ECOLOGY AND POPULATION BIOLOGY
Seasonal Abundance and Population Structure of Chinavia hilaris and
Nezara viridula (Hemiptera: Pentatomidae) in Georgia Farmscapes
Containing Corn, Cotton, Peanut, and Soybean
JOHN J. HERBERT1,2
AND
MICHAEL D. TOEWS
University of Georgia, Department of Entomology, 2360 Rainwater Road, Tifton, GA 31793-5766
Ann. Entomol. Soc. Am. 105(4): 582Ð591 (2012); DOI: http://dx.doi.org/10.1603/AN12008
ABSTRACT Stink bugs are economically important pests that damage a wide range of crops in the
southeastern United States. Stink bug feeding on developing cotton (Gossypium hirsutum L.) bolls may
result in reduced yield and loss of Þber quality; similarly, feeding on developing soybean [Glycine max
(L.) Merr.] pods can reduce yield and seed quality. During 2009 and 2010, the seasonal abundance
and reproductive biology of Chinavia hilaris (Say) and Nezara viridula (L.) were investigated in
replicated 1.62Ð2.83-ha farmscapes containing equal proportions of corn (Zea mays L.), cotton, peanut
(Arachis hypogaea L.), and soybean. Crops were sampled weekly by using whole plant examinations
in corn and sweep net sampling in cotton, peanut, and soybean. In 2010 only, adults were dissected
to rate their reproductive development, and nymphs were classiÞed to instar. No C. hilaris adults or
nymphs ever were observed in corn; however, nymphs were observed in cotton and soybean during
late September with peak abundance occurring just after the autumnal equinox. The peak of late-instar
nymphs was followed within 2 wk by a peak of nonreproductive adults. More adults were observed
in soybean than cotton. In contrast, N. viridula nymphs and adults were found across all crops and had
multiple generations throughout the growing season. Results from this study indicate that C. hilaris
and N. viridula are different in voltinism, phenology, and use of hosts. These data provide knowledge
of stink bug biology and population ecology at the landscape level and are useful for designing and
implementing stink bug management programs.
KEY WORDS voltinism, population dynamics, host plant, reproductive diapause, integrated pest
management
Phytophagous stink bugs (Hemiptera: Pentatominae)
are economically important insect pests of fruit, vegetable, nut, and Þeld crops. In agricultural production
in the southeastern United States, the term stink bug
generally refers to any member of a species complex
that includes Chinavia hilaris (Say), Euschistus quadrator (Rolston), Euschistus servus (Say), Euschistus
tristigmus (Say), and Nezara viridula (L.). Stink bugs
feed by piercing host fruiting structures with piercing
and sucking mouthparts and then extracting ßuids
(McPherson et al. 1994). Plant injury from feeding can
result from loss of turgor pressure, injection of destructive digestive enzymes, injection of plant pathogenic organisms, delayed crop maturation, or subsequent deformation or abortion of fruiting bodies
(Panizzi and Slansky 1985, Barbour et al. 1990, Medrano et al. 2009). In addition, edible plant parts can
take on a bitter taste or pithy texture as a result of stink
Mention of trade names or commercial products in this publication
is solely for the purpose of providing speciÞc information and does not
imply recommendation or endorsement by the University of Georgia.
1 Current Address: Department of Biology, Georgia Southern University, P.O. Box 8042, Statesboro, GA 30460-8042.
2 Corresponding author, e-mail: [email protected].
bug feeding (Callahan et al. 1960). Nationwide economic losses to stink bugs are difÞcult to quantify, but
studies show that losses in cotton production can exceed 222,000 bales annually (Williams 2006); with lint
at 90 cents per pound that would extrapolate to more
than $95 million. Nault and Speese (2002) estimated
that stink bugs injured 26 Ð39% of spring tomatoes in
Virginia. Finally, estimates of stink bug control costs
plus crop damage were estimated at $28.29/ha in Mississippi (Musser and Catchot 2008).
C. hilaris is a highly polyphagous stink bug species
that is a North American native with a range in the
United States from New England west to the PaciÞc
Coast, down to the southwest and across the southeastern states (McPherson 1982). Several authors observed that this species favors woody shrubs and trees
such as elderberry and black cherry before entering
and reproducing in cultivated soybean later in the year
(Jones and Sullivan 1982, McPherson 1982). Morrill
(1910) reported that C. hilaris was a pest of cotton in
the mid-1800s. The species is reported to be univoltine
north of 45⬚ latitude, but bivoltine in Illinois, Kansas,
and South Carolina (Wilde 1969, Jones and Sullivan
1982, Javahery 1990, McPherson and Tecic 1997). At
0013-8746/12/0582Ð0591$04.00/0 䉷 2012 Entomological Society of America
July 2012
HERBERT AND TOEWS: REPRODUCTIVE BIOLOGY OF STINK BUGS
least one report states that adults overwinter and the
species has an obligatory diapause period of 2Ð3 mo
(Javahery 1990). C. hilaris also is known as Acrosternum hilare (Say), but has been formally resolved to C.
hilaris in Brazil (Schwertner and Grazia 2007) and the
same taxonomic revision is supported by a North
American Pentatomid taxonomist (David Rider, personal communication).
