Monitoring of brown stink bug (Hemiptera: Pentatomidae ...download.xuebalib.com/8ey6hYR3dDug.pdf · Arizona, Tucson, AZ 85138, USA; 5College of Engineering, China Agricultural University,

This is an Accepted Article that has been peer-reviewed and approved for publication in the Insect

Science but has yet to undergo copy-editing and proof correction. Please cite this article as doi:

10.1111/1744-7917.12545.

This article is protected by copyright. All rights reserved.

Author running head: X. Z. Ni et al.

Title running head: Predicting brown stink bug abundance in corn

Correspondence: Xin-Zhi Ni, USDA-ARS, Crop Genetics and Breeding Research Unit, University of

Georgia Tifton Campus, 2747 Davis Road, Bldg. #1,Tifton, GA 31793-0748, USA. Tel: +1 (229) 387-

2340; fax: +1 (229) 387-2321; email: [email protected]

ORIGINAL ARTICLE

Monitoring of brown stink bug (Hemiptera: Pentatomidae) population

dynamics in corn to predict its abundance using weather data

Xin-Zhi Ni1, Ted E. Cottrell2, G. David Buntin3, Xian-Chun Li4, Wei Wang5 and Hong Zhuang6

1USDA-ARS, Crop Genetics and Breeding Research Unit, Tifton, GA 31793, USA; 2USDA-ARS,

Southeastern Fruit and Tree Nut Research Laboratory, Byron, GA 31008, USA; 3Department of

Entomology, University of Georgia, Griffin, GA 30223, USA; 4Department of Entomology, University of

Arizona, Tucson, AZ 85138, USA; 5College of Engineering, China Agricultural University, No. 17

Tsinghua E. Road, Beijing, 100083, China and 6USDA-ARS, Quality and Safety Assessment Research

Unit, Athens, GA, USA

Abstract The brown stink bug (BSB), Euschistus servus (Say) (Hemiptera: Pentatomidae), is a serious

economic pest of corn production in the Southeastern U. S. The BSB population dynamics was

https://doi.org/10.1111/1744-7917.12545

https://doi.org/10.1111/1744-7917.12545


2

monitored for 17 weeks from tasseling to pre-harvest of corn plants (i.e., late May to mid-

September) using pheromone traps in three corn fields from 2005 to 2009. The trap data showed

two peaks in early June and mid-August, respectively. The relationship between trap catch and pre-

growing season weather data was examined using correlation and stepwise multiple factor

regression analyses. Weather indices used for the analyses were accumulated growing degree day

(AGDD), number of days with minimum temperature below 0°C (Subz), accumulated daily maximum

(AMaxT) and minimum temperatures (AMinT) and rainfall (ARain). The weather indices were

calculated with lower (10°C) and upper (35°C) as biological thresholds. The parameters used in

regression analysis were seasonal abundance (or overall mean of BSB adult catch) (BSBm), number

of BSB adults caught at a peak (PeakBSB), and peak week (Peakwk). The BSBm was negatively

related to high temperature (AmaxT or AGDD) consistently, whereas 1stPeakBSB was positively

correlated to both ARain and Subz, irrespective of weather data durations (the first 4, 4.5 and 5

months). In contrast, the 7-month weather data (AGDD7) were negatively correlated to the BSBm

only, but not correlated to the 2nd PeakBSB. The 5-year monitoring study demonstrated that

weather data can be used to predict the BSB abundance at its first peak in tasseling corn fields in the

southeastern U.S. states.

Key words Euschistus servus; first trap catch peak; pheromone trap catch; population dynamics;

stepwise regression modeling; weekly mean

The steady increase in the acreage of transgenic Bacillus thuringiensis (Bt) corn and cotton in the

southeastern coastal plain has reduced insecticide applications in cotton production for the

bollworm, also known as the corn earworm, Helicoverpa zea (Boddie) (Lepidoptera: Noctuidae)

management. The reduction in insecticide applications following the introduction of transgenic Bt


3

crops has resulted in increased population density and damage by piercing-sucking hemipteran pests

of corn (Zea mays L.), cotton (Gossypium hirsutum L.), and soybean [Glycine max (L.) Merrill] in the

Midwest (Koch & Pahs, 2014) and the southeastern U. S. (Bundy & McPherson, 2000; McPherson &

McPherson, 2000; Smith et al., 2009; Herbert & Toews, 2011). The three predominant

phytophagous stink bug species (Hemiptera: Pentatomidae) on corn, cotton, and soybean

production in Georgia are the brown stink bug (BSB), Euschistus servus (Say), the southern green

stink bug, Nezara viridula (L.), and the green stink bug, Chinavia hilare (Say) (Tillman, 2010, 2011;

Herbert & Toews, 2011).

However, limited information is available concerning the variation in seasonal stink bug

population dynamics in corn, cotton and soybean fields across the southeastern U.S., which impedes

the development of an effective integrated pest management program targeting stink bugs in these

ecosystems. The introduction of transgenic crops has created a unique opportunity to greatly

improve IPM programs in these ecosystems by bridging the knowledge gap between genetics and

ecology of both the crops and pests (Ni et al., 2014a). Because plant-feeding pentatomid adults are

highly mobile (or elusive when disturbed), and are polyphagous insects, monitoring their populations

on various host crop plants can be difficult, and costly (Cullen & Zalom, 2000; Kamminga et al.,

2009), which frequently causes a lack of precision for estimating population dynamics and

preventing and/or reducing their damage to crops.

