ARTICLE IN PRESS

Crop Protection 24 (2005) 991–998

www.elsevier.com/locate/cropro

Yield components and quality of rice in response to graminaceous

weed density and rice stink bug populations
K.V. Tindalla,, B.J. Williamsb, M.J. Stouta, J.P. Geaghanc, B.R. Leonardd, E.P. Webstere
a
LSU AgCenter, Department of Entomology, 402 Life Science Bldg., Baton Rouge, LA 70803, USA
b
LSU AgCenter, Northeast Research Station, P.O. Box 438, St. Joseph, LA 71366, USA
c
LSU AgCenter, Department of Experimental Statistics, 67 Ag. Admin. Bldg., Baton Rouge, LA 70803, USA
d
LSU AgCenter, Macon Ridge Research Station, 212-A Macon Ridge Road, Winnsboro, LA 71295, USA
e
LSU AgCenter, Agronomy Department, 104 M. B. Sturgis Hall, Baton Rouge, LA 70803, USA
Received 29 June 2004; received in revised form 26 January 2005; accepted 28 January 2005

Abstract

Field experiments were conducted in 2002 and 2003 to investigate weed density, its relationship to rice stink bug (Oebalus pugnax,
F.) populations, and damage to rice caused by stink bugs. Graminaceous weeds examined were barnyardgrass, Echinochloa crus-
galli Beauv., Amazon sprangletop, Leptochloa panicoides (Presl.) Hitchc., broadleaf signalgrass, Brachiaria platyphylla Nash., and
large crabgrass, Digitaria sanguinalis, (L.). Rice seed weight, percent filled seed, percent pecky rice, milling quality, and yield were
measured. Data showed that 13–23 weeds/1 m2 was associated with an increase of one rice stink bug per plot. Weeds served as hosts
of rice stink bugs prior to panicle emergence of rice; consequently, rice stink bugs infested rice early in the grain filling process and
reduced the percentage of filled seeds. One hundred weeds/1 m2 caused a 1% increase in pecky rice, and for every 1% pecky rice,
milling quality was reduced by 0.5%. Plots not treated with insecticide had significantly more non-filled seeds, pecky rice, and
broken kernels than treated plots. Neither weeds nor insects at the densities observed in this test appeared to effect seed weight. Rice
stink bug damage did not significantly contribute to yield losses greater than weeds in the absence of rice stink bugs. Rice stink bugs
had more of an affect on the quality of rice rather than the yield. Results reported here suggest that late season weed control may be
important in terms of indirect losses in grain quality associated with increased populations of rice stink bug.
r 2005 Elsevier Ltd. All rights reserved.

Keywords: Rice; Oryza sativa L.; Rice stink bugs; Oebalus pugnax (F.); Graminaceous weeds; Insect–weed interactions

1. Introduction effects on insect populations (Andow, 1991). Weeds can

Weed and insect pests are important problems faced
by rice producers worldwide. Typically research focuses
on these pests individually; however, there are numerous
ways in which they interact in crop fields. Insects can
feed on the vegetative and/or reproductive plant tissues
of weeds, possibly reducing the seed bank for following
years (Meyer and Root, 1993; Honek et al., 2003). The
presence of weeds can have both positive and negative
Corresponding author. Tel.: +1 225 578 1850;
fax: +1 225 578 1643.
E-mail address: ktindall@agcenter.lsu.edu (K.V. Tindall).
be used as alternate hosts and serve as a source of
infestation, or insects may prefer to feed on weeds and
not damage crop plants. Weeds may also interfere with
the ability of an insect to locate the crop plant.
Additionally, weeds provide a nectar source for
parasitoids and create a more diverse ecosystem with
more beneficial insects present to suppress insect pest
populations.
The rice stink bug, Oebalus pugnax (F.), is an
important insect pest of rice. Female rice stink bugs
lay two rows of barrel-shaped green eggs that turn red as
they mature on plant foliage or panicles (Odglen and
Warren, 1962; McPherson and McPherson, 2000). Early