N. viridula is believed to be of Ethiopian origin, but
is now cosmopolitan throughout tropical and subtropical regions of the world. In the United States, it ranges
from Florida and Georgia west through Texas and
Oklahoma, with an additional population in California
(Hoffman et al. 1987, McPherson et al. 1994). This
species is highly polyphagous and has three to Þve
generations per year, depending on latitude (Todd
1989). Kiritani (1963) described changes in the female
reproductive system of N. viridula and how they predicted life history in Japan. As a native to the tropics,
cold temperature during the winter appears to significantly decrease overwintering populations (Jones
and Sullivan 1981, Todd 1989). Movement and dispersal of this species in southeastern farmscapes recently
have been characterized (Tillman et al. 2009, Toews
and Shurley 2009, Olson et al. 2011).
The authors previously published a report on the
seasonal abundance and population structure of E.
servus in southeastern U.S. farmscapes. This report
addresses Þndings with a focus on C. hilaris and N.
viridula. The objectives of this project were to 1) more
fully understand the sequence of host utilization by
adult and immature stink bugs in southeastern farmscapes, and 2) examine physiological changes in adults
as a function of host and time of year.
Materials and Methods
Plot Layout. The study was conducted in 2009 and
2010 on three University of Georgia experiment stations. Plots were established at Midville (32⬚ 52⬘34.63⬙
N, 82⬚ 13⬘23.46⬙ W); Plains (32⬚ 02⬘23.53⬙ N, 84⬚
22⬘11.96⬙ W); and Tifton (31⬚ 31⬘25.34⬙ N, 83⬚ 32⬘52.96⬙
W), GA. Plot size varied by location between 1.62 and
2.83 ha in total area, based on available land. Subplot
layout and sampling protocols followed the procedures outlined in Herbert and Toews (2011). Brießy,
each main plot was subdivided into four subplots that
were equal in area. Two subplots were rectangular in
shape and located on opposite sides of the plot,
whereas the remaining two subplots were square in
shape and located between the outside subplots. The
approximate plot layout can be visualized by looking
at the letter “H”, where the two vertical lines represent
the rectangular plots and the horizontal line represents the midline that separates the middle two subplots. The width of the rectangular subplots was approximately one-half the widths of the center
subplots, thereby giving each subplot equal area. Each
subplot was planted with corn, cotton, peanut, or soybean such that cotton was always planted in one of the
center plots and corn was always planted in one of the
outside rectangular plots. Crops were rotated the next
583
year and cotton was grown in the opposite center plot
and corn was grown on the opposite rectangular plot.
Soybean and peanut were randomly assigned to the
two remaining subplots. Sampling points were arranged on a 14.2-m grid that was rotated 45⬚ such that
the diagonals of the grid were parallel with the orientation of the crop rows (see Herbert and Toews
2011). Each sampling point was marked in the Þeld
using a 2.4-m Þberglass pole with a plastic ßag (Agri
Drain Corp., Adair, IA). The total number of sampling
points varied with increasing plot size within a range
of 10 Ð12 samples per subplot. To prevent oversampling individual plants, sample rows were alternated
by week so that individual plants were only sampled
once every 3 wk. Brießy, the sample ßag marked the
general area where samples were obtained, but actual
samples were obtained from two rows located on opposite sides of the ßag. Those rows, termed sampled
rows, were rotated among three adjacent rows on
either side of the ßag such that the sample rows were
always four rows apart (i.e., row 3 east of the ßag and
row 1 west of the ßag during week 1, followed by row
2 east of the ßag and row 2 west of the ßag during
week 2).
Crop Management. All crops were grown using University of Georgia Cooperative Extension recommended practices, except that no insecticides were
applied after planting. Crops were grown using conventional tillage on a 0.9-m row spacing under irrigation. Planting dates, timing of crop management, and
harvest dates varied slightly with local weather conditions. Although the same cultivars were planted at
each site within year, commercial availability forced
using different cultivars between years. These cultivars were selected based on Extension recommended
practices and included RR and Bt corn hybrids (120 d
maturity), B2RF cotton germplasm (long season maturity), and modern peanut and soybean (maturity
group VII) cultivars. A comprehensive list of cultivars,
planting dates, and harvest dates is provided in Herbert and Toews (2011).
Crop maturity matched expected benchmarks in
development for southern Georgia. Corn began tasseling during the Þrst week in June and then reached
black layer in the last week of July. Cotton began
producing squares in July and then produced bolls
starting in the third week of July before bolls began
opening during the last week of August. Peanut began
ßowering in mid-June and then started pegging during
the Þrst week of July. Soybean began ßowering during
the last week of June and then began setting pods in
mid-July. Mature pods (stage R-7) were observed
starting in October.
Data Collection. Stink bugs were sampled weekly,
starting in mid-June. Starting just before tassel and
ending when the corn reached black layer, corn was
sampled using whole plant visual searches (10 plants
per sample row for 20 plants in total per sample). On
the Þrst sample date cotton was producing squares,
peanut was in early ßowering stages, and soybean was
still in vegetative development. These crops were sampled using 20 sweeps (38.1-cm sweep net) per row for
584
ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA
Vol. 105, no. 4
2
Mean adult C. hilaris per sample
A
Spring
Autumn
Summer
1
0
B
Cotton
Soybean
4
3
2
1
0
May
Jun
Jul
Aug
Sep
Oct
Nov
Fig. 1. Mean abundance of C. hilaris adults by crop and sampling date for 2009 (A) and 2010 (B). Dashed vertical lines
indicate the summer solstice and autumnal equinox.
