Using Stable Isotopes and Visual Gut Examination to Determine the Diet

AN ABSTRACT OF THE THESIS OF

Laurel A. Moulton for the degree of Master of Science in Horticulture presented on August 5, 2011

Title: Using Stable Isotopes and Visual Gut Examination to Determine the Diet Composition of Pterostichus melanarius (Coleoptera: Carabidae) in Western Oregon Vegetable Row Crops

Abstract approved:

______

Ronald E. Peachey

The omnivorous ground beetle, Pterostichus melanarius Illager (Coleoptera: Carabidae), is a ubiquitous inhabitant of western Oregon crop land. Because it is often the most abundant ground beetle, and attains densities that could lead to serious impacts on its prey sources, I sought to understand the overall diet composition, and specifically the importance of weed seeds in its diet at three sites in western Oregon. I compared the use of stable isotope analysis and manual gut dissection as tools for determining diet composition, and examined the influence of site characteristics.

Visual gut inspection identified broad groups of food items that included arthropods, worms and plant material, while the analysis of δ13C and δ15N signatures revealed more defined prey groups including small ground beetles, spiders, soil larvae and C3 and C4 seeds. Visual examination of gut contents, and comparing the δ13C and δ15N signatures of potential prey sources with gut content each lead to diet compositions that varied between sites and seasons. However, both techniques indicated that the primary contributors to the diet of P. melanarius are ground dwelling arthropods and earthworms. Specifically, the δ13C and δ15N signatures of small ground beetles, earthworms, spiders, centipedes and soil dwelling larvae made the largest contributions to the δ13C and δ15N signature of the diet at each of the sites. Weed seeds did not contribute significantly to the beetle’s diet, however there is preliminary evidence that suggests further examination of the larval diet may reveal that seeds are important at that stage of life. I did find overall differences in δ13C and δ15N diet signatures between sites—the isotopic signatures at the two most geographically similar sites were different than those at the third site. Because general categories of food that were included in the diet did not differ between sites, we attribute isotopic differences to as yet undetermined site characteristics.

© Copyright by Laurel A. Moulton August 5, 2011 All Rights Reserved Using Stable Isotopes and Visual Gut Examination to Determine the Diet Composition of Pterostichus melanarius (Coleoptera: Carabidae) in Western Oregon Vegetable Row Crops

by Laurel A. Moulton

A THESIS

submitted to

Oregon State University

in partial fulfillment of the requirements for the degree of

Master of Science

Presented August 5, 2011 Commencement June 2012

Master of Science thesis of Laurel A. Moulton presented on August 5, 2011.

APPROVED:

______Major Professor, representing Horticulture

______Head of the Department of Horticulture

______Dean of the Graduate School

I understand that my thesis will become part of the permanent collection of Oregon State University libraries. My signature below authorizes release of my thesis to any reader upon request.

______Laurel A. Moulton, Author

ACKNOWLEDGEMENTS

I would like to send my sincere thanks to everybody who supported me throughout my Master’s study and research process. First, I would like to express gratitude to my advisor, Ed Peachey, who invited me into the world of weeds and the life of ground beetles in agricultural lands. He gave me the freedom and trust to dive into a topic that pushed the boundaries of my experience. I would also like to thank him for setting an example for all of us-- of integrity, hard work and dedication to family, Willamette Valley farmers, and the weed management community.

I am grateful to all of my committee members. I wish to thank Gail Lagellotto for introducing me to the use of stable isotopes in the insect world, and for providing resources and ideas for my work. Special thanks to Peter McEvoy for his dedication and insight into insect ecology and many good ideas and discussions about my work. I would also like to thank John Lambrinos for his support, use of lab equipment and excellent comments. A special thanks to my whole committee for and having confidence in my work.

I would also like to thank other members of our lab. Thank you to Jess Green for being a true friend and providing support in my field work, beetle identification, writing and general wellbeing. I would also like to thank those who set the ground work for my research and provided dedicated assistance in the field: Jessica Pitcher, Alysia Greco, Nikki Marshall and other helpers along the way. I am indebted to Mike Russell for his statistical guidance and patience.

All of my work would not be possible without the additional support of other members of the Horticulture Department at Oregon State University. I will never forget the kindness and support of Anita Azarenko in helping me take steps towards my dream of teaching, and providing support so that I could extend my studies. Members of the office staff made my life many times easier with their expertise and endless patience: special thanks to Lee Ann Julson, Gina Hashagen, Viki Meinke and Wayne Wood.

Finally, I would like to express love and thanks to my family for their unwavering love, and their confidence in my abilities. I owe my love of horticulture, the outdoors, and University Extension programs to my parents Curt and Martha Moulton. I am also grateful for their encouragement and many visits and phone calls. I would like to thank my husband John McMillan for waiting for me up in Washington as I took two long years to complete my studies, and to gain some much anticipated teaching experience. I don’t know where I would be without his love, support, and occasional technical assistance. Last but not least, I would like to send my sincere love and thanks to Grandpa Kirby for always pushing me higher.

TABLE OF CONTENTS

Page Chapter 1. Introduction…………………………………………….. 1

Chapter 2. Literature Review………………………………………. 3 Vegetable production and weed control in western Oregon 4 Pterostichus melanarius species description……………… 6 Digestive system and eating habits………………………... 7 Molecular diet analysis methods………………………….. 11 Chapter 3. Materials and Methods…………………………………. 18

Chapter 4. Results…………………………………………………... 25 Tables and Figures…………………………………………. 30

Chapter 5. Discussion………………………………………………. 41

References………………………………………………………….. 49

LIST OF FIGURES

Figure Page

1a. Frequency distribution of the δ13C signatures of P. melanarius foregut contents at Benton, Linn and Marion County research sites……………………….. 36

1b. Frequency distribution of the δ15N signatures of P. melanarius foregut contents at Benton, Linn and Marion County research sites……………………….. 37

2a. Mean isotopic composition and standard error for potential prey sources and gut contents of P. melanarius at Benton site in July, August and September………………………………………………… 38

2b. Mean isotopic composition and standard error for potential prey sources and gut contents of P. melanarius at Linn site in July, August and September……………………………………………. 39

2c. Mean isotopic composition and standard error for potential prey sources and gut contents of P. melanarius at Marion site in July, August and September…………………………………………………. 40

LIST OF TABLES

Table Page

1. Weed seeds germinated from surface-sampled soil at each site. Weeds are listed by family, and species when possible……………………………………………… 30

Arthropod prey items trapped at each site listed by order, family and species when possible……………...... 31

2. Percent of beetles examined containing identifiable parts from food categories including arthropods, earth worms, seeds, other plant material, or having completely empty foreguts…………………………………. 32

3. Pearsons correlation coefficients and associated p values for combinations of items found in the foregut, and combinations of materials identified in the foreguts of beetles, and site characteristics……………… 33

4. Reported P-values from MPRR comparisons……………….. 34

5. Estimated ranges of prey contribution to the diet of P. melanarius by site and sample period based on IsoSource stable isotope mixing model for partitioning excess sources……………………………………………….. 35

Using stable isotopes and visual gut examination to determine the diet composition of Pterostichus melanarius (Coleoptera: Carabidae) in western Oregon vegetable row crops

Chapter 1

Introduction 2

Weed seed bank reduction by invertebrate seed predators and the site characteristics that may enhance seed predator populations in arable land have been explored recently in western Oregon and other northern temperate regions (Marino et al., 1997; Menalled et al., 2000; Orrock et al., 2003; Green, 2010). In these areas ground beetles (Coleoptera: carabidae) are often identified by researchers as the most important invertebrate post-dispersal seed predator on agricultural land (Honek et al., 2005). Seed predation may be a significant source of weed seed mortality (Crawley, 1992), and it is likely that seed eating ground beetles are key post-dispersal seed predators in the Pacific Northwest.

In European crop fields the ground beetle Pterostichus melanarius is often found at densities that may be high enough to affect prey populations (McKemey et al., 2003), and the same is true for some areas in western Oregon. Efforts have been made to evaluate the impact it has on weed seeds in this area (Green, 2010; Peachey & Greco, unpublished), and it has been demonstrated that this beetle removes seeds from seed stations placed in crop fields. However, it is still unclear whether seed removal can be equated to seed destruction or consumption by these beetles. Like many species of ground beetle, P. melanarius includes and may prefer other prey sources including slugs, worms and aphids in its diet. Understanding the seasonal composition of the diet of P. melanarius will lead to a better understanding of the role the species can play as a biocontrol agent of weeds and invertebrate pests in agricultural crops. The results of this work in western Oregon may also serve as an example for further research in other temperate locations.

The main objectives of the current research are to: 1) Determine diet composition of P. melanarius in western Oregon vegetable production fields, and specifically to evaluate the inclusion of weed seeds; 2) Compare the results of manual gut dissection and stable isotope analysis as diet analysis tools for ground beetles; 3) Determine if diet composition is the same between three sites in western Oregon, and over the duration of the beetles adult active season; 4) Evaluate correlations between diet composition and factors including prey availability and site characteristics.

Using stable isotopes and visual gut examination to determine the diet composition of Pterostichus melanarius (Coleoptera: Carabidae) in western Oregon vegetable row crops

Chapter 2

Literature Review 4

Vegetable production and weed control in western Oregon

Nationwide, Oregon is among the top four producers of vegetable row crops. It is ranked 2nd in snap beans production, 3rd in onions and 4th in the production of green peas, potatoes and sweet corn. Other top vegetable crops include vegetable seeds, squash, sugar beets and tomatoes. These crops help fuel the state‘s domestic and international agricultural market, the second largest sector of Oregon‘s economy (USDA, 2010).

One of the largest pest management costs borne by vegetable growers is managing weeds. In the Pacific Northwest, summer annual weeds like pigweed, wild proso millet and hairy nightshade are the primary concern. Summer annual weeds germinate in the spring or summer, and usually complete their life cycle before the onset of winter, causing heavy crop losses in summer crops when left unchecked. Their seeds are capable of persisting in the seed bank for several years, with one year worth of seed-shed potentially leading to weed infestations for years or sometimes decades in the future (Mohler, 2001). Depending on the crop and the method of production (e.g. organic or conventional), between 2 and 30% of the total cost of production goes towards managing these summer annuals and other weeds (Cross et al., 1988a,b; Seavert et al., 2008a,b,c,d; Seavert et al., 2010a,b,c,d). In the mid 1990‘s, it was estimated that $6 billion was being spent annually in the United States on weed control measures including herbicides, tillage and cultivation. During that time 553,000 pounds of herbicide active ingredients was applied each year to crop lands (Rinehold & Jenkins, 1993). Even with all the tools available, weeds still cost US farmers more than $4 billion in lost crops annually (Bridges & Anderson, 1992, as cited in Liebman, 2001).

Common weed control strategies can be grouped into two general categories that target most of the life stages of weeds. The first is reducing the number of seeds that enter the seed bank, or that make their way into the emergence zone. These strategies most often employ herbicides, mowing and hand removal to prevent weeds from maturing and shedding seeds. This category also includes tillage that buries weed seeds below the emergence zone. The second category includes reducing weed recruitment from seeds that are already in the seed bank. The most common strategies include the use of pre-emergent herbicides; creating a stale seed bed where tillage is used to encourage weeds to germinate so that they can be destroyed prior to planting a crop; and no-till strategies that leave weed seeds on the surface 5

where it is assumed that they lose viability faster than when they are buried (Mohler, 2001). Other techniques that can reduce the existing seed bank are less explored. Management tactics including soil solarization can be used in limited circumstances to effectively sterilize soil down to 5cm (Peachey et al., 2001), and flaming is successful in reducing seed germination in some situations (Melander & Rasmussen, 2001; Boyd et al., 2006). Seeds are also lost from the seed bank through consumption by vertebrate and invertebrate seed predators and by various fungi and bacteria (Mohler, 2001; Tooley & Brust, 2002).

