Effects of Landscape, Intraguild Interactions, and a Neonicotinoid on Natural Enemy and Pest Interactions in Soybeans

Home , Apolygus lucorum, Cyclocephala lurida, Hypsipyla robusta, Lasius neoniger, Latridiidae, Myrmecina difficulta

University of Kentucky UKnowledge

Theses and Dissertations--Entomology Entomology

2016

EFFECTS OF LANDSCAPE, INTRAGUILD INTERACTIONS, AND A NEONICOTINOID ON NATURAL ENEMY AND PEST INTERACTIONS IN SOYBEANS

Hannah J. Penn University of Kentucky, [email protected] Author ORCID Identifier: http://orcid.org/0000-0002-3692-5991 Digital Object Identifier: https://doi.org/10.13023/ETD.2016.441

Right click to open a feedback form in a new tab to let us know how this document benefits ou.y

Recommended Citation Penn, Hannah J., "EFFECTS OF LANDSCAPE, INTRAGUILD INTERACTIONS, AND A NEONICOTINOID ON NATURAL ENEMY AND PEST INTERACTIONS IN SOYBEANS" (2016). Theses and Dissertations-- Entomology. 30. https://uknowledge.uky.edu/entomology_etds/30

This Doctoral Dissertation is brought to you for free and open access by the Entomology at UKnowledge. It has been accepted for inclusion in Theses and Dissertations--Entomology by an authorized administrator of UKnowledge. For more information, please contact [email protected]. STUDENT AGREEMENT:

I represent that my thesis or dissertation and abstract are my original work. Proper attribution has been given to all outside sources. I understand that I am solely responsible for obtaining any needed copyright permissions. I have obtained needed written permission statement(s) from the owner(s) of each third-party copyrighted matter to be included in my work, allowing electronic distribution (if such use is not permitted by the fair use doctrine) which will be submitted to UKnowledge as Additional File.

I hereby grant to The University of Kentucky and its agents the irrevocable, non-exclusive, and royalty-free license to archive and make accessible my work in whole or in part in all forms of media, now or hereafter known. I agree that the document mentioned above may be made available immediately for worldwide access unless an embargo applies.

I retain all other ownership rights to the copyright of my work. I also retain the right to use in future works (such as articles or books) all or part of my work. I understand that I am free to register the copyright to my work.

REVIEW, APPROVAL AND ACCEPTANCE

The document mentioned above has been reviewed and accepted by the student’s advisor, on behalf of the advisory committee, and by the Director of Graduate Studies (DGS), on behalf of the program; we verify that this is the final, approved version of the student’s thesis including all changes required by the advisory committee. The undersigned agree to abide by the statements above.

Hannah J. Penn, Student

Dr. John J. Obrycki, Major Professor

Dr. Charles Fox, Director of Graduate Studies

EFFECTS OF LANDSCAPE, INTRAGUILD INTERACTIONS, AND A NEONICOTINOID ON NATURAL ENEMY AND PEST INTERACTIONS IN SOYBEANS

______

DISSERTATION ______

A dissertation submitted in partial fulfillment of the Requirements for the degree of Doctor of Philosophy in Entomology in The College of Agriculture, Food and Environment At the University of Kentucky

Hannah Joy Penn

Lexington Kentucky

Director: Dr. John J. Obrycki, Professor of Entomology

Lexington Kentucky

2016

ABSTRACT OF DISSERTATION

EFFECTS OF LANDSCAPE, INTRAGUILD INTERACTIONS, AND A NEONICOTINOID ON NATURAL ENEMY AND PEST INTERACTIONS IN SOYBEANS

Demand for food, fuel, and fiber has been increasing, escalating the intensification of agriculture during the past fifty years. A more comprehensive understanding of the impact of landscapes on sustainable agriculture production is required to meet the continual increase in human demand. This not only includes how chemical inputs are used but also how cultivated and surrounding landscapes are managed for ecosystem services. This research explains how land cover on landscape and farm scales impact ant and spider-mediated pest suppression. I successfully developed and optimized molecular methods to test ant gut contents from both laboratory and field-caught specimens. A multi-year field study indicated that spatial scale played a significant role in both ant and pest abundance and diversity when land cover at the landscape scale was analyzed, where an increase in scale generally correlated with an increase in the intensity of the relationship between landscape fragmentation and community response. I found that ant and pest populations can potentially be managed at the landscape scale via specific cover types and habitat fragmentation. Farm scale field margin composition significantly influenced spider and pest within field abundances and spatial associations, particularly early in the season. In a field cage study, ants were found to interfere with reductions in pest-induced leaf damage caused by spiders within the soybean system despite the fact that the pest was non-honeydew producing. Finally, I found that chemical disturbance via imidacloprid and fungicide seed treatments did not impact pest populations or ant predation rates but did significantly increase ant diversity by decreasing the abundance of a dominant ant species. However, there were impacts found on ant individuals within a lab mesocosm where imidacloprid induced sub-lethal intoxication but fungicides were lethal to Tetramorium caespitum.

KEYWORDS: Agroecosystem, Araneae, Food Webs, Formicidae, Landscape Ecology

___Hannah J. Penn ______

___December 2, 2016______Date

EFFECTS OF LANDSCAPE, INTRAGUILD INTERACTIONS, AND A NEONICOTINOID ON NATURAL ENEMY AND PEST INTERACTIONS IN SOYBEANS

Hannah Joy Penn

___John Obrycki ___ Director of Dissertation

__Charles Fox______Director of Graduate Studies

___ December 2, 2016______

THIS WORK IS DEDICATED TO MY PARENTS WHO ENCOURAGED ME TO DEVOTE MYSELF, EVEN AS A CHILD, TO MY ENTOMOLOGICAL ENDEAVORS.

ACKNOWLEDGMENTS

The following dissertation, while an individual work, benefited from the insights and direction of many people. First my Dissertation Chair, Dr. John Obrycki, who exhibits the leadership and scholarship to which I aspire, particularly in times of academic crisis. In addition, laboratory technician and fellow student Kacie Athey provided technical training, comments on experimental designs, assistance with experimental execution, and invaluable thoughts on every presentation and paper developed from this dissertation. Dr. Katelyn Kowles, Dr. Michael Sitvarin, Dr. Jamin Dreyer, Dr. Eric Chapman, and Dr. Brian Lee also provided timely and instructive comments on projects and papers during the dissertation process. I would also like to thank Dr. Douglas Johnson, Dr. Chad Lee, and Amanda Martin for assisting in the farmer recruitment and the on-farm research aspects of this dissertation. I would also like to thank Dr. Gary Coovert for ant identification assistance and Dr. Mary Gardiner and Dr. Dave Crowder for helpful comments regarding GIS. I wish to thank the complete Dissertation Committee, and outside reader, respectively: Dr. Ricardo Bessin, Dr. Daniel Potter, Dr. John Snyder, and Dr. Brian Lee. In addition to the technical and thought-provoking comments from those cited above, I was equally supported by friends and family. My professional and personal associate, Jerrod Penn, provided encouragement, collaborative projects, the sushi and bagels required for moral and physical stamina throughout the dissertation process as well as a sounding board for ideas, statistical analyses, and grievances. My mother, Rita McKenrick, worked tirelessly since I was in middle school to ensure that I had every professional opportunity that I could ever require. My father, Daniel McKenrick fielded untimely phone calls regarding unconventional experimental engineering and execution. Finally, I wish to thank the growers who graciously allowed me to sample their fields, discuss their farming histories, and drink coffee in their homes. Their willingness to assist an unknown student allowed this dissertation to come to completion.

vii

Table of Contents

ACKNOWLEDGMENTS ...... vii Table of Tables ...... xii Table of Figures ...... xiv Chapter 1 INTRODUCTION ...... 1 1.1 General Background ...... 2 1.1.1 Soybean Production ...... 2 1.1.2 Insect Pests of Soybean ...... 3 1.2 Biological Control ...... 3 1.2.1 Importance of Generalist Predators ...... 4 1.2.2 Intraguild Competition ...... 7 1.3 Molecular Gut Content Analysis ...... 8 1.4 Landscape Ecology ...... 9 1.4.1 Impact of Landscape on Biodiversity...... 10 1.4.2 Landscape Impacts on Biological Control ...... 11 Chapter 2 OVERCOMING PCR INHIBITION DURING DNA-BASED GUT CONTENT ANALYSIS OF ANTS ...... 14 2. 1 Summary ...... 14 2. 2 Introduction ...... 15 2. 3 Materials and Methods ...... 18 2.3.1 Insects ...... 18 2.3.2 Dissection Treatments ...... 20 2.3.3 DNA Extraction and PCR ...... 21 2.3.4 Statistical Analyses ...... 25 2.4 Results ...... 26 2.5 Discussion ...... 28 Chapter 3 DETECTION OF DNA IN APHID HONEYDEW: A SOURCE OF ERROR FOR MOLECULAR GUT CONTENT ANALYSIS? ...... 38 3.1 Summary ...... 38 3.2 Introduction ...... 39 3.3 Materials and Methods ...... 42 3.3.1 Insects ...... 42

viii

3.3.2 Honeydew Collection ...... 43 3.3.3 DNA Extraction and PCR Protocols ...... 44 3.4 Results ...... 45 3.5 Discussion ...... 46 Chapter 4 LAND COVER DIVERSITY INCREASES ANT SPATIAL AGGREGATION TO PREY AND CONSUMPTION ...... 49 4.1 Summary ...... 49 4.2 Introduction ...... 49 4.3 Materials and Methods ...... 52 4.3.1 Study Sites and Sample Collection ...... 52 4.3.2 Landscape Analysis ...... 53 4.3.3 Spatial Aggregation Analysis ...... 54 4.3.4 Molecular Gut Content Analysis ...... 55 4.3.5 Statistical Analyses ...... 57 4.4 Results ...... 58 4.4.1 Landscape Analysis ...... 58 4.4.2 Spatial Aggregation Analysis ...... 60 4.4.3 Molecular Gut Content Analysis ...... 62 4.5 Discussion ...... 63 Chapter 5 FIELD MARGIN COMPOSITION INFLUENCES EARLY SEASON PEST AND SPIDER SPATIAL ASSOCIATIONS ...... 90 5.1 Summary ...... 90 5.2 Introduction ...... 91 5.3 Materials and Methods ...... 95 5.3.1 Study Sites, Sample Collection, and Field Margin Composition ...... 95 5.3.2 Spatial Analysis ...... 95 5.3.3 Statistical Analyses ...... 96 5.4 Results ...... 97 5.4.1 Pest and Spider Abundances ...... 97 5.4.2 Spatial Patterns of Spider and Pest Communities ...... 99 5.4.3 Spatial and Temporal Overlap of Spiders and Pests ...... 100 5.5 Discussion ...... 102 5.5.1 Field Margins on Spider and Pest Abundances ...... 102

5.5.2 Field Margins on Spider and Pest Spatial Aggregations ...... 103 5.5.3 Field Margins on Spider and Pest Spatial Overlap ...... 104 5.5.4 Abiotic Conditions on Spider and Pest Spatial Patterns ...... 104 5.5.5 Application of Field Margin by SADIE Interactions ...... 105 Chapter 6 NON-CONSUMPTIVE EFFECTS OF ANTS AND SPIDERS INFLUENCE INTRAGUILD INTERACTIONS AND SUBSEQUENT PEST DAMAGE IN SOYBEANS ...... 117 6.1 Summary ...... 117 6.2 Introduction ...... 118 6.3 Materials and Methods ...... 121 6.3.1 Experimental Field Cage Set-up ...... 121 6.3.2 Experimental Treatments ...... 122 6.3.3 Arthropod Observations ...... 123 6.3.4 Molecular Gut Content Analysis ...... 124 6.3.5 Final Plant Damage Analysis ...... 125 6.3.6 Statistical Analyses ...... 125 6.4 Results ...... 126 6.4.1Arthropod Observations ...... 126 6.4.2 Recovery of Cloverworms and Spiders ...... 128 6.4.3 Molecular Gut Content Analysis of Spiders ...... 128 6.4.4 Final Plant Damage...... 129 6.5 Discussion ...... 130 Chapter 7 IMIDACLOPRID SEED TREATMENTS AFFECT INDIVIDUAL ANT BEHAVIOR AND COMMUNITY STRUCTURE BUT NOT EGG PREDATION, PEST ABUNDANCE, OR SOYBEAN YIELD ...... 138 7.1 Summary ...... 138 7.2 Introduction ...... 139 7.3 Experimental Methods ...... 141 7.3.1 Assessment of Lethal and Sublethal Effects in the Laboratory ...... 141 7.3.2 Field Experiments ...... 144 7.4 Results ...... 148 7.4.1 Lethal and Sublethal Effects in the Laboratory ...... 148 7.4.2 Ant Community Composition ...... 149

7.4.3 Pest Abundance ...... 150 7.4.4 Predation ...... 150 7.4.5 Plant Growth and Yield ...... 151 7.5 Discussion ...... 151 Chapter 8 CONCLUSIONS AND BROADER IMPACTS ...... 162 REFERENCES ...... 167 VITA ...... 187

Table of Tables Table 2-1 Non-target species tested for cross-reactivity with the pea aphid primers (Apisum-219-F and Apisum-428-R) designed for this study...... 31 Table 4-1 Landscape attribute abbreviations and definitions used within the landscape analyses component of this study (taken from Fragstats Help 4.2. Where the class scale are computed for every class type in the landscape, and landscape scale is computed for entire patch mosaics where the resulting output is a single value for each field at each scale...... 67 Table 4-2 Non-target species tested for cross-reactivity with the Colaspis brunnea (CB348F and CB531R), Chinavia hilaris (AH276F and AH390R), and Hypena scabra (HS517F and HS598R) primers designed for this study...... 68 Table 4-3 Soybean pest species primers designed for ant molecular gut content testing...... 71 Table 4-4 Parameter estimates from regression results (* indicates P < 0.05) where ant and pest abundance were analyzed using PROC GENMOD with a negative binomial distribution against landscape and class metrics, community metrics, and abiotic conditions at 0.5km and 3.0km scales with the individual field as a random effect...... 72 Table 4-5 Parameter estimates from regression results (* indicates P < 0.05) where ant and pest diversity were analyzed using PROC GENMOD with a negative binomial distribution against landscape and class metrics, community metrics and abiotic conditions at 0.5km and 3.0km scales with the individual field as a random effect...... 73 Table 4-6 Regression results (* indicates P < 0.05) where ant and pest abundance were analyzed using PROC GENMOD with a negative binomial distribution against individual cover type (USDA- defined categories) PLANDs at 0.5km and 3.0km scales. Empty cells indicate covers that were not observed in that buffer size...... 74 Table 4-7 Regression results (* indicates P < 0.05) where ant and pest diversity were analyzed using PROC GENMOD with a negative binomial distribution against individual cover type (USDA- defined categories) PLANDs at 0.5km and 3.0km scales. Empty cells indicate covers that were not observed in that buffer size...... 75 Table 4-8 Spatial Analysis by Distance IndicEs (SADIE) results for ant and pest communities by field and generalized collection time. P_a indicates the P value for overall spatial dispersal, P_vj indicates the P value for positive aggregations/ clusters, while P_vi indicates the P value for negative aggregations/gaps...... 76 Table 4-9 Regression results (* indicates P < 0.05) where ant and pest SADIE (significant spatial trends) were analyzed using PROC LOGISTIC regression against landscape and class metrics, community metrics, and abiotic conditions at 0.5km and 3.0km scales with the individual field as a random effect...... 79 Table 4-10 Spatial Analysis by Distance IndicEs (SADIE) association (NA) results for ants to pests by field and generalized collection time. Chi_P indicates the χ value output where positive indicates that the two groups are clustered together positively and negative indicates that the two groups are avoiding each other (gaps). The probability is the association P value of χ statistic. . 80 Table 4-11 Regression results (* indicates P < 0.05) where ant and pest SADIE associations (significant spatial trends) between ants and pests were analyzed using PROC LOGISTIC regression against landscape and class metrics, community metrics and abiotic conditions at 0.5km and 3.0km scales with the individual field as a random effect...... 82 Table 4-12 Molecular gut content results by ant species tested for Colaspis brunnea, Chivaria hilaris, and Hypena scabra...... 83 Table 4-13 Regression results (* indicates P < 0.05) where ant molecular gut contents (DNA Yes/No?) were analyzed using PROC LOGISTIC regression against landscape and class metrics, community metrics, and abiotic conditions at 0.5km and 3.0km scales with the individual field as a random effect...... 84 Table 5-1 Parameter estimates of margin types (multinomial logistic regression) with aggregated pest and spider abundances as the corresponding independent variables controlled for abiotic conditions and field site...... 108 xii

Table 5-2 Parameter estimates of margin types (multinomial logistic regression) with pest abundances as the corresponding independent variables controlled for abiotic conditions and field site...... 109 Table 6-1 Main effect test (ANOVA) for presence of ants (Ln) and cloverworms (HS) on spider location...... 133 Table 6-2 Main effect test (ANOVA) for presence of ants (Ln), spiders (Os), and cloverworms (Hs) on the proportion of leaf damage...... 134 Table 7-1 Regression results per ant species when seed treatment (imidacloprid and fungicide) are taken into consideration with sampling date...... 155 Table 7-2 Regression results per pest species when seed treatment (imidacloprid and fungicide) are taken into consideration with sampling date...... 156

xiii

Table of Figures Figure 1-1 Generalized diagram of the connections established by the research in this dissertation...... 13 Figure 2-1 Generalized diagram of ant body segments and internal organs screened by molecular gut content analysis to quantify the frequency of inhibition by each representative part of the ant...... 34 Figure 2-2 Diagram of dissection stages where (A) whole ant bodies were analyzed, (B) followed by their respective body segments as described in Fig. 1, and (C) the organs within the gaster once the inhibition was determined to reside primarily within that body segment ...... 35 Figure 2-3 Percentage of samples testing positive for A. pisum using extractions of ants spiked with target prey (A. pisum) DNA. The total number of samples testing positive is given before the bracketed total number of samples tested at the base of each column...... 36 Figure 2-4 Results of the T. caespitum feeding trial. Closed dots represent the percentage of samples at each time point (0–24 h) testing positive for A. pisum DNA. Dashed lines represent the upper and lower 95% confidence intervals ...... 37 Figure 3-1 Depicts the percentage of samples testing positive for aphid species-specific DNA for each aphid species individually. “KW” represents Kim Wipe. Treatments including “Bodies” are pooled positive controls, “KW” and “Leaf” treatments are negative controls, and treatments containing “Honeydew” are samples containing only substrate material and honeydew. No variation was found within treatment types, therefore, no errors are included. (A) Percentage of Sitobion avenae samples testing positive for DNA using species-specific primers. (B) Percentage of Rhopalosiphum padi samples testing positive for DNA using species-specific primers. (C) Percentage of Acyrthosiphon pisum samples testing positive for DNA using species-specific primers...... 48 Figure 4-1 Three sampled field sites (dots), with 0.5 km and 3.0 km buffers around the corresponding year’s Crop Data Layer. Sites indicate the range in landscape diversity (SIDI), average percentage of landscape per cover type (PLAND), total edge length (TE), and cover type clumpiness (CLUMPY). The mapping projection (WGS 1984) causes buffers to appear oblong. 85 Figure 4-2 The average percentage of the landscape (PLAND) by USDA Crop Data Layer land cover classification for (A) 0.5 and (B) 3.0 km buffers...... 86 Figure 4-3 Graph of the parameter estimates associated with (A) ant and pest abundance with landscape diversity (SIDI), (B) ant and pest diversity with SIDI, (C) ant and pest abundance with the class metrics of average percentage of landscape (PLAND) and average total edge distance (TE) and average clumpiness of covers (CLUMPY), and (D) ant and pest diversity with PLAND, TE, and CLUMPY. Positive parameter estimates indicate variables that increase the respective arthropod community while negative estimates indicate the opposite effect...... 87 Figure 4-4 Relationship of ant and pest community abundances using SADIE association analyses for one select field with significant aggregations of (A) ants and (B) pests and (C) significant association of ants to pests. Increased color saturation indicates an increase in population density or aggregation within the field while a lack of color indicates gaps of the populations within the fields. In (C), purple saturation indicates spatial associations of ant and pest aggregations...... 88 Figure 4-5 Graph of the parameter estimates associated for ant SADIEs, pest SADIEs, SADIE association of ants to pests, and predation (PCR) with the landscape metrics of diversity (SIDI) and fragmentation (PLAND, TE, and CLUMPY) for (A) 0.5 km buffers and (B) 3.0 km buffers. Positive parameter estimates indicate variables that increase the respective metric while negative estimates indicate the opposite effect...... 89 Figure 5-1 Potential field margin distance by within field species interactions of soybean pests and predators...... 110 Figure 5-2 Field border examples of (A) other crop, grass/weed, and tree/shrub and (B) soybean, tree/shrub, and road...... 111

xiv

Figure 5-3 (A) The relative proportion of spider families represented within the samples (n = 1282) and (B) the relative proportion of the pest groups (some families, some species) found. 112 Figure 5-4 Parameter estimates of aggregate spider and pest within field spatial aggregations given field margin type for (A) 0-10 m and (B) 10-20 m...... 113 Figure 5-5 Odds ratios for spider and pest community variables in how they contribute to the likelihood of seeing a positive spatial association between spiders and pests. Where spider and pest abundances are per unit odds ratios and SADIE variables are categorical with presence: absence. Any value < 1 indicates a decrease in likelihood while and value > 1 indicates an increase in likelihood...... 114 Figure 5-6 The number of fields containing significantly positive spatial associations between aggregate pest and spider communities over the growing season...... 115 Figure 5-7 SADIE association map indicating general spider by pest association in relation to margin type at 0-10 m. Aggregate pests are represented in blue and spiders in red. Associated patches are purple and gaps are white...... 116 Figure 6-1 The total number of striped lynx spiders, Oxyopes salticus (Hentz), observed at each location (legend) by treatment containing spiders over four days. Os represents Oxyopes salticus, Ln Lasius neoniger, and Hs Hypena scabra...... 135 Figure 6-2 (A) The proportion of the initial fifteen second instar green cloverworms including pupa, Hypena scabra Fabricius (Hs), and (B) lynx spiders, Oxyopes salticus (Hentz) (Os), per cage recovered. Ln indicates the presence of the cornfield ant, Lasius neoniger L. Differences between treatments were determined by the Tukey (HSD) test...... 136 Figure 6-3 The proportion of leaf area damaged after four days. Ln indicates the presence of cornfield ants, Lasius neoniger L.; Os indicates the presence of striped lynx spiders, Oxyopes salticus (Hentz), and Hs indicates the presence of green cloverworms, Hypena scabra Fabricius. Lowercase letters above standard error indicate significant differences between treatments as denoted by the Tukey (HSD) test...... 137 Figure 7-1 Diagram of replication and steps involved in the testing via exposure and behavioral assays of lethal and sublethal effects of seed treatments on Tetramorium caespitum L...... 157 Figure 7-2 The proportion of ants (Tetramorium caespitum L.) classified as normal, intoxicated (unable to right itself and excessive grooming), or dead after 48 h exposure to seed treatments in the laboratory...... 158 Figure 7-3 Change in the average number of lines crossed per minute over 10 min between pre- and post-exposure by ants (Tetramorium caespitum L.) as an indication of sublethal impacts of imidacloprid on ant behavior...... 159 Figure 7-4 Ant community composition metrics –species richness, evenness, and Shannon diversity – as observed between seed treatments in the field...... 160 Figure 7-5 (A) Survival log-rank curves for ant (Formicidae) discovery time of egg masses within 1-5 h of exposure in the field and (B) Number of black cutworm (Agrotis ipsilon (Hufnagel)) eggs surviving in the field after 5 h...... 161

Chapter 1 INTRODUCTION

This chapter provides an introduction to relevant concepts and organismal ecology to motivate the work found within this dissertation. In this dissertation, mapping techniques through the use of Geographic Information Systems (GIS) technology, GPS, and Spatial Analysis by Distance IndicEs (SADIE) were combined with modified molecular gut content analysis (identification of DNA by

Polymerase Chain Reaction (PCR) for characterization of predation events in

Chapter 2 and Chapter 3) to provide an spatially-referenced examination of how land cover alters the within field dispersal of ants and spiders and subsequent movement to and consumption of soybean pests (Chapter 4 and Chapter 5).

Field cage trials were used to further examine ant interactions with a soybean pest in response to intraguild competition provided by a common, co-occurring spider species (Chapter 6). Finally, a lab and field study was conducted to determine if a seed treatment, used on much of the soybean acreage in the

United States, impact ant communities so as to be detrimental to soybean yields and ant-mediated biological control (Chapter 7). Each component connection

(Figure 1-1 arrows) integrates community dynamics (i.e. abundance, diversity) of the predators and pests, the spatial and temporal overlap of these populations, and the subsequent, often consumptive, interactions taking place within conventionally-managed Kentucky soybeans.

1.1 General Background

Demand for food, fuel, and fiber has been increasing, escalating global agricultural intensification (Godfray et al. 2010) though arable land has been declining (Hobbs 2007) and pests have developed resistance to control methods. A greater understanding of the impact of landscape cover on agricultural production is required to meet the continual increase in human demand as these covers determine what insects will thrive. This research will help explain how land cover at multiple scales impacts arthropod communities that, in turn, dictate ecosystem services (i.e. biological control of pest species) in agricultural areas.

1.1.1 Soybean Production

Soybeans are an ecologically and economically relevant crop worldwide, representing 58% of oilseed production, 35% from the United States (ASA 2016).

Soybeans in the United States are second only to field corn in economic relevance, selling at an average of $10.70 per bushel in 2010 (ASA 2016). The continental United States cultivates 77.4 million acres of soybean, accounting for

30% of planted agricultural land with 30 U.S. states growing over 3.3 billion bushels (1.59 billion exported). Soybean crops contribute 25% of Kentucky’s agricultural income (ASA 2016), approximately $1 billion US grown on 1.84 million acres.

1.1.2 Insect Pests of Soybean

A variety of insect pest species are found in soybeans with densities increasing throughout the season as leaf area increases (Price 1976). Invasive species such as the soybean aphid Aphis glycines Matsumura (Hemiptera:

Aphididae), the brown marmorated stink bug Halyomorpha halys (Hemiptera:

Pentatomidae), and the kudzu bug Megacopta cribraria F. (Hemiptera:

Plataspidae) have made soybean pest control more expensive and time- consuming in the United States (Takasu and Hirose 1986, Ragsdale et al. 2007,

Khrimian et al. 2008, Greene et al. 2012). Not only do these pests cause feeding damage, but species like A. glycines and Myzus persicae Sulza (Hemiptera:

Aphididae) are also capable vectors of devastating plant viruses (e.g. soybean mosaic virus) (Hill et al. 2001, Ragsdale et al. 2011). Current chemical control methods in Kentucky include broad-spectrum sprays and neonicotinoid seed treatments for preventative uptake into seedlings (Johnson et al. 2015, Johnson

2016).

1.2 Biological Control

Agricultural monocultures are greatly impacted by arthropods because they contain a surprising number of both pests and predators (Welch et al. 2012).

Pests adapt to commonly used insecticides which can prove harmful for both people and the environment (Brown 1978). To subvert this chemical resistance, biological control practices can be implemented to make agricultural systems more attractive to natural enemies (Barbosa and Castellanos 2005). Simple

systems like soybeans have low plant diversity (Frank et al., 2011) and fast prey turnover (Costamagna et al. 2007), allowing for the facilitation of pest suppression by predators (Barbosa and Castellanos 2005) and stimulation of plant growth (Wiedenmann and Smith 1997, Law and Rosenheim 2011). These predators can either be introduced into the agricultural system or encouraged to stay within the system (conservation biological control) via environmental manipulations such as hedgerows, weed strips, and mulch applications

(Wissinger 1997, Blubaugh and Kaplan 2015).

1.2.1 Importance of Generalist Predators

As biological control agents, natural enemies must be rapid colonizers, maintain a presence as pests decline, and be opportunistic feeders (Symondson et al. 2002). Not only do generalist predators (no specific prey requirements

(Koss and Snyder 2005)) fulfill these requirements, they do well in complex landscapes and can rapidly adjust to environmental changes and prey availability

(Welch et al. 2012, Welch and Harwood 2014). Generalist predators also exhibit pest control early in a growing season due to the generalist’s ability to consume alternative prey items such as springtails, allowing predator populations to increase in abundance before pest species multiply (Butler and O'Neil 2007).

Although alternative prey items can influence the pest suppression of other species, the ability of general predators to feed on pests for the entirety of the field season is invaluable (Barbosa and Castellanos 2005, Prasad and Snyder

2010). Arthropod generalist predators are an important component of biological

control, and in soybean can maintain a low, but constant, rate of predation under field conditions (O'Neil and Wiedenmann 1987).

Ants (Hymenoptera: Formicidae), a plentiful and adaptable generalist predator group (Sudd and Franks 2013), are both primary and secondary consumers that provide important ecosystem services such as soil enrichment, pest suppression, pollination, and seed dispersal (Huxley 1991, Woodell et al.

1991, Way and Khoo 1992, Wang et al. 1995, Perfecto and Castineiras 1998).

Ants have been proven to be excellent biological control agents because colonies are not easily satiated, colonies do not relocate due to food shortages, individual satiation does not impact the predatory behaviors of the individual, ants are responsive to aggregations of prey items, and ants forage for long periods seasonally and diurnally (Paulson and Akre 1992). Ant predation is also not restricted to prey life stage or colony lifespan (Perfecto 1991, Fleet and Young

2000). Biological control using methods to enhance ant presence have been successful in controlling fly populations and enhancing fruit quality and production in several crops including maize, African mahogany and a variety of fruit trees (Lagnaoui et al. 2000, Mele and Vo The 2002, Peng and Christian

2005a, b, Peng et al. 2011). However, there are large gaps in knowledge of ant biological control within the continental United States including aspects of landscape variability and how it relates to ant predation rates (Crowder and

Snyder 2010).

The propensity of ants to protect plants from herbivore damage can be mitigated by the tendency of some ant species to farm honeydew-producing

hemipterans (Cushman et al. 1991, Wyckhuys et al. 2007) and be pests themselves (Banks et al. 1991). Although this activity can increase herbivore numbers on plants (Herbert and Horn 2008), it protects plants by increasing ant presence which reduces mold infestations caused by excess honeydew (Buckley

1990). Alternatively, such an ant presence has been known to interfere with other predators (intraguild competition), both specialist and generalist in hemipteran- infested agricultural systems such as citrus (Buckley 1990, Bownes et al. 2014,

Calabuig et al. 2015). Ant aggression towards other predators in agricultural systems can be especially pronounced when those ant species are invasive such as the Argentine (Linepithema humile Mayr (Hymenoptera: Formicidae)) and red imported fire ants (Solenopsis invicta Buren (Hymenoptera: Formicidae)) (Human and Gordon 1996, Heterick et al. 2002, Epperson and Allen 2010).

Another such generalist predator, spiders, have received some attention in recent years for their role as biological control agents in agricultural systems. The ability of spiders to disperse great distances makes them excellent biological control in areas of high disturbance such as crop fields (Clark et al. 2004). Young

(1990) studied spiders in agroecosystems and found 614 species and 192 genera in 26 families with 62 and 78 species in the families Lycosidae and

Linyphiidae respectively. Even though there are many species of spiders common in the U.S., only two guilds of spiders exist– web-builders and wandering spiders (Uetz et al. 1999). All spiders are predatory primarily on insects (Riechert and Lockley 1984), but according to Riechert and Harp (1987), four attributes determine how well spiders capture prey – prey availability, the

nutritional value of prey, spider body temperature, and risk of predation. Spiders also normally require behavioral cues in order to select an optimal foraging site and initiate attack and feeding on prey items (Riechert and Harp 1987).

Maloney et al. (2003) found that spiders exert top-down control of herbivores such as Hemipterans, Coleopterans, and Dipterans (Riechert and

Lawrence 1997). However, they also found that spiders are more sensitive to insecticides than many of their prey species. Minimal usage of chemical controls and applications of mulch for shelter allow spider populations to flourish in cropping systems. The impacts spiders have on pest populations often fluctuate with climate and spider population density (Nyffeler and Sunderland 2003). For example, a study by Vichitbandha and Wise (2002) speculated that low populations of pest species or low ambient densities of spiders were to blame for the lack of difference in soybean yields in treatments with and without spiders.

However, little is known about spider ecology in relation to differences in landscape cover and composition and needs to be examined in further detail.

1.2.2 Intraguild Competition

Intraguild competition is a possible complication of using generalist predators for pest control. This can come in the form of predation or parasitism of other predators, directly competing for similar prey items, or altering foraging behavior due to the presence of another predator (Polis et al. 1989, Rosenheim

2007). The resulting influence of intraguild interference can cause fluctuations in predator abundances and, therefore, pest populations (Symondson et al. 2002).

Generally, a greater diversity of terrestrial arthropod predators increases the likelihood of intraguild interference (Snyder and Wise 2001). But the influence this interference has on herbivore predation rates depends on the predator species composition, density, and sex ratios (Takizawa and Snyder 2011).

Intraguild interference can play a large part in determining the predator population dynamics of a field (Sanders and Platner 2007). Ant and spiders, both generalist predators found in agricultural systems, share the same trophic level and are found in high densities in similar habitats (Bruning 1991, Halaj et al.

1997). Although intraguild competition is always a concern with generalist predators used as biological control, this should not dissuade growers from using generalist predators as biological control agents because this interference does not necessarily inhibit the ability of predators to suppress pests (Rosenheim et al.

1993, Lang 2003, Law and Rosenheim 2011) as the available food supply and foraging territory as determined by landscape cover and fragmentation around crop fields could mediate ant by spider interactions.

1.3 Molecular Gut Content Analysis

Predators require regulation of multiple nutrients by balancing their intake of various prey (e.g. aphids) since prey items contain varying nutrient content

(Fagan et al. 2003, Mayntz et al. 2005). This differential preference for food items could influence the feeding habits of predators under field conditions. Ants, in particular, have been known to both farm aphids for honeydew and prey upon the same aphid colony to meet their dietary needs (Hölldobler and Wilson 1990);

however, farming aphids is detrimental to plant health whereas consuming aphids is not. Therefore, knowing whether an ant has benefitted the pest population or the plants is important for determining which ant species are suitable for biological control agents. One way of elucidating what predators are consuming is by estimating the within field abundances of all potential prey items and assuming that predators will consume them at approximately the encounter rate (Thomas et al. 2009, Schmidt et al. 2012); but not all predators interact with prey species this way (Athey et al. 2016). Therefore, the use of molecular gut content analysis or detection of prey species DNA in predator guts can be used to reliably identify predator-prey linkages (Hoogendoorn and Heimpel 2002,

Symondson 2002, Harwood et al. 2004, King et al. 2008).

1.4 Landscape Ecology

Since agricultural fields typically require large amounts of inputs and compose a large percentage of the landscape, they are subjected to a great deal of influence by landscape heterogeneity (Altieri 1999), the effects of which will be looked at in some detail in this study. Landscape structure and heterogeneity are comprised of several components – land cover (i.e. forest, grasslands), land use

(i.e. grasslands for hay versus pasture), and fragmentation (Fahrig and Jonsen

1998, Fahrig et al. 2011). These elements work in conjunction with each other to determine the landscape diversity (number of cover types/uses) and connectedness (size and fragmentation of each type/cover) that a given species can use. The scale of these interactions is also vital to consider as area

requirements of each species depend on the species size, nutritional requirements, and dispersal mechanisms (Gabriel et al. 2006).

1.4.1 Impact of Landscape on Biodiversity

The increasing cultivation of arable land in North America is dramatically altering landscape composition, and negatively impacting plant diversity and arthropod diversity (Chapin III et al. 2000, Hooper et al. 2000, Henle et al. 2008).

The degree of arthropod biodiversity in agroecosystems depends on the vegetative diversity of the field, the permanence of the cropping system, degree of disturbance, and the degree of isolation from natural areas (Altieri 1999).

Chisholm et al. (2011) found that multiple levels of arthropod diversity are associated with high quality, low disturbance landscapes. Overall, landscapes with less farmed land have significantly more arthropod diversity than cultivated fields (Clough et al. 2005, Clough et al. 2014). Several studies have shown little difference in diversity due to organic or conventional farming methods. What they did find is that the surrounding landscape had a significant impact on in-field diversity whether the system was organic or conventional (Gibson et al. 2007).

Studies have also found that increasing the amount of corn or decreasing overall landscape diversity in a given area will result in a decrease in natural pest control in the adjacent soybean fields (Landis et al. 2008, Gardiner et al. 2009, Gardiner et al. 2010). Landscape types can enable these pests to spread and flourish as differing rates and needs to be considered when dealing with pest control issues

(Bianchi et al. 2006).

Not only does cultivated land typically decrease the amount of arthropod biodiversity but even within such systems, cultivated field margins are especially barren (Meek et al. 2002). The interface between cropland and all other land covers is of great importance because it is here that predators and herbivores enter a field. The edge effect is defined as “the impact on species richness in the area between two adjoining habitats” (Dauber and Wolters 2004). However, exactly how to measure or define an edge is difficult. This could be due to the confounding effects of the differences in matrix or direct and indirect biological differences in and around an edge (Murcia 1995). These edge interfaces with the field can be beneficial and detrimental for farmers – allowing an influx of predators but also of pest species (Tscharntke et al. 2005). Despite the impact of edge effects, on the whole, landscape composition has the much more significant impact on diversity (Donovan et al. 1997, Gabriel et al. 2006).

1.4.2 Landscape Impacts on Biological Control

Biological control needs to be considered when making large-scale landscape management practices (Zhang et al. 2007) because predators react differently to disturbance at different spatial scales (Steffan-Dewenter et al. 2002,

Frank et al. 2011). Habitat management to supplement pest control involves the control of not just the crop and farmlands but also the surrounding landscapes

(Landis et al. 2000). Altieri (1999) noted that high crop diversity and mixing crops temporally and spatially is important to maintaining low pest populations, particularly when small plots of crops are intermingled with uncultivated lands

because these areas are vital to maintaining the population of predators within heavily disturbed fields (Benton et al. 2003, Schmidt et al. 2005, Gardiner et al.

2009, Gardiner et al. 2010). For instance, Bianchi et al. (2006) found that natural enemy populations were 80% enhanced within herbaceous habitats and 71% within woody habitats around agricultural fields. Habitat and landscape composition impact the level of pest suppression by predators (Madsen et al.

2004) by protecting higher trophic levels from local extinction due to disturbance regimes by providing refuge areas (Finke and Snyder 2010) or by supplementing the generalist diets with alternative prey items, allowing predators to live longer with greater fecundity (Landis et al. 2000). Therefore, the more that is known about how landscape management and composition impacts predator diets and intraguild interactions in crops, the more conservation biological control can be manipulated for specific field sites and pest problems.

Figure 1-1 Generalized diagram of the connections established by the research in this dissertation.

Chapter 2 OVERCOMING PCR INHIBITION DURING DNA-BASED GUT CONTENT ANALYSIS OF ANTS

Chapter contents published in Penn, HJ, EG Chapman, and JD Harwood. 2016. Overcoming PCR Inhibition for Ant DNA Gut Content Analysis. Environmental Entomology (doi: 10.1093/ee/nvw090).

2. 1 Summary

Generalist predators play an important role in many terrestrial systems, especially within agricultural settings, and ants (Hymenoptera: Formicidae) often constitute important linkages of these food webs, as they are abundant and influential in these ecosystems. Molecular gut content analysis provides a means of delineating food web linkages of ants based on the presence of prey DNA within their guts. Although this method can provide insight, its use on ants has been limited, potentially due to inhibition when amplifying gut content DNA. We designed a series of experiments to determine those ant organs responsible for inhibition and identified variation in inhibition among three species (Tetramorium caespitum (L.), Solenopsis invicta Buren, and Camponotus floridanus (Buckley)).