In addition, BSB populations fluctuate greatly from year to year in the southeastern U. S., and trap

catch at population peaks can be critical for the timing of needed insecticide applications. Thus, a

seasonal monitoring study was initiated to describe BSB population dynamics in corn fields from

2005 through 2009 on three farms near Tifton, Georgia. The objectives of the experiment were in

twofold: (i) to document the number of BSB aggregation pheromone trap catch peaks in a corn field


4

from tasseling to pre-harvest; and (ii) to develop a model utilizing weather data (i.e., temperature

and rainfall) to specifically predict the number of BSB caught at the two peaks of trap catches and

the abundance of BSB on corn plants in a growing season.

Materials and methods

Insects and plants

Because no significant difference in damage to soybean was detected among the three species of

stink bugs, including N. viridula, C. hilare, and E. servus (or BSB) (Jones, 1979), and the BSB was

usually the dominant species on corn plants at the Tifton, GA location (Ni et al., 2010), the BSB was

chosen as the subject for the study. Commercial transgenic Bt corn hybrids and the planting dates

for the three corn field on three farms during the five year experimental period were recorded

(Supplementary Table 1). The fields planted for this study on the three research farms (i.e.,

Belflower, Lang-Rigdon, and Gibbs Farms) near Tifton, Georgia were surrounded by either riparian

(or woody) areas at the edge of the farms, fallow fields, or peanut, cotton, or sorghum fields, which

is very similar to the cropping systems as described previously by Ni et al. (2016). Standard

agronomic practices and University of Georgia Extension recommendations (Lee et al., 2007) were

applied to maintain the experimental fields (approximately 0.4 ha). Supplemental irrigation was

applied on an as-needed basis, and no insecticides were applied to the experimental corn fields.


5

Pheromone lure

The pheromone traps used in this study have been described in detail by Mizell and Tedders

(1995) and Cottrell et al. (2000). The aggregation pheromone [i.e., methyl (2E, 4Z)-decadienoate]

and yellow color of the pyramidal trap are the two main chemical and visual cues for attracting the

BSB adults. For each trap, a cattle ear-tag treated with pyrethroid insecticides [(13% piperonyl

butoxide, and 10% λ-cyhalothrin) by Schering-Plough Animal Health Corp., Summit, New Jersey] was

placed in the pheromone trap to kill the insects captured in the trap.

Trap locations and monitoring in a corn field

A total of 10 traps were placed evenly throughout a corn field with three traps on two opposing

sides of the field and four traps through the center of the fields, as shown by Ni et al. (2016).

Because stink bugs prefer to feed on developing fruits/seeds of plants, and low level of stink bug

damage can be seen on seedling stage of corn plants with high level of overwintering population (Ni

et al., 2010; Ni et al., 2016), monitoring of the BSB population in the corn fields started when the

plants were close to flowering or at tasseling (or VT) stage. The corn planting dates varied because of

variation in weather conditions from year to year (Supplementary Table 1). The monitoring period

stopped after 17 weeks at pre-harvest of the corn. The trap catch was monitored twice per wk, and

the total number of BSBs caught per wk was recorded. Because of the sexes of the BSB adults were

not recorded in all years, and very few nymphs were caught in the traps, no analyses were

performed either between two sexes of adults or trap catch of nymphs throughout the season.

Three parameters relating to insect monitoring (as described in Table 1) were: (i) overall mean

number of BSB adults per trap per wk for the 17-week period (BSBm); (ii) the number of BSB caught


6

at the first or second peak (PeakBSB); and (iii) the week when the peak of trap catch occurred

(Peakwk).

Weather data collection and degree day calculation

For the practical adaptation feasibility, a total of five weather-related parameters were used in

data analysis (Table 1). The five parameters were: (1) the number of days with a minimum

temperature below 0°C (Subz) during winter, i.e., from the first to the last day of the frost between

November of the previous year to March of the monitoring year; (2) Accumulated daily maximum

temperature (AMaxT); (3) Accumulated daily minimum temperature (AMinT); (4) Accumulated daily

rainfall (ARain); and (5) accumulated growing degree days (AGDD).

The daily maximum and minimum temperatures and precipitation data from the Tifton location

were retrieved from www.GeorgiaWeather.net for calculation of the AGDD, which was the sum of

daily growing degree day (GDD). The AGDD was calculated using 10°C as lower developmental

threshold, and 35°C as the upper threshold (Murray, 2008). Because the daily maximum (or

minimum) temperature frequently lasts a relatively short period of time (i.e., minutes in some cases)

in a 24 hours period, the adjustment of the developmental threshold is necessary to increase the

precision of the AGDD calculation in corroboration with the biological significance of insect exposure

to weather conditions (Murray, 2008). Thus, the GDD was calculated with the following five

scenarios; (1) when daily maximum and minimum temperatures are below the lower threshold

(10°C), GDD = (lower threshold-lower threshold)/2, or 0; (2) when maximum temperature is above

lower threshold, and minimum temperature is below the lower threshold, GDD = (maximum - lower

threshold)/2-lower threshold; (3) when maximum and minimum temperatures are between the

upper and lower threshold, GDD = (maximum + minimum)/2-lower threshold; (4) when maximum


7

temperature is above and minimum temperature is below the upper threshold, GDD = (upper

threshold+ minimum)/2-lower threshold; and 5) when daily minimum and maximum temperatures

are above the upper threshold, GDD = (upper threshold-lower threshold), or 25 for the current

study.