Pterostichus melanarius species description

The common black ground beetle, Pterostichus melanarius Illager (Coleopera: Carabidae), is native to Western Europe but has spread to most northern temperate regions. The species can be found in a wide range of habitats-- from lowland to subalpine -- but it is best known for its strong colonizing abilities, and thriving in disturbed areas such as agricultural landscapes and urban communities. Fully two thirds of arable sites in Europe that have been surveyed have yielded P. melanarius, and it was the third most common species reported in eastern European fields (Lövei & Sarospataki, 1990). In some cases, when P. melanarius is dominant, the overall diversity of smaller ground beetles declines. In North America, P. melanarius has expanded its range to include most of the northern US, and southern regions of Canada. It is frequently the most common large carabid caught in pitfall traps in arable land in both Europe and North America (Thiele, 1977; Symondson et al., 1996; Luff, 2002), and it is the most abundant beetle in vegetable crop fields in the Willamette Valley of Oregon (Peachey et al., Unpublished).

In the Willamette Valley population, P. melanarius adults range from 1.5-2cm in length. They are active between April and November. This beetle is considered a fall breeder, with its mating season lasting from July through October. During the winter both adults and larvae retreat underground, up to 20cm deep in the soil (Larsson, 1939; Larochelle & Lariviere, 2003). They are considered dimorphic brachypterous species with some adults having wings that appear to be fully developed, and others having reduced vestigial wings (Kromp, 1999). Larochelle and Lariviere (2003) note that macroptery is relatively uncommon, but that it occurs more often in North American populations than European populations indicating the recent dispersal in North America. They are generally considered to be flightless, and there are no accounts disproving this in North American literature. Male P. melanarius tend to dominate pitfall trap catches because males are more mobile while looking for females, however catches early in the spring when the beetles are just emerging yield an almost one to one ratio of males to females (Paill, 2004).

Digestive system and eating habits

Ground beetles have a simple digestive system: essentially a tube extending from mouth to anus, divided into three primary sections. The foregut, composed of the pharynx, esophagus crop and proventriculus (gizzard), functions as a delivery system and storage area for food items before they pass through the stomodaeal valve to the midgut. The midgut, a tube of uniform diameter, is the primary location of digestion and absorption of nutrients. The pyloric valve leads from the midgut to the intestine, rectum and anus in the hindgut. In P. melanarius food is stored in the crop, a simple enlargement of the foregut, before it passes to the midgut for digestion. Hard parts such as seed coats, worm chaetae and chitin often remain in the proventriculus and rectum (Pollet & Desender, 1990). Very little digestion takes place in the foregut (Borror et al., 1989).

Food passage through the digestive system of P. melanarius is driven by time, and to a lesser extent by temperature. Between 17 and 20°C, the crop remains fullest between 2 and 6 hours after a meal, but virtually all food items will have passed from the foregut after 24 hours (10-20% capacity). Digestion is somewhat slower at 10° C with the crop averaging 40% capacity after 24 hours (Pollet & Desender, 1990).

Diet composition and gut fullness change over the season, and also differ between male and female beetles. Symondson (et al., 2000) found earthworm protein in 26% of all beetles tested in June, 28% in July and 45% in August. At the same time the overall quantity of food consumed (crop weight) decreased. Another study looked at liquid food material contained in the foreguts of P. melanarius. Results showed that the volume of liquid food material also varied by season, with larger volumes found in April-June than July-September (Paill, 2004). Aside from seasonal differences, female beetles collected in pitfall traps generally had fuller crops than males. In Paill‘s (2004) investigation, females contained an average of 3.9 ml (SD= 3.1ml) of fluid food material in their foregut, while males contained an average of 2.9ml (SD= 1.9ml).

The eating habits of P. melanarius have been studied extensively in agricultural settings. This beetle is highly polyphagous and generally feeds on prey items it encounters on the surface of the soil (Lee & Edwards, 2011; Green & Peachey, 2011, unpublished data). It has been shown through gut dissection, molecular techniques, and field and lab observations to consume a variety of invertebrates, seeds and scavenged items. One study was able to identify up to 4 different types of prey at once in the guts of P. melanarius, though the gut contents most often contained one type of food item (Harper et al., 2005). Gut dissections reveal prey items such as spiders (Araneae: Linyphiidae), rove beetles (Coleoptera: staphylinidae), flies (Diptera: Scatophagidae, Dolochopodidae), earthworms, aphids, and less specific items such as unidentifiable plant material, cuticle and invertebrate eggs (Sunderland, 1975). Pollet and Desender (1985) recorded a total of 49 different prey taxa in the guts of P. melanarius. Molecular techniques have been developed to identify known species of aphids, dipteran larvae, crickets, weevils, slugs and slug eggs, and several species of earth worms in the guts of P. melanarius (Symondson et al., 1996; Symondson et al., 1999; Zaidi et al., 1999; Symondson et al., 2000; Oberholzer & Frank 2003; Wallace, 2004; Harper et al., 2005; Winder et al., 2005). Lab feeding trials add crane fly larvae and potato beetle larvae and grass seed to the list of potential prey items (Johnson & Cameron, 1969; Kromp, 1999).

Molecular analyses have focused on identifying and quantifying specific invertebrate prey types: mainly slugs, worms and aphids. These studies seek to describe P. melanarius’ interactions with crop pests, and they show that carabid activity can be linked to reductions in aphid and slug populations (Symondson et al., 1996; Winder et al., 2005). Molecular techniques are similarly used to identify different species of earthworm prey, but no data is available connecting worm densities with carabid activity. Symondson (et al., 1996) determined that slugs are a frequent item in P. melanarius’ diet. He also noted that the presence of slug haemolymph in a beetle‘s digestive system did not correlate with crop (foregut) weight, indicating that other unknown food items were consumed regularly. Another study found that an average of 28.4% of beetles sampled tested positive for slug remains, with 39.7% of females and 22.2% of males respectively. Accompanying lab-based feeding trials established the beetles preferred smaller slugs and slug eggs and seldom attacked live slugs that were larger than 15mg (McKemey et al., 2001; Oberholzer & Frank, 2003). However others found that size choices observed in the lab do not necessarily translate to patterns in the 9

field (McKemy et al., 2003). The proportion of beetles testing positive for slug remains correlated strongly with the density of newly hatched slugs, peaking in late spring and early autumn (Oberholzer and Frank, 2003; Paill, 2004). P. melanarius are also frequent consumers of earthworms, with some studies showing that 36- 40% of all beetle samples contained worm protein. One author found that earthworm protein increased as the overall biomass of the foregut decreased, and suggested that earthworms were only consumed in the absence of other prey (Symondson et al., 2000; Harper et al., 2005).

Pterostichus melanarius follow olfactory clues to locate some insect prey using sensors located on their antennae (Wheater, 1989). In one unique case, they responded to alarm pheromone that many species of aphids release when they are disturbed. Over a 6-year investigation, Sunderland and Vickerman (1980) found that 17% of P. melanarius had consumed aphids. They never found the beetles climbing up into the plant canopy to obtain prey, but sensitivity to aphid pheromones could explain P. melanarius’ feeding success. Alarm pheromones cause aphid populations to disperse from host plants when they perceive danger, often dropping to the ground. P. melanarius has demonstrated a strong attraction to the same pheromone, finding the fallen aphids quite accessible (Nault & Montgomery, 1979; Kielty et al., 1996). In other trials the beetles have responded to olfactory stimulus from invertebrates including slugs, crickets and fly larvae (Thomas et al. 2008). Little is known about the role olfactory signals might play in the detection of plant based food, though Wheater‘s (1989) aphid study also indicated that the beetles were attracted to the scent of crude extracts from wheat seedlings.

Although most gut analyses have focused on invertebrate portions of P. melanarius’ diet, these beetles also have been identified as possible seed predators. In laboratory tests Johnson and Cramer (1969) offered P. melanarius food including various life stages of a species of turf grass weevil (Hyperodes spp.), grass foliage, and grass seed. The beetles generally preferred the weevil larvae and pupae, but the authors made an important observation: the beetles would often choose to feed on seeds even in the presence of these other ―preferred‖ food items. Jaspar-Versali and Jeuniaux (1987) found that P. melanarius has higher cellulase enzyme activity in its digestive system than other purely carnivorous beetles, suggesting that they also consume plant material. One study offers evidence that this apparent 10

choice leads to more reproductive success. When fecundity was measured in relation to diet composition, beetles that were fed a diet that was high in carbohydrates and low in protein produced fewer, but measurably larger eggs. The same study found that larvae hatching from larger eggs survived longer than those hatching from smaller eggs when they were starved for food (Wallin, et al., 1992). In this study the high carbohydrate diet was intended to represent the properties that plant material would bring to a polyphagous diet. Unfortunately, the authors used cat food that was high in carbohydrates instead of plant material. This design decision made it impossible to determine whether fecundity results could be extrapolated to the consumption of seeds or other unprocessed plant material, or if they were due to some other nutritional property of the cat food.

Diet studies using molecular techniques and manual gut dissection to identify or quantify seed predation in carabid beetles are uncommon, and are generally replaced by seed removal experiments in the field and lab- based feeding trials. Using these methods, some researchers have identified carabid beetles as the most important invertebrate post-dispersal seed predator on agricultural land (Honek et al., 2005), though many others simply attribute seed removal generally to the activity of herbivorous and generalist invertebrates (Marino et al., 1997; Menalled et al., 1999; Westerman et al., 2003; Marino et al., 2005). Relationships between carabid activity and seed predation are varied. Some studies report seed predation decreasing with the elimination of carabid predators (Honek et al., 2005; Peachey, Unpublished data) and other authors others have found that reduction in weed seeds from predation is not associated with overall carabid activity (Saska et al., 2008). These differences are most likely the result of confounding factors caused by the beetle‘s nocturnal lifestyle and opportunistic or seasonally changing polyphagous feeding habits (Lys, 1995; Cardina et al., 1996; Menalled, 2000; Mauchline et al., 2005). To overcome the limitations of field observations, laboratory-based feeding trials are employed to establish food preference in some species (Honek et al., 2005; Martinkova et al., 2006; Marshall, 2008). Unfortunately, these findings are also limited because lab studies cannot mimic the spatial and temporal availability and variety of food that occurs in the field.

Molecular diet analysis methods

The diet composition and predation habits of generalist arthropods are some of the most difficult ecological interactions to study. Their feeding routines are often cryptic -- nocturnal or occurring inside plant tissue, or under soil, debris or dense plant cover-- and many of them don‘t consume hard parts from their prey (Greenstone, 1996). Visual examination of gut contents cannot identify prey remains that do not include hard parts, and laboratory feeding trials cannot mimic the diversity of prey that is available in the field, or other conditions that lead to prey choice under natural conditions. Even scavengers and other polyphagous feeders do not consume items solely based on availability. Feeding behaviors of generalist arthropods may be influenced by spatial and temporal distribution of food items, growth stage of the predator or prey, defensive mechanisms or toxins, nutritional value of the food item, and changing availability of food items in relation to others (Symondson, 2002). Post-mortem examination of gut contents using molecular or immunological techniques does not require disturbance of the organisms environment or activities, and can potentially detect the signatures of liquid food or prey items without recognizable parts.

Immunological techniques are currently the most common tools for prey detection in arthropods, but molecular methods such as Polymerase Chain Reactions (PCR) that use DNA (Symondson, 2001) are becoming more accessible. Stable isotopes are commonly used in some fields such as the study of marine influenced diets, and aquatic ecosystems, but are becoming more widespread in the study of arthropod food webs (eg. Newsom et al., 2007).