No body segment, other than the gaster, caused significant inhibition. Following dissection, we determined that within the gaster, the digestive tract and crop cause significant levels of inhibition. We found significant differences in the frequency of inhibition between the three species tested, with inhibition most evident in T. caespitum. The most effective method to prevent inhibition before

DNA extraction was to exude crop contents and crop structures onto UV- sterilized tissue. However, if extracted samples exhibit inhibition, addition of bovine serum albumin to PCR reagents will overcome this problem. These methods will circumvent gut content inhibition within selected species of ants,

thereby allowing more detailed and reliable studies of ant food webs. As little is known about the prevalence of this inhibition in other species, it is recommended that the protocols in this study are used until otherwise shown to be unnecessary.

2. 2 Introduction

As voracious predators, ants have the potential to exert significant pressure on their herbivorous prey as they exhibit a wide array of feeding preferences and can consume arthropods of all sizes (Hölldobler and Wilson

1990). Their ubiquity in urban, forested, and agricultural areas, coupled with their diverse feeding habits, suggests their consumption patterns could have important ramifications for food web dynamics (Philpott et al. 2010). For instance, Formica exsectoides Forel (Hymenoptera: Formicidae) has been identified as a major predator of forest pests in pine stands (Campbell 1990), significantly reducing the population of gypsy moth (Lymantria dispar dispar L. (Lepidoptera: Erebidae)), jack pine budworm (Choristoneura pinus Freeman (Lepidoptera: Tortricidae)), redheaded pine sawfly (Neodiprion lecontei (Fitch) (Hymenoptera: Diprionidae)), and white pine weevil (Pissodes strobe Peck (Coleoptera: Curculionidae)).

Furthermore, multiple ant species are important predators of tropical agricultural pests as diverse as coffee berry borer (Hypothenemus hampei (Ferrari)

(Coleoptera: Curculionidae)) in coffee (Gonthier et al. 2013, Trible and Carroll

2014), fall armyworm (Spodoptera frugiperda (J. E. Smith) (Lepidoptera:

Noctuidae)) and corn leafhopper (Dalbulus maidis (De Long and Wolcott)

(Hemiptera: Cicadellidae)) in maize (Perfecto 1990), the oriental fruit fly

(Bactrocera dorsalis (Hendel) (Diptera: Tephritidae)) in citrus (Diame et al. 2015) and mango leaf-hoppers (Idioscopus nitidulus (Walker) (Hemiptera:

Cicadellidae)) in mango (Peng and Christian 2005a). However, their interactions within the food webs of many other systems remain unclear. To better characterize the impact of ants as predators, it is imperative to map the structure of their food webs and establish general consumption patterns (Choate and

Drummond 2011) because, as with other generalist predators, predation frequency is not always correlated to prey availability.

Methods previously used to characterize ant food webs have often been approximate and vague. Visual assays of ant foraging have been extremely difficult to conduct, especially on night-foraging ants, and provide relatively unreliable estimates of predation (Agarwal et al. 2007). Stable-isotope studies have long been employed in ant systems but only estimate the general trophic position of an ant species and do not determine the specific prey being consumed (Blüthgen et al. 2003) or the structure of food webs. Molecular gut content analysis, a postmortem method of identifying trophic linkages, is highly sensitive at identifying prey-specific DNA within predator guts (Symondson 2002,

Sheppard and Harwood 2005, King et al. 2008, Weber and Lundgren 2009). This approach could be an excellent diagnostic tool to delineate trophic linkages and estimate predation patterns of ants on key prey species including crop pests, forest insects, and species occurring in the urban environment.

The few studies utilizing molecular gut content analysis on ants have revealed the difficulties in obtaining prey DNA from ant samples using whole body

extractions. Fournier et al. (2008) tested 24 ant foragers for glassy-winged sharpshooter (Homalodisca vitripennis (Germar) (Hemiptera: Cicadellidae)) DNA using PCR but none tested positive, despite the discrepancy of 12% testing positive using enzyme-linked immunosorbent assays for the same prey.

Similarly, a study investigating predation of Camponotus pennsylvanicus

(DeGeer) (Hymenoptera: Formicidae) on red oak borers (Enaphalodes rufulus

(Haldeman) (Coleoptera: Cerambycidae)) revealed very few field-caught specimens tested positive (1.7%) despite high levels of predation being observed on artificially placed egg masses (Muilenburg et al. 2008). These studies did not examine if ant gut content DNA extraction and amplification were being inhibited, although this phenomenon has been found in the ground beetle Poecilus versicolor (Sturm) (Coleoptera: Carabidae) (Juen and Traugott 2006). Within ants, specifically, Gadau (2009) postulated that PCR inhibition might be due to interference from glands within the gaster, where the majority of the adult ant digestive tract is located. The uncertainty in detection of DNA in ant guts and unreliability of data thus obtained indicate the importance of characterizing the effect of ant organs on DNA detection and amplification success.

To determine if, and which, body segments and organs inhibit extraction and/or PCR of prey DNA from ant guts, we conducted a series of experiments to characterize detection success. Based on the assumptions of Gadau (2009), we hypothesized that the gaster, containing the poison and Dufour’s glands, causes a degree of PCR inhibition in ants that could lead to incorrect interpretation of molecular gut-content data from the field. Furthermore, we examined whether

this effect could be observed in three species of ants (Tetramorium caespitum,

Solenopsis invicta, and Camponotus floridanus) and determined two efficient mechanisms to overcome inhibition, when present.

2. 3 Materials and Methods

2.3.1 Insects

Colonies of Tetramorium caespitum (L.) (Hymenoptera: Formicidae) were selected, based on their prevalence in urban and agricultural settings (López and

Potter 2000, 2003). Solenopsis invicta Buren (Hymenoptera: Formicidae), an invasive species, was identified as an optimal ant to use due to its predaceous habits and potential use for biological control in urban and agricultural systems within the southeastern USA (Adams et al. 1983, Eubanks 2001, Eubanks et al.

2002), and Camponotus floridanus (Buckley) (Hymenoptera: Formicidae) was selected based on its common use as a model species for molecular work

(Gadau et al. 1996, Simola et al. 2013). T. caespitum, S. invicta, and C. floridanus were collected from field sites in Lexington, Kentucky, U.S.A. (38° 6'

55.0794" N, 84° 30' 20.52" W), St. Cloud, Florida, U.S.A. (28° 10' 39.36" N, 81°

14' 37.6794" W), and Micanopy, Florida, U.S.A. (29° 29' 57.1194" N, 82° 19'

45.12" W), respectively. Five colonies within 0.2 hectares from each other were collected from each species with the number of individuals sampled in proportion to the colony size. Specimens were collected using sterilized soft-forceps, individually stored in autoclaved 1.5 ml microcentrifuge tubes with 95% ethanol, and maintained at -20°C until molecular screening.

In addition to preserved wild-caught specimens, live T. caespitum colonies for use in determining the amount of liquid housed in the crop and DNA decay rate trials were excavated from the Spindletop Research Farm, Lexington,

Kentucky, U.S.A., and allowed to settle into 19 liter (5 gallon) buckets for 3 days before brood, workers, and queens were sorted into laboratory nests (Smith and

Tschinkel 2009). Colonies had access to water and were fed a diet of honey

(ALDI, Inc., Batavia, Illinois, U.S.A.), peanut butter (ALDI, Inc., Batavia, Illinois,

U.S.A.), and freshly killed Drosophila melanogaster Meigen (Diptera:

Drosophilidae) in order to maintain colony function without introduction of experimental prey DNA (A. pisum). All colonies were maintained under controlled conditions of 25 ± 1 °C, 65 ± 5% RH, and 10: 14 L:D cycle.

Pea aphids, Acyrthosiphon pisum (Harris) (Hemiptera: Aphididae), were selected as a target pest based on relevance as a model organism in molecular studies (Ferrari et al. 2012, Russell et al. 2013) and abundance in Kentucky

(Pass and Parr 1971, Green et al. 2015). Additionally, no A. pisum were observed near the collection sites of ant colonies, ensuring that the ants did not have A. pisum DNA in their guts prior to testing, Aphids were removed from colonies already established on Vicia faba (L.) (Fabales: Fabaceae) in greenhouse colonies at the University of Kentucky in Lexington, Kentucky. All colonies were kept under controlled conditions of 25 ± 1 °C, 65 ± 5% RH, and

16:8 L:D cycle.

2.3.2 Dissection Treatments

Dissections were conducted to optimize the extraction and amplification of prey DNA from ant guts and determine if, and which, major body segments and corresponding organ systems were inhibiting these processes. Preserved specimens of T. caespitum, S. invicta, and C. floridanus were used for this experiment (n = 90 per species). Stored ants were surface sterilized. The treatments included whole bodies, major body segments (head, mesosoma, legs, and gaster), and organs within the gaster (crop including proventriculus, entire gut tract, Dufour’s gland, and poison gland) (Figure 2-1).

To determine if the organs were inhibiting extraction and amplification of small quantities of non-ant DNA, a controlled amount of prey DNA was added to all dissection treatments (Figure 2-2). The quantity of prey mixture used was determined by allowing live ants from five T. caespitum laboratory colonies to feed on the A. pisum mixture (freshly killed and ground A. pisum, 0.5 mL deionized water, and orange neon gel food color (Betty Crocker®, Signature

Brands, LLC., Ocala, Florida, U.S.A.)). This prey combination was used to encourage forager feeding and subsequent identification (food coloring) in lieu of foragers taking whole prey items back to the nest for larval consumption.

Foragers were allowed to feed until satiated (approximately two to three minutes depending on forager size) and began moving away from the food source, then immediately killed in 95% ethanol to obtain a maximum crop capacity and reduce variation introduced by ingestion time, and surface sterilized to minimize external contamination (Andrews 2013). Their gasters were separated from the rest of

their body, and the contents of the crop extracted through the petiole onto UV- sterilized Kim Wipes (Kimberly-Clark Professional®, Roswell, Georgia, U.S.A.) using sterile forceps. In every instance, the crop contents were colored orange by the food coloring, indicating that the prey mixture had reached the crop by the time of dissection. Colored areas of the Kim Wipes were measured for three ants from each colony and compared to known amounts of aphid mixture pipetted onto Kim Wipes. The diameter of the known volume of A. pisum mixture that matched the average of the crop contents in diameter was selected (0.05 µL) and was used to determine the quantity of prey DNA used to spike all dissection samples.

A set of samples of crop contents without the presence of the crop structure was obtained from T. caespitum (n = 15). Colonies of T. caespitum were starved for a period of 7 days before the testing then provided the A. pisum mixture and killed as above. Within this time period, no carbohydrate sources were provided for foraging worker consumption and any colony food stores would not contain easily accessible carbohydrates or prey containing A. pisum DNA.

Crop contents were immediately removed, the volumetric data obtained (as above) and used as a comparison against crop dissection treatments.

2.3.3 DNA Extraction and PCR

Immediately following collection, samples were homogenized in 180 µL of tissue lysis buffer. Total DNA was extracted from all samples using Qiagen

DNeasy Blood and Tissue Kits© (Qiagen Inc., Valencia, California, U.S.A.)

following the animal tissue protocol. Samples on Kim Wipe papers were extracted with the paper intact. DNA was then stored in clean 1.5 mL microcentrifuge tubes at -20°C until PCR analysis. A positive control (A. pisum) was extracted with each set of dissected treatments.

To facilitate the design of A. pisum primers and test amplification rates of ant gut content DNA, a data matrix was constructed that contained cytochrome c oxidase subunit I (COI) sequences from A. pisum and a variety of aphid taxa

(Chapman et al. 2010). Our goal was to design a primer pair that would amplify

A. pisum DNA without amplifying DNA from T. caespitum, S. invicta, or C. floridanus. COI was amplified from colony specimens of A. pisum with the general COI primers LCO-1498 (Folmer et al. 1994) and HCO-700me (Breton et al. 2006). Polymerase chain reactions (total volume = 25 µL) consisted of 1X

Takara buffer (Takara Bio Inc. Shiga, Japan), 0.2 mM of each dNTP, 0.2 mM of each primer, 0.625 U Takara Ex Taq and template DNA (1 µL of total DNA). PCR was carried out using a Bio-Rad C1000 thermal cycler (Bio-Rad Laboratories,

Hercules, California, U.S.A.). The PCR cycling protocol for Takara reagents was an initial denaturation of 94 °C for 1 minute followed by 40 cycles of 94 °C for 60 s, 40 °C for 60 s, and 72 °C for 60 s. PCRs were cycle sequenced using labeled dideoxynucleotides (ABI Big-Dye Terminator mix v. 3.0; Applied Biosystems,

Foster City, California, U.S.A.; ABI sequencer). The separation of cycle sequencing reaction products was done in Applied Biosystems 3730XL or 3730

DNA Analyzers at the Advanced Genetic Technology Center at the University of

Kentucky.

A. pisum primers were designed with Primer3 (Rozen and Skaletsky 2000) and the forward primer (Apisum-219-F) sequence being 5’ –

GTCCTGATATATCATTTCCTCGC – 3’ and the reverse (Apisum-428-R) 5’ –

AAATTGATGAAATTCCTGCTAGG – 3’. The primers have an optimal annealing temperature of 63 °C and produce a 210 bp amplicon. The PCR cycling protocol for Takara reagents was 94 °C for 1 minute followed by 40 cycles of 94 °C for 45 s, 63 °C for 45 s, and 72 °C for 30 s. Primers were tested against 184 non-target samples to ensure specificity, and DNA half-life of A. pisum DNA in T. caespitum guts calculated using the abovementioned primers.

The total DNA in all samples was first amplified with the above A. pisum- specific primers. All PCRs (12.5 µL total volume) consisted of 1X Takara buffer

(Takara Bio Inc. Shiga, Japan), 0.2 mM of each dNTP, 0.2 mM of each primer,

0.31 U Takara Ex TaqTM and template DNA (1 µL of total DNA). PCRs were carried out as above. To determine the viability of the extractions not testing positive for A. pisum DNA, samples were screened using general COI primers

Jerry (Simon et al. 1994) and Ben3R (Villesen et al. 2004a). These reactions were carried out using Bio-Rad PTC-200 and C1000 thermal cyclers (Bio-Rad

Laboratories, Hercules, California, U.S.A.). The PCR cycling protocol for the general COI primers Jerry and Ben3R (with Takara reagents as above) was 94

°C for 1 minute followed by 35 cycles of 95 °C for 60 s, 47 °C for 60 s, and 72 °C for 90 s. Following amplification, reaction success was determined by electrophoresis of 10 µL PCR product pre-stained with GelRed nucleic acid gel stain (1X Biotium, Hayward, California, U.S.A.) on 2% SeaKem agarose (Lonza,

Rockland, Maine, U.S.A.). In all cases, sets of PCRs contained one positive control of only A. pisum DNA and one negative control of all reagents without the addition of sample DNA. Additionally, any reactions not testing positive with the general COI primers (n = 2) were assumed to be unreliable and were not factored into the results. Addition of bovine serum albumin (BSA), as used in previous studies (Juen and Traugott 2006), was also examined for prevention of

PCR inhibition in dissection treatments that tested negative for A. pisum primers but positive with general COI primers. Samples of all ant species meeting this criteria were treated with 0, 0.23, 0.25, 0.27, 0.30, and 0.33 µL BSA per 1 µL of

DNA (n = 36) (Hoogendoorn and Heimpel 2001, Egert et al. 2004) during PCR.

Feeding trials were undertaken to determine the rate of amplification success for ants after consumption and subsequent digestion of A. pisum. For the feeding trials, colonies of T. caespitum were starved for a period of 7 days before the testing began. Colonies were then provided with A. pisum mixture as described in detail in the Methods then immediately removed and transferred into individual plastic containers with a water supply. Twenty ants were immediately processed as below and the remaining ants were then given time to digest their meals for 1, 2, 4, 8, 12, 16, and 24 h (n = 16-20 ants/time period). Ants were killed and stored in 95% ethanol at -20°C until extraction. Following collection of foragers, crop contents excluding all body parts were obtained. This was done by separating the gaster from the rest of the body, and squeezing out the contents of the crop through the petiole onto UV-sterilized Kim Wipe papers. The portion

of the papers containing crop contents was excised, extracted, and analyzed in the same fashion as the previously discussed dissection treatments.

2.3.4 Statistical Analyses

Frequency of inhibition in whole body extractions of each ant species was compared against A. pisum positive control samples. A second analysis examined which of the major body segments elicited inhibition including ant species, body segment, and species by segment interaction terms. To determine which organs in the targeted body segment were causing inhibition, a third model was completed using species, organ, and species by organ interaction terms.

The inhibition rates of the crop were then compared to those of the crop contents

(without the organ structure included) for T. caespitum. Finally, the influence of

BSA addition to samples previously exhibiting inhibition was analyzed in a similar fashion using BSA presence/absence, ant species, and organ/body segment as model terms. All five models were completed using Firth’s penalized-likelihood logistic regression using PROC LOGISTIC in SAS 9.4 (SAS Institute, Cary, North

Carolina) (Firth 1993, Fort and Lambert-Lacroix 2005, Wang 2014). The detectability of A. pisum DNA in T. caespitum guts over time (feeding trial) was calculated using a Probit model (PROC PROBIT) in PC SAS version 9.4 (SAS

Institute, Cary, North Carolina, U.S.A.) with 95% confidence intervals (Payton et al. 2003, Greenstone et al. 2007).

2.4 Results

Primer Design. The A. pisum primers did not amplify any of the ant species or their previous food items (D. melanogaster) during non-target testing and amplified A. pisum DNA with 100% success. Of the non-target species tested, seven (Aphididae: Sitobian avenae, Thomisidae, Alydidae, Lygaeidae,

Rhyparachromidae, and Pentatomidae) samples exhibited cross-reactivity with the A. pisum primers; six of the seven being Hemiptera (Table 2-1). However, this cross-reactivity was irrelevant for this study because ants were not exposed to these species in the laboratory.

Gut Content Inhibition Sources and DNA Degradation. PCR success was significantly lower in the whole body extractions than in the A. pisum positive controls (χ² = 14.40, P < 0.01) (Figure 2-3). Ant species identity was also a significant factor in inhibition at the whole-body level, with T. caespitum exhibiting increased inhibition compared to C. floridanus and S. invicta (χ² = 9.99, P <

0.01). Since some level of inhibition (average amplification success < 100%) of whole body extractions was found in all ant species, inhibition in the major body segments (head, mesosoma, gaster, and legs) was examined. Only the gaster differed in amplification success from the positive controls (χ² = 5.46, P = 0.02)

(Figure 2-3). As with whole body samples, there were observed species-level effects for the major body segments, with the T. caespitum by gaster interaction term being significant (χ² = 4.09, P = 0.04). With all of these treatment differences, T. caespitum appeared to be the driving variable.

Examination of organs within the gaster revealed that the crop exhibited the lowest levels of amplification success (χ² = 6.61, P = 0.01) (Figure 2-3). Again, there was a species effect at the organ-level, with the T. caespitum (χ² = 4.92, P

= 0.03) and T. caespitum by gut tract interaction (χ² = 4.92, P = 0.03) terms being significant. When dissection treatments containing parts of the gut (i.e. whole body, gaster, crop, and gut dissections) were pooled, 99% of S. invicta samples permitted DNA amplification, whereas 90% of C. floridanus and only 45% of T. caespitum samples were successful. Again, these treatment differences were dependent upon T. caespitum.

Amplification success of extracted T. caespitum crop contents with crop structure on Kim Wipes was significantly greater than T. caespitum samples with the crop included (χ² = 6.05, P = 0.01). All samples without the inclusion of the crop structure tested positive for A. pisum DNA. The optimal quantity of BSA was determined to be 0.25 µL / 1 µL total DNA per PCR since it was the minimum amount necessary to reach 100% amplification success. This quantity of BSA was found to increase amplification from 0% to 100% in samples previously testing negative (χ² = 29.65, P < 0.01). This indicates that the inhibition was not occurring during sample extraction but during PCR and that samples (especially

T. caespitum) can be saved during PCR using a combination of these two methods. The degradation rate in detection for the A. pisum DNA was significantly different from zero (χ² = 39.34, P < 0.01). A. pisum DNA could be amplified for up to 7.4 h for 50% of sampled ants (Figure 2-4) and at 24 h, prey

DNA was no longer detectable in the crops of the ant foragers.

2.5 Discussion

It has been recognized that obtaining gut content DNA from ants can be exceedingly difficult (Strehl and Gadau 2004, Gadau 2009). Our results corroborate this for one of the species tested, indicating that PCR inhibition is attributable to a component of the digestive system not the poison or Dufour’s glands as we hypothesized. Since inhibition in this species is due to components of the digestive tract, these results indicate that considerable care should be taken with T. caespitum and possible other untested species gut dissections when material is being used for molecular gut content analysis. This is especially true given the optimal situation these samples were analyzed under – with ingestion at time 0 and full crop contents. Given the results of the DNA degradation study (DNA half-life of 7.5 hours) and the potential for live ant organs to interfere differently, these results should be taken as conservative estimates of inhibition within this species.

Interestingly, we found differences in the levels of PCR inhibition between the three different ant species. While all species tested elicited some level of inhibition (average amplification success < 100% for whole body extractions), there were clear differences in the overall impact species had on the reliability of

DNA-based gut content analysis. This could be due to differences in endosymbiotic bacteria (Zientz et al. 2005) or digestion-related enzymes secreted by these ants (Ayre 1967, Went et al. 1972). Since the crop contents, once removed from the organ itself, exhibited no inhibition whatsoever, the interference might be due to the organ structure and not digestive tract

secretions per se. Further study is necessary to tease apart the particular compound causing this inhibition and better understand why inhibition rates vary so drastically between the tested ant species. Our results indicate that field- collected T. caespitum should be either dissected to exclude inhibitory organs influencing the results or be screened with BSA addition during PCR to significantly reduce the risk of false negatives. Without such safeguards, the existence of false-negatives could underestimate predation rates.

Despite the presence of PCR inhibition within the one of the ant species screened, these data provide two effective methods to overcome this inhibition.

Exuding crop contents onto sterile Kim Wipes is reliable and adds little time or expense to the overall process of testing ant samples for prey DNA. However, this method has the potential to shorten the detection period of prey DNA as it will not include DNA that has moved from the crop into the rest of the digestive tract. Additionally, this method must be used before extractions are completed and is, therefore, not helpful for previously extracted samples. However, the addition of BSA to samples already containing inhibitory agents, such as whole body extractions, may be helpful. BSA has been used in this way for the spotted lady beetles (Coleomegilla maculata (DeGeer) (Coleoptera: Coccinellidae)) and the spined soldier bug (Podisus maculiventris (Say) (Hemiptera: Pentatomidae))

(Hoogendoorn and Heimpel 2001, Greenstone et al. 2007) and is a valuable reagent to include in molecular gut content analysis. Both of these methods are useful, relatively easy to implement, and provide valuable solutions for different situations.

Our results suggest that at least several species of ants can easily be studied using molecular gut content analysis if appropriate care is taken. These data provide the methodology and confirmation of approaches to allow more accurate delineation of ant food webs and illumination of their biological control potential in agricultural, forestry, and urban settings. For instance, research has revealed the relative contribution of T. caespitum as an urban forager and waste- remover (Youngsteadt et al. 2014, Penick et al. 2015), as well as being an effective predator of turf grass pests (López and Potter 2000, 2003). These relationships, and other ant food web dynamics, can be further examined using molecular gut content analysis now that a methodology has been validated to reliably examine gut content. Until further investigation into the inhibition rates of other species and the mechanism of inhibition, un-tested ants should be assumed to exhibit inhibition and preventative measures, such as the protocols mentioned here, taken.

Table 2-1 Non-target species tested for cross-reactivity with the pea aphid primers (Apisum-219-F and Apisum-428-R) designed for this study. Order Family Species No. No. Tested Positive Araneae Araneidae 4 0 Araneae Lycosidae 2 0 Araneae Oxyopidae 2 0 Araneae Salticidae 3 0 Araneae Tetragnathidae 1 0 Araneae Thomisidae 2 1 Coleoptera Anthicidae 2 1 Coleoptera Carabidae 2 0 Coleoptera Chrysomelidae 2 0 Coleoptera Coccinellidae 3 0 Coleoptera Coccinellidae Scymnus sp. 1 0 Coleoptera Curculionidae 1 0 Coleoptera Elateridae 1 0 Coleoptera Latridiidae 1 0 Coleoptera Phalacridae 1 0 Coleoptera Staphylinidae 2 0 Diptera Agromyzidae 1 0 Diptera Agromyzidae Liriomyza sp. 1 0 Diptera Anthomyiidae 2 0 Diptera Anthomyiidae Delia antiqua 2 0 Diptera Brachycera 2 0 Diptera Chironomidae 1 0 Diptera Chloropidae Thaumatomyia 1 0 glabra Diptera Dolichopodidae Dolichopus finitus 1 0 Diptera Drosophilidae 3 0 Diptera Heleomyzidae 1 0 Diptera Hybotidae Platypalpus sp. 3 0 Diptera Syrphidae 1 0 Gastropoda Discidae Anguispira alternata 1 0 Gastropoda Polygyridae Mesodon zaletus 1 0 Hemiptera Aleyrodidae Bemisia tabaci 2 0 Hemiptera Alydidae 3 3 Hemiptera Aphididae 1 0 Hemiptera Aphididae Brevicoryne 3 0 brassicae Hemiptera Aphididae Capitophorus 3 0 eleagni

Table 2-1 continued

Order Family Species No. No. Tested Positive Hemiptera Aphididae Carolina rhois 2 0 Hemiptera Aphididae Chaetosiphon 1 0 thomasi Hemiptera Aphididae Hyadaphis 3 0 pseudobrassicae Hemiptera Aphididae Illinoia goldamaryae 2 0 Hemiptera Aphididae Macrosiphon 2 0 euphorbiae Hemiptera Aphididae Macrosiphon sp. 2 0 Hemiptera Aphididae Myzocallis 3 0 asclepiadis Hemiptera Aphididae Myzus persicae 6 0 Hemiptera Aphididae Pleotrichophorus 2 0 pseudopatonkus Hemiptera Aphididae Rhodobium 1 0 porosum Hemiptera Aphididae Rhopalosiphun padi 9 0 Hemiptera Aphididae Sitobion avenae 21 17 Hemiptera Aphididae Uroleucon 3 0 gravicorne Hemiptera Aphididae Uroleucon 1 0 helianthicola Hemiptera Aphididae Uroleucon 4 0 macgillivrayae Hemiptera Cicadellidae 4 0 Hemiptera Coccidae Coccus hesperidum 1 0 Hemiptera Coccidae Eulecanium 1 0 cerasorum Hemiptera Coccidae Neolecanium 1 0 cornuparvum Hemiptera Coccidae Parthenolecanium 1 0 quercifex Hemiptera Coccidae Pulvinaria 1 0 innumerabilis Hemiptera Corimelaenidae 1 0 Hemiptera Cynidae 4 0 Hemiptera Geocoridae 2 0 Hemiptera Lygaeidae 1 0 Hemiptera Miridae 1 0 Hemiptera Nabidae 1 0 Hemiptera Pentatomidae Euschistus servus 2 0

Table 2-1 continued

Order Family Species No. No. Tested Positive Hemiptera Pentatomidae Nezara viridula 3 0 Hemiptera Pentatomidae Oebalus pugnax 3 0 Hemiptera Pentatomidae 6 2 Hemiptera Psyllidae Bactericera 1 0 cockerelli Hemiptera Psyllidae Cacopsylla pyricola 1 0 Hemiptera Reduviidae 1 0 Hemiptera Rhyparochromidae 1 1 Hymenoptera Argidae 1 0 Hymenoptera Formicidae Camponotus 2 0 floridanus Hymenoptera Formicidae 1 0 Hymenoptera Formicidae Pheidole tysoni 4 0 Hymenoptera Formicidae Solenopsis invicta 2 0 Hymenoptera Formicidae Tetramorium 9 0 caespitum Hymenoptera Platygastridae 1 0 Lepidoptera Noctuidae Trichoplusia sp. 1 0 Primates Hominidae Homo sapiens 1 0

Figure 2-1 Generalized diagram of ant body segments and internal organs screened by molecular gut content analysis to quantify the frequency of inhibition by each representative part of the ant.

Figure 2-2 Diagram of dissection stages where (A) whole ant bodies were analyzed, (B) followed by their respective body segments as described in Fig. 1, and (C) the organs within the gaster once the inhibition was determined to reside primarily within that body segment

Figure 2-3 Percentage of samples testing positive for A. pisum using extractions of ants spiked with target prey (A. pisum) DNA. The total number of samples testing positive is given before the bracketed total number of samples tested at the base of each column.

Figure 2-4 Results of the T. caespitum feeding trial. Closed dots represent the percentage of samples at each time point (0–24 h) testing positive for A. pisum DNA. Dashed lines represent the upper and lower 95% confidence intervals Copyright © Hannah Joy Penn 2016

Chapter 3 DETECTION OF DNA IN APHID HONEYDEW: A SOURCE OF ERROR FOR MOLECULAR GUT CONTENT ANALYSIS?

Chapter contents published in Penn, HJ and JD Harwood. 2016. Detection of DNA in aphid honeydew: A source of error for molecular gut content analysis? Journal of the Kansas Entomological Society, 89(1), 85-91.

3.1 Summary

Aphids (Hemiptera: Aphididae) are pests of a vast array of crops globally, contributing to significant yield loss annually. Biological control, utilizing natural enemies to suppress focal pest species, has often been integrated into aphid management practices. To ensure that arthropod predators are consuming the target aphid, and ultimately have an impact on pest population dynamics, it is important to gather information on consumption patterns in space and time.

Molecular gut content analysis can be used to assess the strength of trophic linkages between predators and prey by detecting the presence of species- specific DNA (including aphid DNA) in their gut. However, many generalist predators readily consume aphid honeydew as an additional nutritive resource beyond direct consumption of prey. This raises the important question of whether aphid honeydew is responsible for false-positive predation events being recorded using DNA-based gut content analysis, should honeydew contain detectable quantities of aphid DNA. To evaluate this potential source of error, we examined honeydew of three species of aphids, Acyrthosiphon pisum, Sitobion avenae, and Rhopalosiphum padi, for aphid DNA using standard gut content analysis methods. After extensive testing, we revealed that aphid honeydew did not contain any detectable aphid DNA for any of the species, nor did honeydew inhibit the amplification of aphid DNA in positive control experiments. These data

indicate that aphid predation rates previously reported are not influenced by direct consumption of honeydew. This study, therefore, shows that using DNA- based gut content analysis is a reliable method to measure aphid predation in the field, even when predators are exposed to and consume, honeydew.

3.2 Introduction

Over one-hundred aphid (Hemiptera: Aphididae) species are considered to be important crop pests (Blackman and Eastop 2007), attacking every major crop worldwide (Peters et al. 1990). Extensive damage results from aphid feeding (Kieckhefer and Kantack 1988, Quisenberry and Ni 2007), honeydew- induced sooty mold (Reynolds 1999), and the vectoring of plant disease (Katis et al. 2007). This accounts for extensive yield loss with significant implications for growers (Brewer and Elliott 2004, Kim et al. 2008), making control measures critical for financial viability of crop production. Chemical controls have been used to manage aphid pests, but are considered unsustainable due to the creation of resistant aphid populations (Foster et al. 2007) and inefficiencies when dealing with aphid-vectored virus damage (Dedryver et al. 2010). One alternative, or complementary, approach for management of aphids is biological control whereby pest populations are maintained below an economic threshold, utilizing arthropod natural enemies that have been introduced or augmented into the surrounding environment through conservation management (Symondson et al.

2002, Dedryver et al. 2010).

Biological control is commonly used in agricultural settings where aphid populations are problematic. For instance, the release of generalist predators and parasitoids can control green peach aphid (Myzus persicae (Sulzer)

(Hemiptera: Aphididae)) (Messelink et al. 2013), and the cereal aphids

Rhopalosiphum padi (L.) (Hemiptera: Aphididae), Sitobion avenae (Fabricius)

(Hemiptera: Aphididae), and Metopolophium dirhodum (Walker) (Hemiptera:

Aphididae) (Sigsgaard 2000). Furthermore, crop yields have increased in response to biological control of aphids by predators (Östman et al. 2003), but there is considerable variation in the effectiveness of predator-mediated control of aphid populations (Dedryver et al. 2010). Such variability raises the important question of monitoring the consumption habits of predators to enable accurate estimation of likely effectiveness of different parts of the predator community.

Predation events are frequently documented through molecular gut content analysis, where the presence of prey-specific DNA is identified following consumption in the field (Symondson 2002, King et al. 2008, Furlong 2015). This method has been applied to characterize the feeding habits of generalist predators, including consumption of aphid pests (Chen et al. 2000, Harwood and

Obrycki 2005a, Harwood et al. 2007, Boreau de Roincé et al. 2013, Gomez‐Polo et al. 2014, Rondoni et al. 2014, Choate and Lundgren 2015). In many instances, the interactions suggest dietary selection by predators that can be advantageous for biological control. Furthermore, the breadth of published studies indicates the ubiquity of testing predator gut contents for aphid DNA in many different cropping systems.

Generalist predators are widely reported as consuming the sugar and amino acid-filled aphid honeydew (Canard 2001, Offenberg 2001), which is readily available from the producers directly or from the environment around producers (Lundgren 2009). Predatory arthropods that consume the honeydew by-products of hemipterans may gain added nutritional benefit when fed a combination of prey items and honeydew or similar sugar solutions (Evans 2000,

Hogervorst et al. 2007, Gibb and Johansson 2010, Pfannenstiel and Patt 2012,

Pfannenstiel 2015). This indicates that predators studied using molecular gut content analysis have a high likelihood of consuming both the target species and hemipteran-provided honeydew, posing the risk that gut contents screen positive for aphid DNA when only honeydew was consumed. This is a particular concern given that aphid and whitefly honeydew can contain viral and bacterial DNA (Hu et al. 1996, Thompson et al. 2003, Leroy et al. 2011, Roopa et al. 2014) and could also include aphid DNA. Sint et al. (2015) revealed that DNA from spider feces contains both spider and prey DNA, indicating that arthropod bodily secretions often contain enough DNA to be recognized using standard molecular techniques. Aphid DNA in honeydew, if present and amplified using standard molecular gut content analysis techniques, could result in less defined trophic networks within crop fields and communities where aphid predation is being assessed. The presence of aphid DNA from honeydew consumption would also indicate that consumption patterns by generalist predators are skewed, with predation events likely to be overestimated. The goal of this study was to determine if aphid DNA is detectible in aphid honeydew using methods common

to molecular gut content analysis. Based on the abovementioned references that suggest DNA is present in honeydew, we hypothesize that aphid honeydew, without any aphid body part present, will contain aphid DNA resulting in positively skewed predation data.

3.3 Materials and Methods

3.3.1 Insects

Sitobion avenae (Fabricius) (Hemiptera: Aphididae) and Rhopalosiphum padi (L.) (Hemiptera: Aphididae) were collected from winter wheat fields at the

University of Kentucky Research and Education Center in Princeton, Kentucky,

USA (GPS Coordinates: 37° 6' 4" N, 87° 51' 11" W). Colonies were maintained in the greenhouse on Pembroke variety soft red winter wheat (Triticum aestivum L.

(Poales: Poaceae)) (Clemons Ag Supply, Springfield, Kentucky, USA) under controlled conditions (described below). Acyrthosiphon pisum (Harris)

(Hemiptera: Aphididae) were collected from colonies already established on Red

Mammoth Clover (Trifolium pratense L. (Fabales: Fabaceae)) (Johnny’s Selected

Seed, Winslow, Maine, USA) in the lab. All colonies were maintained under controlled conditions of 25 ± 1 °C, 16: 8 L:D cycle and 65 ± 5% RH. Each aphid species had three actively maintained colonies; each consisted of six potted host plants within a 60 × 60 × 60 cm Bug Dorm cage (Megaview Science, Taichung,

Taiwan).

3.3.2 Honeydew Collection

Honeydew was collected via two substrates: Kim Wipes (Kimberly-Clark

Professional, Roswell, Georgia, USA) and host leaves of either winter wheat or clover depending on the species of aphid. For the Kim Wipe substrate, UV- sterilized Kim Wipe papers (6.5 cm²) were placed into UV-sterilized plastic petri dishes (8.5 cm diameter × 1.5 cm deep). Twenty aphids were transferred via soft- bristled paint brush (sterilized with deionized water, 95% EtOH, and DNA Away wash (Molecular Bioproducts, Inc, San Diego, California, USA)) into the petri dish containing the sterilized Kim Wipe. After one hour, the aphids were carefully removed and the Kim Wipe examined for aphid body parts (removed if found), transferred into an autoclaved and UV-sterilized 1.5 ml microcentrifuge tube, and the DNA extracted immediately (see protocol below). For the host-plant leaf substrate, leaves of the honeydew-covered host plants were selected (i.e. leaves with visible signs of sticky droplets and residue) from laboratory colonies. Ten leaves per colony (n = 3 per species) were removed, honeydew-covered areas excised using UV-sterilized scissors and DNA was extracted immediately

(protocol below). Ten replicates using each substrate were completed for three different colonies of each aphid species (n = 30 per substrate/species). Negative controls consisted of extracted substrates not previously exposed to aphids.

Positive controls included six aphid legs, one small aphid, one large aphid, and five large aphids of each species (n = 2 per control type / species). Additional positive controls included extraction of six aphid legs of each species in the presence of each substrate without honeydew (n = 6 per substrate type /

species) and in the presence of each substrate with honeydew produced by the same species as the legs (n = 2 per substrate type / species).

3.3.3 DNA Extraction and PCR Protocols

The abovementioned samples were immediately homogenized in 180 µL of tissue lysis buffer and samples containing substrates were crushed during this step using a micro pestle (Sigma-Aldrich, St. Louis, Missouri, USA). The total

DNA was extracted from all samples using Qiagen DNeasy Blood and Tissue

Kits© (Qiagen Inc., Valencia, California, USA) following standard animal tissue protocols. DNA was stored in clean 1.5 mL microcentrifuge tubes at -20°C until

PCR analysis.

The total DNA in all samples was amplified with general COI primers

(mlCOlintF and jgHCO2198) (Leray et al. 2013). Samples that yielded positive bands within the treatment and controls using the general COI primers were subjected to amplification with species specific primers – S. avenae (EgaCOIIF2 and EgaCOIIR) (Chen et al. 2000), R. padi (BcoaCOIIF4 and BcoaCOIIR2)

(Chen et al. 2000), and A. pisum (Apisum-219-F and Apisum-428-R) (Penn et al.

2016).

All PCR reactions (12.5 µL) consisted of 1X Takara buffer (Takara Bio

Inc., Shiga, Japan), 0.2 mM of each dNTP, 0.2 mM of each primer, 1.25 U

Takara Ex TaqTM and template DNA (1 µL of total DNA). PCR reactions were carried out using Bio-Rad PTC-200 and C1000 thermal cyclers (Bio-Rad

Laboratories, Hercules, California, USA). The PCR cycling protocol for the

general COI primer pair was 94 °C for 1 minute followed by 25 cycles of 95 °C for

10 s, 46 °C for 30 s, and 72 °C for 60 s. The PCR cycling protocol for the S. avenae primer pair was 94 °C for 1 minute followed by 40 cycles of 94 °C for 30 s, 56 °C for 30 s, and 72 °C for 45 s. The cycling protocol for R. padi primer pair only differed from the S. avenae pair with an annealing temperature of 55 °C.

The PCR cycling protocol for the A. pisum primer pair was 94 °C for 1 minute followed by 40 cycles of 94 °C for 45 s, 63 °C for 45 s, and 72 °C for 30 s.

