The Pennsylvania State University

The Graduate School

College of Agricultural Sciences

USING POPULATION STRUCTURE AND PHENOLOGY TO ADVANCE INSECT

MANAGEMENT IN DIVERSIFIED VEGETABLE AGROECOSYSTEMS

A Dissertation in

Entomology

Amanda C. Bachmann

 2012 Amanda C. Bachmann

Submitted in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

December 2012

The dissertation of Amanda C. Bachmann was reviewed and approved* by the following:

Shelby J. Fleischer Professor of Entomology Dissertation Advisor Chair of Committee

Michael Saunders Professor of Entomology

John Tooker Assistant Professor of Entomology

Douglas Miller Associate Professor of Geography

Andrew Michel Assistant Professor of Entomology The Ohio State University Special Member

Gary Felton Department Head, Professor of Entomology

*Signatures are on file in the Graduate School

iii

ABSTRACT

Population structure and phenology can both be used to advance insect management in diverse vegetable agroecosystems. I describe the alate aphid species composition in Northeastern

US processing snap beans and update the list of aphid species found in Pennsylvania. The alate aphid community in Pennsylvania snap bean fields is diverse and contains members that are efficient vectors of economically important plant viruses. One of these aphids, Aphis glycines, is largely present in the state as a result of migration and dispersal from areas with high densities of its overwintering host, Rhamnus cathartica. Using genetic tools and air-flow trajectory models to investigate the natal sources of A. glycines in Pennsylvania, I found A. glycines populations sampled in Pennsylvania had high levels of genotypic diversity. This is indicative of being sourced from many natal populations. They were also genetically similar to some populations in the Midwest. Matching A. glycines clones were found between PA, NY and VA indicating some level of long distance movement, which I attempted to model using air-flow trajectories. In addition to working with aphids, I set out to validate early season activity and growing season phenology models for three pests of cucurbits on land that is transitioning to organic production on research farms in Pennsylvania, Iowa and Kentucky. Modeling the phenology of the striped cucumber beetle and squash bug was challenging due to the fact that they overwinter as adults. I was able to demonstrate a successful early season activity monitoring tool, and used air- temperature degree-days from weather stations and development data from previous studies on both insects to estimate the number of field generations and discuss challenges for their control. I also described the phenology of the squash vine borer in geographic areas not represented in previous studies.

TABLE OF CONTENTS

Abstract ……………………………………………………………………………………...…iii Table of Contents ………………………………………………………………………………iv List of Figures ………………………………………………………………………..…..…….vi List of Tables …………………………………………………………………………………..ix Acknowledgements …………………………………………………………………..….....…..xi

Chapter 1 Introduction ...... 1

Aphids ...... 4 Aphids as vectors of plant viruses ...... 7 Aphids and snap beans ...... 8 Introduced species ...... 9 Insect movement ...... 10 Phenology and pest management ...... 13 Dissertation Objectives ...... 16 References ...... 17

Chapter 2 Alate aphid species composition in Northeastern US processing snap beans and an update to historical lists ...... 20

Introduction ...... 20 Methods ...... 24 Slide mounting protocol ...... 25 Results ...... 25 Discussion ...... 27 References ...... 28 Figures and Tables ...... 31

Chapter 3 Estimating natal sources of Aphis glycines using molecular markers and airflow trajectories ...... 43

Introduction ...... 43 Aphis glycines life history ...... 43 Factors limiting the range of Aphis glycines ...... 44 Aphis glycines in Pennsylvania ...... 46 Soybean management ...... 46 Aerobiology ...... 47 Molecular tools for population identification ...... 49 Objectives ...... 50 Methods ...... 50 Field collection ...... 50 Laboratory Methods ...... 51 Statistical Methods ...... 51 HYSPLIT methods ...... 52 Results ...... 53 Spatial ...... 53 Temporal ...... 55

Aerobiology ...... 56 Discussion ...... 56 References ...... 58 Figures and Tables ...... 62

Chapter 4 Speciation and population structure of Therioaphis trifolii ...... 74

Introduction ...... 74 Methods ...... 75 Results ...... 75 Discussion ...... 76 References ...... 76 Figures and Tables ...... 78

Chapter 5 Phenology model validation of pests of cucurbits ...... 80

Introduction ...... 80 Methods ...... 83 Early season activity...... 83 In-season phenology ...... 85 Squash Vine Borer ...... 86 Meteorological Data ...... 86 Results ...... 88 Striped Cucumber Beetle ...... 88 Squash Bug ...... 89 Squash Vine Borer ...... 90 Discussion ...... 90 References ...... 94 Figures and Tables ...... 96

Chapter 6 Conclusions ...... 111

References ...... 119 Appendix HYSPLIT Screenshots ...... 121

LIST OF FIGURES

Figure 2-1. Individual-based rarefaction curves showing aphid species accumulation in PA and NY...... 41

Figure 2-2. Individual-based rarefaction curve showing aphid species accumulation from the combining of samples from PA and NY (solid line) and the 95% confidence intervals for the curve (dashed lines)...... 42

Figure 2-3. Proportion of aphids from the pan trapping collection in PA and NY that use herbaceous plants, trees, or crops as primary hosts. Host associations for North America characterized from Blackman and Eastop (1994, 2000, and 2006)...... 42

Figure 3-1. Map of A. glycines collection locations. Sites with a black circle were used in 2009. Sites with a black diamond were used in 2010. Sites with a black star are 2009 collections from Orantes et al (2012). Rock Springs was used in both years...... 68

Figure 3-2. Principal component analysis based on Fst of the aphid populations collected in Pennsylvania, New York, and Virginia 2009 showing spatial (primary axis) and temporal (secondary axis) differentiation. The dotted lines group the populations that separated spatially...... 69

Figure 3-3. Principal component analysis based on Fst of the aphid populations collected in Pennsylvania, New York, and Canada in 2010 showing spatial (primary axis) and temporal (secondary axis) differentiation. The dotted lines group the populations that separated spatially...... 70

Figure 3-4. Principal component analysis based on Fst of populations collected in Pennsylvania, New York, Virginia, and the Midwest in 2009 showing spatial (primary axis) and temporal (secondary axis) differentiation. The dotted lines group the populations that separated spatially...... 71

Figure 3-5. Principal component analysis based on Fst of populations collected in Pennsylvania, New York and Canada in 2010 and the Midwest sites 2009 showing spatial (primary axis) and temporal (secondary axis) differentiation. The dotted lines group the populations that separated spatially...... 72

Figure 3-6. Examples of HYSPLIT forward trajectory maps over a 48 hour time period, clockwise from top left; PA to VA 7/13/2009 (score 0.67), NY to VA 7/8/2009 (score 0.11), PA to NY 7/10/2009 (score of 0.52), and PA 7/12/2009 (score of 0)...... 73

Figure 4-1. The relationship between haplotypes derived using the LR primer for T. trifolii collected in Pennsylvania and Serbia. In the web, each line as well as each small dot is one base pair of difference...... 78

Figure 4-2. The relationship between haplotypes derived using the SR primer for T. trifolii collected in Pennsylvania and Serbia. In the web, each line as well as each small dot is one base pair of difference...... 79

vii

Figure 5-1. Gompertz curve showing early season recruitment predictions for PA and KY 2010 – 11 using cumulative degree days base 55 F for Acalymma vittatum...... 99

Figure 5-2. Gompertz curve showing early season recruitment predictions for PA 2010 - 2012 using calendar day for Acalymma vittatum...... 100

Figure 5-3. Gompertz curve showing early season recruitment predictions for KY 2010 - 2012 using calendar day for Acalymma vittatum...... 101

Figure 5-4. Average number of A. vittatum per plant during the growing season for PA 2011 with overlay showing the biofix and projected development times for the first and second field generations ...... 102

Figure 5-5. Average number of A. vittatum per plant during the growing season for Kentucky with overlay showing the biofix and projected development ...... 103

Figure 5-6. Average number of A. vittatum per plant during the growing season for Iowa with overlay showing the biofix and projected development...... 104

Figure 5-7. Average number of squash bug adults, juveniles and egg masses per plant during the growing season for Pennsylvania. The horizontal bar shows the biofix and projected egg to adult development time, and the solid vertical lines indicate the critical photperiod for diapause induction...... 105

Figure 5-8. Average number of squash bug adults, juveniles and egg masses per plant during the growing season for Kentucky. The horizontal bar shows the biofix and projected egg to adult development time, and the two vertical lines indicate the timing of the critical photperiod for diapause induction...... 106

Figure 5-9. Average number of A. tristis adults, juveniles and egg masses per plant during the growing season for Iowa. The horizontal bar shows the biofix and projected egg to adult development time, and the two vertical lines indicate the timing of the critical photperiod for diapause induction...... 107

Figure 5-10. Observed and predicted accumulation of A. vittatum in Pennsylvania during the growing season. Predicted lines were generated using the equation y = 0.9913*exp(-exp(3.3906 – 0.014*x)) where x is cumulative degree days base 55 F. The division between first and second field generation was estimated as 793.6 cumulative degree days base 55 F after the biofix...... 108

Figure 5-11. Observed and predicted accumulation of A. vittatum in Iowa during the growing season. Predicted lines were generated using the equation y = 0.9913*exp(- exp(3.3906 – 0.014*x)) where x is cumulative degree days base 55 F. The division between first and second field generation was estimated as 793.6 dd55 after the biofix...... 109

viii

Figure 5-12. Observed and predicted accumulation of A. vittatum in Kentucky during the growing season. Predicted lines were generated using the equation y = 0.9913*exp(- exp(3.3906 – 0.014*x)) where x is cumulative degree days base 55 F. The division between first and second field generation was estimated as 793.6 dd55 after the biofix...... 110

LIST OF TABLES

Table 2-1. Alate aphid species representing > 1 % of the capture from water pan traps in commercial snap bean fields in PA (2004 – 2006) and NY (2002 – 2006). Derived from Table 1 in Nault et al (2009)...... 31

Table 2-2. Species of alate aphids with host associations, collected from water pan traps in commercial snap bean fields in PA (2004 – 2006) and NY (2002 – 2006), and from similar traps in orchards in central PA (2003-2004, Wallis et al. 2005). Host associations for North America from Blackman and Eastop (1994 [AWT], 2000 [AWC], and 2006 [HPS])...... 32

Table 2-3. New aphid records from PA reported in Nault et al (2009) and/or Wallis et al (2005) but not found in Pepper (1965)...... 35

Table 2-4. Species in six subfamilies of the family Aphididae occurring in PA...... 35

Table 2-5. Species in the subfamily Aphidinae, tribe Macrosiphini occurring in PA ...... 36

Table 2-6. Species in the subfamily Aphidinae, tribe Aphidini occurring in PA ...... 38

Table 2-7. Species in the subfamily Calaphidinae occurring in PA ...... 39

Table 2-8. Species in the subfamily Chaitophorinae occurring in PA...... 39

Table 2-9. Species in the subfamily Drepanosiphinae occurring in PA ...... 39

Table 2-10. Species in the subfamily Eriosomatinae occurring in PA ...... 40

Table 2-11. Species in the subfamily Lachninae occurring in PA ...... 40

Table 3-1. Aphid collection dates and locations for 2009 and 2010...... 62

Table 3-2. Fst values for Pennsylvania, Virginia and New York 2009. None of the Fst values were significant at P < 0.05...... 62

Table 3-3. Fst values for Pennsylvania (all collection dates combined), Virginia and New York 2009. Significant values (P < 0.05) indicated in bold...... 62

Table 3-4. Genotypic diversity for aphid populations collected in Pennsylvania, Virginia, and New York 2009...... 63

Table 3-5. Fst values for Pennsylvania, Virginia, New York and Midwest sites 2009. Significant values (P < 0.05) indicated in bold...... 64

Table 3-6. Fst values for Pennsylvania, Canada, New York (2010) and Midwest sites 2009. Significant values (P < 0.05) indicated in bold...... 65

Table 3-7. Fst values for PA, NY, and sites in Canada from 2010 collections. Significant values (P < 0.05) indicated in bold...... 66

Table 3-8. Genotypic diversity for aphid populations collected in Pennsylvania, Canada, and New York 2010...... 66

Table 3-9. Score from 0 to 1 of forward trajectories that cross the target location from the HYSPLIT maps. 142 date/location scenarios were evaluated and dates where none of the trajectories crossed the target location (score of 0) are not shown on this table. Maps were generated from July 1-31 2009 for PA to VA and PA to NY, and July 1 – August 9 2009 for NY to VA and NY to PA...... 67

Table 5-1: Early season recruitment monitoring dates and locations for 2010 - 2012 in KY, IA, and PA...... 96

Table 5-2: Summary of life history parameters for Acalymma vittatum, Anasa tristis and Melittia cucurbitae from the literature (see footnotes) and this study (*)...... 96

Table 5-3. Mean calendar day and cumulative degree day of first capture for Acalymma vittatum (SCB) and Anasa tristis (SB) on trap flats. N is the number of years...... 97

Table 5-4. Mean calendar day and cumulative degree day of first capture for Acalymma vittatum (SCB) and Anasa tristis (SB) in the phenology plot. N is the number of years...... 97

Table 5-5. Mean calendar day and cumulative degree day base 50 F of first capture for Melittia cucurbitae ...... 97

Table 5-6. Parameter estimates for the Gompertz equations modeling early season recruitment of Acalymma vittatum to trap flats using degree day...... 98

Table 5-7. Parameter estimates for the Gompertz equations modeling early season recruitment of Acalymma vittatum to trap flats using calendar day...... 98

ACKNOWLEDGEMENTS

I would like to thank my advisor Shelby Fleischer and my committee; Andy Michel,

Doug Miller, Mike Saunders, and John Tooker for their guidance and advice during my time at

Penn State. Special thanks go to Andy Michel and his lab for their hospitality and assistance for a significant portion of my soybean aphid data analysis efforts.

Thanks also go out to my officemates (Lori Shapiro and Eric Bohnenblust), and assorted entoids (Kerry Mauck, Sean Halloran, Alexis Barbarin, Christy Harris, and Emily Kuhns) for their friendship and support. A special shout-out goes to Tracy Conklin, my entomology sister who shared her love for Canadian television, gardening, hatchbacks, and culinary adventures with me. Ride forever.

Thank you to the Department of Entomology and the faculty and staff therein for supporting me academically and financially.

The support of my family was integral to my completion of this degree and I would like to thank; my husband, Josh, my parents, Darlene and Clarence, my sister Jaime and her husband

Doug, Uncle Les, Aunt Ardie and Uncle Roe, Megan and Pat, and Dee for their support.

This dissertation is dedicated to my late grandmother, Frances Keller, who always asked me if I was still “working with bugs.”

Chapter 1

Introduction

Pennsylvania is part of the diverse forested Northeastern landscape. Farms in this region compete with urbanization for deforested land, resulting in a varied patchwork of land-use patterns. Land-use patterns on these farms are dominated by field crops, but vegetable crops

(both fresh market and processing) are important. For example, the state is one of the top ten producers of fresh market sweet corn ($35.5 million in 2009), processing snap beans, fresh- market peppers, fresh-market tomatoes, pumpkins, and fresh-market squash. Surveys representing the approximately 4,300 vegetable farms reporting in the state census reveal from 5 to 20 crops on a given farm, and often multiple cultivars and planting dates of several of the crop species. Generally, these farms are small (median farm size in Pennsylvania is 65 acres) and comprised of multiple small fields (USDA NASS, 2009). Furthermore, both fresh and processing markets are relevant. Fresh market vegetables are integral to the local economy in many areas of

Pennsylvania in the form of local farm stands, farmers’ markets, or community supported agriculture organizations. The majority of farmers utilize some form of fresh-marketing outlet, and within this sector, organic production is increasing.

Pest management in vegetable crops is challenging due to this diversity of crops, the diversity of pests that can affect them, and the diversity of marketing systems. Vegetable crops in diverse landscapes are particularly challenging because they are often not the dominant crop and thus subject to control measures around them. For example, European corn borer (Ostrinia nubilalis) is predominantly a pest of corn, but can also damage green peppers. If surrounding corn

2 fields are inadequately managed or inhospitable to ECB (in the case of Bt varieties) nearby farms with a crop of peppers are at risk.

In a diverse landscape comprised of many small farms and fields, pest management cannot focus on one pest at a time, and must also consider movement at multiple scales from plant to plant, to state to state. At a local level, growers need to know what pests are in their field

(accomplished by in-field scouting or monitoring traps) to optimize treatment plans. Pest populations in one field can move into adjacent fields under their own volition or they can sometimes be transported by stronger forces, like prevailing winds and weather systems. This ability to move long distances makes state-wide and regional pest monitoring networks important for alerting growers to pest issues that might impact them in the near future.

Insect management in vegetable crops must often also consider the potential of vectoring of plant pathogens. The definition of an insect species as a pest rises, and can expand to multiple crops, if that species has the capacity to vector a pathogen. Many vegetable crops grown in PA are susceptible to vectored pathogens. Examples include snap beans and a host of aphid-vectored viruses (including cucumber mosaic, bean yellow mosaic, clover yellow vein), and cucurbits and the bacterial pathogens Erwinia tracheiphila and Serratia marcescens (vectored by cucumber beetles and squash bugs, respectively).

Habitat diversity encourages populations of beneficial insects and pollinators, but can also serve as a reservoir for pests and plant diseases. Diversified landscapes are home to increasingly diverse communities of beneficial insects, pests, and the plant diseases associated with them. Insect communities are constantly changing due to migration and dispersal of new species and selection due to changing management practices. Conventional and organic farms will have different pest challenges because of the different control methods available to each of them. With the increased interest in organic production, growers are facing challenges in controlling insect pests without conventional insecticides.

3 A challenge, then, becomes one of contributing useful information to an agricultural system comprised of multiple crops and multiple cultivars. I employ two avenues. My first avenue towards advancing insect management in diverse vegetable cropping systems is to consider two processes that are influencing insect management in most agricultural systems: population dynamics of introduced non-native species, and population phenology. In the first case, my focus is on research contributions to define population genetic structure, with the aim of also contributing to knowledge about movement and speciation. In the second, my focus is on development and validation studies that will contribute towards development of Extension methods. In all cases, I am working with insect species that are important, in part, because of their capacity to vector plant pathogens in vegetable crops.

Second, I propose to contribute work relevant to two crops: snap beans and cucurbits.

Taking the top 20 crops listed by vegetable growers in surveys, you can group them into four plant families. One, the grasses, includes only one species - sweet corn. A second, the legumes, includes multiple types of beans. By including either a grass or legume, I am working in a processing crop and with plant species and insect communities that extend readily into the larger field-crop dominated land-use pattern. The second two plant families - Solanaceae and

Cucurbitaceae – tend to be vegetable crops that harbor insect communities predominantly comprised of specialists perhaps due to their many defensive compounds aimed at preventing herbivory. Within Cucurbitaceae, the genera Cucurbita, Cucumis, and Citrulis each serve as the host to specialist pests that preferentially consume them over one of the other genera. In this thesis, I am focusing on two vegetable crops important to Pennsylvania growers; snap beans, and cucurbits. Snap beans are economically valuable mainly as a processing crop. In 2009, 27,660 tons were harvested for processing at a value of over $7 million (USDA NASS, 2009). Cucurbit crops, in Pennsylvania, are comprised of winter and summer squashes, pumpkins, melons, and cucumbers. Organic cucurbit production is increasing as consumers seek out a wider range of

4 organic produce grown in their home state. By choosing these two crops, I am learning about and contributing towards the advancement of insect management to two sectors of vegetable crops in

Pennsylvania.

Leaving the description of vegetable production in Pennsylvania, this introduction will cover the major themes and systems included in this dissertation beginning with an overview of the aphids.

Aphids

Aphids are prevalent in many areas of the world. As a group, aphids appeared approximately 170 -150 MYA in the Early to Late Jurassic and their radiation from conifers to angiosperms occurred approximately 40 MYA (Grimaldi and Engel, 2005). There are approximately 4800 described species, with 1400 (~29%) of them present in North America.

Surprisingly, aphid biodiversity is greater in temperate regions than in tropical areas. This is the opposite trend from other arthropods, and Dixon et al. (1987) hypothesized that it may be the case because aphids are very host specific, inefficient at finding hosts, and their method of reproduction is costly, thus requiring prolonged feeding. They are phloem-feeders, and must spend an extensive amount of time feeding in order to get adequate nutrition from their host plants. Most aphid species feed on very specific hosts, with one of the most polyphagous (Myzus persicae) being recorded on less than one percent of all plant species (Blackman and Eastop,

2000).

Aphids are notable for their ability to reproduce asexually (parthenogenesis). Their lifecycles are described in two ways – whether or not they alternate hosts and whether or not they reproduce sexually at any time during their life cycle. Aphids are either heteroecious or monoecious, meaning that they alternate hosts (heteroecious, ~10% of species [Grimaldi and

5 Engel, 2005]) or stay on one host (monoecious) throughout the year. Reproduction is defined as either holocyclic (asexual reproduction interrupted by a sexual phase) or anholocyclic

(reproducing only parthenogenetically). Aphid host plants are described as being primary (the host where sexual reproduction occurs if the species is holocyclic, or the host where the entire lifecycle occurs if the species is anholocyclic) or secondary (the host where asexual reproduction occurs if the species is holocyclic).

Female aphids that give birth to live young are called viviparae. The young are clones of the parent aphid. The female sexual morphs that lay eggs are called oviparae. The eggs overwinter on the primary host and when they hatch, the resulting apterous (wingless) females are called fundatrices, as they are the founders of the colony. The ability of aphids to give birth to live young and their short generation time (as little as 5 days), means that aphid populations can increase rapidly on a host plant. With this rapid increase, crowding, and declining host quality due to the aphid feeding or due to changes in host development, apterous females will produce young that mature into alate (winged) adults. These adult females leave the crowded or lower quality host in search of a new host plant. Male alate sexual morphs are produced in late summer or early fall along with the female oviparae possibly due to photoperiod cues, host effects or an interaction of factors. These two morphs return to the primary host, where sexual reproduction occurs.