Immunological analysis Immunological methods use antibodies that are created to bind to molecules (antigens) that are specific to one prey item or a taxonomically related group of prey items. Monoclonal antibodies can be specific to genus, species or even the life stage of a prey item in the gut contents of a consumer. Polyclonal antibodies are cheaper and faster to produce than monoclonal antibodies, but they are less specific and require extensive testing to eliminate cross reactivity with non-target items. For example, if a polyclonal antibody is designed to bind with a molecule of one mollusk, it will also react with other species of mollusk. The strength of cross reactivity is directly related to taxonomic distance from the target mollusk 12

(Greenstone, 1996). Antibodies that bind to the desired antigen can be identified by precipitation or another means of separation from the solution, or by activating a colored dye as in Enzyme Linked Immunosorbent Assays (ELISA). Immunological methods are employed when specificity (being able to differentiate the target from non target antigens) and sensitivity (being able to detect very small quantities of the desired antigen) are the desired outcomes. Some useful outcomes include detecting whether a generalist predator has consumed a specific prey item or target pest (Hagler & Naranjo, 1996). They have also been used to differentiate a specific parasitoid within its host without giving a false positive for a closely related coexisting parasitoid (Stewart & Greenstone, 1996) and identifying specific species of slug or earthworm in the guts of ground beetles (Symondson et al., 1996; Symondson et al., 1999; Symondson et al., 2000; Symondson et al., 2002ab). Detection times for these techniques vary widely depending on temperature, meal size, presence of other prey, host type and host size. Hagler and Naranjo (1997) found detection times of less than 2 hours after meal consumption for one species of lady beetle, but another study found that detection times for aphids consumed by the spider species Lepthyphantes tenuis varied by sex, and ranged up to 193 hours (Harwood et al., 2001). The variability in detection time and sensitivity due to the factors listed above makes immunological analyses insufficient for quantification of prey consumption (Hagler & Naranjo, 1997).

DNA-based analysis PCR analyses are becoming more common as specific primers and genetic sequences from all different life forms are being cataloged and becoming available at a low cost (Chen et al., 2000). PCR is effective in detecting specific prey whether it has been eaten fresh (predation) or scavenged, though the method cannot differentiate between the two sources (Juen & Traugott, 2005). Early gut analysis attempts using PCR focused on identifying single prey species (Chen et al., 2000); a lengthy process involving multiple primers and separate PCR runs was required to identify multiple prey items. Recent work has demonstrated that Multiplex PCR can be effective tool to screen for multiple prey items at the same time (Harper et al., 2005). Detection times for prey DNA are variable. Juen and Traugott (2005) found that prey could be detected 32 hours after feeding in 50% of the predators that had been fed known items, and consistently detected prey in 80% up to 16 hours after feeding. Harper et al. (2005) found that 50% detection windows for prey in the guts of P. melanarius ranged between 30 13

and 96 hours after consuming slugs, aphids, weevils and earthworms. The predators in these two studies were kept at relatively cool temperatures (16°C); prey detection times may be significantly shorter under warm conditions (Hoogendorn & Heimpel, 2001). PCR analysis is not generally useful for quantifying the number of prey that are consumed (Greenstone, 1996) and cannot be used to differentiate between life stages of an organism like immunological techniques can (Chen et al., 2000 and citations within). Though detection times can be long, DNA denatures faster in predator guts than other proteins do, and so it is useful to design primers for short DNA sequences (Gookool et al., 1993).

Stable isotopes In the field of animal and insect science stable isotope analysis has historically been used to determine diet composition and geographic sources of food. In diet studies, stable isotopes are most useful in determining the composition when an organism‘s diet includes two or three known food items or categories of prey with distinct isotopic signatures. For example stable isotope analysis was used to quantify the relative contribution of aphids that fed on different C3 and C4 crop plants (e.g. alfalfa, wheat, and maize) to the diet of ladybird beetles (Ostrom et al., 1997). Painter (et al., 2009) quantified the relative importance of multiple moth species to the diet of spotted bats. In geographic source partitioning, stable isotope analysis has been used to determine the contribution of marine and terrestrial based foods in the diets of bird species including gulls, murrelets, owls, cormorants and alcids (Hobson & Clark, 1992). More recently the technique has become a valuable tool for investigating trophic relationships in insect and other communities (e.g. Bluthgen et al., 2003; Tooker & Hanks, 2004; Langellotto et al., 2006).

In addition to investigating the composition of an organism‘s diet, stable isotopes can detect changes in diet throughout the organism‘s life, and differentiate between organisms that eat mixed diets from those that have switched diets (Gratton & Forbes, 2006). Stable isotope analysis is generally used to evaluate various tissues in an organism‘s body, but can also be used directly on gut contents if only recent feeding habits are of interest. Diet switching lab experiments have been carried out to determine the assimilation of δC13 and δN15 in different tissues. Generally an organism is fed a carbon rich diet for a period of time and then switched 14

to a nitrogen rich diet, or switched between otherwise isotopically distinct diets. After the switch tissues are sampled repeatedly over the duration of the experiment to establish isotopic replacement curves (Tieszen et al., 1983; Hobson & Clark, 1992; Gratton & Forbes, 2006). Based on the use of stable isotope ratios in different body tissues to determine an organism‘s lifelong eating habits, Bearhop et al. (2004) proposed that stable isotopes could be a valuable tool in determining an organism‘s trophic niche width. They note that traditional niche width studies only capture a snapshot of an organism‘s diet, but analyzing stable isotope ratios in different tissues in conjunction with conventional measures, including prey preference and habitat choice, can lead to a detailed understanding of niche utilization over a lifetime.

This tool can also help expand knowledge about individual taxon that may be difficult to observe. For example Langellotto (et al. 2005) found evidence suggesting differences in the life history of one facultative hyperparasitoid between research sites. Facultative hyperparasitoids may act as a primary parasitoid on an herbivore host, or it can use a parasitoid of the herbivore as a host (i.e. act as a hyperparisitoid). In this case it acted solely as a primary parasitoid at one site, but functioned primarily as a hyperparasitoid at the other site.

Stable isotopes are atoms of the same element that have the same number of protons and electrons, but vary slightly in the number of neutrons; the closer in number the protons and neutrons are, the more stable they are. There are about 300 known stable isotopes throughout the periodic table. Usually the lighter ones (most commonly H, O, C, N and S) are used for ecological research because the addition of one neutron causes the largest percent increase in atomic mass for the lightest elements. Carbon and nitrogen are most frequently used because they are common in organic compounds (Sulzman, 2007).

Two isotopes of one element can participate in the same chemical reactions, and therefore are qualitatively similar. Quantitative differences in reaction rates and bond strengths are caused by differences in atomic mass (Sulzman, 2007). One result of these differences is that lighter isotopes like δ14N and δ12C are excreted at a faster rate than the heavier isotopes δ15N and δ13C. This inequality in excretion rates results in fractionation, or changing ratios, of the heavy and light isotope concentrations in plant and animal tissues. Most carbon fractionation occurs as atmospheric carbon is converted to carbohydrates during 15

photosynthesis. Some carbon fractionation also occurs as animals and insects respire and 13 release CO2 back into the atmosphere. Deniro & Epstein (1978) found that progressive δ C 13 enrichment in a consumer was matched by δ C depletion in respired CO2. Nitrogen isotopes fractionate as organic matter is consumed, and the components are assimilated and excreted (Deniro & Epstein, 1981). Theoretically, with each trophic step, there is progressive enrichment of the heavier isotopes, and changing ratios of C to N.

Isotopic differences between materials can be extremely small, and so the ratio of the heavy isotope to the light isotope (i.e. δ15N/ δ14N or δ13C/ δ12C) in the test material is reported permil (‰) in relation to the ratio of the heavy and light isotopes in a set of international standards. Typical delta notation is used,

δX‰ = (Rsample/Rstandard - 1) * 1000 were X represents the heavy isotope δ15N or δ13C, and R is the corresponding ratio of δ15N/ 14 13 12 δ N or δ C/ δ C in the sample (Rsample), or in the reference material (Rstandard). Positive δX‰ indicate that the sample is more enriched in the heavy isotope than the international standard, and negative numbers indicate that a sample is more depleted than the standard (Sulzman, 2007).

Applying fractionation rates of nitrogen and carbon isotopes that have been estimated in one community to another has some limitations in. In a study examining diets ranging from birds to fish, and insects the ratio of δ15N to δ14N increased approximately 1.4±0.21‰ from diet to consumer for organisms raised on an invertebrate diet, 2.2±0.30‰ for those raised on plant or algal diets, and 3.3±0.26‰ for consumers raised on higher protein diets. The ratio of δ13C to δ12C changed little between tropic levels, showing general increases 0.5 ± 0.13% (McCutchan et al., 2003). This and other studies suggest that nitrogen enrichment is the best delineator of trophic steps. Tooker and Hanks (2004) point out that most nitrogen enrichment models developed using predators of marine, aquatic and invertebrate ecosystems and may not be as effective in some arthropod food webs. Langellotto (et al. 2006) determined that fractionation leading to an increase in the δ13C signature between trophic steps was a better tool than nitrogen enrichment for establishing trophic interactions of an insect parasitoid 16

community. In that instance there was a 1.11‰ (SED= 0.25) increase in δ13C from host plant to herbivore, and a 0.82‰ (SED= 0.21) increase from herbivore to primary parasitoid. In a wasp-predator-parasitoid community that occurs in two different aster species investigators only examined nitrogen enrichment to determine trophic interactions. They found δ15N enrichment between host plant and gall wasp, and between gall wasp and predator, but not between the predator and parasitoid in communities inhabiting one species of aster. The same insect assemblage found in a related aster species did not have the same enrichment pattern. In this case, food web assemblages for the two aster species were collected at different sites, and the authors attribute differences in stable isotope signatures to local environmental conditions (Tooker & Hanks, 2004). Stable isotopes are a valuable tool for determining trophic interactions, but care must be taken when comparing results from one species or community to results from another.

Differences in δC13 and δN15 signatures and enrichment patterns are also variable in simple herbivorous relationships. For example the Asian soybean aphid (Aphis glycines), and the corn leaf aphid (Rhopalosiphum maidis) do not show elevated levels of δC13 compared to their hosts. Instead, they have levels similar to those expected for their C3 and C4 host plants (Gratton & Forbes, 2006). Similarly, Daugherty and Briggs (2007) found that leaf hopper adults and nymphs in a pear orchard did not differ in δC13 from their host plant, but δN15 did differ significantly. Another herbivore, the pear psyllid (Cacopsylla pyricola), did not differ in either δC13 or δN15 from its plant host. Differences in species such as diet, modes of excretion and local environmental factors may also complicate enrichment rates (Vanderklift & Ponsard, 2003).

The purpose of the current study is to investigate the recent and seasonal diet of P. melanarius by directly analyzing the contents of the gut. For this reason, it is not necessary to sample the isotopic fractioning differences between tissue types. Directly analyzing gut content, or in some cases fecal matter (Painter et al., 2009), has the potential to skirt some problems with stable isotopes. Different tissues in an organism may assimilate isotopes from prey at different rates, and whole body analysis may result in signatures based on past consumption rather than current. In some cases male and female insects also have the potential to differ in prey assimilation rates. Gratton and Forbes (2006) used δC13 to study the prey 17

assimilation rates of two species of lady beetle. They found that the Harmonia axyridis males and females had the same assimilation rates, but Coccinella septempunctata males and females had distinctly different rates of prey assimilation. 18

Using stable isotopes and visual gut examination to determine the diet composition of Pterostichus melanarius (Coleoptera: Carabidae) in western Oregon vegetable row crops

Chapter 3

Materials and Methods 19

Sites

Research sites were located in Marion, Linn and Benton Counties in the Willamette Valley in western Oregon. The Marion and Linn County sites were located on the flood plain of the North Fork Santiam River near the City of Stayton. The Marion County site is located at 44° 47‘ 30‖ N, 122° 48‘ 40‖ W at an elevation of 130m. The soil is classified as Cloquato silt loam with patches of river cobble and gravel throughout (USDA NRCS, Web soil survey). The Linn County site is located at 44° 45‘ 58‖ N, 122° 49‘ 41‖ W, at an elevation of 127m. The soil is classified as Saturn variant silt loam, with a gravel component. These two sites are managed using conventional techniques, conservation tillage and winter cover crops. Crop rotations occur on a 4-year cycle that alternates 4 years of perennial rye seed crop with 4 years of strip tilled corn and bush beans. During the main sample season of this investigation both sites were planted in sweet corn, with snap beans and a winter cover crop of red clover previous. The Marion County field is irrigated with water drawn from the river starting in late May, and then switched to wastewater from the nearby Norpac vegetable processing plant as it becomes available later in the summer. Irrigation with vegetable processing wastewater continues into November after the crop has been harvested and the cover crop planted. The Linn County site draws all of its irrigation water from the river.