Following amplification, the bands for all primers were visualized on 2% SeaKem agarose (Lonza, Rockland, Maine, USA). Each PCR reaction contained one positive and one negative control to determine success.

3.4 Results

No evidence was found to indicate that aphid honeydew from any of the three species contained detectable quantities of aphid DNA using the abovementioned extraction protocol (Figure 3-1). All substrate negative controls exhibited no DNA amplification using general COI or species-specific primers.

We also revealed that neither aphid honeydew, the substrate used (Kim Wipes and leaf material), or the combination thereof prevented the extraction of aphid

DNA from specimens, with all extracted positive control samples exhibiting 100% amplification success using general COI primers and species-specific aphid primers. The treatment type combinations (positive controls, negative controls, and honeydew treatments) were perfect predictors of PCR outcomes (100%, 0%,

and 100% aphid-DNA amplification success respectively), therefore, no statistical analyses were necessary.

3.5 Discussion

There was no evidence gathered to suggest that honeydew from the three species tested contained detectable quantities of aphid DNA. Furthermore, no differences were observed between the three aphid species for any treatment type despite the inherent variation of body size and quantity of honeydew produced. One possible explanation for the lack of aphid DNA in the honeydew is the extensive filtration systems with which aphids process the sugar-rich honeydew (Douglas 2003). Alternatively, trace amounts of DNA could be present but require finer methods of extraction and amplification since bacterial and viral

DNA has been obtained from honeydew (Leroy et al. 2011, Roopa et al. 2014).

These data indicate that consumption of aphid honeydew by generalist predators will not incorrectly ascribe a predation event using molecular gut content analysis, thereby rejecting our original hypothesis.

Additionally, the presence of aphid honeydew did not hinder the amplification of aphid DNA, indicating that honeydew is not a PCR inhibitor. This is further substantiated by several studies where whole aphids were consumed by predators and those predators tested positive for aphid DNA, even though aphid bodies contain honeydew that has yet to be excreted (Greenstone and

Shufran 2003, Traugott et al. 2012, Chapman et al. 2013). Our study, in addition

to this previously published research, confirms that aphid honeydew does not create false negatives in a DNA-based molecular gut content system.

Given the ubiquity of aphids and other honeydew-producing hemipterans within agroecosystems, it is common for generalist predators, including those providing valuable biological control services, to encounter and/or consume honeydew directly from the producer or the environment (Lundgren 2009).

Although honeydew consumption by predators could be pervasive, our data indicate that aphid honeydew from three species did not interfere with molecular gut content analysis (false positive or negative) using standard DNA extraction methods. It can be concluded, therefore, that honeydew does not interfere with these approaches for field-based analysis of aphid predation and, potentially, other honeydew-producing hemipterans.

Figure 3-1 Depicts the percentage of samples testing positive for aphid species-specific DNA for each aphid species individually. “KW” represents Kim Wipe. Treatments including “Bodies” are pooled positive controls, “KW” and “Leaf” treatments are negative controls, and treatments containing “Honeydew” are samples containing only substrate material and honeydew. No variation was found within treatment types, therefore, no errors are included. (A) Percentage of Sitobion avenae samples testing positive for DNA using species-specific primers. (B) Percentage of Rhopalosiphum padi samples testing positive for DNA using species-specific primers. (C) Percentage of Acyrthosiphon pisum samples testing positive for DNA using species-specific primers.

Chapter 4 LAND COVER DIVERSITY INCREASES ANT SPATIAL AGGREGATION TO PREY AND CONSUMPTION

Chapter contents prepared for publication in Penn, HJ, KJ Athey, and BD Lee. Land cover diversity increases predator spatial aggregation to prey and consumption. Ecology Letters.

4.1 Summary

Increasing plant diversity within agroecosystems is purported to increase arthropod diversity and is dependent on land cover and fragmentation. Ants, important providers of ecosystem services such as biocontrol, are vital to consider as they are susceptible to landscape-level changes. We determined the relationship between land cover diversity and fragmentation on the within field spatial associations of ants to pests as well as resulting predation events by combining mapping and molecular tools. We found that increased land cover diversity and decreased fragmentation increased ant abundance, spatial associations to pests, and resulting predation. Specific land covers influenced ant and pest abundance and diversity more at the smaller scale, indicating that land use close to fields is of greater importance to biocontrol. These results indicate that geospatial techniques and molecular gut content analysis can be successfully combined to determine that land use impacts arthropod predator communities and the resulting predation events in agroecosystems.

4.2 Introduction

As agroecosystems encompass large portions of the terrestrial surface of

Earth it is important to consider how landscapes in and around this habitat influence the productivity of crops and the flora and fauna living therein (Duelli

1990, Forman 1995). Landscape alteration, through a combination of overall loss of non-crop habitat and an increase in habitat fragmentation, has impacted overall biodiversity by inducing local extinction (Wiens and Milne 1989, Golden and Crist 2000, Steffan-Dewenter et al. 2002, Fahrig 2003, Gove et al. 2009).

Intensified agroecosystems, subjected to influence by landscape alteration

(Chapin III et al. 2000), can, therefore, be thought of as relatively poor habitats to maintain local diversity (Altieri 1999) even though they provide plentiful resources for specific pest species (O'Rourke et al. 2011).

In these systems, the surrounding non-agricultural land becomes an important resource for beneficial species living in and adjacent to crops (Dyer and Landis 1997) as well as the crops themselves by sustaining beneficial species providing necessary ecosystem services such as pollination, soil aeration, and pest suppression (Thomas et al. 1991, Weibull et al. 2000).

Landscape properties have a particularly large impact on pest suppression or biocontrol provided by arthropod predators (Madsen et al. 2004), the average value of which is an estimated $8 billion annually in the US (Isaacs et al. 2009).

To manage pest populations via biocontrol by enacting large-scale landscape management practices (Zhang et al. 2007, Rusch et al. 2016), spatial scale is particularly important (Fahrig and Jonsen 1998, Steffan-Dewenter et al. 2002,

Frank et al. 2011) as landscape management to supplement biocontrol involves the manipulation of on-farm land and the surrounding landscapes (Landis et al.

2000, Gabriel et al. 2006). Environmental supports provided through landscape- level enhancement of non-agricultural lands increase predator populations by

providing refuge from crop management-induced disturbances (Finke and

Snyder 2010). Additionally, these non-crop areas can bolster predator populations by providing alternative prey items and plant-based food resources, thus allowing predators to live longer with higher fecundity (Landis et al. 2000,

Woinarski et al. 2002). The more that is known about how land cover diversity and fragmentation impact predator abundance and diet in agricultural areas, the more likely biocontrol can be effectively implemented (Simpson et al. 2011).

Landscape diversity, both in terms of cover composition and configuration or fragmentation (Fahrig et al. 2011), aids in biocontrol by alleviating the pressure monocultures put on predator populations (Grez et al. 2014) and by decreasing the overall use of pesticides (Meehan et al. 2011). Altieri (1999) noted that high crop diversity and mixing crops in time and space is important to maintain low pest population, and this was best achieved with small plots of crops intermingled with uncultivated lands because these areas maintain predators within disturbed fields (Rodenhouse et al. 1992, Benton et al. 2003). In addition to predator abundance, the efficiency of prey-capture rates was found to be associated with habitat changes (Macfadyen et al. 2011), indicating that land cover and fragmentation changes influence not only the abundance of predators and pests but also their interactions in space and time.

Using ants (Hymenoptera: Formicidae), important predators in agroecosystems and indicators of ecosystem health (Campbell 1990, Perfecto

1991, Paulson and Akre 1992, Zenger and Gibb 2001a, Del Toro et al. 2012), as focal predators, we wanted to answer four questions. One, does an increased

landscape (cover composition) diversity and decreased fragmentation provide a corresponding increase in within crop ant abundance and diversity? Within this context, does the scale at which the landscape is analyzed important? Second, does an increase in land cover diversity and ant abundance and diversity result in more within field aggregations of ants to soybean arthropod pests? And, finally, do these aspects of landscape diversity, ant populations, and spatial associations of ants to pests result in an increase in actual ant predation of selected soybean pests? By combining geospatial and molecular techniques, we expected that increased land cover diversity and decreased overall fragmentation will increase the predator abundance and subsequent spatial overlap and predation events within the cropping system as ant foraging patterns are temporally and spatially variable (Cushman and Whitham 1991, Lee et al. 2011).

4.3 Materials and Methods

4.3.1 Study Sites and Sample Collection

Fields (n = 23) with availability dictated by grower participation were sampled repeatedly during the growing seasons of 2012 - 2014 in western/central Kentucky. Soybean fields were planted in late May to early June with minimal to no-tillage regimes, no irrigation, and managed by growers under an annual three crop rotation with corn and winter wheat (Skillman 2001). To estimate ant and soybean pest communities and associated predation events, several collection methods (n = 1282 total; pitfall traps, sweep-net, and hand samples) were used, with counts combined posthoc to ensure a complete

picture of the arthropod community (Lang 2000). Non-baited pitfall traps (dia. 9.5 cm, depth 12 cm) with rain guards and containing ethylene glycol for specimen preservation were placed in a grid pattern (9-16 locations based on field dimensions) between crop rows starting in the outermost row of every field.

Collections at each site occurred over 24 hours, once per month from June to

August. At each pitfall location, both sweep-net (ten figure eights) and hand samples (1 m2 quadrats over a period of two minutes) were conducted.

Specimens were separated in the field with a pooter into individual 1.5 mL centrifuge tubes containing 95% ethanol and stored at -20ºC. Ants and pests were identified to species, and all other arthropods identified to order or family when possible (AAS 2005, Coovert 2005, Johnson et al. 2015); Shannon diversity index was calculated for both communities per sample location within a field (Magurran 1988).

4.3.2 Landscape Analysis

Landscape factors were determined using geospatial data gathered in the field using a handheld GPS unit (Magellan eXplorist 610, MiTAC, Santa Clara,

California, U.S.A.). The Crop Data Layer in WGS84 projection (USDA NASS

2012-2014) was obtained for the entire Commonwealth of Kentucky for each year of the study. Using ArcGIS desktop v.10.0.sp.5 (Esri, Redlands, California,

U.S.A.), maps of all field sites by year were paired with the corresponding crop data layer with 0.5 km (small scale) and 3.0 km (large scale) circular buffers established around each site to approximate the dispersal distance range for

ants after mating flights (Dauber and Wolters 2005). The land cover within these buffers was exported in metric units to Fragstats 4.0 (McGarigal et al. 2012) for analysis of both landscape and class features (Table 4-1) using no specified sampling strategy and an 8-celled neighbor rule. The landscape-level metrics were determined for each site/buffer combination and included total edge length

(TE) and Simpson’s diversity index (SIDI) (Li et al. 2001, Yeh and Huang 2009,

Malaviya et al. 2010). The class-level metrics obtained included percentage of the landscape (PLAND) and average patch clumpiness (CLUMPY). PLAND, TE, and CLUMPY represent the fragmentation of the landscape where fragmentation increases if PLAND and CLUMPY decrease but TE increases (Di Orio et al.

2005, Midha and Mathur 2010).

4.3.3 Spatial Aggregation Analysis

Within field spatial distributions of both ant and pest communities were analyzed using Spatial Analysis by Distance IndicEs (SADIE) (Perry 1995, Perry et al. 1999, Winder et al. 2001, Perry and Dixon 2002, Perry et al. 2002). These data were input into software (SADIEShell 2.0 with QuickSADIE© K.F. Conrad) to obtain non-parametric values of aggregation within the field. The ant and pest populations were then analyzed in comparison to each other (Windor et al. 2005,

Holland et al. 2007) using SADIE association analysis (N_AShell 1.0 with

QuickAssociation 2.0© K.F. Conrad). All output cluster files were mapped using

Surfer v9.0 (Golden Software, LLC, Golden, Colorado, U.S.A.) for all fields at all time periods for ants, pests, and their association.

4.3.4 Molecular Gut Content Analysis

Molecular gut-content analyses, an excellent way to determine trophic linkages and underutilized for ant predation (King et al. 2008, Choate and

Drummond 2011), was completed following collection and identification. Samples

(n = 684) of the six most abundant ant species were prepared for DNA extraction after Penn et al. (2016). Legs were removed from two commonly collected pest specimens – the grape colaspis beetle Colaspis brunnea F. (Coleoptera:

Chrysomelidae) and the green stinkbug Chinavia hilaris (Say) (Hemiptera:

Pentatomidae). Other species such as the green cloverworm Hypena scabra F.

(Lepidoptera: Erebidae) and the Japanese beetle Popillia japonica Newman

(Coleoptera: Scarabaeidae) were non-target tested but did not have reliable COI primer locations. Total DNA was extracted from all samples using DNeasy Blood and Tissue Kits© (Qiagen Inc., Valencia, California, U.S.A.) following the animal tissue protocol.

To design soybean pest primers, a data matrix was constructed containing cytochrome c oxidase subunit I (COI) sequences from pest and ant species collected (Genbank accession numbers KX687920-KX687976). DNA was amplified from the legs of specimens with the general COI primers LCO-1490

(Folmer et al. 1994) and HCO-700me (Breton et al. 2006). Polymerase chain reactions (PCR) consisted of 0.5 µg/µL bovine serum albumin (New England

Biolabs, Inc. Ipswich, Massachusetts, USA), 1X Takara buffer (Takara Bio Inc.

Shiga, Japan), 0.2 mM of each dNTP, 0.2 mM of each primer, 0.625 U Takara Ex

Taq and template DNA (1 µL of total DNA). PCRs were carried out using a Bio-

Rad C1000 thermal cycler (Bio-Rad Laboratories, Hercules, California, U.S.A.).

Reaction success was determined by visualizing bands on a 2% SeaKem agarose (Lonza, Rockland, Maine, U.S.A.) gel stained with GelRed nucleic acid gel stain (1X; Biotium, Hayward, California, USA). DNA sequencing was performed at Beckman Coulter Genomics (Danvers, Massachusetts, U.S.A.). All primers were designed with Primer3 (Rozen and Skaletsky 2000) and tested against 193 non-target samples to ensure specificity (Table 4-2). No suitable primer region was located in COI for C. hilaris and primers were designed using the mitochondrial gene, 16S (16Sbr-H (5'- CCG GTC TGA ACT CAG ATC ACG

T -3') and 16Sar-L (5'- CGC CTG TT ATC AAA AAC AT -3') (Palumi et al. 1991)).

All ants were tested with pest-specific primers designed herein (Table

4-3). All reaction conditions were identical to the COI protocol except the PCR cycling conditions. For the C. brunnea primers, the protocol was 94 °C for 1 minute followed by 40 cycles of 94 °C for 45 s, 62 °C for 45 s, and 72 °C for 30 s.

The protocol for the C. hilaris primers was the same except for an annealing temperature of 53 °C. To determine extraction success for those samples not testing positive for any tested primers, we utilized general COI primers Jerry

(Simon et al. 1994) and Ben3R (Villesen et al. 2004b). Reagents were identical as above. Any reactions not testing positive with the general COI primers (n = 8) were excluded from further analysis.

4.3.5 Statistical Analyses

Statistical analyses at 0.5 and 3.0 km scales were completed in SAS 9.4

(SAS Institute, Cary, North Carolina) where all regressions included a factor for field and repeated measures in addition to abiotic variables (average temperature and wind speed) from the USDA Kentucky Weekly Crop Report (2012-2014). We determined the interactions of ant and pest abundance and diversity in relation to land cover using PROC GENMOD with a negative binomial distribution against the landscape data in stages – landscape and general class-level and specific cover type PLANDs (Jonsen and Fahrig 1997, Campbell 2015). To establish the connections between within field spatial aggregation and land use, the binary

(Significant spatial association? Yes/No) ant, pest, and ant to pest SADIEs were analyzed using PROC LOGISTIC with a binomial distribution with variables including the above landscape variables and ant and pest variables. Due to limited degrees of freedom, the SADIE, and association regressions were run with landscape, class, and community variables but not specific class PLANDs as the maximum likelihood estimate did not exist when these variables were included in the model. To determine the relationship between predation and land cover and SADIE variables, binary PCR outcomes were analyzed with PROC

LOGISTIC with a binomial distribution against SADIE, landscape, and ant and pest community data.

4.4 Results

4.4.1 Landscape Analysis

The landscape matrix surrounding the fields had varied cover composition

(Figure 4-1). Out of a possible 133 cover type categories, 0.5 km buffers had a mean of 7.40 types while 3.0 km buffers had a mean of 14.22 types (Figure 4-2).

The most commonly observed cover types (presence/absence not area) at both scales for all fields were open water and soybean while the least common were clover/wildflower, double-cropped corn/soybean, oat, popcorn, and walnut. In terms of the average area around each field, grass/pasture, deciduous forest, and corn dominated the landscapes when they were present. Each cover type was observed on average in the 0.5 km buffers of 6.26 fields and in the 3.0 km buffers of 12.3 fields.

As land cover fragmentation (CLUMPY) decreased, ant abundance increased (0.5 km: χ² = 8.91, P < 0.01; 3.0 km: χ² = 14.67, P < 0.01) though ant diversity appeared relatively unaffected (0.5 km: χ² = 3.89, P = 0.05; 3.0 km: χ² =

0.38, P = 0.54) (Figure 4-3). Ant and pest diversity was not influenced directly by the general landscape metrics; however, pest abundance (Avg. PLAND, 0.5 km:

χ² = 19.41, P < 0.01; 3.0 km: χ² = 4.20, P = 0.04; Avg. CLUMPY, 0.5 km: χ² =

5.59, P = 0.02; 3.0 km: χ² = 0.01, P = 0.94) was significantly correlated with the same fragmentation metrics as ant abundance (Avg. PLAND, 0.5 km: χ² = 13.49,

P < 0.01; 3.0 km: χ² = 11.10, P < 0.01). Both ant and pest communities were dependent upon scale (Table 4-4, Table 4-5), where decreased scale size resulted in greater impacts of landscape variables on abundance and diversity.

Scale played an important role in how specific land cover types interacted with diversity and abundance of both ants and pests (Table 4-6, Table 4-7). The smaller scale was more influential (i.e. more land covers had significant correlations) on diversity rather than the abundance of ants (2288, 3917 AICc, respectively) and pests (1471, 4427 AICc, respectively). The same was true at the larger scale – land covers were more likely to be associated with ant (2272,

3866 AICc, respectively) and pest (1463, 4370 AICc, respectively) diversity rather than abundance. Additionally, more cover types were responsible for significant associations with either ant or pest communities at the small scale than at the large. For instance, at the small scale, 9 of 15 cover types were significantly correlated with ant and pest abundance, only five of which overlapped between ants and pests. Of these five, four overlapped with the same interaction direction (i.e. parameter estimate sign), and only one had the opposite

(for instance, some covers increased ant and decreased pest abundance).

However, this trend diminished at the 3.0 km scale with 14 of 21 cover types in the regression significantly influencing either ant or pest abundance, only three were significant for both ants and pests and had opposite interactions each time.

When cover types were analyzed in relation to the diversity of ants and pests, a similar trend was found. At the small scale, of the15 covers possible, only seven and four had significant impacts on ant and pest diversity respectively, with no overlap between communities. Again, the impact of specific land use PLANDs was not as strong at the 3.0 km scale where three variables were correlated with ants and none with pests.

In general, cover types that had significant correlations with the respective communities had similar influences on ant and pest abundance and diversity at the smaller scale, but opposite influences at the larger scale. At the small scale, deciduous forest (ant: χ² = 7.66, P < 0.01; pest: χ² = 19.83, P < 0.01), developed land (ant: χ² = 10.66, P < 0.01; pest: χ² = 37.61, P < 0.01), and soybean (ant: χ² =

17.31, P < 0.01; pest: χ² = 24.66, P < 0.01) had significant negative interactions with both ant and pest abundance and open water positively influenced ant but not pest abundance (ant: χ² = 4.82, P = 0.03; pest: χ² = 15.91, P < 0.01). At the large scale, developed (ant: χ² = 14.49, P < 0.01; pest: χ² = 3.78, P = 0.05) and non-alfalfa hay (ant: χ² = 18.52, P < 0.01; pest: χ² = 4.41, P = 0.04) increased ant while decreasing pest abundance. The only exception to this, was an increased presence of sod/grass seed in the larger landscape significantly increased pest and decreased ant abundance (ant: χ² = 6.51, P = 0.01; pest: χ² = 6.05, P =

0.01).

4.4.2 Spatial Aggregation Analysis

The SADIE data showed a field effect on the likelihood that ants and pests were spatially aggregated within their respective communities (Table 4-8, Table

4-9); for most fields, this association was not time dependent. Of the 23 fields tested, eight had significant aggregations of ants and, six had significant aggregations of pests. When the spatial association was measured per each time period between the ant and pest communities, five of the fields were significantly aggregated (Figure 4-4) while no fields contained communities that actively

avoided each other (Table 4-10, Table 4-11). When the SADIE associations of ant and pest were analyzed in relation to landscape and community composition metrics, several influences were determined. Increased land cover diversity

(SIDI, 0.5 km: χ² = 102.71, P < 0.01; 3.0 km: χ² = 0.10, P = 0.75) increased the likelihood of ant aggregations at only the smaller scale. Ant within field aggregations were significantly positively associated with increases in ant diversity (0.5 km: χ² = 5.50, P = 0.02; 3.0 km: χ² = 12.76, P < 0.01). Pest aggregations were more likely to occur with decreased fragmentation (PLAND,

0.5 km: χ² = 49.26, P < 0.01; 3.0 km: χ² = 31.05, P < 0.01; TE, 0.5 km: χ² =

132.46, P < 0.01; 3.0 km: χ² = 28.87, P < 0.01; CLUMPY, 0.5 km: χ² = 6.34, P =

0.01; 3.0 km: χ² = 101.21, P < 0.01) and decreased land cover diversity (0.5 km:

χ² = 157.73, P < 0.01; 3.0 km: χ² = 0.58, P = 0.56) . Pest aggregations were also positively associated with the presence of ant aggregations (0.5 km: χ² = 148.43,

P < 0.01; 3.0 km: χ² = 133.80, P < 0.01) but were negatively impacted by an increase in ant abundance (0.5 km: χ² = 4.29, P = 0.04; 3.0 km: χ² = 5.66, P =

0.02). When the spatial association of ants to pests was determined, decreased fragmentation ( PLAND, 0.5 km: χ² = 57.11, P < 0.01; 3.0 km: χ² = 23.87, P <

0.01; TE, 0.5 km: χ² = 52.21, P < 0.01; 3.0 km: χ² = 12.63, P < 0.01; CLUMPY,

0.5 km: χ² = 42.48, P < 0.01; 3.0 km: χ² = 108.86, P < 0.01) and increased land cover diversity (0.5 km: χ² = 21.24, P < 0.01; 3.0 km: χ² = 67.73, P < 0.01) increased the likelihood of positive spatial associations (Table 4-11). However, the intensity and/or direction of these associations, particularly land cover

diversity (SIDI) changed at the 3.0 km scale, with increases in SIDI decreasing the likelihood of aggregation of ants to pests (Figure 4-5).

4.4.3 Molecular Gut Content Analysis

The designed pest primers did not amplify any of the target ant species nor any other specimens during non-target testing and amplified their respective species’ DNA with 100% success (Table 4-2). Of the six most abundant of the 35 field-collected ant species (Lasius neoniger Emery, Monomorium minimum

Buckley, Pheidole tysoni Forel, Solenopsis molesta Say, Tapinoma sessile (Say),

Tetramorium caespitum (Linnaeus)), L. neoniger, P. tysoni, and S. molesta tested positive for any pest DNA. L. neoniger was found positive for C. brunnea

(13.4%) and H. scabra (2.9%). P. tysoni (36.8%) and S. molesta (13.8%) both only tested positive for H. scabra; no ant species tested positive for C. hilaris

(Table 4-12). When compared with SADIE association data, we found that 5 of the 5 fields with significant spatial associations of ants to pests also contained at least one ant testing positive for the target pests (Table 4-13).

The number of ants testing positive for pest DNA, was compared with landscape and community composition metrics. Increased land cover diversity at both scales increased the likelihood of a positive gut content (0.5 km: χ² = 42.43,

P < 0.01; 3.0 km: χ² = 6.81, P = 0.01) while the alternative was found to be true for decreases in fragmentation (PLAND, 0.5 km: χ² = 51.18, P < 0.01; 3.0 km: χ²

= 4.51, P = 0.03; TE, 0.5 km: χ² = 93.91, P < 0.01; 3.0 km: χ² = 18.82, P < 0.01;

P = 0.04; CLUMPY, 0.5 km: χ² = 81.17, P < 0.01; 3.0 km: χ² = 4.92, P = 0.03).

An increase in ant diversity indicated an increase in ant predation but only at the smaller scale (0.5 km: χ² = 3.45, P = 0.06; 3.0 km: χ² = 6.17, P = 0.01). The increased occurrence of spatial aggregation of ants to pests at both scales was found to significantly increase the likelihood of positive gut contents (0.5 km: χ² =

74.23, P < 0.01; 3.0 km: χ² = 58.35, P < 0.01).

4.5 Discussion

Landscape factors such as cover diversity and fragmentation alter the abundance and diversity of different species, particularly arthropod predators

(Marino and Landis 1996, Gardiner et al. 2009). The ant data support this trend, where increased land cover diversity (SIDI) and decreased fragmentation

(PLAND, TE, and CLUMPY) resulted in greater ant abundance. Ant populations were impacted less by landscape factors as scale increased, potentially due to the dispersal distances often covered by ants and the supplemental resources that areas near the target crop fields provided (Carroll and Janzen 1973, Diaz

1992, Buczkowski and VanWeelden 2010). Differential land cover (PLANDs) interactions with ant and pest populations were also potentially mediated by resource provisioning (Holland and Fahrig 2000, Phillips and Gardiner 2016) and dispersal (Wratten et al. 2003). Arthropod interactions cover types indicate the differential use of patch resources (Lagerlof and Wallin 1993) such as refugia

(Pfiffner and Luka 2000) and alternative food resources (Tscharntke et al. 2005).

For instance, “other hay/ non-alfalfa” was positively associated with ant abundance while they were negatively association with pest abundance; while

sod/grass areas benefitted pests to the detriment of the ant community. Other highly disturbed cover types such as pasture, soybean, and tobacco and densely wooded covers like deciduous and evergreen forest exerted negative influences on both ants and pests. All interactions with cover types were scale dependent in direction and magnitude, unsurprising since distance from focal field and area within the buffer changed as scale increased.

The SADIE data revealed that landscape variables interacted oppositely with ant and pest communities. When ants and pests are present in large quantities in the same field, they are likely to be aggregated to each other within that field which is not necessarily usual for generalist predators in agroecosystems (Pearce and Zalucki 2005). Decreased fragmentation increased the likelihood of pest aggregations within the field, while increased land cover diversity increased the likelihood of ant aggregations. SADIE association data indicated that ant to pest aggregations are dictated by a combination of land cover diversity and fragmentation, unsurprising as these associations relied on both ant and pest communities. When the spatial associations of ants to pests occurred, it corresponded with ant predation on pest species as indicated by the molecular data. These predation events were further influenced by increased ant abundance but not diversity as well as decreased fragmentation but not increased land cover diversity.

Predator biodiversity has been thought to increase biological control measures as an increase in diversity might be correlated with an increase of predator niche-breadth, resulting in greater predation rates (Casula et al. 2006,

Finke and Snyder 2010, Crowder and Jabbour 2014). Our study did not uphold this hypothesis - increased ant abundance but not diversity resulted in an increase in spatial aggregation and predation as indicated by molecular gut contents. The data, therefore, support the idea that certain predator species are more important to pest predation than biodiversity (Straub and Snyder 2006), particularly L. neoniger. Although this ant is known to farm hemipterans detrimental to crop systems (Schwartzberg et al. 2010), it is also known to be a highly voracious predator (Whitcomb et al. 1972, Zenger and Gibb 2001b).

Additionally, this species provided added ecosystem services of encouraging soil aeration as its colonies are often spread out and peak in size during the middle of summer when soil health is important to the crops (Wang et al. 1995).

This study indicates that landscape factors such as cover diversity, habitat fragmentation, cover classifications, and spatial scale significantly influence predator communities in agroecosystems. We found that as land cover diversity increased around agroecosystems, so does the ant abundance within that agricultural setting; the inverse held true for the combined pest community.

Additionally, these two communities were influenced by separate land cover classifications at the smaller scale, indicating that habitats providing resources to ants and pests were distinctly different and distance-dependent. The resulting changes in ant abundance correlated with increased spatial aggregation of ant communities to each other and to prey items (crop pests) within agroecosystems.

This information was corroborated by the molecular gut content analysis that indicated ants were actively preying upon select pest species in the same fields

in which spatial aggregations were occurring. These trends could be used in the future to further our understanding of the scope of landscape ecology and how to harness the combination of geospatial and molecular methods to measure biocontrol for more sustainable agroecosystems.

Table 4-1 Landscape attribute abbreviations and definitions used within the landscape analyses component of this study (taken from Fragstats Help 4.2. Where the class scale are computed for every class type in the landscape, and landscape scale is computed for entire patch mosaics where the resulting output is a single value for each field at each scale.

Abbreviation Scale Used Metric Definition Type SIDI Landscape Diversity Simpson’s Diversity Index, 0 ≤ SIDI ≤ 1, 1 minus the sum across all patch types of the proportional abundance of each patch type squared. PLAND Class Area Percentage of the landscape, 0 ≤ PLAND ≤ 100, proportion abundance of each patch type in the landscape. This is a measure of landscape composition. 67 TE Class, Area Total edge (m), TE ≥ 0, the sum of the lengths (m) of

Landscape all edge segments involving the corresponding patch type. CLUMPY Class Aggregation Clumpiness index (percentage), -1 ≤ CLUMPY ≤ 1, the proportion deviation of the proportion of like adjacencies involving the corresponding class from that expected under spatially random distribution.

Table 4-2 Non-target species tested for cross-reactivity with the Colaspis brunnea (CB348F and CB531R), Chinavia hilaris (AH276F and AH390R), and Hypena scabra (HS517F and HS598R) primers designed for this study.

Order Family Species No. Tested Araneae Linyphiidae Tennesseellum formica 1 Araneae Salticidae 1 Araneae Tetragnathidae Glenognatha foxi 1 Araneae Tetragnathidae 2 Araneae 9 Coleoptera Aderidae 1 Coleoptera Anthicidae 1 Coleoptera Carabidae 2 Coleoptera Chrysomelidae Colaspis brunnea 10 Coleoptera Chrysomelidae Diabrotica undecimpunctata 1 Coleoptera Chrysomelidae 1 Coleoptera Coccinellidae Coccinella septempunctata 1 Coleoptera Coccinellidae Coleomegilla maculata 1 Coleoptera Coccinellidae Hippodamia convergens 1 Coleoptera Coccinellidae Scymnus sp. 1 Coleoptera Coccinellidae 2 Coleoptera Curculionidae Hypothenemus hampei 1 Coleoptera Lathridiidae 1 Coleoptera Meloidae 1 Coleoptera Melyridae 1 Coleoptera Scarabaeidae Popillia japonica 10 Coleoptera Staphylinidae 2 Diptera Agromyzidae 1 Diptera Anthomyiidae Delia antiqua 1 Diptera Anthomyiidae 5 Diptera Anthomyzidae 1 Diptera Chironomidae 2 Diptera Chloropidae Thaumatomyia glabra 1 Diptera Chloropidae 1 Diptera Dolichopodidae 2 Diptera Drosophilidae 5 Diptera Hybotidae 2 Diptera Syrphidae 3 Diptera Tipulidae 1 Diptera 1 Gastropoda Polygyridae Mesodon zaletus 1 Hemiptera Aleyrodidae Bemisia tabaci 1

Table 4-2 continued Order Family Species No. Tested Hemiptera Aphididae Brevicoryne brassicae 1 Hemiptera Aphididae Capitophorus eleagni 1 Hemiptera Aphididae Hyadaphis pseudo 1 brassicae Hemiptera Aphididae Myzus persicae 1 Hemiptera Aphididae 1 Hemiptera Cicadellidae 2 Hemiptera Coccidae Coccus hesperidum 1 Hemiptera Coccidae Eulecanium cerasorum 1 Hemiptera Coccidae Neolecanium cornuparvum 1 Hemiptera Coccidae Parthenolecanium quercifex 1 Hemiptera Coccidae Pseudococcus maritimus 1 Hemiptera Coccidae Pulvinaria innumerabilis 1 Hemiptera Cydnidae Sehirus cinctus 1 Hemiptera Geocoridae 1 Hemiptera Nabidae 2 Hemiptera Nabidae 1 Hemiptera Pentatomidae Euschistus servus 1 Hemiptera Pentatomidae Nezara viridula 1 Hemiptera Pentatomidae Oebalus pugnax 1 Hemiptera Pentatomidae 2 Hemiptera Psyllidae Psylla pyri 1 Hemiptera Reduviidae 3 Hemiptera Rhyparachromidae 1 Hemiptera Thyreocoridae 1 Hemiptera Triozidae Bactericera cockerelli 1 Heterobranchia Discidae Aniguispira alternata 1 Hymenoptera Argidae 1 Hymenoptera Bethylidae Prorops nasuta 1 Hymenoptera Braconidae 4 Hymenoptera Chalcididae 1 Hymenoptera Crabronidae 2 Hymenoptera Eulophidae Phymastichus coffea 1 Hymenoptera Figitidae 1 Hymenoptera Formicidae Camponotus floridanus 3 Hymenoptera Formicidae Lasius neoniger 10 Hymenoptera Formicidae Monomorium minimum 10 Hymenoptera Formicidae Solenopsis invicta 3 Hymenoptera Formicidae Solenopsis molesta 10

Table 4-2 continued Order Family Species No. Tested Hymenoptera Formicidae Tapinoma sessile 6 Hymenoptera Formicidae Tapinoma sp 1 Hymenoptera Formicidae Tetramorium caespitum 3 Hymenoptera Formicidae 1 Hymenoptera Ichneumonidae 4 Hymenoptera Pompilidae 1 Lepidoptera Erebidae Hypena scabra 10 Lepidoptera Noctuidae 1 Mantidae 1 Neuroptera Hemerobiidae 2 Orthoptera Tettigoniidae 1 Orthoptera 1 Primates Hominidae Homo sapiens 1 Psocoptera 1 Thysanoptera Phlaeothripidae Karnyothrips flavipes 1 Thysanoptera Thripidae Frankliniella occidentalis 1

Table 4-3 Soybean pest species primers designed for ant molecular gut content testing.

Target Species Primer Name Primer Sequence (5’ to 3’) Size (bp)

C. brunnea CB348F CAAATTTATAATTCAATCGTAACG 229

CB531R CTGTTCATCCAGTTCCTACTCCG

C. hilaris AH276F AGACCCTATAGAATTTTATTTTAAAG 114

AH390R CCTAAAAATAATTATATTTAAACC

H. scabra HS517F CCCTTCTTTAACCCTTCTAATTTCC 129

HS598R GAAAAAATAGCTAAATCTACAGATCT

Table 4-4 Parameter estimates from regression results (* indicates P < 0.05) where ant and pest abundance were analyzed using PROC GENMOD with a negative binomial distribution against landscape and class metrics, community metrics, and abiotic conditions at 0.5km and 3.0km scales with the individual field as a random effect. 0.5 KM Buffer 3.0 KM Buffer Metric Metric Scale Ants Pests Ants Pests Time N/A 0.00 0.00 -0.04 -0.03 Field Area (Ha) N/A 0.01 0.00 -0.04* 0.01 Avg. Temp. (C) N/A 0.08* 0.03* 0.05* 0.05* Avg. Wind (mph) N/A 0.03 0.01 0.01 0.03 Ant Abundance N/A -0.02 -0.01 Ant Diversity N/A 1.06* -0.16* 1.09* -0.17* 72 Pest Abundance N/A -0.02 0.01 Pest Diversity N/A -0.19 1.74* -0.06 1.73* TE Landscape 0.00 0.00 0.01* 0.00 SIDI Landscape 1.23* 0.33 1.17* -1.72* Avg. PLAND Class -0.08* -0.06* 0.25* -0.11* Avg. CLUMPY Class 1.04* 0.56* 5.23* -0.06

Table 4-5 Parameter estimates from regression results (* indicates P < 0.05) where ant and pest diversity were analyzed using PROC GENMOD with a negative binomial distribution against landscape and class metrics, community metrics and abiotic conditions at 0.5km and 3.0km scales with the individual field as a random effect. 0.5 KM Buffer 3.0 KM Buffer Metric Metric Scale Ants Pests Ants Pests Time N/A -0.02 0.11 -0.03 0.18* Field Area (Ha) N/A 0.00 -0.02 -0.01 0.03* Avg. Temp. (C) N/A -0.04* -0.04 -0.04* -0.02 Avg. Wind (mph) N/A -0.04 -0.05 -0.03 0.00 Ant Abundance N/A 0.09* -0.05 0.10* -0.05 Ant Diversity N/A -0.21* -0.27* Pest Abundance N/A -0.04 0.24* -0.04 0.13* N/A -0.26 -0.25 73 Pest Diversity

TE Landscape 0.00 0.00 0.00 0.00 SIDI Landscape -0.54 0.89 -0.47 -1.09 Avg. PLAND Class -0.03 0.01 -0.08 0.46* Avg. CLUMPY Class 0.62* 0.32 0.54 4.69*

Table 4-6 Regression results (* indicates P < 0.05) where ant and pest abundance were analyzed using PROC GENMOD with a negative binomial distribution against individual cover type (USDA-defined categories) PLANDs at 0.5km and 3.0km scales. Empty cells indicate covers that were not observed in that buffer size. 0.5KM 3.0KM Cover Type Ants Pests Ants Pests Alfalfa -1.01* -0.28 -36.04* 8.71 Clover/Wildflower 0.00 0.00 -36.09 15.39 Corn 0.01 0.04* 2.18* -0.49 Dbl. Crop -0.04* 0.00 -0.58 0.15 Wheat/Soy Deciduous Forest -0.12* -0.13* 1.61* -0.30 Developed -0.09* -0.11* 3.27* -0.99* 0.39 -0.67 -122.04* 33.49 74 Evergreen Forest

Grass/Pasture -0.03* -0.02* 0.66* -0.16 Herbaceous 0.06 0.47* -29.31 17.89 Wetlands Oats 6794* -1460 Open Water 0.48* -0.37* 22.50* -7.66 Other Hay/ Non- 1.19* 4.79* -1.49* Alfalfa Popcorn -10398 4008 Shrubland 0.86 3.07* 97.87* -28.03 Sod/Gass -1091* 409.20* Sorghum -1054* 337.25 Soybean -0.04* -0.03* 1.25* -0.64* Tobacco -10.41* -1.74 64.36* -9.50 Walnut -3667 276.90 Wheat 116.37 -26.52 Woody Wetlands -0.55 -2.72* 26.64* -7.54

Table 4-7 Regression results (* indicates P < 0.05) where ant and pest diversity were analyzed using PROC GENMOD with a negative binomial distribution against individual cover type (USDA-defined categories) PLANDs at 0.5km and 3.0km scales. Empty cells indicate covers that were not observed in that buffer size. 0.5KM 3.0KM Class Type Ants Pests Ants Pests Alfalfa -0.38 -0.10 -21.36* 5.31 Clover/Wildflower 0.00 0.00 -37.84 1.97 Corn 0.01 0.05* 1.25 0.00 Dbl. Crop Wheat/Soy -0.05* 0.01 -0.21 0.01 Deciduous Forest -0.09* -0.07 0.57 -0.16 Developed, Medium -0.07* -0.05 1.95 -0.46 Evergreen Forest 0.13 -2.90* -25.75 27.35 Grass/Pasture -0.02* 0.00 0.48 -0.01 75

Herbaceous -0.08 0.34 -39.92 -7.98 Wetlands Oats 4104* -596.03 Open Water 0.44* -0.28 -2.67 -9.17 Other Hay/ Non- 0.39 0.45 1.40 -1.40 Alfalfa Popcorn -11620 -18.01 Shrubland 1.92 3.82* 44.29 -19.00 Sod/Gass -895.83 503.23 Sorghum -329.42 224.87 Soybean -0.03* -0.01 1.46* -0.21 Tobacco -5.59* -1.87 24.54 -0.74 Walnut 2973 2224 Wheat 11.32 -36.64 Woody Wetlands -0.46 -2.58* 1.62 -6.92

Table 4-8 Spatial Analysis by Distance IndicEs (SADIE) results for ant and pest communities by field and generalized collection time. P_a indicates the P value for overall spatial dispersal, P_vj indicates the P value for positive aggregations/ clusters, while P_vi indicates the P value for negative aggregations/gaps.