Aphids are relatively weak fliers. They are capable of directional flight, but are also small enough that they can get lifted by updrafts and carried by weather systems. When landing on a potential host plant, they use a tasting probe where they insert their stylet (mouthpart derived from the mandible and maxilla) into the epidermal layer of the plant and ingest some of the fluid from the punctured epidermal cell (Ng and Faulk, 2006). If the host is suitable, they will commence with a feeding probe in which the stylet is threaded between the epidermal cells and into the vascular tissues of the plant. The feeding probe takes much longer to initiate, and once

6 the stylet is in place the aphid will remain sessile for long periods of time. If disturbed during feeding, the aphid has to extricate its stylet carefully or risk damaging it.

In the ecosystem, aphids are a plentiful source of food for polyphagous arthropod predators and a source of sugar (in the form of the honeydew they excrete) for ants, fungi, and even native birds (Grimaldi and Engel, 2005). Coccinelids, syrphids, lacewings, and other predaceous Hemipterans are some of the insects that will prey on aphid colonies. These generalist predators are well studied in agricultural systems as a means of biocontrol for aphids. Ant interactions with aphids can be symbiotic (ants tending aphids, actively protecting them from predators or the elements) or more casual (no active tending or protection but still consuming the honeydew). Some ants will move aphids to different parts of a plant or even actively defend their colonies from predators. Aphids are good continual sources of honeydew (excreted plant phloem) since they must feed almost constantly to extract enough amino acids from the phloem.

Aphids host on trees, shrubs, herbaceous plants and crops. They are most economically important as crop pests, with some species being highly polyphagous and impacting multiple crops. In addition to the feeding damage associated with aphids removing phloem, the honeydew they excrete onto plants creates ideal conditions for the growth of sooty mold.

Some aphids are gall formers, inducing host plants to make protective enclosures for their colonies out of their host plants. There are even a few species that make soldier morphs with enlarged front legs to defend colonies from attack.

Aphids have intracellular symbiotic bacteria from the genus Buchnera. Buchnera produces amino acids for the aphid that it cannot get from phloem feeding alone. The bacteria is also associated with producing proteins necessary for some viruses to be transmitted by aphids

7 Aphids as vectors of plant viruses

In addition to their feeding damage due to the removal of phloem on heavily infested plants, aphids can also damage plants by being efficient vectors of plant viruses. There are two general types of relationships between aphids and the viruses they transmit. The first type is persistent transmission, which can be replicative or non-replicative. Viruses that are persistently transmitted are obtained by an aphid from infected tissue during a feeding episode. The virus moves through the stylet and enters the aphid’s digestive system, eventually passing through the gut lining (and in the case of replicative viruses, multiplying in the midgut cells) and entering the hemocoel (Gray and Banerjee, 1999). The virus must return to the salivary glands to infect a new host. From the salivary glands, the virus is injected into another plant when the aphid feeds again.

Persistent viruses are notable for not necessarily being detrimental to the aphid, but once infected the aphid will vector them for life (Gray and Banerjee, 1999).

The second type of virus transmission is non-persistent. These viruses are obtained quickly by their aphid vector during short tasting probes which only puncture the epidermal cells and last a few seconds or minutes versus the hours necessary for feeding. The viruses stick to the stylet lining by binding to helper component proteins (a protein that binds to the stylet wall and the virus coat protein) or directly to the stylet depending on the type of virus (Ng and Falk, 2006).

They will remain there until they are flushed out during another tasting probe, which from the perspective of the virus would ideally be on a new host. Non-persistently transmitted viruses are vectored for only a short time, minutes to hours, by the aphid due to their association with the stylet lining (Ng and Falk, 2006). Both types of aphid/virus relationship can be economically important. For example, aphids can vector cucumber mosaic virus (CMV) in a non-persistent manner, which causes serious damage to cucurbits and legumes.

8 Aphids and snap beans

Aphids are rarely direct pests of snap beans in Pennsylvania, but they are important vectors of plant viruses. Snap beans are a relatively short season crop with multiple plantings.

The shorter growing time means that few plantings will have sufficient aphid populations develop to damaging levels, however disease transmission as the alates migrate through fields is still a risk. In addition to the cultural control that is partially provided by a relatively short-season crop, aphids are maintained below damaging population thresholds through natural biological control, seed treatments, and foliar insecticide use. Other factors that are important in predicting and mitigating the effects of viral epidemics include timing of infection, presence of the virus in the ecosystem, and movement of competent vector species.

Viruses in snap beans can be economically important. In the early 2000s, severe plant stunting, flower abortion, mild to severe mosaic patterns on leaves, rolling of leaves, yellowing of leaves, stunted pods, necrotic streaks on pods and necrosis of internal pod tissues and reduced yields were observed in beans from the upper Midwest, parts of Canada, and into New York.

These symptoms were typical of virus infection and followed the presence of large numbers of the soybean aphid (A. glycines), which at that time was a new invasive species. Further testing determined that a legume strain of cucumber mosaic virus (CMV) was the cause of these symptoms in snap bean. The recently introduced aphid, A.glycines was particularly important because it was found to be a competent vector of CMV in single-aphid assays and also captured in large numbers in snap bean and surrounding fields (Gildow et al. 2008, Nault et al. 2009).

9 Introduced species

In general, the introduction of species – be they plant, invertebrate, or vertebrate – to new areas has increased in prevalence with the increase in global trade and travel. In order to be considered introduced, a species must be transported outside of its native range and establish itself in its new range. Introduced insect species of recent note in PA or nearby include emerald ash borer, Asian longhorned beetle, brown marmorated stink bug, spotted wing drosophila,

Hemlock wooly adelgid, gypsy moth, Western bean cutworm, small hive beetle, and the multicolored Asian lady beetle. Introduced species are the cause of novel and unexpected problems in vegetable crops when they first appear and establish. As they become established it is important to understand their movement and life history in order to effectively manage them.

Aphis glycines is a relatively new introduced species from Asia, being first recorded in the US in 2000 (Ragsdale et al 2004). It is mainly a problem on soybean, but it is also a competent virus-vector making it important in other legumes, like snap bean. When A. glycines was introduced, both the primary and secondary hosts, Glycines max (soybean) and Rhamnus spp.

(buckthorn) respectively, were already established in the United States. Soybean is an important field crop grown on over 69 million acres nationally, with the greatest concentration in the

American Midwest (USDA NASS, 2009). Rhamnus spp. are understory and edge trees and include both native and non-native (notably Rhamnus cathartica) species. This facilitated colonization and undetected spread until populations reached the point where they were causing economically significant crop damage.

Therioaphis trifolii is another species that was introduced to the United States and became important because of its abundance and preference for a legume crop. Like A. glycines, it is a competent vector of CMV in snap bean (Gildow et al. 2008, Nault et al. 2009). The spotted alfalfa aphid (referred to as SAA and formerly of the species Therioaphis maculata) was first

10 reported in the United States in 1954, in New Mexico on alfalfa. It spread rapidly through the

Southwest and on to the Eastern states (Dickson 1959). The population in North America was noted to have biotypes with varying resistance to organophosphate insecticides (Berg and

Ridland, 1981). Its introduction and subsequent unexpected damage to crops mirrors that of A. glycines. Since it has been in the U.S. for over 50 years, its populations have equilibrated due to effective management strategies and it is no longer the cause of extreme economic loss in its host crops of alfalfa and clover. T. trifolii can serve as a model invasive aphid species and potentially give insight into the future population structure of A. glycines.

Insect movement

Insects, especially those capable of strong directional flight, can move long distances without the aid of human travel and trade (as was probably the case for both A. glycines and T. trifolii). An extreme example is that of the Monarch butterfly which migrates from the United

States (as far north as New England) back to its overwintering grounds in Mexico. The adults cover the over 3000 mile distance in ~4 months. Insects can also avail themselves of the prevailing winds and travel long distances as aerial plankton within weather systems.

Alate aphids end up in weather systems, with over 140,000 individuals collected from suction traps in a 4 year period (Schmidt et al. 2012). This ability to travel with weather systems

(whether directly or indirectly), allows aphids to reach new areas with possibly previously uninfested host plants. With A. glycines this long distance movement is especially important for legume-growing areas of the country that were not in the range of the original introduction and/or do not have the preferred primary host on which the aphid can overwinter.

Air flow trajectory models can be used to better understand insect movement, especially those of insects (like aphids) that are not strong fliers. HYSPLIT (Hybrid Single Particle

11 Lagrangian Integrated Trajectory Model) is a publicly availably computer model from the

National Oceanic and Atomospheric Administration (NOAA) to calculate simple air parcel trajectories (Draxler and Rolph 2012). This model can calculate simple forward and back trajectories from single or multiple starting points for a single particle or plume. HYSPLIT is used by meteorologists and others to calculate or forecast particle dispersion.

Rhamnus cathartica (common buckthorn) is the North American overwintering host for

Aphis glycines, which is an economically important pest of soybeans. In its native range, R. davurica is the overwintering host for A. glycines in its native range, but it is not present in North

America. R. cathartica distribution is thought to be one of the limiting factors defining where A. glycines can maintain local populations. Determining its range is important when assessing the contributions of migrants versus a local population for IPM.

Rhamnus cathartica is a shrub/small tree (2-6 m) that is native to Europe and Asia

(Archibold et al. 1997, Converse 1985). It was introduced to North America in the 1800s for windbreaks and hedgerows because of its wide tolerance for many environmental conditions and eventually escaped from cultivation. Now, R. cathartica commonly invades forests and open fields where the soil is moist (Archibold et al. 1997, Kurylo et al. 2007). It can out-compete native species in shaded areas to become a dominant species in the understory (Archibold et al.

1997). Most of the fruit falls beneath the parent tree where seedlings survive well, but some are eaten and dispersed by birds, mice, and white-tailed deer (Archibold et al. 1997, Myers et al.

2004). When R. cathartica is removed from an area the native vegetation returns quickly as long as the seed bank is viable, indicating that R. cathartica may be allelopathic (Bodreau 1992).

Two experiments confirmed R. cathartica as a suitable host for Aphis glycines (Yoo et al.

2005, Voegtlin et al. 2005). On plants in outdoor cages, the soybean aphid over-wintered successfully, developed colonies with viable eggs and alates, on three Rhamnus species (R. cathartica, R. alnifolia, R. lanceolata: Voegtlin et al. 2005). Alates were also observed feeding on

12 R. cathartica, but were not even observed on the non-Rhamnus species in the experiment

(Voegtlin et al. 2005). Yoo et al. (2005) confirmed R. cathartica and R. alnifolia as suitable hosts for A. glycines through a single-choice experiment. The aphids were induced to produce sexual morphs in the laboratory using an autumnal light and temperature regime. Gynoparae and oviparae were caged with a member of the Rhamnaceae family and a soybean plant, and they survived the longest and produced more nymphs on R. cathartica and R. alnifolia. The other

Rhamnaceae taxa were unsuitable hosts because the soybean aphid could not complete development on them without an alternative plant choice.

R. cathartica is the more widespread of the suitable host plants and had the highest fecundity and survival of aphids. Thus, it is most likely to be a major part of the soybean aphid lifecycle in the United States (Voegtlin et al. 2005, Yoo et al. 2005).

Current IPM tools and plant databases do not accurately represent the distribution of R. cathartica in Pennsylvania. The PAPIPE (http://pa-pipe.zedxinc.com/) operates under the assumption that the entire state is suitable habitat in its A. glycines phenology model. Other sites, like PLANTS Database (http://plants.usda.gov/java/), do not have records from every county and one specimen is enough to consider that county suitable for the species.

Central Pennsylvania appears to be a transition zone for this species. R. cathartica exists locally in disturbed forests and along trails, which is similar to what McCay et al. (2009) found in

New York, but not in the same density and/or distribution as some high A.glycines population areas like the Midwest or Northeast. The Forest Inventory Analysis (http://www.fia.fs.fed.us/) does not include buckthorn in its surveys and it is important to understand what is limiting R. cathartica in Pennsylvania and areas to the south. For the purposes of this thesis, I am using this information about the range of R. cathartica as an impetus to investigate the influence of long distance A. glycines migrants on soybeans grown in Pennsylvania.

13 Phenology and pest management

Phenology is simply defined as the timing of biological events. Understanding the phenology of pest species can improve management practices by timing planting for when a pest is absent or dormant and applying control measures when the pest is most vulnerable. Phenology models are complicated by the fact that they require the estimation or discovery of what environmental cues the insects are using in their development and the impact of each cue.

Phenology models driven by temperature are well-developed for many crop pests (Brown

1982, Hansen 2006, Hodgson et al 2011,Tobin et al 2001). Pesticide use and control measures can be more accurately deployed when growers are well informed about risk level, which must incorporate pest phenology models. These are usually developed from laboratory studies that measure development of non-diapausing stages under controlled temperatures. In the case of cucurbit crops, however, phenology models for the two most important pest species – the striped cucumber beetle (Acalymma vittatum) and the squash bug (Anasa tristis) – exist in the literature for portions of their life stages, but components that define diapause initiation and termination are missing, and no models have been validated under field conditions.

Both the striped cucumber beetle and squash bug overwinter as adults, move among crop patches, and both insect species are vectors of bacterial plant pathogens. In general, pests that overwinter as adults, or those that exhibit extensive movement, are more challenging to work with when trying to utilize phenology models. Field-validation of phenology models for these species, and incorporation of phenology models useful to growers in Extension programs, will need to consider these characteristics about overwintering, movement, and vector-capacity.

In conventional systems, these pests are controlled with neonicotinoid pesticides

(Fleischer et al. 1998) as drenches when transplanting seedlings, as seed treatments, or as foliar sprays during the season. Certified organic producers are restricted to using organically certified

14 products to control these pests, and cannot rely on the same pesticides that are used by conventional growers. Knowledge of pest phenology will allow growers to more effectively deploy organic and cultural controls like approved pesticides, row covers, soil applied biocontrols targeting larval stages (Ellers-Kirk and Fleischer, 2006) and adjusted planting dates.

Striped cucumber beetle is an important pest of cucurbits in any production system because of its direct feeding damage to vines and fruit as well as its ability to vector Erwinia tracheiphila, the causal agent of bacterial wilt. Striped cucumber beetle overwinters as an adult in leaf litter and/or crop residue and resumes activity in the spring. When it becomes active in the spring is a point for investigation. Radin and Drummond (1994) suggest that the beetles can be active on any day with an average temperature above 12 C in Maine, and Lewis et al. (1990) found beetle activity on trap flats when temperatures were above 18 C. Previous trapping efforts in Pennsylvania at Rock Springs caught beetles in emergence cages in mid-May (Fleischer, unpublished). The cue that causes these beetles to enter diapause in the fall is unknown, and they do not have a synchronous spring emergence.

There are control measures directed towards both the larval and adult stages of this insect

(Ellers-Kirk et al. 2006). Systemic and foliar pesticide sprays are effective in conventional systems against the adults. Larval control can be achieved by growing cucurbits on black plastic with drip irrigation and also introducing entomopathogenic nematodes through the drip line

(Ellers-Kirk et al 2000).

Squash bug is an important pest of Cucurbita but not Cucumis. This insect also inflicts feeding damage on vines and fruit, and can vector yellow vine decline (S. marcescens), another cucurbit disease. It also overwinters as an adult and has a non-synchronous start to spring activity.

During the growing season, the lifecycle occurs above ground, making it relatively easy to observe egg masses, juvenile instars, and subsequent adult generations. Common cucurbit cultivation practices include the use of mulch (usually plastic), which provides harborage for

15 squash bug adults and nymphs, thus increasing pest pressure (Cartwright et al. 1990). It is unknown how squash bugs terminate diapause, but it is known that the bugs enter diapause due to photoperiod cues (Decker and Yeargan 2008). The critical photoperiod is between 14:10 (L:D) and 14.5:9.5 (Nechols 1988).

Squash Vine Borer (Melittia cucurbitae, Lepidoptera: Sesiidae) is the third pest in this study. In Pennsylvania, squash vine borer has mainly been a problem for small-scale growers or home gardeners, which increases its importance to small-scale organic growers. This uneven influence may be the result of the scale at which the crop is grown and yields are expected by both audiences. On a small farm, if a few squash vine borers are present they could decimate a small planting, while the same number of insects would cause damage that is barely noticeable in a large field. This insect is most damaging as a caterpillar, which burrows into the cucurbit vine, and the resulting feeding damage can potentially kill the entire vine. The most effective way to control this insect is to spray when egg masses and/or first instars are present before they enter the vine. Adults can be monitored with pheromone traps. Squash vine borer overwinters as a late instar larva or pupae and has an extended emergence through the growing season. It requires

1687.5 degree days base 50° Fahrenheit to complete development (Canhilal et al. 2006).

In the case of these cucurbit pests, predicting early season activity may be useful for optimizing planting time to give the plants a head start with limited pest pressure. For striped cucumber beetle and squash bug, we are defining early season activity as the recruitment of the pest to a host. These two pests are active in the spring before crops are planted, and monitoring their presence can estimate the intensity of in-season pest pressure. Once cucurbit crops are planted, in-field monitoring identifies a biofix for each pest. Using degree day developmental requirements from the literature for each pest, we can use the biofix, temperature-dependent development models, and forecast and 30-year climate records to estimate the emergence of subsequent field generations. Since cucurbits require the services of pollinators, predicting pest

16 levels in the growing season is important for timing pest control measures while still allowing for adequate pollination.

Dissertation Objectives

The goal of my thesis is to use population structure and phenology to advance insect management in diverse vegetable agroecosystems. My work will be presented in four chapters:

In Chapter 2 I describe the alate aphid species composition in Northeastern US processing snap beans and update the list of aphid species found in Pennsylvania. In addition to describing the alate aphid community in Pennsylvaina snap beans, I used the J. O. Pepper aphid slide collection in the Frost Entomological museum, previously published works (Pepper 1965,

Wallis et al 2005), and the results of an aphid trapping survey in snap beans to generate a comprehensive review of aphid species present in Pennsylvania. Two of the most abundant aphid species in our survey of snap bean fields were A. glycines and T. trifolii, and they were selected for further investigation.

Chapter 3 and 4 improve our understanding of population structure of two invasive aphid species relevant to virus transmission in snap beans. Because of the absences of large quantities of the primary host (Rhamnus spp.) in Pennsylvania, I hypothesize that the A. glycines population found in Pennsylvania during the growing season is largely influenced by migrants. I am using genetic tools and air-flow trajectory models to investigate the natal sources of A. glycines in

Pennsylvania. In Chapter 4, I am also using genetic tools to understand the possible origin of T. trifolii.

In Chapter 5 I set out to validate phenology models for pests of cucurbits. I monitored the early season activity and growing season phenology of three pests on land that is transitioning to organic production on research farms in Pennsylvania, Iowa and Kentucky on two commercially

17 important cucurbit crops (muskmelon – Cucumis melo and butternut squash – Cucurbita moschata) to better inform organic management practices. I am also testing our ability to create accurate phenology models of these pests using air-temperature degree-days from weather stations and development times from the existing literature.

References

Archibold OW, Brooks D, Delanoy L (1997) An investigation of the invasive shrub European buckthorn, Rhamnus cathartica L., near Saskatoon, Saskatchewan. Canadian Field Naturalist 111:617-621.

Berg GN, Ridland PM (1981) Characterisation of the spotted alfalfa aphid population in Victoria. Journal of the Australian Entomological Society 20:317-318.

Blackman R, Eastop V (2000) Aphids on the World’s Crops: An Identification and Information Guide, 2nd ed.

Bodreau D (1992) Buckthorn research and control at Pipestone National Monument (Minnesota). Restoration and Management Notes 10:94-95.

Brown GC (1982) A generalized phonological forecast model for European corn borer. Journal of the Kansas Entomological Society 55: 625-638.

Canhilal R, Carner GR, Griffin RP, et al. (2006) Life history of the squash vine borer, Melittia cucurbitae (Harris) (Lepidoptera: Sesiidae) in South Carolina. Journal of Agricultural and Urban Entomology 23:1-6.

Cartwright B, Palumbo JC, Fargo WS (1990) Influence of crop mulches and row covers on the population dynamics of the squash bug (Heteroptera: Coreidae) on summer squash. Journal of Economic Entomology 83:1988-1993.

Converse CK (1985) Element Stewardship Abstract for Rhamnus cathartica, Rhamnus frangula (syn. Frangula alnus). The Nature Conservancy.

Decker KB, Yeargan KV (2008) Seasonal phenology and natural enemies of the squash bug (Hemiptera: Coreidae) in Kentucky. Environmental Entomology 37:670-8.

Dickson RC (1959) On the identity of the spotted alfalfa aphid in North America. Annals of the Entomological Society of America 52:63-68.

18 Dixon AFG, Kindlmann P, Leps J, Holman J (1987) Why are there so few species of aphids, especially in the tropics. The American Naturalist 129:580-592.

Draxler RR, Rolph GD (2012) HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) Model access via NOAA ARL READY. NOAA Air Resources Laboratory, Silver Spring MD. http://ready.arl.noaa.gov/HYSPLIT.php.

Ellers-Kirk CD, Fleischer SJ (2006) Development and life table of Acalymma vittatum (Coleoptera: Chrysomelidae), a vector of Erwinia tracheiphila in Cucurbits. Environmental Entomology 35:875-880.

Fleischer SJ, Orzoleck MD, deMackiewicz D, Otjen L (1998) Imidacloprid effects on Acalymma vittatum (Coleoptera:Chrysomelidae) and bacterial wilt in cantaloupe. Journal of Economic Entomology 91: 940-949.

Gildow FE, Shah DA, Sackett WM, et al. (2008) Transmission efficiency of cucumber mosaic virus by aphids associated with virus epidemics in snap bean. Phytopathology 98:1233-41.

Gray SM, Banerjee N (1999) Mechanisms of arthropod transmission of plant and animal viruses. Microbiology and Molecular Biology Reviews 63:128-48.

Grimaldi D, Engel M (2005) The Evolution of the Insects. Cambridge University Press.

Hansen LM (2006) Models for spring migration of two aphid species Sitobion avenae (F.) and Rhopalosiphum padi (L.) infesting cereals in areas where they are entirely holocyclic. Agricultural and Forest Entomology 8:83–88.

Hodgson JA, Thomas CD, Oliver TH, et al. (2010) Predicting insect phenology across space and time. Global Change Biology 17:1289–1300.

Kurylo JS, Knight KS, Stewart JR, Endress AG (2007) Rhamnus cathartica: Native and naturalized distribution and habitat preferences. Journal of the Torrey Botanitcal Society 134:420-430.