The Benton County site was located near the town of Philomath, OR on the Marys River flood plain (44° 31‘ 42‖N, 123° 22‘ 06‖ W; elevation 83m). The soil is classified as Malabon silty clay loam (USDA NRCS, Web soil survey). The 1.5 Ha field has been managed organically since 1987. The field hosts a diverse vegetable rotation that has included winter squash, brassicas, corn, potatoes and other minor crops during the past three growing seasons. The field was planted in a mix of sweet corn, kale and turnips for the main sampling period in the summer of 2010, and a mixture of Delicata and Butternut squash the previous season. The field is bounded to the east and south by an access road, grassy strip and riparian forest, to the west by a grassy strip and blackberry hedges. The field is irrigated through most of the summer with water that is drawn from the Mary‘s River.

Beetle trapping

Samples of P. melanarius were collected from three sites during 3 sample periods that started on July 12, August 16 and September 13 of 2010. Beetles are present at these sites between June and October, and trapping was carried out during the most active season. At each site between 15 and 30 pitfall traps were deployed randomly within a 60 x 120 meter plot. They were opened in the evening, and emptied before 11 AM the following day. Trapping continued for up to 5 days, until 20 male and 20 female beetles were collected at each site. Ten each of the male and female beetles were dedicated to stable isotope analysis and 10 to visual gut inspection. Beetle samples were kept cool during transport, and then frozen at -5°C until they could be processed.

P. melanarius third instar larvae were collected at one site in March 2011 in order to investigate the larval diet. Individual larvae were located by looking for burrow entrances among the roots of corn stubble. The larvae were removed by digging up the corn root balls, and hand sorting.

Prey Availability

The availability of potential prey items was established at each site during the three sample periods. Specific prey categories including slugs, earthworms, soil arthropods and seeds were investigated. In addition to measuring availability, prey items from each site were analyzed for stable isotope signatures that could be compared against the stable isotope signatures found in the foreguts of the beetles.

Worms and slugs were extracted from the soil using an aqueous mustard powder solution as an irritant. Ten replicate samples were collected per site using plastic circles with an area of 573cm2 pressed into the ground. A total of 8 liters of mustard solution (3.5g mustard power/ 1 L tap water) was poured gradually into each circle, allowing each addition to soak in before more was added. The earthworms and slugs were rinsed in fresh water (Zwahlen et al., 2003), counted, weighed and identified to species when possible. The circles were left in the field over night and checked again the following morning (Dreves, personal communication). The mustard vermifuge method is the most successful when soil moisture 21

levels are moderate and earthworms are most active. The method is best for earthworms such as Lumbricus terrestris that inhabit permanent vertical burrows (Muramoto & Werner, 2002). The accuracy of this method was assessed by collecting and hand sorting the top 25cm of soil from three random vermifuge samples at each site during each sample period. This check revealed that the current vermifuge method was successful in extracting between 45 and 90% of the worms in the upper 25 cm of soil. Another investigator found that mustard suspension vermifuge extracted an average of 61.4%±4.2 of the worms that were extracted by sorting and counting (Lawrence & Bowers, 2002). This method was also recommended for slug sampling, however it did not yield a significant quantity of slugs at these sites.

The activity density and diversity of other soil-dwelling arthropods was established by identifying and enumerating non-Pterostichus arthropods that fell in the pitfall traps. These arthropods included members of the Carabidae and other families of beetles, spiders, centipedes, millipedes, and various incidental species. Samples of each potential arthropod prey item were collected for stable isotope analysis.

The availability of viable seeds near or on the soil surface was approximated by collecting 10 soil samples at each site in July, August and September. The top layer of soil was scraped from the surface of a meter square plot to equal 1 liter of soil. The soil samples were broken up and spread on top of a 5 cm peat moss base in a standard greenhouse flat. The flats were watered regularly and left to germinate in a moderate greenhouse until there was no further germination for five days. Once the weed seedlings grew their first true leaves, they were identified to genus or species, counted and removed from the trays.

Gut dissection

Beetles were dissected in order to identify food items in their upper digestive tract. The forgut, including the crop and proventriculus, was removed and assigned a fullness factor between 1 and 4 (with 1 being empty, and 4 being completely full and distended). The contents were examined under a dissecting microscope at up to 100x for the presence of identifiable prey. Items were identified in broad categories including insect, worm, mollusk, 22

seed, other plant material and unknown. After thorough examination, Lugol‘s iodine solution

(K2KI) was added to detect any plant material that contained starch.

Stable isotope analysis

The foregut contents of adult P. melanarius, whole P. melanarius larvae, and items that were available as prey (2-6 samples of each species or prey category) that were collected between July and September of 2010 were analyzed for stable isotope signatures. The foregut, composed of the crop and proventriculus, was removed from each beetle and dried for a minimum of 24 hours in a 60°C oven. In order to attain proper sample weight and reduce sample variability each sample was composed of the homogenized foreguts of three male or three female beetles. P. melanarius larvae samples were composed of individual whole larvae that were dried and homogenized before sampling for analysis. Insect prey items were analyzed whole, or in multiples to attain proper sample weight (for example each spider sample was composed of 2-3 whole spiders) (McNabb et al., 2001). Worm samples were composed of slices taken randomly from the body. Seeds were combined in C3 and C4 groups and homogenized. It was assumed that no changes occurred in the isotope ratio between the fresh prey items , and the little-digested contents of the beetle‘s foregut that I analyzed. Each 1 ±0.1 μg sample was encased in a tin capsule, and flash combusted at >1000°C using a Carlo Erba NA1500 elemental analyzer. The isotopic composition (δ13C and δ15N) of the resulting

CO2 and N2 gases were analyzed by continuous-flow mass spectrometry using a DeltaPlusXL isotope ratio mass spectrometer. Carbon and nitrogen stable isotope ratios were analyzed simultaneously for most samples; however, the C/N ratios of seeds were too high which necessitated measuring the carbon and nitrogen stable isotopes separately. Measured values were normalized to the VPDB and AIR scales for δ13C and δ15N respectively, using the international standards USGS40(L-glutamic Acid), IAEA-CH-6 (sucrose) and IAEA-N2 (ammonium sulfate) which were included in all batches of samples. The international standard IAEA600 (caffeine, expected d13C = -27.77‰, d15N = +1.0‰) was also include in each batch of samples as an internal check (n=30, δ13C =-27.7 ±.07, δ15N = 1.0±0.1).

Data analysis

The density and biomass of earthworms, relative abundance of ground dwelling arthropods, and the density of germinable weed seeds near the soil surface during July, August and September of 2010 was established. Worm data are expressed as the density of individuals or the weight of worm biomass per meter square (25cm deep); seed data are expressed as the number of germinable seeds in the top 1cm of soil per meter square; arthropods were expressed as a mean number of individuals caught per trap night.

The frequency of prey group identification during visual inspection of beetle foregut contents are averaged over all beetles at each site during each sampling period. Pearson product-moment correlation coefficients were calculated to assess the relationship between site characteristics and observed foregut contents in the captured beetles.

To compare the δ13C and δ15N values of foregut contents between male and female beetles, and between the three sites and sampling dates Welch modified 2- sample t-tests, and Analysis of Variance with Bonferroni simultaneous confidence intervals were used. Kruskal- Wallis rank sum tests were employed for comparisons between δ13C values on different sampling dates at the Benton County site because of problematic residuals that could not be improved with standard data transformations. A non parametric multiresponse permutation procedure was used (MRPP; PC-Ord 6 software) to make pairwise comparisons between the δ13C and δ15N values of the beetle gut contents and the values of potential prey and non prey items. This procedure did not require distributional assumptions, and identifies matches and non-matches using analysis of a Euclidean distance matrix (McCune & Grace, 2002). Prey items with δ13C and δ15N that were indistinguishable (p> 0.05) from the δ13C and δ15N were considered the most likely prey items.

Stable isotope mixing models

The resulting δ13C and δ15N signatures were analyzed using IsoSource mixing model software. This software is used to estimate the range of dietary contributions of possible prey sources when the sources exceed the number of isotope ratios analyzed plus 1 (Phillips & 24

Gregg, 2003). Using increments of 1%, all the possible combinations of prey items between 0 and 100% that are consistent with the isotopic mass balance of the gut contents (within a small range of tolerance) were identified. Prey combinations that fell within the tolerance range were considered isotopically feasible diets. A minimum tolerance level of 0.14‰ was calculated for the IsoSource model based on the greatest isotopic difference between prey sources (C3 and C4 seeds), and a 1% source increment used to run the model (tolerance= 1% of the greatest isotopic difference, ~14‰ δ13C). The actual tolerance levels used for modeling ranged between the minimum of 0.14‰ and 1.9 ‰ because of limitations in the data.

Using stable isotopes and visual gut examination to determine the diet composition of Pterostichus melanarius (Coleoptera: Carabidae) in western Oregon vegetable row crops

Chapter 4

Results 26

Prey availability

The weed seed bank in the top centimeter of the soil was greatest at the Benton County site, the only organic farm included in the study. When sampled in July this site had a germinable seed bank that averaged 607 (SD= 468) seeds/m2 in the top centimeter of soil, while the Linn and Marion sites averaged 31 (SD= 36) and 2 (SD= 6) seeds/m2. The weed species richness was also greatest at the Benton county site. There were 15 species of weed seeds germinated and identified at the Benton site, 3 at the Linn site and 5 at the Marion site (Table 1).

An average 94% of all the potential arthropod prey items trapped throughout the summer were ground beetles or non- insect arthropods (mostly Linyphiidae and Lycosidae spiders, as well as centipedes). The Benton County site had the highest taxonomic richness of arthropods, ranging between 10 and 15 genera/orders per sample period; Marion ranged between 5 and 11, and the Linn site had 7 genera/orders found during each of the sample periods. There were 24 total genera/orders found at the Benton site, 17 at the Marion site, and 14 identified at the Linn site (Table 2).

The earthworm fauna at the study sites was composed mainly of the common night crawler, Lumbricus terrestris. These worms live in permanent vertical burrows, and come up at night to scavenge organic materials on the soil surface, making themselves available as prey to foraging P. melanarius. The Linn County site had extremely high worm densities, respectively averaging 134 (SD= 46), 391 (SD= 259) and 284 (SD= 158) worms/m2 in July, August and September. The Benton and Marion sites had substantially fewer worms averaging 17 (SD= 26), 14 (SD= 16) and 38 (SD= 27), and 19 (SD= 23), 49 (SD= 47) and113 (SD= 102) worms/m2 across the same time period.

Gut content: visual observations

Visual examination revealed that arthropod and worm remains were the most common, identifiable food material contained in the foregut of P. melanarius at all sites through the summer and fall of 2010 (Table 3) . The Benton County site had the most frequent observation of arthropods with an average of 75% (SD= 8%) of beetles containing arthropod 27

remains over the three sample periods, followed by Marion (67%, SD= 11%) and Linn County (59%, SD= 15%). Worms were identified in an average 21% (SD= 3%), 37% (SD= 11%) and 20% (SD= 20%) of beetles in Benton, Linn and Marion County with other prey items including seeds, and other plant material being represented substantially less. Beetles with empty foreguts accounted for an average of 6% (SD= 7%), 15% (SD= 5%) and 11% (SD= 13%) of the beetles examined at the Benton, Linn and Marion County sites.

Pearsons correlation coefficients (Table 4) suggest strong positive relationships between available worm biomass and worm material identified in gut contents (r= 0.75, p= 0.02), and between arthropod richness and the number of beetles containing arthropod remains (r= 0.69, p= 0.04). There was a strong negative correlation between availability of worm biomass and arthropod material found in the guts (r= -0.80, p= 0.01), and a weaker negative correlation between arthropod richness, activity density of ground beetles and presence of worm remains in gut contents (r= -0.59, p= 0.01; r= -0.48, p= 0.2). Coefficients suggest weaker correlations between presence of arthropod remains, and arthropod diversity and activity density of small ground beetles (r= 0.66, p=.05; r= 0.61, p= 0.08).