40 sweeps in total per sample. Sampling in cotton and
peanut continued until the cotton bolls at the bottom
of the plant were open (early September); however,
sampling in soybean continued for several more weeks
until the leaves turned color (mid-October) . After
collection, stink bugs immediately were transferred to
labeled 3.78-liter plastic bags or 15-ml vials and transported to the laboratory for further processing.
Upon arrival to the laboratory, samples were processed and dissected using the procedures outlined in
Herbert and Toews (2011). Brießy, samples collected
in 2009 were held at ⫺20⬚C, whereas samples from
2010 were held at 4.4⬚C for up to 48 h before processing. The species determination, life stage (nymph or
adult), and total number of stink bugs were recorded
by sample site. In 2010, nymphs were additionally
classiÞed by individual instar and adult stink bugs were
dissected to assess their gonadal development as nonreproductive, intermediate, or reproductive (Toscano
and Stern 1980, Esquivel 2009, Herbert and Toews
2011).
Presentation and Statistics. Figures representing
the seasonal abundance and phenology of N. viridula
and C. hilaris were constructed by calculating the
overall mean generalized across locations and sampling points within individual crops. Standard errors
for the means were proportional to the mean and are
not included in the Þgures to allow better visualization
of trends. Seasonal abundance and phenology were
subdivided into time of year by vertical dashed lines
that represent the summer solstice (21 June) and the
autumnal equinox (22 September) (U.S. Naval Observatory, Washington, DC).
Stink bug counts were analyzed with a Poisson Regression model by using PROC GLIMMIX in SAS 9.2
(SAS Institute 2008). The statistical model was constructed using the principles of Stokes et al. (2000). A
repeated measures analysis was used to compare the
abundance of adult stink bugs in cotton, peanut, and
soybean for all sampling dates that overlapped. Samples collected within a given crop were considered
subsamples, location was a random effect, and both
year and crop were Þxed effects. The model included
the interaction between year and crop. Corn intentionally was analyzed separately from cotton, peanut,
and soybean because sampling techniques differed
(i.e., whole plant sampling versus sweep net sampling). Means were back transformed using the ILINK
function in the LSMEANS statement and compared
using TukeyÕs honestly signiÞcant difference (HSD)
(SAS Institute 2008).
Data collected in 2010 contained information about
sex and gonadal development of each adult stink bug.
Analyses on reproductive development by species
were conducted by season (spring, summer, or fall).
Each analysis was a repeated measures Poisson regression that used counts as the response variable. The
statistical model included Þxed effects of reproductive
stage, sex, and crop. Location was a random effect and
samples within a crop at one location were treated as
subsamples; treatment means were separated using
TukeyÕs HSD (SAS Institute 2008).
Results
C. hilaris. Regardless of year, C. hilaris populations
were very low in corn and peanut. Neither adults nor
nymphs were encountered until after the summer
solstice on 21 June, thus none were observed in corn.
C. hilaris observations from peanut were not graphed
or analyzed because the abundance was extremely
low; samples in peanut from 2009 contained three
adults and two nymphs in total for the entire sampling
period, whereas samples from 2010 contained six
adults and six nymphs in total. In soybean, C. hilaris
adults exhibited a small peak in abundance during
mid-July for both 2009 and 2010. In contrast, C. hilaris
adults had a much larger peak in soybean during late
September and early October, which coincided with
the autumnal equinox (Fig. 1). In cotton, C. hilaris
adults remained at low levels from late July through
mid-September and there were no sharp increases in
July 2012
HERBERT AND TOEWS: REPRODUCTIVE BIOLOGY OF STINK BUGS
Mean adult C. hilarisI per sample
0.5
A
Spring
585
Autumn
Summer
0.0
B
Non-reproductive
Intermediate
Reproductive
2.0
1.5
1.0
0.5
0.0
May
Jun
Jul
Aug
Sep
Oct
Nov
Fig. 2. Mean abundance of C. hilaris adults by crop and reproductive stage for cotton (A) and soybean (B). Dashed
vertical lines indicate the summer solstice and autumnal equinox.
abundance during those dates. In 2009 only, the abundance of C. hilaris increased slightly in preharvest
October cotton (Fig. 1). Data on the phenology of C.
hilaris adults broken down by sex and reproductive
stage did not show any differences between males and
females (P ⬎ 0.05). The summer population in cotton
and soybean did not contain any nonreproductive
adults. However, the population radically changed to
exclusively nonreproductive individuals starting 2 wk
before the autumnal equinox (Fig. 2). There were no
reproductive adults captured after the third week in
September.
Repeated measures analyses comparing the number
of adult C. hilaris in cotton and soybean showed differences for the main effects of year (F ⫽ 8.25; df ⫽
1, 8; P ⫽ 0.02) and crop (F ⫽ 60.96; df ⫽ 1, 8; P ⬍ 0.01)
and there was no interaction between crop and year
(P ⬎ 0.05). There were greater numbers of adult C.
hilaris per sample in 2010 (0.229 ⫾ 0.021; mean ⫾
SEM) compared with 2009 (0.146 ⫾ 0.018; mean ⫾
SEM). There were also more adults detected in soybean (0.335 ⫾ 0.027; mean ⫾ SEM) compared with
cotton (0.100 ⫾ 0.013; mean ⫾ SEM).