Ants Pests Time Field P_a P_vj P_vi P_a P_vj P_vi 1 A12 0.333 0.2564 0 0.5897 0.5385 0.0513 1 A13 0.3367 0.993 0.7558 0.7605 0.6204 0.8208 1 A14 0.4272 0.901 0.6868 0.2522 0.861 0.8229 1 B12 0.2564 0.1795 0.1026 0.7179 0.4615 0.5897 1 B13 NA NA NA 0.764 0.923 0.6296 1 B14 0.8093 0.922 0.6399 0.052 1.104 0.2214 1 C12 0.9744 0.8462 0.1538 0.9487 0.8462 0.9744 1 C13 0.8705 0.867 0.7305 0.538 0.973 0.4853 1 C14 0.5212 0.981 0.479 0.7795 0.986 0.431 1 D12 0.9744 0.8718 0.8974 0.9744 0.9744 1 1 D13 0.2936 1.264 0.0605 0.2631 0.977 0.476 1 D14 NA NA NA 0.0799 1.118 0.1847 1 E13 0.8016 0.994 1 0.0618 1.425 0.0741 1 E14 0.704 0.986 0.4981 0.1988 1.183 0.1133 1 F13 0.8792 0.9 0.7295 0.3335 0.916 0.6737 1 F14 0.419 1.016 0.3638 0.3734 0.983 0.4486 1 G13 0.3367 0.993 0.7558 0.4766 0.991 0.4463 1 I12 0.2051 0.1795 0.1282 0.7949 0.8205 0.8974 1 O12 0.2564 0.1282 0.3846 0.7179 0.7179 0.6923 1 P12 0.9487 0.8974 0.8974 0.8462 0.5897 0.8205 1 T12 0.6923 0.641 0.1538 0.3333 0.7692 0.3077 1 U12 0.9231 0.9231 0.4359 0.7692 0.8718 0.8718 1 V12 0.7179 0.5641 0.6923 0.4103 0.5385 0.6667 2 A12 0.8718 0.8205 0.9231 0.1538 0.2051 0.2821 2 A14 0.2572 0.941 0.58 0.2645 0.996 0.4041 2 B12 0.2821 0.3077 0.359 0.6667 0.5897 0.5128 2 B14 0.0045 1.213 0.1029 0.0065 1.339 0.0277 2 C12 0.7436 0.641 0.9487 0.4103 0.3333 0.641 2 C14 0.6548 0.857 0.883 0.9583 0.803 0.9397 2 D12 0.0256 0.0256 0.0769 0.7436 0.7949 0.8205 2 D14 0.0072 1.424 0.0156 0.8384 0.902 0.7438 2 E14 0.5614 0.912 0.6866 0.9341 0.799 0.94 2 F14 0.2125 1.045 0.3432 0.9005 0.902 0.7077 2 I12 0.3846 0.4103 0.5385 0.2564 0.4872 0.1282 2 O12 0.0256 0.0256 0 0.4615 0.6154 0.641 2 P12 0.2821 0.4359 0.2051 0.1026 0.1026 0.0256

Table 4-8 continued Ants Pests Time Field P_a P_vj P_vi P_a P_vj P_vi 2 U12 0.1282 0.2051 0.3846 0.3846 0.4872 0.4103 2 V12 0.2821 0.2308 0.3846 0.5897 0.6667 0.7949 3 A12 0.7692 0.6154 0.7179 0.1026 0.0769 0.1026 3 A14 0.3352 0.855 0.8668 0.6657 0.78 0.9683 3 B14 0.2343 1.112 0.1946 0.7954 1.03 0.3653 3 C14 0.9824 0.821 0.9678 0.24 1.056 0.2966 3 D14 0.5401 1.001 0.6301 0.3266 0.965 0.5232 3 E14 0.4961 1.067 0.0399 0.4138 1.037 0.3189 3 F14 0.1564 1.08 0.2197 0.0065 1.449 0.0112 3 U12 0.0769 0.0513 0.1026 0.8205 0.9231 0.8718 4 A14 0.1666 0.848 0.8731 0.5602 0.355 0.325 4 B14 0.2381 0.469 0.5098 0.7374 0.797 0.9219 4 C14 0.2805 0.908 0.7263 0.3104 1.016 0.3694 4 D14 0.2172 0.967 0.1047 0.3665 0.937 0.604 4 E14 0.85 0.842 0.9435 0.2921 0.966 0.5219 4 F14 0.2985 0.984 0.4691 0.1741 1.124 0.1817 1 A12 0.333 0.2564 0 0.5897 0.5385 0.0513 1 A13 0.3367 0.993 0.7558 0.7605 0.6204 0.8208 1 A14 0.4272 0.901 0.6868 0.2522 0.861 0.8229 1 B12 0.2564 0.1795 0.1026 0.7179 0.4615 0.5897 1 B13 NA NA NA 0.764 0.923 0.6296 1 B14 0.8093 0.922 0.6399 0.052 1.104 0.2214 1 C12 0.9744 0.8462 0.1538 0.9487 0.8462 0.9744 1 C13 0.8705 0.867 0.7305 0.538 0.973 0.4853 1 C14 0.5212 0.981 0.479 0.7795 0.986 0.431 1 D12 0.9744 0.8718 0.8974 0.9744 0.9744 1 1 D13 0.2936 1.264 0.0605 0.2631 0.977 0.476 1 D14 NA NA NA 0.0799 1.118 0.1847 1 E13 0.8016 0.994 1 0.0618 1.425 0.0741 1 E14 0.704 0.986 0.4981 0.1988 1.183 0.1133 1 F13 0.8792 0.9 0.7295 0.3335 0.916 0.6737 1 F14 0.419 1.016 0.3638 0.3734 0.983 0.4486 1 G13 0.3367 0.993 0.7558 0.4766 0.991 0.4463 1 I12 0.2051 0.1795 0.1282 0.7949 0.8205 0.8974 1 O12 0.2564 0.1282 0.3846 0.7179 0.7179 0.6923 1 P12 0.9487 0.8974 0.8974 0.8462 0.5897 0.8205 1 T12 0.6923 0.641 0.1538 0.3333 0.7692 0.3077 1 U12 0.9231 0.9231 0.4359 0.7692 0.8718 0.8718 1 V12 0.7179 0.5641 0.6923 0.4103 0.5385 0.6667

Table 4-8 continued Ants Pests Time Field P_a P_vj P_vi P_a P_vj P_vi 2 A12 0.8718 0.8205 0.9231 0.1538 0.2051 0.2821 2 A14 0.2572 0.941 0.58 0.2645 0.996 0.4041 2 B12 0.2821 0.3077 0.359 0.6667 0.5897 0.5128 2 B14 0.0045 1.213 0.1029 0.0065 1.339 0.0277 2 C12 0.7436 0.641 0.9487 0.4103 0.3333 0.641 2 C14 0.6548 0.857 0.883 0.9583 0.803 0.9397 2 D12 0.0256 0.0256 0.0769 0.7436 0.7949 0.8205 2 D14 0.0072 1.424 0.0156 0.8384 0.902 0.7438 2 E14 0.5614 0.912 0.6866 0.9341 0.799 0.94 2 F14 0.2125 1.045 0.3432 0.9005 0.902 0.7077 2 I12 0.3846 0.4103 0.5385 0.2564 0.4872 0.1282 2 O12 0.0256 0.0256 0 0.4615 0.6154 0.641 2 P12 0.2821 0.4359 0.2051 0.1026 0.1026 0.0256 2 U12 0.1282 0.2051 0.3846 0.3846 0.4872 0.4103 2 V12 0.2821 0.2308 0.3846 0.5897 0.6667 0.7949 3 A12 0.7692 0.6154 0.7179 0.1026 0.0769 0.1026 3 A14 0.3352 0.855 0.8668 0.6657 0.78 0.9683 3 B14 0.2343 1.112 0.1946 0.7954 1.03 0.3653 3 C14 0.9824 0.821 0.9678 0.24 1.056 0.2966 3 D14 0.5401 1.001 0.6301 0.3266 0.965 0.5232 3 E14 0.4961 1.067 0.0399 0.4138 1.037 0.3189 3 F14 0.1564 1.08 0.2197 0.0065 1.449 0.0112 3 U12 0.0769 0.0513 0.1026 0.8205 0.9231 0.8718 4 A14 0.1666 0.848 0.8731 0.5602 0.355 0.325 4 B14 0.2381 0.469 0.5098 0.7374 0.797 0.9219 4 C14 0.2805 0.908 0.7263 0.3104 1.016 0.3694 4 D14 0.2172 0.967 0.1047 0.3665 0.937 0.604 4 E14 0.85 0.842 0.9435 0.2921 0.966 0.5219 4 F14 0.2985 0.984 0.4691 0.1741 1.124 0.1817

Table 4-9 Regression results (* indicates P < 0.05) where ant and pest SADIE (significant spatial trends) were analyzed using PROC LOGISTIC regression against landscape and class metrics, community metrics, and abiotic conditions at 0.5km and 3.0km scales with the individual field as a random effect. 0.5 KM Buffer 3.0 KM Buffer Metric Metric Scale Ants Pests Ants Pests Time N/A 0.46* 0.14 0.00 0.41* Field Area (Ha) N/A -0.05* -0.26* -0.21* -0.24* Avg. Temp. (C) N/A -0.02 0.05 -0.06* -0.05 Avg. Wind (mph) N/A -0.38* 0.75* -0.52* 0.40* Ant Abundance N/A 0.01 -0.12* -0.02 -0.09* Ant Diversity N/A 0.42* -0.14 0.27* 0.16 Ant SADIE N/A 1.85* 1.57* Pest Abundance N/A 0.05 0.06 0.14* 0.01 79

Pest Diversity N/A -0.71* 0.22 -0.28 -0.37 Pest SADIE N/A 1.69* 0.81* TE Landscape -0.04* 0.05* 0.00 0.02* SIDI Landscape 10.67* -20.68* 0.27 0.58 Avg. PLAND Class -0.19* 0.33* -1.88* 1.23* Avg. CLUMPY Class -2.42* 2.16* -9.44* 43.47*

Table 4-10 Spatial Analysis by Distance IndicEs (SADIE) association (NA) results for ants to pests by field and generalized collection time. Chi_P indicates the χ value output where positive indicates that the two groups are clustered together positively and negative indicates that the two groups are avoiding each other (gaps). The probability is the association P value of χ statistic. Time Field Chi_P Probability 1 A12 0.1552 0.3608 1 A13 0.3732 0.2629 1 A14 0.6292 0.0572 1 B12 0.6106 0.0008 1 B13 0 1 1 B14 0.3604 0.2149 1 C12 -0.5828 0.8467 1 C13 0.2695 0.2112 1 C14 -0.2244 0.566 1 D12 0.3832 0.1726 1 D13 -0.0262 0.5118 1 D14 0 1 1 E13 -0.8016 0.9499 1 E14 0.2886 0.3446 1 F13 -0.4238 0.7671 1 F14 0.308 0.2673 1 G13 0.1394 0.4549 1 I12 -0.5129 1 1 O12 -0.0725 0.574 1 P12 0.534 0.033 1 T12 -0.2267 0.7566 1 U12 0.1639 0.286 1 V12 0.0246 0.4745 2 A12 0.1167 0.4266 2 A14 0.2348 0.2608 2 B12 0.4089 0.0153 2 B14 -0.2507 0.7382 2 C12 -0.2396 0.5906 2 C14 -0.3057 0.6554 2 D12 0.4495 0.1741 2 D14 -0.1325 0.59 2 E14 0.7482 0.0524 2 F14 0 1 2 I12 0.1767 0.3062 2 O12 0.1783 0.3054

Table 4-10 continued Time Field Chi_P Probability 2 P12 0.6522 0.0115 2 U12 -0.0645 0.5889 2 V12 0.047 0.4544 3 A14 0.482 0.0971 3 B14 0.0963 0.4166 3 C14 -0.044 0.5526 3 D14 -0.4601 0.6787 3 E14 -0.5441 0.7899 3 F14 0.4931 0.2164 4 A14 0.0406 0.4609 4 B14 -0.1134 0.6351 4 C14 -0.0156 0.5138 4 D14 0.0349 0.481 4 E14 0.0336 0.4864 4 F14 0.3399 0.2881

Table 4-11 Regression results (* indicates P < 0.05) where ant and pest SADIE associations (significant spatial trends) between ants and pests were analyzed using PROC LOGISTIC regression against landscape and class metrics, community metrics and abiotic conditions at 0.5km and 3.0km scales with the individual field as a random effect. Metric Metric Scale 0.5 KM Buffer 3.0 KM Buffer Time N/A 1.76* 0.49* Field Area (Ha) N/A 1.98* 0.94* Avg. Temp. (C) N/A -1.72* -0.70* Avg. Wind (mph) N/A 3.42* 1.14* Ant Abundance N/A -0.07 -0.09 Ant Diversity N/A 0.03 0.04 Ant SADIE N/A -22.72* -0.92* Pest Abundance N/A 0.25 -0.16* Pest Diversity N/A 0.64 0.40 82 Pest SADIE N/A -4.41* -3.82*

TE Landscape -0.54* -0.05* SIDI Landscape 39.19* -21.62* Avg. PLAND Class -13.90* 1.79* Avg. CLUMPY Class -81.50* 138.80*

Table 4-12 Molecular gut content results by ant species tested for Colaspis brunnea, Chivaria hilaris, and Hypena scabra. Ant species tested C. brunnea C. hilaris H. scabra Lasius neoniger (n=413) 13.4% 0% 2.9% Monomorium minimum (n=92) 0% 0% 0% Tapinoma sessile (n=92) 0% 0% 0% Tetramorium caespitum (n=92) 0% 0% 0% Pheidole tysoni (n=38) 0% 0% 36.8%

Solenopsis molesta (n=29) 0% 0% 13.8%

Table 4-13 Regression results (* indicates P < 0.05) where ant molecular gut contents (DNA Yes/No?) were analyzed using PROC LOGISTIC regression against landscape and class metrics, community metrics, and abiotic conditions at 0.5km and 3.0km scales with the individual field as a random effect. Metric Metric Scale 0.5 KM Buffer 3.0 KM Buffer Time N/A 1.03* 0.81* Field Area (Ha) N/A 0.11* -0.07* Avg. Temp. (C) N/A 0.12* 0.02 Avg. Wind (mph) N/A 0.45* 0.44* Ant Abundance N/A 0.15* 0.18* Ant Diversity N/A 0.35 0.35* Ant SADIE N/A 1.07* -0.75* Pest Abundance N/A -0.07 -0.04 Pest Diversity N/A 0.37 0.74* 84 Pest SADIE N/A 0.02 0.57* SADIE NA N/A 1.81* 1.56* TE Landscape 0.10* 0.01* SIDI Landscape -11.95* -2.48* Avg. PLAND Class 0.46* -0.37* Avg. CLUMPY Class 8.75* 1.46*

Figure 4-1 Three sampled field sites (dots), with 0.5 km and 3.0 km buffers around the corresponding year’s Crop Data Layer. Sites indicate the range in landscape diversity (SIDI), average percentage of landscape per cover type (PLAND), total edge length (TE), and cover type clumpiness (CLUMPY). The mapping projection (WGS 1984) causes buffers to appear oblong.

Figure 4-2 The average percentage of the landscape (PLAND) by USDA Crop Data Layer land cover classification for (A) 0.5 and (B) 3.0 km buffers.

Figure 4-3 Graph of the parameter estimates associated with (A) ant and pest abundance with landscape diversity (SIDI), (B) ant and pest diversity with SIDI, (C) ant and pest abundance with the class metrics of average percentage of landscape (PLAND) and average total edge distance (TE) and average clumpiness of covers (CLUMPY), and (D) ant and pest diversity with PLAND, TE, and CLUMPY. Positive parameter estimates indicate variables that increase the respective arthropod community while negative estimates indicate the opposite effect.

Figure 4-4 Relationship of ant and pest community abundances using SADIE association analyses for one select field with significant aggregations of (A) ants and (B) pests and (C) significant association of ants to pests. Increased color saturation indicates an increase in population density or aggregation within the field while a lack of color indicates gaps of the populations within the fields. In (C), purple saturation indicates spatial associations of ant and pest aggregations.

Figure 4-5 Graph of the parameter estimates associated for ant SADIEs, pest SADIEs, SADIE association of ants to pests, and predation (PCR) with the landscape metrics of diversity (SIDI) and fragmentation (PLAND, TE, and CLUMPY) for (A) 0.5 km buffers and (B) 3.0 km buffers. Positive parameter estimates indicate variables that increase the respective metric while negative estimates indicate the opposite effect.

Chapter 5 FIELD MARGIN COMPOSITION INFLUENCES EARLY SEASON PEST AND SPIDER SPATIAL ASSOCIATIONS

Chapter contents prepared for publication in Penn, HJ. Field margin composition influences early season pest and spider spatial associations. Journal of Applied Ecology.

5.1 Summary

Crop field margins provide benefits to growers by provisioning alternative food resources and refuges for predators. Pests are known to interact with margins in different ways depending on host plant requirements and dispersal modes, altering their within field interactions with predators. This study quantified soybean pest and spider abundances in relation to margin composition and scale. Within field spatial distributions of soybean pests and spiders in relation to field margins, community abundances, and abiotic conditions were determined to elucidate potential pest-spider interactions. Pest and spider abundances in soybean fields were positively correlated with each other and with increased wind speed. Pest abundance increased if field margins were located from 0-10 m but not 10-20 m and were composed of soybeans. Spider abundances only benefitted from the presence of wooded margins 10-20 m from the field. Both pest and spider Spatial Analysis by Distance IndicEs were impacted by margin type and distance interactions. Pests were more likely to exhibit non-random spatial patterns when pest abundances were small, early in the season. When spider spatial associations to pests were analyzed, the likelihood of spider aggregation increased with the presence of pest aggregations. Wooded and grass margins from 10-20 m also increased the chances of positive spatial overlap between pests and spiders. Our results indicate that pest and spider

abundance in soybeans was dictated by a combination of abundance, land use, and abiotic conditions. Woody areas proved to be a source for spider populations; and this effect was increased with greater wind speed, potentially allowing greater spider dispersal. Additionally, spiders were more likely to overlap with pest populations earlier in the season, indicating that early season predation might be occurring as spatial overlap has been shown to indicate predation in previous studies. To increase the likelihood of predation events via conservation biocontrol, growers should ensure that soybean fields are not immediately adjacent to each other but are separated by another margin type, preferably a tree/shrub line.

5.2 Introduction

Agriculture covers 38% of global arable land (Clay 2004, Donald and Evans

2006), which is concerning since landscape simplification via agricultural intensification leads to biodiversity (Altieri 1999) and associated ecosystem service loss (Menalled et al. 1999, Macfadyen et al. 2011) as well as an increase in dis-services (Zhang et al. 2007, Meehan et al. 2011). When landscapes no longer include high-quality habitat for beneficial species, an increase in mortality and a loss of species richness is exhibited (Chisholm et al. 2011), but when landscapes contain patches of undisturbed habitat, beneficial species (i.e. natural enemies and pollinators) can flourish (Gardiner et al. 2010, Batary et al. 2011,

Cusser et al. 2016).

Within agroecosystems, such patches often involve unmanaged field margins (Potts et al. 2016) as these areas incur little to no cost to growers compared to curated natural areas (Nelson et al. 1998, Brodt et al. 2008). Margin composition is especially influential as field edges are the gateway for arthropods into agricultural areas after disturbance events (i.e. planting, chemical application, harvest) (Altieri and Schmidt 1986, Buddle et al. 2004). The composition of field margins can exert pressure on species interactions in multiple ways – by altering movements, inducing species-specific mortality, changing interspecific interactions between edge and field matrices, and by providing novel spaces for interactions (Fagan et al. 1999). By changing the content and structure of these margins, species diversity and subsequent ecosystem functioning such as biological control within adjacent agricultural fields can be manipulated (Thomas et al. 1991, Dauber and Wolters 2005,

Chaplin-Kramer et al. 2011) via this physical or interaction barrier (Hawkes 1973,

Bowden and Dean 1977, Wratten et al. 2003).

Unmanaged field margins have the potential to act as a barrier for pest dispersal between crop fields by two means as proposed by theoretical models

(Bhar and Fahrig 1998). The first mechanism is by blocking the path into a field with an inhospitable environment where the arthropod is without high-quality food, mates, or necessary abiotic conditions. For example, Hawkes (1973) reported that the height and permeability of crop borders, and the extent to which they acted as windbreaks, influenced adult cabbage root fly in-field abundance and aggregation. Additionally, Holland and Fahrig (2000) and Wratten et al.

(2003) found differential effects of wooded areas like tree lines and shrubs on overall insect dispersal rates. The second is accomplished by trapping the arthropod within the border with preferred host plants or mate access, luring it away from the field (Stinner et al. 1983). However, given the lack of consistency between systems, it is critically important to determine how different categories of vegetative borders impact pest species dispersal into agroecosystems.

Besides posing as a barrier to or reservoir for pest species, field margins can supplement natural enemies (conservation biological control) (Bianchi et al.

2005, Kremen et al. 2011). Not only do margins provide refuge from disturbance

(Landis et al. 2000, Lee et al. 2001) and spaces for overwintering (Holland et al.

2008), they also provide alternative food resources such as non-pest prey items

(Harwood and Obrycki 2005b, Amaral et al. 2013) and plant-based foods such as pollen and extra-floral nectar (Dyer and Landis 1997, Landis et al. 2005). For instance, natural enemy populations can be increased by up to 80% and 71% when given access to herbaceous habitats and woody habitats respectively in addition to cropland (Bedford and Usher 1994, Bianchi et al. 2006). Increased natural enemy populations often result in greater biocontrol of pest arthropods

(Brewer and Elliott 2004) and weeds (Booman et al. 2009, Birthisel et al. 2014,

Birthisel et al. 2015).

One particular group of natural enemies that can readily disperse from field margin habitats are spiders, generalist predators within agroecosystems

(Nyffeler et al. 1987, Young 1990, Maloney et al. 2003, Kuusk and Ekbom 2012) and contributors to non-consumptive control of pest damage (Rypstra and

Buddle 2013, Penn and Athey In prep.). For instance, margin composition surrounding orchards has been shown to have a significant impact on spider abundance and subsequent suppression of codling moth (Lefebvre et al. 2016).

The resources provided by non-crop habitats are important for spider community maintenance (Taylor and Pfannenstiel 2008, Nyffeler et al. 2016), enabling these populations to move into agricultural fields early in the season for pest suppression (Wissinger 1997, Buddle et al. 2004, Gavish-Regev et al. 2008).

This movement of natural enemies into crop fields is especially important for suppressing early season pest populations before they can build up to higher numbers (Harwood et al. 2007).

This study determined how field margins at two spatial scales altered the within field abundance of arthropod pest species and spiders (Figure 5-1). I also assessed the impact of field margins on the likelihood of within field spatial aggregations of pests and spiders and between these groups, as the spatial overlap of predators and prey can indicate corresponding predation events

(Winder et al. 2001, Bell et al. 2010, Zhao et al. 2013). The results of this study inform both how to manage field margins for pest species-specific control but also how to supplement conservation biological control efforts by promoting spider populations.

5.3 Materials and Methods

5.3.1 Study Sites, Sample Collection, and Field Margin Composition

Grower managed soybean fields (n = 23) throughout central and western

Kentucky were sampled from 2012-2014 according to methods in Chapter 4, page 52 (Penn et al. In review), using pooled wet pitfall traps, sweep samples, and hand samples (n = 1282) for an overall view of the spider and arthropod pest community (Lang 2000). Pests from all collection methods were identified to species or family when applicable and spiders were identified to family (AAS

2005, Johnson et al. 2015). Field margins were evaluated for general composition from 0-10 m and 10-20 m from the outermost row of soybeans in the focal field. All fields were observed to have soybean, grassy areas (unmanaged, weedy, not wooded), or wooded areas from 0-10 m and soybean, grassy areas, wooded areas (trees and dense shrub), other crop (typically corn or tobacco), or roadways from 10-20 m from the field margin (Figure 5-2).

5.3.2 Spatial Analysis

Within field spatial patterns of pests (aggregate and by species/family) and spiders were analyzed according to methods described in Chapter 4, page 54

(Windor et al. 2005, Thomson and Hoffmann 2013, Penn et al. In review) using

Spatial Analysis by Distance IndicEs (SADIE) (Perry 1995, Perry et al. 1999,

Winder et al. 2001, Perry and Dixon 2002, Perry et al. 2002) in SADIEShell 2.0 with QuickSADIE (Copyright © 2008, Kelvin F. Conrad). Aggregate and species/group specific pest and aggregate spider within field patterns were

analyzed in comparison to each other (Windor et al. 2005, Holland et al. 2007) using SADIE association analysis (N_AShell 1.0 with QuickAssociation 2.0,

(Golden Software, LLC, Golden, Colorado, U.S.A.) for all fields and collection dates. The presence of significant (P < 0.05) spatial patterns (both patches/aggregations and gaps) were coded as binary data (i.e. present or absent).

5.3.3 Statistical Analyses

To answer the target questions, several sets of regressions were run using JMP 12.0 (SAS Institute, Cary, North Carolina). First, the effect of presence or absence of field margin types at the two different spatial scales was analyzed using a GLM with Poisson distribution and log link function (O’Hara and Kotze

2010) in association with overall spider abundance, pest abundance, and species/group-specific pest abundance while controlling for abiotic conditions

(precipitation, average temperature, average daily wind speed, and field area).

The presence or absence of significant pest (both aggregate and species/group specific) and spider spatial patterns was regressed within a nominal logistic model against field margin classifications and distances as well as the previously stated abiotic factors and community abundances with posthoc odds ratio (OR) calculations (present to absent comparison in categorical variables; per unit increase in continuous variables). Finally, to determine what factors influenced

the spider spatial associations with pests, a final set of nominal logistic regressions with posthoc OR calculations was completed using the binary SADIE data for spiders, pests, each pest species/group, field margin composition and distance, community abundances, and the abiotic variables (Agerbo et al. 2007,

Rita and Komonen 2008). All regressions controlled for collection time and field of origin (random variable).

5.4 Results

5.4.1 Pest and Spider Abundances

The most commonly found insect pests (alphabetical order) included

Acrididae (Orthroptera), Aphididae (Hemiptera), Colaspis brunnea F.

(Coleoptera: Chrysomelidae), Cerotoma trifurcata (Forster) (Coleoptera:

Chrysomelidae), Dectes sayi Dillon and Dillon (Coleoptera: Cerambycidae),

Hypena scabra F. (Lepidoptera: Erebidae), Pentatomidae (Hemiptera), Popillia japonica Newman (Coleoptera: Scarabaeidae), and Spissistilus festinus (Say)

(Hemiptera: Membracidae) (Figure 5-3). The most commonly found spider families were Linyphiidae (Araneae), Lycosidae (Araneae), Oxyopidae

(Araneae), and Salticidae (Araneae).

Aggregate pest (with spiders: χ² = 115.41, df = 1, P < 0.01) and spider

(with pests: χ² = 85.05, df = 1, P < 0.01) abundances were positively correlated with each other, indicating potential habitat or resource preference overlap as these populations exhibited opposite interactions with average daily temperature

(pests: estimate = -0.05, χ² = 26.19, df = 1, P < 0.01; spiders: estimate = 0.06, χ²

= 46.58, df = 1, P < 0.01) and no interaction with precipitation (pests: χ² = 1.64, df

= 1, P = 0.28; spiders: χ² = 1.96, df = 1, P = 0.16). Average wind speed per day was a significant predictor of all abundances, with increased speed positively impacting abundance (pests: χ² = 62.95, df = 1, P < 0.01; spiders: χ² = 6.33, df =

1, P = 0.02). Field size altered pest and spider populations oppositely - larger fields decreased aggregate pest abundance (χ² = 24.44, df = 1, P < 0.01) while spider abundance increased (χ² = 8.39, df = 1, P < 0.01).

Margin types and distance had significant impacts on both aggregate and species/family pest abundance in addition to spider abundance (Table 5-1).

Generally, the closer the margin to the focal field, the greater number of pests were impacted (average number of 4 species/families per margin type impacted at 0-10 m and 2 species/families at 10-20 m) (Table 5-2). Soybean margins were the exception, where 0-10 m positively impacted abundances of the aggregate pests (χ² = 27.37, df = 1, P < 0.01) but negatively impacted them from 10-20 m

(χ² = 39.22, df = 1, P < 0.01). Some margin by pest interactions of note include wooded/shrub areas increasing Pentatomid abundance (0-10 m: χ² = 4.28, df =

1, P = 0.03) and weedy margins increasing P. japonica (0-10 m: χ² = 10.89, df =

1, P < 0.01) and Acridid abundances (0-10 m: χ² = 19.03, df = 1, P < 0.01). C. brunnea and C. trifurcata had no significant interactions with any margin at any distance. Spider abundances were positively impacted by 10-20 m wooded margins (χ² = 7.89, df = 1, P < 0.01) and not by any other type or distance.

5.4.2 Spatial Patterns of Spider and Pest Communities

Non-random spatial arrangements of pests were negatively impacted by an increase in pest abundance (OR = 0.88, P = 0.02) but positively by spider abundance (OR = 1.11, P = 0.03). The pest species/families C. brunnea (OR =

0.92, P = 0.03), Pentatomids (OR = 0.85, P = 0.02), and S. festinus (OR = 0.81,

P < 0.01) were negatively associated with increases in non-random spatial arrangements when soybean pest abundance increased. P. japonica (OR = 1.06,

P = 0.04) saw the opposite trend, and all other pests did not see any association.

Only the C. trifurcata spatial association (OR = 2.71, P < 0.01) was significantly related to the presence of aggregate pest patches within the field. The incidence of significant spider aggregations within the field exhibited significant influences on the likelihood of within field clustering of Acrididae (OR = 15.17, P < 0.01), C. brunnea (OR = 5.75, P < 0.01), H. scabra (OR = 0.17, P < 0.01), and P. japonica

(OR = 3.61, P < 0.01). In general, non-random spatial distributions of aggregated pests were altered by abiotic variables such as precipitation (OR = 0.99, P <

0.01), field area (OR = 0.92, P = 0.04), and wind speed (OR = 4.54, P < 0.01).

Spider spatial distributions were not impacted by aggregate soybean pest

(OR = 1.05, P = 0.13) or spider abundance (OR = 0.97, P = 0.45), but were impacted by the abiotic variables field area (OR = 0.95, P = 0.03) and wind speed (OR = 2.6, P < 0.01). Higher spider abundance was positively associated with Acrididae (OR = 1.10, P = 0.01) and S. festinus (OR = 1.16, P < 0.01) within field aggregations but were revealed to not be associated with any other pest species’ likelihood of spatial aggregation.

The aggregate pest response was variable in response to spatial scale, with the interaction with soybean margins changing from positive (OR = 1.28, P =

0.54) to negative (OR = 0.62, P = 0.22) and vice versa with grassy/weedy margins (0-10 m: OR = 0.68, P = 0.33; 10-20 m: OR = 1.85, P = 0.12) with the increase in distance (0-10 to 10-20 m). Additionally, the response to wooded borders was no longer significant at the greater distance (0-10 m: OR = 4.54, P <

0.01; 10-20 m: OR = 1.76, P = 0.21). Spider aggregations were less likely to occur when wooded areas from 0-10 m (OR = 0.51, P = 0.04) were present.

Potentially, these distance differences between aggregate spider and pest spatial aggregations would indicate the spatial distances they require to establish within a field and benefit from margin-contained resources. However, more clarity can be gained by evaluating the pest specific interactions with margins. We see that distance from the field is still an important factor (Figure 5-4).

5.4.3 Spatial and Temporal Overlap of Spiders and Pests

The spatial associations between aggregate pest and spider populations are more interesting. Influencing abiotic variables included precipitation (OR =

60.95, P < 0.01), average temperature (OR = 8.72e-12, P < 0.01), and field area

(OR = 3196.04, P < 0.01). There was an overwhelming influence of the pest spatial aggregations (OR = 42.79, P < 0.01) rather than the spider aggregations

(OR = 1.33e-10, P < 0.01), as the presence of pest aggregations significantly predicted more spider-pest overlaps (Figure 5-5). This trend was probably related to lower pest abundance as well since the number of significant positive

100

spatial associations between pests and spiders decreased throughout each field season (Figure 5-6). When specific pest groups were analyzed in relation to the spider by aggregate pest spatial associations, certain groups played a more significant role in this trend, further explaining the margin impacts on the spatial association. The pest by spider spatial association was primarily driven by the aggregations of Acridids (OR = 9.79e-9, P < 0.01), aphids (OR = 2.69e-8, P =

0.02), C. brunnea (OR = 6.44, P < 0.01), H. scabra (OR = 4.66e-8, P = 0.02), and

Pentatomids (OR = 6.8e-8, P = 0.02). This trend was clearer when we examined the pest species/family by spider spatial associations and how those impact the overall likelihood of aggregate pest by spider associations. Acridid × spider interactions (OR = 2.7e+10, P < 0.01) positively impacted this likelihood while H. scabra × spider interactions (OR = 3.82e-19, P < 0.01) negatively altered the association.

In terms of the spatial association between pest and spider populations, an indicator of potential predatory interactions, only one margin had an influence at the smaller scale while three influenced this relationship at the larger scale. At the 0-10 m scale, only wooded areas were associated with the spider to pest spatial aggregations and negatively so (OR = 0.13, P < 0.01), but the presence of wooded margins at 10-20 m was positively associated with a spatial overlap of these two groups (OR = 6.30, P < 0.01). Grassy/weedy margins at 10-20 m (OR

= 5.14, P < 0.01) had the opposite impact on this association than did the other crop (non-soybean) edges (OR = 2.01e-10, P = 0.04), where grassy areas exhibited a positive impact (Figure 5-7).

101

5.5 Discussion

5.5.1 Field Margins on Spider and Pest Abundances

The abundance data indicated that both spider and aggregate pest abundances were influenced by margin composition and scale. For instance, pests were impacted by more of the margin types (5 margins) than were spiders

(1 margin); but in both instances, scale dictated whether or not that margin stayed significant or changed the directional (positive or negative) influence on abundance. Spiders were only influenced by wooded margins at the further distance from the field, unsurprising as spiders have been known to balloon into agricultural fields from forested areas (Hatley and Macmahon 1980, Clough et al.

2005). Aggregate pest abundance was influenced by soybean margins at both scales, with the 0-10 m margins resulting in increased abundance while 10-20 m decreased abundance. This indicates that the pests were moving between fields when fields were immediately adjacent (0-10 m), forming a corridor, but that they are unable to overcome non-soybean margins if a soybean field was at the 10-20 m distance (Bhar and Fahrig 1998). The negative impact of wooded margins at both scales on pest abundances could potentially be due to the physical barrier posed by these margins (Wratten et al. 2003) or by the increase these same margins caused in spider abundance (Rand et al. 2006) The trend of the aggregate pest abundance was also mirrored by individual pest species/families in relation to the soybean margins at 0-10 and 10-20 m scales, indicating that this was a more general phenomena.

102

5.5.2 Field Margins on Spider and Pest Spatial Aggregations

The SADIE data for spider or pest aggregations exhibited similar trends with the abundance data; margin type and scale had differential impacts on spiders and aggregate pest aggregation likelihoods. Interestingly, margin type interactions with spider SADIEs were more similar to the margin interactions with prey abundance instead of spider abundance, potentially indicating that spiders clustered in space and time based on pest abundance in any given area. This was supported by spider SADIE likelihood as it was increased by the same margin type that increased aggregate pest abundance. Pest SADIE likelihood was only increased by wooded (0-10 m) margins and appeared to be driven by

Pentatomid abundance and interactions with wooded margins. Interestingly, pest

SADIE likelihood was diminished with an increase in aggregate pest abundance, indicating that smaller and earlier populations of pests were more likely to be significantly clustered than larger older populations. We also saw that abundance and within field, aggregation is not necessarily positively associated as we might imagine. In the case of spiders, there was no association between abundance and within field aggregations. However, with pests, increases in abundances actually decreased the likelihood of pest aggregations within the field, indicating that competition and subsequent dispersal could have been occurring (Denno and Peterson 1995, Applebaum and Heifetz 1999). This was especially important to consider given the spider × pest overlap, where this interaction is dependent upon the presence of pest aggregations. Potential biological control services provided by spiders when these spatial overlaps occur (Windor et al. 2005) are

103

more likely to happen when pest populations are low as is the case in early season predation (Athey et al. 2016).

5.5.3 Field Margins on Spider and Pest Spatial Overlap

Spider by pest SADIE association likelihood was only influenced by two border types – wooded margins and grassy/weedy margins. We found that the wooded margins closer to the field had a negative influence on the aggregation whereas the wooded areas further from the field positively influenced the association of spider to pest aggregations. This result re-emphasizes that pest communities appear to be altered more so by immediately adjacent edges and that this trend cascades into the spider-pest spatial interactions. Furthermore, the greater likelihood of pest aggregations increased the likelihood of spider to pest aggregations, indicating that early season predation (where early is defined by when any given pest enters the field) is at play in these associations since pest aggregations were negatively impacted by increases in pest abundance (Athey et al. 2016).

5.5.4 Abiotic Conditions on Spider and Pest Spatial Patterns

In addition to the impact that abundance and field margin type and distance had on spider and pest communities and spatial associations, abiotic factors contributed greatly. Spiders and pests had opposite interactions with average temperature and field area, where an increase in either increased spider abundance but not pest abundance. Interestingly, neither spider nor pest SADIEs

104

was altered by average temperature and both were negatively impacted by field area. The temperature data again revealed that early-season populations of pests were the force behind the spatial associations since pest populations generally increased with temperature and time. Field area interactions with

SADIE data also mirrored the strong pest interactions with 0-10 m margins; if a field was smaller, there was a greater chance that any given pest will encounter a non-soybean field margin. Wind speed was the only abiotic factor to increase both spider and pest abundances and spatial associations both within their respective communities and between the communities. As many spiders and pest species can disperse aerially, increased wind speed might spread these species across an agricultural landscape quickly, making source margins (in the case of spiders) and barrier margins (in the case of pests) important for controlling arthropod abundances and the subsequent within field spatial interactions of pests and predators.

5.5.5 Application of Field Margin by SADIE Interactions

Our results are similar to previous findings in agroecosystems – where wooded edge habitats were beneficial to natural enemy abundance within fields

(Altieri and Schmidt 1986, Alderweireldt 1989, Phillips and Gardiner 2016).

Additionally, these wooded field margins are extremely important for conservation biological control for generalist such as spiders as these areas often house species and resources that other margin types such as grassy areas do no provide (Bedford and Usher 1994). Although such margins are often

105

suspected by growers to increase pest species (Burton et al. 2008), our results indicated that woody margins actually decrease pest abundance either through physical barriers or by encouraging natural enemies such as spiders to abound

(Kremen et al. 2011, Zhao et al. 2013, Rusch et al. 2016), ultimately benefiting crop production (Losey and Vaughan 2006, Cullen et al. 2008). These interactions depended, in part on pest species identity as various pest groups prefer different host plants, disperse via different mechanisms, and respond to abiotic conditions differently (Blau and Stinner 1983, Stinner et al. 1983, Zaller et al. 2008).