Because weather data varied from year to year, AGDD was calculated using two biological fix

(biofix) dates (i.e., the first frost day in winter time and the calendar year) to identify the best set of

the AGDD data for the regression model. Within a biofix date, the AGDD was calculated in four

durations, i.e., the first 4, 4.5, 5 and 7 months. While AGDD data from 4, 4.5, and 5 months were

used for predicting the number of brown stink bugs at the first peak, the 7 month data were used to

predict the trap catch at the 2nd peak at pre-harvest. Because variation in weather data usually

occurs between January and April in the southeastern U.S., which is accompanied with the SBS

overwintering period, the AGDD data and other weather-related indices from the first 4, 4.5, and 5

months close to the first trap catch peak (coincide with corn plant tasseling) were utilized for

stepwise multiple factor regression modeling. The use of three durations with a 15 days interval was

to identify the best model(s) with precision in predicting the trap catch at the first peak in the

upcoming growing season of a year.

Experimental design and data analysis

The experiment utilized a split-split-plot design with split in both time (five years) and space (three

farms within each year) according to Cochran and Cox (1957). Each of the three corn fields on

different farms per year were considered as the main plot, and the 17-week of the monitoring

period was considered split in time. The three farms were considered replications. The ten traps

were considered nested within a field, and placed equidistant in each field by assigning three traps


8

on each longitudinal edge of a field and four traps throughout the center of a field (Ni et al., 2016).

The location assignment of the traps was to assess the effect of the field edge on the BSB trap catch.

The weekly mean of the BSB trap catch was subjected to analysis of variance (ANOVA) using a mixed

effects model [PROC MIXED procedure of the SAS software package (SAS Institute, 2012)]. In the

ANOVA, three variables [i.e., year, week, and trap location (edge versus interior)] were used as fixed

factors for the PROC MIXED procedure with a REPEATED statement, whereas the 10 traps nested

within a field were considered random factors. The difference in weekly trap catch within a year, and

among the years, as well as the effect of trap location (i.e., edge versus interior) in a corn field were

also compared. The trap catch data were analyzed after logarithm transformation, because of the

great variation throughout a growing season. All graphs were generated using Sigma Plot® (version

11.0) (SYSTAT, Richmond, CA). The correlation between the trap catch parameters (i.e., BSBm,

peakBSB, and Peakwk) and five weather data parameters (as described in Table 2) were examined

using PROC CORR procedure (SAS Institute, 2012). After the trap catch data had been subjected to

the Shapiro-Wilk test for normality as described by Peng (2009), the stepwise regression analysis

(PROC REG) was performed to assess the relationship between the BSB trap catch and weather data

(SAS Institute, 2012). The regression model was validated using the original five-year data (2005-

2009), correlation between the observed and predicted values were calculated. In addition, the

prediction of the BSB abundance (i.e., 1st peakbsbv, and bsbmv) was performed using the regression

model with the 4.5 months in the most recent 7-year weather data (from 2010 to 2016) at Tifton,

Georgia location.

Results

Seasonal BSB population dynamics


9

The overall pooled trap catch data from 2005 through 2009 (n = 2550) was significantly different

among the 17 weeks (F = 46.51; df = 16, 128; P = 0.0001) (Fig. 1), among the five years (F = 97.21; df

= 4, 32; P = 0.0001) (Fig. 2), and between the edge and interior of a corn field (F = 9.76; df = 1, 8; P =

0.01). The weekly trap catch was also affected by year × week (F = 7.79; df = 64, 512; P = 0.0001),

and year × location (F = 11.61; df = 4, 32; P = 0.01) interactions, whereas the trap catch data were

not affected by week × location, or year × week × location interactions (P > 0.10). The pooled data of

all five years showed that the trap catch at the edge was greater (4.71 ± 0.13, n = 2040) than that in

the interior of a field (3.91 ± 0.23, n = 510). When the trap catch was further analyzed within a year

(n = 510), trap catch differed among years for the 17-week period every year (Figs. 2A-E), but there

was no significant difference in trap catch between corn field edge and interior within a year, except

in 2007 and 2009. Both pooled and annual data showed two peaks of the BSB trap catch in corn

fields from late May to mid-September (Figs. 1 and 2). The first peak occurred at the end of May to

early June while the second peak occurred in mid-August at pre-harvest. In addition, because trap

catch data only recorded the first peak partially in three years (2006-2008) as shown in Figs. 2B, C,

and D, respectively, the peak week was not further examined in regression analysis. Because Ni et al.