Lewis PA, Lampman RL, Metcalf RL (1990) Kairomonal attractants for Acalymma vittatum (Coleoptera: Chrysomelidae). Environmental Entomology 19(1): 8-14.

Manglitz G, Calkins C, Walstrom R, et al. (1966) Holocyclic strain of the spotted alfalfa aphid in Nebraska and adjacent states. Journal of Economic Entomology 59:636-639.

McCay TS, McCay DH, Caragiulo AV, Mandel TL (2009) Demography and distribution of the invasive Rhamnus cathartica in habitats of a fragmented landscape. Journal of the Torrey Botanical Society 136:110-121.

Myers JA, Vellend M, Gardescu S, Marks PL (2004) Seed dispersal by white-tailed deer: implications for long-distance dispersal, invasion, and migration of plants in eastern North America. Oecologia 139:35-44.

19 Nault BA, Shah DA, Straight KE, et al. (2009) Modeling temporal trends in aphid vector dispersal and cucumber mosaic virus epidemics in snap bean. Environmental Entomology 38:1347-59.

Nechols JR (1988) Photoperiodic Responses of the Squash Bug (Heteroptera: Coreidae): Diapause Induction and Maintenance. Environmental Entomology 17:427-431.

Pepper JO (1965) A List of the Pennsylvania Aphididae and Their Host Plants (Homoptera). Transactions of the American Entomological Society 91:181-231.

Radin AM, Drummond FA (1994) Patterns of initial colonization of cucurbits, reproductive activity, and dispersion of striped cucumber beetle, Acalymma vittatum (F.) (Coleoptera: Chrysomelidae). Journal of Agricultural Entomology 11:115-123.

Ragsdale DW, Voegtlin DJ, O’Neil RJ (2004) Soybean Aphid Biology in North America. Annals of the Entomological Society of America 97:204–208.

Schmidt NP, O’Neal ME, Anderson PF, et al. (2012) Spatial distribution of Aphis glycines (Hemiptera: Aphididae): a summary of the suction trap network. Journal of Economic Entomology 105:259-271.

Tobin PC, Nagarkatti S, Saunders MC (2001) Modeling development in grape berry moth (Lepidoptera: Tortricidae). Environmental Entomology 30(4): 692-699.

USDA National Agricultural Statistics Service. 2009. 2007 Census of Agriculture: Pennsylvania State and County Data.

USDA. National Agricultural Statistics Service - PA Office. PA Ag Snapshot 2011. http://www.nass.usda.gov/Statistics_by_State/Pennsylvania/Publications/Annual_Statistical _Bulletin/Snapshot.pdf Accessed: May 23, 2012.

Voegtlin DJ, O’Neil RJ, Graves WR, et al. (2005) Potential winter hosts of soybean aphid. Annals of the Entomological Society of America 98:690-693.

Wallis CM, Fleischer SJ, Luster D, Gildow FE (2005) Aphid (Hemiptera: Aphididae) species composition and potential aphid vectors of plum pox virus in Pennsylvania peach orchards. Journal of Economic Entomology 98:1441-50.

Yoo HJS, O’Neil RJ, Voegtlin DJ, Graves WR (2005) Host plant suitability of Rhamnaceae for soybean aphid (Homoptera: Aphididae). Annals of the Entomological Society of America 98:926-930.

Chapter 2

Alate aphid species composition in Northeastern US processing snap beans and an update to historical lists

Introduction

Aphids are a small, but diverse group of insects with an origin in the Jurassic period and a total of 4800 species world-wide (Grimaldi and Engel 2005, Dixon 1985a, Dixon 1985b). They are primarily phloem feeders and in high numbers this can be damaging to their host plant.

Aphids excrete excess carbohydrates from their diet of phloem sap, and sooty mold fungi which develop on this substrate can result in unmarketable produce. Also, aphids can serve as the host resource for an array of generalist predators and specialist parasitoids in agroecosystems.

Some species of aphids are also important pests because of their ability to vector plant viruses. There are two types of relationships between aphids and the viruses they transmit. The first type is persistent transmission, which can be replicative or non-replicative. Viruses that are persistently transmitted are obtained by an aphid from infected tissue during a feeding episode.

The virus moves through the stylet and enters the aphid’s digestive system, eventually passing through the gut lining (and in the case of replicative viruses, multiplying in the midgut cells) and entering the hemocoel (Gray and Banerjee 1999). The virus must return to the salivary glands to infect a new host. From the salivary glands, the virus is injected into another plant when the aphid feeds again. Persistent viruses are notable for not necessarily being detrimental to the aphid, but once infected the aphid will vector them for life (Gray and Banerjee 1999).

The second type of virus transmission is non-persistent. These viruses are obtained quickly by their aphid vector during short tasting probes which only puncture the epidermal cells

21 and last a few seconds or minutes versus the hours necessary for feeding. The viruses stick to the stylet lining by binding to helper component proteins (a protein that binds to the stylet wall and the virus coat protein) or directly to the stylet depending on the type of virus (Ng and Falk 2006).

They will remain there until they are flushed out during another tasting probe, which from the perspective of the virus would ideally be on a new host. Non-persistently transmitted viruses are vectored for only a short time, minutes to hours, by the aphid due to their association with the stylet lining (Ng and Falk 2006). Both types of aphid/virus relationship can be economically important. For example, aphids can vector cucumber mosaic virus (CMV) in a non-persistent manner, which causes serious damage to cucurbits and legumes.

Aphids can reproduce parthenogenically, sexually, or using both methods depending on the species and can spend the entire year on one host or alternate hosts. Both the type of reproduction and host choice are used to categorize aphid life cycles. For reproduction, aphid species are holocyclic (parthenogenesis interrupted by sexual phase), or anholocyclic (year round parthenogenesis). For host selection, aphid species are heteroecious (host alternating), or monoecious (non-host alternating). Unlike most other insect groups, aphids are more diverse in temperate zones than in the tropics (Dixon 1985a). Dixon et al. (1987) hypothesized that this is the case because of their host selectivity, the cost of parthenogenesis, and their lack of efficiency when locating host plants. As Pennsylvania and New York are situated in the Northeastern region of the United States and contain diverse landscapes of deciduous and coniferous forest, agriculture, and developed land, one would expect both states to be species-rich when it comes to aphids.

In Pennsylvania, J.O. Pepper was an entomologist specializing in aphids and actively collected them for most of the 20th century. His collections were centered on his home in central

Pennsylvania, State College, and included much of the surrounding forest and farmland. The bulk of his slides resulting from his collection efforts are housed in the Frost Entomological Museum

22 (University Park, PA), and he also contributed slides to the United States National Collection

(Beltsville, MD). Pepper (1965) reported 345 species in a published list of the aphids of

Pennsylvania and their host plants. To date, this is the most comprehensive published list of aphids for the state. However, since taxonomy and systematics are in flux, the names that Pepper published are currently out of date and in need of revision. Pepper’s contemporary in New York was M. Leonard. Leonard (1963) is a list of the aphids in New York which was updated with

Leonard (1968), ultimately reporting over 430 aphid species in the state.

The ability of aphids to transmit viruses recently became important in Pennsylvania orchards because of the appearance of an invasive species of virus, plum pox virus (Wallis et al

2005). Since it was unknown which aphid species (or group of species) could serve as the primary vector of plum pox, it was important to set up a trapping system in PA orchards to determine what species were present. The resulting aphid captures were slide mounted and identified in order to determine what aphid species were visiting the trees and possibly vectoring the viruses. Once the species were identified, colonies could be set up and each species could be tested to see which were the most efficient at vectoring the viruses (Gildow et al. 2008).

In 2003, snap bean crops in Pennsylvania had virus-like symptoms (leaf mosaic and blistering, deformed pods) and experienced dramatic yield loss. This coincided with the appearance of a newly invasive aphid, Aphis glycines, and earlier reports of similar epidemics in the Midwest and East (Larsen et al. 2002). Damage from plant viruses in snap beans was previously documented, but large scale epidemics were novel for the region and further investigation was necessary to determine the vector(s).

Ten species of aphid have been recorded on Phaseolus spp. (Blackman and Eastop,

2000). Of the aphids listed, six are found in PA (Aphis craccivora, A. fabae, A. gossypii,

Macrosiphum euphorbiae, Myzus persicae, and Smynthurodes betae). Aphid feeding is not as severe a problem on snap beans as feeding by the defoliating pests (ex Mexican bean beetle,

23 cabbage looper), but aphids are important vectors of plant viruses. The shorter growing time means that few plantings will have sufficient aphid populations develop to damaging levels, however pathogen transmission as the alate aphids migrate through fields is still a risk. In addition to the cultural control that is partially provided by a relatively short-season crop, aphids are maintained below damaging population thresholds through natural biological control, seed treatments, and foliar insecticide use. Other factors that are important in predicting and mitigating the effects of viral epidemics include timing of infection, presence of the virus in the ecosystem, and movement of competent vector species.

In Pennsylvania, late planted beans are the most susceptible to CMV (Nault et al 2004, and 2009). By July, aphid colonies on surrounding hosts are producing alates that are moving through and tasting plants in multiple fields. Since A. glycines especially do not use snap beans as one of their secondary hosts (Ragsdale et al. 2004), the chances that they will cause significant feeding damage are low. Damage occurs when a virus infected aphid tastes or feeds on an uninfected plant, passing on the pathogen. If infection occurs during flowering and pod set, it is likely that it will affect yield. Infection after pod set is less likely to affect yield as the virus causes the most damage to the growing points (Jones et al 2008, Zitter and Murphy 2009).

To determine the species composition of the aphid community that could contribute to

CMV transmission in snap beans, we collected alate (winged) aphids from local snap bean fields from 2004 through 2006 in Pennsylvania and 2002 through 2006 in New York, identified them to species, estimated species richness and collated the results with those of previous collections.

Methods

Detailed methods for alate aphid collection in snap bean fields in Pennsylvania and New

York were published in Nault et al. (2009). To summarize, we used water pan traps baited with a green ceramic tile (Webb et al 1994) and filled with a 20% propylene glycol solution in late- planted snap bean fields in both states from 2002 – 2006 in NY and 2004 – 2006 in PA. Traps were installed in a total of 56 fields in western NY (12 each year, except for 2004 which had 8 fields) and 18 fields in Centre county PA (6 each year). The traps in Centre County formed an approximately 40 mile transect in the southern portion of the county roughly following state routes 45 and 192. The traps were checked weekly for aphids from the early trifoliate stage (early to mid July) until field harvest. Aphids collected in NY were identified by R. Eckel (RVWE

Consulting, Frenchtown, NJ), and aphids from Pennsylvania were slide mounted and identified by W. Sackett and A. Bachmann using keys by Smith et al. (1992) and Blackman and Eastop

(2000). The aphid species list resulting from all of the trap catches was used for the species rarefaction curves. Voucher specimens are located at the New York State Agricultural

Experiment Station in Geneva, NY, and Dr. Fleischer’s lab collection in the Department of

Entomology, Pennsylvania State University, University Park, PA.

Species rarefaction curves were calculated for the PA and NY collections individually and combined using EstimateS (Colwell 2005).

To generate a complete list of the aphids of PA, I searched the J. O. Pepper aphid slide collection housed at the Frost Entomological Museum (University Park, PA) and listed every species present and compared that with the aphids recorded in Pepper (1965). I searched the slide

25 collection in addition to using Pepper (1965) because Pepper continued to collect aphids and make slides into the late 1980s, but did not publish any updates to his original 1965 paper.

Because the collection and Pepper (1965) contained aphid species names from the early 20th century, I consulted two online aphid databases to ensure that the final list used the most current nomenclature (Aphid Species File – http://aphidspeciesfile.org, accessed April 22, 2012 and the

United States Collection of Aphididae - http://www.sel.barc.usda.gov/aphid/aphframe.htm, accessed April 22, 2012). I combined my findings from the Pepper collection material with the results of our pan trapping study and Wallis et al (2005) to create a more current list of the aphids of PA (Tables 2-3 – 2-11).

Slide mounting protocol

Aphids collected from the water pan traps in Pennsylvania were prepared for identification by slide mounting individuals. Aphids were stored in 70% EtOH, then transferred to KOH and heated for 1 hour or until clear. Cleared aphids were rinsed for 10 minutes each in a sequence of 95% EtOH, absolute EtOH, and clove oil. Once rinsed, each aphid was placed on a drop of Canada balsam on a glass slide and positioned to expose diagnostic features before a coverslip was placed on top.

Results

We caught and identified a total of 8821 aphids from PA and NY, with 7484 from NY and 1337 from PA. A total of 97 species were caught; 71 from NY and 41 from PA. Only 254

(2.8%) of the aphids were unable to be identified. Of the aphids captured, those species

26 representing 1% or greater of the total number caught in either state are listed in Table 2-1

(originally published in Nault et al 2009) with their abundances.

Our efforts resulted in a list of aphids found in the PA and NY snap bean fields (Tables 2-

3 – 2-11) and we used Blackman and Eastop (1994, 2000, and 2006) to determine host ranges for the species. From this host information we estimated that 61 percent of the species trapped in snap beans in both states were most likely coming in from the surrounding forests as their hosts are woody, not herbaceous, species (Figure 2-3).

Combining the list of aphids collected in this study as well as Wallis et al 2005 (which was conducted in the same geographic region, but in peach orchards) with the list published by

J.O. Pepper in 1965, I developed a new, more comprehensive list of the aphids present in PA. I found 8 species present in our collections that were not present in the slide collection housed in the Frost Entomological Museum (University Park, PA) or published in Pepper (1965) (Table 2-

3). One of these aphids, Aphis glycines, was introduced to the US around the turn of the 21st century and is now present in our agroecosystem.

More aphids overall were trapped in New York than Pennsylvania (Figure 2-1). Species accumulations followed asymptotic patterns suggesting reasonably adequate sampling of the aphid species present as alates in commercial snap bean fields. Overall, there were fewer aphids collected in PA but based on the rarefaction curve there were a similar number of total species represented in a sample of the same number of individuals (Figure 2-1, at 1250 individuals there would be 45 species sampled in PA and 50 in NY). Based on the historical collections reported by

Pepper, there are approximately 350 aphid species in PA. Historical reports by Leonard suggest that there are approximately 430 aphid species in NY.

27 Discussion

Our passive trapping in snap bean fields alone yielded a surprisingly high percentage of the species present throughout PA and NY (~14% and ~18% respectively). Our sampling method concentrated on only one habitat (commercial snap bean fields), but did intercept aphids moving from the surrounding forests and hedgerows. The high degree of landscape heterogeneity and crop diversity in the trapping areas includes plants that serve as hosts for many of the species that represented less than 1% of the total capture (Pfleeger et al. 2006). These aphids were captured in very small numbers (mostly singletons), and are not important contributers to the plant virus epidemics reported by Wallis et al (2005) and Nault et al (2009).

Of the aphids we captured, two species were especially notable; T. trifolii which comprised 31.8% of the identified aphids, and A. glycines which represented 18.2 % of the identified aphids. Both of these aphids were introduced to North America (A. glycines from Asia and T. trifolii from Europe) and were quite destructive to crops immediately after their introduction (soybean and alfalfa, respectively). A. glycines continues to cause significant economic damage in soybean. While not known to colonize Phaseoulus spp., both species were determined to be competent vectors of the legume strain of CMV (Gildow et al 2008).

The Pepper (1965) aphid list in addition to the Pepper slide collection allowed us to compile a comprehensive list of the aphids present in PA, but the nomenclature was in need of updating. Our efforts to update the nomenclature, and incorporate our more recent sampling efforts resulted in a modern list of aphids of PA that includes recently introduced species (like A. glycines).

The intermittent appearance of CMV in central Pennsylvania snap bean crops could be a result of our unique agricultural landscape. Our agricultural fields are located in valleys bordered by the low, but steep, forested ridges of the Appalachian Mountains. Our ridge and valley system

28 might be acting like a filter, keeping CMV out for most of the season. We did not search for a

CMV reservoir outside of testing a few alfalfa fields, which were also negative for CMV. It is possible, that much like our A. glycines population, legume strains of CMV are also a migrant species. If this is the case, migrating aphids may be scrubbed of virions when they land in one of our many bordering forests containing many non-host plants.

References

Blackman R, Eastop V (2000) Aphids on the World’s Crops: An Identification and Information Guide, 2nd ed. Wiley, Chichester.

Blackman R, Eastop V (2006) Aphids on the World’s Herbaceous Plants and Shrubs: An Identification Guide. Wiley, Chichester.

Blackman R, Eastop V (1994) Aphids on the World’s Trees: An Identification and Information Guide. Wiley, Chichester.

Colwell RK (2005) EstimateS: Statistical estimation of species richness and shared species from samples Version 8.2.0.

Dixon, AFG (1985a) Aphid Ecology. Blackie & Son Ltd

Dixon AFG (1985b) Structure of aphid populations. Annual Review of Entomology 30(1):155- 174.

Dixon AFG, Kindlmann P, Leps J, Holman J (1987) Why are there so few species of aphids, especially in the tropics. The American Naturalist 129:580-592.

Favret C. Aphid Species File Version 1.0/4.1. http://aphid.speciesfile.org. Accessed 22 Apr 2012

Gildow FE, Shah DA, Sackett WM, et al. (2008) Transmission efficiency of Cucumber mosaic virus by aphids associated with virus epidemics in snap bean. Phytopathology 98:1233-41.

Gray SM, Banerjee N (1999) Mechanisms of arthropod transmission of plant and animal viruses. Microbiology and Molecular Biology Reviews 63:128-48.

Grimaldi D, Engel M (2005) The Evolution of the Insects. Cambridge University Press.

29 Jones RAC, Coutts BA, Latham LJ, and McKirdy SJ (2008) Cucumber mosaic virus infection of chickpea stands: temporal and spatial patterns of spread and yield-limiting potential. Plant Pathology 57: 842-853.

Larsen RC, Miklas PN, Eastwell KC, et al. (2002) A virus disease complex devastating late season snap bean production in the Midwest. Annual Report of the Bean Improvement Coop 45: 36-37.

Leonard MD (1963) A list of the aphids of New York. Proceedings of the Rochester Academy of Science 10:289-428.

Miller G. United States National Collection of Aphididae. http://www.sel.barc.usda.gov/aphid/aphframe.htm. Accessed 22 Apr 2012

Nault BA, Shah DA, Dillard HR, McFaul AC (2004) Seasonal and spatial dynamics of alate aphid dispersal in snap bean fields in proximity to alfalfa and implications for virus management. Environmental Entomology 33:1593-1601.

Nault BA, Shah DA, Straight KE, et al. (2009) Modeling temporal trends in aphid vector dispersal and cucumber mosaic virus epidemics in snap bean. Environmental Entomology 38:1347-59.

Ng JCK, Falk BW (2006) Virus-vector interactions mediating nonpersistent and semipersistent transmission of plant viruses. Annual review of Phytopathology 44:183-212.

Pepper JO (1965) A list of the Pennsylvania Aphididae and their host plants (Homoptera). Transactions of the American Entomological Society 91:181-231.

Pfleeger TG, Olszyk D, Burdick CA et al. (2006) Using a geographic information system to identify areas with potential for off-target pesticide exposure. Environmental Toxicology and Chemistry 25(8): 2250-2259.

Smith C, Eckel R, Lampert E (1992) A key to many of the common alate aphids of North Carolina (Aphididae: Homoptera). North Carolina Agriculture Research Service Technical Bulletin 299.

Ragsdale DW, Voegtlin DJ, and O’Neil RJ (2004) Soybean aphid biology in North America. Annals of the Entomological Society of America 97(2): 204-208.

Webb SE, Kok-Yokomi ML, Voegtlin DJ (1994) Effect of trap color on species composition of alate aphids (Homoptera: Aphididae) caught over watermelon plants. The Florida Entomologist 77:146-154.

30 Zitter TA, and Murphy JF (2009) Cucumber mosaic virus. The Plant Health Instructor. http://www.apsnet.org/edcenter/intropp/lessons/viruses/Pages/Cucumbermosaic.aspx. Accessed 28 May 2012.

