ABSTRACT DEL POZO-VALDIVIA, ALEJANDRO IVAN. Effects Of

ABSTRACT

DEL POZO-VALDIVIA, ALEJANDRO IVAN. Effects of Agronomic Practices on the Invasive Soybean Pest Megacopta cribraria (Hemiptera: Plataspidae) (Under the direction of Dr. Dominic D. Reisig).

Megacopta cribraria (F.) (Hemiptera: Plataspidae), also known as the kudzu bug, is a pest of soybean, Glycine max (L.) Merr., in the U.S. First-generation adults were effectively reared on soybean under greenhouse conditions, indicating that overwintering individuals could bypass feeding on kudzu, Pueraria montana Loureiro (Merrill) variety lobata

(Willdenow), and develop on early planted soybean if available. In soybean, more adults were caught on white sticky cards between 13:00 – 15:00 hours, implying that peak flight activity of kudzu bug occurs during this period; when cards were visited from 09:00 – 17:00 hours. Kudzu bugs aggregated in the middle portion of the main stem of soybean plants when scouting was conducted from 09:00 – 12:00 hours.

Currently, insecticides are the primary tactic used to manage this insect in commercial soybean fields. The main goals of these studies were to expand the management toolbox for the kudzu bug in soybean, provide information on how production practices in soybean affect kudzu bug, and develop information that will help growers select practices to minimize kudzu bug density and injury. Research from a two-state, two-year, multi-site study indicated that kudzu bug density was influenced by planting date and insecticide use in soybean; kudzu bug density was 55% lower in untreated soybean planted during June, compared with untreated plots planted during April. Kudzu bug population density was not consistently influenced by soybean maturity group. Furthermore, kudzu bug injury negatively affected soybean yield, with a reduction up to 20%, when insecticide protected plots were compared with untreated plots. Soil tillage was found to affect kudzu bug density in soybean. A two-year and multi- site experiment in NC indicated that more kudzu bugs were found in soybean plots that received conventional tillage, in which less than 5% of the soil was covered by the previous crop residue, compared with soybean planted in plots receiving conservation tillage which had more than 85% of the ground cover with crop residue. Tillage affected soybean yield, with conventionally tilled plots yielding higher compared with plots under reduced-tillage conditions. Kudzu bug injury is one of the factors impacting yield, when insecticide treated plots were compared to untreated plots; however, the effect of tillage on yield was independent from kudzu bug densities.

Tillage was used to manipulate the ground cover of experimental soybean plots. Plots ranged from those with almost no ground cover in conventionally tilled plots, to ~100% ground cover in plots where cereal rye, Secale cereale L. stubble was rolled on top of the soil. Although, the effects of tillage in this study were confounded with the effects of ground cover, there was a negative relationship between kudzu bug adult density and amount of ground cover prior to soybean canopy closure. I hypothesize light reflectance from both soil and soybean plants may be used as visual cue for kudzu bug to identify the location of a suitable host plant. Higher light reflectance profiles were measured from experimental plots with more crop residue, across a range of ground covers. A relationship between spectral indexes and kudzu bugs, prior to soybean canopy closure, may indicate that light reflectance may affect the establishment of these insects. Follow up research may corroborate my hypothesis, and may propose the use of highly reflective mulches as a potential management tactic for organic soybean growers. Population densities of kudzu bug were inconsistently affected by row spacing in soybean. Row spacing modified plant architecture and subsequently may have influenced potential visual cues for kudzu bug to find soybean plants. Complementary research on how planting date and row spacing may affect kudzu bug population dynamics in soybean is needed to clarify previous inconsistent results.

by Alejandro Ivan Del Pozo-Valdivia

A dissertation submitted to the Graduate Faculty of North Carolina State University in partial fulfillment of the requirements for the degree of Doctor of Philosophy

Entomology

Raleigh, North Carolina

2016

APPROVED BY:

______Dr. Dominic D. Reisig, Co-chair Dr. Jack S. Bacheler, Co-chair

______Dr. Clyde E. Sorenson Dr. George G. Kennedy

______Dr. Mark R. Abney

DEDICATION

To my mom, mamita Silvia,

To papa Bernardino and mama Hilda,

To my wife Christie and my son Joshua,

Y porque tengo el orgullo de ser Peruano!

BIOGRAPHY

Alejandro is originally from Lima, Perú. He received his B.S. – Agronomy from La

Molina National Agrarian University in 2004. Upon completion of his degree, Alejandro worked at one of the largest asparagus producers in Perú and was in charge of the design and implementation of an asparagus IPM program. Alejandro obtained an M.S. in Entomology from Washington State University, under the direction of Dr. John Brown. His research focused on describing the life history of an important defoliator pest in hybrid poplars.

Alejandro pursued his Ph.D. at North Carolina State University, under the direction of Dr.

Dominic Reisig. His project focused on investigating how production practices influence the soybean pest, Megacopta cribraria, also called kudzu bug. The ultimate goal of Alejandro’s research was to provide practical solutions to manage this new soybean pest in the

Southeastern U.S. Alejandro has been active within the Entomological Society of America, served as the Student Representative to the Plant-Insect Ecosystem Governing Council, and presented results of his research at branch and national meetings. He was the recipient of the

2012 Entomological Foundation’s Larry Larson Graduate Student Award for Leadership in

Applied Entomology. Alejandro was also actively supporting multicultural events at the

Office of International Services and the Latino American Student Association at NCSU.

When Alejandro is not studying insects, he enjoys travelling with his wife Christie and his son Joshua, camping, watching movies, salsa dancing, and playing table tennis, soccer, and volleyball.

iii

ACKNOWLEDGMENTS

I am indebted to Dr. Dominic Reisig for giving me the opportunity to conduct my doctoral research as a member of his program and helped me achieve the fulfillment of this professional goal of mine. Dr. Reisig was instrumental to generate my research proposal, improve my research and statistical skills, and make me a better scientist. His guidance, willingness to always answer my questions, example of how to work in the field, and constructive criticism helped me finalize each research project that I planned.

I thank my committee members, Dr. Jack Bacheler, Dr. Clyde Sorenson, Dr. George

Kennedy, and Dr. Mark Abney for their guidance, their constructive criticism, and their support to ultimately improve my research. I also want to thank Dan Mott, Steven Roberson, and Clifton Moore for their help while working in the field. Their willingness to share their expertise in field experimentation and eagerness to help make them the best collaborators that I have ever worked with. Additional acknowledgments for every person and funding agency that helped me conducting my research are located at each chapter of this dissertation.

Last but not least, I thank Brad Fritz, Pete Nelson, and Andrew Rodstrom for helping me navigate through school and graduate student life. A special thanks to Christie Almeyda, my lovely wife, who encouraged me to pursue graduate school as a personal dream; and to my son Joshua Del Pozo for teaching me how to be a better human being.

TABLE OF CONTENTS

LIST OF TABLES……………………………………………………………………….. vii

LIST OF FIGURES……………………………………………………………………… x

SECTION ONE

CHAPTER ONE: Introduction…………………………………………………... 1 References………………………………………………………………... 3

SECTION TWO

CHAPTER TWO: First-Generation Megacopta cribraria (Hemiptera: Plataspidae) Can Develop on Soybeans…………………………………………. 5 Abstract…………………………………………………………………... 5 Introduction……………………………………………………………..... 5 Materials and Methods………………………………………………….... 7 Results………………………………………………………………...... 8 Discussion……………………………………………………………...... 9 Acknowledgements……………………………………………………..... 10 References………………………………………………………………... 10

CHAPTER THREE: Diel Flight Activity and Intra-Plant Distribution of Megacopta cribraria (Hemiptera: Plataspidae) Adults in Soybean Abstract…………………………………………………………………... 14 Introduction………………………………………………………………. 15 Materials and Methods…………………………………………………… 18 Results……………………………………………………………………. 22 Discussion………………………………………………………………... 24 Acknowledgments………………………………………………………... 27 References………………………………………………………………... 28

SECTION THREE

CHAPTER FOUR: Megacopta cribraria (Hemiptera: Plataspidae) Population Dynamics in Soybeans as Influenced by Planting Date, Maturity Group, and Insecticide Use…………………………………………….. 36 Abstract…………………………………………………………………... 36 Introduction………………………………………………………………. 37 Materials and Methods…………………………………………………… 40 Results……………………………………………………………………. 48 Discussion………………………………………………………………... 53

Acknowledgements………………………………………………………. 61 References……………………………………………………………...... 61

CHAPTER FIVE: Interaction of Tillage, Maturity Group, and Insecticide Use on Megacopta cribraria (Hemiptera: Plataspidae) Populations in Doublecropped Soybeans………………………………………… 84 Abstract…………………………………………………………………... 84 Introduction………………………………………………………………. 85 Materials and Methods…………………………………………………… 89 Results……………………………………………………………………. 94 Discussion………………………………………………………………... 96 Acknowledgments………………………………………………………... 103 References………………………………………………………………... 104

CHAPTER SIX: Effect of Ground Cover and Row Spacing on Light Reflectance and Megacopta cribraria (Hemiptera: Plataspidae) Populations in Soybean…………………………………………………………... 117 Abstract…………………………………………………………………... 117 Introduction………………………………………………………………. 118 Materials and Methods…………………………………………………… 122 Results……………………………………………………………………. 130 Discussion………………………………………………………………... 136 Acknowledgments………………………………………………………... 144 References………………………………………………………………... 144

SECTION FOUR

CHAPTER SEVEN: Conclusions………………………………………………... 158 References………………………………………………………………... 162

LIST OF TABLES

SECTION TWO

CHAPTER TWO Table 1: Mean ± standard error (SE) number of eggs per mass, mean egg development time (in days ± SE) per each egg mass, minimum and maximum number of eggs per mass laid by Megacopta cribraria female, and mean (± SE) percentage of nymph survival after 24 h eclosion in oviposition chambers…………………...……. 12

Table 2: Mean developmental time (in days ± SE) of Megacopta cribraria reared on caged potted soybean plants under greenhouse conditions (28°C, 60% RH, 14:10 [L:D])….... 13

SECTION THREE

CHAPTER FOUR Table 1: Analysis of variance results for influence of planting date and maturity group on presence of Megacopta cribraria egg masses and adult abundance in soybean plants at V5 growth stage or younger……………………. 68

Table 2: Effect of planting date on significant response variables in South Carolina during 2012 and 2013. Each row in the table represents a separate statistical analysis. Means ± standard error (SE) sharing the same letters are not statistically different (α > 0.05)………………………….. 69

Table 3: Analysis of variance for planting date and maturity group effects on Megacopta cribraria for insecticide treated and untreated 2012 South Carolina trials………… 70

Table 4: Analysis of variance for planting date, maturity group, and insecticide effects on Megacopta cribraria in South Carolina in 2013………………………………………….. 71

Table 5: Analysis of variance results for planting date, maturity group, and insecticide effects on Megacopta cribraria egg mass numbers, cumulative insect days for nymphs and adults per sweep, and cumulative nymphs and adults per plant in North Carolina using combined data

vii

from 2012 and 2013……...………………………………. 72

Tables 6: Significant effect of maturity group (Roman numerals) on significant response variables in North and South Carolina during 2012 and 2013, using Megacopta cribraria and soybean yield data. Each row in the table represents a separate statistical analysis. Means ± standard error (SE) sharing the same letters are not statistically different (α > 0.05)…………………….. …… 74

Table 7: Regression analyses between soybean yield (kg/ha) [response variable] and Megacopta cribraria (cumulative adults and nymphs/sweep) [independent variable] in NC and SC locations for 2012 and 2013. Minimum (Min), maximum (Max), mean, and standard error (SE) values are also presented for each variable…… 75

Supplementary Table 1: Significant effect of planting date and maturity group (Roman numerals) on cumulative insect days for adults/sweep in the insecticide treated trial located in South Carolina and planted during April, May, and July of 2012. Means ± standard error (SE) sharing the same letters are not statistically different (α > 0.05)……………………………………………….... 76

Supplementary Table 2: Significant effect of planting date and insecticide use regime on soybean yield (kg/ha) in South Carolina during 2013. Means ± standard error (SE) sharing the same letters are not statistically different (α > 0.05)……………………………………….. 77

CHAPTER FIVE Table 1: Analysis of variance results for the effects of tillage and maturity group on densities of Megacopta cribraria in soybean plants at V5 growth stage or younger, the effects of tillage, maturity group, and insecticide use on densities of M. cribraria in soybean plants V6 growth stage or older, and the effects of tillage, maturity group, and insecticide use on soybean yield in North Carolina using combined data from 2012 and 2013……………………………………..……... 109

Table 2: Effect of tillage on significant response variables in North Carolina using combined data from 2012 and 2013. Each row in the table represents a separate

viii

statistical analysis………………………………………… 111

Table 3: Effect of maturity group on significant response variables in North Carolina using combined data from 2012 and 2013. Each row in the table represents a separate statistical analysis………………………………. 112

Table 4: Effect of insecticide regime on significant response variables in North Carolina using combined data from 2012 and 2013. Each row in the table represents a separate statistical analysis……………………………... 113

LIST OF FIGURES

SECTION TWO

CHAPTER THREE

Figure 1. Proportion of adults of Megacopta cribraria captured hourly (X-axis) on white sticky cards and by sampling dates (horizontal panels). Cards were placed around experimental fields at the Sandhills Research Station and a commercial soybean field near Gibson, NC, and monitored during June (on the 11th and 13th) and August (on the 6th and 13th) in 2013……………………………… 33

Figure 2. Intra-plant distribution of adult aggregations of Megacopta cribraria in soybean. Visual inspections of plants were conducted once during June at a commercial soybean field near Gibson, NC; and once during August at the Sandhills Research Station. Plant canopy was divided vertically into upper, middle and lower thirds. Adult aggregations were defined as two or more adults grouped together with no clear space separation between/among them. Means sharing the same letter are not statistically significant (α > 0.05)……………………. 34

Figure 3. Scatter plot showing the relationship between plant height (cm) of soybean (X-axis), planted during April using four different maturity group (MG) varieties (labeled “IV”, “V”, “VI”, and “VII” on the vertical panels), and densities of adults per plant (Y-axis) recorded on 20 June 2013 (labeled “June” on the horizontal panels) and on 14 August 2013 (labeled “August” on the horizontal panels). This type of plot was chosen to show all data point, including the outliers…….. 35

SECTION THREE

CHAPTER FOUR

Figure 1. Significant effect of the interaction between planting date (X-axis) and insecticide use (treated plots = black bars, untreated plots = white bars) on Megacopta

cribraria egg masses (panels on the left column), on cumulative insect days for nymphs per sweep (panels on the center column) and cumulative insect days for adults / sweep (panels on the right column) from soybean plants older than V5 growth stage in North and South Carolina during 2012 and 2013. Means sharing the same letters are not statistically different (α > 0.05). Separate analyses are presented at each cell level. Asterisk indicates that mean separation was not performed during the analysis because the interaction between planting date and maturity group (single asterisk), and the effect of maturity group (double asterisk) had an effect on the analyzed variables, rather than the effect of planting date alone. The mean separation for the interaction and the single effect are presented in-text and in Table 4 respectively……………. 78

Figure 2. Significant effects of the interaction between planting date and maturity group on Megacopta cribraria egg masses laid in soybean at V5 growth stage or younger (black bars, upper case letters) and adults (grey bars, lower case letters) in North Carolina during 2012 and 2013 and South Carolina in 2013. Means sharing the same letters are not statistically different (α > 0.05). Separate analyses are presented at each row level. The asterisk at the lower panel indicates that separation of means was not performed for egg masses during the analysis of the interaction because the effect of planting date, rather than maturity group, was significant on M. cribraria egg mass abundance. Mean separations for this effect are presented in Table 2………………………. 80

Figure 3. Significant effects of the interaction between planting date (X-axis) and insecticide use (treated plots = black bars, untreated plots = white bars) on Megacopta cribraria nymphs (upper panel) and adults (lower panel) per soybean plant older than V5 growth stage in North Carolina during 2012 and 2013. Means sharing the same letters are not statistically different (α > 0.05). Separate analyses are presented at each row level……….. 81

Figure 4. Significant effect of the interaction between planting date and maturity group (roman numerals on X-axis) on soybean yield in North and South Carolina during

2012 and 2013. Means sharing the same letters are not statistically different (α > 0.05). Separate analyses are presented at each row level………………………………. 82

Figure 5. Significant effect of the interaction between maturity group (Roman numerals) and insecticide regime (X-axis) on soybean yield in North Carolina during 2012 and 2013 and South Carolina in 2013. Means sharing the same letters are not statistically different (α > 0.05). Separate analyses are presented at each row level………………………………………………………. 83

CHAPTER FIVE

Figure 1. Schematic representation of the experimental layout for each field at the Scotland County location. Individual small cells indicate the position of each experimental plot planted in June under the two tillage conditions with four replications. Larger cells left blank indicate experimental plots not included in this study. In this example, the field on the left was conventionally disc- tilled and the one on the right had the previous crop residue left on top of the soil before planting soybean (reduce-till). Roman numbers indicate the soybean maturity group planted at each plot. Insecticide was aggressively sprayed on plots shaded with grey and untreated controls were located at the white plots…....….. 114

Figure 2. Significant effect of the interaction between insecticide treatment regime (X-axis, treated plots = black bars, untreated plots = white bars) and tillage systems (in each panel) on Megacopta cribraria cumulative insect days for nymphs per sweep from soybean plants older than V5 growth stage in North Carolina, using the combined data from 2012 and 2013. Means sharing the same letters are not statistically different (α > 0.05).…….. 115

Figure 3. Scatter plots representing soybean yield (kg/ha, dependent variable) and tillage system (independent variable) using combined data from North Carolina in 2012 and 2013. Soybean yield was divided in two separate data sets, based on the covariate Megacopta cribraria infestation levels: 1) yields when cumulative M. cribraria were less than 1 bug/sweep or ‘below

xii

threshold’ (crosses), and 2) yields when cumulative M. cribraria were more or equal to 1 bug/sweep or ‘above threshold’ (circles). Each data set was used for an analysis of covariance, where dotted line represent the analysis for the first data set, and solid line represent the analysis for the second data set. Slopes between the two regressions were similar (F = 3.19; P = 0.0751) and intercepts were different (F = 5.14; P = 0.0242)……… 116

CHAPTER SIX

Figure 1. Megacopta cribraria nymphs (black bars) and adults (grey bars) per plant, recorded from soybean planted in narrow (17.78 cm) and wide (96.52 cm) rows during the last three sampling dates (horizontal panels) at Caswell Research Station, during 2014. Error bars were calculated using ± one standard error of the mean. Earlier sampling dates are not included since kudzu bug densities (adults and/or nymphs) did not statistically vary during those dates……………………………………………………... 150

Figure 2. Megacopta cribraria adults per sweep, recorded from soybean planted under three types of tillage (conventionally tilled, reduced-till, and rolled-rye, vertical panels), using two row spacings (x-axis, narrow - 38.1 cm and wide - 76.2 cm), and at two experimental locations (Sandhills and Caswell Research Stations, horizontal panels) after canopy closure during 2015. Error bars were calculated using ± one standard error of the mean. Means sharing the same letter are not statistically significant (α > 0.05)………………………………………………… 151

Figure 3. Percentage of ground cover by previous crop residue (cereal rye stubble and straw) in (a) four treatments including, conventional tillage, reduced tillage, rye mowed on top of plots, and rye rolled on top of plots at Caswell Research Station, Kinston, NC during 2014; and (b) three treatments including, conventional tillage, reduced tillage, and rye rolled on top of plots at two locations (Caswell Research Station, Kinston, and Sandhills Research Station, Jackson Springs, NC) during 2015. Measurements were directly following soybean planting. Means sharing the same letter are not statistically different (α > 0.05)………………………….. 152

xiii

Figure 4. Relationship between Megacopta cribraria adults per sweep (expressed as log10-base) and percentage of ground cover in soybean prior canopy closure at the Caswell Research Station, 2014 (a) (P = 0.0336; R2 = 0.23; ground cover = -0.0005 × log10 (adults/sweep) + 0.076), and at the Sandhills Research Station, 2015 (b) 2 (P = 0.0269; R = 0.10; ground cover = -0.0005 × log10 (adults/sweep) + 0.038). During 2015, there were no kudzu bug adults recorded at the Caswell Research Station prior canopy closure……………………………... 153

Figure 5. Average light reflectance calculated using measurements taken on 11 June 2014 from inside (a) and outside (b) of experimental soybean plots including, conventionally tilled (×), under reduced-till (squares), with mowed rye (triangles), and with rolled rye (upside-down triangles) at Caswell Research Station, Kinston, NC. Measurements from inside plots recorded reflectance coming from both small plants (before canopy closure) and ground cover; and from the outside, reflectance coming from the ground cover only with no plants. Average reflectance calculated for kudzu, Pueraria montana, (circles) was taken 16 June 2014 on North Carolina State University campus, Raleigh, NC………………………………………………. 154

Figure 6. Scatter plots showing the relationship between selected spectral indexes (Y-axis) and percentage of ground cover (X-axis), under four different tillage types (horizontal panels; circle=conventionally tilled, plus=reduced tillage, cross=mowed rye, and triangle=rolled rye) at the Caswell Research Station, Kinston, NC, during 2009. These indexes were calculated from light reflectance collected prior soybean canopy closure, and taken from the outside of experimental plots (separation alley with tillage treatment and no soybean plants). Mean separations are not shown in this figure, but they are described in the result section……………………………………………... 155

Figure 7. Contour plots showing the variation of Megacopta cribraria adults (expressed as log10-based and fitted as isolines), based on the Photochemical Reflectance Index (X-axis, expressed as log10-base) calculated from inside (tillage treatment + soybean plants, left) and from the outside experimental plots (separation alley with tillage treatment and no soybean plants, right), and the percent

xiv

of ground cover (Y-axis). Red and blue colors indicate high and low values of M. cribraria adults, respectively. Light reflectance was taken prior to soybean canopy closure at the Caswell Research Station, Kinston, NC, during 2014………………………………………………. 156

Figure 8. Contour plots showing the variation of Megacopta cribraria adults (expressed as log10-based and fitted as isolines), based on three spectral indexes (X-axis, expressed as log10-base): (a) Normalized Difference Vegetation Index, (b) Damage Sensitive Spectral Index 1, and (c) Aphid Index, calculated from the outside of experimental plots (separation alley with tillage treatment and no soybean plants) and the percent of ground cover (Y-axis). Red and blue colors indicate high and low values of M. cribraria adults, respectively. Light reflectance was taken prior to soybean canopy closure at the Caswell Research Station, Kinston, NC, during 2014………………………………………………. 157

SECTION FOUR

CHAPTER SEVEN Figure 1. Flow chart summarizing findings from this dissertation. Rectangular boxes indicate a production practice or a process during the growing season. Diamonds represent decisions preceded by a question. Soybean fields can be characterized as having higher Megacopta cribraria (KB) infestation risk (black box) or lower KB infestation risk (grey boxes). The action threshold (*) for M. cribraria is one nymph/sweep, based on Seiter et al. (2015)…………………………………………………….. 164

SECTION ONE

CHAPTER ONE

Introduction

Soybean, Glycine max (L.) Merr., is an annual herbaceous plant that belongs to the

Fabaceae family, and has the ability to fix atmospheric nitrogen in association with nodule- forming bacteria from the genus Rhizobium (Singh 2010). Originally from Southeast Asia, soybean was domesticated by Chinese farmers during 1,100 BC, and brought to the U.S. during the late 1700’s from Europe (Martin et al. 2002, Singh 2010). Currently, soybean is one of the most economically important crops in the U.S. (Heatherly and Hodges 1999,

Martin et al. 2002, Singh 2010). It is the second most planted crop after corn, Zea mays L., with 33.9 million hectares planted in the U.S. during 2014 (SoyStats 2016). Soybean farm gate value for the U.S. has been annually estimated around $40 billion since 2012 (SoyStats

2016); soybean seed is processed mainly for oil and animal feed (Singh 2010).

Several factors negatively impact soybean yield, including an incorrect selection of the planting date (Heatherly and Hodges 1999, Martin et al. 2002), or insect pest outbreaks

(Heatherly and Hodges 1999, Musser et al. 2015). Corn earworm, Helicoverpa zea (Boddie)

(Lepidoptera: Noctuidae) and stink bugs (Hemiptera: Pentatomidae) are considered the most devastating insect pests in southern U.S. soybeans, based on yield losses and control cost associated with these insects (Musser et al. 2014, Musser et al. 2015).

After becoming established in Georgia in 2009, the invasive kudzu bug, Megacopta cribraria (F.) (Hemiptera: Plataspidae) rapidly became an economically important pest of

U.S. soybean. From a survey of seven states, the area of soybean sprayed to control kudzu

bug increased 6.0 fold from 2012 to 2013 across AL, MS, NC, TN and VA area combined, except for AR and LA (Musser et al. 2014). The kudzu bug-infested area increased almost

1.6 fold in 2014, from 233,434 to 359,834 ha across these seven states; however, the insecticide treated area only increased 1.3 fold from 2013 to 2014 based on surveyed fields and Southeast entomologist’ estimates (Musser et al. 2015). Recent research on kudzu bug has aided our understanding of the damage potential of this insect in soybean. For example, this insect is been reported to reduce soybean yield up to 60% under field-cage conditions

(Seiter et al. 2013). Currently, whole-field insecticide application is used to control kudzu bug in soybean. In order to reduce the number of insecticide applications used to manage kudzu bug, a collaborative research was undertaken to determine an effective action threshold. This effort discovered that an action threshold of one nymph per sweep, was the most cost-effective action threshold tested and required only a single insecticide application to prevent soybean yield loss (Seiter et al. 2015).

The rationale behind the research presented in this dissertation investigates the hypothesis that cultural control is a key component of a kudzu bug integrated pest management, as was demonstrated by Kogan (1998) in a Midwest soybean system. As a newly introduced species, kudzu bug represents a threat to soybean production. The main goal of this research was to generate information about this southeast U.S. invader. Research questions pivoted around the main hypothesis that current soybean production practices create a kudzu bug ‘friendly’ environment. This research was specifically designed to better understand the biology and life history of kudzu bug in soybean, analyze agronomical practices that are favorable and/or detrimental to this insect infestations, and to propose modifications of cultural practices that favor the establishment of this pest in soybean.

This dissertation is divided in four sections. Each section includes a chapter or a group of chapters with a common research topic. This chapter, the introduction, constitutes section one, and explains the rationale behind this research and its organization. Section two includes chapters two and three, addressing biology and life history of the kudzu bug. Section three includes chapters four, five, and six, presenting information on how production practices, such as planting date or tillage, influence kudzu bug populations in soybean.

Section four includes only chapter seven with the overall conclusions of this research.

The rationale behind section two, “Biology and Life History of the Kudzu Bug in

Soybean”, was that there was limited information on the life cycle and in-field activity of this insect. I expected that a better understanding of the life history of the kudzu bug would improve scouting procedures and management practices. The third section, “Influence of

Soybean Production Practices on Kudzu Bug”, addresses the fact that whole-field applications of broad-spectrum insecticides are the currently accepted management tactic for kudzu bug. Relying solely on insecticides has negative environmental and resistance management consequences, and is not a sustainable practice. Alternatively, the manipulation of production practices may negatively impact pest populations, including kudzu bug. The objectives of the research from this section were to identify which practices favor the establishment of this pest, to direct scouting efforts, and to identify specific production practices that impact kudzu bug levels.

References

Heatherly L. G. and H. F. Hodges. 1999. Soybean production in the midsouth. CRC Press

LLC. Boca Raton, FL.

Kogan, M. 1998. Integrated pest management: historical perspectives and contemporary

development. Annu. Rev. Entomol. 43: 243-270.

Martin, J. H., R. P. Waldren, and D. L. Stamp. 2002. Principles of field crop production.

Person Prentice Hall Inc. Upper Saddle River, NJ.

Musser, F. R., A. L. Catchot, Jr., J. A. Davis, D. A. Herbert, Jr., G. M. Lorenz, T. Reed,

D. D. Reisig, and S. D. Stewart. 2014. 2013 Soybean insect losses in the Southern

US. Midsouth Entomol. 7: 15-28.

Musser, F. R., A. L. Catchot, Jr., J. A. Davis, D. A. Herbert, Jr., G. M. Lorenz, T. Reed,

D. D. Reisig, and S. D. Stewart. 2015. 2014 Soybean insect losses in the Southern

US. Midsouth Entomol. 8: 35-48.

Seiter, N. J., J. K. Greene, and F.P.F. Reay-Jones. 2013. Reduction of soybean yield

components by Megacopta cribraria (Hemiptera: Plataspidae). J. Econ. Entomol.

106: 1676-1683.

Seiter, N. J., A. I. Del Pozo-Valdivia, J. K. Greene, F.P.F. Reay-Jones, P. M. Roberts,

and D. D. Reisig. 2015. Action thresholds for managing Megacopta cribraria

(Hemiptera: Plataspidae) in soybean based on sweep-net sampling. J. Econ. Entomol.

108: 1818-1829.

Singh, G. 2010. The soybean: botany, production, and uses. CABI Editors. Wallingford, UK.

SoyStats. 2016. Planting data and value. The American Soybean Association.

http://www.soystats.com (accessed 18 April 2016).

SECTION TWO

CHAPTER TWO1

First-Generation Megacopta cribraria (Hemiptera: Plataspidae)

Can Develop on Soybeans

1Published in Journal of Economic Entomology, 2013, Volume 106, pages 533 - 535

Abstract

Megacopta cribraria (F.) (Hemiptera: Plataspidae) was first reported in 2009 near

Atlanta, Georgia. The insect undergoes two generations per year. The first-generation is reported mainly in kudzu during May and June, with the second establishing on both kudzu and soybean during July and August. A greenhouse study was conducted to determine the suitability of two legumes as hosts for first generation M. cribraria. First generation M. cribraria successfully developed on caged potted soybean plants. Conversely, snap beans were not a suitable host under the conditions of this study. A range of 45 to 50 days was needed to transition from the egg to adult on soybean plants (28°C, 60% RH). Although this study was limited to the greenhouse, kudzu may not be an obligate host for the development of first-generation M. cribraria. An important implication of this finding is the establishment for this pest on spring-planted soybean and for the possible expanded geographic range for this pest beyond that of kudzu.

Introduction

Megacopta cribraria (F.) (Hemiptera: Plataspidae) is an exotic species in the United

States, first reported near Atlanta, Georgia in 2009 (Suiter et al. 2010). It is a piercing-

sucking herbivore feeding on legumes and undergoing two generations per year in the southeastern U.S. (Eger et al. 2010, Suiter et al. 2010, Zhang et al. 2012). In the Southeast, overwintering M. cribraria adults become active when temperatures warm in the spring; the insects disperse to kudzu [Pueraria montana Loureiro (Merrill) variety lobata (Willdenow),

Fabaceae] where they aggregate, feed, mate and lay eggs. The first-generation nymphs then develop on kudzu during May and June (Zhang et al. 2012). Adults from the first generation that originate in kudzu are known to invade soybeans [Glycine max Merrill, Fabaceae] in

July where they subsequently deposit egg masses (Greene 2010).

Megacopta cribraria has spread rapidly and in relatively high densities, expanding its range across the Southeast and into the Midsouth. Soybeans represent an important commodity crop in the U.S. (USDA 2007), and previous experience managing high abundances of M. cribraria demonstrate that this insect can become a serious economic challenge to profitable soybean production (Zhang et al. 2012), possibly requiring multiple insecticide applications for successful management.

Megacopta cribraria will oviposit on a variety of hosts, although nymphs will only develop into adults on soybeans and kudzu according to Zhang et al. (2012). Zhang et al.

(2012) also showed that M. cribraria oviposition, survival and development was much greater on kudzu than soybeans. Of the eggs oviposited on soybean, only 4% developed into adults. Field observations indicate that soybeans are excellent hosts for second-generation

M. cribraria (J. Greene and P. Roberts, personal comm.). Because of the limited information on this species’ development on soybean and other host plants, we conducted a greenhouse study to determine if soybean and snap bean [Phaseolus vulgaris L., Fabaceae] plants were

suitable hosts for the first-generation of M. cribraria and to determine stadia durations for the five reported instars.

