International Research Journal of Microbiology (IRJM) (ISSN: 2141-5463) Vol. 2(8) pp.

285-291, September 2011

Available online http://www.interesjournals.org/IRJM
Copyright 2011 International Research Journals

Full Length Research Paper

Recovery of Bacillus thuringiensis from forest soil after

application for gypsy moth control using susceptible
Insect larvae
Phyllis. A.W. Martin*, Elizabeth A. Mongeon, Dawn E. Gundersen-Rindal and Michael B.
Blackburn
USDA/ARS Invasive Insect Biocontrol and Behavior Laboratory Beltsville, MD 20706
Accepted 08 August, 2011

Bacillus thuringiensis was recovered from forest soil sprayed two years previously with B.
thuringiensis for gypsy moth control by amplifying the bacteria found in the soil on bacterial agar and
feeding this mixed microbial population to tobacco hornworm larvae. These bacteria killed the larvae
when fed on artificial diet. Most of the bacteria recovered from dead insects formed spores and
belonged to the Bacillus cereus group (14) including three B. thuringiensis and one B.
weihenstephanensis. All the B. thuringiensis strains and most of the B. cereus strains were closely
related (<0.001 substitutions/site in the 16S rDNA) to the strain that was sprayed. Ten B. cereus sensu
lato strains differed from the B. thuringiensis strain applied by five or fewer, of the 24 traits tested. The
toxicity to gypsy moth larvae of the three B. thuringiensis strains isolated was similar to the applied B.
thuringiensis strain. Thus, amplification of bacteria present in soil in combination with an insect
susceptible to this strain can recover insect toxic strains that are related to an applied strain.
Keywords: Gypsy

moth,

biological

control,

16S

rDNA

gene

sequences

INTRODUCTION
Bacillus thuringiensis has been applied to many locations
repeatedly over many years to control periodic outbreaks
of gypsy moth, Lymantria dispar (Sharov et al., 2002,
APHIS factsheet 2003). Although there are techniques
that can recover B. thuringiensis from the environment
(Ohba and Aizawa, 1986, Travers et al., 1987), they rely
on isolating the bacteria first and then testing for toxicity.
However, an alternative method can be developed
around assay of mixed cultures grown directly from soil.
When the bacteria occur at low levels, it is often
necessary to have an enrichment step to detect the
bacteria
of interest.
Traditionally
this
has
Involved the use of selective media (Akiba and Katoh

*Corresponding author E-mail: Phyllis.martin@ars.usda.gov;

Phone: 301-504-6331; Fax: 301-504-5725

1986, Bizzari and Bishop 2007, Andrzejczak and Lonc

2008). Alternatively, molecular techniques with specific
probes for the toxin genes could be used, but these
techniques have their own limitations when these
bacteria exist as spores; molecular techniques often
underestimate their abundance due to procedural
difficultly in lysis of the spore stage (Ankolekar et al.,
2009).
In Maryland, at a site in the Frederick Municipal forest,
B. thuringiensis has been sprayed repeatedly over
several years for control of gypsy moth (R. Tichenor,
Maryland Department of Agriculture pers. comm). In a
previous study, we used toxicity to Manduca sexta L. as a
model (Silva et al., 2002), to isolate lepidopteran
pathogens from soil (Martin et al., 2008). In this study, we
describe the use of toxicity to M. sexta as a means to
determine the survival of a commercially applied toxic B.
thuringiensis strain in a forest environment.