Figures and Tables

Percent of Percent of Percent of Species Total Total Total Total Total Total Therioaphis trifolii 2274 30.4 535 40.0 2809 31.8 Aphis glycines 1475 19.7 131 9.8 1606 18.2 Acyrthosiphon pisum 1106 14.8 28 2.1 1134 12.9 Rhopalosiphum maidis 685 9.2 75 5.6 760 8.6 Pemphigus populicaulis 239 3.2 0 0.0 239 2.7 Aphis craccivora 179 2.4 123 9.2 302 3.4 Aphis gossypii 130 1.7 201 15.0 331 3.8 Hayhurstia atriplicis 128 1.7 1 0.1 129 1.5 Lipaphis erysimi 128 1.7 0 0.0 128 1.5 Myzus persicae 97 1.3 26 1.9 123 1.4 Capitophorus eleagni 79 1.1 7 0.5 86 1.0 Aphis sp. 77 1.0 0 0.0 77 0.9 Rhopalosiphum padi 77 1.0 45 3.4 122 1.4 Aphis fabae 15 0.2 14 1.0 29 0.3 Anoecia sp. 1 <0.1 14 1.0 15 0.2 Brachycaudus persicae 2 <0.1 15 1.1 17 0.2 Unknown 216 2.9 38 2.8 254 2.9 Others 576 7.7 84 6.3 660 7.5 Total 7484 100.0 1337 100.0 8821 100.0

Aphid Species PA NY Wallis Primary Host Secondary Host Source Acyrthosiphon kondoi yes yes Leguminosae, Trifoleae, Loteae AWC Acyrthosiphon pisum yes yes yes Leguminosae, Genistae, Trifoleae, Fabae, Hedysareae AWC Amphorophora rubi yes yes Rubus spp. HPS Anoecia corni yes yes yes Cornus sanguinea Gramineae AWT Anoecia cornicola yes Cornus spp. Gramineae AWT Anoecia oenotherae yes Cornus spp. Oenothera biennis AWT Aphis armoraciae yes yes Compositae, Cruciferae, Umbelliferae, Graminae AWC Aphis cephalanthi yes Cephalanthus occidentalis HPS Aphis craccivora yes yes polyphagous, Leguminosae AWC Aphis fabae yes yes yes Euonymous europaeus, Viburnum opulus polyphagous AWC Aphis forbesi yes Fragaria spp. HPS Aphis glycines yes yes Rhamnus spp. Glycine max AWC Aphis gossypii yes yes yes Catalpa, Hibiscus, Celastrus, Rhamnus, Punica polyphagous, cotton, cucurbits all Aphis helianthi yes yes Cornus stolonifera Umbelliferae AWT Aphis lugentis yes Senecio spp., Erigeron sp. HPS Aphis nasturtii yes Rhamnus cathartica, R. alnifolia wide range HPS Aphis oestlundi yes Oenothera biennis HPS Aphis pomi yes Pyroidea AWC Aphis pseudohedrae yes Aphis pulchella yes yes Euphorbia HPS Aphis rubifolii yes Rubus spp. HPS Aphis rumicis yes Rumex spp., Rheum spp. HPS Aphis spiraecola yes yes Citrus, Spiraea spp., polyphagous AWC Aphis viburniphila yes yes Viburnum spp. HPS Aulacorthum solani yes yes Polyphagous HPS Brachycaudus persicae yes yes Prunus persica, P. armeniaca Scrophulariaceae AWT Brevicoryne brassicae yes yes Cruciferae AWC Capitophorus elaeagni yes yes yes Elaeagnus spp. tubuliferous Compositae AWT Capitophorus hippophaes yes yes Elaeagnaceae Polygonum spp., Persicaria AWT spp. Carolinaia rhois yes Rhus glabra, R. typhina Gramineae AWC

Chaitophorus populifolii yes yes Populus spp. AWT Cinara atlantica yes Pinus spp. AWT Drepanaphis acerifoliae yes Acer saccharinum, A. rubrum, A. saccharum AWT Drepanaphis carolinensis yes Acer saccharum, A. rubrum AWT Drepanaphis nigricans yes yes Acer rubrum AWT Drepanaphis sabrinae yes Acer saccharum AWT Drepanosiphum platanoidis yes Acer pseudoplatanus, Acer spp., sycamore AWT Dysaphis plantaginea yes Malus spp., Pyrus Plantago spp. AWT Dysaphis tulipae yes many monocots AWC Eriosoma lanigerum yes Pyroidea, apple, Crataegus, Coloneaster AWC Essigella pini yes yes Pinus spp. AWT Eulachnus rileyi yes yes Pinus spp. AWT Geoica squamosa yes Hayhurstia atriplicis yes yes Chenopodiaceae, Atriplex, Chenopodium spp. HPS Hyadaphis foeniculi yes Lonicera spp. Umbelliferae spp. AWC Hyalopterus pruni yes Prunus domestica, P. armeniaca Phragmites communis, Arundo AWC donax Hydaphias foeniculi yes Lonicera spp. Umbelliferae HPS Hyperomyzus lactucae yes yes Ribes spp. Sonchus spp. AWC Hysteroneura setariae yes Prunus domestica Graminieae AWC Illinoia liriodendri yes Liriodendron tulipifera AWT Kaltenbachiella ulmifusa yes Ulmus rubra Labiatae AWT Lipaphis erysimi yes yes Cruciferae AWC Macrosiphoniella yes Artemisia ludoviciana, A. vulgaris HPS ludovicianae Macrosiphoniella sanborni yes Dendranthema indicum, morifolium, frutescens, AWC Compositae Macrosiphum euphorbiae yes yes yes Rosa spp highly polyphagous, AWC Solanaceae Macrosiphum pallidum yes yes Rosaceae, Rosa spp. AWC Macrosiphum pseudocoryli yes Ostrya virginiana, Corylus spp. AWT Macrosiphum rosae yes Rosa spp. Dipsacaceae AWC Melaphis rhois yes Rhus spp. (glabra, typhina) mosses AWT Monellia caryella yes yes yes Carya spp. AWC Myzus persicae yes yes yes Prunus persica, Prunus spp. polyphagous, over 40 families AWC Nearctaphis bakeri yes yes yes Crataegus, Cydonia, Malus, Pyrus Leguminosae AWC Nearctaphis clydesmithi yes yes Crataegus unknown AWT

Nearctaphis crataegifoliae yes yes yes Crataegus spp Trifolium spp. AWC Ovatus crataegarius yes Crateagus spp. Labiatae esp Mentha AWT Pemphigus populicaulis yes yes Populus deltoides, P. tremuloides unknown AWT Pemphigus yes yes Populus spp. Cruciferae AWC populitransversus Pemphigus populivenae yes yes Populus spp. Chenopodiaceae AWT Periphyllus americanus yes Acer spp. AWT Periphyllus testudinaceus yes Acer spp., Aesculus spp. AWT Phorodon humuli yes Prunus spp. Humulus lupulus (hops) AWC Prociphilus fraxinifolii yes Fraxinus spp. AWT Protaphis middletonii yes yes Compositae, Cruciferae, Graminae, and others HPS Pterocomma bicolor yes Populus spp., Salix spp. AWT Pterocomma smithiae yes Populus spp., Salix spp. AWT Rhodobium porosum yes yes Rhopalomyzus poae yes yes Lonicera alpigena grasses AWC Rhopalosiphoninus yes bulbs (Tulipa, Gladiolus), runners AWC latysiphon Rhopalosiphum maidis yes yes yes Gramineae AWC Rhopalosiphum nymphaeae yes Prunus spp. water plants AWC Rhopalosiphum yes yes Alus, Pyrus, Cotoneaster, Crataegus, Sorbus grasses AWC oxyacanthae Rhopalosiphum padi yes yes yes Prunus virginiana Gramineae AWC Rhopalosiphum yes yes Prunus spp. Gramineae, Cyperaceae, AWC rufiabdominale Solanaceae Schizaphis graminum yes yes Gramineae AWC Sipha flava yes yes Gramineae AWC Sipha glyceriae yes Gramineae AWC Sitobion avenae yes yes Gramineae AWC Tetraneura akinire yes Tetraneura nigriabdominalis yes yes Ulmus spp. Gramineae AWC Therioaphis riehmi yes Melilotus spp. HPS Therioaphis trifolii yes yes yes Leguminoseae AWC Uroleucon anomalae yes Aster novaeangliae HPS Uroleucon pseudambrosiae yes Compositae, Lactuca spp. HPS Utamphorophora crataegi yes Crataegus spp. AWT Vesiculaphis caricis yes Rhododendron spp. Cyperus spp. HPS

Table 2-3. New aphid records from PA reported in Nault et al (2009) and/or Wallis et al (2005) but not found in Pepper (1965). Species Nault et al Wallis et al Acyrthosiphon kondoi yes Aphis armoraciae yes yes Aphis glycines yes Aphis lugentis yes Aphis pulchella yes Lipaphis erysimi yes Nearctaphis clydesmithi yes Tetraneura nigriabdominalis yes

Table 2-4. Species in six subfamilies of the family Aphididae occurring in PA. Subfamily Tribe Species Anoeciinae Anoecia corni Anoecia cornicola Anoecia oenotherae Anoecia setariae Hormaphidinae Cerataphidini Cerataphis brasiliensis Cerataphis lataniae Hormaphidini Hamamelistes spinosus Hormaphis hamamelidis Mindarinae Mindarus abietinus Phyllaphidinae Phyllaphis fagi Stegophylla quercicola Stegophylla quercifoliae Stegophylla quercina Pterocommatinae Fullawaya terricola Plocamaphis flocculosa Pterocomma bicolor Pterocomma medium Pterocomma populifoliae Pterocomma smithiae Saltusaphidinae Saltusaphidini Iziphya flabella Iziphya grandipes Iziphya vittata Strenaphis elongata Thripsaphidini Allaphis verrucosa Subsaltusaphis virginica Thripsaphis ballii

Table 2-5. Species in the subfamily Aphidinae, tribe Macrosiphini occurring in PA

Abstrusomyzus phloxae Dysaphis plantaginea Macrosiphum carpinicolens Pleotrichophorus glandulosus Acuticauda solidaginifoliae Dysaphis tulipae Macrosiphum coryli Pleotrichophorus patonkus Acyrthosiphon kondoi Ericaphis scammelli Macrosiphum cystopteris Pleotrichophorus wasatchii Acyrthosiphon lactucae Ericaphis wakibae Macrosiphum euphorbiae Pseudacaudella rubida Acyrthosiphon malvae Hayhurstia atriplicis Macrosiphum gaurae Rhodobium porosum Acyrthosiphon malvae malvae Hayhurstia atriplicis atriplicis Macrosiphum gei Rhopalosiphoninus latysiphon Acyrthosiphon pisum Hyadaphis foeniculi Macrosiphum geranii Rhopalosiphoninus solani Acyrthosiphon pseudodirhodum Hyalomyzus collinsoniae Macrosiphum lilii Rhopalosiphoninus staphyleae Amphorophora agathonica Hyalomyzus eriobotryae Macrosiphum pallidum Rhopalomyzus lonicerae Amphorophora ampullata laingi Hyalomyzus humboldti Macrosiphum pseudocoryli Rhopalomyzus poae Amphorophora rossi Hyalomyzus mitchellensis Macrosiphum ptericolens Sitobion avenae Amphorophora rubi Hyalomyzus sensoriatus Macrosiphum rosae Sitobion avenae avenae Amphorophora sensoriata Hyadaphis pseudobrassica Macrosiphum siriodentri Uroleucon ambrosiae Aulacorthum solani Hyalopteroides humilis Macrosiphum tiliae Uroleucon ambrosiae ambrosiae Aulacorthum solani solani Hyperomyzus lactucae Mastopoda pteridis Uroleucon anomalae Brachycaudus cardui Hyperomyzus nabali Metopolophium dirhodum Uroleucon caligatum Brachycaudus helichrysi Hyperomyzus picridis Microlophium sibiricum Uroleucon chrysanthemi Brachycaudus knowltoni Idiopterus nephrelepidis Microlophium sibiricum sibiricum Uroleucon chrysopsidicola Brachycaudus persicae Illinoia azaleae Microparsus desmodiorum Uroleucon erigeronense Brachycaudus rociadae Illinoia azaleae azaleae Microparsus kislankoi Uroleucon eupatoricolens Brachycaudus rumexicolens Illinoia azaleae kalmiaflora Microparsus olivei Uroleucon eupatorifoliae Brachycaudus schwartzi Illinoia borealis Microparsus singularis Uroleucon floricola Brachycorynella asparagi Illinoia canadensis Muscaphis musci Uroleucon gravicorne Brevicoryne brassicae Illinoia goldmaryae Myzaphis rosarum Uroleucon helianthicola Cachryphora canadensis Illinoia liriodendri Myzodium modestum Uroleucon illini Cachryphora serotinae Illinoia pepperi Myzus avenae Uroleucon impatiensicolens Capitophorus carduinus Illinoia rhokalaza Myzus cerasi Uroleucon lanceolatum Capitophorus elaeagni Illinoia richardsi Myzus formosanus Uroleucon leonardi

Capitophorus hippophaes Illinoia rubicola Myzus lythri Uroleucon luteolum Carolinaia caricis Illinoia spiraecola Myzus ornatus Uroleucon nigrotibium Carolinaia howardii Linosiphon sanguinarium Myzus persicae Uroleucon nigrotuberculatum Carolinaia rhois Liosomaphis berberidis Nasonovia aquilegiae Uroleucon obscuricaudatum Catamergus kickapoo Lipaphis erysimi Nasonovia compositellae Uroleucon paucosensoriatum Cavariella aegopodii Lipaphis pseudobrassicae Nasonovia cynosbati Uroleucon pepperi Cavariella cicutae Longicaudus trirhodus trirhodus Nasonovia heucherae Uroleucon pieloui Cavariella hendersoni Macrosiphoniella abrotani Nasonovia purpurascens Uroleucon pseudambrosiae Cavariella pastinacae Macrosiphoniella cystopteris Nasonovia ribisnigri Uroleucon rudbeckiae Cavariella salicis Macrosiphoniella frigidicola Nearctaphis bakeri Uroleucon rurale Cavariella theobaldi Macrosiphoniella leucanthemi Nearctaphis clydesmithi Uroleucon russellae Ceruraphis eriophori Macrosiphoniella ludovicianae Nearctaphis crataegifoliae Uroleucon sonchellum Ceruraphis viburnicola Macrosiphoniella millefolii Nearctaphis crataegifoliae crataegifoliae Uroleucon sonchi Chaetosiphon fragaefolii Macrosiphoniella pennsylvanica Neomyzus circumflexus Uroleucon sonchi sonchi Chaetosiphon minor Macrosiphoniella sanborni Neotoxoptera formosana Uroleucon taraxaci Chaetosiphon minor minor Macrosiphoniella siriodendri Neotoxoptera violae Uroleucon tardae Chaetosiphon tetrarhodum Macrosiphoniella subterranea Ovatus crataegarius Uroleucon tuataiae Coloradoa rufomaculata Macrosiphoniella tanacetaria Papulaphis sleesmani Utamphorophora crataegi Cryptomyzus ribis Macrosiphoniella tapuskae Phorodon humuli Utamphorophora humboldti Decorosiphon corynothrix Macrosiphum adianti Pleotrichophorus ambrosiae Diuraphis holci Macrosiphum californicum Pleotrichophorus asterifoliae

Table 2-6. Species in the subfamily Aphidinae, tribe Aphidini occurring in PA

Aphis armoraciae Aphis helianthi Aphis varians Aphis angelicae Aphis illinoisensis Aphis vernoniae Aphis asclepiadis Aphis impatientis Aphis viburniphila Aphis caliginosa Aphis lugentis Protaphis knowltoni Aphis carduella Aphis maculatae Protaphis middletonii Aphis cephalanthi Aphis melliferum Sanbornia juniperi Aphis coreopsidis Aphis middletonii Toxoptera viridirubra Aphis cornifoliae Aphis nasturtii Hyalopterus pruni Aphis craccivora Aphis neilliae Hysteroneura setariae Aphis debilicornis Aphis nerii Pseudasiphonaphis corni Aphis decepta Aphis oenotherae Rhopalosiphum cerasifoliae Aphis fabae Aphis oenotherae sanborni Rhopalosiphum enigmae Aphis farinosa Aphis oestlundi Rhopalosiphum maidis Aphis feminea Aphis pawneepae Rhopalosiphum musae Aphis folsomii Aphis pomi Rhopalosiphum niger Aphis forbesi Aphis pseudohedrae Rhopalosiphum nymphaeae Aphis frangulae Aphis pulchella Rhopalosiphum oxyacanthae Aphis gerardiae Aphis rubicola Rhopalosiphum padi Aphis glycines Aphis rubifolii Rhopalosiphum parvae Aphis gossypii Aphis rumicis Rhopalosiphum rufiabdominale Aphis hamamelidis Aphis sambuci Rhopalosiphum sanguinarium Aphis hederae Aphis spiraecola Schizaphis graminum Aphis hederae pseudohederae Aphis spiraephila Schizaphis nigra

39 Table 2-7. Species in the subfamily Calaphidinae occurring in PA

Tribe Species Calaphidini Betulaphis quadrituberculata Callipterinella calliptera Euceraphis mucida Calaphis alni Cepegillettea myricae Hannabura alnosa Calaphis betulaecolens Euceraphis betulae Euceraphis punctipennis Calaphis betulella Euceraphis gillettei Calaphis leonardi Euceraphis lineata Panaphidini Hoplochaitophorus heterotrichus Myzocallis melanocera Monellia hispida Hoplochaitophorus quercicola Myzocallis multisetis Monellia microsetosa Lachnochaitophorus obscurus Myzocallis punctata Monelliopsis bisselli Myzocallis alhambra Myzocallis spinosa Monelliopsis caryae Myzocallis asclepiadis Myzocallis synthri Monelliopsis nigropunctata Myzocallis bellus Myzocallis tuberculata Protopterocallis fumipennella Myzocallis castaneae Myzocallis walshii Protopterocallis gigantea Myzocallis castanicola Neosymydobius albasiphus Protopterocallis pergandei Myzocallis coryli Patchia virginiana Pterocallis alnifoliae Myzocallis discolor Tuberculatus punctatellus Therioaphis ononidis Myzocallis exultans Chromaphis juglandicola Therioaphis riehmi Myzocallis frisoni Eucallipterus tiliae Therioaphis trifolii Myzocallis granovskyi Melanocallis caryaefoliae Therioaphis trifolii maculata Myzocallis longiunguis Monellia caryella Tinocallis ulmifolii

Table 2-8. Species in the subfamily Chaitophorinae occurring in PA.

Tribe Species Chaitophorini Chaitophorus longipes Chaitophorus populifolii Chaitophorus viminicola Chaitophorus nigrae Chaitophorus populifolii simpsoni Periphyllus americanus Chaitophorus nigricentrus Chaitophorus pusillus Periphyllus californiensis Chaitophorus nudus Chaitophorus saliniger Periphyllus lyropictus Chaitophorus populicola Chaitophorus stevensis Periphyllus negundinis Chaitophorus populifoliae Chaitophorus viminalis Siphini Sipha elegans Sipha flava Sipha glyceriae

Table 2-9. Species in the subfamily Drepanosiphinae occurring in PA

Drepanaphis acerifoliae Drepanaphis nigricans Drepanaphis simpsoni Drepanaphis carolinensis Drepanaphis parva Drepanaphis spicata Drepanaphis kanzensis Drepanaphis platanoides Drepanosiphum platanoidis Drepanaphis monelli Drepanaphis sabrinae Shenahweum minutum

40 Table 2-10. Species in the subfamily Eriosomatinae occurring in PA

Tribe Species Eriosomatini Colopha graminis Eriosoma lanigerum Eriosoma wilsoni Colopha ulmicola Eriosoma lanuginosum Kaltenbachiella ulmifusa Eriosoma americanum Eriosoma mimicum Tetraneura nigriabdominalis Eriosoma crataegi Eriosoma rileyi Tetraneura ulmi Fordini Forda marginata Geoica pellucida Melaphis rhois Forda olivacea Geoica ultricularia Smynthurodes betae Pemphigini Grylloprociphilus imbricator Pemphigus populitransversus Prociphilus longianus Mordwilkoja vagabunda Pemphigus populivenae Prociphilus probosceus Neoprociphilus aceris Prociphilus americanus Prociphilus tessellatus Pachnypappa pseudobyrsa Prociphilus caryae Thecabius affinis Pemphigus bursarius Prociphilus caryae fitchii Thecabius gravicornis Pemphigus monophagus Prociphilus corrugatans Thecabius populimonilis Pemphigus nortonii Prociphilus erigeronensis Pemphigus populicaulis Prociphilus fraxinifolii

Table 2-11. Species in the subfamily Lachninae occurring in PA

Tribe Species Eulachnini Cinara atlantica Cinara laricifex Cinara taedae Cinara banksiana Cinara laricis Cinara tujafilina Cinara braggii Cinara osborniana Cinara watsoni Cinara canatra Cinara pergandei Essigella pini Cinara costata Cinara pilicornis Eulachnus agilis Cinara cupressi Cinara pinea Eulachnus americanus Cinara fornacula Cinara pinivora Eulachnus rileyi Cinara gracilis Cinara pruinosa Schizolachnus parvus Cinara harmonia Cinara pruinosa pruinosa Schizolachnus piniradiatae Cinara juniperi Cinara spiculosa Cinara juniperivora Cinara strobi Lachnini Lachnus allegheniensis Longistigma trirhodus Tuberolachnus salignus Longistigma caryae Tramini Trama rara

100

50 PA 40

Mean # # speciesMean NY 30

0 0 1000 2000 3000 4000 5000 6000 7000 8000 individuals

Figure 2-1. Individual-based rarefaction curves showing aphid species accumulation in PA and NY.

110

100

Mean # # speciesMean 40

0 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 individuals

Crop 29% Herb/grass Woody Unknown 61%

Chapter 3

Estimating natal sources of Aphis glycines using molecular markers and airflow trajectories

Introduction

Aphis glycines is an important pest of soybean as well as a competent virus vector in other legumes (Gildow et al 2008, Nault et al 2009). Since its introduction to the US around 2000, this invasive species has spread through the Midwest and into the Northeast, wherever soybeans are plentiful (Ragsdale et al 2004, Ragsdale et al 2011). Crop losses in soybean resulting from A. glycines if left untreated have the potential to total $2.4 billion annually (Song et al 2006). These dramatic losses were the first time North American soybean growers had to contend with a significant insect pest, and this spurred research on its life history and effective control methods

(Tilmon et al 2011, Hodgson et al 2012). A. glycines is important because of the physical damage high populations cause to soybean, but it is of special concern in Pennsylvania because it is an efficient vector of legume viruses (Gildow et al. 2008).

Aphis glycines life history

Soybean aphid (Hemiptera: Aphididae) is heteroecious and holocyclic. It alternates between sexual and asexual reproduction (holocyclic) and requires two hosts to complete its lifecycle (heteroecious) – the primary host, Rhamnus spp. (buckthorn) is a woody invasive shrub and the secondary host is soybean (Voegtlin et al 2004 and 2005).

44 Factors limiting the range of Aphis glycines

Rhamnus cathartica is the North American overwintering host for Aphis glycines. In its native range, R. davurica is the overwintering host for A. glycines but it is not present in North

America. R. cathartica distribution is thought to be one of the limiting factors in where A. glycines can maintain local populations. Determining its range is important when assessing the contributions of migrants versus a local population for IPM.

Rhamnus cathartica is a shrub/small tree (2-6 m) that is native to Europe and Asia

(Archibold et al. 1997, Converse 1985). It was introduced to North America in the 1800s for windbreaks and hedgerows because of its tolerance for many environmental conditions and eventually escaped from cultivation. Now, R. cathartica commonly invades forests and open fields where the soil is moist (Archibold et al. 1997, Kurylo et al. 2007). It can out-compete native species in shaded areas to become a dominant species in the understory (Archibold et al.

1997). Most of the fruit falls beneath the parent tree where seedlings survive well, but some are eaten and dispersed by birds, mice, and white-tailed deer (Archibold et al. 1997, Myers et al.

2004). When R. cathartica is removed from an area the native vegetation returns quickly as long as the seed bank is viable, indicating that R. cathartica may be alleopathic (Bodreau 1992).