(2016) reported no difference between trap locations in relation to surrounding harvested rye field

and pine tree nursery, no further analysis was performed in the current study for trap location in

relation to the surrounding crop fields or ecological habitats.

Temperature, rainfall, and growing degree day data

All variables related to weather data (Supplementary Table 2) were calculated based on the

original weather data collected at the Tifton location weather station. The pre-growing season

weather data were used in either biofix dates for the calculation of accumulated degree days


10

(AGDD). When the calendar year was used as the biofix date, the AGDD of the first 4, 4.5, and 5

months of each year was calculated separately for the regression analysis for the first peak, while

the 7-month data were used for the regression analysis of the 2nd peak (Supplementary Table 2).

While all four durations of AGDD calculated based on the calendar biofix date were presented here,

only one of the four durations (4.5 months) was presented when the first frost day was used as

biofix date. All 8 variables related to trap catch and weather data (Supplementary Table 2) were

used for further correlation and stepwise regression analyses.

Correlation of trap catch with weather data

Because weather data vary among years, as well as from January to mid-May, and weather data

between 1 January to 15 May were also close to the first peak of BSB catch, which was started in late

May and ended in mid-September, the correlation analysis between trap catch (i.e., BSBm,

1stPeakBSB, and 1stPeakwk) and weather data for the first 4.5 months were performed (Table 2).

The number of stink bugs caught at the first peak (1stPeakBSB) was positively correlated to the

number of subzero days (Subz) and rainfall (ARain4.5), but the overall mean (BSBm) were negatively

correlated to the AGDD4.5 and AMaxT4.5, respectively. The 1stPeakwk was not correlated to any of

the weather indices (Table 2), so the Peakwk data were not used in regression modeling. The

correlation analyses between trap catch and other weather data durations (i.e., the 4, 5 and 7

months, respectively) showed similar patterns, and the results are not presented here.

Stepwise multiple factor regression modeling


11

The normality test confirmed that the trap catch dataset was not different from the normal

distribution (Shapiro-Wilk test of normality for trap data was not significant, P values = 0.4–0.95).

The stepwise regression analysis was conducted individually for each of the three trap catch data

parameters against the five weather data parameters as described in Table 1. A total of 12 (4

periods of early season weather data x 3 trap catch data parameters) regression analysis procedures

were conducted. However, the weekly mean of the trap catch (BSBm) can be predicted using the

five-year dataset for all weather data periods (i.e., 4, 4.5, 5, and 7 months), and the number of stink

bugs at the first peak (1stPeakBSB) can be predicted using 4, 4.5, and 5 month weather data. The 7-

month data could only predict the overall weekly mean (BSBm), but not stink bug catch at the

second peak (2ndPeakBSB). In addition, when the first frost day was used as the biofix date, the

goodness-of-fit of the regression model was not as good as the previous seven models using only

calendar-year weather data as the biofix date (Table 3). Thus, only one regression model based on

the AGDD data using the first frost day as the biofix date was developed and presented in Table 3.

The stepwise multiple factor regression analyses utilized 2 to 5 steps (Table 3). The overall weekly

mean (BSBm) was negatively correlated to AMaxT or AGDD7, while the trap catch at the first peak

(1stPeakBSB) was positively correlated to the number of Subz days, and ARain (Table 3).

Validation of the linear regression models

The model validation (or predicted) values (i.e., 1stpeakbsbv, bsbmv) were paired with the

corresponding field observations (i.e., 1stPeakBSB and BSBm) for each year, as shown in

Supplementary Table 3. The model prediction of the 1stpeakbsbv and bsbmv using 4.5-month

weather data showed the least deviation between the observed and predicted values shown by the

bolded values of the means and standard errors in Supplementary Table 3. The correlation


12

coefficients between observed and predicted values of the peak data (1stPeakBSB) were greater (r =

0.997, P = 0.0001, n = 15) than the observed and predicted values of weekly means (BSBm) (r =

0.911, P = 0.0001, n = 15). In addition, the 1stPeakBSB was also positively correlated to BSBm (r =

0.649, P = 0.009, n = 15). The validations of the two models using the 7-month weather data were

not performed because the model is of little value at the end of the growing season, which was not

as useful as the six models using the pre-growing season weather data. Similarly, the model using

the first frost day + 4.5-month weather data was not validated because the goodness-of-fit for the

model was lower (r2 < 0.87) than the other six models associated with the first peak (r2 > 0.89) as

shown in Table 3. Furthermore, the BSB abundance in the most recent 7 years (from 2010 to 2016)

was also calculated using the linear regression models with 4.5-month weather data (Supplementary

Table 4). The correlation of two predict values (i.e., 1stpeakbsbv and bsbmv) were not significant (r =

0.742, P = 0.056, n = 7), which could be caused extremely warm winter for 2011–2012 that led to the

bsbmv value being negative in 2012, as shown in Supplementary Table 4.