286 Int. Res. J. Microbiol.

MATERIALS AND METHODS

Bacterial strains and media
Three B. thuringiensis strains isolated from commercial
preparations were used as controls for toxicity and
phenotypic properties. B. thuringiensis subsp. kurstaki,
strain IBL 455, was originally isolated from a 1980
preparation of Dipel (Abbott Laboratories, Chicago, IL).
Strain IBL 451 was isolated from a 2001 preparation of
Foray (Novo Nordisk, Danbury, CT) which was applied
to the Fishing Creek site in 2001. IBL 451 was used as a
control for toxicity testing and for comparison to the
bacteria recovered from the environment. Strain IBL
1410, B. thuringiensis subsp. tenebrionis, was isolated
from a 1991 preparation of Novodur (Mycogen, San
Diego, CA) and was used along with IBL 455 as a
controls for phenotypic tests. IBL 10003, which was IPS
82, the B. thuringiensis subsp. israelensis international
standard fermented in 1982 was used for comparisons of
16S rDNA gene sequences.
Bacteria were grown on L-agar (Atlas, 2003) or RM (
L) for enumeration and recovery.
If crystals were
detected, bacteria were grown on T3 (Travers et al.,
1987) for maximum crystal production for insect assays.
Biochemical media were as described for Bacillus and
modified for B. thuringiensis (Martin et al., 1985; Martin et
al., 2010). Fourteen different media were used to test for
various phenotypic traits including acid production from
glucose, arabinose, xylose, mannitol, mannose, salicin
and sucrose, utilization of citrate, hydrolysis of esculin,
production of protease, amylase, urease, phosopholipase
C, and hemolysin. Only those traits that differed among
isolated strains are discussed.
Additionally, eight antibiotic susceptibilities which we
anticipated might be more variable than other phenotypic
properties (ampicillin, vancomycin, chloramphenicol,
triple sulfa, neomycin, tetracycline and kanamycin) were
tested using disk diffusion (Remel, Lenexa, KS) on Lagar. Zones were measured after incubation at 30C for
24h and levels of sensitivity: sensitive, intermediate and
resistant determined by the information supplied for each
antibiotic.
Soil samples
Twenty soil samples were collected along Fishing Creek
in the City of Frederick(MD) Municipal Forest in July 2004
and stored at ambient temperature in sterile plastic bags.
Soil was suspended 1:10 wt: vol in sterile distilled water,
shaken on a vortex mixer for 30 minutes, and plated for
initial enumeration of bacteria. An undiluted sample of
this soil suspension was spread on RM and incubated at
25C for 48 h, increasing the number of culturable
bacteria available from each sample (Martin et al.,

2008). The bacteria harvested from these plates were fed

to M. sexta larvae. Bacteria recovered from these soil
suspensions harvested after growth were enumerated on
RM.
For comparison to this technique, bacteria isolated from
other soil samples (614) that were previously taken in
Maryland were compared to the samples taken along
Fishing Creek. B. thuringiensis and other sporeformers
were isolated from these samples using acetate selection
(Travers et al., 1987). Phenotypes were tested as
described (Martin et al., 2010). In 2009, 75 additional soil
samples were taken from the same area for comparison
of techniques. Bacteria from these soil samples were
suspended in water, a 100 l sample heated for 3 min at
80C and plated on L-agar. Colonies were grown
overnight, transferred to T3, incubated for 48-72 h at
30C and checked for spores and crystals using
Nomarski optics.
Insects and bioassays
For bioassays, tobacco hornworm diet (THW diet) was
used as re-hydrated freeze dried pellets (Martin 2004,
Martin and Blackburn 2007). Manduca sexta L. eggs
were received from J. Pennington (U. Arizona) and
reared on strips of THW diet at 24 C, 46% RH, and 16:8
light: dark cycle until the 2nd instar. Diet was changed
every 2-4 days.
Sixteen diet pellets were used for each treatment in
bioassays with one diet pellet/well (1.6 cm diameter x 1.6
cm deep) in white plastic bioassay trays (C-D
International, Ocean City, NJ). Pellets were re-hydrated
with 0.3 ml of water (controls) or suspensions containing
extracts of soil or dilutions of bacteria. One tobacco
hornworm, 2nd instar larva, was added to each pellet in
individual wells. Wells were sealed with a clear plastic
film and holes made in the film with insect pins for oxygen
transfer. Larvae were incubated as for rearing. Mortality
was recorded as early as 16, 24 and 48 h.
Gypsy moths were received as egg masses from
USDA/ARS, Otis Air National Guard Base (MA). Eggs
were hatched and larvae reared to 2nd instar on a wheat
germ-based diet (Bell et al. 1981) at 25C on a 16:8 h L:D
cycle without humidity control.
For initial toxicity testing 0.3 ml of each dilution of each
B. thuringiensis strain was added to gypsy moth diet on
16 freeze dried pellets. Mortality was recorded daily.
LC50s were done on crystal forming strains against gypsy
moths using the lepidopteran active IBL 455 and IBL 451
as standard strains for comparison.
Recovery of bacteria from insects
From each treatment with larval mortality, one dead larva