Our results indicated that more intentional management of field margins can benefit growers and conservation efforts (Donald and Evans 2006, Marshall et al. 2006, Gaucherel et al. 2007, Dramstad and Fjellstad 2011, Caro et al.

2016). Altering the plant community within agroecosystems would benefit the arthropod community including natural enemies of pest species by providing a larger number and greater diversity of ecological niches (Barberi et al. 2010) while simultaneously buffering the population of the area by invasive insects

(Jonsson et al. 2010). An increase in ecosystem diversity fueled by vegetative margins would, in turn, promote bird and pollinator population growth and diversity (Henderson et al. 2012, Cusser et al. 2016). Therefore, the use of vegetative field margins has the potential to benefit agroecosystems and grower interests in both the short and long term. We suggest that tree lines within soybean systems be maintained as a form of conservation biological control and

106

to fuel the potential for early season predation of pests, a sustainable form of crop damage mitigation.

107

Table 5-1 Parameter estimates of margin types (multinomial logistic regression) with aggregated pest and spider abundances as the corresponding independent variables controlled for abiotic conditions and field site. Distance Margin Type Pests (all) Spiders

0-10 m Grassy 0.18* -0.06

Soybean 0.22* -0.06

Tree -0.12* 0.06

11-20 m Grassy 0.07 0.01

108 Soybean -0.24* 0.04

Tree -0.12* 0.13*

Other Crop -0.10* 0.05

Roadway 0.10* 0.05

* indicates P < 0.05

Pest Grass Soybean Tree Grass Soybean Tree Other Crop Roadway

Acrididae 0.66* 0.54* -0.25 -0.25 -0.63* -0.70 -0.01 0.07

Aphididae -0.57* -0.21 -0.41* 0.20 0.04 -0.20 -0.38 0.00

C. brunnea -0.05 -0.06 -0.25 0.05 0.15 -0.05 0.07 0.16

109 C. trifurcata 0.10 -0.10 -0.20 0.10 0.03 0.24 -0.08 -0.09

D. sayi -0.24 -0.03 -0.46 0.53 0.11 0.46 0.96* 0.19

H. scabra -0.04 0.08 -0.14 0.17 -0.13 0.24* 0.01 0.27*

M. unipuncta -0.67 8.05 -0.45 1.01* -7.98 1.00* -7.82 0.40

Pentatomidae 0.83* 0.23* 0.15 -0.68* -0.14 -0.97* 0.36* 0.82*

P. japonica 0.36* 0.32* -0.23* -0.03 -0.31* 0.16 -0.11 0.16

S. festinus 0.47* 0.54* 0.14 -0.11 -0.45* -0.24 -0.36 -0.08

* indicates P < 0.05

110

Figure 5-1 Potential field margin distance by within field species interactions of soybean pests and predators.

111

Figure 5-2 Field border examples of (A) other crop, grass/weed, and tree/shrub and (B) soybean, tree/shrub, and road.

112

Figure 5-3 (A) The relative proportion of spider families represented within the samples (n = 1282) and (B) the relative proportion of the pest groups (some families, some species) found.

113

Figure 5-4 Parameter estimates of aggregate spider and pest within field spatial aggregations given field margin type for (A) 0-10 m and (B) 10-20 m.

114

Figure 5-5 Odds ratios for spider and pest community variables in how they contribute to the likelihood of seeing a positive spatial association between spiders and pests. Where spider and pest abundances are per unit odds ratios and SADIE variables are categorical with presence: absence. Any value < 1 indicates a decrease in likelihood while and value > 1 indicates an increase in likelihood.

115

Figure 5-6 The number of fields containing significantly positive spatial associations between aggregate pest and spider communities over the growing season.

116

Figure 5-7 SADIE association map indicating general spider by pest association in relation to margin type at 0-10 m. Aggregate pests are represented in blue and spiders in red. Associated patches are purple and gaps are white.

Chapter 6 NON-CONSUMPTIVE EFFECTS OF ANTS AND SPIDERS INFLUENCE INTRAGUILD INTERACTIONS AND SUBSEQUENT PEST DAMAGE IN SOYBEANS

Chapter contents prepared for publication in Penn, HJ and KJ Athey. Non- consumptive effects of ants and spiders influence intraguild interactions and subsequent pest damage in soybeans. Bulletin of Entomological Research.

6.1 Summary

Biological control practices rely on natural enemies, including generalists, to suppress pest populations even though these natural enemies may consume a wide range prey including other predators, known as intraguild predation. We examined interactions between two generalist predators in soybean agroecosystems in central Kentucky – the striped lynx spider (Oxyopes salticus) and the cornfield ant (Lasius neoniger) – and a herbivore, the green cloverworm

(Hypena scabra). We hypothesized that each predator would reduce green cloverworm populations and leaf damage, but the predators would interfere with each other when both were present. To study these interactions, field cages containing potted soybeans were used to examine eight treatments (control, cloverworm, spider, ant, cloverworm + spider, cloverworm + ant, spider + ant, and cloverworm + spider + ant). When the survival of spiders and cloverworms was analyzed, no treatment effects were found. However, when the proportion of leaf damage was compared, the spider, spider + ant, and spider + cloverworm treatments had significantly less damage than the ant, ant + cloverworm, and ant

+ cloverworm + spider treatments. These results indicated that spiders tended to decrease plant damage while ant presence significantly increased damage. We found that generalist predators, such as L. neoniger, can inhibit biological control

117

due to non-consumptive and intraguild interactions even when the same species in a different system provides substantial levels of pest suppression.

6.2 Introduction

Biological control to prevent plant damage and disease transmission depends on natural enemies found in agroecosystems to actively prey upon and suppress pest populations (Landis et al. 2000, Welch and Harwood 2014). The characteristics and disturbances of agricultural habitats influence various natural enemy species differently, allowing for the alteration of natural enemy community composition and abundance in cropping systems (Schmidt et al. 2005, Gardiner et al. 2010, Jonsson et al. 2015). By enhancing the populations of already occurring natural enemies through conservation biological control, pest suppression can be intensified via non-chemical means, benefiting non-target species (i.e. pollinators, detritivores, and predators) through reductions of insecticide inputs (Landis et al. 2000). Additionally, increased species richness, diversity, and evenness of natural enemies within an agroecosystem can ultimately lead to greater biological control effectiveness (Losey and Denno

1998, Snyder et al. 2006, Straub and Snyder 2008, Finke and Snyder 2010). For instance, the ant species Azteca sericeasur Longino (Hymenoptera: Formicidae) within coffee agroecosystems (Marín et al. 2015) and Lasius niger L.

(Hymenoptera: Formicidae) in grasslands (Schuch et al. 2008) have been found to positively influence spider populations, thereby increasing overall predation levels.

118

The additive or multiplicative effects of multiple generalist predator species on biological control is confounded by intraguild interactions, competition or predation among the predators themselves (Polis et al. 1989, Rosenheim et al.

1995, Snyder and Wise 1999). Predators have the potential to actively interfere with each other, releasing pest species from predation (Prasad and Snyder

2006); this interaction hinges on the identity of and the environmental conditions surrounding the predatory species (Straub and Snyder 2006). The presence of two predators with the tendency to interfere with each other could increase crop damage by either inducing behavioral changes or directly attacking each other.

For instance, a predator could consume another in place of the target pest species (Traugott et al. 2012, Messelink et al. 2013) or an aggressive predator could interfere with the normal behaviors of other predators, preventing the threatened individual from consuming the pest (Eubanks et al. 2002).

Ants and spiders, both abundant generalist predators in many systems including agroecosystems, exhibit many types of intraguild interactions including bi-directional predation and non-consumptive behavioral shifts. Over 100 species of spiders have been found to consume ants regularly (Edwards et al. 1974,

Cushing 2012). For instance, Callilepis nocturna (L.) (Araneae: Gnaphosidae) facultatively feed on Formica spp. and Lasius spp. (Hymenoptera: Formicidae).

Another documented effect of spiders on ants was reported by Gastreich (1999).

Pheidole bicornis Forel (Hymenoptera: Formicidae) did not forage on leaves where the silk of Dipoena banksi Chickering (Araneae: Theridiidae) was present, indicating potential predation risk to the ant. Alternatively, many ant species

119

affect the behavior and survival of spiders; for example, Formica cunicularia

(Latreille) (Hymenoptera: Formicidae) alters densities of Linyphiid (Araneae) spiders - where ants were excluded, spider density increased three-fold.

Additionally, treatments where ants were not excluded saw reduced populations of Lepidopteran larvae, indicating that ants can disrupt spider predation while maintaining high predation rates on herbivorous prey of their own (Sanders and

Platner 2007).

Little is known about the intraguild interactions between ants and spiders within simplified and highly disturbed habitats even though both ants and spiders have been shown to be excellent predators. The cornfield ant, Laius neoniger L.

(Hymenoptera: Formicidae), preys upon black cutworm larvae, Agrotis ipsilon

(Hufnagel) (Lepidoptera: Noctuidae) and Japanese beetle eggs, Popilla japonica

Newman (Coleoptera: Scarabaeidae) in a turfgrass system (López and Potter

2000). While the striped lynx spider, Oxyopes salticus (Hentz) (Araneae:

Oxyopidae), has been well documented as a predator of several pest species in cotton agroecosystems (Nyffeler et al. 1987, Breene et al. 1990, Nyffeler et al.

1992, Nyffeler and Sunderland 2003). However, the addition of both predators within the same area, as is common in agricultural field settings, could exhibit any combination of intraguild interactions.

The objective of this study was to determine the intraguild interactions between ants and spiders within the context of soybean production, specifically, the influence of cornfield ants (Laius neoniger) and striped lynx spiders (Oxyopes salticus) on a soybean pest, green cloverworm (Hypena scabra Fabricius

120

(Lepidoptera: Erebidae)). The plant damage produced by the green cloverworms was analyzed in addition to the survival of cloverworms and spiders to assess the nature of ant-spider intraguild interactions. Within this system, we predicted that each predator would reduce green cloverworm survival and subsequent leaf damage via direct predation (i.e. loss of cloverworm individuals and predation by spiders shown via molecular gut content analysis); but when the predators were present simultaneously, they would interfere with each other, resulting in increased cloverworm survival and leaf damage.

6.3 Materials and Methods

6.3.1 Experimental Field Cage Set-up

The experiment was conducted at the University of Kentucky Spindletop

Research Farm in Lexington, Kentucky, USA (38° 07' 31.7" N 84° 30' 55.4" W).

Field cages were composed of nylon mesh screening (52 × 52 mesh count) fine enough to exclude arthropod entry or exit but permit sunlight and rainfall. The cages (Lumite Inc, Alto, California, USA) measured 1.83 m × 1.83 m × 1.83 m, were secured to the ground with tent stakes and permitted researcher access via a side zipper. All cages were buried up to 20 cm to prevent movement of arthropods in the top layer of soil. Removal of pre-existing arthropods within the cages was completed using a leaf blower (Poulan Pro 25cc Gas Blower/Vac,

Poulan, Charlotte, North Carolina, USA) set to reverse with an insect net attached to the air intake. All arthropods captured in this way were placed outside of the cages. Yellow sticky cards (15 × 15 cm, 1 per cage) were deployed from

121

the top of the field cage both the week before and during the study to capture the remaining non-ground dwelling arthropods. All cages were placed in a grassy field which had not been in crop production for one year. Plant material within the cages was killed via two sequential applications of Killzall (Voluntary Purchasing

Groups, Inc., Bonham, Texas, USA) according to the label recommendations, with all plant material removed manually one week before the study. No other chemical application was made after the final herbicide treatment. Soybeans

(Viking 2265 Organic Soybeans, Johnny’s Selected Seed, Winslow, Maine,

USA), were sown at a rate of 1 seed / cm within plastic planters (15.24 cm ×

60.33 cm × 20.00 cm) in the greenhouse at 25 ± 1 °C, 16: 8 L: D cycle and 65 ±

5% RH until the plants reached a growth stage of R1.

6.3.2 Experimental Treatments

Eight treatments (factorial arrangement) were used to measure the interactions of cornfield ants, striped lynx spiders, and the pest species, green cloverworms, with four replicates of each treatment established. Cages were placed based on the observed presence of ant colonies with one colony per cage. To monitor that the ants remained in the cages after removal of other arthropods and plant material, pitfall traps (diameter 9.5 cm, depth 12 cm) with

Styrofoam rain guards (diameter of 22 cm) containing ethylene glycol were installed in the center of each cage. Pitfall traps were set the week before the cage study as well as the time of the cage study. Two randomly selected planters of greenhouse raised soybeans were placed into each cage three days before

122

the start of the experiment. Planters within treatments not containing ants were painted with Fluon® (INSECT-a-SLIP, BioQuip Product, Inc., Rancho

Dominguez, California, USA) to exclude ants from the soybeans. Spiders and cloverworms were obtained via sweep nets from fields of grass and alfalfa at the

Spindletop Research Farm and immediately placed into field cages to mimic arthropod hunger during field conditions (Breene et al. 1990). Organisms were added to the cages at a rate of 4 adult or sub-adult spiders and 15 second instar cloverworms. The presence of ants (observation of ≥ 1 L. neoniger individual) was also observed via direct observations and pitfall trap data. The organisms were allowed to interact with the soybean plants and each other for four days in order to allow accurate assessment of spider gut contents.

6.3.3 Arthropod Observations

Daily observations of sex/stages of spiders, and counts of the ants, spiders, cloverworm larvae, and other arthropods within the cages were made by one researcher. Non-cloverworm herbivores causing chewing damage on the plants were removed from the cage when found. The within-cage locations (i.e. top of leaf, bottom of leaf, stem, ground, or cage wall/ceiling) of the focal species were recorded. At the end of the exposure time, cloverworms and spiders (both live and dead) were recovered, counted, and stored in 95% ethanol.

Cloverworms found as precocious pupae due to parasitism were placed in plastic condiment containers for observation of parasitoids, which were stored in 95% ethanol upon emergence. Ant presence (≥ 1 individual) or absence during the

123

experimental period was determined using direct observations and pitfall trap data.

6.3.4 Molecular Gut Content Analysis

Following collection, spider samples were homogenized in 180 µL of tissue lysis buffer. Total DNA was extracted from all samples using Qiagen

DNeasy Blood and Tissue Kits© (Qiagen Inc., Valencia, California, U.S.A.) following the animal tissue protocol. DNA was then stored in clean 1.5 mL microcentrifuge tubes at -20°C until PCR analysis. The total DNA in all samples was first amplified with the H. scabra primers (HS517F and HS598R) (Penn et al.

In review). All PCRs (12.5 µL total volume) consisted of 1X Takara buffer (Takara

Bio Inc. Shiga, Japan), 0.2 mM of each dNTP, 0.2 mM of each primer, 0.31 U

Takara Ex TaqTM and template DNA (1 µL of total DNA) (Penn et al. 2016). To determine the viability of the extractions not testing positive for H. scabra DNA, samples were screened using general COI primers Jerry (Simon et al. 1994) and

Ben3R (Villesen et al. 2004b). All reactions were carried out using Bio-Rad PTC-

200 and C1000 thermal cyclers (Bio-Rad Laboratories, Hercules, California,

U.S.A.). The PCR cycling protocol for the general COI primers Jerry and Ben3R

(with Takara reagents as above) was 94 °C for 1 minute followed by 35 cycles of

95 °C for 60 s, 47 °C for 60 s, and 72 °C for 90 s. The PCR cycling protocol for the H. scabra primers (with Takara reagents as above) was 94 °C for 1 minute followed by 40 cycles of 94 °C for 45 s, 62 °C for 45 s, and 72 °C for 30 s.

Following amplification, reaction success was determined by electrophoresis of 5

124

µL PCR product pre-stained with GelRed nucleic acid gel stain (1X Biotium,

Hayward, California, USA) on 2% SeaKem agarose (Lonza, Rockland, Maine,

USA). In all cases, sets of PCRs contained one positive control of only H. scabra

DNA and one negative control of all reagents without the addition of sample

DNA. Additionally, any reactions not testing positive with the general COI primers were assumed to be unreliable and were not factored into the results.

6.3.5 Final Plant Damage Analysis

Following the removal of cloverworms and spiders, all plant stems were snipped at soil level, placed into bags (one per cage), and transported to the lab where plants were stored in a cold room (15 °C) for twelve hours (Breene et al.

1990). A five plant sub-sample was blindly and haphazardly selected out of the bag by hand for every cage, the leaves of which were removed from the stem, placed onto a white background with a scale and photographed using a Canon

EOS digital camera (Canon Inc., Tokyo, Japan). Plant damage was assessed via

ImageJ (National Institutes of Health, Bethesda, Maryland, U.S.A.(Rasband

2016)), where the proportion missing leaf area was measured according to published methods (Schneider et al. 2012, Schindelin et al. 2015).

6.3.6 Statistical Analyses

The location within the cage of observed spiders in relation to treatment

(only those containing spiders), day (to account for weather changes), and sex/stage was analyzed using a multinomial logistic regression with replicate as

125

a random effect. Not enough data was gathered on ant and cloverworm locations to allow for statistical analysis. Survival (proportion of each recovered alive at the end of the study) of spiders and survival and parasitism cloverworms was compared among treatments including those respective organisms using a Tukey

(HSD) test after being arc sine square root transformed. The proportion of leaf area damaged was compared among all treatments using a post-hoc Tukey

(HSD) test after being regressed against presence/absence of non-targets and parasitism rate with replicate being a random effect. Additionally, the main effects of the treatments (ant, cloverworm, and spider presence/absence) were analyzed in all cases using an ANOVA. All analyses were completed in JMP 12.0 (SAS

Institute, Cary, North Carolina, USA).

6.4 Results

6.4.1Arthropod Observations

All cages contained non-target arthropods that had emerged during the experiment or were not removed during cage preparation, the most common of which included Acrididae (Orthoptera), Cicadellidae (Hemiptera), and Colaspis brunnea F. (Coleoptera: Chrysomelidae). Ants were found via visual observations and pitfall traps in all treatments the week prior to, and the week of in ant-inclusion treatments. Additionally, Fluon® application appeared to be effective in preventing ant access to the planters of soybeans throughout the duration of this study. When we analyzed the locations of the spiders in relation to treatment, day, and spider sex/stage we found the model was generally

126

significant (χ² = 59.12, df = 24, P < 0.01). However, neither day (χ² = 0.37, df = 1,

P = 0.54) nor spider sex/stage (χ² = 0.09, df = 3, P = 0.76) were significant variables for explaining spider location. The treatment variable overall was found to be significant (χ² = 36.65, df = 8, P < 0.01) (Figure 6-1). The spider-only treatment was found to increase the number of observations made of spiders on the cage structure (χ² = 10.18, df = 8, P < 0.01). Additionally, treatments containing spider + cloverworm and spider + cloverworm + ant, were found to increase the number of observations of spiders on the bottoms of leaves (χ² =

5.10, df = 4, P = 0.02; χ² = 4.17, df = 4, P = 0.04; respectively), potentially due to prey (i.e. cloverworm) presence. Treatments containing any combination of spiders and ants were not found to influence spider location. According to the main effects ANOVA (Table 6-1), both ant and cloverworm presence increased the likelihood that spiders would be on the underside of leaves (t = 2.23, df = 1, P

= 0.03; t = 4.01, df = 1, P < 0.01; respectively) and decreased the likelihood of spider presence on the cage (t = -3.64, df = 1, P < 0.01; t = -3.55, df = 1, P <

0.01; respectively). However, there was an interaction between ant and cloverworm presence on spider in the main effect analysis, with an increased likelihood of spiders being seen on the cage instead of the plant when ants were present in addition to cloverworms (t = 2.81, df = 1, P < 0.01). Daily observations of cloverworms provided insufficient data for location analyses as cloverworms were difficult to observe without disturbing the system. However, based on the observations obtained during the course of this experiment cloverworms were generally located on the bottom of the soybean leaves when found.

127

6.4.2 Recovery of Cloverworms and Spiders

No significant differences among treatments were found for the proportion recovered of any of the study organisms. When the final proportions of recovered cloverworms were compared among treatments containing cloverworms, no differences were found between treatments (F(3, 12) = 1.29, P = 0.32) (Figure 6-2) or in the main effects ANOVA (overall F(4, 15) = 0.59, P = 0.67). The same result was found for recovered spiders generally as compared between treatments (F(3,

12) = 0.37, P = 0.77) and main effects (overall F(4, 15) = 0.59, P = 0.67), nor were there any differences among treatments based on the sex (F(3, 11) = 0.23, P =

0.87) or stage (F(2, 2) = 0.03, P = 0.96) of the spider. No statistical differences were found in parasitism rates between treatments (F(3, 12) = 1.89, P = 0.19), but were found to be highly correlated with cloverworm presence in the main effects

ANOVA (t = 9.25, df = 1, P < 0.01) but not by any of the predators or associated interaction effects.

6.4.3 Molecular Gut Content Analysis of Spiders

Of all spiders placed into field cages (n = 64), 39 were recovered on day four and analyzed for H. scabra DNA presence. Of these samples, 11 tested positive when H. scabra was present in the cage with O. salticus, but 15 tested positive when H. scabra was not intentionally placed in the field cages. Upon testing with COI primers, the samples (n= 35) were found to have DNA present, indicating that the samples were sound, but that the molecular gut content analyses of this study were inconclusive. Therefore, it remains unknown whether

128

or not O. salticus was preying upon H. scabra rather than harassing them.

However, given the final recovery number of H. scabra, it appears that even if predation was occurring by O. salticus, it was not significant.

6.4.4 Final Plant Damage

The proportion of plant damage was compared among treatments (Figure

6-3), which yielded significant variation between treatments using a Tukey HSD test (F(7, 152) = 5.92, P < 0.01), but parasitism rate (t = 2.03, df = 1, P = 0.06) and non-cloverworm herbivore presence (Acrididae: χ² = 1.16, df = 1, P = 0.28; C. brunnea: χ² = 0.0053, df = 1, P = 0.94) played no part in this effect. In Tukey

HSD tests, only the treatment with ant + spider + cloverworm was significantly different than the soybean only control (P < 0.01). The spider + ant + cloverworm and ant + cloverworm treatments had significantly more damage than the spider- only (P < 0.01; P = 0.02; respectively) and spider + ant (P < 0.01; P = 0.04; respectively) treatments. There was a trend of less plant damage when spiders alone were present with cloverworms than when ants alone were present with cloverworms (P = 0.08). However, when both spiders and ants were present with cloverworms, significantly more damage occurred than when cloverworms were present without the addition of any predators (P < 0.01). When the main effect

ANOVA was analyzed (Table 6-2), similar trends were seen where cloverworm presence increased damage (t = 2.29, df = 1, P = 0.03) and this was exacerbated by an ant × cloverworm interaction (t = 2.44, df = 1, P < 0.02). However, the

129

presence of C. brunnea also increased plant damage in the main effects test (t =

2.32, df = 1, P = 0.03).

6.5 Discussion

Given that few ants were observed actively foraging on the plants within non-exclusion treatments, it might be surprising that a strong ant effect was shown in the leaf damage data. However, the transient nature of ant scouts could mean that more ants would have been observed had a greater length of time been deployed (Wenninger et al. 2016). Alternatively, Buckley (1990) has shown some species of tropical ants are more apt to protect herbivores during nocturnal predation events which were not observed directly in this study. The survival of spiders and cloverworms was not influenced by ant presence, indicating that this interaction was not predacious but a non-consumptive interaction that altered spider and cloverworm behaviors in the presence of ants (Schmitz et al. 1997,

Mestre et al. 2016). This is consistent with a previous study where another territorial ant, Solenopsis invicta Buren, failed to reduce either predator or pest abundance in cotton fields through predation (Sterling et al. 1979). Ants have also been shown to benefit pest species even if said species does not intentionally solicit ant protection via honeydew production (James et al. 1997).

Non-consumptive interactions between predators have been frequently observed; for example, Coccinella septempunctata L. (Coleoptera: Coccinellidae) are repelled from aphid consumption by the presence of Tetramorium caespitum

L. and Lasius niger L. (Hymenoptera: Formicidae) via active harassment but not

130

predation (Katayama and Suzuki 2003). Spiders could have also been sensing chemical cues from the ants present and arresting normal predation behaviors for fear of ant antagonism (Clark et al. 2000). Furthermore, ant harassment of spiders in this experiment could be due in part to the relatively simple system presented to them as this same effect has been observed in laboratory studies examining the intraguild interactions of mirids (Tytthus vagus (Knight)

(Hemiptera: Miridae)) and wolf spiders (Pardosa littoralis Banks (Araneae:

Lycosidae)) (Finke and Denno 2002).

The fact that cloverworm survival did not decrease in spider treatments

(without ants present) indicates that the spiders could have also induced non- consumptive behavioral shifts in cloverworms (Whitehouse et al. 2011, Rypstra and Buddle 2013). The consumption habits of cloverworms could diminish simply due to spider presence (the spiders were present and observed predating other small arthropods (Hemiptera: Ciccadellidae)), resulting in less leaf damage

(Thaler and Griffin 2008). The non-consumptive effects were further substantiated by the spider location trends when the treatments included cloverworms – spiders were often found where cloverworms were likely to be present (i.e. undersides of leaves), indicating that the spiders were near cloverworms during the course of the study. Similar results to ours have been seen in other systems such as damsel bug effects on aphid populations despite prevention of predation (Nelson et al. 2004). The influential non-consumptive effects are still an important consideration for biological control using generalist predators (Ohgushi 2008) as even in the absence of predation on a pest species,

131

the presence of a predator such as our spider could alter food web interactions of the pest (Kéfi et al. 2012, Eubanks and Finke 2014, Majdi et al. 2014), cascading through the system to the benefit of crop plants (Preisser et al. 2005, Preisser and Bolnick 2008).

Our results indicate that intraguild interactions are indeed occurring between spiders and ants within the soybean system as we predicted; when both predators are present simultaneously, plant damage by the pest increases.

However, the interactions between predators and between predators and pest were not mediated by direct predation as we had supposed but via non- consumptive interactions between the predators and between both predators and the pest species. We conclude this based on decreased plant damage despite no differences in spider or cloverworm survival. Furthermore, the consumptive and non-consumptive interactions between predators and pests can change based on the environmental conditions as exhibited by our ant species that inhibited biological control services in this study even though the same species in turfgrass provides substantial levels of pest suppression.

132

Table 6-1 Main effect test (ANOVA) for presence of ants (Ln) and cloverworms (HS) on spider location. Source Spider Location df t-ratio P

Ln Leaf bottom 1 2.23 0.03

Ln Cage 1 -3.64 <0.01

Ln Ground 1 0.79 0.43

Ln Stem 1 1.06 0.29

Hs Leaf bottom 1 4.01 <0.01

Hs Cage 1 -3.55 <0.01

Hs Ground 1 -0.82 0.41

Hs Stem 1 0.72 0.47

Ln*Hs Leaf bottom 1 0.62 0.54

Ln*Hs Cage 1 2.81 <0.01

Ln*Hs Ground 1 -0.40 0.69

Ln*Hs Stem 1 -0.55 0.58

133

Table 6-2 Main effect test (ANOVA) for presence of ants (Ln), spiders (Os), and cloverworms (Hs) on the proportion of leaf damage. Source df t-ratio P

Ln 1 3.28 <0.01

Os 1 -1.91 0.07

Hs 1 2.64 0.01

Ln*Os 1 -0.44 0.66

Ln*Hs 1 2.08 0.05

Os*Hs 1 0.81 0.43

Ln*Os*Hs 1 -0.17 0.86

134

135

Figure 6-1 The total number of striped lynx spiders, Oxyopes salticus (Hentz), observed at each location (legend) by treatment containing spiders over four days. Os represents Oxyopes salticus, Ln Lasius neoniger, and Hs Hypena scabra.

136

Figure 6-2 (A) The proportion of the initial fifteen second instar green cloverworms including pupa, Hypena scabra Fabricius (Hs), and (B) lynx spiders, Oxyopes salticus (Hentz) (Os), per cage recovered. Ln indicates the presence of the cornfield ant, Lasius neoniger L. Differences between treatments were determined by the Tukey (HSD) test.

137

Figure 6-3 The proportion of leaf area damaged after four days. Ln indicates the presence of cornfield ants, Lasius neoniger L.; Os indicates the presence of striped lynx spiders, Oxyopes salticus (Hentz), and Hs indicates the presence of green cloverworms, Hypena scabra Fabricius. Lowercase letters above standard error indicate significant differences between treatments as denoted by the Tukey (HSD) test. Copyright © Hannah Joy Penn 2016

Chapter 7 IMIDACLOPRID SEED TREATMENTS AFFECT INDIVIDUAL ANT BEHAVIOR AND COMMUNITY STRUCTURE BUT NOT EGG PREDATION, PEST ABUNDANCE, OR SOYBEAN YIELD

Chapter contents submitted for publication and under review in Penn, HJ and AM Dale. 2016. Imidacloprid seed treatments affect individual ant behavior and community structure but not egg predation, pest abundance, or soybean yield. Pest Management Science.

7.1 Summary

Neonicotinoid seed treatments are under scrutiny because of their variable efficacy against crop pests and for their potential negative impacts on non-target organisms. Ants provide important biocontrol services in agroecosystems and can be indicators of ecosystem health. This study tested for effects of exposure to imidacloprid plus fungicide or fungicide treated seeds on individual ant survival, locomotion and foraging capabilities, and on field ant community structure, pest abundance, ant predation, and yield. Cohorts of ants exposed to either type of treated seeds had impaired locomotion and a higher incidence of morbidity and mortality but no loss of foraging capacity. In the field, we saw no difference in ant species richness regardless of seed treatment. Blocks with imidacloprid did have higher species evenness and diversity, probably due to variable effects of the insecticide on different ant species, particularly

Tetramorium caespitum. Ant predation on sentinel eggs, pest abundance, and soybean growth, and yield were similar in both treatments. Both seed treatments had lethal and sublethal effects on ant individuals, and the influence of imidacloprid seed coating in the field was manifested in altered ant community composition. Those effects, however, were not strong enough to affect egg predation, pest abundance, or soybean yield in field blocks.

138

7.2 Introduction

Increased use of neonicotinoid insecticides for agricultural pest protection has led to concerns about their potential effects on non-target species (Bortolotti et al. 2003, Bonmatin et al. 2005, Blacquiere et al. 2012, Douglas and Tooker

2015, Rundlof et al. 2015, Thompson et al. 2015). One of the most contentious uses of neonicotinoids is via seed treatments because of their widespread use and potential for residues to be translocated into nectar and pollen and for dust particles from treated seed to drift during planting, both of which could harm pollinators, natural enemies, and other non-target species (Blacquiere et al.

2012, Gibbons et al. 2015, Li et al. 2015, Pecenka and Lundgren 2015, Rundlof et al. 2015, Zhu et al. 2015). Efficacy of neonicotinoid seed treatments has also been questioned. One EPA report concluded that neonicotinoid seed treatments, used on 31% of the soybeans grown in the United States, provide little to no benefit as far as soybean pest control (Myers and Hill 2014), but that conclusion may not apply to all soybean-growing regions within the US. For example, neonicotinoid seed treatments were determined to be economically viable for pest control in mid-south soybeans (North et al. 2016). Kentucky is at the border of the mid-south and mid-west regions and contains soybean pests similar to both regions. Therefore, it is unknown whether or not either of these reports can be relied upon for soybean pest management within Kentucky especially since current seed treatment recommendations in the area do not pertain to neonicotinoids (Johnson 2016).

139

Although sprayed neonicotinoids can be an immediate hazard to non- target arthropods (Stanley and Raine 2016), persistence of residues has the potential to impact non-targets indirectly for several months or longer (Sánchez-

Bayo et al. 2007, Zhang et al. 2015). Neonicotinoids from seed treatments have the potential to permeate the surrounding environment (i.e. soils, water bodies, non-target plants) of a crop field (Krupke et al. 2012, Schaafsma et al. 2016); they have been documented in dust (Pistorius et al. 2015), plant guttation

(Girolami et al. 2009), and surface waters (Huseth and Groves 2014, Main et al.

2014). The longevity of neonicotinoids varies by active ingredient, as well as by abiotic conditions, with imidacloprid being one of the longest lasting within the environment (Liu et al. 2011, Limay-Rios et al. 2016).

Environmental infiltration of neonicotinoids has the potential to impact natural enemies and pollinators while disrupting their services to agroecosystems

(Goulson 2013, Godfray et al. 2014, Chagnon et al. 2015). Ants (Hymenoptera:

Formicidae) are common predators of pests and weed seeds in crop fields and other disturbed habitats, and they also are bioindicators of ecosystem health

(Matlock and de la Cruz 2003, Cantagalli et al. 2014). As many temperate ant species are ground-nesting, there is the potential for neonicotinoid soil infiltration to negatively affect them through ingestion or environmental exposure (Rondeau et al. 2014). Neonicotinoids can be very effective at controlling ant populations via oral baits (Lopatina and Eremina 2010) and as sprays in urban and turf ecosystems (Wang et al. 2015a); however, little is known about how ant

140

communities will interact with neonicotinoids applied as seed treatments in agroecosystems.

In this study, we determined how exposure to imidacloprid via seed dressing impacts individual ant survival and behavior (i.e., distance traveled and foraging success), in-field predation on pest eggs, and ant community richness, evenness, and diversity in soybeans. We also examined mid-season pest abundance and yield in response to the neonicotinoid addition in field conditions.

We expected to find both lethal and sublethal impacts imidacloprid on the ants, resulting in a reduction of ant-mediated predation and community diversity.

Additionally, as Kentucky has many similar pests to the mid-south (including the three-cornered alfalfa hopper, Spissistilus festinus (Say) (Hemiptera:

Membracidae)), we expected the neonicotinoids to effectively control pest populations, increasing soybean yields.

7.3 Experimental Methods

7.3.1 Assessment of Lethal and Sublethal Effects in the Laboratory

Five colonies of Tetramorium caespitum L. (Hymenoptera: Formicidae), a common ant species in Kentucky soybean fields, were excavated from the

Spindletop Research Farm, Lexington, Kentucky, U.S.A., and allowed to settle into individual 19-liter buckets before testing. Colonies had access to water and were fed a diet of honey, peanut butter (ALDI, Inc., Batavia, Illinois, U.S.A.), and freshly killed adult Drosophila melanogaster Meigen (Diptera: Drosophilidae). All colonies were maintained under controlled conditions of 25 ± 1 °C, 65 ± 5% RH,

141

and 14: 10 L:D cycle. Multiple cohorts of three ants from one of five colonies

(drawn from equally) were selected from active foragers and randomly assigned a right leg (front, middle, or rear) to be amputated so that the individual ants could be tracked through the experiment.(Walker and Wineriter 1981) Ants were chilled in a refrigerator (1.6 °C) for 10 min before the leg was removed at the femur with sterilized dissecting scissors. Ants were allowed to groom themselves and acclimate to the newly amputated condition for 5 min. Generalized method order is depicted in Figure 7-1 Diagram of replication and steps involved in the testing via exposure and behavioral assays of lethal and sublethal effects of seed treatments on Tetramorium caespitum L.

7.3.1.1 Pre- Exposure Behavioral Assay

A randomized subset of 75 ants (one per cup) designated for foraging behavior assays was observed before treatment exposure of 48 h. To conduct each test, inner walls of a clean petri dish (12 cm diameter) were painted with

Fluon® (INSECT-a-SLIP, BioQuip Product, Inc., Rancho Dominguez, California,

USA) to keep ants on the bottom of the dish for accurate 2D distance measurements. The dish was placed on top of graph paper where the cell size (1

× 1 cm) was approximately equivalent to the average ant body length. A food source (20 µl sugar water) was pipetted on one side of the dish (determined at random) to promote ant foraging. The selected ant was placed opposite the food source under a clean plastic vial (5 ml) and allowed 5 min to adjust to the petri dish. Vials were removed and ants were given 10 min during which the number

142

of lines crossed per minute was recorded as well as the time it took the ant to find (initial antennation of the droplet) and to consume (initial contact of mouthparts with droplet) the sugar water.

7.3.1.2 Exposure Assay

To determine if the neonicotinoid exposure induced lethal or sublethal reactions in T. caespitum, arenas were constructed of small condiment containers (60 ml) where the lids were pierced with a sewing needle 16 times to prevent condensation while maintaining adequate humidity for ant survival.

Autoclaved sand was provided as a substrate (2.5 cm deep) and was dampened by the addition of 5 ml of distilled water. Each container was provided with a cotton ball of sugar water as a food source. Three seeds (same seeding rate as field component) were placed on top of the sand for ant exposure in the two seed treatments while the control contained only sand. Seeds were glyphosate- resistant soybeans (Fabales: Fabaceae, Glycine max (L.) Merr., Roundup Ready

2 Yield, Monsanto Asgrow variety AG4433) with Accerleron fungicide

(fluxapyroxad, metalaxyl, and pyraclostrobin) seed treatment (Monsanto, St.

Louis, Missouri) with half also treated with a neonicotinoid (imidacloprid) applied at a rate of 1.2 ml kg-¹ seed by a seed supplier (Southern States Cooperative,

Richmond, Virginia).

Following the pre-exposure behavior assay, all ants (n = 225 ants total) were placed into the containers for exposure to the seed treatments and observed then placed on their backs at 1, 2, 3, 4, 5, 24, and 48 h. Three ants

143

from the same colony were placed into each cup (n = 25 cups treatment-1) with colony source being evenly distributed between the cups. During each time point, each ant was classified as dead (D), intoxicated (I) as indicated by sluggish movements and inability to right itself and leg/antennae twitching, or normal (N) as indicated by standard foraging/digging behaviors and ability to immediately right itself after being turned over (Kunkel et al. 2001). The ant status at 48 h was regressed against treatment with the colony of origin as a random effect using multinomial logistic regression in JMP 12.0 (SAS Institute, Cary, North Carolina).

Fisher’s Exact Pairs test was used to compare the proportion of each classification with each seed treatment combination.

7.3.1.3 Post- Exposure Behavioral Assay

The previously tested sub-set of 75 (46 survived) ants designated for foraging behavior assays was observed after treatment exposure of 48 h. The change in the number of lines crossed per minute by each ant (from pre- to post- exposure) was regressed with treatment where the colony of origin was a random effect with a GLM. The difference in time it took the ants to find and consume the food source from pre- to post-exposure was analyzed in the same fashion.

7.3.2 Field Experiments

Field sites were located at the University of Kentucky Spindletop Research

Farm (GPS coordinates 38° 7' 37.2" N, 84° 30' 28.7994" W) and were planted

144

with soybeans after a season of conventionally-managed field corn treated with

Poncho 250 at a rate of 0.25 mg clothiadin kernel-¹. All plots were planted on 8

May 2015, with a row width of 76 cm.(Lee and Herbek 2011) Plots (n = 4; 45 ×

25 m) were no-till and full-season, with the same seeds used in the laboratory experiment. Plots were split in half (15.25 × 15.25m) using a randomized block design to receive imidacloprid. A buffer of four rows without the addition of imidacloprid was planted around and between the blocks to reduce edge effects.

7.3.2.1 Insect Sampling

All sampling locations were located a minimum of 3 m into the treatment block to further reduce edge effects. Wet pitfall traps (n = 16 block-1) for sampling the ant community in a grid pattern across the field consisted of plastic cups (dia.

9.5 cm, depth 12 cm), containing 5 cm of ethylene glycol and covered with foam plates for rain protection (dia. of 22 cm), and were set for 7-d intervals (VC, V1, and R2 plant stages). Once soybean plants reached the R2 stage, sweep net samples (10 figure-8 sweeps) for pests were also taken (R2, R3, and R4 stages) at each grid location (Johnson et al. 2015). All samples were then stored in 95% ethanol in twist-top bags until sorted; the contents of the pitfall samples were not pooled with those of the sweep samples.