Materials and Methods

Colonies were established at North Carolina State University, Raleigh, North

Carolina, from overwintering M. cribraria adults collected from kudzu 20 km northwest of

Lincolnton, NC (35°34'23.13"N, 81°22'27.20"W) on 26 March 2012. Fifteen pairs (male and female) were placed into 478 ml plastic containers (15 cm diameter, 10 cm height), which served as oviposition chambers, where they were provided with fresh kudzu stems and leaves daily. These containers were housed in a room where temperature, relative humidity (RH), and photoperiod were controlled (28 ± 1°C, 55% RH, 14:10 [Light:Dark]). Individual egg masses were collected and placed into Petri (6 cm diameter) dishes from early-April to early-

May. The number of eggs per mass and first instar emergence were recorded from six egg masses per oviposition chamber. After egg hatching, two methods were used to rear nymphs:

1) placing them inside plastic containers (50 ml, 8 cm diameter, 6 cm height) in the growth chamber and 2) placing them on potted plants in a greenhouse (28°C, 60% RH, 14:10 [L:D]) using soybeans (var. ‘AG64730’, Monsanto, St. Louis, MO) and snap bean (var. ‘Caprice’,

Harris Seeds, Rochester, NY). For the first rearing method, we used 16 containers under the same room conditions. Half of the containers were provisioned daily with soybean leaves and the other eight with snap bean leaves. Eight soybean and eight snap bean plants were used to carry out the second portion of this experiment in a greenhouse. First instar nymphs emerging from the same egg mass were placed inside containers and on plants using a fine brush 24 h after eclosion. Each group of nymphs was randomly assigned to containers and

plants for both types of rearing methods. In the greenhouse, plants were caged with 40 × 30 cm plastic cylinders made from clear 3 mm thick clear acrylic sheet (Plexiglass, Blue Hawk,

Wilmington, DE). One end of the cylinder was sealed with a fine cloth and the other side was placed over the top of the potted plants. To maintain host quality, potted plants were replaced every two weeks by moving the insects to the younger plant with at least two true leaves using a fine brush. Stadium was recorded for each rearing method, where overall size of individuals was measured following Zhang et al. (2012) parameters for instar determination.

Results

First-generation of M. cribraria were reared on soybean plants under greenhouse conditions (Table 1 and 2). Females deposited egg masses in clutches with approximately 18 eggs per mass, ranging from 5 to 38 eggs per mass (Table 1). Neonates eclosed as a single batch from the same egg mass. After hatching, a high percentage of nymphs survived (98%) after 24 h inside the small petri dishes (Table 1). Most nymphs also survived (90%) after artificial infestation on potted soybean plants (Table 2).

Under greenhouse conditions, M. cribraria developed from egg to adult in 45 to 50 days (Table 2). No M. cribraria first instar nymphs molted to the second instar when caged on snap beans (data not presented). All first instar nymphs placed inside plastic containers in the growth chamber and on potted snap bean plants in the greenhouse died within 72 h of artificial infestation.

Discussion

On soybean, first-generation M. cribraria transitioned from egg to adult in 45-50 days. In Asia, M. cribraria completes its life cycle in 49 to 90 days (Eger et al. 2010, Zhang et al. 2012). Zhang et al. (2012) provided information about the life history of M. cribraria in the United States, using forest legumes as hosts. They also pointed out that M. cribraria survived poorly on soybean compared to kudzu. In contrast to their study, we were able to rear first-generation M. cribraria with a high survival rate (90%) on potted soybeans under greenhouse conditions. In our study, snap beans were not a suitable host for M. cribraria, confirming previous work showing that not all legumes are suitable hosts (Zhang et al. 2012).

Future studies should explore even more legumes as possible feeding and developmental hosts for this insect.

Rearing attempts using both soybean and snap bean material inside plastic containers were unsuccessful. It is possible that factors such as the confined environment or condensation may have contributed to mortality. Furthermore, host quality may be an important factor in M. cribraria development. In the greenhouse study, plants were replaced every two weeks to maintain host quality. Perhaps the lack of development on soybeans observed in Zhang et al.’s (2012) study was due to decline in host quality, as well as cage effects. In nature, it is conceivable that M. cribraria can disperse to better quality hosts when conspecific density is high, as observed in other sap-feeders such as aphids and leafhoppers

(Awmack and Leather 2002).

Although this study was restricted to a no-choice greenhouse assay, it represents the first demonstrated instance of the development of first-generation M. cribraria on U.S. soybeans. If first-generation M. cribraria can develop on soybean in the field, population

buildups could result in larger second-generation populations on this crop. Furthermore, because kudzu is not an obligate host for first-generation M. cribraria, this pest has the potential to spread beyond the distribution of the kudzu into areas intensively planted to soybeans, such as the Midwest.

Acknowledgements

The authors would like to thank Dr. Mark Abney for providing access to rearing and greenhouse facilities, Dr. Jack Bacheler, and two anonymous reviewers for revising this manuscript. This research was supported by grants from the United Soybean Board and the

North Carolina Soybean Producers Association.

References

Awmack, C. S. and S. R. Leather. 2002. Host Plant Quality and Fecundity in Herbivorous

Insects. Annu. Rev. Entomol. 47: 817-844.

Eger, J. E., Jr., L. M. Ames, D. R. Suiter, T. M. Jenkins, D. A. Rider, and S. E. Halbert.

2010. Occurrence of the Old World bug Megacopta cribraria (Fabricius)

(Heteroptera: Plataspidae) in Georgia: a serious home invader and potential legume

pest. Insecta Mundi 0121: 1-11.

Greene, J. K. 2010. Cotton /Soybean Insect Newsletter. Volume 5, Issue #7.

http://www.clemson.edu/public/rec/edisto/pdf/newsletseven.pdf

Suiter, D. R., J. E. Eger, Jr., W. A. Gardner, R. C. Kemerait, J. N. All, P. M Roberts, J.

K. Greene, L. M. Ames, G. D. Buntin, T. M. Jenkins, and G. K. Douce. 2010.

Discovery and distribution of Megacopta cribraria (Hemiptera: Plataspidae) in

northeast Georgia. J. Integr. Pest Manag. 1: 1-4.

United States Department of Agriculture. 2007. Agricultural census.

http://www.agcensus.usda.gov/Publications/2007/Online_Highlights/Rankings_of_M

arket_Value/North_Carolina/

Zhang, Y., J. L. Hanula, and S. Horn. 2012. The biology and preliminary range of

Megacopta cribraria (Heteroptera: Plataspidae) and its impact on kudzu growth.

Environ. Entomol. 41: 40-50.

Table 1. Mean ± standard error (SE) number of eggs per mass, mean egg development time

(in days ± SE) per each egg mass, minimum and maximum number of eggs per mass laid by

Megacopta cribraria female, and mean (± SE) percentage of nymph survival after 24 h eclosion in oviposition chambers (28°C, 60% RH, 14:10 [L:D]).

Mean Percentage of Minimum Maximum Oviposition number of Length of egg nymph survival number of number of chamber eggs per development after 24 h eggs per mass eggs per mass mass eclosiona 1 16.00 ± 2.14 6 21 8.50 ± 0.34 99 ± 0.01 2 13.17 ± 3.42 8 30 7.83 ± 0.31 99 ± 0.01 3 19.00 ± 2.85 8 28 8.17 ± 0.31 97 ± 0.02 4 17.83 ± 0.31 17 19 8.00 ± 0.00 100 ± 0.00 5 24.33 ± 3.89 17 38 8.33 ± 0.21 95 ± 0.01 6 22.83 ± 3.60 8 30 8.50 ± 0.34 96 ± 0.00 7 11.83 ± 1.14 8 15 7.50 ± 0.22 99 ± 0.01 8 18.17 ± 3.02 6 28 8.00 ± 0.26 98 ± 0.02 9 16.50 ± 3.43 5 25 8.00 ± 0.37 98 ± 0.01 10 11.17 ± 1.62 7 17 7.33 ± 0.21 99 ± 0.01 11 30.00 ± 0.86 28 32 9.00 ± 0.00 97 ± 0.01 12 24.50 ± 1.20 22 28 8.50 ± 0.22 93 ± 0.03 13 17.00 ± 0.63 14 18 8.50 ± 0.00 97 ± 0.02 14 14.17 ± 0.95 12 18 7.67 ± 0.21 98 ± 0.02 15 17.83 ± 0.95 15 22 8.17 ± 0.17 98 ± 0.02 18.29 ± 1.34 12.07 ± 1.73 24.60 ± 1.71 8.10 ± 0.11 98 ± 0.00 aEclosion was considered complete when at least one neonate was observed.

Table 2. Mean developmental time (in days ± SE) of Megacopta cribraria reared on caged potted soybean plants under greenhouse conditions (28°C, 60% RH, 14:10 [L:D]).

M. cribraria stadia Number Number Caged of eggs Length of I II III IV V of final plant per egg egg stage insects mass 1 11 7.00 ± 0.00 7.64 ± 0.15 9.82 ± 0.12 7.82 ± 0.12 8.64 ± 0.24 6.18 ± 0.12 11 2 18 8.00 ± 0.00 7.44 ± 0.12 7.44 ± 0.12 8.44 ± 0.12 8.56 ± 0.12 6.56 ± 0.12 17 3 12 7.00 ± 0.00 7.33 ± 0.14 9.67 ± 0.14 9.67 ± 0.14 6.67 ± 0.28 7.27 ± 0.13 11 4 15 8.00 ± 0.00 7.53 ± 0.13 7.57 ± 0.14 7.79 ± 0.11 7.86 ± 0.28 6.93 ± 0.20 14 5 18 9.00 ± 0.00 8.53 ± 0.12 7.13 ± 0.26 7.44 ± 0.13 8.44 ± 0.13 6.44 ± 0.13 16 6 13 8.00 ± 0.00 8.38 ± 0.27 9.69 ± 0.13 8.38 ± 0.27 7.75 ± 0.13 6.75 ± 0.24 12 7 15 9.00 ± 0.00 8.33 ± 0.25 9.23 ± 0.28 7.62 ± 0.14 7.33 ± 0.14 6.33 ± 0.14 12 8 12 8.00 ± 0.00 7.45 ± 0.16 7.55 ± 0.16 8.55 ± 0.16 7.50 ± 0.16 6.30 ± 0.15 10 8.00 ± 0.27 7.83 ± 0.18 8.51 ± 0.42 8.21 ± 0.25 7.84 ± 0.24 6.60 ± 0.13

CHAPTER THREE

Diel Flight Activity and Intra-Plant Distribution of Megacopta cribraria

(Hemiptera: Plataspidae) Adults in Soybean

Abstract

Megacopta cribraria (F.), the kudzu bug, is a soybean pest in the southeastern U.S.

Accidentally introduced into Georgia in 2009 from Asia, kudzu bug can reduce up to 60% of soybean yield when left uncontrolled. There is limited information on the life history of this invasive pest in soybean. The main goals of this research were to investigate the daily flight activity pattern and intra-plant distribution of kudzu bug adults in soybean. This was accomplished through experiments in two locations in NC during 2013 in which dispersing adult kudzu bugs captured hourly on white sticky cards between 09:00 – 17:00 hours were counted hourly, and adults on plants were visually sampled between 09:00 – 12:00 hours from soybean maturity group IV to VII plants. Adult captures on sticky cards were higher from 13:00 to 15:00 hours across sampling dates, suggesting that dispersal or flight activity peak during this interval. When soybean plants were visually inspected, most of the adults formed aggregations on the main stem, with aggregations most common in the middle section of plants. The number of aggregations per plant, the number of adults per plant, and the male-to-female ratio were not influenced by maturity group. Soybean plant height did not affect adult densities per plant. However, densities varied depending on the date of sampling.

Implications of this research on kudzu bug biology are discussed.

Introduction

The kudzu bug, Megacopta cribraria (F.) (Hemiptera: Plataspidae), is an invasive pest of soybean, Glycine max (L.) Merr., in the U.S. Originally from Southeast Asia (Suiter et al. 2010, Hosokawa et al. 2014), this insect was first reported in nine counties of GA during 2009 (Suiter et al. 2010), and is currently widespread across most states in the

Southeastern U.S. (Gardner 2016). It feeds and reproduces on legumes; preferred hosts include kudzu, Pueraria montana Loureiro (Merr.) variety lobata (Willdenow), Wisteria spp., and some bean species from the genera Vicia and Cajanus (Zhang et al. 2012, Medal et al. 2013, Seiter et al. 2014, Blount et al. 2015, Golec et al. 2015).

Kudzu bug is bivoltine in the southeastern U.S. (Zhang et al. 2012, Seiter et al. 2013).

Adults disperse to kudzu or early-planted soybean from overwintering sites during April, where they feed, mate, and then lay eggs (Zhang et al. 2012, Seiter et al. 2013, Del Pozo-

Valdivia et al. 2016); however, some F0 females, that mated in the previous fall, can lay eggs directly following departure from overwintering sites (Golec and Hu 2015). In the southeastern U.S., the first generation (F1) nymphs are observed on soybean during May and

June, and adults of the first in-field generation are seen on soybean during July and August

(Zhang et al. 2012, Seiter et al. 2013, Del Pozo-Valdivia et al. 2016). A second peak of nymphs and adults can be observed during September and October in kudzu and soybean, respectively (Zhang et al. 2012, Seiter et al. 2013).

Scouting for kudzu bug in soybean is critical for implementing the suggested insecticide action threshold when managing this pest (Seiter et al. 2015). The spatial distribution of the kudzu bug within soybean fields may pose a challenge for accurate

sampling populations of this insect. Aggregations of kudzu bug are most commonly found near edges of soybean fields (Seiter et al. 2013). Hence, sampling plans should incorporate this information to avoid sampling bias. Other unknown factors, such as diel flight activity and intra-plant distribution of this insect, might also influence the sampling efficiency.

It is well-documented that some hemipterans have aggregated intra-plant distributions, reflecting their preference for certain tissues on their host plants. In alfalfa,

Medicago sativa L., Acyrthosiphon kondoi Shinji and Kondo (Hemiptera: Aphididae) is mostly found in stems and leaves along the mid-portion of the canopy (Zarrabi et al. 2005).

The spatial distribution can change as the hemipteran or plant develops. For instance, adults of the tarnish plant bug, Lygus lineolaris (Palisot de Beauvois) (Hemiptera: Miridae) are mostly found on vegetative structures (mainly leaves) of cotton, Gossypium hirsutum L., compared with nymphs that are found in fruiting structures (squares, bolls, and blooms)

(Snodgrass 1998). On the other hand, apterous and alate aphids, Aphis gossypii Glover

(Hemiptera: Aphididae) are uniformly distributed over the cotton plant during the early growth stages, with apterous aphids more common in the bottom and middle sections of the cotton plant as cotton develops (Fernandes et al. 2012).

In soybean, there are some records of how insect densities vary within the plant canopy. For example, the soybean aphid, Aphis glycines Matsumura (Hemiptera: Aphididae), has higher growth rates at the upper nodes of plants, compared with the lower section of the plant (McCornack et al. 2008, Costamagna 2013). Stink bugs (Hemiptera: Pentatomidae) are commonly found in the upper portion of plants. However, when density-dependent effects are exerted, high infestation levels of conspecifics force the bugs to feed in the lower portion of plants (Russin et al. 1987). This has been corroborated in southeastern U.S. soybean,

where approximately 80% of green, Chivania hilaris Say, and brown, Euchistus spp., stink bugs are found within the upper portion of the soybean canopy in Virginia (Owens et al.

2014). Additionally, another southeastern U.S. non-hemipteran pest, the corn earworm,

Helicoverpa zea (Boddie) (Lepidoptera: Noctuidae), prefers to oviposit on terminals of trifoliates that are located in the lower and upper sections of the soybean canopy (Terry et al.

1987). The hypothesis is that a closed canopy, especially at the middle section of the plant, may inhibit moth flight or obscure visual cues, resulting in fewer eggs at the middle of the plant (Terry et al. 1987).

Diel flight activity also varies among hemipterans in the field, even among species in the same family. For example, the mealy bugs Planococcus citri Risso and Planococcus ficus

Signoret (Hemiptera: Coccidae) are more active fliers during early in the morning ~07:00 hours (Levi-Zada et al. 2014). Furthermore, the rugose spiraling whitefly, Aleurodicus rugioperculatus Martin (Hemiptera: Aleyrodidae) is more active on ornamental greenhouse plants between 06:00 to 10:00 hours (Taravati et al. 2014). In contrast, more sweetpotato whiteflies, Bemisia tabaci (Gennadius) are caught towards midday in experimental field plots of cotton, watermelons, and cantaloupe (Bellows et al. 1988). Other hemipterans are generally active during later hours of the day. Adults of the stink bug Bagrada hilaris

(Burmeister) are more active and abundant during the afternoon, specifically between 13:00 and 18:00 hours, in desert cole crops (Huang et al. 2013). Finally, other stink bugs, such as

Euchistus conspersus Uhler, are highly active during the night, with a peak of abundance and mating ~21:00 hours (Krupke et al. 2006).

Literature on the life history of kudzu bug is still incomplete. Since its introduction, no information about diurnal activities of this insect in soybean fields has been reported. The

objectives of this study were to identify periods when during the day adult flight activity peaked in soybean fields, and to investigate the distribution of adults within soybean plants of different maturity groups (MG) during the morning. Maturity group is defined as the categorization of the number of days until soybean flowers under specific photoperiod conditions (Pedersen 2009). Hence, plant phenology was modified in these experiments, since different MGs within a given experiment were planted the same date. I hypothesized that adult M. cribraria flight activity would differ at specific times of the day in soybean fields and that intra-plant distribution may be influenced by MG. Additionally, I was interested in testing if plant height, which I expected to change through time and among

MGs, may also influence kudzu bug adult densities. Ultimately, understanding flight activity and intra-plant distribution of adults might lead to improved scouting procedures and optimized insecticide application timing.

Materials and Methods

Plot Information. Field sites were located in the North Carolina State University

Sandhills Research Station, near Jackson Springs in Montgomery County, NC,

(35°11'06.5"N 79°40'15.8"W) and in a commercial soybean field near Gibson, Scotland

County, NC, (34°44'42.8"N 79°35'15.8"W) during 2013. The east (Sandhills) and south

(commercial field) sides of experimental sites were ~40 m away from the wood line, composed mainly of broad-leaf tree species. Experiments were set up as a randomized complete block design in conventionally disc-tilled fields, with four replications and a single factor, maturity group (MG). For both locations, soybean seed was Roundup Ready (Asgrow,

Monsanto Company, St. Louis, MO) and MGs were IV (variety AG4531), V (AG5503), VI

(AG6132), and VII (AG7502). Experimental plots were four-rows wide and 12.2 m long at both locations. At the Sandhills Research Station, soybean was planted on 16 April 2013, with 0.97 m row-spacing using a four-row cone planter (John Deere model 1750, Deere and

Co., Moline, IL) and a seeding rate of 29 seeds/row meter. At the commercial field near

Gibson, soybean was planted on 18 April 2013, with 0.91 m row-spacing using a two-row disc planter (White model 6700, AGCO Corporation, Duluth, GA) and the same seeding rate

(29 seeds/row meter).

Adult Flight Activity Monitoring. Two 20 × 25 cm white sticky traps (replacement liner for Pherocon IV trap, Trécé Inc., Adair, OK) were fastened at 0.75 m and another at 1.5 m above the ground level, respectively to 4 cm diameter wooden poles; one located at each corner of the two fields. Two additional poles were placed in the middle of the northern and southern side of each field. The sticky surface of each trap faced into the experimental field.

Trap heights were selected to approximate mid-canopy and above the maximum canopy height of mature plants (1.0 m). White sticky cards were used because more kudzu bugs are captured on light-colored sticky cards (Horn and Hanula 2011). The kudzu bug adults were counted and removed from each sticky card hourly from 09:00 to 17:00 hours. This experiment was conducted on 11 June 2013 and 6 August 2013 at the Sandhills Research

Station; and on 13 June 2013 and 13 August 2013 at the commercial field. New sticky cards were placed on each date and at each location before beginning adult monitoring, and cards were removed after 17:00 hours.

Adult Intra-Plant Distribution. Whole-plant visual inspections were conducted to record the number of adult aggregations per plant, number of adults per aggregation, number of adults per plant, plant height (cm), location of aggregations within plant’s canopy, and the

number of males and females in the middle section of soybean plants. An adult aggregation was defined as two or more adults adjacent to each other with no clear separation between/among them. Based on preliminary observations indicating that adults were on the soybean main stem during the morning, visual inspections were performed from 09:00 to

12:00 hours. Three plants were randomly selected from the middle section of the third row in each experimental plot. Using a 1 m ruler, plant height was recorded from the soil surface at the base of the plant to the apex of the main stem. After measuring plant height, the canopy was divided into thirds, representing the upper, middle and lower sections of the canopy. The number of aggregations/plant within each section was recorded as were the number of adults/aggregation and the total number of adults/plant. To document the number of males and females on soybean plants, a sub-sample of the counted adults was collected from the middle section of the plant and placed in a 50 ml clear-plastic centrifuge tube with a lid

(Corning Inc., Corning, NY). Adults were sexed in the field while they were contained inside the centrifuge tube, using morphological characters on the last abdominal sternite (Zhang et al. 2012). Adult monitoring in soybean that came from overwintering using sticky cards and plant visual inspections were performed during June. Monitoring adults from the beginning of the first in-field generation in soybean was conducted during August. Plant inspections were conducted on 20 June 2013 (commercial field only) and on 14 August 2013 (Sandhills

Research Station only).

Data Analysis. Numbers of insects captured on sticky cards were pooled across locations (n=2) and sampling dates (n=2) to form a single data set. Adult captures were averaged between height of the sticky cards on each pole (n=2) to provide adults/card for each position of the pole (n=6). Adults/card were log10-transformed [log10(X+1)] to comply

with the assumptions of the analysis of variance (ANOVA). This data set was analyzed using a mixed model ANOVA (PROC MIXED, SAS version 9.3, SAS Institute 2010, Cary, NC).

The response variable for this analysis was log10-transformed adults/card; the fixed effects were sampling date (n=4), time of the day (n=8), and their interactions. Random effects were location and position of the pole nested with location. A repeated statement was included in the previous analysis to account for the effect of time of day, where the subject was calculated based on the multiplication of date, location, and position of the pole. The covariance structure was selected as compound symmetry, since sticky card evaluations were performed at each hour. To further analyze the interaction between sampling date and time of the day, the option SLICE was included in the LSMEANS statement to test the effect of time of day for each level of sampling date.

For the second part of this study, response variables included number of aggregations found in each plant section (n=3), aggregations per plant, adults per aggregation, male-to- female ratio, adults per plant, and plant height. These were analyzed in separate ANOVAs using PROC MIXED (SAS Institute 2010). All response variables except male-to-female ratio and plant height, were log10-transformed [log10(X+1)] to comply with the assumptions of the ANOVA. In our first test, the response variable was log10-transformed number of aggregations; and the fixed effects were sampling date (n=2), MG (n=4), plant section (n=3) nested within plant (n=3), plant sampled per plot (n=3), and their interactions. For the second, third, and fourth tests, the response variables were log10-transformed number of aggregations, log10-transformed number of adults per aggregation, and male-to-female ratio, respectively; and sampling date (n=2), MG, and their interaction were fixed effects. For the last test, log10-transformed number of adults was the response variable; and sampling date,

MG, plant height nested with MG, and their interactions were fixed effects. Replication nested with location was the only random effect included in all ANOVA tests.

Degrees of freedom for all ANOVA tests were calculated using the procedure of

Kenward and Roger (1997). When a two-way interaction was significant, the option SLICE was included in the LSMEANS statement to test the effect of MG at each sampling date. Post hoc mean separation of the transformed data, male-to-female ratio, and plant height was performed using the Tukey’s test at α ≤ 0.05. Means and standard errors are reported from back-transformed data.

Results

Adult Flight Activity. A total of 132 kudzu bug adults were captured on white sticky cards at the two fields. Adult captures were influenced by the interaction of sampling date and time of the day (F = 4.72; df = 21, 140; P < 0.0001). When controlling for the sampling date effect, adult captures on sticky cards were significantly different from 11:00 – 12:00 (F

= 6.64; P = 0.0003) and from 13:00 – 16:00 hours (13:00 – 14:00: F = 10.32; P < 0.0001,

14:00 – 15:00: F = 22.21; P < 0.0001, 15:00 – 16:00: F = 3.35; P = 0.0209), when monitoring was conducted from 09:00 to 17:00 hours. Generally and across sampling dates, higher adult captures were recorded during 13:00 – 15:00 hours (13:00 – 14:00: 0.77 ± 0.23;

14:00 – 15:00: 0.95 ± 0.25 adults/card), compared with captures recorded from 09:00 – 11:00

(no captures) and from 16:00 – 17:00 hours (0.12 ± 0.05 adults/card) (Fig. 1).

Adult Intra-Plant Distribution. There were 1039 adults on 76 plants over the course of this experiment. The location of adult aggregations throughout the main stem of soybean plants was not influenced by sampling date (F = 0.03; df = 1, 4.99; P = 0.8756), MG (F =

1.79; df = 3, 142; P = 0.1511), plant inspected (F = 0.28; df = 2, 139; P = 0.7536), or any interaction among treatments (P = 0.1906 to 0.9741). However, the total number of aggregations differed among canopy sections within soybean plants (F = 28.80; df = 6, 139;

P < 0.0001). More aggregations were observed in the middle section of the main stem of plants, compared to both the upper and the lower sections (Fig. 2).

There were 148 adult aggregations on 76 soybean plants, and the number of aggregations ranged from zero to eight per plant; the average number of aggregations per plant was 1.95 ± 0.17. From those aggregations, the range of number of adults per aggregation was two to 18 and the average number of adults per aggregation was 5.16 ± 0.32.

The number of adults per aggregation was not influenced by either sampling date (F = 1.01; df = 1, 16.1; P = 0.3292), MG (F = 0.08; df = 3, 16.1; P = 0.9713) or their interaction (F =

0.32; df = 3, 16.1; P = 0.8121). Furthermore, 74% ± 3.29 of adults occurred in aggregations rather than individually on soybean plants, where most of the aggregations (98% ± 1.20) were found on main stems. Number of adult aggregations per plant was influenced by the interaction between sampling date and MG (F = 3.86; df = 3, 63.5; P = 0.0133). However, when controlling for the sampling date effect, the number of aggregations per plant was not affected by MG (P = 0.1637 to 0.4984).

There were 359 males and 234 females collected from adult subsampling. The male- to-female ratio was not affected by sampling date (F = 0.39; df = 1, 4.75; P = 0.5632), or MG

(F = 0.35; df = 3, 61.6; P = 0.7888), nor the interaction between these factors (F = 2.66; df =

3, 61.6; P = 0.0561). Numerically, there was a male-bias ratio across all plant inspections, with more males (4.72 ± 0.34) present per plant in the middle section than females (3.08 ±

0.23).

The average number of adults per plant across the two sampling dates was 13.67 ±

1.98. Adult density per plant was not influenced by sampling date (F = 2.83; df = 1, 59.7; P =

0.0977), MG (F = 0.22; df = 3, 56.4; P = 0.8812), plant height (F = 1.04; df = 4, 56.5; P =

0.3967), nor any interaction (P = 0.6591 to 0.9772). However, taller soybean plants were generally recorded from MGs VI (60.43 ± 1.53) and VII (56.23 ± 2.51), compared to MGs V

(48.38 ± 2.31) and IV (42.33 ± 3.02). More adults were usually observed on moderate-height soybean plants within each MG during both June and August, compared to either relatively shorter or taller plants (Fig. 3). There were few instances where taller MG VI and VII soybean plants had more kudzu bug adults than the moderate-height plants within those MGs during August (Fig. 3). Additionally, numerically more adults were on the main stem of soybean plants during August (16.76 ± 3.59), compared to June (10.25 ± 1.12) in 2013.

Discussion

The flight activity pattern of kudzu bug in soybean changes throughout the day. More kudzu bug adults were caught on sticky cards during 13:00 – 15:00 hours, indicating that this period was when the most flight and dispersion activities occurred in soybean during the dates I sampled. Adult densities per plant did not vary when different soybean phenologies were present at the same time. However, kudzu bug adult densities varied in their vertical distribution on soybean main stems. Most adults (74%) were observed in aggregations, rather than singly, and at the middle section of the main stem of soybean plants, rather than at the top or bottom. These visual plant inspections were taken during the morning (09:00 – 12:00 hours), because most kudzu bug adults were not flying at this time.

Ideally, additional sticky cards would have been placed inside the experimental soybean plots to record potential adults flying towards or away from these fields. Since sticky cards were placed at the perimeter of experimental fields, they might have been exposed to stronger wind gust and lower relative humidity conditions than those within experimental plots or the interior of commercial fields. However, this is the first effort to document at what time of the day kudzu bug adults are flying and dispersing in soybean fields. Results from this study could be complemented with future experiments in which sticky cards are placed inside experimental soybean fields to document if adult flight activity patterns at the canopy level and above the canopy differ. Additional research might also investigate how adult flight activity correlates with kudzu bug distribution within soybean plants during and after the peak of flight activity in the field.

Kudzu bug adult densities on soybean plants were not influenced by maturity group.

This finding is consistent with other studies showing the same trend (Blount et al. 2016, Del

Pozo-Valdivia et al. 2016). The experiment detailed in this study included soybeans ranging from MG IV to VII. This range of maturity groups represented a range of plant phenology and plant architecture at two time points in the season. However, the presence or absence of reproductive tissue did not affect population dynamics of kudzu bug in soybean, consistent with previous findings. In contrast, it is possible that plant architecture plays an important role in how kudzu bugs are distributed within plants. For example, the middle section of the plant might favor the presence of more kudzu bug adult aggregations because the main stem is wider compared to the upper portion of the plant, thereby providing more space and plant tissue on which adults can aggregate. Additionally, the micro-climate in the middle section of the plant could be beneficial for the kudzu bugs. By comparison, the lower portion of

soybean canopy is cooler and more humid, compared with the upper portion of the canopy, which is an average of 2.2ºC warmer and 10.6% dryer (Owen et al. 2013). Furthermore, kudzu bug may avoid predators or parasites in this section of the plant.

My study included different MGs, with different plant heights in the same field, and neither MG nor plant height influence kudzu bug density. Although plant height is one factor that characterizes plant canopy, it is possible that other characteristics besides plant height were confounded with MG and may have affected the attractiveness of soybean to kudzu bug. For instance, angle of insertion of leaves or pubescent levels may have changed potential visual cues for the kudzu bug to find a suitable soybean host plant.

Male-to-female ratio was not affected by MG. Since adult densities (number of males and females) were not directly influenced by MG, and the male-to-female ratio was not influenced either, it is expected that sex ratio would be mediated by either semiochemicals cues, such as an aggregation pheromone, or by a density-dependent factor, such as crowdedness of males on the host. In this study, there was a male-bias ratio in the adult population of kudzu bugs found on the middle section of soybean plants. This finding is consistent with Hibino and Itô (1983), who showed that there were more males than females in adult aggregations of Megacopta punctatissima (Montandon), kudzu bug’s sibling species, in Lespedeza crytobotria (Fabales) in Japan. In my study, the majority of kudzu bug adults were found in aggregations, but the activity of the kudzu bugs in aggregations was not measured. However, it is more likely that kudzu bug adults, recorded in my visual inspections, were either feeding or mating during the morning while aggregating than simply resting on plants. Megacopta punctatissima mainly forms aggregations to facilitate mating

(Hibino 1986). For example, more females of this species engage in mating behavior when

more than two males are present in an aggregation (Hibino 1986). It is also possible that kudzu bug may engage in aggregation to benefit mating. Future research on kudzu bug behavior may complement these observations and confirm when feeding, mating, and resting behaviors occur on soybean plants.

In conclusion, this study showed that kudzu bug adult flight activity varies during

09:00 to 17:00 hours, and most of the adults are found forming aggregations in the middle section of the main stem of soybean plants between 09:00 to 12:00 hours. From these data, I can speculate that flight and dispersion activity of the kudzu bug may be reduced from 09:00 to 12:00 hours in soybean fields, since adults form aggregations on soybean plants during that period of time. It is more likely that higher adult captures between 13:00 and 15:00 hours may represent the highest peak of flight activity of the kudzu bug. It is also possible that adults may not be forming aggregations on soybean plants while dispersing during this time.

However, additional observations are needed to confirm intra-plant soybean adult distribution during the afternoon to complement the flight activity experiment in this study, and to understand the intra-plant distribution of the kudzu bug throughout the day.