Two previous experiments confirmed R. cathartica as a suitable host for Aphis glycines

(Yoo et al. 2005, Voegtlin et al. 2005). On plants in outdoor cages, the soybean aphid overwintered successfully on three Rhamnus species (R. cathartica, R. alnifolia, R. lanceolata:

Voegtlin et al. 2005). Successful over-wintering required developed colonies with hatching eggs and the production of spring migrants. Alates were also observed feeding on R. cathartica

(Voegtlin et al. 2005). Aphids were not observed on the non-Rhamnus species in the experiment

(Voegtlin et al. 2005).

45 Yoo et al. (2005) confirmed R. cathartica and R. alnifolia as suitable hosts for A. glycines through a single-choice experiment. The aphids were induced to produce sexual morphs in the laboratory using an autumnal light and temperature regime. Gynoparae and oviparae were caged with a member of the Rhamnaceae family and a soybean plant. Gynoparae survived the longest and produced more nymphs on R. cathartica and R. alnifolia. The other Rhamnaceae taxa were unsuitable hosts because the soybean aphid could not develop on them in the absence of an alternative plant choice.

Current IPM tools and plant databases do not accurately represent the distribution of R. cathartica in Pennsylvania. The PAPIPE operates under the assumption that the entire state is suitable habitat in its A. glycines phenology model. Other sites, like plants.gov, do not have records from every county and one specimen is enough to consider that county suitable for the species.

In the summer of 2006 I scouted local forests and fencerows adjacent to snap bean fields for R. cathartica. I planned on surveying the forests closest to the six snap bean fields used in our

CMV study (Chapter 2) using transects that ran perpendicular to the edge forest edge. At that point, I wanted to measure how deep into the forest this species was present, but recent papers suggest that it is predominantly an edge species. Ultimately, I did not find any R. cathartica after applying the transect method once to each of the six field sites, although it is present in parks (e.g.

Sunset Park) and backyards in State College.

46 Aphis glycines in Pennsylvania

Initially, an outbreak of cucumber mosaic virus (CMV) in snap beans posed a risk to the snap bean processing crop in Pennsylvania. CMV is a non-persistently transmitted virus that is vectored by multiple aphid species. Soybean aphid is a very competent vector of the legume strain of CMV in snap beans, and is a significant component of the epidemiology of CMV in processing snap beans (Gildow et al. 2008, Nault et al. 2009). This led to further work, focusing on the movement of this virus vector in legumes.

In Pennsylvania, soybean aphid is present during the growing season, but the primary host species, common buckthorn (Rhamnus cathartica), needed for overwintering populations to exist is scarce. The scarcity of common buckthorn in Pennsylvania – especially as compared to its distribution in parts of the Midwest and northern New York, may severely limit the overwintering populations. If this is true, then management can be improved by tracking migration and dispersal, since population establishment in Pennsylvania is influenced heavily by long-distance dispersal as opposed to that from local overwintering hosts.

Soybean management

Current conventional pest management in soybean relies on the use of neonicotinoid seed treatments (e.g. imidacloprid, thiamethoxam). Those systemic seed treatments are effective against aphids, but for limited times. Pesticide trials in the Midwest found that seed treatments could effectively manage A. glycines until the soybean R2 stage or until July – August when the pest pressure increases (Rice and O’Neal 2008 and Schulz et al 2011). With early-season effectiveness, seed treatment value would be greatest when significant populations are appearing during the first stages of plant growth – and this will be much more likely in geographic areas

47 where overwintered populations are occurring (Bahlai et al 2010). In contrast, if we find that most of the early population establishment in Pennsylvania requires successful long-distance migration, then we may find less utility from insecticidal seed treatments, at least for earlier planted crops. Additional management implications for biological control, such as the conservation of natural enemies, exist as well (Brewer and Noma 2010, Costamagna and Landis

2011, Zhang and Swinton 2009). Thus, to better manage soybean aphid in Pennsylvania it helps to understand their population origins: are our summer populations from local colonies or are they migrants from neighboring states?

Aerobiology

Aphids can move locally and over long distances. Alate morphs can be produced as a result of a variety of cues including population density on a host plant, nutritional quality, interaction with natural enemies, and temperature (Müller et al 2001). Alate aphids are weak fliers and can successfully move between plants and fields. Alate aphids are not strong enough to resist updrafts, and individuals that find, or propel, themselves in weather systems can be transported long distances. If the deposition occurs on or near a suitable host, this can seed populations in new areas or add to an existing population.

Air flow trajectory models can be used to better understand insect migrations, especially those of insects (like aphids) that are not strong fliers. HYSPLIT (Hybrid Single Particle

Lagrangian Integrated Trajectory Model) is a computer model to calculate simple air parcel trajectories (Draxler and Rolph 2012). This model can calculate simple forward and back trajectories from single or multiple starting points for a single particle or plume. HYSPLIT is used by meteorologists and others to calculate or forecast particle dispersion. Recent uses include

48 tracking dust storms and volcanic ash plumes after an eruption (Stunder et al 2007, Wang et al

2011).

In agricultural systems, aerobiological models and simulations were used to track and forecast the movement of the fungal spores (Phakopsora pachyrhizi) that cause soybean rust.

Using information about soybean rust biology, a scouting program, and meteorological data, Isard et al (2007) created a model that was used to successfully predict the appearance of the pathogen in the Ohio River valley.

Soybean rust was monitored with a field scouting program set up from Florida to southern Ontario, Canada (IPM-PIPE). A similar program was in effect for soybean aphid for the eastern half of the US and southern Canada from 2007-2010. Each state in the program established sentinel fields and monitored them throughout the soybean growing season for soybean aphid. The weekly counts were reported to a website which displayed a map of the catches. This tool was good for showing where the populations were high and/or increasing, but it did not have any forecast capabilities.

Since aphids, including soybean aphid, fly at speeds lower than the relative movement of the air, when they are transported in an updraft they cannot escape and move with that air packet

(Dixon 1985). This movement of air packets is what HYSPLIT can be used to model. In 2005, a suction trap network using ~6 m tall traps was established in the Midwest to monitor the movement of alate aphids, specifically A. glycines. The network was successful in catching alate aphids and noted peak catches of A. glycines in the summer (late July – mid-August) and fall

(Schmidt et al 2012).

49 Molecular tools for population identification

In order to better understand the origin of soybean aphid in Pennsylvania, we used microsatellite markers and single nucleotide polymorphisms (SNPs) to investigate its population structure. We looked at temporal change in allele frequency by collecting aphids from soybean every other week during the summer of 2009 and 2010 from unsprayed sentinel fields in central

Pennsylvania. Aphids from early August 2009 in Pennsylvania were compared to those collected close to the same time from soybean in New York and Virginia to identify any spatial differences in allele frequency. This was repeated in 2010 with samples from New York, Quebec, and

Ontario. When samples from Pennsylvania matched those in other geographic regions, we used

HYSPLIT to determine temporally relevant weather systems that could have been responsible for their immigration. This information will further our understanding of aphid movement to states that do not have a strong overwintering population.

Michel et al 2009a and 2009b provide the basis for this work. They established effective microsatellites for soybean aphid and looked at their spatial differentiation in the Midwest and

South Korea (part of its native range). Michel et al (2009a) found that much of the genotypic variation was due to collection time. Since the Eastern US was not included in that data, we worked together to establish a sampling plan for Pennsylvania and contacted collaborators in

Ontario, Quebec, New York and Virginia. We hypothesize that Pennsylvania differs from portions of the Midwest regarding soybean aphid in that it does not have widespread overwintering habitat. Thus, early-season populations in Pennsylvania may match with those from the Midwest, which would support the conclusion that these are migrants.

We included the sites in New York and Virginia as positive and negative controls for

Rhamnus presence, respectively. Buckthorn is present in Northern New York in forests and

50 adjacent to snap bean fields. In Virginia, buckthorn is only recorded in one northern county and our collection sites in the southeastern part of the state are well outside of its range.

Objectives

The focus of this study was to investigate the potential of using molecular markers and airflow trajectories to provide an estimation of the natal sources of A. glycines that were found in areas with no consistent over-wintering population. With the molecular evidence serving as a proof of concept, future efforts to track or predict the movement of soybean aphid could include the use of HYSPLIT forecast trajectories. Here, we used archived weather information to compute forward trajectories between points where aphids with matching genotypes were collected during our study.

Methods

Field collection

In Pennsylvania, Aphis glycines was collected weekly in an unsprayed soybean field at the Russell E. Larson Agricultural Research and Extension Center in Centre County (2009 and

2010) and on a Penn State farm near the University Park Airport (2010). At least 30 randomly selected plants > 1 m apart were searched, and any aphids found were collected and preserved in

70% EtOH. Fields were checked weekly for aphids beginning June 1. Collecting began when aphids first appeared in the field (late June or early July) and continued biweekly until September.

Cooperators in other states/provinces (New York, Virginia, Ontario and Quebec) collected aphids

51 in the same manner at one or two time points and sent the samples to University Park (Table 3-1 and Figure 3-1).

Laboratory Methods

The collected aphids were taken to Dr. Michel’s lab at OARDC for analysis as per the methods in Michel et al. 2009a and Michel et al. 2009b. Aphid genomic DNA was extracted using the E.Z.N.A. Tissue DNA kit (Omega Bio-tek, Norcross, GA), isolated, tagged with fluorescent primers, and amplified using PCR (PCR cycling in Michel et al 2009a). Six microsatellites described in Michel et al (2009) were used in this study: Ago 66, Ago 69, Ago 89,

AF 85, AF 181, and AF I. Microsatellite genotyping was performed using the Beckman-Coulter

CEQ8800XL (Fullerton, CA) at the Molecular Cellular and Imaging Center at OARDC (Wooster,

OH) and individual genotypes were manually scored using the CEQ Fragment Analysis software

(version 9.0.25).

Seventeen of the SNPs developed and tested in Bai et al (2010) and Orantes et al (2012) were used on the 2010 samples. The 2009 samples were unavailable to be reanalyzed with the

SNPS because they were lost during the 2010 Wooster tornado.

Statistical Methods

Fst values and the Bonferroni corrected P-values were calculated using Microsatellite

Analyzer 4.05 (Dieringer et al 2003). Fst is the mean reduction in heterozygosity in a subpopulation relative to the entire population. It is the measure of the extent of genetic differentiation among subpopulations. Fst values range from 0, which indicates no differentiation

52 to 1, which indicates complete differentiation. Clone analysis, genetic distance, heterozygosity, and principle component analysis was completed with GenAlEx 6.4 (Peakall and Smouse 2006).

HYSPLIT methods

We calculated trajectories for dates and locations in 2009 and 2010 where we observed matching haplotypes using archived data through the online HYSPLIT program (accessible at: http://ready.arl.noaa.gov/HYSPLIT_traj.php). For this study, we selected one starting location and an ensemble trajectory, which creates 27 trajectories from one point by offsetting each one slightly. This enabled us to estimate the range of possible ending locations from each starting point, as opposed to the simple trajectory which would only give one line from the starting point.

We calculated the archived forward trajectories for 3 scenarios in 2009 and one in 2010

(Table 3-9). The trajectory starting location was one of our fields in Pennsylvania where we found an aphid with a genotype that matched to another location. We used the EDAS 40km data for the relavent time periods, and set up the model parameters (see Appendix A). The following changes were made to the default settings: total run time was 48 hours, starting height was 100 m

AGL (above ground level), yes to ‘plot color trajectories,’ and label interval of 24 hours

(screenshots of the procedure in Appendix A). We produced trajectory maps for all of the dates in the month before the last matching aphid was collected.

We scored the maps we generated on the basis of how many of the forward trajectories crossed the location of the other matching aphid catch. The scores range from 0 to 1, with a score of 0 indicating that none of the trajectories on that map crossed the target location, and a score of

1 indicated that all of trajectories crossed the target location.

53 Results

Spatial

The PCA of the 2009 data indicates clustering over a large spatial scale, where 60% of the total variation was explained along the first axis (Figure 3-2). Populations from VA and NY clustered away from populations in central PA along this axis. However, the pairwise genetic differentiation (Fst) between A. glycines populations in Pennsylvania (PA), Virginia (VA) and

New York (NY) in 2009 ranged from 0.003 to 0.045 and were not significant (Table 3-2). This indicates low genetic differentiation between our sampled subpopulations. If all of the PA collection dates were combined into one population and compared with NY (Table 3-3), there was a low (0.020) but significant Fst value.

When comparing this data with data from the Midwest, significant Fst values emerged

(Table 3-5). The four PA collection dates were all significantly different from the early collections in OH-G, MN-L, SD, OH-T, and ON and the Fst values ranged from 0.056 (PA June

30 vs. OH-G1) to 0.235 (PA July 13 vs. SD-1). The earliest PA collection (June 30) was not significantly different from the early collection dates in MN-R, MI, and WI with Fst values of

0.031, 0.034 and 0.050 respectively. The PCA of the East and Midwest populations shows some clustering on the first axis (58.2% of the variation) with the PA, NY, and VA populations in a group separate from the majority of the Midwest populations (Figure 3-4). Of the Midwest populations, the early South Dakota one is separate from both groups. The second axis (20.6% of the variation) does not appear to separate the populations, except early collection from South

Dakota.

For 2010, we analyzed aphids collected from Pennsylvania, Quebec, Ontario and New

York using 17 SNPs. The first axis of the PCA of these populations explains 47.1% of the

54 variation and indicates some level of differentiation over a large spatial scale with the Canadian populations in a group separate from the US populations (Figure 3-3). This separation was supported by the Fst values which were significantly different between PA and the Canadian samples (Table 3-6, 5 of 17 comparisons with P < 0.05) and also on the whole larger than the Fst values between PA populations (0.336 Sherrington QC vs. Rock Springs Aug 30, 0.097 Rock

Springs Jul 5 vs. Rock Springs Aug 30).

We also compared the 2010 samples from the East (PA, Quebec, Ontario, NY) with the

2009 Midwest samples using 10 SNPs that were shared between both analyses. The PCA of this grouping clearly shows the differentiation between these populations (Figure 3-5). The Eastern populations are all to the left of zero on the first axis (69.7% of the variation) and the Midwest populations are to the right. Within the Midwest group, they do separate temporally on the second axis (12.8% of the variation) with the early collections on the edges of the group and the later collections coalescing at the center. All of the Fst values between the East and the Midwest were significantly different (P<0.05) and ranged from 0.090 to 0.213 (Table 3-5).

We looked at the number of unique (occurring only once) and matching (occurring more than once) genotypes per population to get an assessment of genotypic diversity (Table 3-4 and 3-

8). In addition to the high genotypic diversity reported from the PA sites, genotypic diversity was also high (0.81 to 0.97) for the non-PA sites.

In 2009, 60% (108) of the individuals were distinct clones, while 40% of the individuals were not distinct clones (Table 3-4). Four matching clones were found in both PA and NY, one clone was found in both PA and VA, and one clone was found in both NY and VA. The VA/PA clone was present during all sampling dates in PA and was one of the two most common clones

(both had frequencies of 2.7% or 5 individuals). With the incorporation of SNPs for the 2010 data

(Table 3-8), the number of matching clones decreased. There was only one clone that was shared

55 over a long distance (between Ottawa and PA Airport), and one clone that was shared over a shorter distance (between Rock Springs and Airport, ~15 miles).

Temporal

In 2009, a PCA of the microsatellite data suggested differentiation among collection dates in PA along the second orthogonal axis (Fig. 3-2), with the earliest (Jun 30) collection separating from the latest (Aug 31) collection, but Fst values ranged from 0 to 0.01 (Table 3-2) and there was no significant differences (P > 0.05, Bonferroni corrected). In 2010, a PCA of the

SNP data (Fig 3-3) also shows some separation by collection date with early (Jul 2 and Jul 5) samples separating from the late (Aug 30 and Aug 31) samples along a combination of both axes.

Fst values in 2010 were also low, but ranged up to 0.097, and they were significant (P < 0.05,

Bonferroni corrected) when comparing collections in PA from Jul 2 with Aug 30, and Jul 5 with both Aug 2 and Aug 30 (Table 3-7).

Genotypic diversity among dates in PA samples was high in both years [>= 0.86 when using the microsatellite data in 2009 (Table 3-4), and >= 0.81 when using the SNP data in 2010

(Table 3-7)]. The genotypic diversity decreased from 0.97 to 0.86 over time in 2009 (Table 3-4), but the opposite trend (an increase from 0.81 to 1.00) occurred in 2010 (Table 3-8). Overall, for both years regardless of temporal trend, genotypic diversity was high. In 2009, there were 16 clones found in the Rock Springs collections that were shared temporally. The 2010 Airport samples had clones that were shared over time.

56 Aerobiology

When there were matching clones between different sites (e.g. PA and VA), we used

HYSPLIT to determine if recent weather systems could have been responsible for their immigration. An example of this is shown in Figure 3-6. To obtain this figure, we ran forward ensemble trajectories at 48 hour intervals for the month of July to see if there were any possible weather patterns that could have deposited aphids between our study sites in PA and VA.

Ensemble trajectories are useful because the model initiates multiple trajectories from the same point that are slightly offset resulting in a plume rather than a single path. Figure 3-6 (top left) is a forward trajectory ensemble originating at Rock Springs, PA on July 13 and running for 48 hours at 100 meters above ground level. Eighteen of 27 of the trajectories crossed the collection site in

Virginia, resulting in a score of 0.67 on a scale of 0 to 1. Additional examples are displayed in Fig

3-6 that represent a range of scores.

Discussion

Based on these results, we demonstrated that the genotypic diversity of soybean aphids in

Pennsylvania was very high (0.81 to 1.00). In 2009, diversity decreased over time, but in 2010 the opposite occurred, therefore we could not define a consistent temporal trend in genotypic diversity. Genotypic diversity in a field is influenced by how many different clones initially colonize a field, how successful those clones are, and how many clones colonize the field as the season progresses and their success. Our results showing high initial diversity indicate that we have many clones colonizing our fields, and a few of them are present as the season progresses.

As the season progresses, some aphid clones are maintained, others are new migrants, and some clones die or are not resampled. The asexual reproduction of the aphids on soybean has the

57 potential to saturate a field with successful clones, thus decreasing its diversity, but we did not observe this.

On a spatial scale, it appears that we can use molecular markers to detect long distance movements. As we increased our markers from 6 microsatellites to 17 SNPs, it became harder to find shared clones over long distances. The aphids we collected in PA matched with NY and

Ottawa very few times. We did observe clone matching between PA and VA and NY and VA

(once for each), which suggests that we could use clone-matching to estimate the natal source of

VA aphids since the VA collection site is well outside the range of Rhamnus cathartica. The data suggest that with comparisons to more populations we could successfully estimate the natal source of soybean aphid populations in PA.

After including a dataset containing Midwest collections from 2009 and both microsatellite and SNP information, we could form a more comprehensive picture of aphid movement to PA. The Midwest data analyzed here was first used in Orantes et al (2012). In that paper, the authors observed lower levels of genotypic diversity in the early season collections

(0.68 – 0.97), and overlapping but higher levels in the late season collections (0.87 – 1.00). The genotypic diversity from the collections in PA, NY, Canada (2010) and VA (2009) more closely resembled that of the late season Midwest collections (0.86 – 0.97 in 2009, 0.81 – 1.00 in 2010).

This high genetic diversity and the lack of genetic differentiation between populations sampled in

PA suggest high levels of aphid movement (colonization and recolonization of fields), and the lack of a solid local population colonizing from surrounding buckthorn in the spring. Early season aphid density in soybean fields was found to be best predicted by the amount of buckthorn in the surrounding landscape in close proximity to the fields (Bahlai et al 2010). The low early season colonization densities we observed, combined with the high genotypic diversity, would be consistent with relatively rare colonization events expected from long distance migrants.

58 There are multiple avenues for future work necessary to successfully integrate molecular identification techniques and aerobiological modeling into a useful management and risk assessment tool. One would be the need to have a concerted effort to accurately represent the geographic range of buckthorn, the primary host. Since buckthorn is the limiting factor in where populations can overwinter, understanding its range is key to identifying local sources of the aphid. There are many assumptions inherent to modeling and predicting long distance aphid movement including the assumption that aphids actually enter the air column (demonstrated with the suction trap network, Schmidt et al 2012) and are then deposited at some point along the way.

Programs like HYSPLIT give a good visualization of where air parcels are going and with further investigation could be a useful forecasting tool.

The continued development of molecular techniques to identify aphid populations will be of use with the emergence of soybean cultivars with aphid resistance traits, and the subsequent aphid biotypes with resistance characters of their own. Also, if soybean aphid ever branches out to use any of the other Rhamnus species present in the landscape, these tools could be used to identify biotypes or subspecies. This merging of molecular techniques and aerobiology is not limited to this system, and could be expanded to other economically important pests.

References

Archibold OW, Brooks D, Delanoy L (1997) An investigation of the invasive shrub European buckthorn, Rhamnus cathartica L., near Saskatoon, Saskatchewan. Canadian Field Naturalist 111:617-621.

Bahlai CA, Sikkema S, Hallett RH, et al. (2010) Modeling distribution and abundance of soybean aphid in soybean fields using measurements from the surrounding landscape. Environmental Entomology 39:50–6.

Bai X, Zhang W, Orantes L, et al. (2010) Combining next-generation sequencing strategies for rapid molecular resource development from an invasive aphid species, Aphis glycines. PloS One 5:e11370.

59 Bodreau D (1992) Buckthorn research and control at Pipestone National Monument (Minnesota). Restoration and Management Notes 10:94-95.

Brewer MJ, and Noma T (2010). Habitat affinity of resident natural enemies of the invasive Aphis glycines (Hemiptera: Aphididae), on soybean, with comments on biological control. Journal of Economic Entomology 103(3): 583-596.

Converse CK (1985) Element Stewardship Abstract for Rhamnus cathartica, Rhamnus frangula (syn. Frangula alnus). The Nature Conservancy.

Costamagna AC, and Landis DA (2011) Lack of strong refuges allows top-down control of soybean aphid by generalist natural enemies. Biological Control 57: 184-192.

Dieringer D, Schlštterer C (2003) Microsatellite analyser (MSA): a platform independent analysis tool for large microsatellite data sets. Molecular Ecology Notes. 3 (1): 167-169.