Martin et al. 287

was used to recover bacteria. Each dead larva was

surface sterilized by dipping for 3 s in 10% bleach, rinsed
in sterile distilled water, placed in a sterile plastic bag with
five ml sterile distilled water, and ground in a stomacher
blender (Techmar, Cincinnati, OH) for 60 s on high. The
bacteria isolated from the larval extract were enumerated
on RM. Plates were incubated at 25C for 48 h.
Colonies were enumerated and the predominant colony
type was isolated and checked for the formation of
spores and crystals by phase-contrast microscopy. These
colonies were then characterized by their substrate
utilization profiles and sensitivity to antibiotics.
Statistical analysis
Using protein concentration as dose, the LC50 was
determined using probit analysis (SAS Inc. 2008) for
each strain.
Bacterial Identification and cry analysis
In addition to characterization of spores and crystals by
phase contrast microscopy and strains by substrate
utilization, individual isolates were identified by PCR
amplification and sequencing of conserved 16S rDNA
genes. For each isolate, DNA was purified from a 2 ml
culture grown only 8 hours. DNA was isolated using the
Quantum Prep miniprep kit (BioRad, Hercules, CA) as
specified by the manufacturer for use as a template in
polymerase chain reaction (PCR). Nearly full length 16S
rDNA was amplified for each isolate using primers
universal to prokaryotes, R16F0 and R16R0 (Lee et al.
1993). Thirty-five PCR cycles were conducted in a model
9700 thermocycler (Applied Biosystems, Foster City, CA).
The amplification primers and a nested universal primer
533F (5'-GTGCCAGCMGCCGCGGTAA-3') using 30 sec
denaturation at 94C, 1.5-min annealing at 55C, and 2min primer extension (10-min in final cycle) at 72C.
Bacterial 16S rRNA gene amplicons were sequenced
directly. Products were separated on 1.5% NuSieve
agarose gel (FMC, Rockland, ME) in modified-1X TAE
(0.04 M Tris-acetate and 0.1 mM EDTA), and excised for
sequencing using ABI BigDye V1.1 Automatic
sequencing was carried out on an ABI Prism Model 3100
(Applied Biosystems, Foster City CA). Sequences were
edited and assembled (DNASTAR, SeqMan component);
BLAST (Altschul et al., 1990) searches were conducted
to identify bacterial isolates based on rDNA sequence
homology to known bacteria. Sequences thus obtained
for
16S rDNA were deposited in GenBank with
accession numbers EU168402- EU168418. Strains were
also tested for cry1 gene using degenerate cry1 primers
(Juarez-Perez et al., 1997).

RESULTS
Bacteria grown from various soil samples and fed to M.
sexta killed the larvae within 48 h. Larval mortality was
100% for 17 of 20 samples. While the initial bacterial
concentration applied to diet pellets was high, the larvae
did not noticeably consume any of the diet pellets before
they died. The bacterial concentration averaged 1.11+/0.82 x108cfu (colony forming units)/ diet pellet and ranged
from 9.9 x 106 to 3.0 x 108cfu/ diet pellet. The number of
bacteria recovered from various larvae averaged 3.02 +/2.6 x 107 cfu/larva and ranged from 5.0 x105 to 8.2
x107cfu/larva. For most larvae only a single colony type
was obtained. For others, one colony type predominated.
The major colony type was isolated, characterized by
phenotypic tests and fed to other M. sexta larvae. One
sample (10) had two colony types of approximately the
same numbers. One type made round spores that
swelled the sporangia and the other type made oval
spores and a bipyramidal crystal. Only the bacteria
which had 100% mortality when refed made crystals.
Bacteria (15 of 20) recovered from dead larvae from the
second round were found to be identical in phenotype to
the bacteria isolated initially. Once a pure culture was
isolated it was assigned a strain designation. These
strains were identified to species by 16S rDNA
sequencing.
Relatedness based on 16S rDNA genes
Putative identities based on nearly full 16S rRNA genes
showed that 14 strains belonged to the Bacillus cereus
species complex (Figure 1). Three strains (IBL 1056, IBL
1067, IBL 1445) also formed bipyramidal crystals upon
sporulation (from three different samples 10, 8, and 18)
and were identified as B. thuringiensis. Another spore
former, IBL 1077 (from sample 5), was identified as B.
weihenstephanensis and grew at 4 C. A third spore
forming strain (IBL10B1445), which was co- isolated with
crystal forming IBL 1445, formed spherical to oval spores
that swelled the sporangia and was identified as
Lysinibacillus fusiformis (Ahmed et al., 2007).
The relatedness of these strains based on their 16S
rRNA genes are shown in Figure 1. The B. thuringiensis
and the B. cereus strains clustered closely together
differing by fewer than 5 bases in approximately a 1400
base sequence. Therefore a method for more precisely
differentiating stains was needed.
Relatedness based on phenotypes
Phenotypes (substrate utilization and antibiotic resistance
profiles) were chosen to distinguish the strains in the B.