Discussion

Since the introduction of synoptic population model by Southwood and Comins (1976), a number of

models that use deterministic and stochastic factors influencing population dynamics of both r- and

K-strategist insect pests have been examined in detail in ecological research of insect pests (Huffaker

et al., 1984). In particular, utilization of the AGDD to understand insect biology and physiology and

then to predict insect population dynamics throughout a growing season of a given crop has been

examined extensively in recent decades on a number of insect pests and their host plants for both

agricultural (Higley et al., 1986; Kingsolver, 1989; Murray, 2008) and forestry ecosystems (van Asch

& Visser, 2007). For pentatomid pests, as a group of K-strategists, Kamminga et al. (2009) described


13

a model using weather data to predict C. hilare abundance. They reported that seasonal flight

activities of C. hilare can be predicted by utilizing an 18-year (1990-2007) dataset of black light catch

in Virginia, U.S.A. They described that two weather parameters (i.e., mean monthly precipitation and

number of days below freezing) in the early season (from January to April) were adequate to predict

the weekly mean of adults caught in a black light trap. In addition, the three peaks of the black light

catch throughout the season with 10%, 50%, and 90% of the seasonal catch were identified to occur

at 153, 501 and 1 066 degree days from 1 January of a year (Kamminga et al., 2009). Nielson et al.

(2013) reported that black light traps can be used to monitor the distribution and abundance of an

invasive species - brown marmorated stink bug, Halyomorpha halys (Stål) (Hemiptera:

Pentatomidae) in New Jersey. A network of more than 70 black light traps throughout New Jersey

on vegetable and fruit farms participated in the IPM scouting program from 2004–2011 were utilized

to monitor H. halys distribution and abundance, in addition to key lepidopteran pests [e.g., the

European corn borer, Ostrinia nubilalis Hübner (Lepidoptera: Crambidae), and Helicoverpa zea

(Boddie) (Lepidoptera; Noctuidae)] in New Jersey (Nielson et al., 2013). They determined that 685

degree days with the lower threshold of 14°C was required for female H. halys maturation, which fell

into the 29th and 32nd Julian weeks in southern and northern New Jersey, respectively.

The current study is one of a series of studies striving to understand the biology and ecology of

BSB population dynamics associated with corn production. In addition to understanding the impact

of BSB feeding damage on grain quality in corn production (Ni et al., 2010), stink bug damage on

corn kernels was correlated to aflatoxin contaminations in a corn field during some years, but not

every year (Ni et al., 2011; Ni et al., 2014c). However, the common smut, Ustilago maydis (Persoon)

Roussel, infections of corn ears were not correlated to the BSB carrying smut spores (Ni et al.,

2014b). Diurnal flight activities of the BSB adults assessed using pheromone traps when corn plants

were at tasseling stage (VT), coincides with the harvest of winter grain crops (e.g., wheat and rye)


14

(Reisig, 2011; Ni et al., 2016). The findings from the current study on the BSB (or E. servus), a

bivoltine insect that overwinters as adults (McPherson & McPherson, 2000), are consistent with the

report on C. hilare abundance in relation to weather data (Kamminga et al., 2009). The cause of

positive correlation of the first peak BSB catch to cold and wet wintery weather (i.e., Subz and ARain)

could be the result of high survivorship of sustained overwintering (diapause) BSB adult populations.

In contrast, the negative correlation between the BSB abundance (BSBm) and temperature (i.e.,

AMaxT or AGDD) could be the result of termination of diapause, which led to high mortality under

brief deep frost conditions subsequently. Temperature and photoperiod have been well

documented for inducing and terminating insect diapause in general (Xu et al., 2014). The cold and

wet conditions under short photoperiod in winter or early spring sustained diapause state of the

overwintered adults, which leads to high survivorship, and later high trap catch of the overwintering

adults at the first peak of the BSB monitoring period. In contrast, an extended period with high

temperature during an insect diapause might terminate diapause, and subsequent freezing

temperature with precipitation in winter or early spring could lead to high mortality of a post-

diapause insect population. Thus, the 1stPeakBSB was high following a cold and wet winter and vice

versa. The six linear regression models for the first peak described in the current study could be

readily used to predict the BSB abundance at corn plant tasseling (flowering) time, which is critically

important for insecticide applications to prevent corn ear and kernel damage (Reisig, 2011), as well

as allowing growers to be ready for assessing BSB abundance in other crops (e.g., cotton, peanut,

and soybean) that flower at a later time in the cropping systems of the southeastern U.S. (Greene et

al., 2001; Herbert & Toews, 2011; Tillman, 2011; Temple et al., 2013). The linear regression models

between pheromone trap catch of the BSB adults at the first peak and pre-growing season (4.5

month) weather data described here could be critically valuable to preserve maize and other crop

yield and quality in all southern U.S. states under warm temperate climate with similar cropping


15

systems (e.g., corn/sorghum, cotton, and peanut/soybean as main field crops, and pecan and peach

as main orchard crops). In a similar manner, the model for the second peak of BSB adult catch and

the 7-month weather data could be further examined and utilized to estimate overwintering

population of this bivoltine pest in the southeastern U.S. states. Such information would be valuable

in integrated management of the brown stink bug and other pests in the agricultural ecosystems in

the southeastern coastal plain region of the U.S.