Dixon, AFG (1985) Aphid Ecology. Blackie & Son Ltd

Draxler RR, Rolph GD (2012) HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) Model access via NOAA ARL READY Website (http://ready.arl.noaa.gov/HYSPLIT.php). NOAA Air Resources Laboratory, Silver Spring, MD.

Gildow FE, Shah DA, Sackett WM, et al. (2008) Transmission efficiency of Cucumber mosaic virus by aphids associated with virus epidemics in snap bean. Phytopathology 98:1233-41.

Hodgson EW, McCornack BP, Tilmon K, Knodel JJ (2012) Management recommendations for soybean aphid (Hemiptera: Aphididae) in the United States. Journal of Integrated Pest Management 3:1–10.

Isard SA, Russo JM, and Ariatti A (2007) The integrated aerobiology modeling system applied to the spread of soybean rust into the Ohio River valley during September 2006. Aerobiologia 23: 271-282.

Kurylo JS, Knight KS, Stewart JR, Endress AG (2007) Rhamnus cathartica: Native and naturalized distribution and habitat preferences. Journal of the Torrey Botanitcal Society 134:420-430.

Michel AP, Zhang W, Jung JK, et al. (2009a) Cross-species amplification and polymorphism of microsatellite loci in the soybean aphid, Aphis glycines. Journal of Economic Entomology 102:1389–92.

Michel AP, Zhang W, Kyo Jung J, et al. (2009b) Population genetic structure of Aphis glycines. Environmental Entomology 38:1301–11.

Müller CB, Williams IS, Hardie J (2001) The role of nutrition, crowding and interspecific interactions in the development of winged aphids. Ecological Entomology 26:330–340.

60 Myers JA, Vellend M, Gardescu S, Marks PL (2004) Seed dispersal by white-tailed deer: implications for long-distance dispersal, invasion, and migration of plants in eastern North America. Oecologia 139:35-44.

Nault BA, Shah DA, Straight KE, et al. (2009) Modeling temporal trends in aphid vector dispersal and cucumber mosaic virus epidemics in snap bean. Environmental Entomology 38:1347-59.

Orantes LC, Zhang W, Mian MAR, Michel AP (2012) Maintaining genetic diversity and population panmixia through dispersal and not gene flow in a holocyclic heteroecious aphid species. Heredity 1-8.

Peakall R, Smouse PE (2006) GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research. Molecular Ecology Notes. 6: 288-295.

Ragsdale DW, Landis DA, Brodeur J, et al. (2011) Ecology and management of the soybean aphid in North America. Annual Review of Entomology 56:375–99.

Ragsdale DW, Voegtlin DJ, O’Neil RJ (2004) Soybean Aphid Biology in North America. Annals of the Entomological Society of America 97:204–208.

Rice ME and O’Neal ME (2008) Soybean aphid management field guide 2008.University Extension, Iowa State University.

Schmidt NP, O’Neal ME, Anderson PF et al (2012) Spatial distribution of Aphis glycines (Hemiptera: Aphididae): a summary of the suction trap network. Journal of Economic Entomology 105(1): 259-271.

Schulz T, Thelen KD, and Difonzo C (2011). Neonicontinoid seed treatments for soybeans. Soybean Management and Research Technology, Michigan State University Extension.

Song F, Swinton SM, Difonzo C, et al. (2006) Profitability analysis of soybean aphid control treatments in three North-Central states. Department of Agricultural Economics, Michigan State University.

Stunder BJB, Heffter JL, and Draxler RR (2007) Airborne volcanic ash forecast area reliability. Weather and Forecasting 22: 1132-1139.

Tilmon KJ, Hodgson EW, O’Neal ME, Ragsdale DW (2011) Biology of the soybean aphid, Aphis glycines (Hemiptera: Aphididae) in the United States. Journal of Integrated Pest Management 2:1–7.

Voegtlin DJ, O’Neil RJ, Graves WR (2004) Tests of suitability of overwintering hosts of Aphis glycines: identification of a new host association with Rhamnus alnifolia L’Héritier.” Annals of the Entomological Society of America 97(2): 233-234.

Voegtlin DJ, O’Neil RJ, Graves WR, et al. (2005) Potential winter hosts of soybean aphid. Annals of the Entomological Society of America 98:690-693.

61 Wang Y, Stein AF, Draxler RR et al (2011) Global sand and dust storms in 2008: Observation and HYSPLIT model verification. Atmospheric Environment 45(35): 6388-6381.

Yoo HJS, O’Neil RJ, Voegtlin DJ, Graves WR (2005) Host plant suitability of Rhamnaceae for soybean aphid (Homoptera: Aphididae). Annals of the Entomological Society of America 98:926-930.

Zhang W and Swinton SM (2009) Incorporating natural enemies in an economic threshold for dynamically optimal pest management. Ecological Modelling 220: 1315-1324.

62 Figures and Tables

Table 3-1. Aphid collection dates and locations for 2009 and 2010. Location Year Date(s) GPS coordinates PA – Rock Springs 2009 Jun. 30, Jul. 13, Aug. 10 + 31 40.71. -77.94 NY 2009 Aug. 6 42.89, -77.92 VA 2009 Aug. 12 37.50, -76.60 PA – Rock Springs 2010 Jul. 5, Aug 2 + 30 40.71. -77.94 PA – Airport 2010 Jul. 2, Aug. 2 + 31 40.85, -77.84 NY 2010 Jul. 29 43.17, -78.14 Ottawa, ON 2010 Sept. 1 45.39, -75.72 Hungry Bay, QC 2010 Aug. 26 45.19, -74.15 Sherrington, QC 2010 Aug. 26 45.10, -73.53

Table 3-2. Fst values for Pennsylvania, Virginia and New York 2009. None of the Fst values were significant at P < 0.05. VA NY PA Jun 30 PA Jul 13 PA Aug 10 PA Aug 31 VA -- 0 0.011 0.004 0.024 0.003 NY -- 0.016 0.012 0.045 0.015 PA Jun 30 -- 0 0.002 0.014 PA Jul 13 -- 0 0.002 PA Aug 10 -- 0.010 PA Aug 31 --

Table 3-3. Fst values for Pennsylvania (all collection dates combined), Virginia and New York 2009. Significant values (P < 0.05) indicated in bold. VA NY PA VA -- 0 0.008 NY -- 0.020 PA --

63 Table 3-4. Genotypic diversity for aphid populations collected in Pennsylvania, Virginia, and New York 2009. Population No. matching No. unique Total no. Genotypic genotypes genotypes genotypes Diversity* VA 4 24 28 0.93 NY 8 18 26 0.86 PA Jun 30 13 16 29 0.97 PA Jul 13 11 17 28 0.93 PA Aug 10 10 16 26 0.86 PA Aug 30 9 17 26 0.86 *Genotypic diversity = # genotypes – 1/N – 1

Table 3-5. Fst values for Pennsylvania, Virginia, New York and Midwest sites 2009. Significant values (P < 0.05) indicated in bold.

Abbreviations: MN-R, Minnesota – Rosemont; OH-G, Ohio-Wooster; MI, Michigan; SD, South Dakota; OH-T, Ohio-Cortland; ON, Ontario; WI, Wisconsin. 1 indicates early collection, 2 indicates late collection.

Table 3-6. Fst values for Pennsylvania, Canada, New York (2010) and Midwest sites 2009. Significant values (P < 0.05) indicated in bold.

Abbreviations: MN-R, Minnesota – Rosemont; OH-G, Ohio-Wooster; MI, Michigan; SD, South Dakota; OH-T, Ohio-Cortland; ON, Ontario; WI, Wisconsin; RS, Rock Springs; AP, Airport; NY, New York; SHER, Sherrington; HGY, Hungry Bay. 1 indicates early collection, 2 indicates late collection.

Table 3-7. Fst values for PA, NY, and sites in Canada from 2010 collections. Significant values (P < 0.05) indicated in bold.

Table 3-8. Genotypic diversity for aphid populations collected in Pennsylvania, Canada, and New York 2010. Population No. No. unique Total no. Total no. Genotypic Matching genotypes genotypes samples diversity* genotypes Airport Jul 2 10 16 26 32 0.81 Airport Aug 2 7 22 29 30 0.97 Airport Aug 31 6 19 25 31 0.81 Rock Springs Jul 5 5 21 26 30 0.87 Rock Springs Aug 2 1 28 29 30 0.97 Rock Springs Aug 30 0 32 32 32 1.00 Sherrington Quebec 4 22 26 32 0.81 Ottawa 2 28 30 31 0.97 Hungry Bay Quebec 2 25 27 32 0.84 New York 1 30 31 32 0.97 *Genotypic diversity = # genotypes – 1/N – 1

67 Table 3-9. Score from 0 to 1 of forward trajectories that cross the target location from the HYSPLIT maps. 142 date/location scenarios were evaluated and dates where none of the trajectories crossed the target location (score of 0) are not shown on this table. Maps were generated from July 1-31 2009 for PA to VA and PA to NY, and July 1 – August 9 2009 for NY to VA and NY to PA. Start location Target end location Map date Score 7/5/2009 0.07 7/7/2009 0.04 PA VA 7/8/2009 0.37 7/13/2009 0.67 7/14/2009 0.11 7/8/2009 0.11 NY VA 8/9/2009 0.07 7/8/2009 0.04 NY PA 8/6/2009 0.04 7/1/2009 0.04 7/10/2009 0.52 7/11/2009 0.52 PA NY 7/22/2009 0.15 7/23/2009 0.30 7/25/2009 0.52

0.060 PA Aug. 31

0.040

0.020

PA Aug. 10 PA Jul. 13 VA 0.000

-0.020 PC 2: PC 22.93% of variation

-0.040 PA Jun. 30

-0.060 -0.080 -0.060 -0.040 -0.020 0.000 0.020 0.040 0.060 0.080 PC 1: 60.61% of variation

0.030 New York 0.020 Airport Aug. 31 HB Quebec Airport Ottawa Aug. 2 Rock Springs

0.010 Aug. 30 Sh Quebec 0.000

Rock Springs -0.010 Aug. 2

-0.020

PC 2: 2: 18.08% PC ofvariation Rock Springs -0.030 Jul. 5

-0.040 Airport Jul. 2 -0.050 -0.060 -0.040 -0.020 0.000 0.020 0.040 0.060 PC 1: 47.12% of variation

0.060 MI-2 0.040 OH-G1 MN-R2 MI-1 ON-1 OH-T2 0.020 WI-2

ON-2 0.000 NY MN-R1 SD-2 PA Jun. 30 OH-G2 PA Jul. 13 MN-L2 -0.020 VA OH-T1 PA Aug. 10 WI-1 MN-L1

-0.040 PA Aug. 31 PC 2: 20.6 2: 20.6 PC %variation -0.060

-0.080 SD-1

-0.100 -0.100 -0.050 0.000 0.050 0.100 0.150 PC 1: 58.2 % variation

0.080 OH-T1 0.060 OH-G1 SD-1 0.040 MI-1 SHERQC NY WI-1 0.020 AP-EAR SD-2 MI-2 Ontario RS-LT 0.000 AP-MID AP-LT RS-EAR OH-T2 ON-1 -0.020 HGYQC OH-G2 RS-MID MN-R2 WI-2 ON-2

-0.040 MN-L2 PC 2: 2: 12.75% PC variation

-0.060 MN-L1

-0.080 MN-R1 -0.100 -0.150 -0.100 -0.050 0.000 0.050 0.100 0.150 PC 1: 69.71% variation

Chapter 4

Speciation and population structure of Therioaphis trifolii

Introduction

Therioaphis trifolii (the spotted alfalfa aphid) is an aphid species that was introduced to the United States and became important because of its abundance and preference for a legume crop (alfalfa). Like A. glycines, it is a competent vector of CMV in snap bean (Gildow et al. 2008,

Nault et al. 2009). The spotted alfalfa aphid (referred to as SAA and formerly of the species

Therioaphis maculata) was first observed in the United States in 1954, in New Mexico on alfalfa.

It spread rapidly through the Southwest and on to the Eastern states (Dickson 1958). Its introduction and subsequent unexpected damage to crops mirrors that of A. glycines. Since it has been in the U.S. for over 50 years, its populations have equilibrated and it is no longer the cause of extreme economic loss in its host crops of alfalfa and clover. T. trifolii can serve as a model invasive aphid species and give insight into the future of A. glycines.

Recognition of T. trifolii as a single species is in flux. In 1958, T. trifolii was indicated as the yellow clover aphid (YCA), and SAA was considered a different species (T. maculata).

Examination of morphological characters in Dickson (1958) resulted in the conclusion that SAA populations probably came from a single or small colony introduction. The population in 1958 had also developed pesticide resistance in approximately four generations. Host plant differences between YCA (in clones from red clover) and SAA (in clones from alfalfa) were also observed

(Dickson 1958, Manglitz and Russell 1974). The initial observations of SAA populations described it as anholocyclic, but holocyclic (producing sexual morphs and overwintering as eggs instead of asexual adults) populations were observed in the northern part of its new range

75 (Manglitz 1966). SAA was believed to be introduced from the Mediterranean region where sexual morphs did not occur, but sexual morphs do occur in colder regions of the Old World range.

Current classifications consider both SAA and YCA to be in the same species, T. trifolii

(Blackman and Eastop 2000, 2006, Carver 1978).

We set out to compare T. trifolii samples from Pennsylvania and Serbia to see if there was any evident population structure that would reflect the species’ history of introduction.

Methods

We collected at least 30 aphids from two sites in Pennsylvania (the Russell E. Larsen

Reasearch Farm in Centre County, and the Kretschman farm near Pittsburgh) and obtained aphids collected by Olivera Petrovic-Obradovic in Serbia. Aphids were collected from alfalfa fields and individuals were preserved singly in 70% EtOH. DNA was extracted using the E.Z.N.A. Tissue

DNA kit (Omega Bio-tek, Norcross, GA), and run in a PCR with one of three general insect mitochondrial primers (LR, NS, and SR).

The PCR product was sequenced, and the resulting sequences were aligned in BioEdit

(Hall 1999), and TCS (Clement et al 2000) was used to create a haplotype web.

Results

Two of the primers (LR, SR) produced useable sequences. The haplotype web from TCS for each of the primers resulted in seqences obtained from aphids in Pennsylvania being the ancestral haplotype for this collection (Figures 4-1 and 4-2).

76 Discussion

These results represent a summary of the literature and a preliminary attempt to use genetic tools to elucidate the relationships between T. trifolii in the Old and New World. Our samples were from a limited geographic range, and did not indicate clear population-level differences.

Since T. trifolii was introduced on separate occasions and geographically distinct locations, and then flourished for decades it serves as an interesting potential model for the use of genetic tools to identify population change in an introduced aphid species. There is speculation that A. glycines is moving to new primary hosts which could result in sympatric speciation over time, especially since sexual reproduction is only occurring on the primary host. Future work should include more intensive sampling on clover and alfalfa as well as a broader range of geographic regions, and the development and use of species-specific SNPs or microsatellites.

References

Blackman R, Eastop V (2000) Aphids on the World’s Crops: An Identification and Information Guide, 2nd ed. Wiley, Chichester.

Blackman R, Eastop V (2006) Aphids on the World’s Herbaceous Plants and Shrubs: An Identification Guide. Wiley, Chichester.

Carver M (1978) The scientific nomenclature of the spotted alfalfa aphid (Homoptera: Aphididae). Journal of the Australian Entomological Society 17: 287-288.

Clement M, Posada D, Crandall K. (2000) TCS: a computer program to estimate gene genealogies. Molecular Ecology 9(10): 1657-1660

Dickson RC (1959) On the identity of the spotted alfalfa aphid in North America. Annals of the Entomological Society of America 52:63-68.

Gildow FE, Shah DA, Sackett WM, et al. (2008) Transmission efficiency of Cucumber mosaic virus by aphids associated with virus epidemics in snap bean. Phytopathology 98:1233-41.

77 Hall, TA (1999) BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symposium Series 41: 95–98.

Librado P, Rozas, J (2009) DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25: 1451-1452

Manglitz G, Calkins C, Walstrom R, et al. (1966) Holocyclic strain of the spotted alfalfa aphid in Nebraska and adjacent states. Journal of Economic Entomology 59:636-639.

Manglitz GR, Russell LM (1974) Cross matings between Therioaphis maculata (Buckton) and T. trifolii (Monell) (Hemiptera: Homoptera: Aphididae) and their implications in regard to the taxonomic status of the insects. Proceedings of the Entomological Society of Washington 76(3): 290-296.

Nault BA, Shah DA, Straight KE, et al. (2009) Modeling temporal trends in aphid vector dispersal and cucumber mosaic virus epidemics in snap bean. Environmental Entomology 38:1347-59.

78 Figures and Tables

Chapter 5

Phenology model validation of pests of cucurbits

Introduction

Members of the plant family Cucurbitaceae are widely distributed and well-represented in global agriculture (Bates 1990). In 2007, over ten-thousand acres of cucurbit crops were harvested in Pennsylvania (USDA NASS, 2009). Of these crops, pumpkins take up the most acreage and were valued at over $14.5 million in 2011 (USDA NASS, 2011).

Growing cucurbits is challenging due to the presence of three specialist insect pests; striped cucumber beetle (Acalymma vittatum Fabricius), squash bug (Anasa tristis DeGeer) and squash vine borer (Melittia cucurbitae Harris). Not only do these pests cause direct feeding damage to the plants, but striped cucumber beetle and squash bug each vector bacterial pathogens.

The striped cucumber beetle is important during the growing season because of its direct feeding damage to vines and fruit as well as its ability to vector E. tracheiphila, the causal agent of bacterial wilt. E. tracheiphila is a bacterium that proliferates in the xylem which blocks water flow in the plant causing the characteristic wilting on the leaves on the infected stem. Eventually, the bacteria will spread through the entire plant and kill it. Systemic and foliar pesticide sprays are used in conventional systems against the adults. Larval mortality can be increased by growing cucurbits on black plastic with drip irrigation and introducing entomopathogenic nematodes through the drip line (Ellers-Kirk et al 2000).

The squash bug is an important pest of Cucurbita but not Cucumis species (Bonjour and

Fargo 1989). Squash bugs inflict feeding damage on vines and fruit, and can vector Serratia

81 marcescens, the causal agent of cucurbit yellow vine disease, another bacterial cucurbit disease

(Bruton et al 2003). S. marcescens causes yellowing of the leaves, phloem discoloration and a general decline in plant vigor. During the growing season, the lifecycle occurs above ground, making it easier to observe egg masses, juvenile instars, and subsequent adult generations.

Common cucurbit cultivation practices include the use of mulch (usually plastic), which provides harborage for squash bug adults and nymphs, thus increasing pest pressure (Cartwright et al.

1990).

Squash vine borer adults do not damage cucurbit crops, but the larvae burrow into the cucurbit vine where they feed. Larvae can survive well on Cucurbita pepo and C. maxima cultivars (Howe and Rhodes 1973), and short of growing unsuitable host plants the most effective way to control this insect is to apply foliar insecticides when first instars are present before they enter the vine. These insects overwinter as pupa, and emerging adults can be monitored with pheromone traps. The pheromones in the lures overlap with those of other Sessiid species (Van

Wychen Bennett et al, 2011) sometimes resulting in non-target captures.

Striped cucumber beetle and squash bug overwinter as adults in leaf litter and/or crop residue and resume activity in the spring. When they become active again is not well defined.

Radin and Drummond (1994) suggest that the striped cucumber beetles can be active on any day with an average temperature above 12 C (53. 6 F) in Maine, and Lewis et al. 1990 found beetle activity on flats of C. maxima cv. Blue Hubbard when temperatures were above 18 C (64.4 F).

Previous trapping efforts in Pennsylvania at Rock Springs caught beetles in emergence cages in mid-May (Fleischer, unpublished). The cue that causes these beetles to enter diapause in the fall is unknown. Following striped cucumber beetle adult immigration, we projected the timing of emergence of the first field generation (F1) of adults as 793.6 degree-days base 55 F (Ellers-Kirk and Fleischer 2006). The timing of emergence for subsequent field generations adds the 204.88

82 DD base 55 F preoviposition period to the egg to adult development time (Ellers-Kirk and

Fleischer 2006).

Although it is currently unknown how squash bugs terminate diapause, it is known that the bugs enter diapause when the critical photoperiod decreases between 14:10 (L:D) and

14.5:9.5 (Decker and Yeargan 2008, Nechols 1988). Egg to adult development was determined to be 725.1 degree-days base 60 F (Fargo and Bonjour 1988).

Squash vine borer overwinters as a late instar larva or pupae and has an extended emergence through the growing season. It requires 1687.5 degree days base 50 F to complete development (Canhilal et al. 2006).

In the case of these cucurbit pests, predicting early season activity can be useful for optimizing planting time to give the plants an opportunityto grow past their most susceptible growth stages without pest pressure. For striped cucumber beetle and squash bug, we are defining early season activity as the recruitment of the pest to a flat comprised of cucurbit seedlings. These two pests are active in the spring before crops are planted, and monitoring their presence can estimate the intensity of early-season pest pressure. Once cucurbit crops are planted, in-field monitoring identifies the time of colonization (the biofix) for each pest. Using degree day developmental requirements from the literature for each pest, we can use the biofix, forecast air temperatures, and 30-year climatology records to estimate the emergence of subsequent field generations. Since cucurbits require the services of pollinators, predicting pest levels in the growing season is important for timing pest control measures while still allowing for adequate pollination.

For this research, we used air temperature degree days from January 1 as a starting point, because they are well monitored and accessible through meteorological databases. We are using weather station data since it is widely available and part of several established networks that will exist beyond the length of this study. This longevity will allow a successful model and website to

83 operate in subsequent years without the deployment of additional monitoring devices such as independent temperature data loggers.

We set out to monitor the early season activity and growing season phenology of these pests on land that is transitioning to organic production on research farms in Pennsylvania, Iowa and Kentucky on two commercially important cucurbit crops (muskmelon – Cucumis melo, and butternut squash – Cucurbita moschata) to better inform organic management practices.

For each of these insects, knowledge of pest phenology will allow growers to more effectively deploy organic and cultural controls such as approved pesticides, row covers, and adjust planting and transplanting dates to times when they will be most effective.