0261-2194/$ - see front matter r 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.cropro.2005.01.023
 ARTICLE IN PRESS
992 K.V. Tindall et al. / Crop Protection 24 (2005) 991–998
in grain development, nymphs and adults damage rice
by removing the endosperm from immature kernels,
resulting in non-filled seeds. Feeding in the later stages
of grain development causes atrophied seeds and
reduces the quality of the grains (pecky rice). Pecky rice
is a broad term used to describe the appearance of
discolored kernels that results from a combination of
insect feeding and pathogen infection (Tullis, 1936;
McPherson and McPherson, 2000); several pathogens
have been isolated from pecky rice kernels (Tullis, 1936
[and sources within]; Daughtery and Foster, 1966;
Marchetti and Peterson, 1984; Hollay et al., 1987; Lee
et al., 1993). Lee et al. (1993) provided evidence of an
insect–vector relationship between rice stink bugs and
pathogens. Pecky rice and atrophied seeds reduce grain
quality because kernels are more likely to break during
the milling process (Douglas, 1939; Odglen and Warren,
1962; McPherson and McPherson, 2000). Insecticides
are the primary method of control of rice stink bugs.
Current recommendations suggest that insecticide
should be applied when numbers of rice stink bugs
reach 30 per 100 sweeps during the first 2 weeks of
heading, and 100 stink bugs per 100 sweeps after the first
2 weeks (Ring et al., 1999).
Rice stink bugs primarily feed on monocotyledonous
plants, many of which are common graminaceous weeds
associated with rice fields (Douglas, 1939; Odglen and
Warren, 1962; Nilakhe, 1976; Naresh and Smith, 1984;
McPherson and McPherson, 2000). Barnyardgrass,
Echinochloa crus-galli Beauv., Amazon sprangletop,
Leptochloa panicoides (Presl.) Hitchc., broadleaf signal-
grass, Brachiaria platyphylla Nash., large crabgrass,
Digitaria sanguinalis, (L.), bemudagrass, Cynodon dac-
tylon (L.) Pers., fall panicum, Panicum dichotomiflorum
Michx., and Cyperus spp. are common weeds in rice
agroecosystems (Jordan and Sanders, 1999). These
weeds also serve as alternate hosts for the rice stink
bug (Odglen and Warren, 1962; Nilakhe, 1976; Naresh
and Smith, 1984; McPherson and McPherson, 2000);
therefore, the potential exists for interactions to
occur between weed management and rice stink bug
management.
Previous studies showed that the presence of bar-
nyardgrass in rice fields affected the numbers of rice
stink bugs present on rice (Tindall et al., 2004).
Variations in rice stink bug densities on rice were
detected depending on the phenology of barnyardgrass
relative to rice. When barnyardgrass and rice had
panicles present at the same time; rice stink bug
infestations were lower on rice plants in the presence
of barnyardgrass than in pure stands of rice plants.
However, when barnyardgrass senescence occurred
during panicle emergence of rice, barnyardgrass served
as a source of rice stink bug infestation on the newly
emerging rice panicles. If weeds serve as a source of rice
stink bug infestation, an increase in rice stink bug
damage may be an indirect effect of the presence of
graminaceous weeds. The experiments reported here
were designed to examine the effects of varying densities
of graminaceous weeds on rice stink bug populations
and to determine if the damage from the combination of
rice stink bugs and weeds is greater than damage from
weeds alone.
2. Materials and methods
Experiments were conducted at the Macon Ridge
Branch Station, Winnsboro, LA (Franklin Parish) in
2002 and 2003. ‘Cocodrie’ rice was drill seeded into a
loessial upland soil (Gigger silt loam) at a rate of 112 kg/
ha on May 28, 2002 and May 24, 2003. The drill spacing
was 19 cm, and plots consisted of 8 rows 4.5 m in length
(7 m2/plot). Each plot was separated by a 2 m weed-free
border; weed-free borders were treated pre-emergence
with applications of 0.45 kg/ha of quinclorac and 0.55 kg
AI/ha of clomazone. On June 24, 2002 and June 30,
2003, nitrogen in the form of prilled urea was applied at
126 kg/ha immediately prior to the establishment of
permanent floods. Rice plots were flushed as needed.
The experimental design was a completely rando-
mized design with 36 plots that had varying graminac-
eous weed densities. Graminicides were applied to
assigned plots at various timings and rates to achieve a
range of graminaceous weed densities1. Additionally,
broadleaf weeds and sedges were removed from all plots
by applying 25 g AI/ha halosulfuron to all plots at the
four to five leaf stages of rice. Approximately 2 weeks
prior to panicle emergence of rice, weed density was
estimated for each plot by placing 0.1 m2 quadrants over
two rows of rice. All vegetation within the 0.1 m2 area
was removed and taken to the laboratory. Plants were
categorized by species and counted to determine the
percentage of each weed present and weed density. Two
samples were collected from each plot and averaged to
get an estimate of weed density. In 2002, the study area
had a native infestation of barnyardgrass (57%),
Amazon sprangletop (10%), and broadleaf signal grass
(33%). Weed composition in 2003 consisted of bar-
nyardgrass (38%), Amazon sprangletop (33%), broad-
leaf signalgrass (8%), and large crabgrass (22%).
After individual plots were assessed for weed compo-
sition and density, plots were divided into two groups of
18 plots of similar weed density. One group received
672 g AI/ha lamda-cyhalothrin approximately every 4 to