Reproductive data classiÞed by sex and reproductive condition showed differences in the main effects
by crop (F ⫽ 9.74; df ⫽ 1, 4; P ⫽ 0.04) and reproductive
stage (F ⫽ 31.91; df ⫽ 2, 8; P ⬍ 0.01). There was a
greater number of C. hilaris in soybean than cotton
and a greater number of reproductive adults than
intermediate or nonreproductive (Table 1; Fig. 2).
Table 1. Comparisons for differences in the abundance of C.
hilaris adults by crop and reproductive stage
Effect
Variable
Mean ⫾ SEMa
Crop
Cotton
Soybean
Nonreproductive
Intermediate
Reproductive
0.015 ⫾ 0.005b
0.034 ⫾ 0.003a
0.014 ⫾ 0.004b
0.012 ⫾ 0.003b
0.071 ⫾ 0.008a
Stage
a
Means among crop or stage followed by different letters are
different at P ⬍ 0.05 TukeyÕs HSD.
The analysis on soybean after the autumnal equinox
showed differences among reproductive stages (F ⫽
58.48; df ⫽ 2, 4; P ⬍ 0.01). There were signiÞcantly
more nonreproductive (1.054 ⫾ 0.062; mean ⫾ SEM)
than intermediate (0.013 ⫾ 0.007; mean ⫾ SEM) or
reproductive adults (0.038 ⫾ 0.012; mean ⫾ SEM), but
there were no differences between the mean number
of reproductive and intermediate adults (Fig. 2B).
Phenology of C. hilaris nymphs was consistent between 2009 and 2010. Across years, there was a moderate increase in the number of nymphs in cotton
followed by a sharp increase in the number of
nymphs in soybean during early September (Fig. 3).
Interestingly, no late-stage nymphs (fourth or Þfth
instars) were detected in peanut. The abundance of
nymphs declined in cotton after its initial peak in
August, but continued to increase sharply in soybean until the end of September, and the population
of C. hilaris nymphs attained their greatest numbers
just before the autumnal equinox (Fig. 3C).
Nymphal populations peaked 2 wk before the peak
of nonreproductive adults (Figs. 2, 3). Data from
2010 showed that there were approximately equal
numbers of third, fourth, and Þfth instars in cotton
(Fig. 3B). Soybean data during the peak population
of nymphs from 2010 showed that Þfth instars were
greater in abundance than fourth instars by a factor
of 1.5, and fourth instars were greater in abundance
than third instars by a factor of two (Fig. 3C).
N. viridula. Phenology data for N. viridula adults
showed similarities between 2009 and 2010. The abundance of N. viridula adults in corn peaked within 1 d
of the summer solstice each year (Fig. 4). Smaller
population peaks were recorded in cotton and soybean during the summer months, followed by a sharp
increase in abundance of N. viridula adults in soybean
immediately after the autumnal equinox. Abundance
in peanut was low for the entire experiment for both
years (Fig. 4). Reproductive status data from 2010
showed that males and females from each reproductive stage were present in each crop during the experiment (Fig. 5). Males and females did not differ in
586
ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA
Vol. 105, no. 4
2.0
A
Spring
Autumn
Summer
1.5
Cotton
Soybean
1.0
Mean C. hilaris nymphs per sample
0.5
0.0
B
0.25
0.00
C
First
Second
Third
Fourth
Fifth
3.0
2.5
2.0
1.5
1.0
0.5
0.0
May
Jun
Jul
Aug
Sep
Oct
Nov
Fig. 3. Mean abundance of C. hilaris nymphs by crop for 2009 (A) and instar for 2010 cotton (B) and soybean (C). Dashed
vertical lines indicate the summer solstice an autumnal equinox.
Mean adult N. viridula per sample
their abundance or phenology and are not individually
represented in phenology Þgures.
Statistical analyses of N. viridula adults in cotton,
peanut, and soybean across dates that overlap during
2009 and 2010 showed signiÞcant effects for year (F ⫽
43.85; df ⫽ 1, 12; P ⬍ 0.01); crop (F ⫽ 18.00; df ⫽ 2, 12;
P ⬍ 0.01); and an interaction between year and crop
(F ⫽ 6.95; df ⫽ 2, 12; P ⬍ 0.01); thus separate analyses
were performed for 2009 and 2010. Analysis of 2009
data for all dates where cotton, peanut, and soybean
were sampled on the same day showed differences
among crops (F ⫽ 9.56; df ⫽ 2, 6; P ⫽ 0.01) and
differences among all pairwise comparisons between cotton, peanut, and soybean (Table 2). Results from the 2010 data where cotton, peanut, and
soybean all were sampled on the same day showed
a difference in the effect of crop (F ⫽ 13.17; df ⫽ 2,
6; P ⬍ 0.01), and cotton and soybean had a greater
mean number of adult N. viridula per sample than
peanut (Table 2).