Acknowledgements

The author would like to thank NC State University Sandhills Research Station and

T.G. Gibson in NC for providing access to experimental plots. Dan Mott, Steven Roberson,

Clifton Moore, David Morrison, Jeremy Martin, and Eric Willbanks are gratefully acknowledged for their contribution to this research. This project was funded by the NC

Soybean Grower Association and the United Soybean Board.

References

Bellows, T. S., Jr, T. M. Perring, K. Arakawa, and C. A. Farrar. 1988. Patterns in diel

flight activity of Bemisia tabaci (Homoptera: Aleyrodidae) in cropping systems in

southern California. Environ. Entomol. 17: 225-228.

Blount, J. L., G. D. Buntin, and A. N. Sparks, Jr. 2015. Host Preference of Megacopta

cribraria (Hemiptera: Plataspidae) on Selected Edible Beans and Soybean. J. Econ.

Entomol. 108: 1094–1105.

Blount, J. L., G. D. Buntin, and P. M. Roberts. 2016. Effect of planting date and maturity

group on soybean yield response to injury by kudzu bug, Megacopta cribraria F.

(Hemiptera: Plataspidae). J. Econ. Entomol. 109: 207-212.

Campbell, C. A. M. and R. C. Muir. 2005. Flight activity of the damson–hop aphid,

Phorodon humuli. Ann. Appl. Biol. 147: 109–118.

Costamagna, A. C., B. P. McCornack, and D. W. Ragsdale. 2013. Within-plant bottom-up

effects mediate non-consumptive impacts of top-down control of soybean aphids.

PLoS ONE. 8: e56394.

Del Pozo-Valdivia, A. I., N. J. Seiter, D. D. Reisig, J. K. Greene, F. P. F. Reay-Jones,

and J. S. Bacheler. 2016. Megacopta cribraria (Hemiptera: Plataspidae) population

dynamics in soybeans as influenced by planting date, maturity group, and insecticide

use. J. Econ. Entomol. 109: 1141-1155.

Fernandes, F. S., F. S. Ramalho, J. L. Nascimento Junior, J. B. Malaquias, A. R. B.

Nascimento, C. A. D. Silva, and J. C. Zanuncio. 2012. Within-plant distribution of

cotton aphids, Aphis gossypii Glover (Hemiptera: Aphididae), in Bt and non-Bt cotton

fields. B. Entomol. Res. 102: 79-87.

Gardner, W. A. 2016. Distribution map of the kudzu bug. University of Georgia. Center for

invasive species and ecosystem health.

http://www.kudzubug.org/distribution_map.cfm (accessed 21 March 2016).

Golec, J. R., X. P. Hu, L. Yang, and J. E. Eger. 2015. Kudzu-deprived first-generation

Megacopta cribraria (F.) (Heteroptera: Plataspidae) are capable of developing on

alternative legume species. J. Agric. Urban Entomol. 31: 52–61.

Golec, J. R. and X. P. Hu. 2015. Preoverwintering copulation and female ratio bias: life

history characteristics contributing to the invasiveness and rapid spread of Megacopta

cribraria (Heteroptera: Plataspidae). Environ. Entomol. 44: 411–417.

Hibino, Y., and Y. Itô. 1983. Mating aggregation of a stink bug, Megacopta punctatissima

(Montandon) (Heteroptera: Plataspidae). Res. Popul. Ecol. 25: 180–188.

Hibino, Y. 1986. Female choice for male gregariousness in a stink bug, Megacopta

punctatissima (Montandon) (Heteroptera, Plataspidae). J. Ethol. 4: 91-95.

Horn, S. and J. L. Hanula. 2011. Influence of trap color on collection of the recently-

introduced bean plataspid, Megacopta cribraria (Hemiptera: Plataspidae). J. Entomol.

Sci. 46: 85-87.

Huang, T. and M. D. Toews. 2012. Feeding preference and movement of Nezara viridula

(L.) and Euschistus servus (Say) (Hemiptera: Pentatomidae) on individual cotton

plants. J. Econ. Entomol. 105: 847-853.

Huang, T., D. A. Reed, T. M. Perring, and J. C. Palumbo. 2013. Diel activity and

behavior of Bagrada hilaris (Hemiptera: Pentatomidae) on desert cole crops. J. Econ.

Entomol. 106: 1726-1738.

Kenward, M. G., and J. H. Roger. 1997. Small sample interference for fixed effects from

restricted maximum likelihood. Biometrics 53: 983–997.

Krupke, C. H., V. P. Jones, and J. F. Bruner. 2006. Diel periodicity of Euschistus

conspersus (Heteroptera: Pentatomidae) aggregation, mating, and feeding. Ann.

Entomol. Soc. Am. 99:169-174.

Levi-Zada, A., D. Fefer, M. David, M. Eliyahu, J. C. Franco, A. Protasov, E.

Dunkelblum, and Z. Mendel. 2014. Diel periodicity of pheromone release by

females of Planococcus citri and Planococcus ficus and the temporal flight activity of

their conspecific males. Naturwissenschaften. 101: 671-678.

McCornack, B. P., A. C. Costamagna, and D. W. Ragsdale. 2008. Within-plant

distribution of soybean aphid (Hemiptera: Aphididae) and development of node-

based sample units for estimating whole-plant densities in soybean. J. Econ. Entomol.

101: 1488-1500.

Medal, J., S. Halbert, T. Smith, and A. Santa Cruz. 2013. Suitability of selected plants to

the bean plataspid, Megacopta cribraria (Hemiptera: Plataspidae) in no-choice tests.

Fla. Entomol. 96: 631-633.

Owens, D. R., D. A. Herbert, Jr., T. P. Kuhar, and D. D. Reisig. 2013. Effects of

temperature and relative humidity on the vertical distribution of stink bugs

(Hemiptera: Pentatomidae) within a soybean canopy and implications for field

sampling. J. Entomol. Sci. 48: 90-98.

Pedersen, P. 2009. Soybean growth and development. Iowa State University Extension

Distribution Center. Publication PM1945. Ames. IA.

Russin, J. S., M. B. Layton, D. B. Orr, and D. J. Boethel. 1987. Within-plant distribution

of, and partial compensation for, stink bug (Heteroptera: Pentatomidae) damage to

soybean seeds. J. Econ. Entomol. 80: 215-220.

SAS Institute. 2010. SAS® version 9.3 user’s manual. SAS Institute, Cary, NC.

Seiter, N. J., F. P. F. Reay-Jones, and J. K. Greene. 2013. Within-field spatial distribution

of Megacopta cribraria (Hemiptera: Plataspidae) in soybean (Fabales: Fabaceae).

Environ. Entomol. 42: 1363–1374.

Seiter, N. J., J. K. Greene, and F. P. F. Reay-Jones. 2014. Aggregation and oviposition

preferences of Megacopta cribraria (Hemiptera: Plataspidae) in laboratory bioassays.

J. Entomol. Sci. 49: 331-335.

Snodgrass, G. L. 1998. Distribution of the tarnished plant bug (Heteroptera: Miridae) within

cotton plants. Environ. Entomol. 27: 1089-1093.

Suiter, D. R., J. E. Eger, Jr., W. A. Gardner, R. C. Kemerait, J. N. All, P. M. Roberts, J.

K. Greene, L. M. Ames, G. D. Buntin, T. M. Jenkins, et al. 2010. Discovery and

distribution of Megacopta cribraria (Hemiptera: Heteroptera: Plataspidae) in

northeast Georgia. J. Integr. Pest Manage. 1: F1–F4.

Taravati, S., H. Glenn, and C. Mannion. 2014. Daily flight activity of the rugose spiraling

whitefly (Hemiptera: Aleyrodidae). Fla. Entomol. 97: 1842-1844.

Terry, I. J. R. Bradly, Jr. and J. W. Van Duyn. 1987. Within-plant distribution of

Heliothis zea (Boddie) (Lepidoptera: Noctuidae) eggs on soybeans. Environ.

Entomol. 16: 625-629.

Zarrabi A. A., R. C. Berberet, M. E. Payton, and G. E. Hoard. 2005. Within-plant

distribution of Acyrthosiphon kondoi (Homoptera: Aphididae) on alfalfa. Environ.

Entomol. 34: 193-198.

Zhang, Y., J. L. Hanula, and S. Horn. 2012. The biology and preliminary host range of

Megacopta cribraria (Heteroptera: Plataspidae) and its impact on kudzu growth.

Environ. Entomol. 41: 40–50.

Fig. 1. Proportion of adults of Megacopta cribraria captured hourly (X-axis) on white sticky cards and by sampling dates (horizontal panels). Cards were placed around experimental fields at the Sandhills Research Station and a commercial soybean field near Gibson, NC, and monitored during June (on the 11th and 13th) and August (on the 6th and 13th) in 2013.

Fig. 2. Intra-plant distribution of adult aggregations of Megacopta cribraria in soybean.

Visual inspections of plants were conducted once during June at a commercial soybean field near Gibson, NC; and once during August at the Sandhills Research Station. Plant canopy was divided vertically into upper, middle and lower thirds. Adult aggregations were defined as two or more adults grouped together with no clear space separation between/among them.

Means sharing the same letter are not statistically significant (α > 0.05).

Fig. 3. Scatter plot showing the relationship between plant height (cm) of soybean (X-axis), planted during April using four different maturity group (MG) varieties (labeled “IV”, “V”,

“VI”, and “VII” on the vertical panels), and densities of adults per plant (Y-axis) recorded on

20 June 2013 (labeled “June” on the horizontal panels) and on 14 August 2013 (labeled

“August” on the horizontal panels). This type of plot was chosen to show all data point, including the outliers.

SECTION THREE

CHAPTER FOUR2

Megacopta cribraria (Hemiptera: Plataspidae) Population Dynamics in Soybeans as

Influenced by Planting Date, Maturity Group, and Insecticide Use

2Published in Journal of Economic Entomology, 2016, 109: 1141-1155

Abstract

Since its unintentional introduction in Georgia in 2009, Megacopta cribraria (F.) has spread rapidly throughout the southeastern U.S., mainly feeding and reproducing on kudzu,

Pueraria montana Loureiro (Merr.) variety lobata (Willdenow) and soybeans, Glycine max

(L.) Merr. Megacopta cribraria has become a serious economic pest in soybeans, forcing growers to rely solely on insecticide applications to control this insect. The main objective of this study was to investigate if variation in planting date and maturity group of soybeans had an impact on management of M. cribraria populations. Three experimental fields were located in NC (2) and SC (1), and the tests replicated during 2012 and 2013. Treatments consisted of three planting dates, four maturity groups, and insecticide treated vs. untreated, at each location. More M. cribraria were found in untreated early planted soybeans than late planted soybeans. Generally, maturity group did not influence population densities of M. cribraria. Yield was significantly influenced by the interaction between planting date and maturity group. There was a negative linear relationship between M. cribraria populations and soybean yield. Although early planted soybeans may avoid drought conditions and potentially large populations of defoliators, these fields may be at greater risk for infestation by M. cribraria.

Introduction

Soybean, Glycine max (L.) Merrill, is one of the most important crops in the U.S., planted on over 31 million hectares in 2013, second only to corn, Zea mays L. The value of soybeans was estimated at $41.8 billion of farm-gate value during 2013 (USDA 2015).

Cultural practices, including varietal selection and planting date, can influence soybean yield

(Parker et al. 1981, Board et al. 1996, Khaled et al. 2011). There are three soybean planting systems used in the midsouthern and southeastern U.S.: an early soybean production system, full-season soybeans, and double-cropped soybeans (Heatherly et al. 1998). Farmers using early soybean production systems usually plant soybeans in April. Full-season soybeans are generally planted in late April to May. Additionally, these two systems differ in the suggested maturity group. Maturity group (MG) in soybeans (expressed as Roman numerals from 00 to IX) can be defined as the plant’s capacity to flower under specific photoperiod requirements, and it is expressed as days-to-flowering (Pedersen 2009). Group 00 is the earliest maturity group and it is adapted to northern regions, whereas group IX is the latest and adapted to the southern U.S. (Pedersen 2009). Recommendation for maturity group selection under early soybean production system is to plant only early maturing soybeans

(MG III and IV), compared with the full-season system where late maturing soybeans are selected (MG V – VII) (Bowers 1995, Heatherly and Spurlock 1999). Selection of planting date and maturity group in the full-season soybean production system depends on weather predictions and targeted harvesting time (Frederick et al. 1998). In a double-cropped system, soybeans are planted in June, immediately following the winter grain harvest (Frederick et al.

1998). Usually, MG VI and VII are planted under the double-cropped system (Frederick et al. 1998, Heatherly and Hodges 1999).

Varying soybean planting date can impact not only plant performance and yield potential, but also can influence insect densities in this crop (Buschman et al. 1984). Planting date is known to influence insect densities in many crops such as corn, Zea mays L. (Smith and Riley 1992), wheat, Triticum aestivum L. (Morrill and Kushnak 1999), cowpea, Vignia unguiculata (L.) Walp (Asante et al. 2001), and cotton, Gossypium hirsutum L. (Slosser

1993, Bi et al. 2005). In soybeans, planting date and maturity group have been shown to affect seasonal abundance of insect pests and beneficial arthropods in Arkansas (Tugwell et al. 1973), Georgia (McPherson and Bondari 1991) and Louisiana (Boyd et al. 1997). In general, lower insect pest populations of velvetbean caterpillar (Anticarsia gemmatalis

Hübner, Lepidoptera: Noctuidae), soybean looper (Pseudoplusia includens (Walker),

Lepidoptera: Noctuidae), and stink bugs (Hemiptera: Pentatomidae) are found in early planted and early maturing soybeans (McPherson et al. 2001, Gore et al. 2006). McPherson et al. (2001) suggested that one reason for this phenomenon could be the migratory nature of these defoliators. However, having pods available early in the season can attract non- migratory pests such as stink bugs (McPherson and Bondari 1991, Baur et al. 2000, Smith et al. 2009). Once early planted soybeans senesce, stink bugs migrate to available and suitable host plants, including later planted soybeans (Smith et al. 2009, Herbert and Toews 2011).

Besides having a lower risk of late-season pest outbreaks, planting earlier may allow growers to avoid drought conditions during summer (Frederick et al. 1998, Heatherly and

Hodges 1999, McPherson et al. 2001). The modification of selected cultural practices in soybeans, such as row-spacing, can impact the populations of insect pests (Frederick et al.

1998, Heatherly and Hodges 1999). Soybeans planted on wide rows (row-spacing > 0.76 m) are more susceptible to infestations of Helicoverpa zea Boddie (Lepidoptera: Noctuidae)

compared to narrow rows; canopy architecture influences oviposition, and the lack of canopy closure on wide rows makes the crop more attractive for oviposition when the crop is more susceptible to this pest (Bradley and Van Duyn 1979). Southern U.S. soybean growers have narrowed row spacings over time, reducing the risk of H. zea infestations.

Soybean pest status has changed in some parts of the U.S. since 2010, especially in the Southeast, due to the presence of an additional economic pest in the system. The kudzu bug, Megacopta cribraria (F.) (Hemiptera: Plataspidae), was accidentally introduced into the

U.S. from Asia (Eger et al. 2010, Suiter et al. 2010). First reported in Georgia during 2009,

M. cribraria has spread rapidly throughout the southeastern region of the U.S. (Ruberson et al. 2013), and was recently discovered in Arkansas and Kentucky (Gardner 2015).

Megacopta cribraria is a piercing-sucking insect that undergoes five nymphal stadia in 40-50 days (Eger et al. 2010, Zhang et al. 2012, Del Pozo-Valdivia and Reisig 2013). Females deposit several capsules underneath egg masses containing a γ-protobacterium endosymbiont, Candidatus ishikawaella capsulata (Hosokawa et al. 2006). Following eclosion, first instar nymphs feed on these capsules, acquiring the endosymbiont (Hosokawa et al. 2006). It is hypothesized that this endosymbiont is required by M. cribraria for growth and development on different host plants (Hosokawa et al. 2006, Hosokawa et al. 2007).

Megacopta cribraria has two generations per year in the southeastern U.S. (Ruberson et al. 2013, Seiter et al 2013a). Adults begin to emerge from overwintering sites in early

April. The first generation of adults typically peaks in June and the second peaks in August

(Ruberson et al. 2013, Seiter et al 2013a). Although M. cribraria was initially observed feeding on kudzu, Pueraria montana Loureiro (Merrill) variety lobata (Willdenow), it was recognized that it can also feed and reproduce on soybeans directly from overwintering (Del

Pozo-Valdivia and Reisig 2013). This insect can be found feeding on other legumes, including lespedeza, Lespedeza spp., and wisteria, Wisteria spp. (Eger et al. 2010, Zhang et al. 2012); however, larger populations of M. cribraria have been found on kudzu and soybeans compared with these plants (Ruberson et al. 2013). Research on this insect has demonstrated that M. cribraria will aggregate alone soybean field edges (Seiter et al. 2013a) and can reduce soybean yield up to 60% when left uncontrolled in a confined environment, such as field cages (Seiter et al. 2013b). Furthermore, some early planted soybeans

(especially those planted in April) are more prone to M. cribraria infestations and harbor more throughout the season, likely because they can support both generations of this putatively bivoltine pest (Blount et al. 2016). Currently, insecticide applications are the only short-term solution to manage this pest in soybeans (Seiter et al. 2015a).

The main objective of this study was to determine the impact of varying soybean planting dates and maturity groups on M. cribraria field populations collected by two different scouting procedures (sweep-net sampling and insect density per plant). This two- year, two-state, multi-site field experiment also revealed how M. cribraria populations and soybean yield were affected by the application of selected insecticides.

Materials and Methods

North Carolina. There were two field sites; one was located at the North Carolina

State University Sandhills Research Station, near Jackson Springs, NC, in Montgomery

County and another at a commercial soybean field near Gibson, NC, in Scotland County during 2012 and 2013. Field experiments were set up with a split-split plot design with four replications per site, where the main plot was planting date (three dates), the split-plot was

maturity group (MG) (four groups), and the split-split-plot was insecticide treatment (sprayed or unsprayed). Experimental plots, four-rows wide by 12.2 m long, were planted with 0.97 m row-spacing using a four-row cone planter (John Deere model 1750, Deere and Co., Moline,

IL) in Montgomery County, where the targeted seeding rate was 29 seeds/row meter. In

Scotland County, experimental plots were planted with 0.91 m row-spacing using a two-row disc planter (White model 6700, AGCO Corporation, Duluth, GA). Plot dimensions and seeding rate in Scotland County were the same as Montgomery County plots. For both locations, planting dates were 16-17 April 2012, 17-18 May 2012, 18-19 June 2012, 16-18

April 2013, 15 May 2013, and 17-20 June 2013. For both locations and both years, Roundup

Ready soybean seeds (Asgrow, Monsanto Company, St. Louis, MO) with no insecticide coating were planted with MG IV (variety AG4531), V (AG5503), VI (AG6132), and VII

(AG7502).

Scouting for M. cribraria adults and egg masses/whole-plant was conducted at 14 and

28 days after planting by visual inspection of plants in 0.61m-row samples. During these visual inspections, six samples per plot were taken where total number of M. cribraria adults and eggs were counted. In 2012, visual inspections were not performed in soybeans planted in April. Once plants reached vegetative stage five (V5, Fehr et al. 1971), sweep-net samples were taken every other week, from ~42 days after planting until plants reached reproductive stage seven (R7; Fehr et al. 1971). Twenty sweeps, using a 0.38 m diameter standard sweep net, were taken on each sampling date. The net was plunged into the canopy so that the entire diameter of the net was submerged in the canopy just below the top of a single soybean row.

Numbers of M. cribraria adults and nymphs were recorded during sweeping, as well as other insect pests such as defoliators and stink bugs. Collected insects were released back into the

plots after counting. Newly unfolded top soybean trifoliates were randomly selected from 25 plants in each plot and inspected for M. cribraria eggs. Adult and nymph densities on a per plant basis were recorded from the previous 25 selected plants. Sweep-net and trifoliate samples were taken initially from rows two and three, respectively. At the following sampling date, sample locations were switched, where sweep-nets were taken from row three and trifoliates from row two. Sample locations were kept alternating in this manner during each visit until soybean plants reached R7.

Insecticide treatment regimens were either protected (sprayed every two weeks with insecticide after the first application) or unsprayed. Bifenthrin (Discipline 2EC, AMVAC

Chemical Corp., Los Angeles, CA) was applied at 0.11 kg/ha of active ingredient, in a volume of 93.5 l/ha, using a two-row CO2 backpack sprayer with TX-10 hollow-cone nozzles

(Teejet, Wheaton, IL). Insecticide application was triggered when 10 or more adults/plant

(from V1 – V4) or 0.5 – 1.0 adult/sweep (from V5 – R4) of M. cribraria were found across maturity groups in one planting date. A total of four and five insecticide applications were made during 2012 and 2013, respectively. During 2013, larvae of a heliothine species reached economic threshold at both locations in late July. To minimize yield losses associated with either corn earworm or tobacco budworm, Heliothis virescens (F.)

(Lepidoptera: Noctuidae), flubendiamide (Belt, Bayer CropScience LP, Research Triangle

Park, NC) was applied once to the entire experiment at 0.11 kg/ha of active ingredient, using the same backpack sprayer and volume/ha previously mentioned. Flubendiamide was chosen because, at the rate used in these applications, it had little to no impact on M. cribraria

(Seiter et al. 2015b).

Soybean maturity was reached at different dates, depending on the maturity group and planting date. To make the harvest operation feasible, one harvesting date per site was selected during 2012 and 2013. Soybeans were harvested using a two-row mechanical plot combine harvester (Gleaner model K2, AGCO Corporation, Duluth, GA). The middle two rows of each plot were harvested to calculate yield/ha. Seed shattering was measured immediately after the combine harvested the plots. Since seed shattering was not different among planting dates nor maturity groups at any given year or location (data not shown), seed weight was not corrected. Soybean yield per plot and moisture content were measured to determine yield/ha with 13% moisture content.

South Carolina. Trials were established in 2012 and 2013 at the Clemson University

Edisto Research and Education Center in Barnwell County, SC. Soybeans were planted using a four-row planter (John Deere MaxEmerge II model 7300, Deere and Co., Moline, IL) with a row spacing of 0.97 m at a seeding rate of 25 seeds/row meter. Plots were eight rows (7.7 m) wide by 12.2 m long. Planting dates were 20 April 2012, 18 May 2012, 5 July 2012 (the 5

July planting was originally planted on 22 June, but dry conditions resulted in poor germination and emergence, therefore those plots were replanted), 18 April 2013, 20 May

2013, and 26 June 2013. For both years, Roundup Ready soybean seeds (Asgrow, Monsanto

Company) were planted with MG IV (variety: AG4730), V (AG5732), VI (AG6732), and

VII (AG7532). Seed was treated with 0.13 mg clothianidin per seed (Poncho/Votivo, Bayer

CropScience LP). Although the evidence was mixed, previous field observations indicated that M. cribraria might be attracted to soybeans planted following a neonicotinoid seed treatment. Hence, seeds were treated to increase chances of M. cribraria infestation.

Insecticide treatment regime was either unsprayed or protected from populations of

M. cribraria with a foliar insecticide spray. In the protected treatments, applications of 0.035 kg/ha of λ-cyhalothrin and 0.046 kg/ha of thiamethoxam (Endigo ZC, Syngenta Crop

Protection LLC, Greensboro, NC) were made in a volume of 93.5 l/ha using a high-clearance self-propelled sprayer. Insecticide applications were triggered when presence of M. cribraria nymphs was observed across maturity groups in one planting date. A total of three applications were made during each year of this experiment. Because of the presence of soybean looper and heliothine species during 2013, spinosad (Tracer Naturalyte, Dow

AgroSciences, Indianapolis, IN) was applied to the entire experiment on 24 July and 27

August at a rate of 0.077 kg/ha. Spinosad was chosen because the rate used in these applications had little to no impact on M. cribraria (Seiter et al. 2015b).

The experimental design differed between 2012 and 2013. In 2012, unsprayed and insecticide protected treatments were applied to separate, adjacent experiments, which were each deployed as a randomized complete block design with four replications. All combinations of planting date and maturity group were randomly assigned within each block.

In 2013, a single experiment was conducted as in NC as a split-split plot with four replications, where the main plot was insecticide treatment regime (protected or unsprayed), the split-plot was MG (four groups), and the split-split-plot was planting date (three dates).

The methodology for monitoring M. cribraria was similar to the one followed in NC.

In 2012, visual inspections recorded from each sample on the 14th and the 28th day after planting included only egg masses. In 2013, total number of adults per sample was also determined during visual inspections, along with egg masses. Sweep-net samples were taken every other week, starting from ~42 days after planting and ended when plants reached

reproductive stages six to eight (R6 to R8, Fehr et al. 1971) in 2012. Ten sweeps were taken on each sampling date. The net was placed into the canopy so that at least half of the diameter of the net was submerged in the canopy. Sweep-net samples were taken at 180° across two soybean rows; and then alternated between the second and third and the sixth and seventh rows at each plot. Number of M. cribraria adults and nymphs were counted and recorded from sweep-net samples. From reproductive stage seven (R7) to eight (R8), sweep- net samples were taken from rows one and two or seven and eight (outside rows) to avoid damaging internal rows. In 2013, sweep-net sampling ended at R7 to avoid damaging the outside rows at each plot. To monitor M. cribraria eggs, 25 randomly selected, newly unfolded top soybean trifoliates were inspected for presence of egg masses.

Soybeans were harvested from the four center rows in each plot using a two-row plot combine (model 8-XP, Kincaid Equipment Manufacturing, Haven, KS). Harvest was triggered based on when different planting date and maturity group combinations reached maturity; therefore different planting date and maturity group combinations were selectively harvested at different times as soon as they reached R8. Soybean yield/plot and moisture content were measured in order to determine yield/ha with 13% moisture content.

Data Analysis. Insect evaluations on M. cribraria and soybean yield from NC were pooled together based on similar experimental design and methodologies, creating a single data set that included a new variable named ‘trial’. The variable trial accounted for the interaction of both years (2012 and 2013) and location (Montgomery and Scotland Counties).

Since NC and SC differed in the experimental design and how sweeping (post-V5) was performed, data from SC were analyzed separately from NC. For both NC and SC, data from the first two insect evaluations on M. cribraria (visual inspections) were combined into one

data set for NC and two data sets for SC (one for each year); all three data sets were analyzed using a generalized linear mixed model (PROC GLIMMIX, SAS version 9.3, SAS Institute

2010, Cary, NC) approach. Individual analysis of variance (ANOVA) tests were conducted where response variables were: egg masses/0.61 m-row and adults/0.61 m-row. These analyses incorporated early season data (V5 and earlier), before sweep-net sampling.

Because insecticide treatment had not begun at this time, treatments were organized as a randomized complete block design for SC 2012 (two separate fields); and split-plot for NC and SC 2013. Fixed effects included planting date, maturity group and the interaction between planting date and maturity group. Data distribution was selected as log-normal for egg masses and adults, based on model selection criteria (Littell et al. 2006). Degrees of freedom were calculated using the procedure of Kenward and Roger (1997). The effect of replication nested with trial, trial alone, and replication nested with trial by planting date interaction were included in the random statement for analyzing NC data. Replication was considered nested to account for the new hierarchy of this class after the variable ‘trial’ was originated when NC data were pooled together. The random statement in the SC analyses included effects of replication in 2012 and replication and the planting date by replication interaction in 2013.

Total numbers of M. cribraria egg masses/25 trifoliates, and M. cribraria adults and nymphs/plant were calculated from visual counts (V6 and older). Using the sweep-net data

(V6 and older), cumulative insect days were also calculated to measure the magnitude and duration of infestation of M. cribraria by following the equation from Ruppel (1983):

Insect-days = (Xi+1 – Xi) × [ (Yi + Yi+1) / 2 ] ,

where Xi and Xi+1 were adjacent sampling dates; and Yi and Yi+1 were M. cribraria densities for those adjacent sampling dates. Insect density/plant and cumulative insect days/sweep were log10-transformed [log10(X+1)] to comply with the assumptions of the ANOVA.

Soybean yield and post-V5 sampling data from NC and SC were analyzed using individual mixed model ANOVAs (PROC MIXED, SAS Institute 2010). The response variables were soybean yield, log10-transformed M. cribraria adults/plant, log10-transformed

M. cribraria nymphs/plant, log10-transformed cumulative M. cribraria egg masses/25 trifoliates, log10-transformed cumulative M. cribraria days/sweep for nymphs, and log10- transformed cumulative M. cribraria days/sweep for adults. In the NC statistical model, fixed effects were planting date, maturity group, insecticide regime and their interactions. Random effects were replication nested with trial, trial alone, the interaction between replication nested with trial by planting date, and the interaction of maturity group by replication nested with trial by planting date. In the SC 2012 model, unsprayed and sprayed experiments were analyzed separately with planting date, maturity group, and the interaction between planting date and maturity group as fixed effects, and replication alone as a random effect. In the SC

2013 model, the fixed effects were insecticide, maturity group, planting date, and their interactions; random effects were replication alone, replication by insecticide regime interaction, and the replication by insecticide regime by maturity group interaction. Degrees of freedom from all models were also calculated using the Kenward-Roger’s procedure.

Mean separation post-ANOVA of the transformed data was performed using the Tukey’s test at α ≤ 0.05. Means and standard errors are reported from the back-transformed data.

A regression analysis was performed between soybean yield (response variable) and

M. cribraria densities (independent variable) using PROC REG (SAS Institute 2010).

Transformed insect data were used to comply with assumptions of the regression analysis.

Data from each location and year were analyzed separately. Significance of the linear relationship and coefficients of determination (R2) were calculated.

Results

Planting Date Effect on M. cribraria Densities. Planting date alone influenced the number of egg masses of M. cribraria during 2012 and 2013 in SC, when plants were V5 or younger (Table 1). More egg masses/0.61 m-row were found on soybeans planted during

April compared with June or July planted soybeans (Table 2). Planting date had an effect on the number of egg masses found on the uppermost fully-expanded soybean trifoliates in both untreated and insecticide treated soybeans in 2012 in SC, when plants were V6 or older

(Table 3). The highest numbers of egg masses/25 trifoliates were observed on soybeans planted in April compared with July planted soybeans (Fig. 1). Additionally, egg masses were influenced by planting date during 2013 in SC (Table 4). April planted soybeans had more egg masses than May or June planted soybeans (Table 2).

Planting date also affected cumulative insect days for nymph/sweep in both untreated and insecticide treated soybeans in 2012 in SC when plants were V6 or older (Table 3).

There were more cumulative nymph days per sweep in April and May planted soybeans in the 2012 untreated trial compared with July (Fig. 1). In contrast, more cumulative insect days for nymphs/sweep were experienced from treated plots planted in July, compared to soybeans planted in April and May (Fig. 1). During the same year in the 2012 untreated trial, soybeans planted in May had more cumulative insect days for adults/sweep compared with soybeans planted in April and July (Table 3, Fig. 1).

Densities of M. cribraria egg masses and adults per 0.61m-row were influenced by the interaction of planting date and maturity group in NC during 2012 and 2013, and SC during 2013 (Table 1). When planted during April, MG V and VI soybeans had more egg masses than the MG IV and VII varieties in NC. Also, the MG V and VI varieties planted during April had more egg masses than MG IV, V, and VII varieties planted during June in

NC (Fig. 2). More adults were observed on MG V and VI soybeans planted in April than MG

IV and VII soybeans planted in the same month in NC (Fig. 2). On the contrary, more adults were observed on MG IV soybean plants planted during April in SC, compared with MG V and VII soybeans planted during May or with all MG of soybeans planted during June (Fig.

2). Cumulative insect days for adults/sweep were also affected by the interaction between planting date and maturity group in the insecticide treated trial in 2012 in SC (Table 3).

Maturity groups V and VII soybeans planted in July, and MG IV planted in May had more cumulative insect days for adults/sweep, compared with soybeans planted with MGs V, VI, and VII planted in May or any MG planted in April (Supp. Table 1).

The interaction between planting date and insecticide regime consistently influenced

M. cribraria densities when plants were V6 or older in both NC and SC (Tables 4 and 5).

Cumulative number of egg masses deposited by M. cribraria on soybean trifoliates was influenced by this interaction in NC during 2012 and 2013 (Table 5). More egg masses were found in both insecticide treated and untreated soybeans planted in April and May compared with soybeans planted in June; and more eggs were found in April plantings than May plantings (Fig. 1). There were more egg masses in untreated soybeans than insecticide treated soybeans, except for the June plantings (Fig. 1).