288 Int. Res. J. Microbiol.

Figure1: Dendrogram showing the relationship of Bacillus strains isolated from dead
larvae to known B. thuringiensis strains

cereus group because previous as well as ongoing

research has shown that the closely related strains of B.
thuringiensis can be differentiated using these
characteristics (Martin and Travers 1989, Martin et al.
2010). The three B. thuringiensis strains recovered from
M. sexta larvae differed from the strain that had been
sprayed (IBL 451) by two (IBL 1445 - kanamycin and
tetracycline sensitivity), four (IBL 1056 urease
production, esculin hydrolysis; triple sulfa and kanamycin
sensitivity;) or five (IBL 1067- acid production from
salicin; vancomycin, triple sulfa, neomycin and
tetracycline sensitivity) phenotypic traits (Table 1). The
four most closely related B. cereus strains differed from
the B. thuringiensis strain that had been sprayed by three
(IBL 1050, 1058, 1075 and 1078) phenotypic traits other
than crystal production (Table 1).

Relatedness based on cry genes and toxicity

The initial mixture of B. thuringiensis strain IBL 1445 and
L. fusiformis strain IBL 10B1445 was toxic when fed to M.
sexta larvae. The L. fusiformis strain, however, was not
toxic when fed separately, but the B. thuringiensis strain
was toxic when fed separately. All the crystal forming
strains had cry1 genes by PCR. One strain that did not
form crystals, IBL 1050, also had a cry1 gene.
The toxicity of recovered B. thuringiensis strains to
gypsy moth was not significantly different than the strain
that was applied (IBL 451). The LC50 for the applied strain
(IBL 451) in pure culture was 40ng protein/diet pellet,
12ng protein/diet pellet for IBL 1067 and 38ng/diet pellet
for IBL 1056. IBL 1050 which did not make a crystal killed
60% of the M. sexta larvae at a screening concentration

Martin et al. 289

Table 1: Phenotypes of Bacillus cereus group strains recovered (bold) applied (highlighted) and control strains

Soil no.
10
8
18
12
21
11
19
16
17
14
7
1
9
5

IBL no
451
1445
1067
1056
1050
1439
1075
1058
1078
1080
1054
1065
1048
1068
1077

T
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+

L
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+

U
+
+
+
+
+
+
+
+

S
+
+

A
+
+
+
+
+
+
+
+
+
+
+

E Am van cam ery 3S neo

+ R
S S S S S
+ R
S S S S S
+ R
I
S S I I
- R
S S S R S
+ R
S S S S S
- R
S S
I I I
+ R
S S S S S
+ R
S S S R S
+ R
S S S I S
+ R
S
I
S R I
+ R
S S S I R
- R
S S S I I
- R
S S S I S
+ R
S S S I I
- R
S S S S I

control
control
control

10003
1410
455

+
+
+

+ - - - - - + - + + - + +

R
S
R

S
S
S

S
I
S

S S S
S S I
S S S

tet
S
I
I
S
I
S
I
S
S
S
I
S
S
R
S
S
S
R

kan Species
R
Bt
I
Bt
R
Bt
S
Bt
I
Bc
S
Bc
I
Bc
S
Bc
S
Bc
I
Bc
I
Bc
I
Bc
S
Bc
S
Bc
I
Bw
I
I
I

Bt
Bt
Bt

Abbreviations used, T - amylase production on starch, L- Lecithinase or phospholipase C production, U- production

of urease, S acid production from sucrose, A- acid production from salicin, E- hydrolysis of esculin. Antibiotics
amp- 10 ug ampicillin, van, 30 ug vancomycin, cam 30 ug chloramphenicol, ery, 15 ug erythromycin, 3S 1 mg triple
sulfa, neo 30 ug neomycin, tet, 30 ug tetracycline kan 30 ug kanamycin. Acid production from mannitol, arabinose
and xylose was negative for all strains as was utilization of citrate. All strains were hemolytic, produced proteases
and produced acid from glucose.

of 107cfu/ diet pellet.

production from salicin and esculin hydrolyisis)

phenotypes were recovered from areas previously known
to be sprayed with B. thuringiensis.