Analyses comparing the reproductive stages and sex
of N. viridula adults in corn showed no signiÞcant
effects or interactions. Results from the analysis for N.
viridula adults in cotton, peanut, and soybean during
8
A
Spring
Autumn
Summer
6
Corn
Cotton
Peanut
Soybean
4
2
0
1
B
0
May
Jun
Jul
Aug
Sep
Oct
Nov
Fig. 4. Mean abundance of N. viridula adults by crop and sampling date for 2009 (A) and 2010 (B). Dashed vertical lines
indicate the summer solstice and autumnal equinox.
July 2012
HERBERT AND TOEWS: REPRODUCTIVE BIOLOGY OF STINK BUGS
587
0.30
A
Spring
Autumn
Summer
0.15
Mean number of adult N. viridula per sample
0.00
B
0.6
0.3
2D Graph 5
0.0
0.15
C
0.00
D
Non-reproductive
Intermediate
Reproductive
0.6
0.3
0.0
May
Jun
Jul
Aug
Sep
Oct
Nov
Fig. 5. Mean abundance of N. viridula adults by crop and reproductive stage for corn (A), cotton (B), peanut (C), and
soybean (D). Dashed vertical lines indicate the summer solstice and autumnal equinox.
fore the peak of nonreproductive adults (Figs. 5Ð7).
Data on individual instar indicated that all instars were
present in all crops at sometime during the year (Fig.
7). In soybean, the magnitude of nymphs recovered
was greatest for Þfth instars followed by fourth instars
and then third instars (Fig. 7D). However, capturing
a greater number of Þfth instars in comparison to
fourth instars and third instars was not observed in
corn, cotton, or peanut (Fig. 7AÐC).
the summer months showed signiÞcant effects of crop
(F ⫽ 17.59; df ⫽ 2, 6; P ⬍ 0.01) and reproductive stage
(F ⫽ 12.83; df ⫽ 2, 12; P ⬍ 0.01). There were no
interactions among Þxed effects. More N. viridula
adults were captured in cotton and soybean compared
with peanut, and there were a greater number of
reproductive adults than intermediate and nonreproductive adults (Table 3). In soybean after the autumnal equinox only, there were obvious differences
among nonreproductive (0.232 ⫾ 0.028; mean ⫾
SEM), intermediate 0.047 ⫾ 0.012; mean ⫾ SEM), and
reproductive (0.023 ⫾ 0.008; mean ⫾ SEM) stages
(F ⫽ 27.77; df ⫽ 2, 4; P ⬍ 0.01) (Fig. 5). During autumn,
the mean number of nonreproductive adults was
greater than the number of reproductive and intermediate adults, but there were no differences in abundance of intermediate and reproductive adults.
The phenology of N. viridula nymphs exhibited similar trends in 2009 and 2010, but the overall abundance
in 2009 was more than fourfold greater compared with
2010 (Figs. 6, 7). Nymphs reached their peak abundance in corn during mid-July, followed by progressive captures in soybean, peanut, and cotton (Fig. 7).
Nymphs had their peak population immediately be-
Protocols described here were successful for tracking the life histories of C. hilaris and N. viridula, and
this study is the Þrst to couple progressive nymphal
development by species with the emergence of adults.
Previous studies have documented the seasonal abundance of stink bugs in agronomic and nonagronomic
hosts, but none followed nymphal instar or reproductive maturity of adults (Bundy et al. 1999, Bundy and
McPherson 2000, Tillman et al. 2009). Here, peak
abundance of Þfth instars for C. hilaris and N.
viridula typically was observed the week before a
large portion of nonreproductive adults, indicating
Table 2. Comparisons for differences in the abundance of
C. hilaris adults by year
Table 3. Comparisons for differences in the abundance of
N. viridula adults by crop and reproductive stage
Discussion
Year
Crop
Mean ⫾ SEMa
Effect
Variable
Mean ⫾ SEMa
2009
Cotton
Peanut
Soybean
Cotton
Peanut
Soybean
0.928 ⫾ 0.115b
0.130 ⫾ 0.046c
2.016 ⫾ 0.170a
0.283 ⫾ 0.057a
0.039 ⫾ 0.021b
0.157 ⫾ 0.047a
Crop
Cotton
Peanut
Soybean
Nonreproductive
Intermediate
Reproductive
0.037 ⫾ 0.004a
0.004 ⫾ 0.001b
0.028 ⫾ 0.004a
0.011 ⫾ 0.003b
0.011 ⫾ 0.003b
0.033 ⫾ 0.004a
2010
a
Means among crop within year followed by different letters are
different at P ⬍ 0.05 TukeyÕs HSD.
Stage
a
Means among crop or stage followed by different letters are
different at P ⬍ 0.05 TukeyÕs HSD.
588
ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA
Vol. 105, no. 4
Mean number of N. viridula nymphs per sample
14
Spring
Autumn
Summer
12
Corn
Cotton
Peanut
Soybean
10
8
6
4
2
0
May
Jun
Jul
Aug
Sep
Oct
Nov
Fig. 6. Mean abundance of N. viridula nymphs by crop in 2009. Dashed vertical lines indicate the summer solstice and
autumnal equinox.
that the population of nonreproductive adults was
the result of Þfth instars molting into adults. Thus,
the sampling techniques and use of dissections were
helpful to clearly delineate each generation in each
crop for each species.