Cumulative insect days for both M. cribraria nymphs/sweep and adults/sweep were influenced by the interaction of planting date and insecticide regime during 2012 and 2013 in

NC (Table 5), and in SC during 2013 (Table 4). More insect days for nymphs were calculated for untreated soybeans planted during April or May compared with untreated soybeans planted during June in both NC and SC (Fig. 1). Fewer insect days for nymph/sweep were calculated from any insecticide treated soybean planted on any date in NC; or from soybeans treated with insecticide and planted in April than insecticide treated soybeans planted in June in SC (Fig. 1). Higher insect days for adults accrued on untreated soybeans planted in April compared with untreated soybeans planted during June in both NC and SC (Fig. 1). Fewer insect days for adults/sweep were calculated from untreated soybeans planted in June than

May plantings in NC; and from any of the other insecticide treated soybean planted on any date in SC (Fig. 1). Similar insect days for adults accrued for insecticide treated soybeans planted in April compared to untreated soybeans planted during May in NC, insecticide treated soybeans planted during May compared to untreated soybeans planted during June in

NC, and untreated soybeans planted during April compared with untreated soybeans planted during May in SC (Fig. 1).

The interaction between planting date and insecticide regime also consistently influenced nymph and adult presence in soybeans in NC during 2012 and 2013 when plants were V6 or older (Table 5). There were more nymphs per plant in untreated soybeans planted in April or May, compared to insecticide treated and untreated soybeans planted in June, or insecticide treated soybeans planted in either April or May (Fig. 3). Additionally, untreated soybeans planted in April had the highest densities of adult/plant, compared to treated soybeans planted during April or insecticide treated or untreated soybeans planted in May

and June (Fig. 3). More adults/plant were also observed on insecticide treated soybeans planted in April than insecticide treated soybeans planted in May or June and untreated soybeans planted in June (Fig. 3).

The Effect of Maturity Group on M. cribraria Densities. Maturity group alone influenced the cumulative number of egg masses deposited by M. cribraria on the uppermost fully-expanded soybean trifoliates when plants were V6 growth stage or older in NC (Table

5) and SC in 2013 (Table 4). There were more egg masses/25 trifoliates on MG IV soybeans compared with MG V soybeans in NC; and more eggs on MG V soybeans than on MG VI soybeans in SC (Table 6). Cumulative insect days for adults/sweep were also affected by maturity group alone in SC during 2013 (Table 4). More insect days for adults were accrued on MG IV and V soybeans compared with MG VII soybeans. Additionally, densities of nymphs/plant were also influenced by maturity group when soybeans were V6 or older in NC during 2012 and 2013 (Table 5). More nymphs/plant were found in soybeans planted with

MG IV compared with MGs V and VI, where MG VII soybeans had the same nymph density compared to the rest of MGs (Table 6).

Soybean Yield. The interaction between planting date and maturity group influenced soybean yield in NC (Table 5), the SC insecticide treated trial in 2012 (Table 3), and the SC

2013 (Table 4). Yields (ranging from 3,300 – 3,700 kg/ha) were higher for MG IV, V, and VI soybeans planted in May compared with MG IV soybeans planted in either April or June

(Fig. 4). There were relatively high yields for MG VI and VII soybeans planted in April (Fig.

4). Yields were relatively lower for MG V and VI soybeans planted in June (SC 2012 insecticide treated trial), for MG VII soybeans planted during June (SC 2012 insecticide treated trial and SC 2013), and for MG V soybeans planted during April (SC 2012 insecticide

treated trial) (Fig. 4). Planting date and maturity group alone (Table 3) affected yield in the

SC untreated trial in 2012. Soybeans planted in May and July, usually had higher yields than

April planted soybeans (Table 2). Yields were higher in MG V soybeans, compared to MG

IV and VI soybeans (Table 6).

Soybean yields were influenced by both the interaction between planting date and insecticide regime in SC in 2013 (Table 4), and the interaction between MG and insecticide regime in NC (Table 5) and SC in 2013 (Table 4). Plots yielded more when treated with insecticide and planted either in April or May, compared with the untreated plots planted during April, May, or June, and insecticide treated plots planted in June (Supp. Table 2). In addition, insecticide treated soybeans planted with MG V yielded more, compared with insecticide treated soybean planted with MG IV or untreated soybeans planted with either

MG IV or V in NC (Fig. 5). In SC during 2013, insecticide treated soybeans planted with

MG V and VI soybeans had higher yields compared with insecticide treated soybean planted with MG IV or untreated soybeans planted with either MG IV or V, and untreated soybeans planted with MG IV (Fig. 5).

There was a negative relationship between soybean yield (sometimes log10- transformed to correct for normality of residuals) and log10-transformed abundance of M. cribraria in a majority of locations in NC and SC (Table 7). Higher-yielding trials in NC

(average yield above 3,344.64 kg/ha), including Montgomery County during 2012 and

Scotland County during 2013, with relatively low pressure from M. cribraria (below 2.40 and 10.78 cumulative adults and nymphs/sweep in 2012 and 2013 respectively) failed to show a relationship between soybean yield and insect abundance. Coefficients of

determination (R2) were 0.23 or higher when insect levels were 27.59 cumulative adults and nymphs/sweep or higher, and soybean yield ranged from 2,585.73 to 3,050.16 kg/ha.

Discussion

The present study showed that in the early season, when plants were at V5 growth stage or younger, both planting date and maturity group influenced M. cribraria egg masses and adult densities in soybeans. This study also showed that in soybeans at the V6 growth stage and older, M. cribraria densities, as measured by both cumulative insects per sweep and number of insects per plant, were consistently influenced by the interaction between soybean planting date and insecticide treatment regime.

Adults in the early season, when plants were at V5 growth stage or younger, were assumed to be from the F0 generation (coming directly from overwintering sites), although this study did not address the origin of the adult population. There was not a clear preference of dispersing adults for any MG of soybeans planted during May or June in NC nor any MG of soybean planted during April in SC. Adult densities were more abundant in MG V and VI soybeans planted during April, compared with MGs IV and VII planted during the same month in NC. In addition, more adults were found on soybeans of all maturity groups planted during April in SC, compared to soybeans of all maturity groups planted in June. Fewer adults emigrating from overwintering sites and the existence of an early planted host in the field could have reduced adult densities in June planted soybeans when plants were at the V5 growth stage or younger. We can infer that fewer egg masses were found in June planted soybeans because there were fewer adults in those late planted soybean plots. For instance, fewer eggs were found in MG V soybeans planted during June in NC, compared with MG V

soybeans planted in April. Fewer egg masses were also found in SC soybeans planted in May than April, and fewer were found in soybeans planted in June than May. However, we cannot explain the fact that there were fewer egg masses on MG IV, an indeterminate soybean variety, and MG VII, a determinate soybean variety, planted in April in NC, when compared to MG V and MG VI (both determinate varieties). Since a single variety was used as a proxy for MG effects, one possible explanation could be a varietal effect. It would have been ideal to include two or more different soybean varieties with four levels of maturity group each, to potentially explore the effect of variety on M. cribraria egg mass deposition. We hypothesized that the variety planted in this experiment might have affected M. cribraria densities early in the season during each planting date. Different varieties were planted in SC and adult densities were not assessed in the early season during 2012. Moreover, the effect was only observed in a single year in NC. So it is not possible to test this hypothesis with data we collected in this study.

In soybeans at the V6 growth stage and older, M. cribraria densities were consistently influenced by the interaction between soybean planting date and insecticide treatment regime. Although the assignment of the sub-sub-plot for NC (insecticide treatment regime) and SC (planting date) varied in 2013, the overall findings were similar, despite the different precision levels for each factor. Megacopta cribraria was most prevalent in untreated soybeans planted in April compared to soybeans planted in June and left untreated, corroborating a previous finding from Georgia (Blount et al. 2016). However, our findings were not as pronounced in NC, as those observed in the GA study. Whereas they observed

73% fewer adults in June planted soybeans, we observed only 55% fewer in NC (data from both years combined). Early planted soybeans are one of the first hosts available for

overwintering adults of M. cribraria (Ruberson et al. 2013), which could explain this phenomenon. Hence, more M. cribraria egg masses were laid on these early planted soybeans (April planted) when they were at, or before, the V5 growth, compared with June or

July planted soybeans. Furthermore, there could be an interaction of geography with planting date and infestation rate. One possible factor influencing dispersal might be that M. cribraria breaks overwintering diapause earlier in southern latitudes than northern ones due to changes in temperature or impacts of photoperiodism. Yet another possibility is a density dependent effect, as densities were higher in the GA study and in SC trials, compared to those in NC. It would have been expected that insect’s crowdedness and potential competition for food may have trigger adult dispersal in M. cribraria. Since early planted soybeans are one of the first available host, M. cribraria may not disperse to late planted soybeans until the carrying capacity, unknown for this insect in soybeans, is reached.

Planting date alone influenced M. cribraria densities when it was not possible to incorporate the insecticide treatment in the statistical analysis (SC experiment from 2012).

More M. cribraria eggs masses were always found in April planted soybeans compared to

July planted soybeans at either of the separate insecticide-treated or untreated trials during

SC 2012. Insecticide application to an entire field drastically reduced M. cribraria densities in this trial where approximately ten times fewer cumulative insect days for adults/sweep accrued, compared with the untreated trial during SC 2012. Densities of nymphs and adults of M. cribraria were affected by planting date when insecticide was sprayed as a whole-field application, compared with the adjacent field left untreated during SC 2012. Planting date also affected where egg masses of M. cribraria were laid in the SC 2013 trial, since an attractive host for oviposition was available at different times during the season. However,

planting date alone did not influence subsequent life stages of M. cribraria. Higher mobility and inter-plot movement of adults might have impacted the in-field distribution of M. cribraria in plots with different planting dates and insecticide regimes in the SC 2013 trial.

Rapid oviposition during early growth stages (VC to V2) might have been one of the reasons why early planted soybeans accumulated higher populations of M. cribraria throughout the season; it is also possible that presence of M. cribraria adults early in the season might have attracted later emigrating adults into these experimental fields. Adults already established in the crop, might have produced semiochemicals that influenced movement patterns by directly or indirectly attracting subsequent emigrant adults in the field.

The existence of an aggregation pheromone in M. cribraria has not been proven; however, research and field observations have shown that adults cluster while feeding on host plants

(Seiter et al. 2013a) and that they aggregate for mating (Hibino and Itô 1983). There was only one instance where relatively high numbers of M. cribraria adults were recorded in July

(SC in 2012 insecticide treated trial). There were fewer M. cribraria adults in the beginning of the sampling period in late planted soybeans (5 to 13 adults/10 sweeps), compared to the last sampling dates at the same plots (170 to 280 adults/10 sweeps). Possibly the increase in

M. cribraria adult abundance in late planted soybeans at this trial originated from a late infestation of F1 generation of immigrant M. cribraria adults.

Maturity group alone did not consistently have an effect on densities of M. cribraria when plants were at V6 growth stage and older; our studies included both indeterminate varieties (MG IV) and determinate varieties (MGs V-VII). Maturity group did have a significant effect in SC during 2013, where more M. cribraria adults were found in MG V soybeans compared with MG VII. Because M. cribraria is thought to feed on phloem

components (Zhang et al. 2012), the presence or absence of reproductive plant tissue would not likely be a major driving force to attract M. cribraria to soybeans. However, soluble nutrient content in soybeans can vary in the phloem (Walter and DiFonzio 2007) and xylem

(Krishnan et al. 2011) and this is dependent on plant developmental stage. This, in turn, can influence densities of insects, such as aphids (Walter and DiFonzio 2007). In SC, the peak of

M. cribraria oviposition (9.81 egg masses/25 trifoliates) was recorded on 19 June 2013. At that time, MG V soybeans were at fully flowering (R2), while MG VII soybeans were in vegetative stages (V8 – V9). Higher numbers of egg masses were recorded from MG V soybeans (15.50 egg masses/25 trifoliates), compared with MG VII soybeans (8.25 egg masses/25 trifoliates). Having more egg masses during crop establishment or during early growth stages could have led to a larger first in-field generation (F1 generation) of M. cribraria in MG V plants in SC during 2013.

The interaction between planting date and maturity group consistently influenced soybean yield in NC and SC. In contrast to a related study which did not find an interaction

(Blount et al. 2016), soybeans in our study usually yielded more when planted in May or

April using MGs V or VII variety, respectively. Optimum yield can be achieved when the ideal planting date and maturity group are selected for a specific environment or location (Hu

2013). Environmental conditions, such as average daily temperatures, drought periods, and the length of the day directly impact plant physiology and, ultimately, yield. In this study, soybeans planted in June or April using a MG IV variety usually yielded the least. Soybean yield was also impacted by the interaction of insecticide treatment (which reduced densities of M. cribraria) and planting date in SC in 2013, similar to a previous related study (Blount et al. 2016). For example, soybean yield was reduced by 20.6% in SC in 2013 when May

planted untreated plots were compared to insecticide treated plots. Blount et al. (2016) found a 28.18% reduction when untreated soybeans were compared to insecticide treated soybeans planted in May.

In our study, soybean yield was also influenced by the interaction of insecticide treatment and maturity group in NC (data from both years combined) and SC in 2013, an effect not observed by Blount et al. (2016). Generally in our study, early maturing soybeans

(MGs IV and V) treated with insecticide had higher yields compared to untreated soybeans within the same MGs. This effect was not as consistent with later maturity groups. Earlier maturing soybeans, by definition, reach reproductive status sooner, leaving less time for compensation. This, combined with the stress of M. cribraria feeding throughout the season might have affected plant performance, and ultimately yield, in untreated plots.

There was a positive correlation between soybean yield loss and high numbers of M. cribraria (adults and nymphs) in five of the seven locations in this study. Even when we combined data from all trials in a single analysis (data not shown), the correlation between these two factors was still positive and with a R2 of 0.14. There was no relationship between soybean yield and insect densities in experiments where relatively lower numbers of M. cribraria occurred with relatively high-yielding soybean plots (Montgomery County, NC,

2012 and Scotland County, NC, 2013). Plants can compensate for insect herbivory (Trumble et al. 1993), and soybeans have the capacity to compensate for yield under stressed conditions, including injury caused by insects (Ball et al. 2014). Furthermore, there is likely an interaction among yield potential of the soybean crop, M. cribraria density, and duration of the infestation. We expected to see a reduction on soybean yield in experimental plots with high levels of M. cribraria, where those infestations lasted throughout the season. In the NC

experiments with relatively high yields (ranging from 3,334.64 to 3,601.15 kg/ha), densities of M. cribraria recorded in two locations of this study were very low and always under the proposed action threshold by Seiter et al. (2015a) of one nymph per sweep. Finally, we use the term “relatively high yielding”, because yields were compared to our lower yielding trials

(from 1,509.47 to 3,050.16 kg/ha). The average soybean yield in NC and SC during 2012 and

2013 was 2,471 and 2,084 kg/ha respectively (USDA 2015). Hence, most fields in these areas will likely not exhibit good yield compensation under M. cribraria infestations at or near threshold; since average yield is below to our “relative high yielding” field classification.

This study clearly demonstrates that early planted soybeans are at high risk of having infestations of M. cribraria, a principle corroborated with another study (Blount et al. 2016).

Historically, southeastern US soybeans planted early in the planting window resulted in reduced susceptibility to late-season defoliators. However, early planting now will put soybeans at a higher risk for infestation and yield loss due to M. cribraria if the species continues to be a prominent pest early in the season. Manipulating planting date to manage

M. cribraria could be an important cultural control for this pest in soybeans. Planting soybeans during the middle of May or later will ensure lower M. cribraria infestations, compared to an earlier planting. Lowering the risk of higher infestation of M. cribraria in soybeans will aid to reduce insecticide applications in this crop. Reducing insecticide active ingredient per hectare will decrease the negative effects of whole-field applications on natural enemies (Higley and Boethel 1994). Reducing insecticide applications will also alleviate the cost of labor, equipment, and supplies during the growing season, ultimately increasing farmer’s profits. Modifying the planting date of soybeans aiming to control M.

cribraria might impact yields, if the planting happens outside the proposed and recommended planting window by each region. Selection of different maturity groups might help to compensate for any loss of heat units (Heatherly and Hodges 1999), if planting date would be modified to manage M. cribraria in soybeans.

The mechanisms behind how M. cribraria is attracted to early planted soybeans are not fully understood. Future research should investigate dispersal patterns of this insect (from overwintering sites to soybean fields and movement within soybean fields) in order to elucidate the mechanism(s) for large populations of M. cribraria in early planted soybeans.

Quality of host plant might be another factor influencing infestation levels of M. cribraria in soybeans. A complementary choice-test could be conducted in the field, where M. cribraria coming from overwintering sites would be concurrently exposed to soybean plants at different developmental stages, ranging from late-vegetative to early reproductive stages.

Furthermore, it is also possible that endosymbionts of M. cribraria play an important role in this insect’s “ability” to obtain essential nutrients for survival from a soybean plant. How these endosymbiont bacteria support M. cribraria growth is still unknown, however,

Hosokawa et al. (2006) hypothesized that the endosymbionts might provide essential amino acids and vitamins to the insect.

This study also indicates that planting date and insecticide protection can be manipulated to influence populations of M. cribraria in soybeans. Changing planting date to manage an insect pest in a crop should be examined cautiously. For example, manipulation of planting date and maturity group in soybean may impact yield, independent of M. cribraria densities, as soybean yield potential is influenced by environmental factors such as temperature, rainfall, and day length. The only formal recommendation to alter planting date

for insect management in soybean has been proposed under trap-cropping programs to use early planted soybeans for managing stink bugs (McPherson and Newsom 1984). Hence, ideal planting date and maturity group should be selected considering the geographical region and other factors that aimed for the best plant performance. When planting soybeans earlier in the Southeast, it is important to consider that those fields may be at higher risk of having economically damaging infestations of M. cribraria.

Acknowledgements

Drs. Nick Seiter, Jeremy Greene, and Francis Reay-Jones (Clemson University) are co-authors of this publication. Authors would like to thank T.G. Gibson (NC) for providing access to experimental plots. Dan Mott, Steven Roberson, Clifton Moore, David Morrison,

Jeremy Martin, Eric Willbanks, Brad Fritz (NC State University), James Smoak, and Dan

Robinson (Clemson University) are gratefully acknowledged for their contribution to this research. We also thank Dr. Mark Abney and four anonymous reviewers for their comments on a previous manuscript. This project was funded by the NC Soybean Grower Association and a multi-institution grant received from Clemson University and funded by the United

Soybean Board. Initial observations on the effect of planting date on M. cribraria were provided by Drs. Philip Roberts and Jeremy Greene and were the genesis of these studies.

References

Asante, S. K., M. Tamo, and L.E.N. Jackai. 2001. Integrated management of cowpea

insect pests using elite cultivars, date of planting and minimum insecticide

application. African Crop Sci. J. 9: 655-665.

Ball, R. A., L. C. Purcell, and E. D. Vories. 2014. Short-season soybean yield

compensation in response to population and water regime. Crop Sci. 40: 1070-1078.

Baur, M. E., D. J. Boethel, M. L. Boyd, G. R. Bowers, M. O. Way, L. G. Heatherly, J.

Rabb, and L. Ashlock. 2000. Arthropod populations in early soybean production

systems in the mid-south. Environ. Entomol. 29: 312-328.

Bi, J., D. Lin, K. Lii, and N. C. Toscano. 2005. Impact of cotton planting date and nitrogen

fertilization on Bemisia argentifolii populations. Insect Sci. 12: 31-36.

Blount, J. L., G. D. Buntin, and P. M. Roberts. 2016. Effect of planting date and maturity

group on soybean yield response to injury by kudzu bug, Megacopta cribraria F.

(Hemiptera: Plataspidae). J. Econ. Entomol. 109: 207-212.

Board J. E., W. Zhang, and B. G. Harville. 1996. Yield rankings for soybean cultivars

grown in narrow and wide rows with late planting dates. Agron. J. 88: 240-245.

Bowers, G. R. 1995. An early soybean production system for drought avoidance. J. Prod.

Agric. 8: 112-118.

Boyd, M. L., D. J. Boethel, B. R. Leonard, R. J. Habetz, L. P. Brown, and W. B.

Hallmack. 1997. Seasonal abundance of arthropod populations on selected soybean

varieties grown in early season production systems in Louisiana. Agric. Exp. Station.

Bull. 860. Baton Rouge, LA.

Bradley, J. R., Jr. and J. W. Van Duyn. 1979. Insect pest management in North Carolina

soybeans, pp. 343-354. In: F. T. Corbin (ed.), World Soybean Research Conference

II: Proceedings. Westview Press, Boulder, CO.

Buschman L. L., H. N. Pitre, and H. F. Hodges. 1984. Soybean cultural practices: effects

on populations of geocorids, nabids, and other soybean arthropods. Environ. Entomol.

13: 305-317.

Del Pozo-Valdivia, A. I., and D. D. Reisig. 2013. First-generation Megacopta cribraria

(Hemiptera: Plataspidae) can develop on soybeans. J. Econ. Entomol. 106: 533-535.

Eger, J. E., Jr., L. M. Ames, D. R. Suiter, T. M. Jenkins, D. A. Rider, and S. E. Halbert.

2010. Occurrence of the Old World bug Megacopta cribraria (Fabricius)

(Heteroptera: Plataspidae) in Georgia: a serious home invader and potential legume

pest. Insecta Mundi. 0121: 1-11.

Fehr, W. R., C. E. Caviness, D. T. Burmood, and J. S. Pennington. 1971. Stage of

development descriptions for soybeans, Glycine max (L.) Merrill. Crop Sci. 11: 929-

931.

Frederick, J. R., P. J. Bauer, W. J. Busscher, and G. S. McCutcheon. 1998. Tillage

management for doublecropped soybeans grown in narrow and row width culture.

Crop Sci. 38: 755-762.

Gardner, W. A. 2015. Distribution map of the kudzu bug. University of Georgia. Center for

invasive species and ecosystem health.

http://www.kudzubug.org/distribution_map.cfm

Gore, J., C. A. Abel, J. J. Adamczyk, and G. Snodgrass. 2006. Influence of soybean

planting date and maturity group on stink bug (Heteroptera: Pentatomidae)

populations. Environ. Entomol. 35: 531-536.

Heatherly, L. G., G. Bowers, D. Boethel, M. Baur, J. Rabb, J. Rupe, J. Tyler, M. O.

Way, and L. Ashlock. 1998. Early soybean production system handbook. United

Soybean Board Technology Transfer publication. Chesterfield, MO.

Heatherly L. G. and H. F. Hodges. 1999. Soybean production in the midsouth. CRC

Press LLC. Boca Raton, FL.

Heatherly L. G. and S. R. Spurlock. 1999. Yield and economics of traditional and early

soybean production system (ESPS) seedings in the midsouthern United States. Field

Crop Res. 63: 33-45.

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: 909-918.

Higley, L. G., and D. J. Boethel. 1994. Handbook of soybean insect pests. Entomological

Society of America: Lanham, MD.

Hibino, Y., and Y. Itô. 1983. Mating aggregation of a stink bug, Megacopta punctissimun

(Montandon) (Heteroptera: Plataspidae). Res. Popul. Ecol. 25: 180-188.

Hosokawa, T., Y. Kikuchi, N. Nikoh, M. Shimada, and T. Fukatsu. 2006. Strict host-

symbiont cospeciation and reductive genome evolution in insect gut bacteria. PLOS

Biol. 4: e337.

Hosokawa, T., Y. Kikuchi, M. Shimada, and T. Fukatsu. 2007. Obligate symbiont

involved in pest status of host insect. Proc. R. Soc. B Biol. Sci. 274: 1979-1984.

Hu, M. 2013. Effect of late planting dates, maturity groups, and management systems, on

growth, development and yield of soybean in South Carolina. M.S. thesis. Clemson

University, Clemson, SC.

Kenward, M. G. and J. H. Roger. 1997. Small sample interference for ﬁxed effects from

restricted maximum likelihood. Biometrics 53: 983-997.

Khaled, O., E. Washington, P. Lage, S. K. Kantartzi, D. A. Lightfoot, and M. A.

Kassem. 2011. Comparison of early and conventional soybean production systems

for yield and other agronomic traits. Atlas J. Plant Biol. 1: 1-5.

Krishnan, H. B., S. S. Natarajan, J. O. Bennett, and R. C. Sicher. 2011. Protein and

metabolite composition of xylem sap from field-grown soybeans (Glycine max).

Planta. 233: 921–931

Littell, R. C., W. W. Stroup, G. A. Milliken, R. D. Wolfinger, and O. Schabenberger.

2006. SAS® for mixed models. Second Edition. SAS Institute, Cary, NC.

McPherson, R. M. and L. D. Newsom. 1984. Trap crops for control of stink bugs in

soybeans. J. Georgia Entomol. Soc. 19: 470-480.

McPherson, R. M. and K. Bondari. 1991. Influence of planting date and row width on

abundance of velvet bean caterpillars (Lepidoptera: Noctuidae) and southern green

stink bugs (Heteroptera: Pentatomidae) in soybean. J. Econ. Entomol. 84: 311-316.

McPherson, R. M., M. L. Wells, and C. S. Bundy. 2001. Impact of the early soybean

production system on arthropod pest populations in Georgia. Environ. Entomol. 30:

76-81.

Morril W. L. and G. D. Kushnak. 1999. Planting date influence on the wheat stem sawfly

(Hymenoptera: Cephidae) in spring wheat. J. Agric. Urban Entomol. 16: 123-128.

Parker, M. B., W. H. Marchant, and B. J. Mullinix, Jr. 1981. Date of planting and row

spacing effects on four soybean cultivars. Agron. J. 73: 759-762.

Pedersen, P. 2009. Soybean growth and development. Iowa State University Extension

Distribution Center. Publication PM1945. Ames. IA.

Ruberson, J. R., K. Takasu, G. D. Buntin, J. E. Eger Jr., W. A. Gardner, J. K. Greene,

T. M. Jenkins, W. A. Jones, D. M. Olson, P. M. Roberts, D. R. Suiter, and M. D.

Toews. 2013. From Asian curiosity to eruptive American pest: Megacopta cribraria

(Hemiptera: Plataspidae) and prospects for its biological control. Appl. Entomol Zool.

48: 3-13.

Ruppel, R. F. 1983. Cumulative insect-days as an index of crop protection. J. Econ.

Entomol. 76: 375-377.

SAS Institute. 2010. SAS® version 9.3 user’s manual. SAS Institute, Cary, NC.

Seiter, N. J., F.P.F. Reay-Jones, and J. K. Greene. 2013a. Within-field spatial distribution

of Megacopta cribraria (Hemiptera: Plataspidae) in soybean (Fabales: Fabaceae).

Environ. Entomol. 42: 1363-1374.

Seiter, N. J., J. K. Greene, and F.P.F. Reay-Jones. 2013b. Reduction of soybean yield

components by Megacopta cribraria (Hemiptera: Plataspidae). J. Econ. Entomol.

106: 1676-1683.

Seiter, N. J., A. I. Del Pozo-Valdivia, J. K. Greene, F.P.F. Reay-Jones, P. M. Roberts,

and D. D. Reisig. 2015a. Action thresholds for managing Megacopta cribraria

(Hemiptera: Plataspidae) in soybean based on sweep-net sampling. J. Econ. Entomol.

108: 1818-1829.

Seiter, N. J., J. K. Greene, F.P.F. Reay-Jones, P. M. Roberts, and J. N. All. 2015b.

Insecticidal control of Megacopta cribraria (Hemiptera: Plataspidae) in soybean. J.

Entomol. Sci. 50: 263-283.

Slosser, J. E. 1993. Influence of planting date and insecticide treatment on insect pest

abundance and damage in dryland cotton. J. Econ. Entomol. 86: 1213-1222.

Smith, M. S. and T. J. Riley. 1992. Direct and interactive effects of planting date, irrigation,

and com earworm (Lepidoptera: Noctuidae) damage on aflatoxin production in

preharvest field corn. J. Econ. Entomol. 85: 998-1006.

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

composition, and population dynamics of stink bugs in production fields of early and

late soybean in south Arkansas. J. Econ. Entomol. 102: 229-236.

Suiter, D. R., J. E. Eger, Jr., W. A. Gardner, R. C. Kemerait, J. N. All, P. M. Roberts, J.

K. Greene, L. M. Ames, G. D. Buntin, T. M. Jenkins, and G. K. Douce. 2010.

Discovery and distribution of Megacopta cribraria (Hemiptera: Heteroptera:

Plataspidae) in northeast Georgia. J. Integr. Pest Manage. 1: F1-F4.

Trumble, J. T., D, M. Kolodny-Hirsch, and I. P. Ting. 1993. Plant compensation for

arthropod herbivory. Annu. Rev. Entomol. 38: 93-119.

Tugwell, P., E. P. Rouse, and R. G. Thompson. 1973. Insects in soybeans and a weed host

(Desmodium sp.). Arkansas Exp. Station. Report Series 214. Fayetteville, AR.

(USDA) U.S. Department of Agriculture-National Agricultural Statistics Service. 2015.

Crop production 2014 Summary. USDA, Beltsville, MD.

Walter, A. J. and C. D. DiFonzo. 2007. Soil potassium deficiency affects soybean phloem

nitrogen and soybean aphid populations. Environ. Entomol. 36: 26-33.

Zhang, Y., J. L. Hanula, and S. Horn. 2012. The biology and preliminary host range of

Megacopta cribraria (Heteroptera: Plataspidae) and its impact on kudzu growth.

Environ. Entomol. 41: 40-50.

Table 1. Analysis of variance results for influence of planting date and maturity group on presence of Megacopta cribraria egg masses and adult abundance in soybean plants at V5 growth stage or younger.

Location Response Variable Source of Variation df F P / Year Egg Masses / 0.61 m- Planting Date 2, 192 4.33 0.0145 row Maturity Group 3, 189 8.98 <0.0001 North Carolina Planting Date*Maturity Group 3, 189 7.51 <0.0001 2012 & Adults / 0.61 m-row Planting Date 2, 207 0.05 0.9509 2013 Maturity Group 3, 193.20 9.01 <0.0001 Planting Date*Maturity Group 6, 193.90 10.15 <0.0001 Egg Masses / 0.61 m- Planting Date 2, 77 226.31 <0.0001 South row Carolina Maturity Group 3, 77 1.81 0.1518 2012a Planting Date*Maturity Group 6, 77 0.64 0.6991 Egg Masses / 0.61 m- Planting Date 2, 1 214.15 0.0483 row South Maturity Group 3, 22.13 2.22 0.1138 Carolina Planting Date*Maturity Group 6, 1 10.84 0.2284 2013 Adults / 0.61 m-row Planting Date 2, 72 39.71 <0.0001

Maturity Group 3, 72 4.78 0.0205 Planting Date*Maturity Group 6, 72 2.33 0.0411 aAdult data is not included in this table because densities were not collected at this trial.

Table 2. Effect of planting date on significant response variables in South Carolina during

2012 and 2013. Each row in the table represents a separate statistical analysis. Means ± standard error (SE) sharing the same letters are not statistically different (α > 0.05).

Planting Date (Mean ± SE) Year Location Response Variable April May June / July 2012 South Egg Masses / 0.61 m-row 92.38 ± 5.57 A 25.41 ± 4.71 B 0.03 ± 0.03 C Carolina 2012 SC Yield (kg / ha) 1,244.40 ± 98.61 b 1,609.25 ± 89.62 a 1,674.76 ± 73.71 a Untreateda 2013 South Egg Masses / 0.61 m-row 122.50 ± 14.26 A 79.88 ± 14.08 A 0.31 ± 0.16 B Carolina 2013 South Cumulative Egg Masses / 28.31 ± 2.06 a 10.94 ± 1.03 b 4.56 ± 0.46 c Carolina 25 Trifoliates aIn 2012, there were two separated tests where one field was sprayed with insecticide and the adjacent field was left untreated.

Table 3. Analysis of variance for planting date and maturity group effects on Megacopta cribraria for insecticide treated and untreated 2012 South Carolina trials.