Gut content: stable isotope analysis

There was no detectable effect of sex on the δ13C and δ15N isotopic signatures of foregut contents (δ13C: p= 0.73; δ15N: p= 0.51). Thus all male and female beetle samples were 13 15 pooled for analysis. Site had an impact on both δ C and δ N gut content signatures (F(2,144) 13 =187, p<0.001; F(2, 144) =650, p<0.001) (Figure 1a &b). Specifically, δ C was different between all three sites, while δ15N was only different between the Marion and Benton sites, and Linn and Benton sites, but not between the Marion and Linn sites.

At the individual sites, the δ15N gut content signatures did not differ significantly across sampling periods (Benton: Kruskal-Wallace H(2) = 4, p= 0.14; Linn: F(2, 45) = 0.89, p= 13 0.42; Marion: F(2,49) =2, p= 0.13), but δ C signatures varied between dates at both the Benton 13 and Linn sites (Benton: H(2) = 17, p<0.001 ; Linn: F(2, 49) =18, p<0.001). Specifically, δ C gut signatures differed between July, and August/September, but not between August and September at the Benton site. At the Linn site δ13C signatures differed between July and 28

August/September, but not between August and September. At the Marion site there was no 13 difference in the δ C signature of foregut contents between the three sampling dates (F(2, 49) = 1, p= 0.33).

MPRR analysis did not detect a difference in the δ13C and δ15N values of foregut content and potential prey items including earth worms and various arthropods including small ground beetles (SGB), spiders and centipedes, at the Benton site (p > 0.05), though the matches varied between sample dates (Table 5). Over all sample dates at the Linn site and the July sample date at the Marion site, the δ13C and δ15N values of SGB were not different from foregut contents. Based on these matches, we consider the above arthropods and worms as prey items, but we also include the non-matches in the stable isotope models because they have been confirmed as potential prey items in other studies, and they were readily available at the current study sites.

Stable isotope mixing models

IsoSource stable isotope mixing models determined that soil dwelling arthropods and worms were likely the largest contributors in dietary prey combinations that were consistent with the δ13C and δ15N signatures of gut content of P. melanarius at our sites (Table 6). The number of potential diet combinations was variable between sample dates and sites.

At the Benton County site IsoSource identified 23,870, 46 and 69 possible dietary prey combinations respectively in the July, August and September collection periods. In July centipedes were the largest possible contributor (47-78%), followed by spiders (0-42%), C3 weed seeds (0-28%) and soil larvae (0-26%). In August, worms were the largest contributor (54-72%) followed by soil larvae (28-44%). SGB (44-68%), soil larvae (32-36%) and worms (0-21%) where major contributors in September. At the Linn County site IsoSource identified 290, 1,990 and 100 possible dietary prey combinations for July, August and September. At this site SGB were the largest dietary contributor, with contributions ranging from 65% to 100% for all sampling periods. In July, spiders (15-46%) also were identified as a large potential dietary component. At the Marion county site, IsoSource determined 4, 17 and 145,975 potential dietary prey combinations. SGB were the largest potential contributor for 29

both July (88-89%) and August (91-95%). All of the prey items tested in September were identified as possible diet contributors. Worms (22-50%) had the most defined range, and GDB (0-69%), ant larvae (0-46%), soil larvae (0-36%), C4 (0-37%) and C3 (0-25%) weed seeds also were potentially important.

Larval diet

MPRR analyses did not indicate any prey items that matched the signature of the larvae; however IsoSource models identified 53 potential dietary prey combinations that were consistent with the δ13C and δ15N signatures of the larvae. In those combinations C4 weed seeds (66-72%) were the largest potential contributor followed by worms (22-30%).

Table 1. Weed seeds germinated from surface-sampled soil at each site. Weeds are listed by family, and species when possible.

Benton Linn Marion Family: Amaranthaceae Amaranthus powellii x Family: Asteraceae Senecio vulgaris x x Sonchus sp. x Family: Brassicaceae Capsella bursa-pastoris x Cardamine oligosperma x Family: Caryophyllaceae Stellaria media x x Family: Chenopodiaceae Chenopodium album x Family: Fabaceae Lotus sp. x Trifolium dubium x Trifolium incarnatum x Family: Geraniaceae Geranium sp. x Family: Lamiaceae Lamium purpureum x Family: Poaceae Panicum miliaceum x unknown sp. x x Family: Polygonaceae Polygonum aviculare x Rumex acetosella x Family: Portulacaceae Portulaca oleracea x Family: Scrophulariaceae Veronica persica x x Family: Solanaceae Solanum sp. x Total 15 3 5

Table 2. Arthropod prey items trapped at each site listed by order, family and species when possible.

Benton Linn Marion Order: Coleoptera Carabidae Agonum sp. x x x Amara sp. x x x Anisodactylus sp. x x Bradycellus congener x x x Clivina fossor x x x Harpalus affinis x x x H. pensylvanicus x x x Nebria brevicollis x x Trechus sp. x Curculionidae- Entiminae x Chrysomelidae Diabrotica undecimpunctata x Unknown x Staphylinidae x x x Order: Araneae Linyphiidae x x x Lycosidae x x x Thomisidae x x Order: Chilopoda x x x Order: Collembola x x x Order: Diplopoda x x Order: Isopoda Armadillidium x x Order: Opiliones x x Order: Dermaptera x Order: Hymenoptera Formicidae x Unknown x Order: Diptera (larvae) x x x Order: Heteroptera x Order: Orthoptera Gryllidae x Order: Lepidoptera (larvae) x Total 24 17 14

Table 3. Percent of beetles examined containing identifiable parts from food categories including arthropods, earth worms, seeds, other plant material, or having completely empty foreguts.

Site Benton Linn Marion Sample July August September July August September July August September Period ------% Obs. ------n= 23 n= 21 n= 17 n= 20 n= 20 n= 19 n= 12 n= 14 n= 20 Arthropod 83 67 76 65 70 42 75 71 55 Worm 22 24 18 45 25 42 0 21 40 Seed 4 14 6 0 0 11 0 0 0 Plant 4 10 0 0 0 5 0 21 5 Empty 13 0 6 20 15 11 25 7 0

Table 4. Pearson‘s correlation coefficients and associated p values for combinations of items found in the foregut, and combinations of materials identified in the foreguts of beetles, and site characteristics. Correlation coefficients in bold indicate significant associations (n= 9, p ≤ 0.05). Site characteristics: Available worm biomass= worm biomass in the top 25cm of soil /m2; Seed density= number of germinable seeds in the top 1cm of soil /m2; ground beetle AD= activity per trap night for small ground beetles; arthropod richness= total number of arthropod species per site; arthropod diversity= arthropod diversity as measured by the Shannon-Wiener index; air temperature= mean air temperature at the site (there was no temperature data available for the Marion site, n=6 for this variable).

Foregut contents Arthropod Worm Seed Empty Arthropod - r= -0.70 r= -0.31 r= 0.25 - p= 0.04 p= 0.43 p= 0.51 Worm r= -0.70 - r= 0.12 r= -0.29 p= 0.04 - p= 0.76 p= 0.45 Site characteristics

Available worm biomass r= -0.80 r= 0.75 r= -0.15 r= 0.06 p= 0.01 p= 0.02 p= 0.68 p= 0.88 Seed density r= 0.49 r= -0.24 r= 0.58 r= -0.38 p= 0.18 p= 0.53 p= 0.10 p= 0.31 Ground beetle AD r= 0.61 r= -0.48 r= 0.20 r= 0.16 p= 0.08 p= 0.20 p= 0.60 p= 0.67 Arthropod richness r= 0.69 r= -0.59 r= 0.34 r= 0.11 p= 0.04 p= 0.10 p= 0.37 p= 0.77 Arthropod diversity r= 0.66 r= -0.40 r= 0.44 r= -0.16 p= 0.05 p= 0.28 p= 0.23 p= 0.68 Air temperature r= 0.23 r= 0.30 r= -0.72 r= 0.64 p= 0.66 p= 0.57 p= 0.11 p= 0.17

Table 5. Reported P-values from MPRR comparisons. P-values are corrected using the Bonferroni procedure to reflect the number of pairwise comparisons. P- values significant at the 0.05 level for the Benton site are 0.006, 0.007 and 0.007 respectively for July, August and September; 0.01 for all sample dates at the Linn site; and 0.008, 0.006 and 0.001 at the Marion site. Bolded p-values are not significant, indicating there is no evidence that the prey items δ13C and δ15N isotope signature is different from that of the beetle foregut contents. The prey items associated with bolded values are very likely to be included in the beetle‘s diet. Prey type: worm= all earthworms sampled at each site; beetle= all small ground beetles combined (species included may vary between sites); spiders= all spider species combined; centipedes= one common species of centipede; soil larvae= all soil larvae collected, likely representing Diabrotica sp.; cutworm= one species of Noctuidae; Cranefly= one or more species of Tupulidae larvae; ant larvae= one species of Formicidae; C3 seeds= seeds from weeds that employ C3 photosynthesis; C4 seeds= seeds from weeds (mainly Poaceae) that employ C4 photosynthesis.

Benton Linn Marion

Prey type July August September July August September July August September

Worm 0.0031 0.0366 0.3053 0.0008 0.0000 0.0000 0.0028 0.0000 0.0015

Beetle 0.0098 0.0004 0.0488 0.0278 0.0239 0.1403 0.0249 0.0021 0.0009

Spider 0.0362 0.0011 - 0.0000 0.0000 - 0.0005 0.0000 -

Centipede 0.0974 ------

Soil larvae 0.0008 0.0000 0.0040 - - - - - 0.0012

Cutworm - - 0.0048 ------

Cranefly ------0.0006 -

Ant larvae ------0.0000 0.0000

C3 0.0000 0.0000 0.0003 - - - 0.0001 0.0000 0.0000

C4 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 35

Table 6. Estimated ranges of prey contribution to the diet of P. melanarius by site and sample period based on IsoSource stable isotope mixing model for partitioning excess sources. Bolded values indicate the largest estimated prey contribution to the diet. Prey type: worm= all earthworms sampled at each site; beetle= all small ground beetles combined (species included may vary between sites); spiders= all spider species combined; centipedes= one common species of centipede; soil larvae= all soil larvae collected, likely representing Diabrotica sp.; cutworm= one species of Noctuidae; Cranefly= Tupulidae sp. larvae; ant larvae= one species in the Formicidae family; C3 seeds= seeds from weeds that employ C3 photosynthesis; C4 seeds= seeds from weeds (mainly Poaceae) that employ C4 photosynthesis.

Benton Linn Marion July August September July August September July August September Possible combinations 23,870 46 69 290 1,990 100 4 17 145,975 Tolerance 1 1.6 1 1.6 1.9 1.7 1.4 1.3 0.14 Prey type ------% diet contribution ------Worm 0-8 54-72 0-21 0-3 0-15 0-9 0-2 0-3 22-50 Beetle 0-14 0-1 44-68 53-84 65-100 82-100 88-89 91-95 0-69 Spider 0-42 0-3 - 15-46 0-3 - 1-12 2-9 - Centipede 47-78 ------Soil larvae 0-26 28-44 32-36 - - - - 0-2 0-36 Cutworm - - 0 ------Cranefly ------0-1 - Ant larvae ------0-1 0-46 C3 seeds 0-28 0 0 0 0-2 0-25 C4 seeds 0-1 0-1 0 0-8 0-12 0-18 0 0-1 0-37

Linn Marion

Benton

Frequency

δ13C ‰

Figure 1a. Frequency distribution of the δ13C signatures of P. melanarius foregut contents at Benton (blue), Linn (green) and Marion County (red) research sites. 37

Marion Benton

Linn

Frequency

δ15N ‰

Figure 1b. Frequency distribution of the δ15N signatures of P. melanarius foregut contents at Benton (blue), Linn (green) and Marion County (red) research sites.