C. hilaris was not present in samples from corn for
either year of this study, and previous studies also
indicate that populations of C. hilaris generally are
extremely low in corn when their presence is detected
(Tillman 2010). Data from this study and previous
studies indicated that C. hilaris were not present in
South Carolina soybean or agronomic environments
in southeast Georgia until July, and its absence in corn
0.30
0.15
A
Spring
may be because of the timing of its life cycle (Jones
and Sullivan 1982). C. hilaris has not been captured in
spring emergence traps in southern Georgia, and information regarding the overwintering habits of C.
hilaris in Georgia is lacking (M.D.T., unpublished
data). It is unknown if C. hilaris successfully overwinters in southeastern Georgia or if the late arrival of C.
hilaris in agronomic habitats is because of a generation
that occurs on a nonagronomic host or perhaps immigration from another geographic region. Further
investigations regarding the overwintering of C. hilaris
in Georgia may beneÞt from targeting the known
overwintering habitat of deciduous forest leaf litter,
Autumn
Summer
Mean number of N. viridula nymphs per sample
0.00
0.15
B
0.00
0.15
C
0.00
D
First
Second
Third
Fourth
Fifth
1.5
1.0
0.5
0.0
May
Jun
Jul
Aug
Sep
Oct
Nov
Fig. 7. Mean abundance of N. viridula nymphs by crop and instar in 2010 for corn (A), cotton (B), peanut (C), and
soybean (D). Dashed vertical lines indicate the summer solstice and autumnal equinox.
July 2012
HERBERT AND TOEWS: REPRODUCTIVE BIOLOGY OF STINK BUGS
and nonagronomic hosts such as black cherry, Elderberry, or winged, or dwarf sumac, which are known to
support the Þrst generation of C. hilaris (Jones and
Sullivan 1981, 1982).
Adult and nymph C. hilaris were observed from July
through November in peanut, cotton, and soybean.
The abundance of C. hilaris in peanut was extremely
low for both years, which is consistent with Þndings
from other studies on stink bugs in peanut (Tillman
2008, Tillman et al. 2009). Late-instar nymphs were not
encountered in peanut, and all C. hilaris nymphs sampled from peanut were Þrst, second, or third instar.
Low abundance of adults, coupled with the absence of
late instars, indicates that peanut is not a suitable host
for reproduction or the nymphal development C. hilaris. Adult C. hilaris appeared in cotton and soybean
during July, and the majority of the population was
sexually mature and capable of reproduction, but
nymphs were not detected in either soybean or cotton
until August (Figs. 2, 3). Fifth instars appeared in
cotton and soybean at the same time, and the peak
abundance of Þfth instars in soybean was more than
eight times greater than cotton. These data suggest
that cotton is a suitable host, but cotton is likely a poor
reproductive host and not an optimal host for nymphal
development (Fig. 3). Soybean was clearly the most
suitable host in this study, and the sharp increase in
Þfth instars followed by the appearance of nonreproductive or newly emerged adults shows a distinct
generation (the overwintering generation) emerging from soybean immediately after the autumnal
equinox.
Data from this study showed a single generation of
C. hilaris, and this generation represents the overwintering population. However, previous research on the
emergence of C. hilaris in the southeastern United
States indicated that C. hilaris was bivoltine and has at
least one generation on nonagronomic hosts (Jones
and Sullivan 1981, 1982). Thus, the origin of the population in southeast Georgia that invades, oviposits,
and reproduces in soybean is unknown. Previous studies on the population dynamics of stink bugs that
included C. hilaris also indicate that the peak abundance of adult stink bugs occurred near the autumnal
equinox (Bundy and McPherson 2000; Smith et al.
2008, 2009). The generation that emerges just after the
autumnal equinox is the overwintering population,
and individuals sampled after this emergence show no
signs of reproductive development (Fig. 2). Emergence of an overwintering population near the autumnal equinox also is seen in Euschistus servus, which
is also native to North America (Herbert and Toews
2011). Data from the previous study on E. servus and
this study strongly suggest that E. servus and C. hilaris
are using the same photoperiod cues to initiate overwintering and show the same trend in population
dynamics.
N. viridula was present in corn, cotton, peanut, and
soybean for some duration of the experiment. These
data are consistent with other studies showing the
seasonal abundance of N. viridula in Georgia cotton
farmscapes (Tillman 2008, 2010; Tillman et al. 2009).
589
The presence of Þfth instars immediately followed by
newly emerged nonreproductive adults was observed
in all four crops; these data suggest that all crops were
suitable reproductive hosts for this species, but the
crops do not appear to be equal in their suitability.
Because the samples were obtained using different
methods, population estimates from corn cannot be
directly compared with peanut, soybean, or cotton
using these data, but the relative trends among mean
numbers of reproductive adults, Þfth instars, and
newly emerged adults can be used to infer the relative
suitability of corn. The number of Þfth instars present
in corn was less than the number of reproductive
adults, and the number of newly emerged adults was
much less than either Þfth instars or reproductive
adults. These observations suggest that the population
of N. viridula in corn merely was maintained as opposed to increased.