Soybean pests (with the exception of grasshoppers (Orthoptera:

Acrididae) and weevils (Coleoptera: Curculionidae)) and ants were identified to the species level using representatives from the University of Kentucky

Entomology Museum collection previously identified by Gary Coovert together

145

with the Ants of Ohio identification guide (Coovert 2005). The abundances of collected ants and pests were recorded for each sampling location on each date per field. Additionally, the communities were analyzed in terms of species richness (total number of species found per sample per time), evenness, and

Shannon diversity (McCune and Grace 2002). Species abundance, richness, evenness, and diversity were regressed against plot as a random effect, sample date for repeated measures, and treatment in JMP 12.0 using the generalized linear model (GLM) function and Poisson distribution with log link for count data and a normal distribution for community composition metrics.

7.3.2.2 Estimating Ant Predation in the Field

To estimate the potential influence of imidacloprid seed treatments on ant predation, the removal rates of sentinel egg masses were observed.(Larson et al.

2012) Black cutworm, Agrotis ipsilon (Hufnagel) (Lepidoptera: Noctuidae), eggs were ordered from Benzon Research (Carlisle, Pennsylvania). Groups of 10 eggs on fabric pieces were placed on wooden stakes (15.2 × 3 × 0.25 cm) using double-sided tape and placed in two equidistant rows parallel to field rows (5 stakes row-1) within each block, such that the egg masses were about 3 cm above the ground. As a control for non-ant-mediated egg removal, petroleum jelly was applied to the base of three other stakes with attached eggs and placed between the two testing rows. Egg masses were placed in the field at soybean stage R3 (Mueller and Sisson 2016) and left for 24 hours after which, the number of eggs removed was recorded. During the first 5 h in the field, the egg masses

146

were observed every hour, and the arthropods on the egg masses were identified to the lowest possible level. The experiment was modified and rerun at soybean stage R4 with the final egg count occurring after only 5 h to gain greater resolution for short-term predation rates. The number of eggs remaining was analyzed in relation to treatment with plot and date as random effects, using JMP

12.0 (SAS Institute, Cary, North Carolina) using GLM with a normal distribution.

The rate at which ants discovered the egg masses during both of the 5 h observation periods was analyzed using Cox proportional hazard regression, and the treatments were compared using a survival log-rank test in JMP 12.0.

7.3.2.3 Plant Growth and Yield

The growth stage of plants of each plot was recorded at the same time as sample collections and predation assays. Soybeans were harvested on 3

November 2015, after they reach about 13% moisture content, using a plot combine (Wintersteiger HCGG+HM800, Ried im Innkreis, Austria). The soybean yield was calculated based on three sub-samples of four rows evenly spaced in each plot (12 rows harvested plot-1). Kilograms per hectare were recorded and adjusted for weight using moisture content. The final yield was compared between treatments with the plot as a random effect using JMP 12.0 using a

GLM with a normal distribution, and treatment effects were compared post-hoc using a t-test.

147

7.4 Results

7.4.1 Lethal and Sublethal Effects in the Laboratory

When T. caespitum was exposed to seeds in the laboratory the ants suffered significantly more intoxication and mortality from both types of treated seeds compared to the control (Fisher’s Exact Pairs, df = 2, P < 0.001; Figure

7-2). Mortality was higher from exposure to the fungicide-only treatment than from exposure to the seeds treated with fungicide plus imidacloprid (Fisher’s

Exact Pairs, df = 2, P < 0.0001; Figure 7-2). Cohorts of ants exposed to non- treated seeds had more normally functioning individuals than did those exposed to either of the seed treatments (Fisher’s Exact Pairs, control versus fungicide; df

= 2, P < 0.0001; Fisher’s Exact Pairs, control versus imidacloprid + fungicide; df

= 2, P < 0.001). Exposure to treated seeds also caused sublethal impairment of ants’ behavior. Ants exposed to non-treated seeds experienced no significant change in activity (distance traveled, measured as lines crossed per minute) from pre- to post-exposure (error encompassed zero). Ants exposed to either type of seed treatment were less active, crossing fewer lines per minute (Figure 7-3).

However, only the fungicide treatment was statistically lower than the control

(Fisher’s exact test, df = 2, P = 0.043). There were no treatment effects on foraging success (F = 0.89; df = 2, 7; P = 0.41) or consumption time (F = 2.06; df

= 2, 7; P = 0.22).

148

7.4.2 Ant Community Composition

Ant abundance was similar in the fungicide seed treatment and the fungicide + imidacloprid treatments blocks (χ² = 0.08, df = 1, P = 0.77). However, there was a date effect (χ² = 2164, df = 1, P < 0. 01), with the overall number of ants declining as the season progressed. Additionally, certain blocks had greater abundance of specific species (χ² = 32.28, df = 1, P < 0.01), indicating a location effect. Individual species were analyzed for treatment influences on abundance, and for half of them there was no influence (Table 7-1 Regression results per ant species when seed treatment (imidacloprid and fungicide) are taken into consideration with sampling date. However, for the 8 of 16 species that did exhibit treatment effects, where the parameter estimate was either positive (L. neoniger, P. tysoni, S. molesta) or negative (A. rudis, M. americana, N. faisonensis, P. dentata, and T. caespitum). Although there was a date effect (χ² =

196, P < 0.01), when ant species richness was analyzed in relation to treatment, no associations were found (χ² = 0.003, P = 0.96). However, there were significant treatment and date effects on both evenness (Treatment, χ² = 21.7, P

< 0.01; Time, χ² = 38.4, P < 0.01) and Shannon diversity (Treatment, χ² = 5.72, P

= 0.02; Time, χ² = 95.8, P < 0.01) of the ant communities (Error! Reference source not found.). Presence of the imidacloprid seed treatment increased both metrics, and both were decreased at later dates, with diversity showing a date by treatment interaction (χ² = 5.25, df = 1, P = 0.022).

149

7.4.3 Pest Abundance

No significant pest or disease damage was noted after planting until harvest. No differences were found in overall pest abundance (χ² = 1.08, df = 1,

P = 0.30), richness (χ² = 0.04, df = 1, P = 0.84), evenness (χ² = 0.12, df = 1, P =

0.73), or diversity (χ² = 0.06, df = 1, P = 0.81) between neonicotinoid treated and untreated blocks. The most abundant pests did not differ by treatment (Table 7-2

Regression results per pest species when seed treatment (imidacloprid and fungicide) are taken into consideration with sampling date.. Soybean aphid

(Hemiptera: Aphididae, Aphis glycines Matsumura) was not observed in the field until soybeans reached R7 and was not included in the analyses because its activity was outside of the sampling times. The three-cornered alfalfa hopper,

Spissistilus festinus (Say) (Hemiptera: Membracidae), another relatively common pest of soybeans that is targeted by neonicotinoid seed treatments, also was not observed during this study.

7.4.4 Predation

Lasius neoniger Emery, Pheidole tysoni Forel, Solenopsis molesta (Say), and T. caespitum were the most commonly observed ant species preying on the sentinel egg masses. Aphaenogaster rudis Emery was also observed feeding on the eggs during the 5-h observation periods. The time before some ants discovered the egg masses ranged from 1–5 h, and was unaffected by treatment at readings taken 5 h (χ² = 1.75, df = 1, P = 0.19) and 24 h (χ² = 2.04, df = 1, P =

0.15) after eggs were placed in the field (Figure 7-5). There were no treatment

150

differences in numbers of eggs remaining after 5 h (t = -0.69, df = 1, P = 0.49) or

24 h (t = 0.89, df = 1, P = 0.38; Figure 7-5).

7.4.5 Plant Growth and Yield

Rate of plant growth did not differ at all, therefore, no statistical analyses were run. When soybean yields were regressed against treatment, there were no treatment differences (F = 0.16; df = 1, 23; P = 0.70) or block effects (F = 1.26; df

= 3, 21; P = 0.33).

7.5 Discussion

Several previous studies have documented that insect susceptibility to neonicotinoids is dependent upon dosage, exposure route, and species.

Exposure to neonicotinoids may kill insects outright or may cause altered behaviors via sublethal effects (Barbieri et al. 2013, Galvanho et al. 2013), potentially including foraging (Thiel and Kohler 2016). When we exposed T. caespitum in the laboratory to field dosages of imidacloprid on fungicide-treated seeds, the ants’ mobility was impaired. The fungicide itself used on seed treatments had lethal effects and sublethal influences on ant movements, indicating potentially strong non-target effects on ant communities. Fungicides have been reported to cause death and altered foraging/nesting behaviors in other insects but have not been well studied on ants especially when used in conjunction with insecticides (Chalfant et al. 1977, Yang et al. 2011, Bozdogan

2014, Artz and Pitts-Singer 2015, Obear et al. 2015). As the foraging time and

151

success rates were not impacted in either the laboratory or field assays (Thiel and Kohler 2016), these sublethal effects from imidacloprid and fungicides might impact ant colonies in other ways such as their ability to reproduce or longevity and aggression of a colony (Desneux et al. 2007, Cantagalli et al. 2014, Wang et al. 2015b, Thiel and Kohler 2016). Additionally, while no differences in overall ant predation on eggs were seen in the field component of this study, they could exist on a species level.

Species-level interaction differences with imidacloprid were seen in how ant species’ abundances responded to seed dressing with this neonicotinoid.

Although some species were not affected, pitfall trap captures of others increased or decreased with the addition of imidacloprid to the system. The divergent reactions of these species could be due to ecological patterns of particular species in relation to nesting proximity to treated seeds or dietary preferences. Ants known to tend aphids and scales (i.e. L. neoniger and M. minimum) were not negatively impacted by the imidacloprid seed dressing despite aphids being among the targeted pests of this insecticide. However, no aphids were found in the field until R7, although they were tended by these species after that time.

With the exception of T. caespitum, activity-density of ant species (L. neoniger, S. molesta, and P. tysoni) that are generally regarded as relatively tolerant of disturbance (McGlynn 1999, Buczkowski and Richmond 2012) was not affected by the addition of the imidacloprid seed dressing in the field. Nesting preferences (commonly observed nesting in corn stalk residue) or general ability

152

to react to toxin may have impacted T. caespitum’s ability to recover from seed treatment exposure; however, neither are well studied especially in comparison to other common ant species. As T. caespitum is generally very competitive towards other ants and even ground-nesting bees (Schultz 1982), the decrease in its abundance might have provided release from competition for the other ant species present, resulting in an increase in their activity-density. This would also account for the significant changes in both ant evenness and diversity without changes to the total number of species present (richness) (DeJong 1975, Smith and Wilson 1996). Imidacloprid is only active for a few weeks after planting within the plant itself; potentially, this small exposure window is not adequate (Kunkel et al. 1999, Zenger and Gibb 2001b, Pretorius 2014) to induce larger immediate shifts in ant communities (Liu et al. 2011, Jones et al. 2014, Schaafsma et al.

2016, Xu et al. 2016). But the changes in ant community composition seen in this study suggest some field-level impact on ants.

Given that the treatments had no influence on pest abundance and soybean yields, and that imidacloprid has the potential to harm other beneficial organisms including ground beetles, pollinating Hymenoptera, and earthworms

(Stapel et al. 2000, Li et al. 2015, Yao et al. 2015), those risks need to be considered to determine if the benefits of neonicotinoid seed treatments are worth the costs. The apparent lack of pest reduction or yield increase in our study mirrors the original EPA impact report (Myers and Hill 2014) which questions if neonicotinoid seed treatments for soybeans provide sufficient benefit to warrant their use. However, as the field component of this study only spanned one year

153

in one geographic location, the results cannot be generalized. Most (8/10) of the collected pests in this study are not usually targeted by neonicotinoid seed treatments especially after the plants reach R2, which is consistent with pest management practices in other Midwestern states (Myers and Hill 2014).

However, warmer growing seasons in the future might affect the types of pests present and the speed at which they colonize soybeans, potentially changing the cost-benefit of use of insecticide-treated seeds in Kentucky.

This study indicated that both imidacloprid and fungicides can have lethal and sublethal impacts on ant communities. Individual ant survival and mobility were negatively impacted by both seed treatments, but did not alter foraging ability. The field trials indicated that these effects could result in a community composition shift where more susceptible species decrease in abundance.

However, this community shift, at least in the short term, did not cause different predation rates on sentinel egg sticks or overall changes in yield.

154

Table 7-1 Regression results per ant species when seed treatment (imidacloprid and fungicide) are taken into consideration with sampling date.

Seed Intercept = Fung. Treatment Time Mean Abundance Parameter Parameter Ant Species / Sample Estimate χ² (df) P-value Estimate χ² (df) P-value All Ants 13.28 0.01 0.08 (1) 0.77 -0.87 2164 (1) < 0.01 Aphaenogaster rudis 0.22 -0.20 3.12 (1) 0.08 -1.59 84.08 (1) < 0.01 Camponotus americanus 0.003 9.60 1.39 (1) 0.24 -18.39 2.20 (1) 0.14 Camponotus pennsylvanicus 0.003 9.60 1.39 (1) 0.24 -18.39 2.20 (1) 0.14 Formica biophilica 0.16 -0.02 0.03 (1) 0.85 -1.65 64.43 (1) < 0.01 Formica subsericea 0.02 -0.15 0.16 (1) 0.69 -18.22 15.46 (1) < 0.01 Hypoponera opacior 0.02 -0.35 0.69 (1) 0.41 0.24 0.51 (1) 0.63 Lasius neoniger 5.06 -0.05 5.23 (1) 0.02 -1.05 1114 (1) < 0.01 Myrmecina americana 0.01 9.73 4.16 (1) 0.04 8.6e-15 3.6e-15 (1) 0.99 Myrmica latifrons 0.09 0.29 2.87 (1) 0.10 -0.48 4.87 (1) 0.03 Nylanderia faisonensis 0.18 0.25 4.23 (1) 0.04 -0.62 16.01 (1) < 0.01 Pheidole dentata 0.08 1.30 25.65 (1) < 0.01 -1.20 20.51 (1) < 0.01 Pheidole dentigula 0.01 -0.69 1.92 (1) 0.17 -0.64 1.23 (1) 0.27 Pheidole tysoni 2.85 -0.28 83.66 (1) < 0.01 -0.61 249.1 (1) < 0.01 Ponera pennsylvanica 0.01 9.33 6.87 (1) 0.01 -0.01 0.00 (1) 0.99 Solenopsis molesta 0.18 -0.58 19.39 (1) < 0.01 -0.27 3.24 (1) 0.07 Tetramorium caespitum 4.40 0.21 72.84 (1) < 0.01 -0.87 709.1 (1) < 0.01

155

Table 7-2 Regression results per pest species when seed treatment (imidacloprid and fungicide) are taken into consideration with sampling date.

Seed Intercept = Fung. Treatment Time Mean Abundance Parameter Parameter Pest Species / Sample Estimate χ² (df) P-value Estimate χ² (df) P-value All Pests 13.96 0.03 1.08 (1) 0.30 -0.07 4.53 (1) 0.03 Acrididae 0.83 0.18 2.43 (1) 0.11 -0.06 0.18 (1) 0.67 Cerotoma trifurcata 2.12 0.07 1.03 (1) 0.31 -0.18 4.27 (1) 0.04 Chinavia hilaris 1.37 -0.07 0.44 (1) 0.51 0.63 30.46 (1) < 0.01 Colaspis brunnea 0.52 -0.14 0.39 (1) 0.53 -1.37 41.69 (1) < 0.01 Curculionidae 0.99 0.01 0.01 (1) 0.91 0.72 28.05 (1) < 0.01 Diabrotica unidecimpunctata 0.22 0.27 1.26 (1) 0.26 0.46 2.58 (1) 0.11 Euschistus servus 0.07 0.15 0.14 (1) 0.71 -0.48 0.98 (1) 0.32 Hypena scabra 1.52 0.05 0.15 (1) 0.70 -1.13 92.71 (1) < 0.01 Popillia japonica 6.27 -0.01 0.03 (1) 0.86 -0.01 0.02 (1) 0.88

156

Figure 7-1 Diagram of replication and steps involved in the testing via exposure and behavioral assays of lethal and sublethal effects of seed treatments on Tetramorium caespitum L.

157

158

Figure 7-2 The proportion of ants (Tetramorium caespitum L.) classified as normal, intoxicated (unable to right itself and excessive grooming), or dead after 48 h exposure to seed treatments in the laboratory.

159

Figure 7-3 Change in the average number of lines crossed per minute over 10 min between pre- and post-exposure by ants (Tetramorium caespitum L.) as an indication of sublethal impacts of imidacloprid on ant behavior.

160

Figure 7-4 Ant community composition metrics –species richness, evenness, and Shannon diversity – as observed between seed treatments in the field.

161

Figure 7-5 (A) Survival log-rank curves for ant (Formicidae) discovery time of egg masses within 1-5 h of exposure in the field and (B) Number of black cutworm (Agrotis ipsilon (Hufnagel)) eggs surviving in the field after 5 h.

Chapter 8 CONCLUSIONS AND BROADER IMPACTS

Within this dissertation, I evaluated how land covers at multiple scales can impact ant and spider community abundance and diversity in soybeans in addition to the subsequent spatial and temporal associations with insect pest species. I have shown that molecular gut content analyses via PCR can be an effective way of monitoring ant predation even when honeydew producing insects are present and that these predation events correspond with the spatial and temporal associations with pest species. However, ant interactions can influence pest behavior rather than suppress pest abundance, causing a non-trophic cascade through the system, potentially increasing plant damage. Finally, I have shown that commonly used imidacloprid and fungicide soybean seed treatments can impact ant survival, behavior, and community composition but this does not impact short term predations rates in soybeans.

The first hurdle to researching these topics was the evaluation and optimization of the molecular gut content techniques employed (Gadau 2009).

Initially, I found that field collected ants such as Tetramorium caespitum were not producing adequate PCR results based on observed feeding activity (Penn et al.

2016). I found an optimal method of preventing such false negatives by the dissection of the crop contents and the addition of bovine serum albumin (BSA) to ensure that all field collected samples were not wasted (Chapter 2). The next step to effective molecular gut content analysis of ants was to determine if honeydew, commonly eaten by many ants, contained aphid DNA, potentially providing a false positive for the gut content analyses. However, these fears were put to rest in 162

Chapter 3 (Penn and Harwood 2016). These two chapters of this dissertation have addressed complications associated with molecular gut content analysis of ants, allowing for me to use these techniques for my field studies.

Much of this dissertation used Spatial Analysis by Distance IndicEs (SADIE) to determine ant and spider spatial correlation with soybean pest species (Penn In prep., Penn et al. In review). By using this technique, based in the popular Kriging method but incorporating right-tailed biased count data, I determined if it was worth testing field-collected predators’ guts for pest species’ DNA. For instance, if ant never encountered a grape colaspis beetle in the field (i.e. no overlap both spatially and temporally), then it is a waste of resources to use molecular gut content analysis to determine if predation occurred. By using SADIE, I have further confirmed the prior research that estimating the significant spatial and temporal correlation of populations is useful to determining the likelihood of predation events particularly when coupled with molecular techniques.

Based on Chapter 4 and Chapter 5 results, it appears to be possible to manage multiple predators and pest species within agroecosystems simultaneously. For instance, ants responded most to land covers at a 0.5 km scale whereas spiders and pests could easily be manipulated by on-farm land uses such as field margins. Additionally, the specific land covers influencing predators and pests are compatible. For instance, ant diversity and abundance benefitted from the addition of pasture to the broader landscape, which increased within field spatial overlap of ants to pests. This same area, weedy/grassy patches, also benefited spider spatial overlap with pests early in the season, indicating that

163

incorporating relatively undisturbed grassy areas are beneficial to crop production.

The same holds true for wooded areas which increase spider abundance but decrease pest abundance within the fields. In addition, such landscape manipulations not only benefit growers in terms of pest suppression but also have been shown to decrease soil erosion from wind and rain run-off (Pimentel et al.

1995, Nelson et al. 1998). From my personal conversations with Kentucky soybean growers, many are removing wooded areas and pastures from their farms to increase ease of access into the current field sites and to obtain tax incentives at the state level. This removal is costly to them and could in fact be detrimental to the sustainability of their farmland (Barberi et al. 2010, Batary et al. 2011). Further research into the synergies such landscape components have with crop production, environmental conservation for recreation (i.e. hunting and birding) and restoration (EPA incentives), and aesthetics could persuade land managers to maintain such areas to the benefit of biological control initiatives (Pimentel et al.

1992, Donald and Evans 2006, Henderson et al. 2012).

Synergies between predators are not always possible, so even though both ants and spiders appear to be manipulated by landscapes, they still might interfere with each other in the field. According to the field cage study in Chapter 6, we saw that Lasius neoniger was in fact increasing damage done by the green cloverworm,

Hypena scabra, even in the presence of another common predator –Oxyopes salticus. These interactions might be species specific and vary given different microhabitats (available corn husks versus fully tilled soil or very weedy areas versus highly controlled fields). This study also revealed that both ant and spider

164

influences on soybean pest species are not necessarily consumptive but behavior modifying. Given our knowledge about the SADIE data where ants, spiders, and pests overlap, this might contribute to the plant damage control seen in the field.

So, predators such as ants and spiders might be consuming prey (as seen in

Chapter 4) but also altering prey behaviors that diminish plant damage. However, these behavior alternations go both ways, and the positive influence of ant predation and spider presence might change when ants induce spider behavioral differences.

Finally, ant behavioral alterations do not always manifest in diminished predation on pest species when such alterations are induced by soybean seed treatments (Chapter 7)(Penn and Dale In revision). This is an example of a landscape-wide practice since most soybeans grown in the US have either fungicide or imidacloprid applied to their coats (Myers and Hill 2014, North et al.

2016) and impacts ant community composition. Therefore, when determining landscape influences on ant communities, simply considering land cover is insufficient. Land use and management is also a consideration. This can be seen in other systems where tillage regimes have been found to greatly alter ant diversity (Peck et al. 1998, Pretorius 2014, Tamburini et al. 2016). In the future, integrating knowledge of land cover, land use, community spatial trends, and intraguild interactions in order to fully understand generalist predation by ants and spiders, to contribute to pest suppression in an economically meaningful way

(Harwood et al. 2009, Choate and Drummond 2011, Phillips and Gardiner 2016).

165

This dissertation has combined techniques and principles of landscape and molecular ecology in order to gain a more realistic understanding of how to implement conservation biological control using generalist predators in a soybean system. At the time of writing, this dissertation represents one of the first successful published attempts to integrate molecular gut content analysis with mapping of the spatial and temporal overlaps of predators and pests.

Additionally, it is one of the few successful uses of molecular gut contents using ants (Hymenoptera: Formicidae) as the targeted generalist predator and was only made possible by the molecular methods developed herein. This dissertation also reveals the intraguild interactions between two generalist predators and how to manipulate their within field abundances based on land cover management at different spatial scales. The combination of methods and the integration of land use (neonicotinoid seed treatments) in conjunction with l and land cover will prove to be useful in future sustainable management of farmscapes for pest control and non-target effect remediation as long as grower buy-in can be obtained.

166

REFERENCES AAS. 2005. Spiders of North America: An identification manual, American Arachnological Society, Columbia, MO, USA. Adams, C. T., W. A. Banks, C. S. Lofgren, B. J. Smittle, and D. P. Harlan. 1983. Impact of the red imported fire ant, Solenopsis invicta (Hymenoptera: Formicidae), on the growth and yield of soybeans. J. Econ. Entomol. 76: 1129-1132. Agarwal, V., N. Rastogi, and S. Raju. 2007. Impact of predatory ants on two Lepidopteran insect pests in Indian cauliflower agroecosystems. J. Appl. Entomol. 131: 493-500. Agerbo, E., J. A. C. Sterne, and D. J. Gunnell. 2007. Combining individual and ecological data to determine compositional and contextual socio-economic risk factors for suicide. Soc. Sci. Med. 64: 451-461. Alderweireldt, M. 1989. An ecological analysis of the spider fauna (Araneae) occurring in maize fields, Italian ryegrass fields and their edge zones, by means of different multivariate techniques. Agric. Ecosyst. Environ. 27: 293-306. Altieri, M. A. 1999. The ecological role of biodiversity in agroecosystems. Agric. Ecosyst. Environ. 74: 19-31. Altieri, M. A., and L. L. Schmidt. 1986. The dynamics of colonizing arthropod communities at the interface of abandoned, organic and commercial apple orchards and adjacent woodland habitats. Agric. Ecosyst. Environ. 16: 29-43. Amaral, D. S., M. Venzon, M. V. Duarte, F. F. Sousa, A. Pallini, and J. D. Harwood. 2013. Non-crop vegetation associated with chili pepper agroecosystems promote the abundance and survival of aphid predators. Biol. Control 64: 338-346. Andrews, E. S. 2013. Analyzing arthropods for the presence of bacteria. Curr. Protoc. Microbiol.: 1E. 6.1-1E. 6.14. Applebaum, S. W., and Y. Heifetz. 1999. Density-dependent physiological phase in insects. Annu. Rev. Entomol. 44: 317-341. Artz, D. R., and T. L. Pitts-Singer. 2015. Effects of fungicide and adjuvant sprays on nesting behavior in two managed solitary bees, Osmia lignaria and Megachile rotundata. PloS one 10: e0135688. ASA. 2016. SoyStats 2016. In T. A. S. Association [ed.], http://soystats.com/. Athey, K. J., J. Dreyer, K. K. Kowles, H. J. Penn, M. I. Sitvarin, and J. D. Harwood. 2016. Spring Forward: Molecular detection of early season predation in agroecosystems. Food Webs 9: 25-31. Ayre, G. 1967. The relationships between food and digestive enzymes in five species of ants (Hymenoptera: Formicidae). The Canadian Entomologist 99: 408-411. Banks, W. A., C. T. Adams, and C. S. Lofgren. 1991. Damage to young citrus trees by the red imported fire ant (Hymenoptera: Formicidae). J. Econ. Entomol. 84: 241-246. Barberi, P., G. Burgio, G. Dinelli, A. Moonen, S. Otto, C. Vazzana, and G. Zanin. 2010. Functional biodiversity in the agricultural landscape: Relationships between weeds and arthropod fauna. Weed Res. 50: 388-401. Barbieri, R. F., P. J. Lester, A. S. Miller, and K. G. Ryan. 2013. A neurotoxic pesticide changes the outcome of aggressive interactions between native and invasive ants. Proc. R. Soc. B-Biol. Sci. 280: 20132157. Barbosa, P., and I. Castellanos. 2005. Ecology of predator-prey interactions, Oxford University Press, New York, NY, USA. Batary, P., A. Baldi, D. Kleijn, and T. Tscharntke. 2011. Landscape-moderated biodiversity effects of agri-environmental management: A meta-analysis. Proc. R. Soc. B-Biol. Sci. 278: 1894-1902. Bedford, S., and M. Usher. 1994. Distribution of arthropod species across the margins of farm woodlands. Agric. Ecosyst. Environ. 48: 295-305. Bell, J. R., R. Andrew King, D. A. Bohan, and W. O. C. Symondson. 2010. Spatial co- occurrence networks predict the feeding histories of polyphagous arthropod predators at field scales. Ecography 33: 64-72. Benton, T. G., J. A. Vickery, and J. D. Wilson. 2003. Farmland biodiversity: Is habitat heterogeneity the key? Trends Ecol. Evol. 18: 182-188.

167

Bhar, R., and L. Fahrig. 1998. Local vs. landscape effects of woody field borders as barriers to crop pest movement. Conserv. Ecol. 2: http://www.consecol.org/vol2/iss2/art3/. Bianchi, F. J., C. J. Booij, and T. Tscharntke. 2006. Sustainable pest regulation in agricultural landscapes: A review on landscape composition, biodiversity and natural pest control. Proc. R. Soc. B-Biol. Sci. 273: 1715-1727. Bianchi, F. J. J. A., W. K. R. E. van Wingerden, A. J. Griffioen, M. van der Veen, M. J. J. van der Straten, R. M. A. Wegman, and H. A. M. Meeuwsen. 2005. Landscape factors affecting the control of Mamestra brassicae by natural enemies in brussels sprout. Agric. Ecosyst. Environ. 107: 145-150. Birthisel, S. K., E. R. Gallandt, and R. Jabbour. 2014. Habitat effects on second-order predation of the seed predator Harpalus rufipes and implications for weed seedbank management. Biol. Control 70: 65-72. Birthisel, S. K., E. R. Gallandt, R. Jabbour, and F. A. Drummond. 2015. Habitat and time are more important predictors of weed seed predation than space on a diversified vegetable farm in Maine, USA. Weed Sci. 63: 916-927. Blackman, R. L., and V. F. Eastop. 2007. Taxonomic Issues. In V. Emden, H. Fritz and R. Harrington (eds.), Aphids as crop pests. CAB International, Cambridge, Massachusetts, USA. Blacquiere, T., G. Smagghe, C. A. Van Gestel, and V. Mommaerts. 2012. Neonicotinoids in bees: A review on concentrations, side-effects and risk assessment. Ecotoxicology 21: 973-992. Blau, W. S., and R. E. Stinner. 1983. Temporal flight patterns in Mexican bean beetle (Coleoptera: Coccinellidae) and their relation to weather. Environ. Entomol. 12: 1047- 1054. Blubaugh, C. K., and I. Kaplan. 2015. Tillage compromises weed seed predator activity across developmental stages. Biol. Control 81: 76-82. Blüthgen, N., G. Gebauer, and K. Fiedler. 2003. Disentangling a rainforest food web using stable isotopes: Dietary diversity in a species-rich ant community. Oecologia 137: 426- 435. Bonmatin, J., I. Moineau, R. Charvet, M. Colin, C. Fleche, and E. Bengsch. 2005. Behaviour of imidacloprid in fields. Toxicity for honey bees, pp. 483-494. In E. Lichtfouse, J. Schwarzbauer and D. Robert (eds.), Environmental Chemistry. Springer, Berlin Germany. Booman, G. C., P. Laterra, V. Comparatore, and N. Murillo. 2009. Post-dispersal predation of weed seeds by small vertebrates: Interactive influences of neighbor land use and local environment. Agric. Ecosyst. Environ. 129: 277-285. Boreau de Roincé, C., C. Lavigne, J.-F. Mandrin, C. Rollard, and W. Symondson. 2013. Early-season predation on aphids by winter-active spiders in apple orchards revealed by diagnostic PCR. Bull. Entomol. Res. 103: 148-154. Bortolotti, L., R. Montanari, J. Marcelino, P. Medrzycki, S. Maini, and C. Porrini. 2003. Effects of sub-lethal imidacloprid doses on the homing rate and foraging activity of honey bees. Bull. Insectol. 56: 63-67. Bowden, J., and G. J. W. Dean. 1977. The distribution of flying insects in and near a tall hedgerow. J. Appl. Ecol. 14: 343-354. Bownes, A., S. D. Moore, and M. H. Villet. 2014. My enemy's enemies: recruiting hemipteran- tending generalist ants for biological control in citrus orchards by spatial partitioning of foraging webs. Afr. Entomol. 22: 519-529. Bozdogan, A. M. 2014. Assessment of total risk on non-target organisms in fungicide application for agricultural sustainability. Sustainability 6: 1046-1058. Breene, R. G., W. L. Sterling, and M. Nyffeler. 1990. Efficacy of spider and ant predators on the cotton fleahopper [Hemiptera: Miridae]. Entomophaga 35: 393-401. Breton, S., G. Burger, D. T. Stewart, and P. U. Blier. 2006. Comparative analysis of gender- associated complete mitochondrial genomes in marine mussels (Mytilus spp.). Genetics 172: 1107-1119. Brewer, M. J., and N. C. Elliott. 2004. Biological control of cereal aphids in North America and mediating effects of host plant and habitat manipulations. Annu. Rev. Entomol. 49: 219- 242.

168

Brodt, S., K. Klonsky, L. Jackson, S. B. Brush, and S. Smukler. 2008. Factors affecting adoption of hedgerows and other biodiversity-enhancing features on farms in California, USA. Agroforestry Systems 76: 195-206. Brown, A. W. A. 1978. Ecology of pesticides, John Wiley & Sons., New York, NY, USA. Bruning, A. 1991. The effect of a single colony of the red wood ant, Formica polyctena, on the spider fauna (Araneae) of a beech forest floor. Oecologia 86: 478-483. Buckley, R. 1990. Ants protect tropical homoptera against nocturnal spider predation. Biotropica 22: 207-209. Buczkowski, G., and M. VanWeelden. 2010. Foraging arena size and structural complexity affect the dynamics of food distribution in ant colonies. Environ. Entomol. 39: 1936-1942. Buczkowski, G., and D. S. Richmond. 2012. The effect of urbanization on ant abundance and diversity: A temporal examination of factors affecting biodiversity. PloS one 7: e41729. Buddle, C. M., S. Higgins, and A. L. Rypstra. 2004. Ground-dwelling spider assemblages inhabiting riparian forests and hedgerows in an agricultural landscape. The American Midland Naturalist 151: 15-26. Burton, R., C. Kuczera, and G. Schwarz. 2008. Exploring farmers' cultural resistance to voluntary agri‐environmental schemes. Soc. Ruralis 48: 16-37. Butler, C. D., and R. J. O'Neil. 2007. Life history characteristics of Orius insidiosus (Say) fed diets of soybean aphid, Aphis glycines Matsumura and soybean thrips, Neohydatothrips variabilis (Beach). Biol. Control 40: 339-346. Calabuig, A., F. Garcia-Mari, and A. Pekas. 2015. Ants in citrus: impact on the abundance, species richness, diversity and community structure of predators and parasitoids. Agric Ecosyst Environ 213: 178-185. Campbell, K. U. 2015. Biodiversity of ants and associated mites in constructed grasslands at multiple spatial scales. Miami University. Campbell, N. J. 1990. An evaluation of Formica exsectoides Forel as a potential biological control agent of insect pests of pines. Michigan State University. Canard, M. 2001. Natural food and feeding habits of lacewings, pp. 116-129. In P. McEwen, T. New and A. Whittington (eds.), Lacewings in the crop environment. Cambridge University Press, New York, USA. Cantagalli, L. B., D. A. Lopes, A. L. Paz Barateiro Stuchi, and M. C. Colla Ruvolo- Takasusuki. 2014. Leaf-cutting ants Acromyrmex niger Smith, 1858 (Hymenoptera; Formicidae) used as bioindicators of agrotoxics residues. Acta Biol Colomb 19: 233-239. Caro, G., R. Marrec, B. Gauffre, M. Roncoroni, S. Augiron, and V. Bretagnolle. 2016. Multi- scale effects of agri-environment schemes on carabid beetles in intensive farmland. Agric. Ecosyst. Environ. 229: 48-56. Carroll, C., and D. H. Janzen. 1973. Ecology of foraging by ants. Annu. Rev. Ecol. Syst. 4: 231- 257. Casula, P., A. Wilby, and M. B. Thomas. 2006. Understanding biodiversity effects on prey in multi-enemy systems. Ecol. Lett. 9: 995-1004. Chagnon, M., D. Kreutzweiser, E. A. D. Mitchell, C. A. Morrissey, D. A. Noome, and J. P. Van der Sluijs. 2015. Risks of large-scale use of systemic insecticides to ecosystem functioning and services. Environ. Sci. Pollut. Res. 22: 119-134. Chalfant, R. B., J. W. Todd, W. Kent Taylor, and B. Mullinix. 1977. Laboratory studies on the antifeeding effect of a fungicide, guazatine, on eleven species of phytophagous insects. J. Econ. Entomol. 70: 513-517. Chapin III, F. S., E. S. Zavleta, V. T. Eviner, R. L. Naylor, P. M. Vitousek, H. L. Reynolds, D. U. Hooper, S. Lavorel, O. E. Sala, S. E. Hobbie, M. C. Mack, and S. Diaz. 2000. Consequences of changing biodiversity. Nature 405: 234-242. Chaplin-Kramer, R., M. E. O'Rourke, E. J. Blitzer, and C. Kremen. 2011. A meta-analysis of crop pest and natural enemy response to landscape complexity. Ecol. Lett. 14: 922-932. Chapman, E. G., S. A. Romero, and J. D. Harwood. 2010. Maximizing collection and minimizing risk: does vacuum suction sampling increase the likelihood for misinterpretation of food web connections? Mol. Ecol. Resour. 10: 1023-1033.

169

Chapman, E. G., J. M. Schmidt, K. D. Welch, and J. D. Harwood. 2013. Molecular evidence for dietary selectivity and pest suppression potential in an epigeal spider community in winter wheat. Biol. Control 65: 72-86. Chen, Y., K. Giles, M. Payton, and M. Greenstone. 2000. Identifying key cereal aphid predators by molecular gut analysis. Mol. Ecol. 9: 1887-1898. Chisholm, C., Z. Lindo, and A. Gonzalez. 2011. Metacommunity diversity depends on connectivity and patch arrangement in heterogeneous habitat networks. Ecography 34: 415-424. Choate, B., and F. Drummond. 2011. Ants as biological control agents in agricultural cropping systems. Terrestrial Arthropod Reviews 4: 157-180. Choate, B. A., and J. G. Lundgren. 2015. Invertebrate communities in spring wheat and the identification of cereal aphid predators through molecular gut content analysis. Crop Protect. 77: 110-118. Clark, R. J., R. R. Jackson, and B. Cutler. 2000. Chemical cues from ants influence predatory behavior in Habrocestum pulex, an ant-eating jumping spider (Araneae, Salticidae). J. Arachnol. 28: 309-318. Clark, R. J., P. J. Gerard, and J. M. Mellsop. 2004. Spider biodiversity and density following cultivation in pastures in the Waikato, New Zealand. N. Z. J. Agric. Res. 47: 247-259. Clay, J. 2004. World agriculture and the environment: A commodity-by-commodity guide to impacts and practices, Island Press, Washington, DC, USA. Clough, Y., A. Kruess, D. Kleijn, and T. Tscharntke. 2005. Spider diversity in cereal fields: Comparing factors at local, landscape and regional scales. J. Biogeogr. 32: 2007-2014. Clough, Y., J. Ekroos, A. Báldi, P. Batáry, R. Bommarco, N. Gross, A. Holzschuh, S. Hopfenmüller, E. Knop, M. Kuussaari, R. Lindborg, L. Marini, E. Öckinger, S. G. Potts, J. Pöyry, S. P. M. Roberts, I. Steffan-Dewenter, and H. G. Smith. 2014. Density of insect-pollinated grassland plants decreases with increasing surrounding land-use intensity. Ecol. Lett. 17: 1168-1177. Coovert, G. A. 2005. The ants of Ohio (Hymenoptera: Formicidae), vol. 15, Ohio Biological Survey, Inc., Columbus, OH, USA. Costamagna, A. C., D. A. Landis, and C. D. Difonzo. 2007. Suppression of soybean aphid by generalist predators results in a trophic cascade in soybeans. Ecol. Appl. 17: 441-451. Crowder, D. W., and W. E. Snyder. 2010. Eating their way to the top? Mechanisms underlying the success of invasive insect generalist predators. Biol. Invasions 12: 2857-2876. Crowder, D. W., and R. Jabbour. 2014. Relationships between biodiversity and biological control in agroecosystems: Current status and future challenges. Biol. Control 75: 8-17. Cullen, R., K. D. Warner, M. Jonsson, and S. D. Wratten. 2008. Economics and adoption of conservation biological control. Biol. Control 45: 272-280. Cushing, P. E. 2012. Spider-ant associations: An updated review of myrmecomorphy, myrmecophily, and myrmecophagy in spiders. Psyche: A Journal of Entomology 2012: 23. Cushman, J. H., and T. G. Whitham. 1991. Competition mediating the outcome of a mutualism - Protective services of ants as a limiting resource for Membracids. Am. Nat. 138: 851-865. Cushman, J. H., J. F. Addicott, C. R. Huxley, and D. F. Cutler. 1991. Conditional interactions in ant-plant-herbivore mutualisms, Oxford University Press, New York, NY, USA. Cusser, S., J. L. Neff, and S. Jha. 2016. Natural land cover drives pollinator abundance and richness, leading to reductions in pollen limitation in cotton agroecosystems. Agric. Ecosyst. Environ. 226: 33-42. Dauber, J., and V. Wolters. 2004. Edge effects on ant community structure and species richness in an agricultural landscape. Biodivers. Conserv. 13: 901-915. Dauber, J., and V. Wolters. 2005. Colonization of temperate grassland by ants. Basic Appl. Ecol. 6: 83-91. Dedryver, C. A., A. Le Ralec, and F. Fabre. 2010. The conflicting relationships between aphids and men: A review of aphid damage and control strategies. C. R. Biol. 333: 539-553. DeJong, T. M. 1975. A comparison of three diversity indices based on their components of richness and evenness. Oikos 26: 222-227.