In conclusion, the regression models developed from the current 5-year study can be utilized to

predict the BSB abundance in corn using pre-growing season weather data by targeting the two

peaks. While the predicted stink bug number at the first peak would be valuable in designing

contingent management strategies in corn and other crops of the same year, the predicted stink bug

abundance at the second peak could be used to estimate the overwintering population for this

bivoltine pest, which is valuable for assessing brown stink bug abundance for the following year. In

particular, the two models developed using the first 4.5 month (or pre-growing season) weather

data of a year, would allow growers to predict BSB damage using trap catch at the first peak

(1stPeakBSB) and seasonal BSB abundance (BSBm) before the BSB infestation actually occurs, which

is not only critically important for timely stink bug management on corn, but also on cotton and

soybean production in the southeastern U.S. states.

Acknowledgments

Mention of trade names or commercial products in this article is solely for the purpose of providing

specific information and does not imply recommendation or endorsement by the U. S. Department

of Agriculture. We thank G. Gunawan, K. Da, R. Powell, Jr., J. C. Mullis, and P. M. Tapp (USDA-ARS,


16

Crop Genetics and Breeding Research Unit, Tifton, GA) for their technical assistance during the

experiment. We thank J. K. Westbrook (USDA-ARS Insect Control and Cotton Disease Research Unit,

College Station, TX) and N. C. Elliott (USDA-ARS Plant Science Research Laboratory, Stillwater, OK) for

their critical reviews of the earlier version of the manuscript. We also thank anonymous reviewers

and the editor for their advice and constructive comments that have strengthened the manuscript.

The research was supported in part by the Georgia Agricultural Commodity Commission for Corn.

Disclosure

All authors declare no conflict of interest.

References

Bundy, C.S. and McPherson, R.M. (2000) Dynamics and seasonal abundance of stink bugs

(Heteroptera: Pentatomidae) in a cotton–soybean ecosystem. Journal of Economic

Entomology, 93, 697–706.

Cochran, W.G. and Cox, G.M. (1957) Factorial experiments with main effects confounded: split-plot

designs. Experimental Designs (eds. W.G. Cochran & G.M. Cox), pp. 293–316. John Wiley and

Sons, Inc. New York, NY.

Cottrell, T.E., Yonce, C.E. and Wood, B.W. (2000) Seasonal occurrence and vertical distribution of

Euschistus servus (Say) and Euschistus tristigmus (Say) (Hemiptera: Pentatomidae) in pecan

orchards. Journal of Entomological Science, 35, 421–431.


17

Cullen, E.M. and Zalom F.G. (2000) Phenology-based field monitoring for consperse stink bug

(Hemiptera: Pentatomidae) in processing tomatoes. Environmental Entomology, 29, 560–

567.

Greene, J.K., Turnipseed, S.G., Sullivan, M.J. and May, O.L. (2001) Treatment thresholds for stink

bugs (Hemiptera: Pentatomidae) in cotton. Journal of Economic Entomology, 94, 403–409.

Herbert, J.J., and Toews, M.D. (2011) Seasonal abundance and population structure of brown stink

bug (Hemiptera: Pentatomidae) in farmscapes containing corn, cotton, peanut, and soybean.

Annals of the Entomological Society of America, 104, 909–918.

Higley, L.G., Pedigo, L.P. and Ostlie, K.R. (1986) DegDay: a program for calculating degree-days, and

assumptions behind the degree-day approach. Environmental Entomology, 15, 999–1016.

Huffaker, C.B., Berryman, A.A. and Laing, J.E. (1984) Natural control of insect populations. Ecological

Entomology (eds. C.B. Huffaker & R.L. Rabb), pp. 359–398. John Wiley and Sons, New York.

Kamminga, K.L., Herbert Jr. D.A., Kuhar, T.P. and Brewster, C.C. (2009) Predicting black light trap

catch and flight activity of Acrosternum hilare (Hemiptera: Pentatomidae) adults.

Environmental Entomology, 38, 1716–1723.

Kingsolver, J.G. (1989) Weather and the population dynamics of insects: integrating

physiological and population ecology. Physiological Zoology, 62, 314–334.

Koch, R.L. and Pahs, T. (2014) Species composition, abundance, and seasonal dynamics of stink bugs

(Hemiptera: Pentatomidae) in Minnesota soybean fields. Environmental Entomology, 43,

883–888.


18

Lee, R.D., Harris, Jr. G.H., Prostko, E., Buntin, G.D., Kemerait, R. and Harris, K.G. et al. (2007) A guide

to corn production in Georgia 2008. Georgia Coop. Extension Service Misc. Publication CSS

01-08, 61 pp.

McPherson, J.E. and McPherson, R.M. (2000) Euschistus spp. Stink Bugs of Economic Importance in

America North of Mexico (eds. J.E. McPherson & R.M. McPherson), pp. 101–127. CRC Press,

Boca Raton, FL.

Mizell III, R.F. and Tedders W.L. (1995) A new monitoring method for detection of the stinkbug

complex in pecan orchards. Proceedings of Southeastern Pecan Growers Association, 88, 36–

40.

Murray, M.S. (2008) Using Degree Days to Time Treatments for Insect Pests. Utah State University

Extension and Utah Plant Pest Diagnostic Laboratory IPM-05-08, April 2008, 5 pp.