18 16 14 12 10 8 6 4 Benton: July 2 0

-35 -30 -25 -20 -15 -10 -5 0 18

16 14

12 15

N (‰) 10

8 6 4 Benton: August 2

0 -35 -30 -25 -20 -15 -10 -5 0 18 16 14 12 10 8 6 4 Benton: September 2 0 -35 -30 -25 -20 -15 -10 -5 0

δ13C (‰) Figure 2a. Mean isotopic composition and standard error for potential prey sources and gut contents of P. melanarius at Benton site in July, August and September. Prey items include earthworms, beetles (small ground beetles), spiders, centipedes, soil larvae, C3 weed seeds and C4 weed seeds. δ13C (‰) 39

18 Linn: July 16 14 12 10 8 6 4 2 0 -35 -30 -25 -20 -15 -10 -5 0 18 Linn: August 16

δ 12 15

N (‰) 10

8 6 4 2

0 -35 -30 -25 -20 -15 -10 -5 0 18

Linn: September 16 14 12 10 8 6 4 2 0 -35 -30 -25 -20 -15 -10 -5 0

δ13C (‰) Figure 2b. Mean isotopic composition and standard error for potential prey sources and gut contents of P. melanarius at Linn site in July, August and September. Prey items include earth worms, beetles (small ground beetles), spiders and C4 seeds.

18 Marion July 16 14 12 10 8 6 4 2 0

18 Marion: August 16 14

12 15

N (‰) 10

6 4 2

0 -35 -30 -25 -20 -15 -10 -5 0 18 Marion September 16 14

12 10 8

6 4 2

0 -35 -30 -25 -20 -15 -10 -5 0 -2 13 δ C (‰)

Figure 2c. Mean isotopic composition and standard error for potential prey sources and gut contents of P. melanarius at Marion site in July, August and September. Prey items include earthworms, beetles (small ground beetles), spiders, soil larvae, ant larvae, craneflies, C3 weed seeds and C4 weed seeds.

Using stable isotopes and visual gut examination to determine the diet composition of Pterostichus melanarius (Coleoptera: Carabidae) in western Oregon vegetable row crops

Chapter 5

Discussion 42

I analyzed tissue from the foregut along with its contents to determine the diet of P. melanarius, and whether weed seeds played a significant role in its diet. It is more common for researchers to test δ13C and δ15N values of other body tissues such as elytra or hind-wings to determine the long term integrated diet of beetles, and to rely on more metabolically active tissues such as blood and reproductive organs in mature insects, or chitin in developing insects for recent diet trends (Webb et al., 1998, Gratton & Forbes, 2006). I was most interested in determining the immediate diet composition, and dietary changes during the three months of the adult beetle‘s most active period in the vegetable fields of western Oregon. I assumed the δ13C and δ15N values of the metabolically active foregut tissue would closely reflect those of the recent diet, and that analyzing the foregut and its contents would best reflect food choices at a resolution that could be examined in relation to changing prey availability.

I did not encounter any significant differences in the δ13C and δ15N signatures of the male and female diets. Others have found that male and female diets are similar, but there is evidence that female beetles were less likely to have empty crops, and may consume higher volumes of food than males. It is unclear whether the higher volume is due to the larger body size of the female beetles (Paill, 2004). Additionally, I found that there were no consistent differences in diet composition between the three sites based on gut dissection, MPRR and stable isotope analysis. There were overall differences in δ13C and δ15N signatures between sites—in particular I found that isotopic signatures at the Benton site were different than those at the Linn and Marion sites. Because items that were included in the diet did not differ between sites, I suggest that isotopic differences are related to as yet undetermined site characteristics such as differences in farming practices and inputs used between the farms (Ponsard & Arditi, 2000; Tooker & Hanks, 2004 ).

IsoSource models resulted in extreme numbers of possible diet composition estimates for at least one sample period at each of the sites, and the models often relied on varying tolerance levels in order to obtain feasible combinations for diet contribution proportion estimates (Table 5). Uncertainty in the data that led to these issues could have resulted from missing prey sources, or a combination of small sample sizes and similarities of δ13C and δ15N signatures between many of the prey items (Figures 2a-c). For these reasons, the exact breakdown of the diet contributions of individual prey groups is somewhat vague, as indicated by the broad ranges of possible contributions assessed by IsoSource (Table 5). Isotope mixing 43

models generally work best with either isotopically distinct source (prey) items, or with more similar sources that have a small standard error owing to large sample sizes. Sensitivity analyses conducted by Phillips and Gregg (2001) estimated that a doubling of the stable isotopic distance between sources (e.g. from 2‰ to 4‰ ) reduced uncertainty in resulting diet contribution proportion estimates by half. Since it is unrealistic to broaden existing differences between sources, increasing sample size is the best way to reduce uncertainty in diet contribution models. Even so, the higher contribution, noted by IsoSource, of worms and soil- dwelling arthropods including SGB, centipedes and spiders is consistent with their isotopic similarity with the beetle foregut contents (Table 4 and Figure 2a-c). Based on the consistency between sites and sample dates and corroboration between IsoSource models and MPRR analyses, some combination of these prey groups must constitute a major portion of the diet of P. melanarius.

Stable isotope models show that the δ15N signature of the foregut is slightly, but consistently, enriched over the signature of all potential prey items that were tested (Figure 2a- c). The enrichment likely indicates that there was a food source that was not accounted for by prey sampling, or that the foregut tissue that included in the gut content samples (and associated isotope fractionation) had more of an impact on the overall signature than anticipated (Philips, 2001). In some cases, consumers have similar δ13C signatures to their food sources, and have δ15N signatures that are enriched between 1 and 3‰ relative to the diet (McCutchan et al., 2003). If fractionation of δ14N /δ15N isotopes in the foregut tissue resulted in enrichment that wasn‘t accounted for, we might expect the δ15N signature of the foregut to be elevated over the signatures of the prey items.

One potential food item that sampling efforts may have underestimated was slugs. There was little evidence of slug damage on crops at the Marion and Benton sites, but there was a significant enough slug population during the spring to cause damage, resulting in 30- 40% yield loss in the corn crop at the Linn site (Hendricks 2011, personal communication, July 2011). Even with evidence of heavy damage at this site, only a handful of slugs were trapped between the three sites over the duration of the sampling periods. Low catches may be attributed to factors including several applications of the slug killer, Blue Bombshell (4% Metaldehyde) at the Linn site before sampling began, midsummer aestivation or inadequate sampling effort. Slugs often go through a period of dormancy during the driest months, with 44

some estimates indicating that as little as 5% of the slug population may be active during midsummer (IPPC, 2005). During dormancy they often travel deep into soil cracks to aestivate until rains return in the fall, and indeed some dehydrated slugs were discovered while digging 25 cm soil samples for worm extraction technique validation at the Benton site.

Investigators in Western Europe focus extensively on slugs and slug eggs being important items in the diet of P. melanarius. Studies have found that 11-80% of P. melanarius test positive for slug remains, and that the beetles often aggregate in areas with higher densities of slugs (Symondson et al., 2002; Bohan et al., 2000). The few slugs we did capture were all identified as Deroceras reticulatum (Müller). One study indicated that P. melanarius shows preference for eating the eggs of this species, preferring them over alternative prey including dead crickets, aphids, fly larvae and the eggs of another common slug, although they did not evaluate whether the beetle would prefer adult slugs in the presence of other prey sources (Oberholzer & Frank, 2003). In lab trials, we also found that P. melanarius consumed slug eggs enthusiastically (Marshall, unpublished data). In similar studies the closely related P. madidus preferred earthworms over D. reticulatum when given the choice (Mair & Port, 2001). In favorable conditions, this species of slug is capable of producing two overlapping generations per year, with one laying eggs in late spring, and the other in autumn (South, 1989). Given the above information, it is likely that the common field slug, D. reticulatum, or the eggs of this species could be a missing prey source.

Our findings, that P. melanarius eats a variety of arthropod and other foods, is not unusual. The Pterostichus genus is marked by omnivory and a preference for ―animal‖ food (Forsythe, 1982). In perhaps the most systematic and thorough diet study, Pollet and Desender (1985) identified parts of up to 49 different food items, mainly various arthropods and worms, that were consumed P. melanarius. Food sources besides slugs and other crop pests are frequently referred to as, ―alternative‖ prey sources by more recent investigators. Symondson (et al., 2006 ) found that female beetles that were offered a variety of prey (in this case up to three species of Diptera larvae and two species of earthworms) had increased fitness and egg load over those that were just offered the slug species D. reticulatum.

The alternative food source that we identified most often using MPRR and stable isotope mixing models as a large diet contributor was ―small ground beetles‖ (SGB). At our 45

sites this category included Agonum, Amara, Anisodactylus, Bradycellus, Clivina and Trechus species, though we only specifically compared the isotopic signatures of Clivinia, Bradycellus and Trechus species. In addition to consuming SGB, we frequently observed evidence that P. melanarius consumed the larger Harpalus sp., as noted by piles of orange Harpalus pensylvanicus legs in pitfall traps that contained both species. These findings correspond with observations made by Sunderland (1975), who noted that P. melanarius was a ―catholic feeder,‖ with its main food source in the cereal fields that he studied being other adult Coleoptera. Mitchell (1963) also observed that P. melanarius readily consumed small carabids including Bembidion lampros and Trechus quadristriatus. He attributed a sharp decline in T. quadristriaus in mid July to the emergence of this species, and noted that pitfall trapping for smaller carabids was nearly impossible when P. melanarius were active because of heavy intraguild predation in the traps. In their work examining predation of cabbage maggots by ground beetles, Prasad and Snyder (2004) noted that the addition of P. melanarius to experimental enclosures reduced the predation rates of small ground beetles on the maggots. While P. melanarius did not directly impact the cabbage maggot population, they hypothesized that P. melanarius were reducing the activity density of small ground beetles through intraguild predation.

Other alternative arthropod prey sources identified included spiders, centipedes and soil larvae. It is well documented that P. melanarius consumes the larvae of various Dipteran, Lepidopteran and Coleopteran species (Sunderland, 1975; Pollet & Desender, 1985; Wallace, 2004; Zaidi et al., 1999), however it has been demonstrated that adults will consume food placed on the soil surface more often than when it is buried. In one study examining consumption of black vine weevil (Otiorynchus sulcatus (Fabricius)) larvae, P. melanarius consumed larvae 85% of the time when they were offered on the soil surface, 43% of the time when they were buried 1.3 cm below the surface and 18% of the time when they were buried 5 cm below (Lee & Edwards, unpublished). The soil larvae encountered at our research sites were likely various larval instars and pupae of the spotted cucumber beetle Diabrotica undecimpunctata. They were found under the soil, but were often located within ~1 cm of the soil surface (Moulton, personal observation), and may have been within range of foraging beetles. Spiders have also been noted as common alternative prey, though few studies have documented it. In one example, Davey et al. (2006) found that >40% of the beetles collected tested positive for spider DNA. 46

The consumption of earthworms by P. melanarius has been extensively documented. In one case, 36% of beetles that were tested contained the remains of earthworms (Symondson et al., 2000; Pollet & Desender, 1985), and we observed beetles readily attacking and consuming worms in lab feeding trials. There is evidence that P. melanarius consumes some prey items, like slugs, based on availability (Paill, 2004; Symondson et al., 2002; Symondson et al., 1996), but we cannot confirm the same is true for earthworms. Considering the extremely high levels of worm biomass present at the Linn site, we were surprised to find that worms were not identified as an important food source more frequently there or at the other sites. In fact, we found that worms remains were identified with the lowest frequency in gut dissection trials at the Linn site (Table 2). The authors who found that 36% of the beetles tested positive for worm remains also noted that the amount of earthworm proteins in the beetle foreguts was negatively related to total foregut biomass, suggesting that earthworm consumption increased as overall prey availability declined. They further suggested that beetles were more likely to feed on earthworms in the absence of preferred prey (Symondson et al., 2000).