Trial Response Variable Source of Variation df F P Untreated Cumulative Egg Planting Date 2, 33 481.44 <0.0001 Masses / 25 Trifoliates Maturity Group 3, 33 2.86 0.0519 Planting Date*Maturity Group 6, 33 0.99 0.4511 Cumulative Insect Days Planting Date 2, 36 178.96 <0.0001 for Nymphs / Sweep Maturity Group 3, 36 1.31 0.2874 Planting Date*Maturity Group 6, 36 1.15 0.3517 Cumulative Insect Days Planting Date 2, 33 61.18 <0.0001 for Adults / Sweep Maturity Group 3, 33 2.03 0.1288 Planting Date*Maturity Group 6, 33 2.17 0.0709

Yield (kg / ha) Planting Date 2, 33 11.10 0.0002 Maturity Group 3, 33 5.76 0.0028 Planting Date*Maturity Group 6, 33 1.82 0.1258 Insecticide Cumulative Egg Planting Date 2, 33 113.97 <0.0001 Treated Masses / 25 Trifoliates Maturity Group 3, 33 2.43 0.0825 Planting Date*Maturity Group 6, 33 0.99 0.4488 Cumulative Insect Days Planting Date 2, 33 22.97 <0.0001 for Nymphs / Sweep Maturity Group 3, 33 1.53 0.2239 Planting Date*Maturity Group 6, 33 1.10 0.3839 Cumulative Insect Days Planting Date 2, 33 37.92 <0.0001 for Adults / Sweep Maturity Group 3, 33 2.23 0.1034 Planting Date*Maturity Group 6, 33 4.37 0.0024

Yield (kg / ha) Planting Date 2, 36 162.48 <0.0001 Maturity Group 3, 36 22.66 <0.0001 Planting Date*Maturity Group 6, 36 10.01 <0.0001

Table 4. Analysis of variance for planting date, maturity group, and insecticide effects on

Megacopta cribraria in South Carolina in 2013.

Response Variable Source of Variation df F P Cumulative Egg Masses / Insecticide 1, 66 2.19 0.1894 25 Trifoliates Maturity Group 3, 66 14.31 <0.0001 Insecticide*Maturity Group 3, 66 0.73 0.5389 Planting Date 2, 66 152.32 <0.0001 Planting Date*Insecticide 2, 66 1.84 0.1666 Planting Date*Maturity Group 6, 66 1.89 0.0961 Planting Date*Insecticide*Maturity Group 6, 66 0.15 0.9893 Cumulative Insect Days for Insecticide 1, 72 501.45 <0.0001 Nymphs / Sweep Maturity Group 3, 72 0.17 0.9156 Insecticide*Maturity Group 3, 72 1.38 0.2549 Planting Date 2, 72 2.67 0.0764 Planting Date*Insecticide 2, 72 15.87 <0.0001 Planting Date*Maturity Group 6, 72 0.27 0.9470 Planting Date*Insecticide*Maturity Group 6, 72 0.23 0.9674 Cumulative Insect Days for Insecticide 1, 6 444.55 <0.0001 Adults / Sweep Maturity Group 3, 66 5.10 0.0031 Insecticide*Maturity Group 3, 66 1.01 0.3941 Planting Date 2, 66 25.33 <0.0001 Planting Date*Insecticide 2, 66 8.06 0.0007 Planting Date*Maturity Group 6, 66 0.41 0.8667 Planting Date*Insecticide*Maturity Group 6, 66 0.46 0.8323 Yield (kg / ha) Insecticide 1, 3 96.28 0.0022 Maturity Group 3, 66 23.30 <0.0001 Insecticide*Maturity Group 3, 66 3.88 0.0129 Planting Date 2, 66 26.25 <0.0001 Planting Date*Insecticide 2, 66 8.07 0.0007 Planting Date*Maturity Group 6, 66 9.09 <0.0001 Planting Date*Insecticide*Maturity Group 6, 66 1.74 0.1264

Table 5. Analysis of variance results for planting date, maturity group, and insecticide effects on Megacopta cribraria egg mass numbers, cumulative insect days for nymphs and adults per sweep, and cumulative nymphs and adults per plant in North Carolina using combined data from 2012 and 2013.

Response Variable Source of Variation df F P Cumulative Egg Masses / 25 Planting Date 2, 177 207.36 <0.0001 Trifoliates Maturity Group 3, 177 4.15 0.0071 Planting Date*Maturity Group 6, 177 0.63 0.7036 Insecticide 1, 180 131.84 <0.0001 Insecticide*Planting Date 2, 180 15.37 <0.0001 Insecticide*Maturity Group 3, 180 0.11 0.9526 Insecticide*Planting Date*Maturity Group 6, 180 0.29 0.9425 Cumulative Insect Days for Planting Date 2, 345 3.09 0.0469 Nymphs / Sweep Maturity Group 3, 345 0.66 0.5796 Planting Date*Maturity Group 6, 345 1.02 0.4130 Insecticide 1, 345 98.62 <0.0001 Insecticide*Planting Date 2, 345 5.21 0.0059 Insecticide*Maturity Group 3, 345 0.69 0.5560 Insecticide*Planting Date*Maturity Group 6, 345 0.96 0.4492 Cumulative Insect Days for Planting Date 2, 165 102.44 <0.0001 Adults / Sweep Maturity Group 3, 165 1.98 0.1183 Planting Date*Maturity Group 6, 165 1.82 0.0990 Insecticide 1, 180 302.41 <0.0001 Insecticide*Planting Date 2, 180 4.94 0.0081 Insecticide*Maturity Group 3, 180 2.16 0.0939 Insecticide*Planting Date*Maturity Group 6, 180 1.19 0.3160 Cumulative Nymphs / Plant Planting Date 2, 345 13.60 <0.0001 Maturity Group 3, 345 4.90 0.0024 Planting Date*Maturity Group 6, 345 1.26 0.2772 Insecticide 1, 345 141.13 <0.0001 Insecticide*Planting Date 2, 345 47.67 <0.0001 Insecticide*Maturity Group 3, 345 0.88 0.4538 Insecticide*Planting Date*Maturity Group 6, 345 0.69 0.6583 Cumulative Adults / Plant Planting Date 2, 165 104.29 <0.0001 Maturity Group 3, 165 0.87 0.4597 Planting Date*Maturity Group 6, 165 0.72 0.6324 Insecticide 1, 180 168.11 <0.0001

Table 5. Continued

Insecticide*Planting Date 2, 180 7.06 0.0011 Insecticide*Maturity Group 3, 180 0.22 0.8790 Insecticide*Planting Date*Maturity Group 6, 180 0.42 0.8660 Yield (kg / ha) Planting Date 2, 165 18.93 <0.0001 Maturity Group 3, 165 3.22 0.0241 Planting Date*Maturity Group 6, 165 3.20 0.0054 Insecticide 1, 180 20.19 <0.0001 Insecticide*Planting Date 2, 180 0.57 0.5650 Insecticide*Maturity Group 3, 180 4.60 0.0040 Insecticide*Planting Date*Maturity Group 6, 180 1.61 0.1463

Table 6. Significant effect of maturity group (Roman numerals) on significant response variables in North and South Carolina during 2012 and 2013, using Megacopta cribraria and soybean yield data. Each row in the table represents a separate statistical analysis. Means ± standard error (SE) sharing the same letters are not statistically different (α > 0.05).

Maturity Group (Mean ± SE) Year Location Response Variable IV V VI VII 2012 & North Carolina Cumulative Egg 29.90 ± 3.99 A 21.24 ± 3.32 B 24.91 ± 3.01 AB 22.95 ± 2.70 AB 2013 Masses / 25 Trifoliates 2012 & North Carolina Cumulative 0.95 ± 0.19 a 0.60 ± 0.19 b 0.66 ± 0.20 b 0.70 ± 0.14 ab 2013 Nymphs / Plant

2012 SC Untreateda Yield (kg / ha) 1,303.39 ± 107.47 B 1,746.74 ± 138.99 A 1,415.11 ± 55.82 B 1,572.64 ± 105.17 AB

2013 South Carolina Cumulative Egg 13.42 ± 1.78 b 22.13 ± 3.59 a 11.46 ± 2.19 c 11.42 ± 1.78 bc Masses / 25 Trifoliates

2013 South Carolina Cumulative Insect 570.88 ± 129.23 A 661.03 ± 184.08 A 513.36 ± 141.53 AB 265.75 ± 53.91 B Days / Sweep for Adults aIn 2012, there were two separated tests where one field was sprayed with insecticide and the adjacent field was left untreated.

Table 7. Regression analyses between soybean yield (kg/ha) [response variable] and Megacopta cribraria (cumulative adults and nymphs/sweep) [independent variable] in NC and SC locations for 2012 and 2013. Minimum (Min), maximum (Max), mean, and standard error (SE) values are also presented for each variable.

Regression Analysis Year Location Variable Min Max Mean SE F P R2 Equation

2012 Montgomery Co. Yield (kg/ha) 1,626.81 4,864.26 3,344.64 70.73 0.001 0.9722 0.001 Yield=3,350.04-11.09*log10(adults+nymphs/sweep) NC Adults + nymphs / 0.10 11.30 2.40 0.19 sweep Scotland Co. NC Yield (kg/ha) 675.53 4,084.26 2,576.67 83.61 15.250 0.0002 0.14 Yield=3,080.87-1,261.10*log10(adults+nymphs/sweep) Adults + nymphs / 0.05 6.70 1.83 0.16 sweep Barnwell Co. Yield (kg/ha) 613.02 2,357.76 1,509.47 56.83 7.310 0.0096 0.14 Yield= 2,072.50-141.83*log10(adults+nymphs/sweep) SC Adults + nymphs / 47.80 1007.90 453.36 38.48 (untreated) sweep Barnwell Co. Yield (kg/ha) 1270.83 3830.36 2585.73 122.32 13.350 0.0007 0.23 log10(Yield)=3.57-0.13*log10(adults+nymphs/sweep) SC 3.30 220.10 27.59 6.06 (insecticide Adults + nymphs / treated) sweep 2013 Montgomery Co. Yield (kg/ha) 226.49 3,224.21 1,909.70 63.71 11.840 0.0009 0.11 Yield=2,440.73-412.55*log10(adults+nymphs/sweep) NC Adults + nymphs / 1.80 80.50 18.37 1.55 sweep Scotland Co. NC Yield (kg/ha) 1,692.49 5,973.46 3,601.15 87.19 0.004 0.9522 0.001 log10(Yield)=3.54+0.001*log10(adults+nymphs/sweep) Adults + nymphs / 0.60 86.05 10.78 1.39 sweep Barnwell Co. SC Yield (kg/ha) 111.49 4,412.48 3,050.16 72.10 29.970 <0.0001 0.24 Yield=3,700.79-286.48*log10(adults+nymphs/sweep) Adults + nymphs / 2.00 380.70 69.18 8.64 sweep

Supplementary Table 1. Significant effect of planting date and maturity group (Roman numerals) on cumulative insect days for adults/sweep in the insecticide treated trial located in South Carolina and planted during April, May, and July of 2012. Means ± standard error (SE) sharing the same letters are not statistically different (α > 0.05).

Planting Date / Maturity Group (Mean ± SE)

April Response Variable IV V VI VII Cumulative Insect Days for 161.66 ± 85.48 112.39 ± 52.01 85.18 ± 9.26 114.98 ± 51.28 Adults / Sweep CD D D CD

May Response Variable IV V VI VII Cumulative Insect Days for 813.88 ± 542.87 92.50 ± 32.24 176.59 ± 98.68 131.63 ± 58.20 Adults / Sweep AB D B-D CD

July Response Variable IV V VI VII Cumulative Insect Days for 351.44 ± 99.42 A- 968.71 ± 468.43 301.16 ± 53.63 1,112.61 ± 807.91 Adults / Sweep C A A-D A

Supplementary Table 2. Significant effect of planting date and insecticide use regime on soybean yield (kg/ha) in South Carolina during 2013. Means ± standard error (SE) sharing the same letters are not statistically different (α > 0.05).

Planting Date / Insecticide Regime (Mean ± SE) April May June / July Response Variable Untreated Treated Untreated Treated Untreated Treated

Yield (kg / ha) 2,650.9 ± 253.04 3,657.14 ± 137.94 2,914.30 ± 61.67 3,687.80 ± 81.86 2,536.55 ± 103.38 2,854.26 ± 97.99 BC A B A C BC

Fig. 1. Significant effect of the interaction between planting date (X-axis) and insecticide use

(treated plots = black bars, untreated plots = white bars) on Megacopta cribraria egg masses

(panels on the left column), on cumulative insect days for nymphs per sweep (panels on the center column) and cumulative insect days for adults / sweep (panels on the right column) from soybean plants older than V5 growth stage in North and South Carolina during 2012 and 2013.

Means sharing the same letters are not statistically different (α > 0.05). Separate analyses are presented at each cell level. Asterisk indicates that mean separation was not performed during the analysis because the interaction between planting date and maturity group (single asterisk), and the effect of maturity group (double asterisk) had an effect on the analyzed variables, rather than the effect of planting date alone. The mean separation for the interaction and the single effect are presented in-text and in Table 4 respectively.

Fig. 2. Significant effects of the interaction between planting date and maturity group on

Megacopta cribraria egg masses laid in soybean at V5 growth stage or younger (black bars, upper case letters) and adults (grey bars, lower case letters) in North Carolina during 2012 and 2013 and South Carolina in 2013. Means sharing the same letters are not statistically different (α > 0.05). Separate analyses are presented at each row level. The asterisk at the lower panel indicates that separation of means was not performed for egg masses during the analysis of the interaction because the effect of planting date, rather than maturity group, was significant on M. cribraria egg mass abundance. Mean separations for this effect are presented in Table 2.

Fig. 3. Significant effects of the interaction between planting date (X-axis) and insecticide use (treated plots = black bars, untreated plots = white bars) on Megacopta cribraria nymphs

(upper panel) and adults (lower panel) per soybean plant older than V5 growth stage in North

Carolina during 2012 and 2013. Means sharing the same letters are not statistically different

(α > 0.05). Separate analyses are presented at each row level.

Fig. 4. Significant effect of the interaction between planting date and maturity group (roman numerals on X-axis) on soybean yield in North and South Carolina during 2012 and 2013.

Means sharing the same letters are not statistically different (α > 0.05). Separate analyses are presented at each row level.

Fig. 5. Significant effect of the interaction between maturity group (Roman numerals) and insecticide regime (X-axis) on soybean yield in North Carolina during 2012 and 2013 and

South Carolina in 2013. Means sharing the same letters are not statistically different (α >

0.05). Separate analyses are presented at each row level.

CHAPTER FIVE

Interaction of Tillage, Maturity Group, and Insecticide Use on Megacopta cribraria

(Hemiptera: Plataspidae) Populations in Double Cropped Soybean

Abstract

Megacopta cribraria (F.), also known as the kudzu bug, is a soybean pest in the U.S., and can cause up to a 60% yield reduction if not controlled. Insecticides are commonly used to manage this pest in commercial soybean fields. However, other soybean production practices might also impact kudzu bug populations. This study investigated the effect of soil tillage, maturity group selection, and insecticide use on kudzu bug densities in soybean.

During 2012 and 2013, at two locations each year in NC, varieties of four different soybean maturity groups were planted during June into conventionally tilled plots and into plots with cereal crop residue under reduced tillage conditions (mimicking double crop production).

Plots were further split as insecticide protected and untreated. Four times more kudzu bugs were found in conventionally tilled than reduced till plots throughout the growing season.

Population densities of kudzu bug were inconsistent across maturity group for either of the tillage systems in this study. A 56% reduction of kudzu bug densities was achieved through insecticide treatment, with a nearly 6% increase in yield. Yield effects between tillage systems were confounded by kudzu bug density and other tillage effects; they were higher when kudzu bug density was below the threshold of one insect (adult and/or nymph) per sweep, but were lower when kudzu bug populations were above this threshold. Information on how production practices, including soil tillage, affect kudzu bug populations in soybean may help growers select practices to minimize kudzu bug injury and protect yield.

Introduction

Megacopta cribraria (F.) (Hemiptera: Plataspidae), also known as kudzu bug, is an invasive soybean pest in the Southern U.S. (Suiter et al. 2010). First reported in nine counties in Georgia during 2009, the kudzu bug has rapidly spread across 10 states, including new reports in Florida, Louisiana, Arkansas, Tennessee, Kentucky, and Virginia during 2015

(Gardner 2015). Originally found feeding on kudzu, Pueraria montana Loureiro (Merrill) var. lobata (Willdenow), the kudzu bug was also identified feeding on other legumes, including Wisteria spp.; pigeon pea, Cajanus cajan (L.) Millsp; lima bean, Phaseolus lunatus

L.; pinto bean, Phaseolus vulgaris L.; and mung beans Vigna radiata (L.) R. Wilczek (Zhang et al. 2012, Medal et al. 2013, Blount et al. 2015, Golec et al. 2015).

The insect is thought to be bivoltine, undergoing two in-field generations when suitable host plants are present (Zhang et al. 2012, Seiter et al. 2013). Kudzu bug overwinters under tree bark, in leaf litter, and inside human-made buildings and structures (Suiter et al.

2010). Females coming from overwintering sites will either mate or directly lay eggs in early spring on kudzu (Golec and Hu 2015), a major host in addition to soybean. It has also been shown that the first-generation of kudzu bug can develop on soybean (Del Pozo-Valdivia and

Reisig 2013) and a portion of the overwintering adult population can potentially bypass kudzu if early soybean is present. First-generation adults emerge during late spring and infest kudzu and soybean during summer (Seiter et al. 2013). The more economically damaging second-generation of this insect is generally found during late summer on soybean (Seiter et al. 2013). Kudzu bugs disperse to overwintering sites during late fall when preferred host plants senesce (Suiter et al. 2010). Other factors that could potentially impact kudzu bug

flight to overwintering quarters and diapause, including photoperiod and temperature, have yet to be investigated.

Kudzu bug represents a threat to soybean production in the southern U.S. when left unmanaged, as it has been shown to reduce soybean yield up to 60% by reducing seed weight and number of seeds per pod (Seiter et al. 2013). Currently, whole-field insecticide applications are the primary tactic used to control this pest, and treatment is advised when the recommended action threshold of one nymph per sweep is reached (Seiter et al. 2015a).

Selected cultural practices, such as planting date, also affect population dynamics of this insect in soybean (Blount et al. 2016, Del Pozo-Valdivia et al. 2016). Earlier planted soybean accumulates more insect days (higher insect populations on soybean for longer periods of time), compared with soybean planted later (Del Pozo-Valdivia et al. 2016). It is suggested that early planted soybean plants harbor more kudzu bugs because they are one of the first hosts available for these insects in the spring (Del Pozo-Valdivia et al. 2016). Furthermore, other characteristics of soybean are not known to influence kudzu bug densities. For example, populations of kudzu bug do not consistently vary among different soybean maturity groups (Blount et al. 2016, Del Pozo-Valdivia et al. 2016); maturity groups (MG) represent categorizations of the number of days until bloom under specific photoperiod conditions (Pedersen 2009).

Planting date and varietal selection are not the only cultural practices that influence insect abundance in crops, including soybean (Higley and Boethel 1994). For example, southern U.S. soybean growers have narrowed row spacings over time, reducing the risk of corn earworm, Helicoverpa zea (Boddie) (Lepidoptera: Noctuidae), infestation (Bradley and

Van Duyn 1979). Higher numbers of corn earworms occur in soybean planted on wide rows,

defined as row-spacing at > ~0.76 m. One hypothesis is that canopy architecture influences oviposition. Canopy architecture can influence microclimate, especially when it closes the row middles at which time leaves and branches cover most of the soil surface, effectively intercepting the sun light, preventing soil moisture loss, and, importantly, shading weeds

(Heatherly and Hodges 1999). In the case of corn earworm, the lack of canopy closure on wide rows during flowering makes the crop more attractive for oviposition (Bradley and Van

Duyn 1979).

Soil tillage can also influence the population dynamics of arthropods in crops, including soybean (Stinner and House 1990). There are many different types of tillage employed by soybean growers in the southeastern U.S. The conventional till system is defined as the physical manipulation of the soil to control weeds, incorporate herbicides and fertilizers, and to prepare the seedbed for a good seed-soil contact and uniform plant emergence (Heatherly and Hodges 1999, Singh 2010). On the other hand, in a complete no- till system, all primary tillage for seedbed preparation is removed; planting seeds and weed management under this system is more complex than in the conventional till system

(Heatherly and Hodges 1999, Singh 2010). Reduced tillage systems are intermediate, with residues of the previous crop generally covering 30% or more of the soil surface; examples include ridge-till or mulch-till, where either the planting row or the whole-field is disturbed with fewer tillage passes than conventional tillage (Conservation Technology Information

Center 2002). The double cropping system, where soybean is planted immediately following the winter grain harvest, generally during June, is considered to be an example of conservational or reduced tillage (Frederick et al. 1998) and is common in the southeastern

U.S. In this system, relatively late maturing soybean varieties, such as MG VI and VII are

usually planted (Frederick et al. 1998, Heatherly and Hodges 1999) so that enough vegetative growth can be achieved before flowering.

There are many examples demonstrating how tillage practices influence insect population dynamics in soybean, with examples of positive, negative, and neutral effects. For instance, more potato leafhoppers Empoasca fabae (Harris) (Hemiptera: Cicadellidae)

(Hammond and Stinner 1987, Smith et al. 1988), seed corn maggots Delia platura (Meigen)

(Diptera: Anthomyiidae) (Hammond 1997), and striped flea beetles Phyllotreta striolata (F.)

(Coleoptera: Chrysomelidae) (Smith et al 1988) are found in conventionally tilled soybean compared with no-till soybean. In contrast, more tarnished plant bugs Lygus lineolaris

(Palisot de Beauvois) (Hemiptera: Miridae), bean leaf beetles Cerotoma trifurcata (Förster)

(Coleoptera: Chrysomelidae) (Troxclair and Boethel 1984, Smith et al. 1988), damsel bugs

Nabis spp. (Hemiptera: Nabidae), and predatory ground beetles (Coleoptera: Carabidae)

(House and Stinner 1983) are found in no-till soybean compared with conventionally tilled soybean. In other cases, tillage does not influence the abundance of some soybean insect pests. For example, population densities of green cloverworm, Hypena scabra (F.)

(Lepidoptera: Erebidae), are similar in conventionally tilled and no-till soybeans (Hammond and Stinner 1987, Lam and Pedigo 1998). Similarly, bigeyed bugs, Geocoris spp.

(Hemiptera: Geocoridae), and some damsel bugs Nabis spp., occur in equal frequency independent of tillage in soybean (Funderburk et al. 1988, Lam and Pedigo 1998). However, with exceptions, the general pattern across crops is that insect pest abundance increases when the soil disturbance increases through tillage (Stinner and House 1990).

The main goal of this study was to determine if tillage practices influence kudzu bug population densities in soybean. In this study, I compared soybean planted in conventionally

tilled plots (disc-tilled) with soybean planted into cereal residue under reduced tillage, a typical way that soybean is planted in a double cropping system in the southeastern U.S. I also tested the interactions of tillage with soybean maturity group and insecticide use

(multiple spray versus untreated) for impacts on kudzu bug populations and yield in soybean.

The potential effect of tillage on kudzu bug was measured before and after soybean canopy closure. Finally, I also explored how the interaction between tillage type and levels of kudzu bugs influenced soybean yield in these studies.

Materials and Methods

Plot Information. The first experimental site was located at the North Carolina State

University Sandhills Research Station, near Jackson Springs, Montgomery County, NC and the second site was at a commercial soybean field near Gibson, Scotland County, NC during both 2012 and 2013. There were two separate, but contiguous fields at each experimental site

(Fig. 1). The first experimental site consisted of a field that was conventionally tilled using a disc-harrow (Frontier, model DH1615, Deere and Co., Moline, IL). The conventionally tilled field was adjacent to a second field that had been planted with cereal rye, Secale cereale L., during October of the previous year. The cereal rye was harvested in late May during both years. After harvest, standing cereal rye was mowed, using a rotary cutter (John Deere, model MX6, Deere and Co., Moline, IL) to produce a stubble height of ~10 cm. After mowing, a thick cereal straw residue was left on this field, with residue covering ~85% of the ground. The second experimental site also consisted of both a conventionally and disc-tilled field and an adjacent field planted with wheat, Triticum aestivum L., during October of the previous year. Wheat was harvested in early June during both years, and the straw was also

mowed to produce a stubble height and residue coverage similar to the first experimental site.

Both fields, conventionally tilled and reduced tillage with crop residue, at each experimental location were set up as a split-plot design with four replications, where the main plot was maturity group (MG) IV (variety AG4531, Asgrow, Monsanto Company, St. Louis, MO), V

(AG5503), VI (AG6132), and VII (AG7502), and the split-plot was insecticide treatment

(aggressively protected or unsprayed). Experimental plots in each field were four-rows wide by 12.2 m long, and were planted with 0.97 m row-spacing using a four-row cone planter

(John Deere, model 1750, Deere and Co., Moline, IL) in Montgomery County, at 29 seeds per row meter. In Scotland County, experimental plots were planted with 0.91 m row-spacing using a two-row disc vacuum planter (White, model 6700, AGCO Corporation, Duluth, GA), also at 29 seeds per row meter. Planting occurred between 18-19 June 2012 and 17-20 June

2013 at both locations.

The two insecticide treatment regimens were either aggressively protected, which meant that they were sprayed every two weeks with insecticide, or left unsprayed. The first insecticide treatment was triggered when adult kudzu bug densities initially reached 0.5 to

1.0 adults per sweep in a single maturity group averaged across the four replications.

Bifenthrin (Discipline 2EC, AMVAC Chemical Corp., Los Angeles, CA) was applied at 0.11 kg/ha of active ingredient, in a volume of 93.5 l/ha, using a two-row CO2 backpack sprayer with TX-10 hollow-cone nozzles (Teejet, Wheaton, IL) at 40 psi. A total of two and three insecticide applications were made during 2012 and 2013, respectively. During late July

2013, heliothine pest larvae densities reached the published economic threshold at both locations. To control this potential yield-limiting factor, flubendiamide (0.11 kg/ha of the active ingredient Belt, Bayer CropScience LP, Research Triangle Park, NC) was applied to

both entire fields at each location, using the same equipment and conditions as those used in applications for the aggressive insecticide treatment targeted for kudzu bug. Flubendiamide was chosen since it has no known effect on kudzu bugs when applied at the selected rate

(Seiter et al. 2015b).

Kudzu Bug Sampling Prior to Soybean Canopy Closure. Whole-plants were visually inspected for kudzu bugs at 14 and 28 days after planting. Six visual samples were randomly taken in each plot and consisted of an inspection of all soybean plants located within a 0.61m-row. During these inspections, the total number of soybean plants, kudzu bug adults and egg masses were counted.

Kudzu Bug Sampling After Soybean Canopy Closure. Sweep-net samples were taken every other week, once plants reached V5 (Fehr et al. 1971), and sweeping continued until plants reached R7. At each sampling date, 20 sweeps were taken in each plot using a

0.38m diameter sweep-net. A single soybean row was sampled; the net was plunged into the canopy and the entire net was submerged below the top of plants within the row. During sweeping, the number of kudzu bug adults and nymphs, and other insect pests, such as defoliators and stink bugs, were documented in the field. To monitor the presence of egg masses, 25 newly unfolded top soybean trifoliates were randomly selected and visually inspected in each plot. Sweep-net samples and trifoliates were initially taken from rows two and three, respectively. Sample locations were switched at the following sampling date, where sweep-nets were taken from row three and trifoliates from row two. Location of samples alternated in this manner at each sampling date until plants reached R7. This resulted in sweep-net samplings being taken in a given location no more than once monthly.

Harvest Information. Soybean was harvested at both locations on 4-5 November

2012 and 6-7 November 2013. The middle two rows of each plot were harvested to calculate yield (kg/ha), using a two-row mechanical plot combine harvester (Gleaner model K2,

AGCO Corporation, Duluth, GA). Directly following harvest, seed shattering was measured in the first three replications of each treatment by collecting and weighing seed from a 0.5 ×

0.5 m random area within the plot using a plastic frame. Seed weight was not corrected because seed shattering, expressed as seed weight, was not different among maturity groups in any given year or location (data not presented). Soybean yield (in kilograms) and moisture content (percentage) was measured to determine yield/ha adjusted to 13% seed moisture content.

Data Analysis. Data from insect evaluations and soybean yield were pooled together and organized by ‘trial’ (n=4), where each trial represented two paired fields (conventionally tilled and reduced tillage) in one location during one year (location = Montgomery and

Scotland Counties, year = 2012 and 2013). To account for the effect of tillage, which was not randomized at each location, the tillage factor was nested within each trial and, subsequently, replication was nested within tillage.

Visual inspections of adults and egg masses before soybean canopy closure, were expressed as individuals/meter-row. The first two visual evaluations (14 and 28 DAP) were combined into one data set and analyzed using a generalized linear mixed model approach

(PROC GLIMMIX, SAS version 9.3, SAS Institute 2010, Cary, NC). Separate individual analyses of variances (ANOVA) were conducted using adults/meter-row and egg masses/meter-row as response variables. Treatments for the first two visual evaluations were organized as a randomized complete block design since insecticide applications (split-plots)

were not deployed early in the season. Tillage, maturity group, and the interaction between these two factors were the fixed effects in our model. A Poisson distribution was selected to describe both adult and egg mass data, based on model fit and data distribution criteria

(Littell et al. 2006). The procedure of Kenward and Roger (1997) was used to calculate the degrees of freedom in our models, as an adjustment for the estimator of variance of fixed effects. The effects of trial alone, and replication nested within tillage and trial (tillage × trial) were considered as random. After the ANOVA analysis for the visual inspection data, mean separation of treatment effects was performed using the Tukey’s test at α ≤ 0.05.

Cumulative insect days for nymphs/sweep and adults/sweep after canopy closure were calculated (Ruppel 1983) and indicated the magnitude and duration of the kudzu bug infestation in each treatment. Sweep data, expressed as insect days, and cumulative egg mass counts after soybean canopy closure were log10-transformed [log10(X+1)] to comply with the assumptions of the ANOVA. Individual mixed model ANOVAs (PROC MIXED, SAS

Institute 2010) were calculated for each of the following response variables: soybean yield

(kg/ha), log10-transformed cumulative M. cribraria egg masses/trifoliate, log10-transformed cumulative M. cribraria days for nymphs/sweep, and log10-transformed cumulative M. cribraria days for adults/sweep.

In each statistical model, fixed effects were tillage, maturity group, insecticide regime, and their interactions. Random effects were trial alone, replication nested with tillage and trial (tillage × trial), and the interaction between replication nested with tillage and trial by maturity group. The Kenward-Roger’s procedure (1997) was used to calculate degrees of freedom. Mean separation post-ANOVA of the transformed data was performed using the

Tukey’s test at α ≤ 0.05. Means and standard errors are reported from the back-transformed data.

Data were partitioned using a one bug/sweep threshold, the action threshold confirmed by Seiter et al. (2015a); densities below this threshold are expected not to reduce yield. Combined data were partitioned into two data sets where the first set included soybean yield when plots had one or more kudzu bug adults and/or nymphs per sweep or ‘above threshold’. The second set included plots with fewer than one kudzu bug adult and/or nymph per sweep or ‘below threshold’. Using these data sets, an analysis of covariance (ANCOVA) was run using the PROC GLM procedure (SAS Institute 2010), where soybean yield was the response variable in our model, cumulative kudzu bug number (adults and/or nymphs per sweep) was the covariate, and tillage was the independent variable. To test the similarity of intercepts of the regressions from my ANCOVA, the interaction term between insect densities × tillage type was not included in a separate model. Transformed [log10(X + 1)] insect data were used to comply with assumptions of the ANCOVA analysis.

Results

Kudzu Bug Populations Prior to Soybean Canopy Closure. Kudzu bug population densities were influenced by tillage before plant canopy closure (≤ V5, Table 1). More adults and egg masses per one-meter row were found on soybean plants under conventionally tilled conditions, compared with plots under reduced tillage (Table 2). Maturity group did not influence kudzu bug population densities before the plant canopy closure under either tillage condition (Table 1).