Ni, X.Z., Da, K., Buntin, G.D., Cottrell, T.E., Tillman, P.G. Olson, D.M. et al. (2010) Impact of brown

stink bug (Heteroptera: Pentatomidae) feeding on corn grain yield components and quality.

Journal of Economic Entomology, 103, 2072–2079.

Ni, X.Z., Wilson, J.P., Buntin, G.D., Guo, B., Krakowsky, M.D. Lee, R.D.et al. (2011) Spatial patterns of

aflatoxin levels in relation to ear-feeding insect damage in pre-harvest corn. Toxins, 3, 920–

931.

Ni, X.Z., Lei, Z., He, K., Li, X., Li, X. and Xu, W. (2014a) Integrated pest management is the lucrative

bridge connecting the ever emerging knowledge islands of genetics and ecology. Insect

Science, 21, 537–540.


19

Ni, X.Z., Toews, M.D., Buntin, G.D., Carpenter, J.E., Huffaker, A. Schmelz, E.A. et al. (2014b) Influence

of brown stink bug feeding, planting date and sampling time on common smut infection of

maize. Insect Science, 21, 564–571.

Ni, X.Z., Wilson, J.P., Toews, M.D., Buntin, G.D., Lee, R.D. and Li, X. et al. (2014c) Evaluation of spatial

and temporal patterns of insect damage and aflatoxin level in the pre-harvest corn fields to

improve management tactics. Insect Science, 21, 572–583.

Ni, X., Cottrell, T.E., Toews, M.D., Tillman, P.G. and Buntin, G.D. (2016) Diurnal activities of the brown

stink bug (Hemiptera: Pentatomidae) in and near tasseling corn fields. Journal of

Entomological Science, 51, 226–237.

Nielsen, A.L., Holmstrom, K., Hamilton, G.C., Cambridge, J., Ingerson–Mahar, J. (2013) Use of black

light traps to monitor the abundance, spread, and flight behavior of Halyomorpha halys

(Hemiptera: Pentatomidae). Journal of Economic Entomology, 106, 1495–1502.

Peng, C.Y.J. (2009) Comprehensive descriptive analysis and normality test. In: SAGE Research

Methods, pp. 163–207.

Reisig, D.D. (2011) Insecticidal management and movement of the brown stink bug, Euschistus

servus, in corn. Journal of Insect Science, 11, 168.

SAS Institute (2012) SAS/STAT version 9.4 User’s guide. The SAS Institute, Cary, NC.

Smith, J.F., Luttrell, R.G. and Greene, J.K. (2009) Seasonal abundance, species composition, and

population dynamics of stink bugs in production fields of early and late soybean in South

Arkansas. Journal of Economic Entomology, 102, 229–236.


20

Southwood, T.R.E. and Comins, H.N. (1976) A synoptic population model. Journal of Animal Ecology,

45, 949–965.

Temple, J.H., Davis, J.A., Micinski, S., Hardke, J.T., Price, P. and Leonard, B.R. (2013) Species

composition and seasonal abundance of stink bugs (Hemiptera: Pentatomidae) in Louisiana

Soybean. Environmental Entomology, 42, 648–657.

Tillman, P.G. (2010) Composition and abundance of stink bugs (Heteroptera: Pentatomidae) in corn.

Environmental Entomology, 39, 1765–1774.

Tillman, P.G. (2011) Influence of corn on stink bugs (Heteroptera: Pentatomidae) in subsequent

crops. Environmental Entomology, 40, 1159–1176.

van Asch, M. and Visser, M.E. (2007) Phenology of forest caterpillars and their host trees: the

importance of synchrony. Annual Review of Entomology, 52, 37–55.

Xu, L., Ni, X., Wang, Z. and He, K. (2014) Effects of photoperiod and temperature on diapause

induction in Conogethes punctiferalis (Lepidoptera: Pyralidae). Insect Science, 21, 556–563.

Manuscript received June 20, 2017

Final version received September 15, 2017

Accepted September 24, 2017


21

Table 1 Acronyms of trap catch and weather data used for regression analysis

Abbreviation Variable

Trap catch

BSBm Overall weekly mean of brown stink bugs per trap throughout a season

PeakBSB Number of brown stink bugs recorded at the first or second peak

Peakwk The week when the first or second peak of stink bug catch occurred

Weather data

Subz The number of days with frost, or minimum temperature < 0°C throughout the whole

winter

AGDD Accumulated growing degree day

AMinT Accumulated daily minimum temperature in Celsius

AMaxT Accumulated daily maximum temperature in Celsius

ARain Accumulated rainfall (mm)

This is an Accepted Article that has been peer-reviewed and approved for publication in the Insect

Science but has yet to undergo copy-editing and proof correction. Please cite this article as doi:

10.1111/1744-7917.12545.