5 d after 20% panicle emergence of rice to minimize the
1
Herbicide programs consisted of no herbicide, 224.2, 448.3, and
672.5 g AI/ha clomazone applied pre-emergence, and 448.3 and
672.5 g AI/ha clomazone applied pre-emergence followed by 213 g AI/
ha cyhalofop at the 4–5 leaf rice stage. Herbicide treatments were
arranged in a completely randomized design.
 ARTICLE IN PRESS
K.V. Tindall et al. / Crop Protection 24 (2005) 991–998
effects of rice stink bugs so that individual effects of
weed density could be quantified. The remaining group
was not treated with insecticide. Weed density for
lamda-cyhalothrin treated plots were 0–160 weeds/1 m2
in 2002 and 0–445 weeds/1 m2 in 2003; non-treated plots
had weed densities ranging from 0–225 weeds/1 m2 in
2002 and 0–495 weeds/1 m2 in 2003.
2.1. Data collection
2.1.1. Rice stink bugs
Rice stink bug densities were sampled with a sweep
net (38 cm in diameter) after 50% panicle emergence.
Twenty sweeps were made per plot every 5–7 d for
approximately 3 weeks. Sampling was conducted by
sweeping across all rows for the length of plots. Rice
stink bugs were counted and released in the test area.
Numbers of rice stink bugs collected on each sample
date were averaged to obtain an estimate of rice stink
bug populations over the three week sampling period.
2.1.2. Percent filled seed
Prior to mechanical harvest, 30 panicles from each
plot were hand harvested to prevent loss of non-filled
seeds and seeds were manually removed from the
panicle. Rough rice samples were then divided into
groups of filled or non-filled seeds. A seed was
characterized as non-filled if, when the seed was placed
on its tip and pressure was applied with the thumb, the
palea and lemma folded easily. Filled seeds were run
through an automated seed counter (Count-A-Pak;
Seedburo Equipment Co., Chicago, IL, USA). Non-
filled seeds were manually counted since they were not
detected by the seed counter. Percent filled seeds was
determined.
2.1.3. Seed weight
After rough rice samples were separated into groups
of filled and non-filled seed, 100 filled seeds were
randomly collected and weighed to the nearest mg.
2.1.4. Yield
Yield data were collected using a mechanical harvest-
er on October 2, 2002 and September 30, 2003.
Approximately 225 g of rough rice per plot were
collected at harvest to assess pecky rice and milling
quality. Yield data were log transformed prior to
analysis to meet the assumption of normality.
2.1.5. Pecky rice
Samples of 100 g of rough rice were collected from
mechanically harvested plots. Rough rice was run
through a McGill Sheller (H.T. McGill Inc., Houston,
TX, USA) to remove the paleae and lemmas. Samples
were manually sorted to assess the incidence of pecky
rice. There are several causes of pecky rice (Tullis, 1936;
993
Lee et al., 1993), but only pecky rice caused by stink bug
feeding was recorded. Pecky rice associated with rice
stink bug feeding exhibit circular spots (Lee et al., 1993).
Pecky rice caused by rice stink bugs can range from a
small speck that would easily be removed in the milling
process to a completely pathogen-infected seed. Seeds
were considered pecky if there was any amount of stink
bug injury present; all other seeds were classified as non-
pecky. Both pecky and non-pecky rice were weighed and
weights were used to calculate percent pecky rice.
2.1.6. Milling quality
Rough rice samples of 125 g of mechanically har-
vested seed were run through a McGill Sheller to remove
the paleae and lemmas of seeds. Seeds were then run
through a McGill Miller (H.T. McGill Inc., Houston,
TX, USA) to remove the caryopsis of seeds. Rice was
weighed to the nearest 0.01 g. Milled grains were then
placed on a machine that separates broken kernels from
whole kernels. Whole kernels were weighed to estimate
the milling quality expressed as percent whole kernels.
2.2. Data analysis
All variables were analyzed using analysis of covar-
iance in PROC MIXED (SAS, 1998). Weed density was
used as the continuous variable and insecticide applica-
tion as the categorical variable. The first stage of this
analysis determines if there was a significant effect of
weed density on the Y-variable measured (i.e., does the
slope of the line equal zero?). The second phase of the
analysis incorporates the effect of the insecticide
application into the analysis and compares the intercepts
of the regression lines generated from insecticide treated
and non-treated plots (i.e., is there more damage in non-
treated plots than treated plots?). The final step of the
analysis examines the interaction between weed density
and insecticide application and evaluates the slopes of
regression lines for both treated and non-treated plots
(i.e., is the rate of increase in damage in the non-treated
plots greater than the increase in treated plots?). Data
from 2002 and 2003 were analyzed together and graphs
depict any significant year interactions. The relationship
between percent pecky rice and milling quality was
assessed using PROC REG and PROC CORR in SAS.
3. Results
3.1. Rice stink bugs
In 2002 there were more rice stink bugs present than
in 2003; however the trends were similar for both years
(Fig. 1, Table 1). As weed density increased, populations
of rice stink bug generally increased. Lambda-cyhalo-
thrin suppressed rice stink bugs in the insecticide-treated
 ARTICLE IN PRESS
994 K.V. Tindall et al. / Crop Protection 24 (2005) 991–998