Methods

Early season activity

To monitor the activity of the insects before the field planting of cucurbit crops, we planted smaller containers with cucurbit seeds. The resulting flats containing seedlings and young plants were placed on research farms with a history of growing cucurbit crops on some portion of their land. The goal was to be able to attract and count striped cucumber beetles and squash bugs that were active early in the season prior to the establishment of commercial cucurbit fields.

Each trap flat consisted of one standard 1020 greenhouse flat (28 by 53 cm) planted with

20 – 30 Cucurbita maxima seeds (2010 seeds from Rupp var. ‘Blue Hubbard,’ and Johnny’s

Selected Seeds ‘var. Blue Ballet,’ 2011 seeds from Johnny’s Selected Seeds var. ‘Blue

Hubbard’). These cultivars were selected for the trap flats because of their demonstrated attractiveness as a trap crop for striped cucumber beetles (Adler and Hazzard, 2009). Flats were

84 grown in a greenhouse under ambient light and watered regularly. Traps were deployed 2-3 weeks after seeding, when plants were in the first through third leaf stage.

Ten trap flats were placed in pairs around the Russell E. Larson Research and Education

Center, Pennsylvania Furnace, PA (colloquially referred to as Rock Springs), within approximately 0.5 km of fields that were planted in cucurbits within the last year. In Kentucky and Iowa, trap flats were placed at the University of Kentucky Organic Farming Research and

Extension Unit at the UK Horticulture Research Farm in Lexington KY and the Iowa State

University Horticulture Farm in Gilbert IA, respectively (Table 5-1). In addition to the plants, each flat also had a 7.6 by 12.7 cm yellow sticky card (RSTRIP Gempler’s, Madison, WI) mounted on a 30 cm wooden stake. Flats were checked three times per week for striped cucumber beetle and squash bug activity until a research plot of cucurbits was planted, which was also the typical timing of established commercial cucurbit fields in that area (see Table 5-1 for monitoring duration). Trap flats were replaced weekly or earlier if they were damaged by frost or lack of water. Any striped cucumber beetles or squash bugs found were counted and removed from the flat.

We took the cumulative number of captured beetles to represent the timing of spring emergence activity, and regressed cumulative capture, scaled from 0 to 1, against degree-days base 55 Fahrenheit, using a Gompertz equation (Winsor 1932) calculated with JMP v.8 (JMP

2009). The Gompertz equation is a good model for early season recruitment, because once a beetle is found on a trap flat, additional beetle accumulation happens rapidly (Smyth and Hoffman 2003). For our analysis with JMP the constants k, a, and b are represented as θ1, θ2, and θ3 respectively. In the equation, θ1 is the maximum value that y approaches. θ2 determines the placement of the curve’s first inflection point on the x-axis and θ3 influences the steepness of the slope as the curve approaches the value of θ1. We used both

85 calendar day and cumulative degree day as the independent variable (x) and y is the proportion of the total pest population observed.

In-season phenology

In order to monitor insect pest populations during the growing season, we planted 30.5 row meters each of Cucumis melo var. ‘Strike’ and Cucurbita moschata var. ‘Betternut’ on black plastic with drip irrigation. Planting dates followed local production practices and ranged from mid-May to early June (Table 5-1). Five plants per row were monitored weekly (biweekly if insect pressure was high) for striped cucumber beetle and squash bug using visual inspection of whole plants. Striped cucumber beetle and squash bug adults were counted in addition to the number of squash bug juvenile instars and egg masses. A biofix for each site was defined as two consecutive observations with non-zero counts. Plants were replaced if they died and the plot was not sprayed to control insects.

Two wire cone Harstack traps were set up at each research farm in fields planted with muskmelon and butternut, and baited with squash vine borer pheromone lures (Great Lakes IPM).

Lures were changed every other week. The traps were checked weekly for squash vine borer and caught moths were removed and frozen.

Adult immigration into the phenology plots was estimated in two ways. First, we used the model of spring activity, described above, which essentially is defined by heat unit accumulations in the spring. Secondly, we used the sampling data within the plots to determine when striped cucumber beetles or squash bugs were first present, defined as a mean of >0 counts per plant for two consecutive sampling events. Following adult immigration, we projected the timing of emergence of the first field generation (F1) of striped cucumber beetle adults as 793.6 degree- days base 55 F (Ellers-Kirk and Fleischer 2006). The timing of emergence for subsequent field

86 generations adds the 204.88 DD base 55 F preoviposition period to the egg to adult development time (Eller-Kirk and Fleischer 2006).

For squash bugs, we projected the timing of emergence of the first field generation (F1) of adults as 725.1 degree-days base 60 F (Fargo and Bonjour 1988). After a 140 – 200 DD preoviposition period (Nechols 1987), the F1 adult females were predicted to lay eggs for the next field generation. There is also a photoperiod component to squash bug phenology, with eggs that hatch and nymphs that reach adulthood after the critical photoperiod (14:10 (L:D) and 14.5:9.5

(Decker and Yeargan 2008, Nechols 1988)) preparing for diapause instead of reproduction.

Squash Vine Borer

Using the life table for squash vine borer recorded in Canhilal et al. 2006, we converted the growth chamber data to degree-days base 50 F in order to estimate a predicted adult spring emergence of 754.2 DD base 50 F (Table 5-2), and compared it to observed captures in pheromone traps.

Meteorological Data

Degree days for each location were obtained by the Center for Environmental

Informatics, Penn State University, from meteorological station and model data. The station data came from FAA ASOS (Federal Aviation Administration Automated Surface Observing System) and COOP (Cooperative Observer Program) sites depending on which was closer to the farm location. Both types of data came from NOAA (National Oceanic and Atmospheric

Administration) datastreams. The FAA data consisted of hourly or sub-hourly reports of air temperature, dew point, precipitation, wind speed, wind direction, cloud cover, visibility, and

87 present weather from mainly airport weather stations with a day being from midnight to midnight.

The National Weather Service (NWS), FAA and Department of Defense operate the 967 stations across the United States. The data was averaged into hourly values, which were then averaged from midnight to midnight into daily values.

The COOP data comes from the NOAA datastream as either daily averages or daily maximum and minimum temperatures, which were averaged to get daily temperatures (each day begins at 7 AM). The COOP, started in 1890 under the Organic Act, which created the Weather

Bureau, is now operated by the National Weather Service (NWS). It was formed to provide observational meteorological data that usually includes daily temperature maximums and minimums, snowfall, and 24 hour accumulated precipitation. The goal of the COOP program is to define the climate of the U.S. and to help measure long-term climate change. The observations, usually recorded by volunteers, send monthly temperature and precipitation data to the National

Climatic Data Center (NCDC), where it is digitized, checked, and archived. There are over

11,000 volunteers taking observations across the United States.

The model data were from the NARR (North American Regional Reanalysis) through a

NOAA ftp server. The NCEP (National Centers for Environmental Prediction) NARR is a long- term, dynamically consistent, high-resolution, high-frequency, atmospheric and land surface hydrology dataset for the North American domain. The NARR model uses the very high resolution NCEP Eta Model (32km/45 layer) together with the Regional Data Assimilation

System (RDAS) which, significantly, assimilates precipitation values along with other variables.

Examples of other variables included in the NARR dataset are: air temperature, dew point, wind speed, specific humidity, soil temperature, soil moisture, surface downward longwave and shortwave radiation, and surface evaporation. The output analyses are 3-hourly. We used a program that finds the gridpoint closest to the desired latitude/longitude coordinates of the farm,

88 and uses that value for the desired location. These data are then averaged into daily temperatures.

The NARR is not run in real-time (about a month delayed).

To calculate degree days, we converted the daily averages from degrees Celsius to degrees Fahrenheit and subtracted the appropriate degree threshold (50, 55, or 60) if the daily average was greater than the threshold depending on the insect species.

Results

Striped Cucumber Beetle

Striped cucumber beetles were first detected on average across all three states on calendar day 127.0 (+/- 18.4 days) (Table 5-3). KY reported the beetles the earliest, on calendar day 112.3

(+/- 15.5 days), and the first capture date in IA was the most variable (133.5 +/- 24.8 days).

The average cumulative degree day base 55 F of first detection was 224.7 (+/- 62.9) over all three states. Average cumulative degree day was most consistent for KY, with an error of +/-

33.6 degree days (Table 5-3). The variation of 59.2 degree days in PA is about 5-10 calendar days based on 30 year climate averages.

Time of first detection on the in-season phenology plot was dependent on the date of planting, with the beetles arriving on average 9.5 +/- 3.5 and 40.0 +/- 22.6 days after planting in

PA and IA respectively (Table 5-4).

The values for the regression coefficients from the Gompertz equations were more consistent when degree day was used as the independent variable [standard errors ranging from

0.0013 to 0.4486 using cumulative degree day (Table 5-6) versus 0.0481 to 6.8221using calendar day (Table 5-7)]. For calendar day, a curve calculated for each state individually gave a more

89 accurate representation of early season recruitment for that state than a curve based on all locations combined (Figures 5-2 and 5-3).

Looking at the in-season phenology for each state, it appears that the season-long degree day accumulations are only enough to support the development and emergence of one field generation (Figures 5-4, 5-5, and 5-6).

The observed numbers of striped cucumber beetles per plant in the in-season phenology plots did not show clear peaks defining a separation between the overwintered adults and the first field generation. To further elucidate the seasonal dynamics, we divided the data based on the biofix and egg-adult development time (Figures 5-10, 5-11, 5-12). Of the 5 state/years observed, only one did not appear to have a second generation (Iowa 2011, Figure 5-11)

Squash Bug

Squash bugs were caught during the trap flat sampling, but not in high enough numbers to model their activity. Squash bugs were first detected after 138.5 calendar days (+/- 12.3) and

176.2 (+/- 81.6) cumulative degree days base 60 F (Table 5-3). Calendar day at first detection was the most variable in IA (+/- 24.8 days) and the least variable in PA (+/- 4.0 days). Degree days at first detection were the greatest in KY (241.3) and the least in PA (125.7).

Much like the striped cucumber beetles, in-season phenology plot initial detection was dependent on planting date with the bugs arriving on average 18.0 +/- 4.2 and 35.0 +/- 4.2 days after planting in PA and IA respectively (Table 5-4).

The season-long degree day accumulations for each state are enough to allow for the development of one field generation of squash bugs in Pennsylvania and Iowa (Figures 5-7 and 5-

9), and come close to allowing two field generations in Kentucky (Figure 5-8). In Kentucky, eggs hatching after the critical photoperiod would go into diapause as adults instead of reproducing.

90 Squash Vine Borer

Squash vine borer was first captured later in the spring for all locations at calendar day

169.5 and cumulative degree day base 50 F 1080.2 (Table 5.5).

Discussion

There have been several previous attempts at predicting the timing of and identifying the cues that initiate striped cucumber beetle activity in the spring before commercially cultivated cucurbits are present (Elsey 1988, Lewis et al. 1990, Pair 1997, Radin and Drummond 1994).

This study focused on advancing our ability to estimate early season activity using air temperature degree days. Across a three state region, when cumulative degree days base 55 F was used as the predictor, the first inflection point of a curve modeling recruitment to seedlings used as traps was approximately 150 degree days, and ranged from 140 to160. Looking at the 30-year climate and assuming that beetles begin their activity between 100 and 200 cumulative degree days, we can estimate that as a range of approximately 11 days for PA and 10 days for IA and KY occurring in mid to late May.

The pattern of rapid recruitment of striped cucumber beetle on the trap flats which occurred prior to commercial transplanting on farms was consistent with the findings from Smyth and Hoffman (2003) who measured recruitment during the summer field season. They used male and female beetles and plants with high or low levels of cucurbitacin in different combinations to determine ecological mechanisms of attraction. They found that the beetles oriented upwind, especially on traps containing pioneer males (Smyth and Hoffman 2003). In our study, we had low initial counts on the trap flats, but once beetles were found on a flat, subsequent samplings had much higher counts (some flats had hundreds, especially in Kentucky). The pattern of the

91 Gompertz equation represents this aggregative activity well, by increasing more rapidly than a simple logistic curve (Winsor 1932).

Predicting early season activity of an insect that overwinters as an adult is challenging without knowing the cues that are inducing the end of diapause and factors that affect spring activity. We are using a single variable to attempt to model the start of spring activity for these pests, and using recruitment to a trap flat as a surrogate for measuring early season activity. There are other variables (soil temperature, day length, appearance of other food sources) that could give more precise results once the mechanism for exiting diapause and initiating spring activity is known.

This study also reinforced the trap flat methods used in other studies (Lewis et al. 1990 and Radin and Drummond 1994). Our use of ‘Blue Hubbard’ trap flats well distributed on farms with a history of cucurbit plantings was successful in attracting striped cucumber beetle and squash bug in KY and PA. The lack of trap flat success in Iowa probably stems from the fact that there is less cucurbit pest pressure on that farm and they experienced cold winters before the start of the study.

During the season, the beetle was quite active on the phenology plots. Using the degree day data we have and the information on development time, we can estimate that there is one field generation of striped cucumber beetle in PA. We are assuming that the initial biofix beetles have already completed their preoviposition period, and that overwintered females are capable of ovipositing by the time we observe beetles on commercial plantings. Supporting this assumption, striped cucumber beetles that were dissected over the winter months were found to have undeveloped ovaries, but the percent of females with developing ovaries was high in April in

South Carolina (Elsey 1988).

We could not observe a clean delineation between the field generations or the overwintered adults. This generational overlap makes it a challenge to predict a gap between

92 generations as a safe time to specify management options that require the absence of adults, such as removing row covers for pollination. Adult females are also long-lived, with the potential to survive over 100 days post-emergence and continue laying eggs during that time (Ellers-Kirk and

Fleischer 2006). For management purposes, it may be more accurate to think of the striped cucumber beetle population as a single cohort, comprised of overwintering and field generated adults.

17 to August 4 for Lexington KY, and July 29/30 to August 11/12 for Gilbert IA (visualized by vertical lines in Figures 5-7, 5-8, and 5-9). Squash bugs that hatch after those dates in each location will not lay eggs and will diapause if they make it to adulthood (Nechols 1988). This pest also had a period of extended emergence and activity with no clear peaks between generations, but there was a large spike in nymph numbers in July that would be particularly injurious to crops.

Both the bug and the beetle vector plant diseases, thus making any effort to control them dually important for the control of the diseases. Some of the previous papers on both pests suggest clean cultivation as a way to reduce numbers, and this may not mesh with all organic farming practices. A move away from black plastic mulch would be good thing for growers concerned about squash bug because of how the bugs use the plastic as harborages, as it can increase their survival (Cartwright et al. 1990). Planting a trap crop of a more attractive variety may lure the insects away from the main field, but the trap crop must then be destroyed to remove the pest population.

93 Squash vine borer had the lowest densities of this pest assemblage, but our results give more information into its life history in geographic areas not well-represented in the current literature. This pest also has an extended emergence and adults were captured in the pheromone traps from June through September in South Carolina (Jackson et al 2005). In our study we stopped monitoring the traps after harvest, and had low per night catches in IA and PA.

Especially at our site in PA, the population of squash vine borer might be so low that it is below the detection threshold of our trapping method.This work shows some consistency in time of first capture using degree days (1080.2 +/- 151.6) and calendar days (169.5 +/- 5.5), however our first capture was over 300 degree days later than that reported in Canhilal et al (2006). Our work also took the developmental work in Canhilal et al. 2006 and converted it from days to degree days

(base 50 F), which will help to standardize future efforts.

Assaying the population of striped cucumber beetle to identify beetles that are in diapause is an important area for future work. The Gompertz curve was an effective model for representing striped cucumber beetle recruitment to trap flats and population increases after a biofix.

Phenology models are complicated by the fact that they require the estimation or discovery of what environmental cues the insects are using in their development and the impact of each cue. Other factors of biological importance that could influence phenology include, but are not limited to, photoperiod, soil temperature, and soil moisture, all of which could be investigated in future work.

94 References

Adler LS, Hazzard RV (2009) Comparison of perimeter trap crop varieties: effects on herbivory, pollination, and yield in butternut squash. Environmental Entomology 38:207-15.

Bates DM (1990). Biology and Utilization of the Cucurbitaceae. Cornell University Press. Ithaca, NY.

Bonjour EL, Fargo WS (1989). Host effects on the survival and reproduction of Anasa tristis (Heteroptera: Coreidae). Environmental Entomology 18: 1083 – 85.

Bruton BD, Mitchell F, Fletcher J, Pair SD et al (2003) Serratia marcescens, a phloem- colonizing, squash bug-transmitted bacterium: causal agent of cucurbit yellow vine disease. Plant Disease 87: 937-944.

Decker KB, Yeargan KV (2008) Seasonal phenology and natural enemies of the squash bug (Hemiptera: Coreidae) in Kentucky. Environmental Entomology 37:670-8.

Ellers-Kirk CD, Fleischer SJ (2006) Development and life table of Acalymma vittatum (Coleoptera: Chrysomelidae), a vector of Erwinia tracheiphila in Cucurbits. Environmental Entomology 35:875-880.

Ellers-Kirk CD, Fleischer SJ, Snyder RH, Lynch JP (2000) Potential of entomopathogenic nematodes for biological control of Acalymma vittatum (Coleoptera: Chrysomelidae) in cucumbers grown in conventional and organic soil management systems. Journal of Economic Entomology 93:605-12.

Elsey KD (1988a) Cucumber beetle seasonality in coastal South Carolina. Environmental Entomology 17:496-502.

Elsey KD (1988b) Reproductive diapause in the spotted cucumber beetle. The Florida Entomologist 71:78-83.

Fargo WS, Bonjour EL (1988) Developmental rate of the squash bug- Anasa tristis (Heteroptera: Coreidae)- at constant cemperatures. Environmental Entomology 17:926-929.

Fleischer SJ, Orzolek MD, Mackiewicz DDE, Otjen L (1998) Imidacloprid effects on Acalymma vittata (Coleoptera: Chrysomelidae) and bacterial wilt in cantaloupe. Journal of Economic Entomology 91:940-949.

95 Jackson DM, Canhilal R, Carner GR (2005) Trap monitoring squash vine borers in Cucurbits. Journal of Agricultural and Urban Entomology. 22(1): 27 – 39.

JMP, Version 8. SAS Institute Inc., Cary, NC, 1989-2009.

Lewis PA, Lampman RL, Metcalf RL (1990) Kairomonal attractants for Acalymma vittatum (Coleoptera: Chrysomelidae). Environmental Entomology 19:8-14.

Nechols, J. R. 1987. Voltinism, seasonal reproduction, and diapause in the squash bug (Heteroptera: Coreidae) in Kansas. Environmental Entomology 16 (1): 269-273.

Nechols JR (1988) Photoperiodic responses of the squash bug (Heteroptera: Coreidae): diapause induction and maintenance. Environmental Entomology 17:427-431.

Pair SD (1997) Evaluation of systemically treated squash trap plants and attracticidal baits for early-season control of striped and spotted cucumber beetles (Coleoptera: Chrysomelidae) and squash bug (Hemiptera: Coreidae) in cucurbit crops. Journal of Economic Entomology 90:1307-1314.

Radin AM, Drummond FA (1994) Patterns of initial colonization of cucurbits, ceproductive activity, and dispersion of striped cucumber beetle, Acalymma vittata (F.) (Coleoptera: Chrysomelidae). Journal of Agricultural Entomology 11:115-123.

Smyth RR, Hoffmann MP (2003) A male-produced aggregation pheromone facilitating Acalymma vittatum [F.] (Coleoptera: Chrysomelidae) early-season host plant colonization. Journal of Insect Behavior 16:347-359.

USDA National Agricultural Statistics Service. 2009. 2007 Census of Agriculture: Pennsylvania State and County Data.

Van Wychen Bennett K, Burkness EC, Hutchison WD (2011). Squash vine borer. VegEdge. http://www.vegedge.umn.edu/vegpest/CUCS/vinebor.htm Accessed: Aug 3, 2012.

Winsor CP (1932) The Gompertz curve as a growth curve. Proceedings of the National Academy of Sciences 18:1-8.

96 Figures and Tables

Table 5-1: Early season recruitment monitoring dates and locations for 2010 - 2012 in KY, IA, and PA. State Location GPS Year Start Date End Date In-season coordinates phenology planting PA Pennsylvania 40.714 2010 May 3 June 18 June 1 Furnace -77.948 2011 May 2 June 14 June 8 2012 April 13 June 11 June 4 KY Lexington 37.974 2010 April 26 June 4 -- -84.535 2011 April 3 May 25 -- IA Gilbert 42.108 2010 May 3 June 11 May 25 -93.589 2011 April 20 June 10 May 17

Table 5-2: Summary of life history parameters for Acalymma vittatum, Anasa tristis and Melittia cucurbitae from the literature (see footnotes) and this study (*). Acalymma vittatum Anasa tristis Melittia cucurbitae Spring life stage Adult Adult Late stage pupae of interest Degree-day base 55 F 60 F 50 F Degree days to 132.9 +/- 17.3* 141.7 +/- 9.0* 745.2 DD 1 first detection Egg – Adult 793.6 DD 2 725.1 DD 3 1687.5 DD 1 Development Time (Fahrenheit DD) Preoviposition 204.8 DD 2 140 – 200 DD 4 -- Period 1Canhilal et al 2006 2 Ellers-Kirk and Fleischer 2006 3Fargo and Bonjour 1988 4Nechols 1987

97 Table 5-3. Mean calendar day and cumulative degree day of first capture for Acalymma vittatum (SCB) and Anasa tristis (SB) on trap flats. N is the number of years.