Many studies have focused on some aspect of the diet of P. melanarius, but few focus on their impact on the seed bank. We found little convincing evidence that adult P. melanarius consume seeds in large enough quantities to be detected at our sites, except in the case of a limited number of starch observations in foregut contents. Using IsoSource, both C3 and C4 seeds were identified as accounting for a maximum of 8-37% of the adult dietary signature in some sampling periods at each of the sites, however there was no consistent pattern, and the ranges of possible contribution assessed contained 0, without exception (Table 5). We have observed adults consuming various weed seeds in the lab, and documented them stashing seeds in burrows and under objects on the soil surface in the field (personal observation; Peachey, unpublished data). Further, analysis of digestive enzymes confirms that the presence of cellulase (used to break down cellulose) and amylase (used to break down starch) activity in the beetles gut is higher than expected for an obligate carnivore (Jaspar-Versali 1987).

Few studies have put detailed emphasis on the diet of the P. melanarius larvae. Our diet sampling in this regard was limited by small sample size, and the inclusion of individuals from only one site. Based on this limited sample, IsoSource modeling suggested that C4 weed seeds may account for between 66-72% of the isotopic signature of 3rd instar larvae. It is 47

impossible to draw conclusions, however there are some possible explanations. P. melanarius spend more than 90% of their developmental period from egg through pupal stages as active larvae (Aukema et al., 1996 as cited in Luff 2003). Their high δ13C signature could indicate that the weed seed stashing behavior of adults benefits their offspring by leaving a supply of food to consume as the larvae go through this long development over the winter. Some have noted that P. melanarius larvae have low mobility for at least part of their larval development (Lovi & Sunderland, 1996), and having a ready supply of food could facilitate growth and survival during this period. However, others classify the larvae of P. melanarius as ―soil pore explorers‖ that move about through soil pores and crevices searching for food (Zetto Brandmayr et al., 1998, as cited in Luff 2003). The location where we collected our larvae samples was planted in sweet corn the previous growing season, and all of the larvae we encountered were found in burrows under the decaying roots of the remaining corn stubble. An increase in δ13C signature could be a byproduct of the presence of the corn alone. However during the previous season we also frequently encountered corn root worms, the larvae of Diabrotica sp., which are known to overwinter in the soil of vegetable fields. In analyzing the δ13C and δ15N signatures of prey sources we found that the δ13C signature of these corn root worms was indistinguishable from the carbon signature of C4 seeds (Kruskal-Wallace: H(7)= 7, p= 0.43). In addition to the possibility that the δ13C signature could be explained by consumption of stashed seeds or site effects of the corn crop, it is also feasible that the P. melanarius encounter overwintering corn root worm larvae as they ―explore‖ soil pores. Regardless of the limitations of the current study, there is sufficient evidence to warrant further investigation into seed consumption by P. melanarius in both adult, and particularly larval life stages.

In comparing the use of stable isotope analysis and manual gut dissection for rapid and efficient determination of beetle diets, there are tradeoffs. Similar conclusions about general diet composition can be drawn from the two techniques, and they each require similar amounts of time spent in the lab. However, manual gut dissection is considerably less expensive and doesn‘t require access to specialized equipment including a -20°C freezer, an ultra microbalance scale, and other equipment required for combusting and analyzing samples for isotopic composition. Stable isotope analysis does have the benefit of providing more detailed conclusions about diet composition. For example, manual gut dissection allowed us to identify food items to broad groups including arthropods, worms, seeds, etc. (though it should 48

be noted that more exacting identifications can be made with considerably more effort (eg. Pollet & Desender, 1985). Stable isotope analysis enabled us to determine the possible contributions of more specific items within the broad categories. For example small differences in the δ13C and δ15N signatures of possible prey items helped separate diet contributions by arthropods like spiders, small ground beetles and soil larvae, and we were also able to quantify their possible contribution to P. melanarius’ diet. Stable isotopes are also useful for separating the contributions from plant material versus animal materials, or to separate the influence of plant materials with C3 and C4 photosynthetic pathways. Overall, making more precise identifications that can separate the diet contributions from pests and beneficial animal material (eg. Diabrotica larvae, and spiders), or quantify the contribution of plant material versus animal material is worth the extra cost of stable isotope analysis. If the sole desire is to determine the presence of seeds, manual gut dissection combined with the use of Lugo‘s iodine solution (I2KI) to identify starch is more cost effective.

References:

Bearhop, S., C. E. Adams, S. Waldrons, R. A. Fuller and H. Macleod. 2004. Determining trophic niche width: a novel approach using stable isotope analysis. Journal of Animal Ecology, 73: 1007-1012.

Bilde, T. and S. Toft. 1997. Consumption by carabid beetles of three cereal aphid species relative to other prey types. Entomophaga, 42(1/2): 21-32.

Blüthgen, N., G. Gebauer and K. Fiedler. 2003. Disentangling a rainforest food web using stable isotopes: dietary diversity in a species-rich ant community. Oecologia,137: 426–435.

Bohan, D. A., A. C. Bohan, D. M. Glen, W. O. C. Symondson, C. W. Wiltshire and L. Huges. 2000. Spatial dynamics of predation by carabid beetles on slugs. Journal of Animal Ecology, 69: 367-379.

Borror, D., C. Triplehorn and N. Johnson. 1989. An introduction to the study of insects, 6th edition. Saunders College Publishing, Fort Worth, USA. Pp. 43-46.

Boyd, N. S., E. B. Brennan and S. A. Fennimore. 2006. Stale seedbed techniques for organic vegetable production. Weed Technology, 20(4): 1052-1057.

Cardina, J., H. M. Norquay, B. R. Stinner and D. A. McCartney. 1996. Post dispersal predation of velvetleaf (Abutilon threophrasti) seeds. Weed Science, 44: 534-539.

Chen, Y., K. L. Giles, M. E. Payton and M. H. Greenstone. 2000. Identifying key cereal aphid predators by molecular gut analysis. Molecular Ecology. 9: 1887-1898.

Crawley, M. J. 1992. Seed predators and plant population dynamics. In M. Fenner (ed), Seeds: the ecology of regeneration in plant communities (167-182). CABI Publishing, Oxfordshire, UK.

Cross, T., J. Burt and D. McGrath. 1988a. Enterprise budget: Carrots, Willamette Valley region. Oregon State University Extension Publication EM 8379. Oregon State University, Corvallis, OR.

Cross, T., J. Burt and D. McGrath. 1988b. Enterprise budget: Table beets, Willamette Valley region. Oregon State University Extension Publication EM 8381. Oregon State University, Corvallis, OR.

Daugherty, M. P. and C. J. Briggs. 2007. Multiple sources of isotopic variation in a terrestrial arthropod community: challenges for disentangling food webs. Environmental Entomology, 36(4): 776-791.

DeNiro, M. J. and S. Epstein. 1978. Influence of diet on the distribution of carbon isotopes in animals. Geochimica et Cosmochimica Acta, 42(5): 495-506.

DeNiro, M. J. and S. Epstein..1981. Influence of diet on the distribution of nitrogen isotopes in animals. Geochimica et Cosmochimica Acta, 45(3): 341-351. Forsythe, T. 1982. Feeding mechanisms of certain ground beetles (Coleoptera: Carabidae). The Coleopterists Bulletin. 36(1): 26-73.

Gokool S., C. F. Curtis and D.F. Smith. 1993. Analysis of mosquito bloodmeals by DNA profiling. Medical and Veterinary Entomology, 7: 208-215.

Gratton, C. and A. Forbes. 2006. Changes in δ13C stable isotopes in multiple tissues of insect predators fed isotopically distinct prey. Oecologia, 147: 615-624.

Green, J. M. 2010. Structuring habitat to conserve ground beetles (Coleoptera: Carabidae) and reduce summer annual weeds in agroecosystems (Master‘s thesis). Oregon State University, Corvallis, OR.

Greenstone, M. H. 1996. Serological analysis of arthropod predation: past, present and future. Ch. 11 in The ecology of agricultural pests, Ed. W. O. C. Symondson and J. E. Liddell. Chapman and Hall, London.

Hagler, J. R., S. E. Naranjo. 1996. Using gut content immunoassays to evaluate predaceous biological control agents: a case study. Ch. 16 in The ecology of agricultural pests, Ed. W. O. C. Symondson and J. E. Liddell. Chapman and Hall, London.

Harper, G. L., R. A. King, C. S. Dodd, J. D. Harwood, D. M. Glen, M. W. Bruford and W. O. C. Symondson. 2005. Rapid screening of invertebrate predators for multiple prey DNA targets. Molecular Ecology, 14: 819–827.

Harwood, J. D. 2008. Are sweep net sampling and pitfall trapping compatible with molecular analysis of predation? Environmental Entomology 37(4): 990-995.

Honke, A., Z. Martinkova and P. Saska. 2005. Post-dispersal predation of Taraxicum officinale (dandelion) seed. Journal of Ecology. 93: 345-352.

Honek, A., P. Saska and Z. Martinkova. 2006. Seasonal variation in seed predation by adult carabid beetles. Entomologia Experimentalis et Applicata. 118: 157-162.

Hobson, K. A. and R. G. Clark. 1992. Assessing avian diets using stable isotopes: turnover of δ13C in tissues. The Condor, 94: 181-188.Publishing, Ltd.

Hoogendorn, M. and G. E. Heimpel. 2001. PCR-based gut content analysis of insect predators: using ribosomal ITS-1 fragments from prey to estimate predation frequency. Molecular Ecology, 10: 2059–2067. 51

IPPC. 2005. Slugs. Retrieved from http://oregonstate.edu/dept/nurspest/slugs.htm .

Jaspar-Versali, M. F. and C. Jeuniaux. 1987. The digestive enzymatic equipment of carabid beetles in relation to their ecological niche. Revue d'eìcologie et de biologie du sol, 24(4): 541-547.

Johnson, N. E. and S. R. Cameron. 1969. Phytophagous Ground Beetles. Annals of the Entomological Society of America, 62(4): 909-914.

Juen, A. and M. Traugott. 2005. Detecting predation and scavenging by DNA gut-content analysis: a case study using a soil insect predator-prey system. Oecologia, 142: 344–352.

Julian, J., C. Seavert, E. Peachey, A. Stone, D. McGrath. 2008a. Enterprise budget: Broccoli, organic, processed market, Willamette Valley region. Oregon State University Extension Publication EM 8960-E. Oregon State University, Corvallis, OR.

Julian, J., C. Seavert, E. Peachey, A. Stone, D. McGrath. 2008b. Enterprise budget: Bush beans, organic, processed market, Willamette Valley region. Oregon State University Extension Publication EM 8959-E. Oregon State University, Corvallis, OR,

Julian, J., C. Seavert, E. Peachey, A. Stone, D. McGrath. 2008c. Enterprise budget: Cauliflower, organic, processed market, Willamette Valley region. Oregon State University Extension Publication EM 8961-E. Oregon State University, Corvallis, OR.

Julian, J., C. Seavert, E. Peachey, A. Stone, D. McGrath. 2008d. Enterprise budget: Sweet corn, organic, processed market, Willamette Valley region. Oregon State University Extension Publication EM 8962-E. Oregon State University, Corvallis, OR.

Julian, J., C. Seavert, D. McGrath, R. Reynolds, E. Peachey. 2010a. Enterprise budget: Broccoli, processed market, Willamette Valley region. Oregon State University Extension Publication AEB0003. Oregon State University, Corvallis, OR.

Julian, J., C. Seavert, D. McGrath, R. Reynolds, E. Peachey. 2010b. Enterprise budget: Bush beans, processed market, Willamette Valley region. Oregon State University Extension Publication AEB0004. Oregon State University, Corvallis, OR.

Julian, J., C. Seavert, D. McGrath, R. Reynolds, E. Peachey. 2010c. Enterprise budget: Cauliflower, processed market, Willamette Valley region. Oregon State University Extension Publication AEB0005. Oregon State University, Corvallis, OR.