plots; however, complete control was not obtained (7–9 weed density caused direct yield losses in the percentage
rice stink bugs/20 sweeps [2002], and 1–6 rice stink bugs/ of filled seeds. Additionally, there were more non-filled
20 sweeps [2003]). In the absence of insecticide, the seeds in the plots not treated with insecticide compared
presence of 13 weeds/1 m2 (2002) and 23 weeds/1 m2 to insecticide-treated plots, demonstrating that when not
(2003) caused an average increase of one rice stink bug controlled, rice stink bugs reduced the percentage of
per 20 sweeps over insecticide-treated plots. Addition- filled seeds. The insecticide–weed density interaction
ally, populations of rice stink bugs increased at a greater showed that the rate of decline in percentage of filled
rate in the plots not treated with insecticide than in the seed was more than three times greater in the rice not
insecticide-treated plots. treated with insecticide than in insecticide-treated rice.
In these experiments, a 1% reduction in percent filled
seeds occurred when 74 weeds/1 m2 were present in
3.2. Percent filled seed
insecticide-treated plots; whereas, only 24 weeds/1 m2
reduced percent filled seeds by 1% in the non-treated
Percentage of filled seeds decreased as weed density
plots.
increased at the same rate for both years (Fig. 2, Table
1). The year effect was not significant alone or in any
3.3. Pecky rice
interaction; therefore, data were pooled across years. A
significant effect was observed in both insecticide treated
As weed density increased, the incidence of pecky rice
and plots not treated with insecticide suggesting that
increased. Pecky rice was 1.6 times more common in rice
not treated with insecticide in 2002 than 2003 (Fig. 3,
80
2002 No Insecticide
No. Rice Stink Bugs / 20 Sweeps