Kudzu Bug Populations After Soybean Canopy Closure. Abundance of kudzu bug egg masses was affected by tillage, maturity group, and insecticide regime when soybean plants were at the V6 growth stage or older (Table 1). More egg masses/trifoliate were documented in conventionally tilled plots (Table 2), in MG IV plots (Table 3), and in untreated plots (Table 4), compared with plots under reduced tillage conditions (Table 2),

MG V and VI (Table 3), and insecticide-treated plots (Table 4), respectively. Number of egg masses on soybean MG VII did not differ significantly from the other maturity groups in this study (Table 3). Nymph densities were different among tillage conditions and insecticide regime (Table 1). More insect days for nymphs/sweep accrued in untreated soybean planted in conventionally tilled plots, compared with insecticide-treated soybean planted in conventionally tilled plots (Fig. 2). Insect days for nymphs/sweep were similar from plots under reduced tillage conditions, regardless of the insecticide treatment, and were also similar to plots that were conventionally tilled and insecticide-treated (Fig. 2). Tillage and insecticide affected the cumulative insect days for adults/sweep when soybean plants were

V6 or older (Table 1). There were more cumulative insect days for adults/sweep in soybean planted into conventionally tilled plots (Table 2) and from untreated soybean (Table 4), compared with soybean planted into reduced tillage plots (Table 2) and insecticide-treated plots (Table 4).

Soybean Yield. Stink bugs and defoliators were not present in significant numbers

(defined as reaching the economic threshold and data not presented) in any experiment and potentially-damaging heliothine pests were managed using a selective insecticide application that did not impact kudzu bugs. Tillage, maturity group, and insecticide regime influenced yield (Table 1). Conventionally tilled (Table 2), MG VII (Table 3) and insecticide-treated

soybean yielded higher (Table 4), compared with reduced tillage (Table 2), MG IV and V

(Table 3), and untreated soybean (Table 4). MG VI soybean yields were similar to the other maturity groups in this study (Table 3). When used as a covariate, level of kudzu bug infestation, grouped as above or below the selected threshold, did not affect soybean yield under any of the tillage systems (F = 0.23, P = 0.6344). Moreover, the slopes of the regressions between soybean yield and tillage system were similar between the two groupings (above and below the selected threshold) of kudzu bug infestation (tillage system

× kudzu bug infestation levels: F = 3.19; P = 0.0751). The lack of interaction (similar slopes) indicates that tillage effect on yield was independent from kudzu bug densities when they were grouped as above or below an action threshold of one bug per sweep. Consistent with a significant main effect of tillage on yield, the intercepts for these two regressions were different (F = 5.14; P = 0.0242); the intercept when plots had kudzu bug densities above threshold was 214.12 kg/ha less than the intercept when plots had kudzu bug densities below threshold (Fig. 3).

Discussion

Kudzu bug infestation duration and intensity, measured indirectly using cumulative insect days, was approximately six times higher in conventionally tilled plots compared with plots under reduced tillage conditions. These differences were noted when soybean was planted into different tillage types both early, before canopy closure (V5 growth stage or younger), as well as later, after canopy closure (V6 or older). Perhaps most striking, was that cumulative insect days for nymphs/sweep in untreated reduced tillage plots were low (0.25) and did not differ from those in insecticide-treated plots in either the reduced tillage or the

conventionally tilled plots (Fig. 2). The explanatory mechanism for this phenomenon with tillage is not clear and cannot be addressed using the data from these experiments.

Tillage type and surface cultivation promotes soil aeration, retention of water from rainfall, and mineralization in the soil, where nitrate availability can also increase (House and

Stinner 1983, Martin et al. 2006, Singh 2010). Higher concentrations of nitrates in the soil might affect concentrations of nitrogen in the vascular tissue of plants (House and Stinner

1983); however, understanding the impacts of soil nitrogen in a plant such as soybean may be more difficult than in other systems, since bacteria in root nodules can fix nitrogen for utilization by the soybean plant (Singh 2010). A higher concentration of nitrogen in soybean phloem provides enhanced nutrition to sap feeders, increasing the population growth such as in aphids (Walter and DiFonzo 2007). Consequently, a similar phenomenon may occur with kudzu bug, which likely feeds on phloem, assuming that the soybean plants under conventional tillage were more nitrogen-rich. For instance, results from this study indicated that numbers of egg masses/row-meter were 14.3 times higher in conventionally tilled plots, compared with reduced tillage plots. Furthermore, numbers of egg masses/trifoliate were 3 times higher in conventionally tilled plots, compared with reduced tillage plots. Another potential explanation for the differences in insect densities between tillage systems is that plants in conventionally tilled plots could have been healthier and with a greater biomass, compared with plants in reduced tillage plots. Assuming that more kudzu bug adults in early season implied that soybean plants were more attractive, this attraction may have enhanced their oviposition and or development in conventionally tilled plots, compared to those in reduced tillage plots.

The soil incorporation of crop residue by tillage modifies the light reflectance of the ground surface, due to the lack of crop residue (Nagler et al. 2000) and the exposure of soil particles to direct sunlight (Stoner and Baumgardner 1981, Weidong et al. 2002). Insects, especially some hemipteran herbivores like aphids, are visually attracted to specific light spectra (Costello 1995, Döring and Chittka 2007). The modification of light reflectance of the soil by tillage may be one of the explanations why kudzu bug is more attracted to conventional till plots, compared with plots under reduced tillage. Crop residue on top of the ground may affect soil reflectance, influencing how kudzu bugs alight on the soil.

Alternatively, crop residue could also be a physical barrier to immigrating insects, including kudzu bugs, attempting to crawl up and locate suitable host plants (Sarrantonio and Gallandt

2003). Crop residue can increase background reflectance (Summers et al. 2004), and

“camouflage” the reflectance of soybean plants against this background. In the specific case of double cropped soybean, winter cereal stubble and straw left on the field after harvest may act as a physical barrier early in the season to reduce kudzu bug colonization, movement, and establishment. In my study, the seedling soybean plants were less apparent in the reduced tillage fields compared to the conventionally tilled fields. It is possible that dispersing kudzu bugs visually cue toward conventionally tilled plots early in the season. After soybean canopy closure, light reflectance from the ground becomes more difficult to detect, since plant vegetation is a physical barrier that blocks the soil reflectance. In this study, the effect of tillage on kudzu bugs persisted throughout the season. Hence, I speculate that tillage impacts plant growth and architecture, in addition to influencing light spectra signatures and nutritional qualities of the plant, such as amino acid content. These other potential effects of tillage on soybean plants may have influenced kudzu bug populations later in the season,

when bigger canopies shadowed the crop residue on the experimental plots. Future research could explore the effect of the light reflectance coming from the soil, with and without crop residue, and how it might impact attractiveness of kudzu bug; the variation in plant nitrogen titers in plants grown in conventionally and reduced tillage systems should also be explored in relation to kudzu bug attraction and growth.

Having two adjacent fields with two different tillage systems at each location facilitated soil cultivation in this experiment. However, the physical layout of the tillage factor might have influenced the in-field distribution and migration of adult kudzu bugs among experimental plots. Ideally, tillage treatments should have been completely randomized within the same field. My experimental design was constrained by limitations in the ability to randomize the tillage treatments. To alleviate this issue, the tillage factor was double-nested with trial and replication for statistical analysis. As a consequence of the nesting arrangement, the degrees of freedom for the F test were reduced when tillage type was statistically analyzed as a factor. Reducing the degrees of freedom on the denominator in the F test also lowered the chances to falsely detect a significant effect of tillage on kudzu bug in this study. However, despite the loss of degrees of freedom, the effect of tillage was still significant for all the factors measured.

Although insecticide and tillage had consistent impacts on kudzu bug density and oviposition, maturity group influenced only kudzu bug oviposition when plants were at the

V6 growth stage or older. There were more egg masses per trifoliate in the MG IV and MG

VII varieties compared to the other varieties at these stages. Similar to previous studies

(Blount et al. 2016, Del Pozo-Valdivia et al. 2016), patterns of egg mass numbers were not consistent across maturity groups. Moreover, maturity group did not impact kudzu bug

density (as measured by cumulative insect days) before or after canopy closure. It is possible that varieties with vegetative periods on the extremes, either shorter, as in the MG IV variety, or longer, as in the MG VII variety, could have stimulated more oviposition compared to other maturity groups (MG V and VI). For instance, a full canopy may have been available earlier and more attractive for oviposition from MG IV plants, compared with other MGs.

The duration of the vegetative periods in soybean may be linked to how much nitrogen would be available in the phloem, indirectly influencing host plant quality for the kudzu bug.

However, I did not have more than one variety of each maturity group in this study to evaluate this hypothesis. Hence, the effects seen could have masked a possible varietal effect, since only one variety was used as a proxy for plant growth habit. Moreover, the MG IV variety was indeterminate, producing nodes until beginning of pod fill, in contrast to the MG

V, VI, and VII determinate varieties used in this experiment, that stop vegetative growth soon after flowering starts (Pedersen 2009). Soybean growth habits (determinate or indeterminate) may have influenced plant canopy, and potentially affected the attractiveness of the kudzu bug to host plant to oviposit.

Multiple insecticide applications were an effective tactic to manage populations of kudzu bugs in soybean, regardless of the tillage system. Insecticide treatment resulted in fewer egg masses and cumulative adult insect days. Several applications of insecticides throughout the season provided long-term protection against kudzu bug injury during this study. Insecticide treatments seemed to be unnecessary in the reduced tillage plots (Fig. 2), but based on the selected kudzu bug threshold for this experiment (0.5 to 1.0 adults per sweep), treatments were called for in the conventionally tilled plots. This aggressive control tactic in our experiment was solely utilized for research purposes. Spraying insecticide to

100

control the kudzu bug multiple times during the growing season will increase production costs. Broad spectrum insecticides, like the bifenthrin used in this study, negatively impact beneficial insects (Table 2 in Bradley and Van Duyn 1979), potentially leading to damaging infestations of secondary pests that cause economic damage in soybean (Kogan and

Turnipseed 1987). Insect monitoring through scouting will provide the information about population densities of the kudzu bug in this crop. Seiter et al. (2015a) suggested that a single insecticide application targeting nymphs was sufficient and cost-effective to prevent soybean yield reduction. Adding to this, my results demonstrate that kudzu bug numbers can be reduced as much as an aggressive insecticide management program by planting soybean into cereal grain residue.

Soybean yield was affected by tillage system, maturity group, and insecticide use.

Conventionally tilled plots yielded higher, from MG VII soybean and from insecticide treated plots, compared with plots under reduced tillage, MG IV and V, and untreated plots.

As mentioned earlier, soil tillage can increase mineralization and the concentration of nitrates in the soil (Martin et al. 2006, Singh 2010). Soybean plant performance could have improved due to the presence of additional available nitrogen in the soil released from tillage

(Heatherly and Hodges 1999, Martin et al. 2006, Singh 2010). However, in other environments, especially in sandy soils, soybean can have similar yields when cultivated under conventionally tilled and long-term no-till conditions (Singh 2010). My experiment did not address the effect of long-term no-till conditions on yield, and my plots were located where cereal rye or wheat was planted during the season before the soybean was planted.

Maturity group is also an important variable that influences yield in soybean

(Heatherly and Hodges 1999, Singh 2010). Under the same environment, a MG V variety

101

will reach the reproductive stages faster than a MG VII variety, since it is adapted to shorter photoperiods (Pedersen 2009). Longer vegetative growth, such the ones exhibited by the MG

VII variety in this study, should enhance the duration of photosynthetic activity and, therefore, increase yields (Singh 2010). In this study, kudzu bug injury also influenced yield in soybean, when insecticide treated plots were compared to untreated ones. Insecticide applications reduced kudzu bug pressure and therefore injury, which is known to ameliorate yield loss in soybean (Seiter et al. 2013, Blount et al. 2016, Del Pozo-Valdivia et al. 2016).

Higher soybean yields in this study were recorded not only from conventionally tilled plots, but also from plots with fewer kudzu bugs where insecticide was applied. In contrast, conventionally tilled plots yielded higher but also had more kudzu bugs than reduced tillage plots. Results from the ANCOVA test showed that the effect of tillage on soybean yield was independent from kudzu bug densities. Soybean yield from conventional tilled plots were

18.3% higher than yield from reduced till plots. However, there was a reduction of 214.11 kg/ha on soybean yield when kudzu bug populations were above the selected threshold of one bug/sweep, compared with plots where kudzu bug was below the threshold, representing a 9.8% reduction across tillage systems, regardless of tillage treatment. Kudzu bug injury to soybean has been reported in previous studies (Seiter et al. 2013, Blount et al. 2016, Del

Pozo-Valdivia et al. 2016). However, this experiment demonstrates that potential kudzu bug injury to soybean, impacting on yield, could be modulated by other factors in the system, such as the effect of tillage.

Because there is a higher probability of kudzu bug infestation in soybean when it is grown under a conventional tillage system, modifying tillage practices might be an option to manage kudzu bug populations. However, other outcomes from modifying the tillage system

102

should be also considered in the analysis. Soil tillage will influence seed germination, weed growth, and ultimately influence yield in soybean (Heatherly and Hodges 1999, Martin et al.

2006, Singh 2010). Several benefits from reduced tillage include reducing soil erosion and superficial water run-off (Heatherly and Hodges 1999, Martin et al. 2006). The modification of any production practice should not compromise the yield potential or net income received for a crop. It will be important to know which additional production practices influence kudzu bug populations in soybean. Early planting (Blount et al. 2016, Del Pozo-Valdivia et al. 2016), and conventional tillage will favor conditions for kudzu bug infestation in soybean.

Hence, more intensive scouting efforts should be directed to those fields with these characteristics. Future research could corroborate the impact of tillage and planting date combined with kudzu bug population dynamics in soybean.

Acknowledgements

Author would like to thank T.G. Gibson in NC for providing access to experimental plots. Dan Mott, Steven Roberson, Clifton Moore, David Morrison, Jeremy Martin, Eric

Willbanks, and Brad Fritz (NC State University), are gratefully acknowledged for their contribution to this research. Author also thank Dr. Consuelo Arellano (NC State University) for statistical advice, and Dr. Nick Seiter (University of Arkansas) for his comments on a previous version of this manuscript. This project was funded by the NC Soybean Grower

Association and by the United Soybean Board.

103

References

Blount, J. L., G. D. Buntin, and P. M. Roberts. 2016. Effect of planting date and maturity

group on soybean yield response to injury by kudzu bug, Megacopta cribraria F.

(Hemiptera: Plataspidae). J. Econ. Entomol. 109: 207-212.

Bradley, J. R., Jr. and J. W. Van Duyn. 1979. Insect pest management in North Carolina

soybeans, pp. 343-354. In: F. T. Corbin (ed.), World Soybean Research Conference

II: Proceedings. Westview Press, Boulder, CO.

Conservational Technology Information Center. 2002. Tillage type definitions.

http://www.ctic.purdue.edu/resourcedisplay/322/ (accessed 11 March 2016).

Costello, M. J. 1995. Spectral reflectance from a broccoli crop with vegetation or soil as

background: influence on immigration by Brevicoryne brassicae and Myzus persicae.

Ent. Exp. Appl. 75: 105-114.

Del Pozo-Valdivia, A. I., and D. D. Reisig. 2013. First-generation Megacopta cribraria

(Hemiptera: Plataspidae) can develop on soybeans. J. Econ. Entomol. 106: 533-535.

Del Pozo-Valdivia, A. I., N. J. Seiter, D. D. Reisig, J. K. Greene, F. P. F. Reay-Jones,

and J. S. Bacheler. 2016. Megacopta cribraria (Hemiptera: Plataspidae) population

dynamics in soybeans as influenced by planting date, maturity group, and insecticide

use. J. Econ. Entomol. 109: 1141-1155.

Döring, T. F. and L. Chittka. 2007. Visual ecology of aphids—a critical review on the role

of colours in host finding. Arthropod-Plant Interactions. 1: 3-16.

Fehr, W. R., C. E. Caviness, D. T. Burmood, and J. S. Pennington. 1971. Stage of

development descriptions for soybeans, Glycine max (L.) Merrill. Crop Sci. 11: 929-

931.

104

Frederick, J. R., P. J. Bauer, W. J. Busscher, and G. S. McCutcheon. 1998. Tillage

management for doublecropped soybeans grown in narrow and row width culture.

Crop Sci. 38: 755–762.

Funderburk, J. E., D. L. Wright, and I. D. Tear. 1988. Preplant tillage effects on

population dynamics of soybean insect predators. Crop Sci. 28: 973-977.

Gardner, W. A. 2015. Distribution map of the kudzu bug. University of Georgia. Center for

invasive species and ecosystem health.

http://www.kudzubug.org/distribution_map.cfm (accessed 9 December 2015).

Golec, J. R., X. P. Hu, L. Yang, and J. E. Eger. 2015. Kudzu-deprived first-generation

Megacopta cribraria (F.) (Heteroptera: Plataspidae) are capable of developing on

alternative legume species. J. Agric. Urban Entomol. 31: 52–61.

Golec, J. R. and X. P. Hu. 2015. Preoverwintering copulation and female ratio bias: life

history characteristics contributing to the invasiveness and rapid spread of Megacopta

cribraria (Hemiptera: Plataspidae). Environ. Entomol. 44: 411-417.

Hammond, R. B. and B. R. Stinner. 1987. Soybean foliage insects in conservation tillage

systems: effects of tillage, previous cropping history, and soil insecticide application.

Environ. Entomol. 16: 524-531.

Hammond, R. B. 1997. Long-term conservation tillage studies: impact of no-till on seedcorn

maggot (Diptera: Anthomyiidae). Crop Prot. 16: 221-225.

Heatherly L. G. and H. F. Hodges. 1999. Soybean production in the midsouth. CRC Press

LLC. Boca Raton, FL.

Higley, L. G., and D. J. Boethel. 1994. Handbook of soybean insect pests. Entomological

Society of America. Lanham, MD.

105

House, G. J. and B. J. Stinner. 1983. Arthropods in no-tillage soybean agroecosystems:

community composition and ecosystem interactions. Environ. Manage. 7: 23-28.

Kenward, M. G. and J. H. Roger. 1997. Small sample interference for ﬁxed effects from

restricted maximum likelihood. Biometrics 53: 983-997.

Kogan M. and S. G. Turnipseed. 1987. Ecology and management of soybean arthropods.

Ann. Rev. Entomol. 32: 507-538.

Lam, W. F. and L. P. Pedigo. 1998. Response of soybean insect communities to row width

under crop-residue management systems. Environ. Entomol. 27: 1069-1079.

Littell, R. C., W. W. Stroup, G. A. Milliken, R. D. Wolfinger, and O. Schabenberger.

2006. SAS® for mixed models. Second Edition. SAS Institute, Cary, NC.

Martin, J. H., R. P. Waldren, and D. L. Stamp. 2002. Principles of field crop production.

Person Prentice Hall Inc. Upper Saddle River, NJ.

Medal, J., S. Halbert, T. Smith, and A. Santa Cruz. 2013. Suitability of selected plants to

the bean plataspid, Megacopta cribraria (Hemiptera: Plataspidae) in no-choice tests.

Fla. Entomol. 96: 631-633.

Nagler, P. L., C. S. T. Daughtry, and S. N. Goward. 2000. Plant litter and soil reflectance.

Remote Sens. Environ. 71: 207-25.

Pedersen, P. 2009. Soybean growth and development. Iowa State University Extension

Distribution Center. Publication PM1945. Ames. IA.

Ruppel, R. F. 1983. Cumulative insect-days as an index of crop protection. J. Econ.

Entomol. 76: 375-377.

SAS Institute. 2010. SAS® version 9.3 user’s manual. SAS Institute, Cary, NC.

106

Seiter, N. J., J. K. Greene, and F.P.F. Reay-Jones. 2013. Reduction of soybean yield

components by Megacopta cribraria (Hemiptera: Plataspidae). J. Econ. Entomol.

106: 1676-1683.

Seiter, N. J., A. I. Del Pozo-Valdivia, J. K. Greene, F.P.F. Reay-Jones, P. M. Roberts,

and D. D. Reisig. 2015a. Action thresholds for managing Megacopta cribraria

(Hemiptera: Plataspidae) in soybean based on sweep-net sampling. J. Econ. Entomol.

108: 1818-1829.

Seiter, N. J., J. K. Greene, F.P.F. Reay-Jones, P. M. Roberts, and J. N. All. 2015b.

Insecticidal control of Megacopta cribraria (Hemiptera: Plataspidae) in soybean. J.

Entomol. Sci. 50: 263-283.

Singh, G. 2010. The soybean: botany, production, and uses. CABI Editors. Wallingford, UK.

Smith A. W., R. B. Hammond and B. R. Stinner. 1988. Influence of rye-cover crop

management on soybean foliage arthropods. Environ. Entomol. 17: 109-114.

Stinner, B. R. and G. J. House. 1990. Arthropods and other invertebrates in conservation-

tillage agriculture. Annu. Rev. Entomol. 35: 299-318.

Stoner E. R. and M. F. Baumgardner. 1981. Characteristic Variations in Reflectance of

Surface Soils. Soil Sci. Soc. Am. J. 45: 1161-1165.

Suiter, D. R., J. E. Eger, Jr., W. A. Gardner, R. C. Kemerait, J. N. All, P. M. Roberts, J.

K. Greene, L. M. Ames, G. D. Buntin, T. M. Jenkins, and G. K. Douce. 2010.

Discovery and distribution of Megacopta cribraria (Hemiptera: Heteroptera:

Plataspidae) in northeast Georgia. J. Integr. Pest Manage. 1: F1-F4.

Sarrantonio, M. and E. Gallandt. 2003. The role of cover crops in North American

cropping systems. J. Crop Prod. 8: 53-74.

107

Summers, C. G., J. P. Mitchell, and J. J. Stapleton. 2004. Management of aphid-borne

viruses and Bemisia argentifolii (Homoptera: Aleyrodidae) in zucchini squash by

using UV reflective plastic and wheat straw mulches. Environ. Entomol. 33: 1447-

1457.

Troxclair, N. N., Jr. and D. J. Boethel. 1984. Influence of tillage practices and row spacing

on soybean insect populations in Louisiana. J. Econ. Entomol. 77: 1571-1579.

Walter, A. J. and C. D. DiFonzo. 2007. Soil potassium deficiency affects soybean phloem

nitrogen and soybean aphid populations. Environ. Entomol. 36: 26-33.

Weidong, L., F. Bareta, G. Xingfa, T. Qingxi, Z. Lanfen, and Z. Bing. 2002. Relating soil

surface moisture to reflectance. Remote Sens. Environ. 81: 238-246.

Zhang, Y., J. L. Hanula, and S. Horn. 2012. The biology and preliminary host range of

Megacopta cribraria (Heteroptera: Plataspidae) and its impact on kudzu growth.

Environ. Entomol. 41: 40-50.

108

Table 1. Analysis of variance results for the effects of tillage and maturity group on densities of Megacopta cribraria in soybean plants at V5 growth stage or younger, the effects of tillage, maturity group, and insecticide use on densities of M. cribraria in soybean plants V6 growth stage or older, and the effects of tillage, maturity group, and insecticide use on soybean yield in North Carolina using combined data from 2012 and 2013.

Plant Response Growth Source of Variation df F P Variable Stage Tillage 1, 30.07 70.42 <0.0001 Adults / ≤V5 Maturity Group meter-row 3, 1 22.11 0.1548 Tillage*Maturity Group 3, 1 3.92 0.3517 Egg Tillage 1, 24 15.22 0.0001 Masses / ≤V5 Maturity Group 3, 1 0.19 0.8923 meter-row Tillage*Maturity Group 3, 1 0.87 0.6388 Tillage 1, 27 39.36 <0.0001 Maturity Group 3, 210 4.36 0.0053 Cumulative Tillage*Maturity Group 3, 210 1.04 0.3739 Egg ≥V6 Masses / Insecticide 1, 210 7.34 0.0073 Trifoliate Insecticide*Tillage 1, 210 0.61 0.4364 Insecticide*Maturity Group 3, 210 0.12 0.9460 Insecticide*Tillage*Maturity Group 3, 210 0.34 0.7933 Tillage 1, 27 8.16 0.0081 Cumulative Maturity Group 3, 210 0.91 0.4372 Insect Tillage*Maturity Group 3, 210 1.18 0.3175 Days for ≥V6 Insecticide 1, 210 27.08 <0.0001 Nymphs / Insecticide*Tillage 1, 210 6.23 0.0134 Sweep Insecticide*Maturity Group 3, 210 2.58 0.0546 Insecticide*Tillage*Maturity Group 3, 210 2.39 0.0695 Tillage 1, 27 40.05 <0.0001 Cumulative Maturity Group 3, 210 1.62 0.1847 Insect Tillage*Maturity Group 3, 210 1.87 0.1354 Days for ≥V6 Insecticide 1, 210 99.6 <0.0001 Adults / Insecticide*Tillage 1, 210 2.25 0.1353 Sweep Insecticide*Maturity Group 3, 210 2.32 0.0769 Insecticide*Tillage*Maturity Group 3, 210 0.62 0.6005

109

Table 1. Continued

Tillage 1, 27 8.54 0.0069 Maturity Group 3, 90 5.11 0.0026 Tillage*Maturity Group 3, 90 0.73 0.5373 Yield (kg / --- ha) Insecticide 1, 120 9.57 0.0025 Insecticide *Tillage 1, 120 0.30 0.5877 Insecticide*Maturity Group 3, 120 1.50 0.2189 Insecticide*Tillage*Maturity Group 3, 120 1.14 0.3360

110

Table 2. Effect of tillage on significant response variables in North Carolina using combined data from 2012 and 2013. Each row in the table represents a separate statistical analysis.

Tillage (Mean ± SE)a Response Variable Reduced-till Till Adults / meter-row 0.78 ± 0.19 b 9.14 ± 1.28 a

Egg Masses / meter-row 0.03 ± 0.01 b 0.43 ± 0.07 a

Cumulative Egg Masses / 0.07 ± 0.01 b 0.23 ± 0.02 a Trifoliate

Cumulative Insect Days for 9.46 ± 0.86 b 51.80 ± 7.76 a Adults / Sweep

Yield (kg / ha) 2,086.94 ± 62.87 b 2,555.69 ± 66.01 a

aMeans ± standard error (SE) sharing the same letters are not statistically different (α > 0.05).

111

Table 3. Effect of maturity group on significant response variables in North Carolina using combined data from 2012 and 2013.

Each row in the table represents a separate statistical analysis.

Maturity Group (Mean ± SE)a Response Variable IV V VI VII Cumulative Egg Masses / 0.18 ± 0.02 a 0.12 ± 0.02 b 0.12 ± 0.02 b 0.17 ± 0.02 ab Trifoliate

Yield (kg / ha) 2,137.78 ± 117.44 b 2,215.22 ± 98.77 b 2,388.53 ± 81.95 ab 2,543.73 ± 71.82 a

aMeans ± standard error (SE) sharing the same letters are not statistically different (α > 0.05).

112

Table 4. Effect of insecticide regime on significant response variables in North Carolina using combined data from 2012 and 2013. Each row in the table represents a separate statistical analysis.

Insecticide (Mean ± SE)a Response Variable Treated Untreated Cumulative Egg Masses / 0.13 ± 0.02 b 0.17 ± 0.02 a Trifoliate

Cumulative Insect Days for 18.66 ± 3.66 b 42.30 ± 7.08 a Adults / Sweep

Yield (kg / ha) 2,388.29 ± 66.32 a 2,254.34 ± 68.60 b

aMeans ± standard error (SE) sharing the same letters are not statistically different (α > 0.05).

113

June: Till June: Reduced-till

V V VI VI VII VII IV IV IV IV V V VI VI VII VII

June: Till June: Reduced-till

IV IV VII VII V V VI VI IV IV VII VII V V VI VI

June: Till June: Reduced-till

VI VI IV IV V V VII VII VII VII VI VI IV IV V V

June: Till June: Reduced-till

V V IV IV VII VII VI VI V V IV IV VII VII VI VI

Fig. 1. Schematic representation of the experimental layout for each field at the Scotland

County location. Individual small cells indicate the position of each experimental plot planted in June under the two tillage conditions with four replications. Larger cells left blank indicate experimental plots not included in this study. In this example, the field on the left was conventionally disc-tilled and the one on the right had the previous crop residue left on top of the soil before planting soybean (reduced till). Roman numbers indicate the soybean maturity group planted at each plot. Insecticide was aggressively sprayed on plots shaded with grey and untreated controls were located at the white plots.

114

Fig. 2. Significant effect of the interaction among insecticide treatment regime (X-axis, treated plots = black bars, untreated plots = white bars) and tillage systems (in each panel) on

Megacopta cribraria cumulative insect days for nymphs per sweep from soybean plants older than V5 growth stage in North Carolina, using the combined data from 2012 and 2013.

Means sharing the same letters are not statistically different (α > 0.05).

115

Fig. 3. Scatter plots representing soybean yield (kg/ha, dependent variable) and tillage system (independent variable) using combined data from North Carolina in 2012 and 2013.

Soybean yield was divided in two separate data sets, based on the covariate Megacopta cribraria infestation levels: 1) yields when cumulative M. cribraria were less than 1 bug/sweep or ‘below threshold’ (crosses), and 2) yields when cumulative M. cribraria were more or equal to 1 bug/sweep or ‘above threshold’ (circles). Each data set was used for an analysis of covariance, where dotted line represent the analysis for the first data set, and solid line represent the analysis for the second data set. Slopes between the two regressions were similar (F = 3.19; P = 0.0751) and intercepts were different (F = 5.14; P = 0.0242).

116

CHAPTER SIX

Effect of Tillage and Row Spacing on Ground Cover, Light Reflectance,

and Megacopta cribraria (Hemiptera: Plataspidae)

Populations in Soybean

Abstract

The kudzu bug, Megacopta cribraria (F.) (Hemiptera: Plataspidae), is an exotic soybean pest in the U.S. Planting date, tillage, and insecticides can be used to manipulate kudzu bug density. The ultimate goal of the present study was to investigate whether kudzu bug density in soybean is influenced by tillage and row spacing. Experimental soybean plots were planted during May 2014 and 2015 at two NC locations, using three to four tillage types and two row spacings. Crop residue ranged across tillage types, generating experimental plots with almost no residue (conventionally tilled) to plots covered by cereal rye rolled on the soil surface. In addition to kudzu bug density, net reflectance from soil and crop residue was measured across tillage types using a spectroradiometer. Higher reflectance profiles were documented from plots with higher crop residue. There was a negative relationship between ground cover and kudzu bug density before soybean canopy closure; where ground cover influenced spectral indexes calculated from light reflectance. Kudzu bug densities in general, were not consistently influenced by row spacing. However, they were influenced by row spacing late in the scouting season during September. Further studies should focus on understanding the host-finding process of kudzu bug in soybean.

117

Introduction

Plant residue on the soil surface can be manipulated by soil tillage or by the presence of a cover crop (Martin et al. 2006). Conventional tillage buries most of the previous crop residue, leaving the soil with almost no cover (Martin et al. 2006). On the other hand, conservation tillage with the use of cover crops, favors the presence of plant residue on top of the soil (Conservational Technology Information Center, CTIC, 2002). Conservation tillage programs aim to leave at least 30% or more of crop residue, reducing cultivation passes and soil erosion (CTIC 2002). Increasing numbers of growers have implemented conservation tillage systems to grow several crops in the U.S., including soybean, Glycine max (L). Merr.

The availability of herbicides, herbicide tolerant soybean varieties, and specialized planting equipment has facilitated the shift towards these systems (Hartwig and Ammon 2002).

Soybean grown in a double cropping system also falls under the category of conservation tillage. In this system, seeds are planted during summer after harvesting the previous cereal crop (Heatherly and Hodges 1999). Stubble and straw from the cereal crop are left on the soil and soybean is planted no-till, with increasingly rare exceptions, such as growers who burn or till the stubble.

Cover crops are living ground covers, planted after a main crop, that are killed before the following crop is planted (Hartwig and Ammon 2002). Cover crops can reduce soil and water erosion, augmenting or retaining nitrogen in the soil, reducing soil compaction, and reducing weed pressure (Wyland et al 1996, Hartwig and Ammon 2002). Commonly planted cover crops are cereal rye, Secale cereale L., and various legumes, including red clover,

Trifolium pretense L. (Bottenberg et al. 1997, Sarrantonio and Gallandt 2003). The adoption of cover crops has been slower than the adoption of conservation tillage practices, due to

118

complex economic, biological, and operational issues of planting cover crops (Sarrantonio and Gallandt 2003). However, some organic production systems have adopted both conservation tillage and the use of cover crops. For example, to improve weed control, growers in the southeastern U.S. use roller-crimpers to terminate a cereal rye cover crop, before planting soybean under no-till conditions, to create a thicker layer of rye mulch

(Reberg-Horton et al. 2011).