Comparison with the standard method

DISCUSSION
From the retrospective samples, taken from Maryland
soils by acetate selection, 165 samples contained more
than 10 B. thuringiensis isolates and 189 contained no B.
thuringiensis isolates. Eleven isolates had the same
phenotype as the strain applied (without antibiotic tests).
From 75 soil samples collected in 2009, B. thuringiensis
was recovered from 41 samples. Only 22 samples had
spore forming bacteria which made bipyramidal crystals
and only 13 of these from 4 samples had the same
phenotypic profile (Martin et al. 2010) as the B.
thuringiensis strain applied.
Comparing all the isolates, 725 B. thuringiensis were
isolated along with 304 spore forming bacteria that did
not make crystals. The most common phenotypes are
shown in Table 2. All but 3 of the 67 TLUAE (starch
utilization, lecithinase and urease production, acid

Using a pest insect sensitive to microbial insecticides, the

tobacco hornworm, three distinct crystal forming B.
thuringiensis strains and 12 other strains in the B. cereus
group were recovered from 15 different sample sites.
These sites had been sprayed with a commercial
formulation of B. thuringiensis five different times in the
22 years prior to our sampling (1983, 1986, 1989, 1995
and 2001; R. Tichenor, Maryland Department of
Agriculture, pers. comm.). Two of the three crystal
formers recovered from the application sites (IBL 1445
and IBL 1067) were highly similar to the applied B.
thuringiensis var. kurstaki (Foray; IBL 451), by patterns of
substrate utilization, toxicity, antibiotic susceptibility, and
16S rRNA gene sequence. The third, IBL 1056, differed
in its antibiotic resistance profile and did not

290 Int. Res. J. Microbiol.

Table 2: Common phenotypes of spore forming bacteria isolated in Maryland

Phenotype
TL
TLU
TLAE
TLUAE*
Null (0)
Total of all phenotypes**

Bt
176
82
50
23
7
725

% total Bt
24.3
11.3
6.9
3.2
1.0

Other spore formers

111
31
16
7
69
304

% other spore formers

36.0
10.2
5.3
2.3
22.6

*Phenotype of the strain that was applied

** including rare phenotypes not shown

produce urease. These results indicate that an insect,

such as the tobacco hornworm, can be used to select B.
thuringiensis from mixtures of soil bacteria, and recover
applied strains.
Interestingly, a number of toxic B. cereus (non-crystal
forming isolates) were isolated using our technique, and
except for the lack of a parasporal crystal, a number of
these had phenotypic profiles resembling the applied
strain. Some of these toxic non-crystal forming isolates
had 16S rRNA gene sequences identical to the applied
strain and may represent descendants of the applied
strain that have lost the ability to form a crystal. However,
other strains had 16S rRNA gene sequence that differed
slightly from Foray, indicating that they are not derived
from the applied strain. Although these may have
acquired their pathogenicity from the applied strain via
conjugation (Grohmann et al., 2003), it seems more
probable that they, and the other non-crystal forming
isolates represent naturally occurring varieties of
entomopathogenic B. cereus that have previously
escaped detection with isolation methods based on
selection of crystal forming isolates.
We have recently demonstrated that urease production
is an important trait that appears to improve replication of
B. thuringiensis in gypsy moth larvae (Martin et al., 2009).
The relatively high numbers of urease producers among
our isolates (50%), compared with urease producers in
our general collection (ca. 20%) may reflect differences in
pathogenicity based selection used in this study and
selection based on crystal formation. We plan to
investigate these possibilities using multilocus sequence
typing (Priest et al., 2004) to determine the phylogenetic
relationships between these isolates.
ACKNOWLEDGMENTS
Thanks to L. Liska for rearing insects, and A. Mitchell and
R. Farrar for general technical assistance. Mention of a
trademark or proprietary product does not constitute a
guarantee or warranty of the product by the United States

Department of Agriculture and does not imply its approval

to the exclusion of other products that may also be
suitable.