2002 Insecticide
2003 No Insecticide 85
2003 Insecticide
60
80

75
% Filled Seeds

40

70
20
65

No Insecticide
0 60 Insecticide
0 100 200 300 400 500
Weed Density / 1 m2 55
0 100 200 300 400 500
Fig. 1. Number of rice stink bugs per 20 sweeps on insecticide treated Weed Density / 1 m2
and non-treated rice in response to increasing weed density. 2002 No
Insecticide: y ¼ 34:6063 þ 0:07800x; 2002 Insecticide: y ¼ 8:7103  Fig. 2. Percent filled seeds in insecticide treated and non-treated plots
0:00367x; 2003 No Insecticide: y ¼ 15:1257 þ 0:04325x; 2003 Insecti- in response to increasing weed density. No Insecticide: y ¼ 81:5891 
cide: y ¼ 1:1793 þ 0:01067x: Data were analyzed using analysis of 0:04800x; Insecticide: y ¼ 82:8714  0:01355x: Data were analyzed
covariance; see text for explanation. using analysis of covariance; see text for explanation.

Table 1
ANOVA table presenting the statistical results from analysis of covariance for the variables, the number of rice stink bugs, percent filled seeds and
pecky rice, milling quality, seed weight and yield

No. of RSB % Filled seeds % Pecky rice Milling quality Seed Wt. Yield

F df P F df P F df P F df P F df P F df P

Y 57.96 1.63 o.001 2.60 1.64 .112 10.32 1.63 .002 23.83 1.61 .112 9.72 1.64 .003 4.80 1.63 .032
WD 31.27 1.63 o.001 22.81 1.64 o.001 9.56 1.63 .003 20.65 1.61 o.001 0.07 1.64 .787 55.88 1.63 o.001
Y*WD 6.30 1.63 .015 3.42 1.64 .069 0.56 1.63 .046 3.49 1.61 .067 12.04 1.64 .001 0.02 1.63 .893
I 196.2 1.63 o.001 6.38 1.64 .014 69.32 1.63 o.001 16.29 1.61 o.001 12.42 1.64 .001 2.14 1.63 .148
I*WD 3.48 1.63 .067 4.72 1.64 .034 0.30 1.63 .585 0.06 1.61 .815 0.51 1.64 .476 1.64 1.63 .205
I*Y 19.16 1.63 o.001 2.09 1.64 .153 6.23 1.63 .015 0.00 1.61 .996 17.27 1.64 o.001 0.12 1.63 .728
Y*I*WD 1.94 1.63 .169 0.00 1.64 .982 0.50 1.63 .831 0.06 1.61 .804 0.02 1.64 .902 0.18 1.63 .669