Row spacing may influence the potential yield of some crops, including soybean

(Martin et al. 2006). Plant spacing will also determine plant canopy architecture, influencing the efficiency of sunlight capture, and ultimately affecting photosynthetic performance

(Board and Harville 1996, Andrade et al. 2002). In soybean, row spacings at 76.2 cm (30 inches) or below are recommended for planting from mid-May through June (Brower et al.

2000, Pedersen 2004), the time when most double cropped soybean is planted in the southeastern U.S. Soybean planted in narrower rows during this time period has a higher light interception ratio, and usually yields better than plants under wider row conditions

(Brower et al. 2000, Pedersen 2004).

Soil tillage has an effect on population densities of several arthropods and most of the research indicates that herbivore densities are higher under conventional tillage conditions compared with no-till systems (i.e., Stinner and House 1990). The use of cover crops affects arthropod densities in several crops. For instance, in Italy, fewer European corn borers,

Ostrina nubilalis Hübber (Lepidoptera: Crambidae), and aphids, Aphis fabae Scopoli

(Hemiptera: Aphididae), were found in corn planted following a cover crop (Hartwig and

Ammon 2002). Similarly, fewer economically damaging populations of lepidopteran pests were found in cotton planted into a cover crop, compared with cotton planted without a cover

119

crop (Tillman et al. 2004). Furthermore, fewer leafhoppers, Empoasca fabae (Harris)

(Hemiptera: Cicadellidae), were found in snap bean (Bottenberg et al. 1997) and in soybean

(Miklasiewicz and Hammond 2001) when plots were planted after a cereal rye or wheat cover, respectively, compared with fallow ground. On the other hand, the presence of crop residue from a previous cover crop in snap bean and soybean does not influence densities of bean leaf beetles, Cerotoma trifurcata (Föster) (Coleoptera: Chrysomelidae), western corn rootworm, Diabrotica virgifera virgifera LeConte (Coleoptera: Chrysomelidae) (Bottenberg et al. 1997), and generalist predators, including coccinellids (Coleoptera: Coccinellidae), nabids (Hemiptera: Nabidae), and chrysopids (Neuroptera: Chrysopidae) (Koch et al. 2012).

Row spacing can also impact arthropod densities in crops such as soybean. Seeds planted in narrow rows (less than 30 cm) can increase the numbers of green cloverworm,

Plathypena scabra F. (Lepidoptera: Noctuidae), (McPherson et al. 1988) velvetbean caterpillar, Anticarsia gemmatalis Hübner (Lepidoptera: Noctuidae), and southern green stink bug, Nezara viridula L. (Hemiptera: Pentatomidae) (McPherson and Bondari 1991), compared to wider rows. Furthermore, recent work conducted in Maryland has also shown that leaf-feeding herbivores, including green cloverworm and soybean looper, Pseudoplusia includens (Walker) (Lepidoptera: Noctuidae), are generally more abundant in narrow-spaced soybean (~19 cm), compared with wide-spaced soybean (~76 cm) (Buchanan et al. 2015).

Megacopta cribraria (F.) (Hemiptera: Plataspidae), the kudzu bug, has been an important pest in soybean throughout the southeastern U.S. since 2010 (Eger et al. 2010,

Suiter et al. 2010, Zhang et al. 2012, Seiter et al. 2013). Originally from Asia, kudzu bug was first reported feeding on kudzu, Pueraria montana Loureiro (Merr.) variety lobata

(Willdenow) near Atlanta, GA; however it can also feed and reproduce in other legumes,

120

including Wisteria spp. and pigeon pea Cajanus cajan (L.) Millsp. (Zhang et al. 2012, Blount et al. 2015). This insect has two generations during the year, and the second in-field generation (F2) is generally the most damaging generation if improperly managed (Zhang et al. 2015, Seiter et al. 2013, Fritz et al. 2016). In soybean, farmers are advised to spray an insecticide when populations reach the proposed action threshold of one nymph per sweep, using a conventional 38 cm diameter sweep net (Seiter et al. 2015). Uncontrolled populations of this insect can be damaging, reducing soybean yield up to 60% (Seiter et al. 2013). Early- planted soybean favors higher infestation of kudzu bug (Blount et al. 2016, Del Pozo-

Valdivia et al. 2016). Along with planting date, tillage also has an effect on kudzu bug densities in soybean. Conventionally tilled soybean has higher kudzu bug densities compared with soybean planted into winter rye or wheat, Triticum aestivum L., cereal stubble and straw

(A.I.D., unpublished data). Hence, soybean planted in the double cropping system with reduced tillage should have fewer kudzu bugs, compared with soybean planted using conventional tillage (A.I.D., unpublished data).

In the present study, I evaluated whether tillage and row spacing influence kudzu bug densities in soybean. I hypothesized that kudzu bug population density would be affected differently by tillage treatments; where tillage would result in different degrees of ground cover. Because differences in ground cover have been shown to affect densities of some insects (i.e. whiteflies, Bemisia argentifolii Bellows and Perring, Homoptera: Aleyrodidae, in zucchini squash, Cucurbita pepo L.) (Summers et al. 2004), I also measured light reflectance from the ground surface of my experimental plots. Additionally, I hypothesized that different row spacings could affect kudzu bug densities in soybean. The specific objectives of this study, therefore, were to evaluate the effects of different tillage treatments on kudzu bug

121

densities before and after soybean canopy closure, manipulate the amount of ground cover by tillage type, investigate the effect of ground cover on kudzu bug populations, document any relationship between light reflectance from experimental plots and ground cover, and ascertain the effect of row spacing on kudzu bug populations in soybean.

Materials and Methods

2014 Plot Information. Two experimental soybean fields were located at the Caswell

Research Station, near Kinston, Lenoir County, NC. One was designed to test the effect of tillage type and the other to test the effect of row spacing on kudzu bug densities. The first field was planted with cereal rye, Secale cereale L., as a cover crop, during November in

2013. Before planting soybean during 2014, rye residue was either tilled and soybean was planted, or soybean was planted directly into the residue. This experiment was a randomized complete block design with four replications and 14 m × 14 m plots. The main factor was tillage type with treatments including: 1) conventional tillage, 2) reduced-tillage or “double cropped”, 3) mowed cover crop, and 4) rolled cover crop. Cereal rye in selected conventional tilled plots was killed during March with glyphosate (0.39 liter of active ingredient per ha,

Roundup, Monsanto Company, St. Louis, MO), using a CO2 backpack sprayer with 11002 even flat-fan nozzles (Teejet, Wheaton, IL), at 30 psi, and with a volume of 22 liters per ha.

These plots were then disc tilled (John Deere Model 215, Deere and Co., Moline, IL) on the same day that soybean was planted. Before establishing the remaining treatments, standing cereal rye was killed using glyphosate (same conditions as mentioned before) and a high- clearance self-propelled sprayer (John Deere Model 6700, Deere and Co.) during the last week of April to avoid re-sprouting of the rye. To mimic the amount of residue left by a

122

double cropping system, cereal rye was mowed once using a flail mower (John Deere Model

115, Deere and Co.) before soybean was planted, leaving stubble at a height below 20 cm.

Cereal rye was mowed twice to deploy the third treatment, once right after the rye was killed in April, and again in May, before planting soybean. For the fourth treatment, standing cereal rye was rolled using a curved bar roller-crimper (I & J Manufacturing, Gap, PA) on the same day that soybean was planted, creating a stubble mulch. AG 5831 (Asgrow, Monsanto

Company), Roundup ready soybean variety from maturity group (MG) V, was planted using a no-till planter (John Deere Max Emerge 2 Plus Series, Deere and Co.), equipped with row cleaners (Yetter Manufacturing, Colchester, IL) and fluted discs on each planter box to ensure penetration through the mulch. Sixteen rows were planted in each plot on 13 May

2014, with a row spacing of 76.2 cm, and 21 seeds per row meter (310,000 plants per ha). A

2.5 m wide alley between adjacent plots was established on all sides of each plot. This alley contained the tillage treatment only, with no soybean planted.

The second 2014 experiment was established in a different field and was also arranged in a randomized complete block design, but with seven, instead of four replicates.

This field was conventionally tilled before planting a MG VI Roundup ready soybean (AG

6132, Asgrow, Monsanto Company). Experimental plots were 14 m × 14 m. The main factor was row spacing, with either narrow (17.78 cm) or wide (96.52 cm) rows. Soybean was planted on 21 May 2014, using a grain drill (John Deere Model 8300, Deere and Co.) or a cone planter (John Deere Model 7100, Deere and Co.) for the narrow and wide row treatments, respectively. The number of rows per plot in the narrow row treatment was 75 and was planted so that the target plant population was equivalent to 345,800 seeds per hectare; the wide row treatment had 14 rows per plot, and was also planted so that the target

123

plant population was equivalent to 345,800 seeds per hectare. Stand counts from a three meter section of the middle row of each plot were performed in this field 28 days after planting (DAP) to ensure that plant populations between the narrow and wide row treatments were similar. Based on these counts (data not shown), I randomly removed the equivalent of six plants per row-meter in the wide row treatment to adjust plant density to ~27 per row- meter. The final plant population of both treatments was equivalent to ~300,000 plants per hectare.

2015 Plot Information. Cereal rye was planted during November 2014 at two experimental locations, one at the Sandhills Research Station, near Jackson Springs,

Montgomery County, NC, and another at Caswell Research Station. Experiments at both locations were a split-plot design, with tillage type as the whole plot. Tillage treatments were:

1) conventional till, 2) reduced-till or ‘double cropped’, and 3) rolled rye. Although mowed rye was included during the 2014 experiment, it was dropped from the 2015 experiment because ground cover was similar to the rolled rye during 2014. The sub-plot was row spacing, with either narrow (38.1 cm) or wide (76.2 cm) rows. Experimental plots were 6.4 m wide by 12.2 m long, with 8 and 16 rows for the wide and narrow row treatments, respectively. The experiment at the Sandhills Research Station had seven replications while the trial at the Caswell Research Station had eight. All treatments were established using the same methods described for the 2014 trials. A single Roundup ready soybean variety from

MG V (AG 5533, Asgrow, Monsanto Company) was planted in both locations, using the same methods described for the 2014 trial. Stand counts (data not shown) indicated that plant population was similar between wide and narrow row treatments, equivalent to ~300,000 plants per hectare (narrow: 299,925.60 ± 7,270.70 and wide: 304,414.68 ± 4,593.52). There

124

was a 2 m wide separation alley surrounding plots, containing the tillage type treatment only, where no soybeans were planted.

Ground Cover. Percentage of ground cover was visually estimated during 2014 and

2015, for experiments where tillage was manipulated. Three randomly selected 1 m × 1m quadrats, delineated by a white PVC-pipe frame, were inspected on each plot to measure the ground cover. This was done directly after planting following the procedure described by the manual of sampling vegetation attributes (USDA and USDI 1999), which correspond to percentage of the cereal rye mulch material covering the soil surface contained within the

PVC frame.

Insect Monitoring Prior Canopy Closure. Densities of kudzu bugs were recorded using two methodologies. Whole-plant visual inspections were performed at 14 and 28 DAP.

During these inspections, adults and egg masses of kudzu bugs were counted from three randomly selected samples in each plot, where each sample consisted of a 0.61 meter section of the middle row. The second sampling methodology was sweep-netting. One sample of 20 sweeps, using a 38 cm of diameter sweep net, was performed on a single row (in the wide row spacing) or across two rows (in the narrow row spacing) from the middle of each plot at

49 and 56 DAP. Number of kudzu bug adults and nymphs were counted in each sweep net sample. To monitor the abundance of kudzu bug egg masses on the same day sweep net samples were taken, 15 randomly selected uppermost and fully-formed soybean trifoliates were visually inspected in each plots. Trifoliate samples were taken from a different row from where sweep net samples were taken.

Insect Monitoring After Canopy Closure. Kudzu bug densities were recorded from sweep net samples after canopy closure. The canopy was considered closed when canopies of

125

adjacent plants within and across rows lap such that the ground was completely shaded when the sun was directly overhead. One sample of 20 sweeps, using a sweep net, was taken in each plot every other week, from 63 DAP until plants reached R7 (Fehr et al. 1971), following the methodology used prior to canopy closure. The trifoliate sampling methodology used prior to canopy closure was also used during this time period.

Additionally, whole-plant visual inspections for kudzu bugs in the field were performed in the row spacing experiment during 2014. Fifteen random plants were selected in each plot and the numbers of nymphs and adults per plant were counted. These plant inspections began at 70 DAP and were performed every other week, until plants reached R7.

Field Spectrometry. Light reflectance was taken from the field using a portable hyperspectral spectroradiometer (FieldSpec Series, ASD Inc., Boulder, CO) during 2014.

The spectroradiometer measured reflectance between 350 to 2,500 nm and bands from 400 to

1,000 nm were sampled. Three random samples were taken inside each plot using the spectroradiometer, where both plants and the ground cover treatment were present. An additional sample was taken from the outside of plots in the separation alley, where there were no plants. Calibration with a white panel that reflected nearly 100% of the light was carried out at the beginning of each sampling date and at the end of each replication. Light spectral measurements were taken under clear sky (no clouds) and sunny conditions, between

09:00 and 11:00 hours, equivalent to the zenith of the sun. Measurements were taken twice during the growing season, once prior soybean canopy closure (on 11 June 2014) and after soybean canopy closure (30 July 2014). Additional light spectral measurements were taken on 16 June 2014 from kudzu located on North Carolina State University campus, Raleigh,

NC.

126

Harvest Information. Experimental soybean fields were harvested on 05 November

2014, 22 October 2015 (Sandhills Research Station), and 04 December 2015 (Caswell

Research Station) to calculate yield. Entire plots were harvested using a grain combine (CIH

Model 2144, CNH Industrial, London, UK) in both trials during 2014, and at the Caswell

Research Station trial during 2015. At the Sandhills Research Station during 2015, the middle two rows (wide row treatment) or middle four rows (narrow row treatment) of each plot were harvested, using a two-row mechanical plot combine harvester (Almaco, Nevada,

IA). Soybean seed weight and moisture content in the seed were measured in the field to determine kg/ha adjusted to 13% seed moisture content.

Data Analysis. The first two visual inspections (14 and 28 DAP) of adults and egg masses per meter-row, and number of egg masses per trifoliate prior soybean canopy closure, were not included in the analysis because no kudzu bug adults or eggs were recorded during

2014 and 2015 at any location. Cumulative numbers of kudzu bugs adults per sweep were calculated from samples taken at 49 and 56 DAP prior soybean canopy closure, and were log10-transformed [log10(X + 1)] to comply with the assumptions of the analysis of variance

(ANOVA). Log10-transformed adults/sweep was the response variable in a series of general linear mixed model analyses (PROC MIXED, SAS version 9.3, SAS Institute 2010, Cary,

NC). In the separate experiments investigating tillage type and row spacing were the only fixed effects and replication the only random effect in their respective analyses. In the analysis of the 2015 data model, location, tillage type, row spacing and their interactions were the fixed effects, while replication nested with location and replication nested with location × tillage type were the random effects. To further analyze a significant interaction

127

the SLICE option was included in the LSMEANS statement to test the effect of tillage type across the two locations.

Cumulative adults/sweep, nymphs/sweep, egg masses/trifoliate, adults/plant, and nymphs/plant were calculated from data collected after canopy closure, from 63 DAP until plants reached R7. Response variables, including log10-transformed adults/sweep, log10- transformed nymphs/sweep, log10-transformed egg masses/trifoliate, and soybean yield

(kg/ha) were analyzed in separate ANOVAs. Models were created and analyzed using a general linear mixed model (PROC MIXED; SAS Institute 2010); fixed and random effects were similar to the models using data before the canopy closure. Three additional ANOVA’s

(PROC MIXED; SAS Institute 2010) were used to analyze the effect of row spacing on kudzu bug density per plant during 2014. The response variables were log10-transformed adults/plant, log10-transformed nymphs/plant, and log10-transformed egg masses/trifoliate, respectively. The fixed effects in these models were sampling date, row spacing and their interaction, while replication was the only random effect. A repeated statement was included in the previous three analyses to account for the effect of sampling date, where the subject was the experimental plots. The covariance structure was selected as compound symmetry, since insect densities evaluations were performed every other week. To further analyze a significant interaction between sampling date and row spacing, the option SLICE was included in the LSMEANS statement to test the effect of row spacing across sampling dates.

The proportion of ground covered by the cereal rye mulch was analyzed using a general linear mixed model approach (PROC MIXED, SAS Institute 2010). Two separate individual ANOVAs were conducted using percentage of ground cover as a response variable from the 2014 and 2015 data. Tillage type was the only fixed effect in the model for the 2014

128

analysis. In the 2015 model, location, tillage type, and the interaction between these two factors were the fixed effects. Replication (2014) and replication nested with location (2015) were the only random effect in these two models. For all the statistical models, degrees of freedom were calculated following the procedure of Kenward and Roger (1997). Post hoc mean separation of the transformed data and soybean yield was performed using the Tukey’s test at α ≤ 0.05. Means and standard errors are reported from back-transformed data.

Correlation analyses were performed to elucidate the relationship (i.e. linear vs. non- linear) between ground cover (independent variables) and kudzu bug numbers (dependent variable), using the PROC CORR procedure (SAS Institute 2010). Transformed [log10(X +

1)] insect data were used to comply with assumptions of the correlation analysis, since distribution of calculated residuals was skewed. The first correlation analyzed the relationship between the percentage of ground cover and kudzu bug adult densities prior to soybean canopy closure. The second correlation analyzed the relationship between ground cover and kudzu bug adult densities after canopy closure. All correlation analyses used the

Pearson correlation coefficient (r) to determine significance at α = 0.05. If a correlation between two variables was significant, a linear regression analysis using PROC REG (SAS

Institute 2010) was performed to quantify the relationship.

Light spectral measurements from 2014 were averaged to generate a single reflectance spectrum profile for the inside (plants + ground cover) and the outside (ground cover only) of each experimental plot. Light reflectance was calculated using the process function (no derivative) from the ViewSpecPro 6.2.0 software (ASD Inc. 2008, Boulder,

CO). Reflectance profiles were plotted using the graphing option of the previously mentioned software to visually compare profiles among treatments. Four selected spectral indexes were

129

also calculated using light reflectance, based on previously known indexes that allow the quantification of aphid infestations in wheat (Luo et al. 2013). These spectral indexes were:

1) Normalized Difference Vegetation Index (NDVI), 2) Damage Sensitive Spectral Index 1

(DSSI 1), 3) Photochemical Reflectance Index (PRI), and 4) Aphid Index (AI). These indexes were used to detect differences of reflectance in my experimental plots that were associated with different tillage regimes that resulted in differences in ground cover on the soil surface. These indexes were calculated for reflectance measurements taken prior to, and after plant canopy closure, as well as from within each experimental plot (tillage treatment + soybean plants) and from outside experimental plots (separation alley with tillage treatment and no soybean plants). To test how ground cover and tillage type affected the calculated spectral indexes, separate Analyses of Covariance (ANCOVAs) using the general linear model approach (PROC GLM, SAS Institute 2010) were performed, where each spectral index was considered as a response variable, percentage of ground cover was the covariate, and tillage type was the independent variable. To investigate if both ground cover and spectral indexes affected kudzu bug adult densities, a second set of ANCOVAs tested the model where insect densities were the response variable, the percentage of ground cover was the covariate, and each spectral index were the independent variables. Log10-transformed insect data [log10(X + 1)] and log10-transformed spectral indexes were used to comply with assumptions of the ANCOVA analysis.

Results

Kudzu Bug Densities Prior Canopy Closure. Neither kudzu bug adults nor egg masses were observed during visual inspections of soybean plants at 14 and 28 DAP during

130

both 2014 and 2015. During 2014, kudzu bug adult densities in sweep-net sampling did not differ among tillage types (F = 2.49; df = 3, 9; P = 0.1263). However, numerically more adults were found in conventionally tilled plots (0.07 ± 0.01 adults/sweep), compared to plots where rye was rolled on top (0.04 ± 0.01). Additionally, adult densities did not differ between either row spacing using sweep-net sampling (F = 0.05; df = 1, 5; P = 0.8360); where numerically more adults/sweep (0.28 ± 0.08) were found on soybeans planted in wide rows, compared to soybean planted in narrow rows (0.23 ± 0.05).

During 2015, kudzu bug adult densities were affected by the interaction between location and tillage type (F = 0.0481; df = 2, 39; P = 0.0481). When the effect of location was controlled, there were more adults in the conventionally tilled plots (F = 19.92; P <

0.0001) (0.05 ± 0.02 adults/sweep), compared to either plots under reduced tillage conditions

(0.02 ± 0.01) or to plots with the rye rolled on top (0.01 ± 0.01). Additionally, adult densities were not influenced by row spacing during 2015 (F = 4.00; df = 1, 39; P = 0.0526). No kudzu bug egg masses on trifoliates or nymphs in sweep net samples were observed prior canopy closure during 2014 and 2015.

Kudzu Bug Densities After Canopy Closure. Soybean canopy closure was recorded on 29 July 2014 (Caswell), 03 August 2015 (Sandhills), and 04 August 2015 (Caswell). Egg masses and kudzu bug densities did not vary among tillage types during 2014 (egg masses/trifoliate: F = 2.06; df = 3, 9; P = 0.1756, nymphs/sweep: F = 1.52; df = 3, 9; P =

0.2739, adults/sweep: F = 1.16; df = 3, 12; P = 0.3640). In addition, kudzu bug densities did not vary between row spacings during 2014, from the sweep-net sampling (nymphs/sweep: F

= 2.64; df = 1, 10; P = 0.1351, adults/sweep: F = 0.81; df = 1, 5; P = 0.4100). However, number of egg masses/trifoliate was influenced by the sampling date in 2014 (F = 21.04; df =

131

6, 60; P < 0.0001), where more egg masses were found on 23 July 2014 (63 DAP) (0.21 ±

0.04 egg masses/trifoliate), compared to later sampling dates in August (05 August 2014:

0.05 ± 0.01, 30 August 2014: 0.003 ± 0.003), and September (17 September 2014: no egg masses).

Densities of kudzu bug nymphs and adults per plant were influenced by the interaction between sampling date and row spacing during 2014 (nymphs: F = 2.62; df = 6,

60; P = 0.0256, adults: F = 3.53; df = 6, 60; P = 0.0047). When controlling for sampling date, nymphs were affected by row spacing on soybean planted in narrow rows (F = 7.71; P

< 0.0001), compared with soybean planted in wide rows (F = 1.11; P = 0.3677). Nymph densities were higher on soybean planted in narrow rows, compared to soybean in wider rows (Fig. 1). Additionally, adult densities were affected by the two row spacing treatments

(narrow rows: F = 15.63; P < 0.0001, wide rows: F = 6.83, P < 0.0001), when controlling for the effect of sampling date. Similarly to nymph densities, more adults were found when soybean was planted in narrow rows, compared to soybean in wider rows (Fig. 1)

Number of kudzu bug egg masses/trifoliate and nymphs/sweep were only influenced by the experimental location during 2015 (egg masses: F = 68.18; df = 1, 78; P < 0.0001, nymphs: F = 5.32; df = 1, 39; P = 0.0259). More egg masses per trifoliate and more nymph per sweep were recorded in Sandhills (0.024 ± 0.004 eggs, 0.019 ± 0.004 nymphs), compared to Caswell Research Station (no egg masses recorded, 0.008 ± 0.002 nymphs). On the other hand, kudzu bug adults/sweep were affected by the three-way interaction among location, tillage type, and row spacing during 2015 (F = 4.71; df = 2, 39; P = 0.0148). Higher densities of kudzu bug adults were recorded from plots under conventional tillage, with wide rows, and located in the Sandhills Research Station, compared to plots with narrow rows, planted

132

either in Sandhills or Caswell Research Stations, and under reduced tillage or with rye rolled

(Fig 2).

Ground cover. Percentage of ground cover in experimental soybean plots was successfully manipulated by tillage during 2014 (F = 190.33; df = 3, 12; P < 0.0001), as measured by visual inspection in quadrats. During 2014, conventionally tilled plots had the least amount of ground cover, reduced-till plots had an intermediate amount of ground cover, and plots where the cereal rye was mowed or rolled had the greatest amount of ground cover with very little exposed soil (Fig. 3a). During 2015, there was an interaction of tillage type and location for percentage of ground cover (F = 41.92; df = 3, 161; P < 0.0001). As in 2014, the conventionally tilled plots had the least amount of ground cover at both locations, compared with reduced-till or rolled rye plots (Fig. 3b). The rolled rye treatment covered less of the soil at the Sandhills Research Station (~60%), while the same treatment at the Caswell

Research Station soil coverage exceeded 80% (Fig 3b).

Ground Cover and Kudzu Bug Adult Densities. Prior to canopy closure, there was a negative relationship between kudzu bug adult densities and percentage of ground cover at the Caswell Research Station during 2014 (r = -0.533; P = 0.0336), and at the Sandhills

Research Station during 2015 (r = -0.341; P = 0.0269). More adults were recorded from plots with less ground cover, compared with plots with high amounts of crop residue (Fig. 4).

After soybean canopy closure, there was no relationship between kudzu bug adult densities and ground cover at the Caswell Research Station during 2014 (r = -0.262; P =

0.3275). Similarly during 2015, no relationship between adult densities and ground cover was present at the Sandhills Research Station (r = -0.018; P = 0.9101) or at the Caswell

Research Station (r = -0.012; P = 0.8635).

133

Field Spectrometry and Light Reflectance. Prior to canopy closure, light reflectance recorded from both inside and outside of plots differed among tillage treatments during 2014. Conventionally tilled and reduced-till plots had a lower reflectance profile, compared with plots where the cereal rye was mowed or rolled (data from inside plots, Fig.

5a). Reflectance from conventionally tilled and reduced-till plots followed a similar pattern to the reflectance from sampled kudzu plants, with a small peak near 550 nm in the green color region, and a small depression near 700 nm in the near infrared region (Fig. 5a). On the other hand, the reflectance profile from plots where the cereal rye was either mowed or rolled followed a linear pattern (Fig. 5a). All reflectance from outside the plots, where there were no plants, had a linear pattern (Fig. 5b) and higher values, compared with reflectance collected from inside the plots (Fig. 5). After soybean canopy closure, light reflectance profiles were similar among tillage types and followed a typical light reflectance pattern for a plant, like the kudzu plant spectral profile in Fig. 5, when measurements were taken from inside plots (data not shown).

Light Reflectance, Type of Tillage, and Ground Cover. Prior to soybean canopy closure, each calculated spectral index using light reflectance from within plots, was not affected by tillage type or ground cover (data not shown). On the other hand, NDVI (F =

6.88; P = 0.0132), PRI (F = 4.80; P = 0.0339), and AI (F = 8.77; P = 0.0066) were affected by the interaction between tillage type and ground cover, when calculated using light reflectance from the outside of plots. Spectral indexes calculated from plots under conventional or reduced tillage with lower ground cover had a wider range of values among replications, compared with plots with mowed or rolled rye and higher ground cover (Fig. 6).

Higher NDVI and PRI values were calculated from plots under conventional tillage with

134

almost no crop residue, compared with plots with rolled rye covering ~100% of the ground

(Fig. 6). On the contrary, higher AI values were calculated from plots with rolled rye and higher crop residue, compared with conventionally tilled plot with almost no previous crop residue (Fig. 6). Spectral indexes, calculated after canopy closure and from both inside and outside experimental plots, were not affected by tillage type, ground cover, nor the interaction between these two variables (data not shown).

Kudzu Bug Adult Densities, Light Reflectance, and Ground Cover. Prior to soybean canopy closure, kudzu bug adult densities in experimental plots were consistently influenced by the interaction between PRI and percentage of ground cover, from both reflectance collected inside (F = 9.02; P = 0.0110) and outside plots (F = 10.06; P = 0.0008).

Generally, more adults were found in experimental plots with lower crop residue (from zero to 50% ground cover) and where PRI values were low, compared to plots with higher crop residue and high PRI values (Fig. 7). Additionally, adult densities were only affected by the interaction between each spectral index and percentage of ground cover, when indexes were calculated using reflectance from outside plots (NDVI × ground cover: F = 10.39; P =

0.0073, DSSI 1 × ground cover: F = 14.22; P = 0.0027, AI × ground cover: F = 5.21; P =

0.0414). More kudzu bug adults were usually found in plots with low ground cover and high

NDVI values, compared to plots with higher ground cover and low NDVI values (Fig. 8). On the contrary, higher kudzu bug adult densities were normally found in plots with reduced presence of crop residue and low DSSI 1 and AI values, compared to plots with higher presence of crop residue and high DSSI 1 and AI values (Fig. 8). After soybean canopy closure, kudzu bug adult densities were not affected by any spectral index, calculated from

135

the inside or the outside of plots, ground cover, nor the interaction between these two variables (data not shown).

Soybean Yield. Potential pests other than kudzu bug, including stink bugs and lepidopteran caterpillars, did not reach any economic threshold in any experiment (data not shown). Neither tillage type (F = 3.38; df = 3, 9; P = 0.0679) nor row spacing (F = 0.17; df =

1, 5; P = 0.6950) influenced soybean yield during 2014. Similarly, during 2015 tillage type

(F = 0.33; df = 2, 26.1; P = 0.7211) and row spacing (F = 3.00; df = 1, 38.8; P = 0.0915) did not influence soybean yield. However, yield varied between locations during 2015 (F =

77.96; df = 1, 12.8; P < 0.0001), with higher yields recorded from plots at the Caswell

Research Station (4,568.48 ± 107.84 kg/ha), compared with plots at the Sandhills Research

Station (2,261.99 ± 86.91 kg/ha).

Discussion

Kudzu bug densities were not affected by tillage type, prior to, or after, soybean canopy closure when tillage type was tested alone in one experimental location. On the other hand, when row spacing and tillage were incorporated into a single experiment and the experiment was replicated in two locations, kudzu bug densities were influenced by the interaction between tillage type and location (prior to canopy closure), and by the interaction among tillage type, row spacing, and location (after canopy closure). Crop residue, either cereal stubble or/and straw, and ground cover level from these residues were correlated with tillage in these experimental soybean plots. Ground cover had an effect on kudzu bug adult densities prior soybean canopy closure; where lower adult numbers were recorded in experimental plots with higher levels of ground cover. In addition, tillage type was associated

136

with differences in light reflectance from ground surface in the present study; where the effect on spectral index values differed between tillage types for a given level of ground cover. Additionally, kudzu bug adult densities were influenced by the interaction between ground cover and index level for some spectral indexes, indicating a potential relationship between light reflectance and kudzu bug attractiveness in soybean. Kudzu bug adult densities also varied at the end of the sampling season when row spacing treatments were compared.

However, in general, kudzu bug densities were inconsistently influenced by row spacing in this study.

There were no kudzu bugs recorded at the beginning of the sampling of trials during both years at both locations. Although data were not collected, most adults were located at kudzu patches near my experimental plots, rather than in soybean, at this time of the season.

Kudzu bugs began to disperse in higher numbers into experimental plots after soybean canopy closure. However, even with relatively low initial kudzu bug adult densities (~0.5 times the action threshold), I was able to identify a significant relationship between adult densities and ground cover. It is still unclear why kudzu bug adults, located in kudzu near my experimental fields, did not disperse earlier to my experimental soybean plots. The dispersal cues from kudzu for this insect are still unknown. However, if conspecific density dependent effects are important cues to initiate dispersal, then it is possible that the relatively low kudzu bug numbers in kudzu were not high enough to cause dispersal and subsequent dispersal into my soybean plots (kudzu bug numbers in general were far lower across the Southeast in 2014 and 2015 than in the previous three years). Overall low kudzu bug numbers may have not impacted the quality of kudzu as a host, making it more attractive and suitable for these

137

populations for longer periods of time, kudzu bugs may have not have dispersed from the kudzu.

