growth rates DAPO Box 7777, Metro Manila, Philippines,
of 85%.1 This will surely change dramatically in the next 5–10 p.onolan@cgiar.org.
years as cellular and wireless Internet devices become ever
cheaper, more capable, and increasingly convergent. The em-
powering effects of access to communication by the poor are
Rice yield and quality depend highly on soil nitrogen fertility Simulation model and structure of the system
(Murayama 1979). Yearly fluctuations of temperature and sun-
shine also affect rice growth and yield considerably. Thus, a Outline of the simulation model
decision support system formed by inputting those physical The simulation model employed in the system contains sev-
variables has been anticipated. A decision support system for eral simple submodels that simulate the development and
rice was developed in Australia to calculate the amount of growth of rice plants in daily time steps (Table 1). Most of the
topdressing nitrogen fertilizer based on leaf nitrogen analysis submodels were modified or evolved from the models by pre-
(Angus et al 1996). cursors. The developmental stage of rice plants is estimated
Recently, the number of farmers with more than 10 ha by maximum and minimum air temperature, and is expressed
has increased in Japan, accompanied by land consolidation, in as a continuous variable. The influence of daylength on panicle
which several small fields (0.1 to 0.3 ha) are joined to form a initiation is not considered in this model. The parameters of
large field (1 ha), aiming at increasing production efficiency. the model for soil nitrogen mineralization, such as mineral-
However, this has often engendered variability in soil fertility ization rate constant, mineralization potential, and apparent
within a large field, and thus uniform application of fertilizer activation energy, should be derived from the experimental
results in the lodging of rice plants and considerable spatial data.
variability in grain yield and quality. More information-inten-
sive fertilization strategies, by which nitrogen fertilizers are Structure of the system
precisely applied according to site-specific information about This decision support system is named “RiceNiSMo” (Rice
soil fertility, are needed to match fertilizer inputs to the sea- Nitrogen Simulation Model). It is written in Visual Basic and
sonal pattern of rice crop nutrient demand and soil nitrogen runs on PCs with Windows-OS. The target rice plant of the
supply. Although there was a trial for site-specific nutrient man- system is transplanted Koshihikari, a leading variety in Japan
agement in several Asian countries in cooperation with IRRI with good taste. This system has roughly four functions, as
(Witt and Dobermann 2002), the required precision to over- follows.
come the variability in the reclaimed large field would be much Registration of nitrogen fertilizer. Not only ammonium
higher. Thus, we have developed a computer-based decision sulfate or ammonium chloride but also the compound fertil-
support system based on the idea of site-specific nutrient man- izer, high-analysis mixed fertilizer, coated urea (controlled-
agement to solve the above problems. release urea N) could be registered in this system. Two or three
different types of nitrogen fertilizers, which are mixed for basal
dressing by users, could also be registered as original N fertil-
Table 1. Relationships between simulated items and factors for the simulation.
Wrap-up of Session 20
The importance of information technology (IT) is increasing rap- a low-cost, wireless-based field-monitoring system, Field Server.
idly in agricultural production and research, as demonstrated by Field Server is a Web server installed in fields. When used with a
several examples of IT-related activities with rice. IT, unlike many Field-Monitoring Server (FMS), it allows real-time monitoring to
other technologies, is horizontally applicable in that it has appli- identify wireless LAN hotspots around the FMS. The FMS con-
cations throughout almost all aspects of agriculture. This session sists of a Fieldserver-Engine (a micro Web server for control and
discussed the present status of IT-based activities and provided data acquisition), wireless LAN access point, network camera,
an opportunity to discuss the future of IT applications in rice re- and sensors. By combining these elements, the FMS can be con-
search and production. nected to various sensors and probes directly. A group of Field
During this session, seven oral papers and nine posters Servers can be linked to develop a linked sensor network. Data
were presented. Seven of the oral papers and one poster have acquired from Field Servers by an agent program are stored in
been recommended for inclusion as full papers in the confer- the FMS archive in an XML format and the stored data are linked
ence proceedings. The papers cover the application of IT in pre- to a weather database, accessible through a weather data me-
cision farming, a low-cost wireless-based field-monitoring robot, diator, MetBroker, to provide a consistent interface to several
an integrated database system to efficiently transfer knowledge, agricultural software applications. A group of Field Servers has
lessons learned in an eLearning activity, an integrated system to been used in paddy fields for a long-term experiment without any
accelerate rice breeding, and a high-performance Internet infra- trouble. Several questions arose during the discussion period—
structure to harness IT applications. The following is the sum- particularly about the availability and price of the FMS.
mary of each paper and a discussion from each presentation. A. Otuka and others, in “Prediction of airborne immigration
T. Chosa, in “Data mining using combined yield and quality of rice pest insects,” reported on a new simulation model to pre-
maps of paddy fields,” discussed developments in precision farm- dict the airborne immigration of rice insect pests to Japan. The
ing in recent years, particularly in Japan and some other Asian whitebacked planthopper and the brown planthopper are major
countries. Precision farming involves making optimal decisions pests of rice in eastern Asia, and migrate from southern China to
based on site-specific information, such as fertility, growth, and Japan, primarily in June to July each year. Generally, the predic-
previous crop yields. In spite of the progress in information ac- tion of planthopper migration provides farmers and plant protec-
quisition technology and site-specific management, there are still tion advisers with information about the source of the migration
difficulties in providing optimal recommendations for crop man- as well as where the migration is occurring. Knowing the migra-
agement and no general method for producing recommenda- tion source is important as it helps to understand pest character-
tions has been determined. This paper suggested the need to istics such as biotype, pesticide resistance, etc. The quality of
develop a categorical map as a guideline for making recommen- conventional prediction methods that use 6- or 12-h two-dimen-
dations for crop management in succeeding years by combining sional wind data is limited. The authors adopted a three-dimen-
yield and quality maps of paddy fields. During discussion, the sional simulation method using a boundary layer model to de-
benefits of this approach to farmers were pointed out as one of velop a new method to predict pest immigration. Field observa-
the most important criteria for further development in precision tions of pest immigration over two seasons showed that the new
farming. method could predict immigration of the pest 2 days before its
M. Hirafuji and others, in “A wireless sensor network with arrival with a precision as high as 80%. Compared with the preci-
Field-Monitoring Servers and MetBroker in paddy fields,” discussed sion of the former method, this precision is fairly high, indicating