The population dynamics of N. viridula in peanut
were similar to the dynamics observed in corn. Although corn and peanut may be suitable for nymphal
development, they are relatively poor hosts that do not
signiÞcantly contribute to increases in the overall population of N. viridula in the landscape (Figs. 5A, C; 7A,
C). In contrast, a greater number of Þfth instars compared with reproductive adults and a greater number
of nonreproductive adults compared with the reproductive or ovipositing population were observed in
cotton and soybean (Figs. 5B, D; 7B, D). Observed
dynamics in cotton and soybean suggest that the population of N. viridula increased on these hosts.
The population dynamics of C. hilaris and N.
viridula are similar in that both species show a sharp
increase in newly emerged nonreproductive adults
near the autumnal equinox, but these species also have
some important differences. Data from this experiment and previous studies indicate that C. hilaris is
bivoltine, whereas N.viridula is multivoltine (Alli et al.
1979; Jones and Sullivan 1981, 1982; Negron and Riley
1987; Todd 1989; Musolin and Numata 2003). N.
viridula has progressive generations that show some
degree of successful reproduction in all crops, whereas
C. hilaris showed only one generation that was simultaneous in cotton and soybean. Differences in the
voltinisim of these two species could be because of the
areas in which these species evolved. C. hilaris is
native to North America, whereas N. viridula is native
to the tropics. Observed differences among the life
history characteristics of different stink bug species
highlight the critical importance of correctly identifying specimens and making predictions about future
damage potential.
Ecologically based strategies for managing stink
bug populations beneÞt from understanding when a
particular pest species is likely to be present within
a farmscape. For example, it makes sense to adjust
planting dates or maturity groups to escape damage
from a bivoltine species (e.g., C. hilaris). Conversely, this strategy would not work for a polyvoltine species like N. viridula that has progressive
overlapping generations throughout the growing
season. Similarly, approaches for reducing overall
590
ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA
populations will maximize efÞciency by temporally
targeting relatively nonmobile nymphs at their peak
population density, such as was observed in late fall
in soybean.
Acknowledgments
We gratefully acknowledge Jessica Corbett, Blake
Crabtree, Kevin Frizzel, and Anne Horak for their assistance
sampling stink bugs in the Þeld. We also thank Anthony
Black, David GrifÞn, Stan Jones, and Wesley Stephens for
managing the plots. BASF, Bayer CropScience, Dow AgroSciences, DuPont, Monsanto, and Syngenta AG donated seed
and agricultural chemicals used in the trials. This research
was funded by the USDAÐNIFAÐSpecial Research Grants
Program under awards 2008-34566-18998 and 2009-3456620100.
References Cited
Alli, M. A., A. M. Awadallah, and A. A. El-Rahman. 1979. A
study on the phenology and ecology of the green stink
bug Nezara viridula L. (Heteroptera: Pentatomidae).
J. Appl. Entomol. 88: 476 Ð 483.
Barbour, K. S., J. R. Bradley, Jr., and J. S. Bacheler. 1990.
Reduction in yield and quality of cotton damaged by
green stink bug (Hemiptera: Pentatomidae). J. Econ.
Entomol. 83: 842Ð 845.
Bundy, C. S., and R. McPherson. 2000. Dynamics and seasonal abundance of stink bugs (Heteroptera: Pentatomidae) in a cotton soybean ecosystem. J. Econ. Entomol. 93:
697Ð706.
Bundy, C. S., R. M. McPherson, and G. A. Herzog. 1999.
Stink bugs in cotton: a temporal occurrence, pp. 1038 Ð
1040. In Proceedings, Beltwide Cotton Conferences, 3Ð7
January 1999, Orlando, FL. National Cotton Council of
America, Memphis, TN.
Callahan, P. S., R. Brown, and A. Dearman. 1960. Control of
tomato insect pests in Louisiana. La. Agric. 3: 1Ð3.
Esquivel, J. F. 2009. Stages of gonadal development of the
southern green stink bug (Hemiptera: Pentatomidae):
improved visualization. Ann. Entomol. Soc. Am. 102: 303Ð
309.
Herbert, J. J., and M. D. Toews. 2011. Seasonal abundance
and population structure of brown stink bug (Hemiptera:
Pentatomidae in farmscapes containing corn, cotton, peanut, and soybean. Ann. Entomol. Soc. Am. 104: 909 Ð918.
Hoffman, M. P., L. T. Wilson, and F. G. Zalom. 1987. The
southern green stink bug, Nezara viridula Linnaeus (Heteromptera: Pentatomidae); new location. Pan-Pac. Entomol. 63: 333.
Javahery, M. 1990. Biology and ecological adaption of the
green stink bug (Hemiptera: Pentatomidae) in Quebec
and Ontario. Ann. Entomol. Soc. Am. 83: 201Ð206.
Jones, W. A., and M. J. Sullivan. 1981. Overwintering habitats, spring emergence patterns, and winter mortality of
some South Carolina Hemiptera. Environ. Entomol. 10:
409 Ð 414.
Jones, W. A., and M. J. Sullivan. 1982. Role of host plants in
population dynamics of stink bug pests of soybean in
South Carolina. Environ. Entomol. 11: 867Ð 875.
Kiritani, K. 1963. The change in reproductive system of the
southern green stink bug, Nezara viridula, and its application to forecasting of the seasonal history. Jpn. J. Appl.