Experiments were conducted in 2002 and 2003 (Y) with a range of weed densities (WD) in the presence and absence of insecticide (I).
 ARTICLE IN PRESS
K.V. Tindall et al. / Crop Protection 24 (2005) 991–998 995

22 88
2002 No Insecticide 2002 No Insecticide*
20 2002 Insecticide 2002 Insecticide
2003 No Insecticide 86 2003 No Insecticide
18 2003 Insecticide
2003 Insecticide*
16 84
% Pecky Rice

% Whole Kernels
14 82
12
80
10
8 78
6
76
4
2 74
0 100 200 300 400 500
72
Weed Density / 1 m2
0 100 200 300 400 500
Fig. 3. Percent pecky rice in insecticide treated and non-treated plots Weed Density / 1 m2
in response to increasing weed density. 2002 No Insecticide: y ¼
14:69 þ 0:009943x; 2002 Insecticide: y ¼ 4:7628 þ 0:009943x; 2003 No Fig. 4. Milling quality expressed as whole kernels in insecticide treated
Insecticide: y ¼ 8:9538 þ 0:009943x; 2003 Insecticide: y ¼ 3:7008 þ and non-treated plots in response to increasing weed density. * Lines
0:009943x: Data were analyzed using analysis of covariance; see text for 2002 No Insecticide and 2003 Insecticide overlap. 2002 No
for explanation. Insecticide: y ¼ 84:3231  0:01767x; 2002 Insecticide: y ¼ 87:8891 
0:01767x; 2003 No Insecticide: y ¼ 80:6246  0:01767x; 2003 Insecti-
cide: y ¼ 84:3734  0:01767x: Data were analyzed using analysis of
covariance; see text for explanation.
Table 1). Pecky rice from plots not treated with
insecticide was 2.4 (2003) to 3 (2002) times higher than
that from insecticide-treated plots. Regardless of the 2.4
year or insecticide application, pecky rice increased at a 2002 No Insecticide
2002 Insecticide
rate of 1% for every 100 weeds/1 m2. 2003 No Insecticide
2.3 2003 Insecticide

3.4. Milling quality

Seed Wt (mg)

2.2
Milling quality, expressed as percent whole kernels,
revealed that as weed density increased, quality decreased;
56 weeds/1 m2 reduced the milling quality by 1% (Fig. 4, 2.1
Table 1). The percentage of whole kernels was 3.6 and 3.7
times lower in plots not treated with insecticide than in
2.0
insecticide-treated plots in 2002 and 2003, respectively.
Percent pecky rice and milling quality were significantly 0 100 200 300 400 500
correlated (Pearson correlation coefficient ¼ 0:4924;
Weed Density / 1 m2
Po0:0001). Data suggest that for every 1% of pecky rice,
milling quality declined by nearly 0.5% when milling Fig. 5. Weight of 100 seeds from insecticide treated and non-treated
quality was regressed against percent pecky rice plots in response to increasing weed density. 2002 No Insecticide: y ¼
(y ¼ 0:44772x  4:63; Po0:0001). 2:3422  0:000412x; 2002 Insecticide: y ¼ 2:1845  0:000412x; 2003
No Insecticide: y ¼ 2:1499 þ 0:0001902x; 2003 Insecticide: y ¼
2:1434 þ 0:0001902x: Data were analyzed using analysis of covariance;
3.5. Seed weight see text for explanation.

There was a weak response in seed weights. Seed

weights responded differently in 2002 and 2003 (Fig. 5,
Table 1). In 2002, seed weights decreased as weed
density increased; however, 2430 weeds/1 m2 were esti-
mated to reduce seed weight by 1 mg for both
insecticide-treated and non-insecticide-treated rice. Seed
weights were 0.16 mg lower in non-treated rice than
treated rice. In 2003, seed weight appeared to increase in
response to increasing weed density; the regression
equation estimated that 5260 weeds/1 m2 were needed
to cause an increase of 1 mg for both treated and
non-treated rice. Seed weights were 0.006 mg greater in
the non-treated rice than in the insecticide-treated rice.
The estimated weed densities required to cause an effect
on seed weight were outside the weed densities examined
in these experiments.
3.6. Yield
As weed densities increased, grain yields decreased in
both 2002 and 2003; however, the loss was more severe