170

Del Toro, I., R. R. Ribbons, and S. L. Pelini. 2012. The little things that run the world revisited: a review of ant-mediated ecosystem services and disservices (Hymenoptera: Formicidae). Myrmecol News 17: 133-146. Denno, R. F., and M. A. Peterson. 1995. Density-dependent dispersal and its consequences for population dynamics, pp. 113-130, Population dynamics: new approaches and synthesis. Academic Press, New York, New York, USA. Desneux, N., A. Decourtye, and J.-M. Delpuech. 2007. The sublethal effects of pesticides on beneficial arthropods. Annu. Rev. Entomol. 52: 81-106. Di Orio, A. P., R. Callas, and R. J. Schaefer. 2005. Forty-eight year decline and fragmentation of aspen (Populus tremuloides) in the South Warner Mountains of California. For. Ecol. Manage. 206: 307-313. Diame, L., I. Grechi, J. Y. Rey, C. A. B. Sane, P. Diatta, J. F. Vayssieres, A. Yasmine, H. De Bon, and K. Diarra. 2015. Influence of Oecophylla longinoda Latreille, 1802 (Hymenoptera: Formicidae) on mango infestation by Bactrocera dorsalis (Hendel) (Diptera: Tephritidae) in relation to Senegalese orchard design and management practices. Afr. Entomol. 23: 294-305. Diaz, M. 1992. Spatial and temporal patterns of granivorous ant seed predation in pathy cereal crop areas of central Spain. Oecologia 91: 561-568. Donald, P. F., and A. D. Evans. 2006. Habitat connectivity and matrix restoration: The wider implications of agri‐environment schemes. J. Appl. Ecol. 43: 209-218. Donovan, T. M., P. W. Jones, E. M. Annand, and F. R. Thompson. 1997. Variation in local- scale edge effects: Mechanisms and landscape context. Ecology 78: 2064-2075. Douglas, A. E. 2003. The nutritional physiology of aphids. Adv. Insect Physiol. 31: 73-140. Douglas, M. R., and J. F. Tooker. 2015. Large-scale deployment of seed treatments has driven rapid increase in use of neonicotinoid insecticides and preemptive pest management in U.S. Field crops. Environ. Sci. Technol. 49: 5088-5097. Dramstad, W. E., and W. J. Fjellstad. 2011. Landscapes: Bridging the gaps between science, policy and people. Landscape Urban Plann. 100: 330-332. Duelli, P. 1990. Population movements of arthropods between natural and cultivated areas. Biol. Conserv. 54: 193-207. Dyer, L. E., and D. A. Landis. 1997. Influence of noncrop habitats on the distribution of Eriborus terebrans (Hymenoptera: Ichneumonidae) in cornfields. Environ. Entomol. 26: 924-932. Edwards, G. B., J. F. Carroll, and W. H. Whitcomb. 1974. Stoidis aurata (Araneae: Salticidae), a spider predator of ants. Fla. Entomol. 57: 337-346. Egert, M., S. Marhan, B. Wagner, S. Scheu, and M. W. Friedrich. 2004. Molecular profiling of 16S rRNA genes reveals diet-related differences of microbial communities in soil, gut, and casts of Lumbricus terrestris L. (Oligochaeta: Lumbricidae). FEMS Microbiology Ecology 48: 187-197. Epperson, D. M., and C. R. Allen. 2010. Red imported fire ant impacts on upland arthropods in Southern Mississippi. Am. Midl. Nat. 63: 54-63. Eubanks, M. D. 2001. Estimates of the direct and indirect effects of red imported fire ants on biological control in field crops. Biol. Control 21: 35-43. Eubanks, M. D., and D. L. Finke. 2014. Interaction webs in agroecosystems: Beyond who eats whom. Current Opinion in Insect Science 2: 1-6. Eubanks, M. D., S. A. Blackwell, C. J. Parrish, Z. D. Delamar, and H. Hull-Sanders. 2002. Intraguild predation of beneficial arthropods by red imported fire ants in cotton. Environ. Entomol. 31: 1168-1174. Evans, E. W. 2000. Egg production in response to combined alternative foods by the predator Coccinella transversalis. Entomologia Experimentalis et Applicata 94: 141-147. Fagan, W. F., R. S. Cantrell, and C. Cosner. 1999. How habitat edges change species interactions The American Naturalist 153: 165-182. Fagan, W. F., E. Siemann, C. M. Mitter, R. F. Denno, A. F. Huberty, and H. A. Woods. 2003. Nitrogen in insects: Implications for trophic complexity and species diversification. SICB Annual Meeting & Exhibition Final Program and Abstracts 2003: 168. Fahrig, L. 2003. Effects of habitat fragmentation on biodiversity. Annu. Rev. Ecol. Evol. Syst. 34: 487-515.

171

Fahrig, L., and I. Jonsen. 1998. Effect of habitat patch characteristics on abundance and diversity of insects in an agricultural landscape. Ecosystems 1: 197-205. Fahrig, L., J. Baudry, L. Brotons, F. G. Burel, T. O. Crist, R. J. Fuller, C. Sirami, G. M. Siriwardena, and J. L. Martin. 2011. Functional landscape heterogeneity and animal biodiversity in agricultural landscapes. Ecol. Lett. 14: 101-112. Ferrari, J., J. A. West, S. Via, and H. C. J. Godfray. 2012. Population genetic structure and secondary symbionts in host-associated populations of the pea aphid complex. Evolution; international journal of organic evolution 66: 375-390. Finke, D. L., and R. F. Denno. 2002. Intraguild predation diminished in complex-structured vegetation: Implications for prey suppression. Ecology 83: 643-652. Finke, D. L., and W. E. Snyder. 2010. Conserving the benefits of predator biodiversity. Biol. Conserv. 143: 2260-2269. Firth, D. 1993. Bias reduction of maximum-likelihood estimates. Biometrika 80: 27-38. Fleet, R. R., and B. L. Young. 2000. Facultative Mutualism between Imported Fire Ants (Solenopsis invicta) and a Legume (Senna occidentalis). The Southwestern Naturalist 45: 289-298. Folmer, O., M. Black, W. Hoeh, R. Lutz, and R. Vrijenhoek. 1994. DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Mo.l Mar. Biol. Biotechnol. 3: 294-299. Forman, R. T. T. 1995. Some general principles of landscape and regional ecology. Landscape Ecol. 10: 133-142. Fort, G., and S. Lambert-Lacroix. 2005. Classification using partial least squares with penalized logistic regression. Bioinformatics 21: 1104-1111. Foster, S. P., G. Devine, and A. L. Devonshire. 2007. Insecticide Resistance, pp. 261-285. In V. Emden, H. Fritz and R. Harrington (eds.), Aphids as crop pests. CAB International, Cambridge, Massachusetts, USA. Fournier, V., J. Hagler, K. Daane, J. De León, and R. Groves. 2008. Identifying the predator complex of Homalodisca vitripennis (Hemiptera: Cicadellidae): A comparative study of the efficacy of an ELISA and PCR gut content assay. Oecologia 157: 629-640. Frank, S. D., P. M. Shrewsbury, and R. F. Denno. 2011. Plant versus prey resources: Influence on omnivore behavior and herbivore suppression. Biol. Control 57: 229-235. Furlong, M. J. 2015. Knowing your enemies: Integrating molecular and ecological methods to assess the impact of arthropod predators on crop pests. Insect Science 22: 6-19. Gabriel, D., I. Roschewitz, T. Tscharntke, and C. Thies. 2006. Beta diversity at different spatial scales: Plant communities in organic and conventional agriculture. Ecol. Appl. 16: 2011- 2021. Gadau, J. 2009. DNA isolation from ants. Cold Spring Harbor Protocols 2009: pdb. prot5245. Gadau, J., J. Heinze, B. Hölldobler, and M. Schmid. 1996. Population and colony structure of the carpenter ant Camponotus floridanus. Mol. Ecol. 5: 785-792. Galvanho, J. P., M. P. Carrera, D. D. O. Moreira, M. Erthal, C. P. Silva, and R. I. Samuels. 2013. Imidacloprid inhibits behavioral defences of the leaf-cutting ant Acromyrmex subterraneus subterraneus (Hymenoptera: Formicidae). J. Insect Behav. 26: 1-13. Gardiner, M., D. Landis, C. Gratton, N. Schmidt, M. O’Neal, E. Mueller, J. Chacon, and G. Heimpel. 2010. Landscape composition influences the activity density of carabidae and arachnida in soybean fields. Biol. Control 55: 11-19. Gardiner, M., D. Landis, C. Gratton, C. DiFonzo, M. O'neal, J. Chacon, M. Wayo, N. Schmidt, E. Mueller, and G. Heimpel. 2009. Landscape diversity enhances biological control of an introduced crop pest in the north-central USA. Ecol. Appl. 19: 143-154. Gastreich, K. 1999. Trait-mediated indirect effects of a theridiid spider on an ant-plant mutualism. Ecology 80: 1066-1070. Gaucherel, C., F. Burel, and J. Baudry. 2007. Multiscale and surface pattern analysis of the effect of landscape pattern on carabid beetles distribution. Ecological Indicators 7: 598- 609. Gavish-Regev, E., Y. Lubin, and M. Coll. 2008. Migration patterns and functional groups of spiders in a desert agroecosystem. Ecological Entomology 33: 202-212.

172

Gibb, H., and T. Johansson. 2010. Forest succession and harvesting of hemipteran honeydew by boreal ants. Ann. Zool. Fenn. 47: 99-110. Gibbons, D., C. Morrissey, and P. Mineau. 2015. A review of the direct and indirect effects of neonicotinoids and fipronil on vertebrate wildlife. Environ. Sci. Pollut. Res. 22: 103-118. Gibson, R. H., S. Pearce, R. J. Morris, W. O. C. Symondson, and J. Memmott. 2007. Plant diversity and land use under organic and conventional agriculture: A whole-farm approach. J. Appl. Ecol. 44: 792-803. Girolami, V., L. Mazzon, A. Squartini, N. Mori, M. Marzaro, A. d. Bernardo, M. Greatti, C. Giorio, and A. Tapparo. 2009. Translocation of neonicotinoid insecticides from coated seeds to seedling guttation drops: a novel way of intoxication for bees. J. Econ. Entomol. 102: 1808-1815. Godfray, H. C., J. R. Beddington, I. R. Crute, L. Haddad, D. Lawrence, J. F. Muir, J. Pretty, S. Robinson, S. M. Thomas, and C. Toulmin. 2010. Food security: The challenge of feeding 9 billion people. Science 327: 812-818. Godfray, H. C. J., T. Blacquière, L. M. Field, R. S. Hails, G. Petrokofsky, S. G. Potts, N. E. Raine, A. J. Vanbergen, and A. R. McLean. 2014. A restatement of the natural science evidence base concerning neonicotinoid insecticides and insect pollinators. Proceedings of the Royal Society B: Biological Sciences 281. Golden, D. M., and T. O. Crist. 2000. Experimental effects of habitat fragmentation on rove beetles and ants: Patch area or edge? Oikos 90: 525-538. Gomez‐Polo, P., O. Alomar, C. Castañé, J. G. Lundgren, J. Piñol, and N. Agustí. 2014. Molecular assessment of predation by hoverflies (Diptera: Syrphidae) in Mediterranean lettuce crops. Pest Manage. Sci. 71: 1219-1227. Gonthier, D., K. Ennis, S. Philpott, J. Vandermeer, and I. Perfecto. 2013. Ants defend coffee from berry borer colonization. BioControl 58: 815-820. Goulson, D. 2013. Review: An overview of the environmental risks posed by neonicotinoid insecticides. J. Appl. Ecol. 50: 977-987. Gove, A. D., J. D. Majer, and V. Rico-Gray. 2009. Ant assemblages in isolated trees are more sensitive to species loss and replacement than their woodland counterparts. Basic Appl. Ecol. 10: 187-195. Green, J. D., J. R. Martin, L. Murdock, L. Townsend, and P. Vincelli. 2015. Kentucky integrated crop management manual for field crops "alfalfa". Coorperative Extension Service University of Kentucky College of Agriculture, Food and Environment, Lexington, KY, USA. Greene, J., P. Roberts, W. Gardner, F. Reay-Jones, and N. Seiter. 2012. Kudzu bug identification and control in soybeans. United Soybean Board, Chesterfield: 1-11. Greenstone, M., D. Rowley, D. Weber, M. Payton, and D. Hawthorne. 2007. Feeding mode and prey detectability half-lives in molecular gut-content analysis: An example with two predators of the Colorado potato beetle. Bull. Entomol. Res. 97: 201-209. Greenstone, M. H., and K. A. Shufran. 2003. Spider predation: Species-specific identification of gut contents by polymerase chain reaction. J. Arachnol. 31: 131-134. Grez, A. A., T. Zaviezo, and M. M. Gardiner. 2014. Local predator composition and landscape affects biological control of aphids in alfalfa field. Biol. Control 76: 1-9. Halaj, J., D. Ross, and A. Moldenke. 1997. Negative effects of ant foraging on spiders in Douglas-fir canopies. Oecologia 109: 313-322. Harwood, J. D., and J. J. Obrycki. 2005a. Quantifying aphid predation rates of generalist predators in the field. Eur. J. Entomol. 102: 335-350. Harwood, J. D., and J. J. Obrycki. Year. Published. The role of alternative prey in sustaining predator populations, pp. 453-462. In, Proceedings of the Second International Symposium on Biological Control of Arthropods, 2005b. Citeseer. Harwood, J. D., K. D. Sunderland, and W. O. C. Symondson. 2004. Prey selection by Linyphiid spiders: Molecular tracking of the effects of alternative prey on rates of aphid consumption in the field. Mol. Ecol. 13: 3549-3560. Harwood, J. D., N. Desneux, H. J. Yoo, D. L. Rowley, M. H. Greenstone, J. J. Obrycki, and R. J. O'Neil. 2007. Tracking the role of alternative prey in soybean aphid predation by Orius insidiosus: A molecular approach. Mol. Ecol. 16: 4390-4400.

173

Harwood, J. D., S. W. Phillips, J. Lello, K. D. Sunderland, D. M. Glen, M. W. Bruford, G. L. Harper, and W. O. C. Symondson. 2009. Invertebrate biodiversity affects predator fitness and hence potential to control pests in crops. Biol. Control 51: 499-506. Hatley, C. L., and J. A. Macmahon. 1980. Spider community organization - Seasonal-variation and the role of vegetation architecture. Environ. Entomol. 9: 632-639. Hawkes, C. 1973. Factors affecting the aggregation of the adult cabbage root fly (Erioichia brassicae (Bouché)) at hedges. Bull. Entomol. Res. 63: 661-671. Henderson, I. G., J. M. Holland, J. Storkey, P. Lutman, J. Orson, and J. Simper. 2012. Effects of the proportion and spatial arrangement of un‐cropped land on breeding bird abundance in arable rotations. J. Appl. Ecol. 49: 883-891. Henle, K., D. Alard, J. Clitherow, P. Cobb, L. Firbank, T. Kull, D. McCracken, R. F. A. Moritz, J. Niemelä, M. Rebane, D. Wascher, A. Watt, and J. Young. 2008. Identifying and managing the conflicts between agriculture and biodiversity conservation in Europe – A review. Agric. Ecosyst. Environ. 124: 60-71. Herbert, J. J., and D. J. Horn. 2008. Effect of ant attendance by Monomorium minimum (Muckley) (Hymenoptera: Formicidae) on predation and parasitism of the soybean aphid Aphis glycines Matsumura (Hemiptera: Aphididae). Environ. Entomol. 37: 1258-1263. Heterick, B. E., J. Casella, and J. D. Majer. 2002. Influence of Argentine and coastal brown ant (Hymenoptera: Formicidae) invasions on ant communities in Perth gardens, Western Australia. Urban Ecosyst 4: 277-292. Hill, J. H., R. Alleman, D. B. Hogg, and C. R. Grau. 2001. First report of transmission of soybean mosaic virus and alfalfa mosaic virus by Aphis glycines in the new world. Plant Dis. 85: 561-561. Hobbs, P. R. 2007. Conservation agriculture: What is it and why is it important for future sustainable food production? The Journal of Agricultural Science 145: 127-137. Hogervorst, P. A. M., F. L. Wackers, and J. Romeis. 2007. Detecting nutritional state and food source use in field-collected insects that synthesize honeydew oligosaccharides. Functional Ecology 21: 936-946. Holland, J., and L. Fahrig. 2000. Effect of woody borders on insect density and diversity in crop fields: A landscape-scale analysis. Agric. Ecosyst. Environ. 78: 115-122. Holland, J. M., T. Birkett, and S. Southway. 2008. Contrasting the farm-scale spatio-temporal dynamics of boundary and field overwintering predatory beetles in arable crops. BioControl 54: 19-33. Holland, J. M., L. Winder, C. Woolley, C. J. Alexander, and J. N. Perry. 2007. The spatial dynamics of crop and ground active predatory arthropods and their aphid prey in winter wheat. Bull. Entomol. Res. 94: 419-431. Hölldobler, B., and E. Wilson. 1990. The ants, Harvard University Press, Cambridge, MA, USA. Hoogendoorn, M., and G. E. Heimpel. 2001. PCR‐based gut content analysis of insect predators: using ribosomal ITS‐1 fragments from prey to estimate predation frequency. Mol. Ecol. 10: 2059-2067. Hoogendoorn, M., and G. Heimpel. Year. Published. PCR-based gut content analysis of insect predators: a field study, pp. 14-18. In, Proceedings of the 1st International Symposium on Biological Control of Arthropods, 2002. Hooper, D. U., D. E. Bignell, V. K. Brown, L. Brussaard, J. M. Dangerfield, D. H. Wall, D. A. Wardle, D. C. Coleman, K. E. Giller, P. Lavelle, W. H. Van der Putten, P. C. De Ruiter, J. Rusek, W. L. Silver, J. M. Tiedje, and V. Wolters. 2000. Interactions between aboveground and belowground biodiversity in terrestrial ecosystems: Patterns, mechanisms, and feedbacks. Bioscience 50: 1049-1061. Hu, J. S., M. Wang, D. Sether, W. Xie, and K. W. Leonhardt. 1996. Use of polymerase chain reaction (PCR) to study transmission of banana bunchy top virus by the banana aphid (Pentalonia nigronervosa). Ann. Appl. Biol. 128: 55-64. Human, K. G., and D. M. Gordon. 1996. Exploitation and interference competition between the invasive Argentine ant, Linepithema humile, and native ant species. Oecologia 105: 405- 412.

174

Huseth, A. S., and R. L. Groves. 2014. Environmental fate of soil applied neonicotinoid insecticides in an irrigated potato agroecosystem. PloS one 9: e97081. Huxley, C. 1991. Ants and plants: A diversity of interactions, pp. 1-11. In C. R. Huxley and D. F. Cutler (eds.), Ant-plant interactions. Oxford University Press, New York, NY. Isaacs, R., J. Tuell, A. Fiedler, M. Gardiner, and D. Landis. 2009. Maximizing arthropod- mediated ecosystem services in agricultural landscapes: The role of native plants. Frontiers in Ecology and the Environment 7: 196-203. James, D., M. Stevens, and K. O'Malley. 1997. The impact of foraging ants on populations of Coccus hesperidum L.(Hem., Coccidae) and Aonidiella aurantii (Maskell)(Hem., Diaspididae) in an Australian citrus grove. J. Appl. Entomol. 121: 257-259. Johnson, D. 2016. Insecticide recommendations for soybeans‐ 2016. Coorperative Extension Service University of Kentucky College of Agriculture, Food and Environment, Lexington, Kentucky, USA. Johnson, D. W., L. H. Townsend, J. R. Green, W. W. Witt, D. E. Hershman, L. Murdock, and J. Herbek. 2015. Kentucky integrated crop management manual for field crops "soybeans", pp. 1-32. In P. Lucas [ed.], Kentucky IPM Manuals. University of Kentucky, Lexington, Kentucky. Jones, A., P. Harrington, and G. Turnbull. 2014. Neonicotinoid concentrations in arable soils after seed treatment applications in preceding years. Pest Manage. Sci. 70: 1780-1784. Jonsen, I. D., and L. Fahrig. 1997. Response of generalist and specialist insect herbivores to landscape spatial structure. Landscape Ecol. 12: 185-197. Jonsson, M., S. D. Wratten, D. A. Landis, J.-M. L. Tompkins, and R. Cullen. 2010. Habitat manipulation to mitigate the impacts of invasive arthropod pests. Biol. Invasions 12: 2933-2945. Jonsson, M., C. S. Straub, R. K. Didham, H. L. Buckley, B. S. Case, R. J. Hale, C. Gratton, and S. D. Wratten. 2015. Experimental evidence that the effectiveness of conservation biological control depends on landscape complexity. J. Appl. Ecol. 52: 1274-1282. Juen, A., and M. Traugott. 2006. Amplification facilitators and multiplex PCR: Tools to overcome PCR-inhibition in DNA-gut-content analysis of soil-living invertebrates. Soil Biol. Biochem. 38: 1872-1879. Katayama, N., and N. Suzuki. 2003. Bodyguard effects for aphids of Aphis craccivora Koch (Homoptera: Aphididae) as related to the activity of two ant species, Tetramorium caespitum Linnaeus (Hymenoptera: Formicidae) and Lasius niger L. (Hymenoptera: Formicidae). Applied Entology and Zoology 38: 427-433. Katis, N. I., J. A. Tsitsipis, M. Stevens, and G. Powell. 2007. Transmission of Plant Viruses. In V. Emden, H. Fritz and R. Harrington (eds.), Aphids as crop pests. CAB International, Cambridge, Massachusetts, USA. Kéfi, S., E. L. Berlow, E. A. Wieters, S. A. Navarrete, O. L. Petchey, S. A. Wood, A. Boit, L. N. Joppa, K. D. Lafferty, and R. J. Williams. 2012. More than a meal… Integrating non‐feeding interactions into food webs. Ecol. Lett. 15: 291-300. Khrimian, A., P. W. Shearer, A. Zhang, G. Hamilton, and J. R. Aldrich. 2008. Field trapping of the invasive brown marmorated stink bug, Halyomorpha halys, with geometric isomers of Methyl 2,4,6-Decatrienoate. Journal of Agricultural and Food Chemistry 56: 197-203. Kieckhefer, R. W., and B. H. Kantack. 1988. Yield losses in winter grains caused by cereal aphids (Homoptera: Aphididae) in South Dakota. J. Econ. Entomol. 81: 317-321. Kim, C. S., G. Schaible, L. Garrett, R. Lubowski, and D. Lee. 2008. Economic impacts of the U.S. soybean aphid infestation: a multi-regional competitive dynamic analysis. Agricultural and Resource Economics Review 37: 227-242. King, R. A., D. S. Read, M. Traugott, and W. O. Symondson. 2008. Molecular analysis of predation: A review of best practice for DNA-based approaches. Mol. Ecol. 17: 947-963. Koss, A. M., and W. E. Snyder. 2005. Alternative prey disrupt biocontrol by a guild of generalist predators. Biol. Control 32: 243-251. Kremen, C., C. Pease, R. F. Long, and L. Morandin. 2011. Hedgerows enhance beneficial insects on farms in California's Central Valley. California Agriculture 65: 197-201.

175

Krupke, C. H., G. J. Hunt, B. D. Eitzer, G. Andino, and K. Given. 2012. Multiple routes of pesticide exposure for honey bees living near agricultural fields. PloS one 7: e29268. Kunkel, B. A., D. W. Held, and D. A. Potter. 1999. Impact of halofenozide, imidacloprid, and bendiocarb on beneficial invertebrates and predatory activity in turfgrass. J. Econ. Entomol. 92: 922-930. Kunkel, B. A., D. W. Held, and D. A. Potter. 2001. Lethal and sublethal effects of bendiocarb, halofenozide, and imidacloprid on Harpalus pennsylvanicus (Coleoptera: Carabidae) following different modes of exposure in turfgrass. J. Econ. Entomol. 94: 60-67. Kuusk, A.-K., and B. Ekbom. 2012. Feeding habits of lycosid spiders in field habitats. J. Pest Sci. 85: 253-260. Lagerlof, J., and H. Wallin. 1993. The abundance of arthropods along two field margins with different types of vegetation composition: An experimental study. Agric. Ecosyst. Environ. 43: 141-154. Lagnaoui, A., F. Cisneros, J. Alcazar, and F. Morales. 2000. A sustainable pest management strategy for sweetpotato weevil in Cuba: A success story, Food Fertilizer Technology Center, Taiwan. Landis, D. A., S. D. Wratten, and G. M. Gurr. 2000. Habitat management to conserve natural enemies of arthropod pests in agriculture. Annu. Rev. Entomol. 45: 175-201. Landis, D. A., F. D. Menalled, A. C. Costamagna, and T. K. Wilkinson. 2005. Manipulating plant resources to enhance beneficial arthropods in agricultural landscapes. Weed Sci. 53: 902-908. Landis, D. A., M. M. Gardiner, W. van der Werf, and S. M. Swinton. 2008. Increasing corn for biofuel production reduces biocontrol services in agricultural landscapes. Proc. Natl. Acad. Sci. U. S. A. 105: 20552-20557. Lang, A. 2000. The pitfalls of pitfalls: A comparison of pitfall trap catches and absolute density estimates of epigeal invertebrate predators in arable land. J. Pest Sci. 73: 99-106. Lang, A. 2003. Intraguild interference and biocontrol effects of generalist predators in a winter wheat field. Oecologia 134: 144-153. Larson, J. L., C. T. Redmond, and D. A. Potter. 2012. Comparative impact of an anthranilic diamide and other insecticidal chemistries on beneficial invertebrates and ecosystem services in turfgrass. Pest Manage. Sci. 68: 740-748. Law, Y. H., and J. A. Rosenheim. 2011. Effects of combining an intraguild predator with a cannibalistic intermediate predator on a species-level trophic cascade. Ecology 92: 333- 341. Lee, C., and J. Herbek. 2011. Soybean planting in Kentucky. University of Kentucky College of Agriculture, Food and Environment, Lexington, Kentucky, USA. Lee, J. C., F. D. Menalled, and D. A. Landis. 2001. Refuge habitats modify impact of insecticide disturbance on carabid beetle communities. J. Appl. Ecol. 38: 472-483. Lee, K. M., S. Y. Lee, and R. M. Connolly. 2011. Combining stable isotope enrichment, compartmental modelling and ecological network analysis for quantitative measurement of food web dynamics. Methods Ecol. Evol. 2: 56-65. Lefebvre, M., P. Franck, J.-F. Toubon, J.-C. Bouvier, and C. Lavigne. 2016. The impact of landscape composition on the occurrence of a canopy dwelling spider depends on orchard management. Agric. Ecosyst. Environ. 215: 20-29. Leray, M., N. Agudelo, S. C. Mills, and C. P. Meyer. 2013. Effectiveness of annealing blocking primers versus restriction enzymes for characterization of generalist diets: Unexpected prey revealed in the gut contents of two coral reef fish species. PloS one 8: e58076. Leroy, P. D., A. Sabri, S. Heuskin, P. Thonart, G. Lognay, F. J. Verheggen, F. Francis, Y. Brostaux, G. W. Felton, and E. Haubruge. 2011. Microorganisms from aphid honeydew attract and enhance the efficacy of natural enemies. Nature Communications 2: 1-7. Li, W. D., P. J. Zhang, J. M. Zhang, W. C. Lin, Y. B. Lu, and Y. L. Gao. 2015. Acute and sublethal effects of neonicotinoids and pymetrozine on an important egg parasitoid, Trichogramma ostriniae (Hymenoptera: Trichogrammatidae). Biocontrol Sci. Technol. 25: 121-131. Li, X., L. Lu, G. Cheng, and H. Xiao. 2001. Quantifying landscape structure of the Heihe River Basin, north-west China using FRAGSTATS. J. Arid Environ. 48: 521-535.

176

Limay-Rios, V., L. G. Forero, Y. Xue, J. Smith, T. Baute, and A. Schaafsma. 2016. Neonicotinoid insecticide residues in soil dust and associated parent soil in fields with a history of seed treatment use on crops in southwestern Ontario. Environ. Toxicol. Chem. 35: 303-310. Liu, Z., Y. Dai, G. Huang, Y. Gu, J. Ni, H. Wei, and S. Yuan. 2011. Soil microbial degradation of neonicotinoid insecticides imidacloprid, acetamiprid, thiacloprid and imidaclothiz and its effect on the persistence of bioefficacy against horsebean aphid Aphis craccivora Koch after soil application. Pest Manage. Sci. 67: 1245-1252. Lopatina, Y. V., and O. Y. Eremina. 2010. Use of neonicotinoids for the control of pest ants (Insecta, Hymenoptera, Formicidae). Agrokhimiya: 86-93. López, R., and D. A. Potter. 2000. Ant predation on eggs and larvae of the black cutworm (Lepidoptera: Noctuidae) and Japanese beetle (Coleoptera: Scarabaeidae) in turfgrass. Environ. Entomol. 29: 116-125. López, R., and D. A. Potter. 2003. Biodiversity of ants (Hymenoptera: Formicidae) in golf course and lawn turf habitats in Kentucky. Sociobiology 42: 701-713. Losey, J. E., and R. F. Denno. 1998. Positive predator–predator interactions: Enhanced predation rates and synergistic suppression of aphid populations. Ecology 79: 2143- 2152. Losey, J. E., and M. Vaughan. 2006. The economic value of ecological services provided by insects. Bioscience 56: 311-323. Lundgren, J. G. 2009. Relationships of natural enemies and non-prey foods, vol. 7, Springer Science + Business Media B. V., Berlin, Germany. Macfadyen, S., R. H. Gibson, W. O. C. Symondson, and J. Memmott. 2011. Landscape structure influences modularity patterns in farm food webs: Consequences for pest control. Ecol. Appl. 21: 516-524. Madsen, M., S. Terkildsen, and S. Toft. 2004. Microcosm studies on control of aphids by generalist arthropod predators: Effects of alternative prey. BioControl 49: 483-504. Magurran, A. E. 1988. Ecological diversity and its measurement, Princeton University Press, Princeton, NJ, USA. Main, A. R., J. V. Headley, K. M. Peru, N. L. Michel, A. J. Cessna, and C. A. Morrissey. 2014. Widespread use and frequent detection of neonicotinoid insecticides in wetlands of Canada's Prairie Pothole Region. PloS one 9: e92821. Majdi, N., A. Boiche, W. Traunspurger, and A. Lecerf. 2014. Predator effects on a detritus‐ based food web are primarily mediated by non‐trophic interactions. J. Anim. Ecol. 83: 953-962. Malaviya, S., M. Munsi, G. Oinam, and P. K. Joshi. 2010. Landscape approach for quantifying land use land cover change (1972–2006) and habitat diversity in a mining area in Central India (Bokaro, Jharkhand). Environ. Monit. Assess. 170: 215-229. Maloney, D., F. A. Drummond, and R. Alford. 2003. Spider predation in agroecosystems: Can spiders effectively control pest populations? Technical Bulletin - Maine Agricultural and Forest Experiment Station: iv + 28 pp. Marín, L., D. Jackson, and I. Perfecto. 2015. A positive association between ants and spiders and potential mechanisms driving the pattern. Oikos 124: 1078-1088. Marino, P. C., and D. A. Landis. 1996. Effect of landscape structure on parasitoid diversity and parasitism in agroecosystems. Ecol. Appl. 6: 276-284. Marshall, E. J. P., T. M. West, and D. Kleijn. 2006. Impacts of an agri-environment field margin prescription on the flora and fauna of arable farmland in different landscapes. Agric. Ecosyst. Environ. 113: 36-44. Matlock, R. B., and R. de la Cruz. 2003. Ants as indicators of pesticide impacts in banana. Environ. Entomol. 32: 816-829. Mayntz, D., D. Raubenheimer, M. Salomon, S. Toft, and S. J. Simpson. 2005. Nutrient- specific foraging in invertebrate predators. Science 307: 111-113. McCune, B., and J. B. Grace. 2002. Analysis of ecological communities, MjM Software Design, Gleneden Beach, OR, USA.

177

McGarigal, K., E. Ene, and S. Cushman 2012. FRAGSTATS: Spatial pattern analysis program for categorical and continuous maps computer program, version By McGarigal, K., E. Ene, and S. Cushman. McGlynn, T. P. 1999. The worldwide transfer of ants: Geographical distribution and ecological invasions. J. Biogeogr. 26: 535-548. Meehan, T. D., B. P. Werling, D. A. Landis, and C. Gratton. 2011. Agricultural landscape simplification and insecticide use in the midwestern United States. Proc. Natl. Acad. Sci. U. S. A. 108: 11500-11505. Meek, B., D. Loxton, T. Sparks, R. Pywell, H. Pickett, and M. Nowakowski. 2002. The effect of arable field margin composition on invertebrate biodiversity. Biol. Conserv. 106: 259- 271. Mele, P. v., and T. Vo The. 2002. Observations and farmer experimentation with predatory ants. Leisa 18: 28-29. Menalled, F. D., P. C. Marino, S. H. Gage, and D. A. Landis. 1999. Does agricultural landscape structure affect parasitism and parasitoid diversity? Ecol. Appl. 9: 634-641. Messelink, G. J., C. M. J. Bloemhard, M. W. Sabelis, and A. Janssen. 2013. Biological control of aphids in the presence of thrips and their enemies. BioControl 58: 45-55. Mestre, L., J. Piñol, J. A. Barrientos, and X. Espadaler. 2016. Differential ant exclusion from canopies shows contrasting top-down effects on community structure. Oecologia 180: 193-203. Midha, N., and P. K. Mathur. 2010. Assessment of forest fragmentation in the conservation priority Dudhwa landscape, India using FRAGSTATS computed class level metrics. J. Indian Soc. Remote 38: 487-500. Mueller, D., and A. Sisson. 2016. Corn and soybean field guide, ISU Extension and Outreach, Ames, IA, USA. Muilenburg, V. L., F. L. Goggin, S. L. Hebert, L. Jia, and F. M. Stephen. 2008. Ant predation on red oak borer confirmed by field observation and molecular gut-content analysis. Agric. For. Entomol. 10: 205-213. Murcia, C. 1995. Edge effects in fragmented forests - Implications for conservation. Trends Ecol. Evol. 10: 58-62. Myers, C., and E. Hill. 2014. Benefits of neonicotinoid seed treatments to soybean production. United States Environmental Protection Agency Office of Chemical Safety and Pollution Prevention, Washington, DC, USA. Nelson, E. H., C. E. Matthews, and J. A. Rosenheim. 2004. Predators reduce prey population growth by inducing changes in prey behavior. Ecology 85: 1853-1858. Nelson, R. A., R. A. Cramb, K. M. Mienz, and M. A. Mamicpic. 1998. Cost-benefit analysis of alternative forms of hedgerow intercropping in the Philippine uplands. Agroforestry Systems 39: 241-262. North, J. H., J. Gore, A. L. Catchot, S. D. Stewart, G. M. Lorenz, F. R. Musser, D. R. Cook, D. L. Kerns, and D. M. Dodds. 2016. Value of neonicotinoid insecticide seed treatments in mid-south soybean (Glycine max) production systems. J. Econ. Entomol. 109: 1156- 1160. Nyffeler, M., and K. D. Sunderland. 2003. Composition, abundance and pest control potential of spider communities in agroecosystems: A comparison of European and US studies. Agric. Ecosyst. Environ. 95: 579-612. Nyffeler, M., D. A. Dean, and W. L. Sterling. 1987. Evaluation of the importance of the striped lynx spider, Oxyopes salticus (Araneae: Oxyopidae), as a predator in Texas cotton. Environ. Entomol. 16: 1114-1123. Nyffeler, M., D. A. Dean, and W. L. Sterling. 1992. Diets, feeding specialization, and predatory role of two lynx spiders, Oxyopes salticus and Peucetia viridans (Araneae: Oxyopidae), in a Texas cotton agroecosystem. Environ. Entomol. 21: 1457-1465. Nyffeler, M., E. J. Olson, and W. O. Symondson. 2016. Plant-eating by spiders. J. Arachnol. 44: 15-27. O'Neil, R. J., and R. N. Wiedenmann. 1987. Adaptations of arthropod predators to agricultural systems. Fla. Entomol. 70: 40-48.

178

O'Rourke, M. E., K. Rienzo-Stack, and A. G. Power. 2011. A multi-scale, landscape approach to predicting insect populations in agroecosystems. Ecol. Appl. 21: 1782-1791. O’Hara, R. B., and D. J. Kotze. 2010. Do not log‐transform count data. Methods Ecol. Evol. 1: 118-122. Obear, G. R., A. W. Adesanya, P. J. Liesch, R. C. Williamson, and D. W. Held. 2015. Fungicides affect Japanese beetle Popillia japonica (Coleoptera: Scarabaeidae) egg hatch, larval survival and detoxification enzymes. Pest Manage. Sci. 72: 966-973. Offenberg, J. 2001. Balancing between mutualism and exploitation: The symbiotic interaction between Lasius ants and aphids. Behav. Ecol. Sociobiol. 49: 304-310. Ohgushi, T. 2008. Herbivore‐induced indirect interaction webs on terrestrial plants: The importance of non‐trophic, indirect, and facilitative interactions. Entomologia Experimentalis et Applicata 128: 217-229. Östman, Ö., B. Ekbom, and J. Bengtsson. 2003. Yield increase attributable to aphid predation by ground-living polyphagous natural enemies in spring barley in Sweden. Ecological Economics 45: 149-158. Pass, B. C., and J. C. Parr. 1971. Seasonal occurrence of the pea aphid and a Braconid parasite, Aphidius smithi, in Kentucky. J. Econ. Entomol. 64: 1150-1153. Paulson, G. S., and R. D. Akre. 1992. Evaluating the effectiveness of ants as biological-control agents of pear psylla (Homoptera, Psyllidae). J. Econ. Entomol. 85: 70-73. Payton, M. E., M. H. Greenstone, and N. Schenker. 2003. Overlapping confidence intervals or standard error intervals: What do they mean in terms of statistical significance? Journal of Insect Science 3: 34. Pearce, S., and M. P. Zalucki. 2005. Do predators aggregate in response to pest density in agroecosystems? Assessing within-field spatial patterns. J. Appl. Ecol. 43: 128-140. Pecenka, J. R., and J. G. Lundgren. 2015. Non-target effects of clothianidin on monarch butterflies. The Science of Nature 102: 1-4. Peck, S. L., B. Mcquaid, and C. L. Campbell. 1998. Using ant species (Hymenoptera: Formicidae) as a biological indicator of agroecosystem condition. Environ. Entomol. 27: 1102-1110. Peng, R. K., and K. Christian. 2005a. The control efficacy of the weaver ant, Oecophylla smaragdina (Hymenoptera: Formicidae), on the mango leafhopper, Idioscopus nitidulus (Hemiptera: Cicadellidea) in mango orchards in the Northern Territory. Int. J. Pest Manage. 51: 297-304. Peng, R. K., and K. Christian. 2005b. Integrated pest management in mango orchards in the Northern Territory Australia, using the weaver ant, Oecophylla smaragidina, (Hymenoptera : Formicidae) as a key element. Int. J. Pest Manage. 51: 149-155. Peng, R. K., K. Christian, and D. Reilly. 2011. The effect of weaver ants Oecophylla smaragdina on the shoot borer Hypsipyla robusta on African mahoganies in Australia. Agric. For. Entomol. 13: 165-171. Penick, C. A., A. M. Savage, and R. R. Dunn. 2015. Stable isotopes reveal links between human food inputs and urban ant diets. Proc. R. Soc. B-Biol. Sci. 282: 20142608. Penn, H. J. In prep. Field margin composition influences early season pest and spider spatial associations. J. Appl. Ecol. Penn, H. J., and J. D. Harwood. 2016. Detection of DNA in Aphid Honeydew: A Source of Error for Molecular Gut Content Analysis? J. Kans. Entomol. Soc. 89: 85-91. Penn, H. J., and K. J. Athey. In prep. Non-consumptive effects of ants and spiders influence intraguild interactions and subsequent pest damage in soybeans. Bull. Entomol. Res. Penn, H. J., and A. M. Dale. In revision. Imidacloprid seed treatments affect individual ant behavior and community structure but not egg predation, pest abundance, or soybean yield. Pest Manage. Sci. Penn, H. J., E. G. Chapman, and J. D. Harwood. 2016. Overcoming PCR inhibition during DNA- based gut content analysis of ants. Environ. Entomol.: doi: 10.1093/ee/nvw1090. Penn, H. J., K. J. Athey, and B. D. Lee. In review. Land cover diversity increases ant spatial aggregation to prey and consumption Ecol. Lett.