Entomol. Zool. 7: 327Ð337.
McPherson, J. E. 1982. The Pentatomoidea (Hemiptera) of
northeastern North America with emphasis on the fauna
Vol. 105, no. 4
of Illinois. Southern Illinois University Press, Carbondale
and Edwardsville, IL.
McPherson, J. E., and D. L. Tecic. 1997. Notes on the life
histories of Acrosternum hilare and Cosmopepla bimaculata (Heteroptera: Pentatomidae) in southern Illinois. Gt.
Lakes Entomol. 30: 79 Ð 84.
McPherson, R. M., J. W. Todd, and K. V. Yeargan. 1994.
Stink bugs, pp. 87Ð90. In L. G. Higley and D. J. Boethel
(eds.), Handbook of soybean insect pests. Entomological
Society of America, Lanham, MD.
Medrano, E. G., J. F. Esquivel, R. L. Nichols, and A. A. Bell.
2009. Temporal analysis of cotton boll symptoms resulting from southern green stink bug feeding and transmission of a bacterial pathogen. J. Econ. Entomol. 102: 36 Ð 42.
Morrill, A. W. 1910. Plant-bugs injurious to cotton bolls.
U.S. Dep. Agric. Bur. Entomol. Bull. 86: 1Ð110.
Musolin, D. L., and H. Numata. 2003. Timing of diapause
induction and its life- history consequences in Nezara
viridula: is it costly to expand the distribution range?
Ecol. Entomol. 28: 694 Ð703.
Musser, F. R., and A. Catchot. 2008. Mississippi soybean
insect losses. Midsouth Entomol. 1: 29 Ð36.
Nault, B. A., and J. Speese III. 2002. Major insect pests and
economics of fresh-market tomato in eastern Virginia.
Crop Prot. 21: 359 Ð366.
Negron, J. F., and T. J. Riley. 1987. Southern green stink bug,
Nezara viridula. J. Econ. Entomol. 80: 666 Ð 669.
Olson, D. M., J. R. Ruberson, A. R. Zeilinger, and D. A.
Andow. 2011. Colonization preference of Euschistus servus and Nezara viridula in transgenic cotton varieties,
peanut, and soybean. Entomol. Exp. Appl. 139: 161Ð169.
Panizzi, A. R., and F. Slansky Jr. 1985. Review of phytophagous pentatomids associated with soybean in the Americas. Fla. Entomol. 68: 184 Ð214.
SAS Institute. 2008. SAS/STAT 9.2 userÕs guide. SAS Institute, Cary, NC.
Schwertner, C. F., and J. Grazia. 2007. O gênero Chinavia
Orian (Hemiptera, Pentatomidae, Pentatominae) no Brasil, com chave pictórica para os adultos. Rev. Bras. Entomol. 51: 416 Ð 435.
Smith, J. F., R. G. Luttrell, and J. K. Greene. 2008. Seasonal
abundance of stink bugs (Heteroptera : Pentatomidae)
and other polyphagous species in a multi-crop environment in south Arkansas. J. Entomol. Sci. 43: 1Ð12.
Smith, J. F., R. G. Luttrell, and J. K. Greene. 2009. Seasonal
abundance, species composition, and population dynamics of stink bugs in production Þelds of early and late
soybean in South Arkansas. J. Econ. Entomol. 102: 229 Ð
236.
Stokes, M. E., C. S. Davis, and G. G. Koch. 2000. Categorical
data analysis using the SAS system, 2nd ed. SAS Institute,
Cary, NC.
Tillman, G. 2008. Observations of stink bugs (Heteroptera:
Pentatomidae) ovipositing and feeding on peanuts. J.
Entomol. Sci. 43: 447Ð 452.
Tillman, G. 2010. Composition and abundance of stink bugs
(Heteroptera: Pentatomidae) in corn. Environ. Entomol.
39: 1765Ð1774.
Tillman, G., T. D. Northfield, R. F. Mizell, and T. C. Riddle.
2009. Spatiotemporal patterns and dispersal of stink bugs
(Heteroptera: Pentatomidae) in peanut-cotton farmscapes. Environ. Entomol. 38: 1038 Ð1052.
Todd, J. W. 1989. Ecology and behavior of Nezara viridula.
Annu. Rev. Entomol. 34: 273Ð292.
Toews, M. D., and D. W. Shurley. 2009. Crop juxtaposition
affects cotton Þber quality in Georgia farmscapes. J. Econ.
Entomol. 4: 1515Ð1522.
July 2012
HERBERT AND TOEWS: REPRODUCTIVE BIOLOGY OF STINK BUGS
Toscano, N. C., and V. M. Stern. 1980. Seasonal reproductive
condition of Euschistus conspersus. Ann. Entomol. Soc.
Am. 73: 85Ð 88.
Wilde, G. E. 1969. Photoperiodism in relation to development and reproduction in the green stink bug. J. Econ.
Entomol. 62: 629 Ð 630.
591
Williams, M. R. 2006. Cotton insect losses-2005, pp. 1151Ð
1204. In Proceedings of the Beltwide Cotton Conferences, 3Ð 6 January 2005, San Antonio, TX. National Cotton Council of America, Memphis, TN.
Received 16 January 2012; accepted 11 May 2012.