Pest State n Calendar Day Cumulative Degree Day* PA 3 137.3 +/- 10.7 187.7 +/- 59.2 KY 3 112.3 +/- 15.5 216.3 +/- 33.6 SCB IA 2 133.5 +/- 24.8 292.8 +/- 69.4 All 8 127.0 +/- 18.4 224.7 +/- 62.9 PA 3 146.3 +/- 4.0 125.7 +/- 59.1 KY 3 134.0 +/- 7.9 241.3 +/- 87.6 SB IA 2 133.5 +/- 24.8 154.4 +/- 54.5 All 8 138.5 +/- 12.3 176.2 +/- 81.6 * Cumulative Degree Day base 55 F for SCB, and base 60 F for SB

Table 5-4. Mean calendar day and cumulative degree day of first capture for Acalymma vittatum (SCB) and Anasa tristis (SB) in the phenology plot. N is the number of years. Cumulative Degree Days after Degree Days after Pest State n Calendar Day Day* transplanting transplanting PA 2 167.0 +/- 5.7 523.45 +/- 41.8 9.5 +/- 3.5 555.6 +/- 218.6 KY 2 150.5 +/- 6.4 542.6 +/- 156.1 -- -- SCB IA 2 181.0 +/- 17.0 973.3 +/- 262.1 40.0 +/- 22.6 264.0 +/- 109.0 All 6 166.2 +/- 16.1 679.8 +/- 265.9 24.8 +/- 22.0 409.8 +/- 219.6 PA 1 174 353.7 18.0 +/- 4.2 317.3 +/- 135.4 KY 2 157.5 +/- 16.3 428.1 +/- 275.6 -- -- SB IA 2 176.0 +/- 1.4 564.3 +/- 91.6 35.0 +/- 4.2 133.1 +/- 42.1 All 5 168.2 +/- 12.8 467.7 +/- 172.6 26.5 +/- 10.4 225.2 +/- 134.2 * Cumulative Degree Day base 55 F for SCB, and base 60 F for SB

Table 5-5. Mean calendar day and cumulative degree day base 50 F of first capture for Melittia cucurbitae

State n Calendar Day Cumulative Degree Day PA 2 175.0 +/- 4.2 975.3 +/- 41.8 KY 2 165.0 +/- 5.7 1213.9 +/- 191.1 IA 2 168.5 +/- 0.7 1051.4 +/- 131.0 All 6 169.5 +/- 5.5 1080.2 +/- 151.6

98 Table 5-6. Parameter estimates for the Gompertz equations modeling early season recruitment of Acalymma vittatum to trap flats using degree day. x-variable Scenario Parameter Estimate StdError LowerCL UpperCL

Degree PA and KY θ1 0.9913 0.0481 0.9065 1.0938 Day 2010 - 12 θ2 3.3906 0.4486 2.6487 4.3904 θ3 0.0140 0.0020 0.0107 0.0185 PA and KY θ1 1 0 2010 – 12, θ1 θ2 3.3415 0.3324 2.7491 4.0802 forced to 1 θ3 0.0138 0.0013 0.0115 0.0167

Table 5-7. Parameter estimates for the Gompertz equations modeling early season recruitment of Acalymma vittatum to trap flats using calendar day. x-variable Scenario Parameter Estimate StdError LowerCL UpperCL

Calendar KY 2010 -12 θ1 1.3350 0.2824 0.9774 3.1930 Day θ2 7.9811 2.1841 4.0417 14.3457 θ3 0.0624 0.0189 0.0266 0.1155 PA 2010 -12 θ1 0.9870 0.0784 0.8541 1.2515 θ2 24.7061 6.8221 12.5093 54.3958 θ3 0.1732 0.0481 0.0868 0.3799

Figure 5-1. Gompertz curve showing early season recruitment predictions for PA and KY 2010 – 11 using cumulative degree days base 55 F for Acalymma vittatum.

100

Figure 5-2. Gompertz curve showing early season recruitment predictions for PA 2010 - 2012 using calendar day for Acalymma vittatum.

101

Figure 5-3. Gompertz curve showing early season recruitment predictions for KY 2010 - 2012 using calendar day for Acalymma vittatum.

6/23 7/17 8/4 9/4

1st field 2nd field biofix generation generation emerges oviposition emerges

Figure 5-4. Average number of A. vittatum per plant during the growing season for PA 2011 with overlay showing the biofix and projected development times for the first and second field generations

6/1 6/21 7/8 7/25

8/11

1st field 2nd field biofix generation generation emerges oviposition emerges

5/24 6/14 7/6 7/24 8/9

8/31

1st field 2nd field biofix generation generation emerges oviposition emerges

Figure 5-5. Average number of A. vittatum per plant during the growing season for Kentucky with overlay showing the biofix and projected development

104

6/8 6/28 7/16 8/2 8/19

1st field 2nd field biofix generation generation emergesoviposition emerges

6/2 6/19 7/16 7/19 7/31

1st field 2nd field biofix generation generation emergesoviposition emerges

Figure 5-6. Average number of A. vittatum per plant during the growing season for Iowa with overlay showing the biofix and projected development.

9/8 7/1 7/15 7/25 8/10 1st field generation egg biofix emerges

8/6 6/17 7/5 7/21

1st field generation egg biofix emerges 5/29

8/15 6/22 7/12 7/28

egg biofix

5/31

1st field generation emerges

107

7/18 8/8 8/31

6/25

1st field generation adult biofix emerges

7/21 8/8 9/4

1st field generation 7/1 egg biofix emerges

1.0 0.9 2011 6/23 0.8 7/30 0.7 0.6 0.5 1st gen observed 2nd gen observed 0.4 1st gen predicted

proportion SCB proportion 0.3 2nd gen predicted 0.2 0.1 0.0

cumulative degree days base 55 F

109

1.0 1st gen observed 0.9 2nd gen observed 1st gen predicted 0.8 2010 2nd gen predicted 0.7 0.6 0.5 6/16 7/27 0.4

proportion SCB proportion 0.3 0.2 0.1 0.0 0 250 500 750 1000 1250 1500 1750 2000 2250 2500

1.0 1st gen observed 0.9 1st gen predicted 2011 0.8 0.7 7/16 0.6 0.5 0.4

proportion SCB proportion 0.3 0.2 0.1 0.0 0 250 500 750 1000 1250 1500 1750 2000 2250 2500

Figure 5-11. Observed and predicted accumulation of A. vittatum in Iowa during the growing season. Predicted lines were generated using the equation y = 0.9913*exp(-exp(3.3906 – 0.014*x)) where x is cumulative degree days base 55 F. The division between first and second field generation was estimated as 793.6 dd55 after the biofix.

110

1.0 0.9 2010 0.8 0.7 0.6 0.5 6/10 7/13 0.4

proportion SCB proportion 0.3 1st gen observed 0.2 2nd gen observed 1st gen predicted 0.1 2nd gen predicted 0.0 0 500 1000 1500 2000 2500

1.0 0.9 2011 0.8 0.7 0.6 0.5 5/30 7/9 0.4

proportion SCB proportion 0.3 1st gen observed 0.2 2nd gen observed 1st gen predicted 0.1 2nd gen predicted 0.0 0 500 1000 1500 2000 2500

Figure 5-12. Observed and predicted accumulation of A. vittatum in Kentucky during the growing season. Predicted lines were generated using the equation y = 0.9913*exp(-exp(3.3906 – 0.014*x)) where x is cumulative degree days base 55 F. The division between first and second field generation was estimated as 793.6 dd55 after the biofix.

Chapter 6

Conclusions

The diverse agroecosystem in Pennsylvania presents a challenge to growers to manage the endemic and introduced pests and their associated diseases. Looking at this problem over a large scale, both temporally and spatially, invites the use of new techniques to investigate population movement and phenology.

As identified in Chapter 2, the aphid community in Pennsylvania is diverse even when sampled in an agricultural field. In addition to describing the alate aphid community in PA snap beans, I further investigated the overall biodiversity of aphids in PA by generating a comprehensive review of the aphid species present using the results of the pan trapping study,

Wallis et al (2005), and Pepper (1965).

Sampling only in one habitat, snap bean fields, yielded a surprisingly high percentage of the over 400 species present throughout PA and NY (~14% and ~18% respectively). Although our sampling method was concentrated on one habitat, we did intercept aphids moving from the surrounding forests and hedgerows. The high degree of landscape heterogeneity and crop diversity in the trapping areas includes plants that serve as hosts for many of the species that represented less than 1% of the total capture (Pfleeger et al. 2006). These aphids were captured in very small numbers (mostly singletons), and are not important contributers to the plant virus epidemics reported by Nault et al (2009).

112 introduction (soybean and alfalfa, respectively). While not known to colonize Phaseoulus spp., both species were determined to be competent vectors of the legume strain of CMV (Gildow et al

2008).

The species assemblage in an area is always changing with the introduction of new species. The U.S. saw the introduction of soybean aphid in 2000, and it was eventually found in

Pennsylvania. It is relevant because of its ability to transmit plant viruses as well as causing feeding damage on soybean in high numbers. Because of the absences of large quantities of the primary host in PA, I hypothesized that the soybean aphid population found in PA during the growing season is largely influenced by migrants. I used genetic tools and air-flow trajectory models to investigate the natal sources of A. glycines in Pennsylvania.

Using microsatellite and SNP data, we demonstrated that the genotypic diversity of soybean aphids in Pennsylvania was very high (0.81 to 1.00). In 2009, diversity decreased over time, but in 2010 the opposite occurred, therefore we could not define a consistent temporal trend

113 in genotypic diversity. Genotypic diversity in a field is influenced by how many different clones initially colonize a field, how successful those clones are, and how many clones colonize the field as the season progresses and their success. Our results showing high initial diversity indicate that we have many clones colonizing our fields, and a few of them are present as the season progresses. As the season progresses, some aphid clones are maintained and others are new migrants or die off or are not resampled.

Ottawa very few times, but the data suggest that with comparisons to more populations we could successfully estimate the natal source of soybean aphid populations in PA.

(0.68 – 0.97), and higher levels in the late season collections (0.87 – 1.00). The genotypic diversity from the collections in PA, NY, Canada (2010) and VA (2009) more closely resembled that of the late season Midwest collections (0.86 – 0.97 in 2009, 0.81 – 1.00 in 2010). This high genetic diversity and the lack of genetic differentiation between populations sampled in PA suggest high levels of aphid movement, and the lack of a solid local population colonizing from surrounding buckthorn in the spring. Early season aphid density in soybean fields was found to be best predicted by the amount of buckthorn in the surrounding landscape in close proximity to the fields (Bahlai et al 2010). The low early-season colonization densities we observed combined with the high genotypic diversity would be consistent with relatively rare colonization events expected from long distance migration.

114 There are multiple avenues for future work necessary to successfully integrate molecular identification techniques and aerobiological modeling into a useful management and risk assessment tool. One would be the need to have a concerted effort to accurately represent the range of buckthorn, the primary host. Since buckthorn is the limiting factor in where populations can overwinter, understanding its range is key to identifying local sources of the aphid. There are many assumptions inherent to modeling and predicting long distance aphid movement including the assumption that aphids actually getting into the air column (demonstrated with the suction trap network, Schmidt et al 2012) and then being deposited at some point along the way.

Programs like HYSPLIT give a good visualization of where air parcels are going and with further investigation could be a useful forecasting tool.

The continued development of molecular techniques to identify aphid populations will be of use with the emergence of soybean with aphid resistance traits and the subsequent aphid biotypes with resistance characters of their own. Also, if soybean aphid ever branches out to use any of the other Rhamnus species present in the landscape, these tools could be used to identify biotypes or subspecies. This merging of molecular techniques and aerobiology is not limited to this system, and could be expanded to other economically important pests.

In Chapter 4, I used general insect primers to ask questions about the possible origin of T. trifolii in the United States. The results represent a summary of the literature and a preliminary attempt to use genetic tools to elucidate the relationships between T. trifolii in the Old and New

World. Our samples were from a limited geographic range, and did not indicate clear population- level differences.

In Chapter 5 we set out to monitor the early season activity and growing season phenology of three pests on land that is transitioning to organic production on research farms in

Pennsylvania, Iowa and Kentucky on two commercially important cucurbit crops (muskmelon –

Cucumis melo and butternut squash – Cucurbita moschata) to better inform organic management

115 practices. We also set out to create accurate phenology models of these pests using air- temperature degree-days from weather stations and development times from the existing literature.

This study focused on advancing our ability to estimate early season activity using air temperature degree days. When cumulative degree days base 55 F was used as the predictor, the first inflection point of a curve modeling recruitement to trap seedlings was approximately 150 degree days, and ranged from 140 to160. Looking at the 30-year climatology and assuming that beetles begin their activity between 100 and 200 cumulative degree days, we can estimate that as a range of approximately 11 days for PA and 10 days for IA and KY occurring in mid to late

May.

The pattern of recruitment of striped cucumber beetle on the trap flats which occurred prior to commercial transplanting on farms was consistent with the findings from Smyth and

Hoffman (2003) who measured recruitment during the summer field season. They used male and female beetles and plants with high or low levels of cucurbitacin in different combinations to determine ecological mechanisms of attraction. They found that the beetles oriented upwind, especially on traps containing pioneer males (Smyth and Hoffman 2003). In our study, we had low initial counts on the trap flats, but once beetles were found on a flat, subsequent samplings had much higher counts (some flats had hundreds, especially in Kentucky). The pattern of the

Gompertz equation represents this aggregative activity well, by increasing more rapidly than a simple logistic curve (Winsor 1932).

Predicting early season activity of an insect that overwinters as an adult is challenging without knowing the cues that are inducing the end of diapause and factors that affect spring

116 activity. We are using a single variable to attempt to model the start of spring activity for these pests, but what we are really modeling is the recruitment to a trap flat. There are other variables

(soil temp, day length, appearance of other food sources etc) that could give more precise results once the mechanism for exiting diapause and initiating spring activity is known.

This study also reinforced the trap flat methods used in other studies (Lewis et al. 1990 and Radin and Drummond 1994). Our use of Blue Hubbard trap flats well distributed on farms with a history of cucurbit plantings was successful in attracting striped cucumber beetle and squash bug in KY and PA. The lack of trap flat success in Iowa probably stems from the fact that there is less cucurbit pest pressure on that farm and they experienced cold winters before the start of the study.

South Carolina (Elsey 1988).

We could not observe a clean delineation between the field generations or the overwintered adults. This generational overlap makes it a challenge to predict a gap between generations as a safe time to specify management options that require the absence of adults, such as removing row covers for pollination. Adult females are also long-lived, with the potential to survive over 100 days post-emergence and continue laying eggs during that time (Ellers-Kirk and

Fleischer 2006). For management purposes, it may be more accurate to think of the striped

117 cucumber beetle population as a single cohort, comprised of overwintering and field generated adults.

Squash bug appears to have one field generation per year in PA, KY and IA. KY is the state that comes the closest to having the degree days necessary for a second field generation, but that potential is limited by the critical photoperiod that induces diapause. The critical photoperiod of 14.5: 9.5 to 14:10 (L: D) occurs from July 26/27 to August 9/10 for Rock Springs PA, July 17 to August 4 for Lexington KY, and July 29/30 to August 11/12 for Gilbert IA (visualized by vertical lines in Figures 5-7, 5-8, and 5-9). Squash bugs that hatch after those dates in each location will not lay eggs and will diapause if they make it to adulthood (Nechols 1988). This pest also had a period of extended emergence and activity with no clear peaks between generations, but there was a large spike in nymph numbers in July that would be particularly injurious to crops.

Both the bug and the beetle vector plant diseases, thus making any effort to control them dually important for the control of the diseases. Some of the previous papers on both pests suggest clean cultivation as a way to reduce numbers, and this may not mesh with all organic farming practices. A move away from black plastic mulch, which is used to raise soil temperature and help control weeds, would be beneficial for growers concerned about squash bug because of how the bugs use the plastic as harborages, increasing their survival (Cartwright et al. 1990).

Planting a trap crop of a more attractive variety may lure the insects away from the main field, but the trap crop must then be destroyed to remove the pest population.

Squash vine borer had the lowest densities of this pest assemblage, but our results give more information into its life history in geographic areas not well-represented in the current literature. This pest also has an extended emergence and adults were captured in the pheromone traps from June through September in South Carolina (Jackson et al 2005). In our study we stopped monitoring the traps after harvest, and had low per night catches in IA and PA.

118 Especially at our site in PA, the population of squash vine borer might be so low that it is below the detection threshold of our trapping method.This work shows some consistency in time of first capture using degree days (1080.2 +/- 151.6) and calendar days (169.5 +/- 5.5). Our work also took the developmental work in Canhilal et al. 2006 and converted it from days to degree days

(base 50 F), which will help to standardize future work.

These chapters sought to use population structure and phenology to expand the knowledge of a handful of pest species, both endemic and migratory in nature that impact vegetable crop production in Pennsylvania. The alate aphid community in Pennsylvania snap bean fields is diverse and contains members that are efficient vectors of economically important plant viruses. One of these aphids, A. glycines, is largely present in the state as a result of migration from areas with high densities of its overwintering host, R. cathartica. A. glycines populations found in Pennsylvania had high levels of genotypic diversity, which was indicative of being sourced from many natal populations, and they were genetically similar to some populations in the Midwest. Matching A. glycines clones were found between PA, NY and VA indicating some level of long distance movement, which we attempted to model using air-flow trajectories. The work with T. trifolii was not conclusive, but it provides the basis for an avenue

119 of future research into the population structure of a recently introduced aphid species that could help to answer additional questions about A. glycines. Modeling the phenology of the striped cucumber beetle and squash bug was challenging due to the fact that they overwinter as adults.

We were able to demonstrate a successful early season activity monitoring tool, and use development data from previous studies on both insects to estimate the number of field generations and discuss challenges for their control in three states. We also described the phenology of the squash vine borer in geographic areas not represented in previous studies. These results can be integrated into management strategies on both conventional and organic cucurbit- growing farms.

References

Ellers-Kirk CD, Fleischer SJ (2006) Development and life table of Acalymma vittatum (Coleoptera: Chrysomelidae), a vector of Erwinia tracheiphila in Cucurbits. Environmental Entomology 35:875-880.

Elsey KD (1988) Cucumber beetle seasonality in coastal South Carolina. Environmental Entomology 17:496-502.

Gildow FE, Shah DA, Sackett WM, et al. (2008) Transmission efficiency of cucumber mosaic virus by aphids associated with virus epidemics in snap bean. Phytopathology 98:1233-41.

Jackson DM, Canhilal R, Carner GR (2005) Trap monitoring squash vine borers in Cucurbits. Journal of Agricultural and Urban Entomology. 22(1): 27 – 39.

Lewis PA, Lampman RL, Metcalf RL (1990) Kairomonal attractants for Acalymma vittatum (Coleoptera: Chrysomelidae). Environmental Entomology 19(1): 8-14.

120 Nault BA, Shah DA, Straight KE, et al. (2009) Modeling temporal trends in aphid vector dispersal and cucumber mosaic virus epidemics in snap bean. Environmental Entomology 38:1347-59.

Nechols JR (1988) Photoperiodic Responses of the Squash Bug (Heteroptera: Coreidae): Diapause Induction and Maintenance. Environmental Entomology 17:427-431.

Orantes LC, Zhang W, Mian MAR, Michel AP (2012) Maintaining genetic diversity and population panmixia through dispersal and not gene flow in a holocyclic heteroecious aphid species. Heredity 1-8.

Pepper JO (1965) A List of the Pennsylvania Aphididae and Their Host Plants (Homoptera). Transactions of the American Entomological Society 91:181-231.

Schmidt NP, O’Neal ME, Anderson PF, et al. (2012) Spatial distribution of Aphis glycines (Hemiptera: Aphididae): a summary of the suction trap network. Journal of Economic Entomology 105:259-271.

Winsor CP (1932) The Gompertz curve as a growth curve. Proceedings of the National Academy of Sciences 18:1-8.

Appendix

HYSPLIT Screenshots

On the next screen, we entered the starting location (one of our fields where we found an aphid with a genotype that matched to another location) in decimal degrees.

We used the EDAS 40km data and selected the archived data we needed (i.e. jul09.01 is the first 15 days of July 2009).

122

After the archived weather data is selected, we set up the model paraments.We changed the following from the default settings: total run time was 48 hours, starting height was 100 m

AGL (above ground level), yes to ‘plot color trajectories,’ and label interval of 24 hours. Once these parameters are set, click ‘request trajectory.’

123

124 Once the computations are complete, click the link for the GIF image and save it. After saving the

GIF, return to the results screen and click ‘modify this scenario and rerun the model.’ This returns to the model parameters screen and we can select the next day and rerun the model.

We continued doing this for all of the dates in the month before the last matching aphid was collected. When we reached the 15th of the month, we went back to the beginning and selected the data file for the last half of the month (e.g. jul09.02 is from July 16 to 31 2009).

VITA Amanda C. Bachmann

Education Ph.D. 2012, Entomology, Pennsylvania State University B.S. 2006, Biology, Case Western Reserve University

Publications Albro, S.L., S.M. Petersen, A.C. Bachmann, and P.B. Drewa. 2008. Effects of fragmentation on juvenile morphology of Acer saccharum Marsh. (sugar maple) in temperate forests of northeastern Ohio, USA. Forest Ecology and Management. Vol. 254: 233-238.

Nault, B.A., D.A. Shah, K.E. Straight, A.C. Bachmann, W.M. Sackett, H.R. Dillard, S.J. Fleischer, and F.E. Gildow. 2009. Modeling temporal trends in aphid vector dispersal and Cucumber mosaic virus epidemics in snap bean. Environmental Entomology. Vol. 38: 1347-1359.

Bachmann, A., S.J. Fleischer, and W.S. Smiles. Evaluation of foliar insecticides for the control of Lepidopterans, 2009. 2010. Arthropod Management Tests. Vol 35.

Teaching and Extension Invited Speaker: Penn State Extension Master Gardeners of Centre County Monthly Meeting. March 21, 2011. Quebec Horticulture Days. Dec. 3, 2009. Saint Remi, Quebec. Ohio Agricultural Research and Development Center/Ohio State University, Department of Entomology seminar. Sept. 29, 2009. Wooster, OH. Potter County Crops Day. March 11, 2009. Ulysses, PA. Extension Education meeting for processing snap beans for Hanover Foods. Feb. 23, 2007. Centre Hall, PA. Teaching Assistant: ENT 497A Evolution of Insects – Fall 2011 ENT 202 Insect Connections – Spring 2008, Fall 2008 Instructor: ENT 316 Field Crop Entomology – Spring 2009 Assistant coach, State College Area High School Science Olympiad team. 2006 – 2012.

Grants and Awards Penn State College of Agricultural Sciences Graduate Student Grant. April 1 – June 30 2010. Evans Family Award for Graduate Student Extension Achievement. Penn State College of Agricultural Sciences. Spring 2010. Lloyd E. Adams Memorial Grant-in-Aid. Penn State College of Agricultural Sciences. Fall 2009.