REFERENCES
Akiba Y, Katoh K (1986). Microbial ecology of Bacillus thuringiensis V.
Selective medium for Bacillus thuringiensis vegetative cells. Appl.
Entomol. Zool. 21: 210215.
Ahmed I, Yokota A, Yamazoe A, Fujiwara T (2007). Proposal of
Lysinibacillus boroitolerans gen. nov. sp. nov., and transfer of
Bacillus fusiformis to Lysinibacillus fusiformis comb. nov. and Bacillus
sphaericus to Lysinibacillus sphaericus comb. nov. Intern. J. Syst.
Evol. Microbiol. 57:117-1125.
Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990). Basic
local alignment search tool. J. Mol. Biol. 215: 403-410.
Andrzejczak S, Lonc E (2008). Selective isolation of Bacillus
thuringiensis from soil by the use of L-serine as a minimal media
supplement. Pol. J. Microbiol. 57:333-335.
APHIS factsheet (2003). Gypsy moth: slow the spread.
http://www.docstoc.com/docs/5427202/APHIS-Plant-Protection-andQuarantine-Factsheet-April Gypsy-Moth. Accessed February 4, 2010.
Atlas RM (2004). Handbook of microbiological media. Third edition.
CRC Press, Inc. Boca Raton, FL. Pp 904.
Bell RA, Owens CD, Shapiro M, Tardif JGR (1981). Development of
mass rearing technology, in The gypsy moth: Research toward
integrated pest management (Doane CC, McManus ML, Eds.) U.S.
Department of Agriculture Technical Bulletin 1584, Washington, pp
599-633.
Bizzarri MF, AH Bishop (2007). Recovery of Bacillus thuringiensis in
vegetative form from the phylloplane of clover (Trifolium hybridum)
during a growing season. J. Invertebr. Pathol. 94: 38-47.
Grohmann E, Muth G, Espinosa M (2003). Conjugative plasmid transfer
in Gram-positive bacteria. Microbiol. Mol. Biol. Rev. 67: 277-301.
Juarez-Perez VM, Ferrandis MD, Frutos R (1997). PCR-based
approach for detection of novel Bacillus thuringiensis cry genes. Appl.
Environ. Microbiol. 63:29973002.
Lee IM, Hammond RW, Davis RE, Gundersen DE (1993). Universal
amplification and analysis of pathogen 16S rDNA for classification
and identification of mycoplasma like organisms. Mol. Plant Pathol.
83: 834-842.
Martin PAW, Haransky EB, Travers RS, Reichelderfer CF (1985). Rapid
biochemical testing of large numbers of Bacillus thuringiensis isolates
using agar dots. BioTechniques 3:386-392.
Martin PAW, Travers RS (1989). Worldwide abundance and distribution
of Bacillus thuringiensis isolates. Appl. Environ. Microbiol. 55: 24372442.
Martin PAW (2004). A freeze-dried diet to test pathogens of Colorado

Martin et al. 291

potato beetle. Biol. Cont. 29: 09-14.

Martin PAW, Blackburn MB (2007). Using combinatorics to screen
Bacillus thuringiensis isolates for toxicity against Manduca sexta and
Plutella xylostella. Biol. Cont. 42:226-232.
Martin PAW, Mongeon EA, Gundersen-Rindal DE (2008). Microbial
combinatorics: a simplified approach to isolating insecticidal bacteria.
Biocont. Sci. Technol. 18: 291-305.
Martin PAW, Farrar Jr RR, Blackburn MB (2009). Survival of diverse
Bacillus thuringiensis strains in gypsy moth (Lepidoptera:
Lymantriidae) is correlated with urease production. Biol. Cont.
51:147-151.
Martin PAW, Gundersen-Rindal DE, Blackburn MB (2010). Distribution
of phenotypes among Bacillus thuringiensis strains. Syst. Appl.
Microbiol. 33:204-208.
Ohba M, Aizawa K (1986). Insect toxicity of Bacillus thuringiensis
isolated from soils of Japan. J. Invertebr. Pathol. 47:12-20.

Priest FG, Barker M, Baillie LWJ, Holmes EC, Maiden MCJ (2004).
Population structure and evolution of the Bacillus cereus group J.
Bacteriol. 186: 79597970.
SAS Institute Inc. (2008). SAS OnlineDoc7. Version 9.1. SAS Institute
Inc. Cary, NC.
Sharov AA, Leonard D, Liebhold AM, Clemens NS (2002). Evaluation of
preventive treatments in low-density gypsy moth populations using
pheromone traps. J. Econ. Entomol. 95: 1205-1215.
Silva CP, Waterfield NR, J Daborn P, Dean P, Chilver T, Au CPY,
Sharma S, Potter U, Reynolds SE, ffrench-Constant RH (2002).
Bacterial infection of a model insect: Photorhabdus luminescens and
Manduca sexta. Cell Microbiol. 4: 329-339.
Travers RS, Martin PAW, Reichelderfer CF (1987). Selective process
for efficient isolation of soil Bacillus spp. Appl. Environ. Microbiol.
53:1263-1266