No kudzu bug egg masses were observed on the uppermost and fully-formed soybean trifoliates at the Caswell Research Station during 2015. Despite this result, I was able to record kudzu bug nymphs from these plots in sweep-net samples. Consequently, the number of uppermost, fully-formed soybean trifoliates sampled in this study was not adequate to detect egg masses under low population densities were apparently present (~10 times below the action threshold at this location during 2015) in soybean. On the other hand, sweep- netting seems to be a more reliable sampling methodology to estimate the relative abundance of kudzu bug nymphs and adults at densities lower than the action threshold (Stubbins et al.

2014).

Soybean row spacing did not consistently affect kudzu bug abundance in the present study. After canopy closure during 2014, more kudzu bug nymphs were recorded from soybean planted on wide rows, compared with narrow row soybean. This trend did not continue throughout the season, since more kudzu bug adults were recorded from narrow soybean during the last two evaluations (Fig. 1). A potential explanation for this discrepancy could be dispersal. Perhaps more adults migrated to experimental plots later in the season when plant architecture of narrow row soybean favored their establishment. On the other hand, during 2015, more kudzu bug adults were recorded on soybean planted in wide rows under conventional tillage, compared with narrow soybean under the same tillage type in one location compared to the other. At the location where kudzu bugs were more abundant, the effect of tillage on kudzu bug numbers depended on row spacing. I speculate that the 2015 experimental design, where tillage type and row spacing were included in the same trial

138

together, could have impacted the visual cues present for emigrant kudzu bugs, such as plant architecture, or light reflectance from soybean plants and the ground surface. Potential modification of visual cues by tillage type and row spacing in these plots could have ultimately influenced kudzu bug attraction to soybean plants.

Geographic location had also an impact on kudzu bugs during 2015. The western location in NC (Sandhills Research Station) had higher kudzu bug populations, compared to an eastern location (Caswell Research Station). Based on the recent kudzu bug invasion range distribution, it seems like this range moved from the point of origin in GA to north and east into VA during the early years of the invasion (2009 – 2013), and to west into AR and south into FL during the past three years (Gardner 2016). This ‘invasion wave’ may have caused a lag between northern and southern, as well as eastern and western kudzu bug populations, since the southern and western ones established first and had the opportunity to build up in numbers. Higher populations, based on location, may have impacted kudzu bug population dynamics, ultimately influencing potential density-dependent dispersal factors, like the previously mentioned in kudzu. In this study, the western location in NC had higher kudzu bug densities early in the season and prior to soybean canopy closure, indicating that more kudzu bug adults dispersed earlier in this location, compared to the eastern site located

~200 km away from the western location.

The percentage of ground cover was manipulated by tillage type. In the present study, conventional tillage plots had almost no crop residue on the soil surface compared with almost complete soil coverage when cereal rye was rolled on the soil surface. Cover crops and crop residue can change the visual or olfactory cues, or create mechanical barriers to movement, that interfere with the host finding process of insects (Sarrantonio and Gallandt

139

2003). In this study, fewer kudzu bug adults were recorded in plots with higher ground cover, compared with densities from conventionally tilled plots, suggesting that ground cover may have an effect on the establishment of kudzu bug early in the season.

Since finding the location of a suitable host plant can be driven by several visual and chemical cues (Speight et al. 2008), several factors may be influencing the host finding process for kudzu bugs in soybean. In the present study, I investigated how tillage type modified the presence of crop residue, and how crop residue might affect light reflectance from the ground surface. One factor, reflectance, could potentially impact the host finding process of kudzu bug in plots with a different amounts and qualities of ground cover. This study confirms that presence of crop residue in soybean fields modified light reflectance, and may have affected the initial establishment of kudzu bug in this crop. This is confirmed by the significant and negative relationship between kudzu bug adults and ground cover prior soybean canopy closure over the course of two years at two different locations, and by the variation of kudzu bug densities due to light reflectance (analyzed as spectral indexes) and ground cover.

Light reflectance is one of the component visual cues used to find potential host plants in insects, especially in hemipterans such as aphids (Speight et al. 2008, Luo et al.

2013). Light reflectance will change depending on the composition of the reflecting surface.

For instance, variation in moisture content from soil samples generates different reflectance profiles (Lobell and Asner 2002). In the present study, the presence of uncovered soil in conventionally tilled or reduced-till plots, may have changed the light reflectance in those treatments, compared to plots with mowed and rolled rye where residue was covering most of the ground. Soil particles and their constituents can absorb light (Lobell and Asner 2002),

140

generating lower reflectance values from conventionally tilled and reduced-till plots. When cereal rye residue covered most of the soil in this study, it reflected more light, in contrast with the greater light absorption in conventionally tilled plots. Previous research has shown that cereal straw on the surface of zucchini squash plots is highly reflective, compared with squash planted on bare ground (Summers et al. 2004). The light reflectance profiles taken in the conventionally tilled plots had a small peak near 550 nm. Because these plots had almost no ground cover, some reflectance near this area of the light spectrum likely indicates some green reflectance due to the presence of chlorophyll in soybean plants (Nansen and Elliot

2016). On the other hand, the reflectance profile from rolled rye plots was a straight line with no peak near 550 nm. It is likely that the cereal rye cover saturated the reflectance profile of soybean plants in those plots, since wheat straw reflects more light than bare ground

(Summers et al. 2004).

Several spectral indexes were calculated in the present study to detect potential changes in ground surface or soybean canopy reflectance from different tillage types. The

Photochemical Reflectance Index (PRI) was the only index tested was consistently associated with differences in kudzu bug densities among different ground covers. PRI measures the efficiency of photosynthetic radiation use in plants (Gamon et al. 1997). Lower PRI values indicate that selected plants might be under stress or have a nutrient deficiency (Gamon et al.

1997). In this study, lower PRI values were calculated from rolled rye plots before soybean canopy closure. Perhaps soybean plants located at those plots were under stress, or perhaps reflectance profiles of the cover cereal rye may be similar to stressed plants. However, there was no visual indication that soybean plants were under stress. Additional indexes, such as the Aphid Index (AI), were also associated with kudzu bug populations across different

141

levels of ground cover. The AI is a variation of the Structure Insensitive Pigment Index and indirectly measures the physiology and phenology of a plant (Mirik et al. 2006). Higher AI values may indicate that selected plants were senescing or under stress (Mirik et al 2006).

More kudzu bugs were found at plots where AI values were low, such as in conventionally tilled and reduced-till plots. Having a low AI indicates the presence of green, healthy, and potential young plants (Mirik et al 2006). Hence, the presence of cereal rye residue on top of plot may have ‘camouflaged’ soybean plants or mimicked the signal of unhealthy plants in those plots. Additionally, I can speculate that the light reflectance from the background

(ground cover only) may be influencing the attractiveness of kudzu bugs to specific plots, since spectral index values calculated only from outside of experimental plots (ground cover with no soybean plants) were correlated with kudzu bug adult densities prior to soybean canopy closure. It also seems that this potential effect of ground cover and light reflectance on kudzu bug attractiveness did not last through canopy closure, since there was no relationship between light reflectance and kudzu bug densities later in the season.

Soybean yield was not impacted by tillage type or row spacing treatment at any given location during the two years of this study. These results contradicted my previous research, where soybean planted under conventionally tilled plots had higher yields, compared with soybeans planted under a reduced-till system (A.I.D., unpublished data). In this study, planting date was earlier (in May), compared with the unpublished research study

(in June); kudzu bug pressure was lower, and higher precipitation was recorded during 2014 and 2015 (this study), compared with 2012 and 2013 during the unpublished field trials. I suspect that soybean yield compensation was higher during this study, since stress factors such as adverse weather or insect injury were low during 2014 and 2015.

142

Future research should examine the effects of manipulating additional potential factors influencing the host finding process of kudzu bug mediated by visual cues. Additional experiments could test how light reflectance may influence kudzu bug attraction to soybean by changing the color on the ground surface of plots under the same tillage type, identifying specific wavelengths that attract kudzu bugs, or comparing straw and high reflectance plastic mulches. If relationships between light reflectance and insect densities can be established in this system, the use of highly reflective mulch could alter the attractiveness of kudzu bugs to host plants, making this type of mulch a potential control tactic for organic soybean farmers.

In conclusion, tillage type changed the amount of ground cover, impacting light reflectance in my experimental soybean plots. Moreover, there may be several factors influencing kudzu bug host finding and establishment processes in soybean. Light reflectance may be one of those factors, affecting how kudzu bug visually cues on soybean. Crop residue may also affect the microclimate at the ground level, generating an adverse environment for kudzu bug, due to the potential presence of predators or pathogens, as in other systems mentioned by Stinner and House (1990). However, it is more likely that lower kudzu bug densities will be found in soybean planted under a conservation tillage system with an elevated amount of crop residue on the ground, compared with soybean planted under conventional tillage with almost no crop residue. Furthermore, kudzu bug densities could be influenced by row spacing later in the growing season, when the risk of kudzu bug injury and associated yield loss would be reduced. Managing kudzu bug populations later in the season may not be economically advisable, since most of yield will not be reduced after the seed- filling period (R5 to R7) (Singh 2010).

143

Acknowledgments

Author would like to thank Dr. Chris Reberg-Horton, Rachel Atwell, Sarah Seehaver

(NC State University Department of Crop Science), Roy Maitland, Evan Taylor, Jeremy

Martins, Joshua Meyers, and personnel from the Caswell and Sandhills Research Stations for helping coordinating equipment availability and site preparation; and David Williams (NCSU

Department of Soil Science) for conducting the field spectrometry measurements. Author also thank Dan Mott, Steven Roberson, Clifton Moore, James Lawrence, Christopher

McBennett, Ernesto Sobrevilla, and Axel Gonzales for their assistance during planting and field scouting. This study was partially funded by the North Carolina Soybean Producer

Association and the United Soybean Board.

References

Andrade, F. H., P. Calviño, A. Cirilo, and P. Barbieri. 2002. Yield responses to narrow

rows depend on increased radiation interception. Agron. J. 94: 975-980.

ASD Inc. 2008. ViewSpec Pro version 6.2.0 user’s manual. ASD Inc., Boulder, CO.

Blount, J. L., G. D. Buntin, and A. N. Sparks, Jr. 2015. Host preference of Megacopta

cribraria (Hemiptera: Plataspidae) on selected edible beans and soybean. J. Econ.

Entomol. 108: 1094–1105.

Blount, J. L., G. D. Buntin, and P. M. Roberts. 2016. Effect of planting date and maturity

group on soybean yield response to injury by kudzu bug, Megacopta cribraria F.

(Hemiptera: Plataspidae). J. Econ. Entomol. 109: 207-212.

Board, J. E. and B. G. Harville. 1996. Growth dynamics during the vegetative period

affects yield of narrow-row, late-planted soybean. Agron. J. 88: 567-572.

144

Bottenberg, H., J. Masiunas, C. Eastman, and D. M. Eastburn. 1997. The impact of rye

cover crops on weeds, insects, and diseases in snap bean cropping systems. J. Sustain.

Agr. 9: 131-155.

Bowers, G. R., J. L. Rabb, L. O. Ashlock, and J. B. Santini. 2000. Row spacing in the

early soybean production system. Agron. J. 92: 524-531.

Buchanan, A. L., E. Zobel, J. Hinds, A. Rosario-Lebron, and C. R. R. Hooks. 2015. Can

row spacing influence arthropod communities in soybean? Implications for early and

late planting. Environ. Entomol. 44: 557-561.

Conservational Technology Information Center. 2002. Tillage type definitions.

http://www.ctic.purdue.edu/resourcedisplay/322/ (accessed 17 March 2016).

Del Pozo-Valdivia, A. I., N. J. Seiter, D. D. Reisig, J. K. Greene, F. P. F. Reay-Jones,

and J. S. Bacheler. 2016. Megacopta cribraria (Hemiptera: Plataspidae) population

dynamics in soybeans as influenced by planting date, maturity group, and insecticide

use. J. Econ. Entomol. pii: tow020.

Eger, J. E., Jr., L. M. Ames, D. R. Suiter, T. M. Jenkins, D. A. Rider, and S. E. Halbert.

2010. Occurrence of the Old World bug Megacopta cribraria (Fabricius)

(Heteroptera: Plataspidae) in Georgia: a serious home invader and potential legume

pest. Insecta Mundi 0121: 1–11.

Fehr, W. R., C. E. Caviness, D. T. Burmood, and J. S. Pennington. 1971. Stage of

development descriptions for soybeans, Glycine max (L.) Merrill. Crop Sci. 11: 929-

931.

145

Fritz, B., D. D. Reisig, C. E. Sorenson, A. I. Del Pozo-Valdivia, T. E. Carter, Jr. 2016.

Host plant resistance to Megacopta cribraria (Hemiptera: Plataspidae) in diverse

soybean germplasm maturity groups V through VIII. J. Econ. Entomol. pii: tow036.

Gamon, J. A., L. Serrano, and J. S. Surfus. 1997. The photochemical reflectance index: an

optical indicator of photosynthetic radiation use efficiency across species, functional

types, and nutrient levels. Oecologia. 112: 492-501.

Gardner, W. A. 2016. Distribution map of the kudzu bug. University of Georgia. Center for

invasive species and ecosystem health.

http://www.kudzubug.org/distribution_map.cfm (accessed 21 March 2016).

Hartwig N. L. and H. U. Ammon. 2002. Cover crops and living mulches. Weed Sci. 50:

688-699.

Heatherly L. G. and H. F. Hodges. 1999. Soybean production in the midsouth. CRC Press

LLC. Boca Raton, FL.

Kenward, M. G., and J. H. Roger. 1997. Small sample interference for fixed effects from

restricted maximum likelihood. Biometrics 53: 983–997.

Koch, R. L., P. M. Porter, M. M. Harbur, M. D. Abrahamson, K. A. G. Wyckhys, D. W.

Ragsdale, K. Buckman, Z. Sezen, and G. E. Heimpel. 2012. Response of soybean

insects to an autumn-seeded rye cover crop. Environ. Entomol. 41: 750-760.

Lobell, D. B. and G. P. Asner. 2002. Moisture effects on soil reflectance. Soil Sci. Soc. Am.

J. 66: 722-727.

Luo, J., W. Huang, L. Yuan, C. Zhao, S. Du, J. Zhang, and J. Zhao. 2013. Evaluation of

spectral indices and continuous wavelet analysis to quantify aphid infestation in

wheat. Precision Agric. 14: 151-161.

146

Martin, J. H., R. P. Waldren, and D. L. Stamp. 2006. Principles of field crop production.

Person Prentice Hall Inc. Upper Saddle River, NJ.

McPherson, R. M., G. W. Zehnder, and J. C. Smith. 1988. Influence of cultivar, planting

date and row width on abundance of green cloverworms (Lepidoptera: Noctuidae)

and green stink bugs (Heteroptera: Pentatomidae) in soybean. J. Entomol. Sci. 23:

305-313.

McPherson, R. M. and K. Bondari. 1991. Influence of planting date and row width on

abundance of velvetbean caterpillars (Lepidoptera: Noctuidae) and southern green

stink bugs (Heteroptera: Pentatomidae) in soybean. J. Econ. Entomol. 84: 311-316.

Miklasiewicz, T. J. and R. B. Hammond. 2001. Density of potato leafhopper (Homoptera:

Cicadellidae) in response to soybean-wheat cropping systems. Environ. Entomol. 30:

204-214.

Mirik, M., G. J. Michels, Jr., S. Kassymzhanova Mirik, N. C. Elliott, and V. Catana.

2006. Spectral sensing of aphid (Hemiptera: Aphididae) density using field

spectrometry and radiometry. Turk. J. Agri. For. 30: 421-428.

Nansen, C. and N. Elliot. 2016. Remote sensing and reflectance profiling in Entomology.

Annu. Rev. Entomol. 61: 139-158.

Pedersen, P. 2004. Row spacing in soybeans. Extension Bulletin. Iowa State University,

Ames, IA.

Reberg-Horton, S. C., J. M. Grossman, T. S. Kornecki, A. D. Meijer, A. J. Price, G. T.

Place, and T. M. Webster. 2011. Utilizing cover crop mulches to reduce tillage in

organic systems in the southeastern USA. Renew. Agri. Food Syst. 27: 41-48.

147

Sarrantonio, M. and E. Gallandt. 2003. The role of cover crops in North American

cropping systems. J. Crop Prod. 8: 53-74.

SAS Institute. 2010. SAS® version 9.3 user’s manual. SAS Institute, Cary, NC.

Seiter, N. J., J. K. Greene, and F.P.F. Reay-Jones. 2013. Reduction of soybean yield

components by Megacopta cribraria (Hemiptera: Plataspidae). J. Econ. Entomol.

106: 1676-1683.

Seiter, N. J., A. I. Del Pozo-Valdivia, J. K. Greene, F. P. F. Reay-Jones, P. M. Roberts,

and D. D. Reisig. 2015. Action thresholds for managing Megacopta cribraria

(Hemiptera: Plataspidae) in soybean based on sweep-net sampling. J. Econ. Entomol.

108: 1818–1829.

Singh, G. 2010. The soybean: botany, production, and uses. CABI Editors. Wallingford, UK.

Speight, M. R., M. D. Hunter, and A. D. Watt. 2008. Ecology of insects, concepts and

applications. Second Edition. Wiley-Blackwell, Oxford, UK.

Stinner, B. R. and G. J. House. 1990. Arthropods and other invertebrates in conservation-

tillage agriculture. Annu. Rev. Entomol. 35: 299-318.

Stubbins, F. L., N. J. Seiter, J. K. Greene, and F. P. F. Reay-Jones. 2014. Developing

sampling plans for the invasive Megacopta cribraria (Hemiptera: Plataspidae) in

soybean. J. Econ. Entomol. 107: 2213-2221.

Suiter, D. R., J. E. Eger, Jr., W. A. Gardner, R. C. Kemerait, J. N. All, P. M. Roberts, J.

K. Greene, L. M. Ames, G. D. Buntin, T. M. Jenkins, et al. 2010. Discovery and

distribution of Megacopta cribraria (Hemiptera: Heteroptera: Plataspidae) in northeast

Georgia. J. Integr. Pest Manage. 1: F1–F4.

148

Summers, C. G., J. P. Mitchell, and J. J. Stapleton. 2004. Management of aphid-borne

viruses and Bemisia argentifolii (Homoptera: Aleyrodidae) in zucchini squash by

using UV reflective plastic and wheat straw mulches. Environ. Entomol. 33: 1447-

1457.

Tillman G., H. Schomberg, S. Phatak, B. Mullinix, S. Lachnicht, P. Timper, and D.

Olson. 2004. Influence of cover crops on insect pests and predators in conservation

tillage cotton. J. Econ. Entomol. 97: 1217-1232.

U.S. Department of Agriculture (USDA) and U.S. Department of the Interior (USDI).

1999. Sampling vegetation attributes - Interagency technical reference. Bureau of

Land Management National Business Center, Denver, CO.

Wyland, L. J., L. E. Jackson, W. E. Chaney, K. Klonsky, S. T. Koike, B. Kimple. 1996.

Winter cover crops in a vegetable cropping system: impacts on nitrate leaching, soil

water, crop yield, pests and management costs. Agri. Ecosyst. Environ. 59: 1-17.

Zhang, Y., J. L. Hanula, and S. Horn. 2012. The biology and preliminary host range of

Megacopta cribraria (Heteroptera: Plataspidae) and its impact on kudzu growth.

Environ. Entomol. 41: 40–50.

149

Fig. 1. Megacopta cribraria nymphs (black bars) and adults (grey bars) per plant, recorded from soybean planted in narrow (17.78 cm) and wide (96.52 cm) rows during the last three sampling dates (horizontal panels) at Caswell Research Station, during 2014. Error bars were calculated using ± one standard error of the mean. Earlier sampling dates are not included since kudzu bug densities (adults and/or nymphs) did not statistically vary during those dates.

150

Fig. 2. Megacopta cribraria adults per sweep, recorded from soybean planted under three types of tillage (conventionally tilled, reduced-till, and rolled-rye, vertical panels), using two row spacings (x-axis, narrow - 38.1 cm and wide - 76.2 cm), and at two experimental locations (Sandhills and Caswell Research Stations, horizontal panels) after canopy closure during 2015. Error bars were calculated using ± one standard error of the mean. Means sharing the same letter are not statistically significant (α > 0.05).

151

(a) (b)

Fig. 3. Percentage of ground cover by previous crop residue (cereal rye stubble and straw) in

(a) four treatments including, conventional tillage, reduced tillage, rye mowed on top of plots, and rye rolled on top of plots at Caswell Research Station, Kinston, NC during 2014; and (b) three treatments including, conventional tillage, reduced tillage, and rye rolled on top of plots at two locations (Caswell Research Station, Kinston, and Sandhills Research Station,

Jackson Springs, NC) during 2015. Measurements were directly following soybean planting.

Means sharing the same letter are not statistically different (α > 0.05).

152

(a) (b)

Fig. 4. Relationship between Megacopta cribraria adults per sweep (expressed as log10-base) and percentage of ground cover in soybean prior canopy closure at the Caswell Research

2 Station, 2014 (a) (P = 0.0336; R = 0.23; ground cover = -0.0005 × log10 (adults/sweep) +

0.076), and at the Sandhills Research Station, 2015 (b) (P = 0.0269; R2 = 0.10; ground cover

= -0.0005 × log10 (adults/sweep) + 0.038). During 2015, there were no kudzu bug adults recorded at the Caswell Research Station prior canopy closure.

153

Fig. 5. Average light reflectance calculated using measurements taken on 11 June 2014 from inside (a) and outside (b) of experimental soybean plots including, conventionally tilled (×), under reduced-till (squares), with mowed rye (triangles), and with rolled rye (upside-down triangles) at Caswell Research Station, Kinston, NC. Measurements from inside plots recorded reflectance coming from both small plants (before canopy closure) and ground cover; and from the outside, reflectance coming from the ground cover only with no plants.

Average reflectance calculated for kudzu, Pueraria montana, (circles) was taken 16 June

2014 on North Carolina State University campus, Raleigh, NC.

154

Fig. 6. Scatter plots showing the relationship between selected spectral indexes (Y-axis) and percentage of ground cover (X-axis), under four different tillage types (horizontal panels; circle=conventionally tilled, plus=reduced tillage, cross=mowed rye, and triangle=rolled rye) at the Caswell Research Station, Kinston, NC, during 2009. These indexes were calculated from light reflectance collected prior soybean canopy closure, and taken from the outside of experimental plots (separation alley with tillage treatment and no soybean plants). Mean separations are not shown in this figure, but they are described in the result section.

155

Fig. 7. Contour plots showing the variation of Megacopta cribraria adults (expressed as log10-based and fitted as isolines), based on the Photochemical Reflectance Index (X-axis, expressed as log10-base) calculated from inside (tillage treatment + soybean plants, left) and from the outside experimental plots (separation alley with tillage treatment and no soybean plants, right), and the percent of ground cover (Y-axis). Red and blue colors indicate high and low values of M. cribraria adults, respectively. Light reflectance was taken prior to soybean canopy closure at the Caswell Research Station, Kinston, NC, during 2014.

156

Fig. 8. Contour plots showing the variation of Megacopta cribraria adults (expressed as log10-based and fitted as isolines), based on three spectral indexes (X-axis, expressed as log10-base): (a) Normalized Difference Vegetation Index, (b) Damage Sensitive

Spectral Index 1, and (c) Aphid Index, calculated from the outside of experimental plots (separation alley with tillage treatment and no soybean plants) and the percent of ground cover (Y-axis). Red and blue colors indicate high and low values of M. cribraria adults, respectively. Light reflectance was taken prior to soybean canopy closure at the Caswell Research Station, Kinston, NC, during 2014.

157

SECTION FIVE

CHAPTER SEVEN

Conclusions

First-generation kudzu bug, Megacopta cribraria (F.) (Hemiptera: Plataspidae), successfully transitioned from egg to adult in 45-50 days on soybean, Glycine max (L.) Merr., under greenhouse conditions. My no-choice experiments were the first evidence demonstrating that kudzu, Pueraria montana Loureiro (Merrill) variety lobata (Willdenow), was not an obligate host for first-generation kudzu bug. Since individuals from the first-generation can feed and reproduce on soybean, second-generation population densities have the potential to increase from reproducing first-generation individuals within the same field. Furthermore, the geographic distribution of kudzu bug may not be restricted to areas where kudzu is located; hence, this insect could potentially move and infest soybean planted in the Midwest, beyond the current kudzu distribution. My greenhouse studies also demonstrated that this insect was not able to survive on snap beans, Phaseolus vulgaris L., indicating that not all legumes are suitable hosts for the kudzu bug. A later study confirmed that only a few legumes, besides kudzu and soybean, are suitable reproductive hosts for this insect (Blount et al. 2015).

Kudzu bug flight was most apparent within soybean fields from 13:00 to 15:00 hours, based on sticky card monitoring. Lower captures during the morning may imply that kudzu bugs were not dispersing within experimental fields, since I was able to document during that time that they were mostly aggregating on soybean plants. Kudzu bug density on soybean was not influenced by soybean phenology, resulted from the variation of soybean maturity groups (MG)

158

in the same field. In general, most kudzu bugs formed aggregations in the middle section of the soybean main stem, regardless of plant height and MG, when visually inspected from 09:00 to

12:00 hours. More kudzu bug adults were found in the middle section of soybean plants in

August compared with June during these visual inspections; and from those adults, there were more males than females per plant (1.53 male-to-female ratio). I speculate, based on previous findings from Hibino (1986), that kudzu bug aggregations are correlated with mating behavior; in this scenario females will have several males to choose from before and while engaging in mating. Additional research on kudzu bug behavior, designed to test this hypothesis and to determine when this insect feeds on soybean, mates, oviposits, disperses, etc., is warranted.

Kudzu bug density is influenced by several soybean production practices, including planting date, insecticide use, soil tillage, and ground cover. On the other hand, populations of this pest were not consistently affected by selection of soybean maturity group (MG) or row spacing. More kudzu bugs were found on soybean planted during April or May, as well as those that were not treated with insecticide independent of planting date. In contrast, fewer kudzu bugs were found on soybean planted during June, as well as those that were treated with insecticide independent of planting date. The reasons why kudzu bug reaches higher numbers in early planted soybean are not fully understood. However, I speculate that early planted soybean is one of the first apparent hosts for overwintering adults and they can colonize them earlier. Early planted soybean may remain as suitable host for kudzu bug population growth for a greater portion of the kudzu bug activity period than later planted soybean. Additionally, established adults on early planted soybean may signal the presence of a suitable plant, by producing an aggregation pheromone (Wertheim et al. 2005), and subsequently attract more immigrant

159

overwintering adults to this early planted soybean fields. The purpose of such aggregation could be to facilitate mating, or to avoid parasitization or predation through herding (Hamilton 1971); other explanations are also possible.

More kudzu bugs were also found on soybean planted under conventional tillage with very little residue from the previous crop, compared with soybean planted under conservational tillage with abundant crop residue. Presence of crop residue modified the light reflectance of selected experimental plots. Although one experimental design confounded tillage effects with those of light reflectance due to residue, I hypothesize that visual cues might be major drivers during the host finding process of the kudzu bug. The relationship between kudzu bug density and light reflectance found in these studies prior to canopy closure, and expressed as spectral indexes, provides some initial support for my hypothesis. Perhaps immigrant kudzu bugs cannot effectively identify soybean plants planted under conservational tillage; previous crop residue, specifically cereal straw and stubble, might be acting as highly reflective mulch. This type of mulch therefore may be impeding kudzu bug recognition of soybean/soil reflectance signatures, impeding them to align with the rows, and forcing them to continue their search for a plant host.

Future research on this topic, incorporating modifications of the color of the background, or using highly reflective plastic mulches, may provide further clarification of how kudzu bugs visually recognize soybean plants. I envision that these studies could provide the basis for explanations into the possibility of using remote sensing to identify areas at risk for kudzu bug infestations in commercial soybean fields, using the spectral indexes such as the Photochemical

Reflectance Index (Gamon et al. 1997) or the Aphid Index (Mirik et al. 2006).

160

Soybean maturity group and row spacing inconsistently affected population densities of kudzu bug. Hence, although different phenologies of soybean on the same experimental field were available at any given time, these did not influence kudzu bug in a predictable manner.

Presence or absence of reproductive structures may not influence kudzu bug density, since they are considered phloem feeders. However, soybean nutrient titers change in the plant at different growth stages (Walter and DiFonzio 2007) and may influence the production of plant volatiles.

These volatiles could be potential chemical cues modulating the host finding process of the kudzu bug. Subsequently, the modification of row spacing alters soybean plant architecture, including main stem diameter and distance between nodes. I believe that there must be other visual cues beside plant height or soybean plant phenology that kudzu bug relies on to find a suitable host, including leaf shape and other plant architectural features. Future research could measure other plant canopy architecture features in soybean and identify if they influence the attractiveness of kudzu bug.

Soybean yield was influenced by planting date, maturity group, insecticide use, and tillage system. Higher soybean yields were recorded from soybean planted during May using a

MG VII, compared with soybean planted during April using a MG IV. Kudzu bug injury reduced soybean yield up to 20% in my studies, when insecticide treated plots were compared with untreated plots. When planted in June, conventionally tilled soybean had a higher yield than reduced-till soybean. It is possible that soil cultivation could have influenced the availability of nutrients in the soil (Singh 2010) and ultimately affected plant growth and seed production.

Based on these results, a combination of select production practices can be expected to influence the possibility of having a higher or a lower risk of kudzu bug infestation in soybean

161

(Fig. 1). Soybean growers could initially direct their scouting efforts to only those “higher kudzu bug risk” fields (Fig. 1, black box). “Lower kudzu bug risk” fields (Fig. 1, grey boxes) should be scouted once kudzu bugs are observed in “higher kudzu bug risk” fields; since the expectation would be to have higher kudzu bug densities in “higher kudzu bug risk” fields, compared with

“lower kudzu bug risk” fields. This directed scouting proposal may ultimately reduce scouting labor and cost. Growers should be advised to plant in May or later, reduce cultivation passes, and increase conservational tillage practices to minimize kudzu bug infestation risk. However, the selection of any agronomical practice should ultimately provide a favorable environment to express the maximum yield potential of the crop, which may include the adoption of production practices that create a higher risk environment for kudzu bug infestation.

References

Blount, J. L., G. D. Buntin, and A. N. Sparks, Jr. 2015. Host preference of Megacopta

cribraria (Hemiptera: Plataspidae) on selected edible beans and soybean. J. Econ.

Entomol. 108: 1094–1105.

Gamon, J. A., L. Serrano, and J. S. Surfus. 1997. The photochemical reflectance index: an

optical indicator of photosynthetic radiation use efficiency across species, functional

types, and nutrient levels. Oecologia. 112: 492-501.

Hamilton, W. D. 1971. Geometry for the selfish herd. J. Theor. Biol. 31: 295-311.

Hibino, Y. 1986. Female choice for male gregariousness in a stink bug, Megacopta

punctissimum (Montandon) (Heteroptera, Plataspidae). J. Ethol. 4: 91-95.

162

Mirik, M., G. J. Michels, Jr., S. Kassymzhanova Mirik, N. C. Elliott, and V. Catana. 2006.

Spectral sensing of aphid (Hemiptera: Aphididae) density using field spectrometry and

radiometry. Turk. J. Agri. For. 30: 421-428.

Seiter, N. J., A. I. Del Pozo-Valdivia, J. K. Greene, F.P.F. Reay-Jones, P. M. Roberts, and

D. D. Reisig. 2015. Action thresholds for managing Megacopta cribraria (Hemiptera:

Plataspidae) in soybean based on sweep-net sampling. J. Econ. Entomol. 108: 1818-1829.

Singh, G. 2010. The soybean: botany, production, and uses. CABI Editors. Wallingford, UK.

Walter, A. J. and C. D. DiFonzo. 2007. Soil potassium deficiency affects soybean phloem

nitrogen and soybean aphid populations. Environ. Entomol. 36: 26-33.

Wertheim, B., E. A. van Baalen, M. Dicke, and L. E. M. Vet. 2005. Pheromone-mediated

aggregation in nonsocial arthropods: an evolutionary ecology perspective. Annu. Rev.

Entomol. 50: 321-346.

163

Fig. 1. Flow chart summarizing findings from this dissertation. Rectangular boxes indicate a production practice or a process during the growing season. Diamonds represent decisions preceded by a question. Soybean fields can be characterized as having higher Megacopta cribraria (KB) infestation risk (black box) or lower KB infestation risk (grey boxes). The action threshold (*) for M. cribraria is one nymph/sweep, based on Seiter et al. (2015). 164