996 K.V. Tindall et al. / Crop Protection 24 (2005) 991–998
9.0
log[yield (kg / ha)]
8.5
8.0
7.5
7.0
0 100 200 300
Weed Density / 1 m2
Fig. 6. Yield loss in response to increasing weed density. 2002: y ¼
8:6214  0:002978x; 2003: y ¼ 9:0414  0:002978x: Data were ana-
lyzed using analysis of covariance; see text for explanation.
in 2002 (Fig. 6, Table 1). Regardless of year, yield losses
were ca. 1% per 30 weeds/1 m2. Data from the present
study showed that yield losses were 163 and 248 kg/ha in
2002 and 2003, respectively. Insecticide had no effect on
yield losses.
4. Discussion
Our results demonstrate the influence of insect–weed
interactions and their impact on rice stink bug damage
to rice. Increases in weed densities led to increases in
densities of rice stink bugs and rice injury. Fewer filled
seeds, more pecky rice, and lower milling quality were
observed in plots with higher densities of weeds than
those with lower densities. Although there was a
decrease in filled seeds in plots not treated with
insecticide, yield losses were not significantly different
than that for insecticide-treated weedy rice, suggesting
that weeds were the more important of these two pests
with respect to reducing grain yields in these experi-
ments.
The data for percent filled seeds revealed several
pieces of information. First, the presence of weeds,
irrespective of insects, reduced the amount of filled
seeds. Donald and Khan (1996) found similar results
with spring wheat; numbers of seeds per spike were
reduced as densities of thistles increased. Second, there
were more non-filled seeds in plots not treated with
insecticide compared to that in insecticide-treated plots,
suggesting that when not controlled, rice stink bugs also
reduced the amount of filled seeds. The reduction in
percent filled seeds was 3.6 times greater in non-treated
rice than in treated rice. Previous research has shown
that rice stink bugs feed on weeds in areas of rice fields
that have not yet produced panicles and that this
behavior could lead to earlier infestations on rice plants
ARTICLE IN PRESS
in weedy fields (Tindall et al., 2004). Weeds in this study
2002 had seed heads emerge approximately 3.5 weeks before
2003
rice panicle emergence, and rice stink bugs were
observed in weedy areas of rice prior to panicle
emergence of rice. Seed heads of weeds were beginning
senesce as rice panicles began to emerge (personal
observation). Therefore it is likely that instead of rice
stink bugs emigrating from weeds to locate a suitable
host, rice stink bugs were able to infest adjacent rice
plants at vulnerable stages of development, anthesis and
early grain filling. These findings support previous
results that showed rice infested with rice stink bugs at
400 500 anthesis resulted in severe injury and prevented further
grain development (Lee et al., 1993). Rice stink bug
infestations 1 day after anthesis reduced filled seeds by
approximately 40% (Patel and Stout, unpublished
data).
Pecky rice also increased as weed density increased in
both insecticide-treated and non-treated rice. In non-
treated plots, the increase in pecky rice could be
explained by the increase in rice stink bug densities.
The increase in pecky rice in plots treated with
insecticide, however, was not expected. This result
suggests that weeds may also have played an important
role in the amount of pecky rice detected in these
experiments. Weeds may serve as a source of inoculum
and their presence may enhance populations of patho-
gens. A review of published literature (Tindall, 2004) on
the host range of pathogens isolated from discolored
rice kernels showed that 57% of these pathogens are
known to infect several genera of common weeds of rice.
Moreover, several pathogens that have been isolated
from discolored rice kernels have also been isolated
from seeds of Echinochloa spp. of weeds (Huelma et al.,
1996) and Curvularia lunata, one of the more commonly
isolated pathogens, has also been documented to
discolor seeds of an Echinochloa sp. (Joshi and Gupta,
1980).
Although weed plants in these experiments were not
sampled for pathogens, it is likely that these organisms
were present on weeds and that rice stink bugs fed on
seed heads of weeds infected with pathogens. Several
pathogens have been collected from stylets (mouth-
parts), saliva, and feeding sheaths (saliva remaining at a
feeding site) of rice stink bugs (Hollay et al., 1987; Lee et
al., 1993); therefore, pathogens may have remained on
their stylets after feeding on diseased alternate hosts.
After coming in contact with pathogens and moving to
rice, rice stink bugs could transfer pathogens when
feeding on rice, causing the incidence of pecky rice
observed in these studies. Lee et al. (1993) found
13–80% of kernels subjected to simulated rice stink
bug feeding in the presence of pathogens were dis-
colored, whereas only 2% of kernels were discolored in
the presence of pathogens alone. Therefore, though
possible for pecky rice to occur in the presence of only a
 ARTICLE IN PRESS
K.V. Tindall et al. / Crop Protection 24 (2005) 991–998
pathogen, penetration of pathogens is enhanced in the
presence of rice stink bugs. This hypothesis also
supports the fact that fungicide applications do not
reduce the incidence of pecky rice (Lee et al., 1993). If
rice stink bugs are a major contributing factor of pecky
rice and they are capable of migrating into a field, even if
a producer applies a fungicide, pecky rice could be
present if rice stink bugs acquired pathogens from
alternate hosts some distance away. The phenomenon
that infected weeds can serve as a source of inoculum for
insect transmitted diseases has been previously docu-
mented with black nightshade, Solanum nigrum L.
and the green peach aphid, Myzus persicae (Sulzer.)
in small plot studies with bell peppers (Fereres et al.,
1996).
Pecky rice is known to affect milling quality of rice
(Douglas, 1939; Odglen and Warren, 1962; McPherson
and McPherson, 2000); therefore, a negative relation-
ship between these two variables was expected. Regres-
sion analysis suggests that every one percentage of pecky
rice reduced milling quality by half a percent. These data
probably underestimate losses because samples exam-
ined in these experiments included minor rice stink bug
damage that would probably not be considered pecky
rice at a commercial mill; kernels with minor injury are
less likely to break during the milling process than those
with severe rice stink bug damage.
The effect of weed density on rice yields is well
documented (Smith, 1988) and yield losses were an
expected result. Smith (1988) and Tindall et al. (2003)
showed that yields were reduced by 65–71 kg/ha for
every one barnyardgrass plant/m2. Experiments relating
barnyardgrass density to rice yields have shown yield