179

Perfecto, I. 1990. Indirect and direct effects in a tropical agroecosystem: The maize-pest-ant system in Nicaragua. Ecology 71: 2125-2134. Perfecto, I. 1991. Ants (Hymenoptera, Formicidae) as natural control agents of pests in irrigated maize in Nicaragua. J. Econ. Entomol. 84: 65-70. Perfecto, I., and A. Castineiras. 1998. Deployment of the predaceous ants and their conservation in agroecosystems, Academic Press, Cambridge, MA, USA. Perry, J. N. 1995. Spatial-analysis by distance indexes. J. Anim. Ecol. 64: 303-314. Perry, J. N., and P. M. Dixon. 2002. A new method to measure spatial association for ecological count data. Ecoscience 9: 133-141. Perry, J. N., L. Winder, J. M. Holland, and R. D. Alston. 1999. Red-blue plots for detecting clusters in count data. Ecol. Lett. 2: 106-113. Perry, J. N., A. M. Liebhold, M. S. Rosenberg, J. L. Dungan, M. Miriti, A. Jakomulska, and S. Citron-Pousty. 2002. Illustrations and guidelines for selecting statistical methods for quantifying spatial pattern in ecological data. Ecography 25: 578-600. Peters, D., J. Webster, and C. Chlouber. 1990. Proceedings. Aphid-plant interactions: Populations to molecules. Agricultural Experiment Station, Oklahoma State University. Pfannenstiel, R. S. 2015. Extended survival of spiders (Aranaeae) feeding on whitefly (Homoptera: Aleyrodidae) honeydew. Journal of Entomological Science 50: 110-118. Pfannenstiel, R. S., and J. M. Patt. 2012. Feeding on nectar and honeydew sugars improves survivorship of two nocturnal cursorial spiders. Biol. Control 63: 231-236. Pfiffner, L., and H. Luka. 2000. Overwintering of arthropods in soils of arable fields and adjacent semi-natural habitats. Agric. Ecosyst. Environ. 78: 215-222. Phillips, B. W., and M. M. Gardiner. 2016. Does local habitat management or large-scale landscape composition alter the biocontrol services provided to pumpkin agroecosystems? Biol. Control 92: 181-194. Philpott, S. M., I. Perfecto, I. Armbrecht, and C. L. Parr. 2010. Ant diversity and function in disturbed and changing habitats, pp. 137-157. In L. Lach, C. Parr and K. Abbott (eds.), Ant Ecology. Oxford University Press, New York, NY, USA. Pimentel, D., U. Stachow, D. A. Takacs, H. W. Brubaker, A. R. Dumas, J. J. Meaney, J. A. O'Neil, D. E. Onsi, and D. B. Corzilius. 1992. Conserving biological diversity in agricultural/forestry systems. Bioscience: 354-362. Pimentel, D., C. Harvey, P. Resosudarmo, K. Sinclair, D. Kurz, M. McNair, S. Crist, L. Shpritz, L. Fitton, R. Saffouri, and R. Blair. 1995. Environmental and economic costs of soil erosion and conservation benefits. Science 267: 1117-1123. Pistorius, J., A. Wehner, M. Kriszan, H. Bargen, S. Knäbe, O. Klein, M. Frommberger, M. Stähler, and U. Heimbach. 2015. Application of predefined doses of neonicotinoid containing dusts in field trials and acute effects on honey bees. Bull. Insectol. 68: 161- 172. Polis, G. A., C. A. Myers, and R. D. Holt. 1989. The ecology and evolution of intraguild predation: Potential competitors that eat each other. Annu. Rev. Ecol. Syst. 20: 297-330. Potts, J. R., T. Hillen, and M. A. Lewis. 2016. The "edge effect" phenomenon: Deriving population abundance patterns from individual animal movement decisions. Theor. Ecol. 9: 233-247. Prasad, R. P., and W. E. Snyder. 2006. Polyphagy complicates conservation biological control that targets generalist predators. J. Appl. Ecol. 43: 343-352. Prasad, R. P., and W. E. Snyder. 2010. A non-trophic interaction chain links predators in different spatial niches. Oecologia 162: 747-753. Preisser, E. L., and D. I. Bolnick. 2008. The many faces of fear: Comparing the pathways and impacts of nonconsumptive predator effects on prey populations. PloS one 3: e2465. Preisser, E. L., D. I. Bolnick, and M. F. Benard. 2005. Scared to death? The effects of intimidation and consumption in predator–prey interactions. Ecology 86: 501-509. Pretorius, R. J. 2014. The effect of agricultural practices on sugar beet root aphid (Pemphigus betae Doane) and beneficial epigeal arthropods. Doctor of Philosophy Dissertation, University of Nebraska-Lincoln Lincoln, Nebraska, USA. Price, P. W. 1976. Colonization of crops by arthropods - Nonequilibrium communities in soybean fields. Environ. Entomol. 5: 605-611.

180

Quisenberry, S. S., and X. Ni. 2007. Feeding Injury. In V. Emden, H. Fritz and R. Harrington (eds.), Aphids as crop pests. CAB International, Cambridge, Massachusetts, USA. Ragsdale, D. W., D. A. Landis, J. Brodeur, G. E. Heimpel, and N. Desneux. 2011. Ecology and management of the soybean aphid in North America. Annu. Rev. Entomol. 56: 375- 399. Ragsdale, D. W., B. P. McCornack, R. C. Venette, B. D. Potter, I. V. Macrae, E. W. Hodgson, M. E. O'Neal, K. D. Johnson, R. J. O'Neil, C. D. Difonzo, T. E. Hunt, P. A. Glogoza, and E. M. Cullen. 2007. Economic threshold for soybean aphid (Hemiptera : Aphididae). J. Econ. Entomol. 100: 1258-1267. Rand, T. A., J. M. Tylianakis, and T. Tscharntke. 2006. Spillover edge effects: The dispersal of agriculturally subsidized insect natural enemies into adjacent natural habitats. Ecol. Lett. 9: 603-614. Rasband, W. S. 2016. ImageJ computer program, version By Rasband, W. S., Bethesda, Maryland, USA. Reynolds, D. R. 1999. Capnodium citri: The sooty mold fungi comprising the taxon concept. Mycopathologia 148: 141-147. Riechert, S., and J. Harp. 1987. Nutritional ecology of spiders, pp. 645-672. In J. Slansky, F. and J. G. Rodriguez (eds.), Nutritional Ecology of Insects, Mites, and Spiders. John Wiley & Sons, New York. Riechert, S. E., and T. Lockley. 1984. Spiders as biological-control agents. Annu. Rev. Entomol. 29: 299-320. Riechert, S. E., and K. Lawrence. 1997. Test for predation effects of single versus multiple species of generalist predators: spiders and their insect prey. Entomologia Experimentalis et Applicata 84: 147-155. Rita, H., and A. Komonen. 2008. Odds ratio: An ecologically sound tool to compare proportions. Ann. Zool. Fenn. 45: 66-72. Rodenhouse, N. L., G. W. Barrett, D. M. Zimmerman, and J. C. Kemp. 1992. Effects of uncultivated corridors on arthropod abundances and crop yields in soybean agroecosystems. Agric. Ecosyst. Environ. 38: 179-191. Rondeau, G., F. Sanchez-Bayo, H. A. Tennekes, A. Decourtye, R. Ramirez-Romero, and N. Desneux. 2014. Delayed and time-cumulative toxicity of imidacloprid in bees, ants and termites. Sci. Rep. 4: 8. Rondoni, G., K. J. Athey, J. D. Harwood, E. Conti, C. Ricci, and J. J. Obrycki. 2014. Development and application of molecular gut‐content analysis to detect aphid and coccinellid predation by Harmonia axyridis (Coleoptera: Coccinellidae) in Italy. Insect Sci. 22: 719-730. Roopa, H., K. Rebijith, R. Asokan, R. Mahmood, and K. K. NK. 2014. Isolation and identification of culturable bacteria from honeydew of whitefly, Bemisia tabaci (G.)(Hemiptera: Aleyrodidae). Meta gene 2: 114-122. Rosenheim, J. A. 2007. Intraguild predation: New theoretical and empirical perspectives. Ecology 88: 2679-2680. Rosenheim, J. A., L. R. Wilhoit, and C. A. Armer. 1993. Influence of intraguild predation among generalist insect predators on the suppression of an herbivore population. Oecologia 96: 439-449. Rosenheim, J. A., H. K. Kaya, L. E. Ehler, J. J. Marois, and B. A. Jaffee. 1995. Intraguild predation among biological-control agents - Theory and evidence. Biol. Control 5: 303- 335. Rozen, S., and H. Skaletsky. 2000. Primer3 on the WWW for general users and for biologist programmers, pp. 365-386. In S. Misener and S. A. Krawetz (eds.), Bioinformatics Methods and Protocols. Humana Press, Totowa, NJ, USA. Rundlof, M., G. K. S. Andersson, R. Bommarco, I. Fries, V. Hederstrom, L. Herbertsson, O. Jonsson, B. K. Klatt, T. R. Pedersen, J. Yourstone, and H. G. Smith. 2015. Seed coating with a neonicotinoid insecticide negatively affects wild bees. Nature 521: 77-80. Rusch, A., R. Chaplin-Kramer, M. M. Gardiner, V. Hawro, J. Holland, D. Landis, C. Thies, T. Tscharntke, W. W. Weisser, C. Winqvist, M. Woltz, and R. Bommarco. 2016.

181

Agricultural landscape simplification reduces natural pest control: A quantitative synthesis. Agric. Ecosyst. Environ. 221: 198-204. Russell, J. A., S. Weldon, A. H. Smith, K. L. Kim, Y. Hu, P. Łukasik, S. Doll, I. Anastopoulos, M. Novin, and K. M. Oliver. 2013. Uncovering symbiont‐driven genetic diversity across North American pea aphids. Mol. Ecol. 22: 2045-2059. Rypstra, A. L., and C. M. Buddle. 2013. Spider silk reduces insect herbivory. Biol. Lett. 9: doi: 10.1098/rsbl.2012.0948. Sánchez-Bayo, F., H. Yamashita, R. Osaka, M. Yoneda, and K. Goka. 2007. Ecological effects of imidacloprid on arthropod communities in and around a vegetable crop. Journal of Environmental Science and Health, Part B 42: 279-286. Sanders, D., and C. Platner. 2007. Intraguild interactions between spiders and ants and top- down control in a grassland food web. Oecologia 150: 611-624. Schaafsma, A., V. Limay-Rios, Y. G. Xue, J. Smith, and T. Baute. 2016. Field-scale examination of neonicotinoid insecticide persistence in soil as a result of seed treatment use in commercial maize (corn) fields in southwestern Ontario. Environ. Toxicol. Chem. 35: 295-302. Schindelin, J., C. T. Rueden, M. C. Hiner, and K. W. Eliceiri. 2015. The ImageJ ecosystem: An open platform for biomedical image analysis. Mol. Reprod. Dev. 82: 518-529. Schmidt, J., J. Harwood, and A. Rypstra. 2012. Foraging activity of a dominant epigeal predator: molecular evidence for the effect of prey density on consumption. Oikos 121: 1715-1724. Schmidt, M. H., I. Roschewitz, C. Thies, and T. Tscharntke. 2005. Differential effects of landscape and management on diversity and density of ground-dwelling farmland spiders. J. Appl. Ecol. 42: 281-287. Schmitz, O. J., A. P. Beckerman, and K. M. O’Brien. 1997. Behaviorally mediated trophic cascades: Effects of predation risk on food web interactions. Ecology 78: 1388-1399. Schneider, C. A., W. S. Rasband, and K. W. Eliceiri. 2012. NIH image to ImageJ: 25 years of image analysis. Nat. Methods 9: 671-675. Schuch, S., C. Platner, and D. Sanders. 2008. Potential positive effect of the ant species Lasius niger on linyphiid spiders. J. Appl. Entomol. 132: 375-381. Schultz, G. W. 1982. Soil-dropping behavior of the pavement ant, Tetramorium caespitum (l.) (Hymenoptera: Formicidae) against the alkali bee (Hymenoptera: Halictidae). J. Kans. Entomol. Soc. 55: 277-282. Schwartzberg, E. G., D. W. Johnson, and G. C. Brown. 2010. The influence of Lasius neoniger (Hymenoptera: Formicidae) on population growth and biomass of Aphis glycines (Hemiptera: Aphididae) in soybeans. Environ. Entomol. 39: 1788-1793. Sheppard, S. K., and J. D. Harwood. 2005. Advances in molecular ecology: Tracking trophic links through predator-prey food-webs. Functional Ecology 19: 751-762. Sigsgaard, L. 2000. The temperature-dependent duration of development and parasitism of three cereal aphid parasitoids, Aphidius ervi, A. rhopalosiphi, and Praon volucre. Entomologia Experimentalis et Applicata 95: 173-184. Simola, D. F., C. Ye, N. S. Mutti, K. Dolezal, R. Bonasio, J. Liebig, D. Reinberg, and S. L. Berger. 2013. A chromatin link to caste identity in the carpenter ant Camponotus floridanus. Genome Res. 23: 486-496. Simon, C., F. Frati, A. Beckenbach, B. Crespi, H. Liu, and P. Flook. 1994. Evolution, weighting, and phylogenetic utility of mitochondrial gene-sequences and a compilation of conserved polymerase chain-reaction primers. Ann. Entomol. Soc. Am. 87: 651-701. Simpson, M., G. M. Gurr, A. T. Simmons, S. D. Wratten, D. G. James, G. Leeson, H. I. Nicol, and G. Orre‐Gordon. 2011. Attract and reward: Combining chemical ecology and habitat manipulation to enhance biological control in field crops. J. Appl. Ecol. 48: 580- 590. Sint, D., I. Thurner, R. Kaufmann, and M. Traugott. 2015. Sparing spiders: Faeces as a non- invasive source of DNA. Front Zool 12: 5. Skillman, L. 2001. Crop rotation important part of farming. University of Kentucky College of Agriculture, Food and Environment, Princeton, Kentucky, USA. Smith, B., and J. B. Wilson. 1996. A consumer's guide to evenness indices. Oikos: 70-82. 182

Smith, C. R., and W. R. Tschinkel. 2009. Collecting live ant specimens (colony sampling). Cold Spring Harbor Protocols 2009: pdb. prot5239. Snyder, W. E., and D. H. Wise. 1999. Predator interference and the establishment of generalist predator populations for biocontrol. Biol. Control 15: 283-292. Snyder, W. E., and D. H. Wise. 2001. Contrasting trophic cascades generated by a community of generalist predators. Ecology 82: 1571-1583. Snyder, W. E., G. B. Snyder, D. L. Finke, and C. S. Straub. 2006. Predator biodiversity strengthens herbivore suppression. Ecol. Lett. 9: 789-796. Stanley, D. A., and N. E. Raine. 2016. Chronic exposure to a neonicotinoid pesticide alters the interactions between bumblebees and wild plants. Functional Ecology 30: 1132-1139. Stapel, J. O., A. M. Cortesero, and W. J. Lewis. 2000. Disruptive sublethal effects of insecticides on biological control: Altered foraging ability and life span of a parasitoid after feeding on extrafloral nectar of cotton treated with systemic insecticides. Biol. Control 17: 243-249. Steffan-Dewenter, I., U. Munzenberg, C. Burger, C. Thies, and T. Tscharntke. 2002. Scale- dependent effects of landscape context on three pollinator guilds. Ecology 83: 1421- 1432. Sterling, W. L., D. Jones, and D. A. Dean. 1979. Failure of the red imported fire ant to reduce entomophagous insect and spider abundance in a cotton agroecosystem. Environ. Entomol. 8: 976-981. Stinner, R. E., C. S. Barfield, J. L. Stimac, and L. Dohse. 1983. Dispersal and movement of insect pests. Annu. Rev. Entomol. 28: 319-335. Straub, C. S., and W. E. Snyder. 2006. Species identity dominates the relationship between predator biodiversity and herbivore suppression. Ecology 87: 277-282. Straub, C. S., and W. E. Snyder. 2008. Increasing enemy biodiversity strengthens herbivore suppression on two plant species. Ecology 89: 1605-1615. Strehl, C. P., and J. Gadau. 2004. Cladistic analysis of paleo-island populations of the Florida harvester ant (Hymenoptera; Formicidae) based upon divergence of mitochondrial DNA sequences. Fla. Entomol. 87: 576-581. Sudd, J. H., and N. R. Franks. 2013. The behavioural ecology of ants, Springer Science & Business Media, Berlin, Germany. Symondson, W. 2002. Molecular identification of prey in predator diets. Mol. Ecol. 11: 627-641. Symondson, W., K. Sunderland, and M. Greenstone. 2002. Can generalist predators be effective biocontrol agents? Annu. Rev. Entomol. 47: 561-594. Takasu, K., and Y. Hirose. 1986. Kudzu-vine community as a breeding site of Ooencyrtus nezarae Ishi (Hymenoptera: Encyrtidae), an egg parasitoid of bugs attacking soybean. Jpn J Appl Entomol Zool 30: 302-304. Takizawa, T., and W. E. Snyder. 2011. Predator biodiversity increases the survivorship of juvenile predators. Oecologia 166: 723-730. Tamburini, G., S. De Simone, M. Sigura, F. Boscutti, and L. Marini. 2016. Conservation tillage mitigates the negative effect of landscape simplification on biological control. J. Appl. Ecol. 53: 233-241. Taylor, R. M., and R. S. Pfannenstiel. 2008. Nectar feeding by wandering spiders on cotton plants. Environ. Entomol. 37: 996-1002. Thaler, J. S., and C. A. M. Griffin. 2008. Relative importance of consumptive and non- consumptive effects of predators on prey and plant damage: The influence of herbivore ontogeny. Entomologia Experimentalis et Applicata 128: 34-40. Thiel, S., and H. R. Kohler. 2016. A sublethal imidacloprid concentration alters foraging and competition behaviour of ants. Ecotoxicology 25: 814-823. Thomas, M., S. Wratten, and N. Sotherton. 1991. Creation of 'island' habitats in farmland to manipulate populations of beneficial arthropods: Predator densities and emigration. J. Appl. Ecol.: 906-917. Thomas, R., J. Harwood, D. Glen, and W. Symondson. 2009. Tracking predator density dependence and subterranean predation by carabid larvae on slugs using monoclonal antibodies. Ecological Entomology 34: 569-579.

183

Thompson, H. M., S. Wilkins, S. Harkin, S. Milner, and K. F. A. Walters. 2015. Neonicotinoids and bumblebees (Bombus terrestris): Effects on nectar consumption in individual workers. Pest Manage. Sci. 71: 946-950. Thompson, J. R., S. Wetzel, M. Klerks, D. Vašková, C. Schoen, J. Špak, and W. Jelkmann. 2003. Multiplex RT-PCR detection of four aphid-borne strawberry viruses in Fragaria spp. in combination with a plant mRNA specific internal control. J. Virol. Methods 111: 85-93. Thomson, L. J., and A. A. Hoffmann. 2013. Spatial scale of benefits from adjacent woody vegetation on natural enemies within vineyards. Biol. Control 64: 57-65. Traugott, M., J. Bell, L. Raso, D. Sint, and W. O. Symondson. 2012. Generalist predators disrupt parasitoid aphid control by direct and coincidental intraguild predation. Bull. Entomol. Res. 102: 239-247. Trible, W., and R. O. N. Carroll. 2014. Manipulating tropical fire ants to reduce the coffee berry borer. Ecological Entomology 39: 603-609. Tscharntke, T., A. M. Klein, A. Kruess, I. Steffan-Dewenter, and C. Thies. 2005. Landscape perspectives on agricultural intensification and biodiversity – Ecosystem service management. Ecol. Lett. 8: 857-874. Uetz, G. W., J. Halaj, and A. B. Cady. 1999. Guild structure of spiders in major crops. J. Arachnol. 27: 270-280. Vichitbandha, P., and D. H. Wise. 2002. A field experiment on the effectiveness of spiders and carabid beetles as biocontrol agents in soybean. Agric. For. Entomol. 4: 31-38. Villesen, P., U. G. Mueller, T. R. Schulz, M. M. Adams, and A. C. Bouck. 2004a. Evolution of ant-cultivar specialization and cultivar switching in apterostigma fungus-growing ants. Evolution; international journal of organic evolution 58: 2252-2265. Villesen, P., U. G. Mueller, T. R. Schultz, R. M. Adams, and A. C. Bouck. 2004b. Evolution of ant‐cultivar specialization and cultivar switching in Apterostigma fungus‐growing ants. Evolution; international journal of organic evolution 58: 2252-2265. Walker, T. J., and S. A. Wineriter. 1981. Marking techniques for recognizing individual insects. The Florida Entomologist 64: 18-29. Wang, D., K. McSweeney, B. Lowery, and J. M. Norman. 1995. Nest structure of ant Lasius neoniger Emery and its implications to soil modification. Geoderma 66: 259-272. Wang, L., L. Zeng, and J. Chen. 2015a. Impact of imidacloprid on new queens of imported fire ants, Solenopsis invicta (Hymenoptera: Formicidae). Sci. Rep. 5: 8. Wang, L., L. Zeng, and J. Chen. 2015b. Sublethal effect of imidacloprid on Solenopsis invicta (Hymenoptera: Formicidae) feeding, digging, and foraging behavior. Environ. Entomol. 44: 1544-1552. Wang, X. 2014. Firth logistic regression for rare variant association tests. Frontiers in Genetics 5: 187. Way, M. J., and K. C. Khoo. 1992. Role of ants in pest management. Annu. Rev. Entomol. 37: 479-503. Weber, D. C., and J. G. Lundgren. 2009. Assessing the trophic ecology of the Coccinellidae: their roles as predators and as prey. Biol. Control 51: 199-214. Weibull, A.-C., J. Bengtsson, and E. Nohlgren. 2000. Diversity of butterflies in the agricultural landscape: The role of farming system and landscape heterogeneity. Ecography 23: 743- 750. Welch, K. D., and J. D. Harwood. 2014. Temporal dynamics of natural enemy–pest interactions in a changing environment. Biol. Control 75: 18-27. Welch, K. D., R. S. Pfannenstiel, and J. D. Harwood. 2012. The role of generalist predators in terrestrial food webs: Lessons for agricultural pest management, Wiley-Blackwell. Wenninger, A., T. N. Kim, B. J. Spiesman, and C. Gratton. 2016. Contrasting foraging patterns: Testing resource-concentration and dilution effects with pollinators and seed predators. Insects 7: doi: 10.3390/insects7020023. Went, F. W., J. Wheeler, and G. C. Wheeler. 1972. Feeding and digestion in some ants (Veromessor and Manica). Bioscience 22: 82-88. Whitcomb, W. H., H. A. Denmark, A. P. Bhatkar, and G. L. Greene. 1972. Preliminary studies on the ants of Florida soybean fields. Fla. Entomol. 55: 129-142.

184

Whitehouse, M. E. A., S. Mansfield, M. C. Barnett, and K. Broughton. 2011. From lynx spiders to cotton: Behaviourally mediated predator effects over four trophic levels. Austral Ecol. 36: 687-697. Wiedenmann, R. N., and J. Smith. 1997. Attributes of natural enemies in ephemeral crop habitats. Biol. Control 10: 16-22. Wiens, J. A., and B. T. Milne. 1989. Scaling of ‘landscapes’ in landscape ecology, or, landscape ecology from a beetle’s perspective. Landscape Ecol. 3: 87-96. Winder, L., C. J. Alexander, J. M. Holland, C. Woolley, and J. N. Perry. 2001. Modelling the dynamic spatio-temporal response of predators to transient prey patches in the field. Ecol. Lett. 4: 568-576. Windor, L., C. J. Alexander, J. M. Holland, W. O. C. Symondson, J. N. Perry, and C. Woolley. 2005. Predatory activity and spatial pattern: The response of generalist carabids and their aphid prey. J. Anim. Ecol. 74: 443-454. Wissinger, S. A. 1997. Cyclic colonization in predictably ephemeral habitats: A template for biological control in annual crop systems. Biol. Control 10: 4-15. Woinarski, J. C., A. N. Andersen, T. B. Churchill, and A. J. Ash. 2002. Response of ant and terrestrial spider assemblages to pastoral and military land use, and to landscape position, in a tropical savanna woodland in northern Australia. Austral Ecol. 27: 324-333. Woodell, S. R. J., T. J. King, C. R. Huxley, and D. F. Cutler. 1991. The influence of mound- building ants on British lowland vegetation, Oxford University Press. Wratten, S. D., M. H. Bowie, J. M. Hickman, A. M. Evans, J. R. Sedcole, and J. M. Tylianakis. 2003. Field boundaries as barriers to movement of hover flies (Diptera: Syrphidae) in cultivated land. Oecologia 134: 605-611. Wyckhuys, K. A. G., R. L. Koch, and G. E. Heimpel. 2007. Physical and ant-mediated refuges from parasitism: Implications for non-target effects in biological control. Biol. Control 40: 306-313. Xu, T. B., D. G. Dyer, L. L. McConnell, S. Bondarenko, R. Allen, and O. Heinemann. 2016. Clothianidin in agricultural soils and uptake into corn pollen and canola nectar after multiyear seed treatment applications. Environ. Toxicol. Chem. 35: 311-321. Yang, C., C. Hamel, V. Vujanovic, and Y. Gan. 2011. Fungicide: modes of action and possible impact on nontarget microorganisms. ISRN Ecology 2011. Yao, F.-L., Y. Zheng, J.-W. Zhao, N. Desneux, Y.-X. He, and Q.-Y. Weng. 2015. Lethal and sublethal effects of thiamethoxam on the whitefly predator Serangium japonicum (Coleoptera: Coccinellidae) through different exposure routes. Chemosphere 128: 49-55. Yeh, C.-T., and S.-L. Huang. 2009. Investigating spatiotemporal patterns of landscape diversity in response to urbanization. Landscape Urban Plann. 93: 151-162. Young, O. P. E., G.B. 1990. Spiders in United States field crops and their potential effect on crop pests. J. Arachnol. 18: 1-27. Youngsteadt, E., R. C. Henderson, A. M. Savage, A. F. Ernst, R. R. Dunn, and S. D. Frank. 2014. Habitat and species identity, not diversity, predict the extent of refuse consumption by urban arthropods. Global Change Biology 21: 1103-1115. Zaller, J., D. Moser, T. Drapela, C. Schmoger, and T. Frank. 2008. Effect of within-field and landscape factors on insect damage in winter oilseed rape. Agric. Ecosyst. Environ. 123: 233-238. Zenger, J., and T. Gibb. 2001a. Identification and impact of egg predators of Cyclocephala lurida and Popillia japonica (Coleoptera: Scarabaeidae) in turfgrass. Environ. Entomol. 30: 425- 430. Zenger, J., and T. Gibb. 2001b. Impact of four insecticides on Japanese beetle (Coleoptera: Scarabaeidae) egg predators and white grubs in turfgrass. J. Econ. Entomol. 94: 145- 149. Zhang, W., T. H. Ricketts, C. Kremen, K. Carney, and S. M. Swinton. 2007. Ecosystem services and dis-services to agriculture. Ecological Economics 64: 253-260. Zhang, Z., X. Zhang, F. Liu, and W. Mu. 2015. Insecticide susceptibility of the green plant bug, Apolygus lucorum Meyer-Dur (Homoptera: Miridae) and two predatory arthropods. J Plant Prot Res 55: 362-370.

185

Zhao, Z.-H., C. Hui, D.-H. He, and F. Ge. 2013. Effects of position within wheat field and adjacent habitats on the density and diversity of cereal aphids and their natural enemies. BioControl 58: 765-776. Zhu, Y. C., J. Adamczyk, T. Rinderer, J. Yao, R. Danka, R. Luttrell, and J. Gore. 2015. Spray toxicity and risk potential of 42 commonly used formulations of row crop pesticides to adult honey bees (Hymenoptera: Apidae). J. Econ. Entomol. 108: 2640-2647. Zientz, E., H. Feldhaar, S. Stoll, and R. Gross. 2005. Insights into the microbial world associated with ants. Arch. Microbiol. 184: 199-206.

186

VITA

Name: Hannah J. Penn (née McKenrick); orcid.org/0000-0002-3692-5991 Affiliation: Department of Entomology, University of Kentucky

Education: Graduate Certificate, Applied Statistics: UK, Lexington, KY; Completed: May 2013 BS, Entomology and Nematology (Cum laude): University of Florida (UF); Completed: May 2011

Awards:  2nd place PhD 10-minute talk competition, North Central Branch of the Entomological Society of America (NCB-ESA), 2016  International PEO Scholar Award Nomination, KY AM Chapter, 2014  2nd place PhD 10-minute talk competition, Entomological Society of America (ESA), 2013  1st place PhD Poster competition, NCB-ESA, 2013  Daniel R. Reedy Quality of Achievement Award, University of Kentucky, 2011, 2012, 2013

Scholarships and Fellowships:  UK Graduate School Academic Year Fellowship. 2016. $20,000 plus tuition  Kentucky Opportunity Fellowship. 2015. $15,000  UK Graduate School Travel Grant. 2012, 2013, 2014, 2015. $400  NCB-ESA Presidential Student Travel Scholarship. 2013, 2015, 2016. $250  NSF Graduate Research Fellowship. 2011. $30,000 plus tuition  UF College of Ag and Life Sciences Scholarship. 2009, 2010. $500  UF Entomology Club Scholarship. 2009, 2010. $500  Florida Bright Futures Scholarship. 2008, 2009, 2010. Tuition  Robert Byrd Scholarship. 2008, 2009, 2010. $3,500

Grants:  UK Student Sustainability Council (SSC), Science and Business of Insects, 2015, ($1658)  Southern SARE Graduate Student Grant, Evaluation of soybean field border management on immigration of insect pests and arthropod predator abundance, 2013, 2014  Kentucky Soybean Board, Impact of border vegetation on soybean pest abundance and movements, 2013, 2014

187

Professional Service:  International Congress of Entomology (ICE), abstract editor, student volunteer, 2016  UK Dept. of Entomology, Faculty Search Committee, 2015  UK SSC: Director of Development, 2015-2016; at-large member 2014-present  Ohio Valley Entomological Association (OVEA): Governing Board Member 2015-2016; President, 2014-2015; President-elect 2013-2014  NCB-ESA Student Activities Committee: Treasurer, 2015–present; Linnaean Games Committee Student Representative, 2014-2016  ESA Annual Meeting: Student volunteer, 2013; Student presentation judge, 2014  UK Entomology H. Garman Entomology Club: Vice President, 2014-2015; Treasurer, 2011-2013; Honey Bee Committee, 2011-2016  UF Entomology Club: Webmaster, 2009-2011  Reviewer of 6 manuscripts from Biological Control (2), Environmental Entomology (2), Biotropica (1), and The Florida Entomologist (1)

Public Outreach: • Bugs, bugs, bugs, Insect Presentation, 2016 • Watershed Watch Kentucky: Phase 1 certified sampler, 2016 • Living Arts & Science Center Discovery Night, 2016 • Interview on Green Talks, WRFL 88.1, 2015 • Pest Control Short Course volunteer, 2015 • Floracliff and Maywoods Bioblitz volunteer, 2015 • Science Fair Judge (grades K-12), Kentucky American Water Science Fair, 2012, 2015 • Science Fair Judge (grades 6-8), Keystone Heights Science Fair, 2011 • Explorium of Lexington, Insect Presentations, 2011-2014 • Bug Walk, UK Arboretum, 2013, 2014, 2016 • Scott County, KY Library, Insect Presentation, 2013 • Caterpillar Outreach (grades K-12) Insect Presentations, 2011 • UF Bug Fest Insect Festival, 2011 • Gator Encounter (grades 9-12) Insect Presentation, 2011 • IDEAL Fair (grades 9-12) Insect Presentation, 2011 • Terwilliger Elementary School (grades 1-2) Insect Presentations, 2010

Professional Societies: • Kansas Entomological Society, 2015-present • Kentucky Academy of Science, 2014-present • Ohio Valley Entomological Association, 2013-present • UK Chapter, Gamma Sigma Delta, Honor Society of Agriculture, 2013- present • International Organization of Biological Control, 2012-present • Entomological Society of America, 2003-present • Florida Entomological Society, 2006-2011

188

Research and Extension Publications: Penn, HJ, KJ Athey, and BD Lee. Land cover diversity increases predator spatial aggregation to prey and consumption. Ecology Letters (in review). Penn, JM, HJ Penn, MF Potter, and W Hu. 2016. Bed bugs and hotels: Traveler insights and implications for the industry. American Entomologist (in review). Penn, HJ and AM Dale. 2016. Imidacloprid seed treatments affect individual ant behavior and community structure but not egg predation, pest abundance, or soybean yield. Pest Management Science (in revision). Penn, HJ, EG Chapman, and JD Harwood. 2016. Overcoming PCR Inhibition for Ant DNA Gut Content Analysis. Environmental Entomology (doi: 10.1093/ee/nvw090). Athey, KJ, J Dreyer, KA Kowles, HJ Penn, MI Sitvarin, and JD Harwood. 2016. Spring Forward: Molecular detection of early season predation in agroecosystems. Food Webs (doi:10.1016/j.fooweb.2016.06.001). Penn, HJ and JD Harwood. 2016. Detection of DNA in aphid honeydew: A source of error for molecular gut content analysis? Journal of the Kansas Entomological Society, 89(1), 85-91. Penn, HJ. 2015. Ants in Kentucky landscapes and gardens. UK EntFact-458.

Research Presentations: Invited Seminars Penn, HJ. 2015. Landscape effects on ant food webs in agroecosystems. Department of Biology Eco-Lunch Seminar, UK, Lexington, KY. Penn, HJ. 2014. Methods, Mighty to Miniscule: Landscapes to Guts. Plant Science GSO Seminar, South Dakota State University, Brookings, SD.

Conference and Extension Oral Presentations Penn, JM, HJ Penn, and Wuyang Hu. 2016. Public attitudes on monarch conservation. OVEA, Lafayette, IN. Penn, HJ. 2015. Multi-scale land use effects on ant predation in agroecosystems. International Congress of Entomology, Orlando, FL. Penn, HJ. 2016. Influence of neonicotinoid seed treatments and field margins on Kentucky soybean pests. Corn and soybean field day, UK REC, Princeton, KY. Penn, HJ, AM Dale, and JD Harwood. 2016. Influence of neonicotinoid seed treatments on ant communities and predation services in soybeans. NCB- ESA, Cleveland, OH. Penn, HJ, KJ Athey, and JD Harwood. 2015. Direct and indirect interactions of ants, spiders, and soybean pests. ESA, Minneapolis, MN. Dale, A, HJ Penn, and JD Harwood. 2015. Effects of neonicotinoid seed treatments on ant behavior and survival. OVEA, Lexington, KY. Penn, HJ and JD Harwood. 2015. Effects of land use and field margins on ant community composition in agroecosystems. Ecological Society of America, Baltimore, MD.

189

Penn, HJ, EG Chapman, and JD Harwood. 2015. Delineating ant food webs: Overcoming PCR Inhibition. NCB-ESA, Manhattan, KS. Penn, HJ and JD Harwood. 2014. Field border vegetation alters Kentucky soybean pest immigration and movements. ESA, Portland, OR; OVEA, Columbus, OH. Kowles, KA, HJ Penn, DW Johnson, and JD Harwood. 2014. DNA detection methods in aphid honeydew: Implications for wheat biological control. ESA, Portland, OR. Penn, HJ and JD Harwood. 2014. Land use and field edges drive movements of spiders in soybean agroecosystems. NCB-ESA, Des Moines, IA. Penn, HJ, KA Kowles and JD Harwood. 2013. Spatial relationships between ants, prey and border vegetation in a soybean agroecosystem. ESA, Austin, TX; OVEA, Indianapolis, IN. Penn, HJ and JD Harwood. 2013. Impact of landscape heterogeneity on ant richness and evenness in an agroecosystem. Virginia Tech Forest Resources and Environmental Conservation Graduate Student Research Symposium, Blacksburg, VA. McKenrick, HJ, SM Wilder, ST Behmer, RE Gold, and MD Eubanks. 2009. The attraction of red imported fire ants (Solenopsis invicta) to soft baits that vary in nutritional content. ESA, Indianapolis, IN.

Poster Presentations Penn, HJ and JD Harwood. 2013. Impacts of potential prey availability on within field movement of spiders in agroecosystems. NCB-ESA, Rapid City, SD. Penn, HJ and JD Harwood. 2013. Using DGGE to delineate prey consumption patterns in a generalist ant (Formicidae: Tetramorium caespitum) colony. Second International Symposium on the Molecular Detection of Trophic Interactions, Lexington, KY. Penn, HJ and JD Harwood. 2012. Impacts of landscape heterogeneity on ant community composition in soybean fields. ESA, Knoxville, TN. McKenrick, HJ, R Bethel, and J Resasco. 2011. Comparison of ant species richness and evenness in Savanna river plant site, South Carolina, from 1976 to 2010. UF Undergraduate Research Assistantship Program Symposium, Gainesville, FL; ESA, Reno, NV. Harwood, JD, CD Allen, EG Chapman, KJ Johansen, KA Kowles, HJ McKenrick, JA Peterson, JM Schmidt, KD Welch, and TD Whitney. 2011. Disentangling the spider’s web: Insights from complex terrestrial ecosystems. ESA, Reno, NV. Eubanks, MD, HJ McKenrick, L Campbell, and SM Wilder. 2011. Potential efficacy of a novel bait for control of the red imported fire ant. Imported Fire Ant Conference, Galveston, TX. McKenrick, HJ, SM Wilder, ST Behmer, RE Gold, and MD Eubanks. 2009. The attraction of red imported fire ants (Solenopsis invicta) to soft baits that vary in nutritional content. TAMU Undergraduate Research Symposium, College Station, TX.

190

McKenrick, HJ and OE Liburd. 2007. Behavior and susceptibility of blueberry leaf beetles, Colaspis pseudofavosa, to conventional and reduced-risk insecticides. UF Student Science Training Program Symposium, Gainesville, Fl.

191