Table 2 Pearson’s Correlation Coefficients of 8 indices of the five year (n = 5) data for the 4.5-month

weather data from 1 January to 15 May between 2005 and 2009

Subz AGDD4.5 AMaxT4.5 AminT4.5 ARain4.5 1stpeakBSB 1stpeakwk

AGDD4.5 −0.42

0.49

AMaxT4.5 −0.62 0.96

0.26 0.009

AMinT4.5 −0.28 0.68 0.59

0.64 0.21 0.29

ARain4.5 0.62 −0.34 −0.55 0.10

0.26 0.57 0.34 0.88

1stpeakBSB 0.89 −0.45 −0.67 −0.17 0.90

0.04 0.44 0.21 0.78 0.04

1stpeakwk 0.66 −0.79 −0.86 −0.25 0.45 0.60

0.22 0.11 0.06 0.68 0.44 0.29

BSBm 0.54 −0.96 −0.99 −0.58 0.58 0.65 0.80

https://doi.org/10.1111/1744-7917.12545

https://doi.org/10.1111/1744-7917.12545


23

0.35 0.009 0.002 0.30 0.30 0.23 0.11

In each table cell, top value = r, bottom value = P of the MANOVA statement of the SAS

software. The bolded values indicate significant correlation with P < 0.05.

Table 3 Linear regression models based on different durations of weather data.

Biofix date Regression equation† Statistics of stepwise

regression (P < 0.05)

Calendar

year 4 months

BSBm = 49.03 − 0.019×(AMaxT4) r2 = 0.887; 2 steps

1stPeakBSB = −13.20 + 0.825×(Subz) +

0.034×(ARain4) r2 = 0.996; 3 steps

4.5 months BSBm = 57.19 − 0.019×(AMaxT4.5) r2 = 0.973; 2 steps

1stPeakBSB = −13.44 + 0.896×(Subz) +

0.030×(ARain4.5) r2 = 0.994; 3 steps

5 months BSBm = 55.07 − 0.016×(AMaxT5) r2 = 0.914; 4 steps

1stPeakBSB = − 9.70 + 0.683×(Subz) +

0.027×(ARain5) r2 = 0.995; 5 steps

7 months BSBm = 43.82 − 0.025×(AGDD7) r2 = 0.865; 2 steps

First frost

day 4.5 months 1stPeakBSB = −9.86 + 0.045×(ARain) r2 = 0.819; 4 steps

†In regression equation column, BSBm = overall weekly mean of brown stink bugs per trap

throughout the 17-week monitoring period; 1stPeakBSB = number of brown stink bugs at the first

peak; AGDD = accumulated growing degree days of the first 7 months; AMaxT or AMinT =


24

accumulated maximum or minimum daily temperature readings; and ARain = accumulated rainfall

data. Subz = number of days with minimum temperature below 0°C.

Figure Captions

Trap Catch of Brown Stink Bugs (2005-2009, n =150)

Week (AGDD)

1 (9

51.9

3)

2 (1

052.

78)

3 (1

156.

57)

4 (1

273.

66)

5 (1

388.

03)

6 (1

506.

42)

7 (1

625.

39)

8 (1

738.

71)

9 (1

860.

39)

10 (1

982.

29)

11 (2

102.

96)

12 (2

226.

97)

13 (2

351.

22)

14 (2

472.

75)

15 (2

590.

17)

16 (2

704.

77)

17 (2

812.

67)

Num

be

r o

f b

row

n s

tink b

ug

s p

er

tra

p

0

2

4

6

8

10

12

Fig. 1 Weekly mean of trap captured E. servus adults (or BSBm) during the 17-week monitoring

period from 2005 to 2009 with 30 traps per year (10 traps for each of the three fields) (n = 150). The

error bars represent the standard error of the mean. Week (AGDD) denotes the five-year mean of

AGDD (calculated from 1 January of each year) at the beginning of the weekly monitoring period.


25

E) 2009 (n = 30)

Week

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

0

5

10

15

20

25

30

C) 2007 (n = 30)

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Nu

mb

er

of b

row

n s

tink

bu

gs

pe

r tr

ap

0

5

10

15

20

25

30

D) 2008 (n = 30)

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

0

5

10

15

20

25

30

B) 2006 (n = 30)

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

0

5

10

15

20

25

300 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

0

5

10

15

20

25

30

Week: F = 15.44; df = 16,128; P < 0.0001

Trap location: F = 5.74; df = 1,8; P = 0.04

Week*location: F = 0.73; df = 16,128, P = 0.76

A) 2005 (n = 30)

Week: F = 26.01; df = 16,128; P < 0.0001



Week: F = 7.10; df = 16,128; P < 0.0001

Trap location: F = 18.27; df = 1,8; P < 0.003


Week: F = 14.18; df = 16,128; P < 0.0001



Week: F = 16.30; df = 16,128; P < 0.0001



Fig. 2 Weekly mean of trap captured E. servus adults during the 17-week monitoring period with 10

traps in each of the three corn fields (n = 30) from 2005 to 2009, respectively. The error bars

represent the standard error of the mean.

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页文献云下载图书馆入口外文数据库大全疑难文献辅助工具

http://www.xuebalib.com/cloud/

http://www.xuebalib.com/

http://www.xuebalib.com/cloud/


http://www.xuebalib.com/vip.html

http://www.xuebalib.com/db.php

http://www.xuebalib.com/zixun/2014-08-15/44.html


Documents

Monitoring of brown stink bug (Hemiptera: Pentatomidae ...download.xuebalib.com/8ey6hYR3dDug.pdf · Arizona, Tucson, AZ 85138, USA; 5College of Engineering, China Agricultural University,