losses result from a reduction in the number of tillers,
panicles, and seeds per panicle (Tindall et al., 2003).
Yield losses from the present study were 163 and 248 kg/
ha in 2002 and 2003, respectively. Yield losses from this
study are likely higher than those from previous studies
since a complex of graminaceous weeds was examined,
and not an individual weed species. On the other hand,
applications of insecticides to control rice stink bug had
no significant effect on grain yield losses. This suggests
that, although there was a reduction in percent filled
seeds as a result of rice stink bugs, the reduction in filled
seeds did not significantly contribute to an overall yield
loss. Harper et al. (1993) also found that grain yield
losses from rice stink bugs were not significant and that
reduction in quality was the major loss attributed to rice
stink bugs.
Rice stink bug populations were influenced by weed
density; however, rice stink bugs appeared to have a
larger effect on quality than on yield of rice. Weed
density was shown to have the greatest impact on yield.
Available literature suggests that graminaceous weeds,
like barnyardgrass and broadleaf signalgrass, are more
competitive with rice prior to panicle emergence (Smith,
997
1974; McGregor et al., 1988; Azmi and Mashhor, 1992).
Graminaceous weeds can also interfere with rice during
the late season by increasing numbers of rice stink bugs
and may also serve as a source of inoculum of pathogens
that cause to pecky rice. Therefore, weed management
throughout the season appears to be important to
maximize rice yield and quality.
Acknowledgments
We would like to express appreciation to Boris
Castro, Don Groth, Donnie Miller, and Boyd Padgett
for their critical reviews of the manuscript. We would
also like to thank John Bernhardt for his assistance with
the determination of pecky rice and the milling lab at the
LSU AgCenter Rice Research Station for assistance
with milling rice samples. The LA Rice Research and
Promotion Board and USDA Southern Region IPM
Grants Program (Project No. LAB03567) funded this
research. Approved for publication by the director of
the LSU AgCenter, manuscript number 04-26-0301.
References
Andow, D.A., 1991. Vegetational diversity and arthropod population
response. Annu. Rev. Entomol. 36, 561–586.
Azmi, M., Mashhor, M., 1992. Competition of barnyardgrass
(Echinochloa crus-galli (L.) Beauv.) in direct seeded rice. Proceed-
ings of the Third International Conference on Plant Protection in
the Tropics, Genting Highlands, Malaysia, 20–23 March 1990, vol.
6, pp. 224–229.
Daughtery, D.M., Foster, J.E., 1966. Organism of yeast-spot disease
isolated from rice damaged by rice stink bug. J. Econ. Entomol. 59,
1282–1283.
Donald, W.W., Khan, M., 1996. Canada thistle (Cirsium arvense)
effects on yield components of spring wheat (Triticum aestivum).
Weed Sci. 44, 114–121.
Douglas, W.A., 1939. Studies of rice stinkbug populations with special
reference to local migration. J. Econ. Entomol. 32, 300–303.
Fereres, A., Avilla, C., Collar, J.L., Duque, M., Fernández-Quinta-
nilla, C., 1996. Impact of various yield-reducing agents on open-
field sweet peppers. Environ. Entomol. 25, 983–986.
Harper, J.K., Way, M.O., Dress, B.M., Rister, M.E., Mjelde, J.M.,
1993. Damage function analysis for the rice stink bug (Hemiptera:
Pentatomidae). J. Econ. Entomol. 86, 1250–1258.
Hollay, M.E., Smith, C.M., Robinson, J.F., 1987. Structure and
formation of feeding sheaths of rice stink bug (Heteroptera:
Pentatomidae) on rice grains and their association with fungi. Ann.
Entomol. Soc. Am. 80, 212–216.
Honek, A., Martinkova, Z., Jarosik, V., 2003. Ground beetles
(Carabidae) as seed predators. Eur. J. Entomol. 100, 531–544.
Huelma, C.C., Moody, K., Mew, T.W., 1996. Weed seeds in rice seed
shipments: a case study. Int. J. Pest Manage. 42, 147–150.
Jordan, D., Sanders, D.E., 1999. Weed management. In: , Louisiana
Rice Production Handbook, vol. 2321. The LA Cooperative
Extension Service Publishers, Baton Rouge, LA, pp. 37–50.
Joshi, D.N., Gupta, S.C., 1980. Studies on seed mycoflora and its role
in causing diseases of Echinochloa frumentacea. Ind. Phytopath. 33,
433–435.
 ARTICLE IN PRESS
998 K.V. Tindall et al. / Crop Protection 24 (2005) 991–998
Lee, F.N., Tugwell, N.P., Fannah, S.J., Weidemann, G.J., 1993. Role
of fungi vectored by rice stink bug (Heteroptera: Pentatomidae) in
discoloration of rice kernels. J. Econ. Entomol. 86, 549–556.
Marchetti, M.A., Peterson, H.D., 1984. The role of Bipolaris oryzae in
floral abortion and kernel discoloration. Plant Dis. 68, 288–291.
McGregor Jr., J.T., Smith Jr., R.J., Talbert, R.E., 1988. Broadleaf
signalgrass (Brachiaria platyphylla) duration of interference in rice
(Oryza sativa). Weed Sci. 36, 747–750.
McPherson, J.E., McPherson, R.M., 2000. Oebalus spp. In: Stink Bugs
of Economic Importance in America North of Mexico. CRC Press,
Boca Raton, pp. 141–158.
Meyer, G.A., Root, R.B., 1993. Effects of herbivorous insects and soil
fertility on reproduction of goldenrod. Ecology 74, 1117–1128.
Naresh, J.S., Smith, C.M., 1984. Feeding preference of the rice stink
bug on annual grasses and sedges. Entomol. Exp. Appl. 35, 89–92.
Nilakhe, S.S., 1976. Overwintering, survival, fecundity, and mating
behavior of the rice stink bug. Ann. Entomol. Soc. Am. 69, 717–720.
Odglen, G.E., Warren, L.O., 1962. The rice stink bug Oebalus pugnax
F. in: Arkansas. University of Arkansas, Fayetteville Ag. Exp.
Station Report Series 107.
Ring, D.R., Barbour, J., Rice, W.C., Stout, M., Muegge, M., 1999.
Insect management. In: Louisiana Rice Production Handbook,
pub. 2321. The LA Cooperative Extension Service Publishers,
Baton Rouge, LA, pp. 85–93.
SAS Institute, 1998. User’s manual, version 7.0. SAS Institute, Cary,
NC.
Smith Jr., R.J., 1974. Competition of barnyard grass with rice
cultivars. Weed Science 22, 423–426.
Smith, R.J., 1988. Weed thresholds in southern U.S. rice, Oryza sativa.
Weed Technol. 2, 232–241.
Tindall, K.V., 2004. Investigation of insect–weed interactions and
insect and weed management practices in the rice agroecosystem.
Ph.D. Dissertation, Louisiana State University, Baton Rouge, LA.
Tindall, K.V., Williams, B.J., Webster, E.P., Stout, M.J., 2003. Effects
of barnyardgrass density on rice yields. In: Proceedings
Southern Weed Science Society Annual Meeting, Houston, TX.
January 27–29, 2003, Southern Weed Science Society, Raleigh,
pp. 306.
Tindall, K.V., Stout, M.J., Williams, B.J., 2004. Effects of the presence
of barnyardgrass on rice water weevil (Coleoptera: Curculionidae)
and rice stink bug (Hemiptera: Pentatomidae) populations on rice.
Environ. Entomol. 33, 720–726.
Tullis, E.C., 1936. Fungi isolated from discolored rice kernels. USDA
Technology Bulletin, vol. 540. pp. 11.