Julian, J., C. Seavert, D. McGrath, R. Reynolds, E. Peachey. 2010d. Enterprise budget: Sweet corn, processed market, Willamette Valley region. Oregon State University Extension Publication AEB0006. Oregon State University, Corvallis, OR. 52

Kielty, J. P., L. J. Allen-Williams, N. Underwood and E. A. Eastwood. 1996. Behavioral responses of three species of ground beetle (Coleoptera: Carabidae) to olfactory cues associated with prey and habitat. Journal of Insect Behavior, 9(2): 237-250.

King, R. A., D. S. Read, M. Traugott and W. O. C. Symondson. 2008. Molecular analysis of predation: a review of best practice for DNA-based approaches.

Kromp, B. 1999. Carabid beetles in sustainable agriculture: a review on pest control efficacy, cultivation impacts and enhancement. Agriculture, Ecosystems and Environment 74: 187–228.

Langellotto, G. A., J. A. Rosenheim and M. R. Williams. 2005. Enhanced carbon enrichment in parasitoids (Hymenoptera): A stable isotope study. Annals of the Entomological Society of America, 98(2): 205-213.

Langellotto, G. A., J. A. Rosenheim and M. R. Williams. 2006. Assessing trophic interactions in a guild of primary parasitoids and facultative hyperparasitoids: stable isotope analysis. Oecologia, 150: 291-299.

Larochelle, A., and M. C. Lariviere. 2003. A natural history of the ground beetles (Coleoptera: Carabidae) of America north of Mexico. Pensoft, SoÞa, Bulgaria.

Larsson, S. G. 1939. Entwicklungstypen und entwicklungszeiten der danischen Carabiden. Entomologiske Meddelelser 20: 277-560.

Lawrence, A. Paulson, M. A. Bowers. 2002. A test of the ‗hot‘ mustard extraction method of sampling earthworms. Soil Biology and Biochemistry, 34: 549-552.

Lee, J. and D. L. Edwards. 2011. Impact of predatory carabids on below- and above-ground pests and yield in strawberry. Manuscript submitted for publication in Biocontrol.

Liebman, M. 2001. Weed management: a need for ecological approaches. In M. Liebman, C. L. Mohler and C. P. Staver (eds.), Ecological Management of Agricultural Weeds (pp. 1-30). Cambridge University Press, Cambridge, UK.

Lövei, G. L., Sarospataki, M., 1990. Carabids in agricultural fields in Eastern Europe. In: Stork, N. E. (Ed.), The Role of Ground Beetles in Ecological and Environmental Research. Intercept Ltd, Hanover, UK, pp. 87–93.

Luff, M. L. 2002. Carabid assemblage organization and species composition. In J. M. Holland (Ed.), Agroecology of Carabid Beetles (pp. 41-79). Andover, UK: Intercept Limited.

Luff, M. L. 2003. Biology and ecology of immature stages of ground beetles (Carabidae). Proceedings of the 11th European Carabidologists Meeting. European Carabidology. Pp. 183- 208.

Lys, J. A. and W. Nentwig. 1994. Improvement of the overwintering sites for Carabidae, Staphylinidae and Araneae by strip-management in a cereal field. Pedobiologia. 38: 238-242.

Mair J. and G. R. Port. 2001. Predation on the slug Deroceras reticulatum by the carabid beetles Pterostichus madidus and Nebria brevicollis in the presence of alternative prey. Agricultural and Forest Entomology, 3: 169-174.

Marino, P. C., K. L. Gross, D. A. Landis. 1997. Weed seed loss due to predation in Michigan maize fields. Agriculture, Ecosystems & Environment, 66 (3): 189-196.

Marino, P. C., P. R. Westerman, C. Pinkert, W. van der Werf. 2005. Influence of seed density and aggregation on post-dispersal weed seed predation in cereal fields. Agriculture, Ecosystems and the Environment, 106: 17-25.

Marshall, Nicole. 2008. Feeding preferences of Pterostichus melanarius: lab based cafeteria feeding trials. OSU undergraduate honors research project.

Martinkova, Z., P. Saska and A. Honek. 2006. Consumption of fresh and buried seed by ground beetles (Coleoptera: Carabidae). European Journal of Entomology. 103: 361-364.

Mauchline, A. L., S. J. Watson, V. K. Brown and R. J. Froud-Williams. 2005. Post-dispersal seed predation of non-target weeds in arable crops. Weed Research, 45: 157-164.

McCune, B. and J. B. Grace. 2002. Analysis of ecological communities. Gleneden Beach , OR: MjM Software.

McCutchan, J. H., W. M. Lewis, C. Kendall and C. C. McGrath. 2003. Variation in trophic shift for stable isotope ratios of carbon, nitrogen and sulfur. Oikos, 102: 378-390.

McKemey, A. R., W. O. C. Symondson and D. M. Glen. 2003. Predation and prey size choice by the carabid beetle Pterostichus melanarius (Coleoptera: Carabidae): the dangers of extrapolating from laboratory to field. Bulletin of Entomological Research. 93: 227–234.

McKemey, A. R., W. O. C. Symondson, D. M. Glen and P. Brain. 2001. Effect of slug size on predation by Pterostichus melanarius (Coleoptera: Carabidae). Biocontrol Science and Technology, 11: 83-93.

McNabb, D. M., J. Halaj and D. H. Wise. 2001. Inferring trophic positions of generalist predators and their linkage to the detrital food web in agroecosystems: a stable isotope analysis. Pedobiologica, 45: 289-297.

Melander, B. and G. Rasmussen. 2001. Effects of cultural methods and physical weed control on intrarow weed numbers, manual weeding and marketable yield in direct-sown leek and bulb onion. Weed Research, 41: 491–508.

Menalled, F. D., P. C. Marino, K. A. Renner, D. A. Landis. 1999. Post-dispersal weed seed predation in Michigan crop fields as a function of agricultural landscape structure. Agriculture, ecosystems and environment. 77: 193-202.

Mitchell, B. 1963. Ecology of two carabid beetles, Bembidion lampros (Herbst) and Trechus quadristriatus (Schrank). Journal of Animal Ecology, 32(3): 377-392.

Mohler, C. L. 2001. Weed life history: identifying vulnerabilities. In M. Liebman, C. L. Mohler and C. P. Staver (eds.), Ecological Management of Agricultural Weeds (pp.40-84). Cambridge University Press, Cambridge, UK.

Muramoto, J. and M. R. Werner. 2002. Mustard powder vermifuge: an alternative to formalin expulsion method for earthworm sampling. Edaphologia, 70: 7-11.

Nault, L. R. and M. E. Montgomery. 1979. Aphid alarm pheromones. Miscellaneous Publications, Entomological Society of America, 11(3): 23-31.

Newsom, S. D., C. Martinez del Rio, S. Bearhop and D. Phillips. 2007. A niche for isotopic ecology. Frontiers in ecology and the environment, 5(8): 429-436.

Oberholzer F. and T. Frank. 2003. Predation by the carabid beetles Pterostichus melanarius and Poecilus cupreus on slugs and slug eggs. Biocontrol Science and Technology, 13: 99-110.

Orrock, J.L., B.J. Danielson, M. J. Burns and D.J. Levey. 2003. Spatial ecology of predator- prey interactions: corridors and patch shape influence seed predation. Ecology, 84(10): 2589- 2599.

Ostrom, P. H., · M. Colunga-Garcia, and S. H. Gage. 1997. Establishing pathways of energy flow for insect predators using stable isotope ratios: field and laboratory evidence. Oecologia, 109: 108–113.

Paill, W. 2004. Slug feeding in the carabid beetle Pterostichus melanarius: seasonality and dependence on prey size. Journal of Molluscan Studies 70(2): 203-205.

Painter, M. L., C. L. Chambers, M. Siders, R. R. Doucett, J. O. Witaker Jr., and D. L. Phillips. 2009. Diet of spotted bats (Euderma maculatum) in Arizona as indicated by fecal analysis and stable isotopes.Canadian Journal of Zoology, 87: 865-875.

Peachey, R. E., J. N. Pinkerton, K. L. Ivors, M. L. Miller and L.W. Moore. 2001. Effect of soil solarization, cover crops, and Metham on field emergence and survival of buried annual bluegrass (Poa annua) Seeds. Weed Technology, 15(1): 81-88.

Pollet, M. and Desender K. 1985. Adult and larval feeding ecology in Pterostichus melanarius Ill. (Coleoptera, Carabidae). Medelingen van de Faculteit Landbouwwetenschappen Rijksuniversiteit Gent, 50: 581-594.

Pollet, M. and K. Desender. 1990. Investigating the food passage in Pterostichus melanarius (Coleoptera:Carabidae): An attempt to explain its feeding behavior. Medelingen van de Faculteit Landbouwwetenschappen Rijksuniversiteit Gent, 55(2b): 527-540.

Ponsard, S. and R. Arditi. 2000. What can stable isotopes tell (δ13C and δ13N) tell us about the food web of soil macro-invertebrates? Ecology, 81(3): 852–864.

Rinehold, J. and J. J. Jenkins. 1996. Pesticide use survey: Oregon Pesticide use estimates for vegetable crops, 1993. Oregon State University Extension Publication EM8643. 128p.

South, A. 1989. A comparison of the life cycles of the slugs Deroceras reticulatum (Müller) and Arion intermedius (Normand) on permanent pasture. Journal of Molluscan Studies, 55(1): 9-22.

Stewart, M. K. and M. H. Greenstone. 1996. Serological diagnosis of parasitism: a monoclonal antibody-based immunodot assay for Microplitis croceipes (Hymenoptera: Braconidae). Ch. 12 in W. O. C. Symondson and J. E. Liddell (Eds.) The ecology of agricultural pests.. Chapman and Hall, London.

Saska, P., W. van der Werf, E. de Vries and P. R. Westerman. 2008. Spatial and temporal patterns of carabid activity-density in cereals do not explain levels of predation on weed seeds. Bulletin of Entomological Research. 98: 169-181.

Schmidt, Olaf, James P. Curry, Jens Dyckmans, Emilia Rota, Charles M. Scrimgeour. 2004. Dual stable isotope analysis (δ13C and δ13N) of soil invertebrates and their food sources. Pedobiologia. 48: 171-180.

Sheppard, S. K., and J. D. Harwood. 2005. Advances in molecular ecology: tracking trophic links through predator-prey food-webs. Functional Ecology. 19: 751-762.

Sunderland K. D. and G. P. Vickerman. 1980. Aphid feeding by some polyphagous predators in relation to aphid density in cereal fields. Journal of Applied Ecology, 17: 389-396.

Sulzman, E. W. 2007. Stable isotope chemistry and measurement: a primer. Ch. 1 in Michener, R. and K. Lajtha (Eds.) Stable isotopes in ecology and environmental science, Second edition.. Blackwell Publishing. Malden, MA. pp. 1-21.

Symondson, W. O. C. 2002. Molecular identification of prey in predator diets. Molecular Ecology, 11: 627-641.

Symondson, W. O. C., M. L. Erickson and J. E. Liddell. 1999. Development of a monoclonal antibody for the detection and quantification of predation on slugs within the Arion hortensis agg. (Mollusca: Pulmonata). Biological Control, 16: 274:282.

Symondson, W. O. C., D. M. Glen, A. R. Ives, C. J. Langdon and C. W. Wiltshire. 2002. Dynamics of the relationship between a generalist predator and slugs over five years. Ecology, 83(1): 137-147.

Symondson, W. O. C., D. M. Glen, C. W. Wiltshire, C. J. Langdon and J. E. Liddell. 1996. Effects of cultivation techniques and methods of straw disposal on predation by Pterostichus melanarius (Coleoptera: Carabidae) upon slugs (Gastropoda: Pulmonata) in an arable field. Journal of Applied Ecology, 33: 741-753.

Symondson, W. O. C., D. M. Glen, M. L. Erickson, J. E. Liddell and C. J. Langoon. 2000. Do earthworms help sustain the slug predator Pterostichus melanarius (Coleoptera: Carabidae) within crops? Investigation using monoclonal antibodies. Molecular Ecology, 9: 1279-1292.

Symondson, W. O. C., S. Cesarini, P. W. Dodd, G. L. Harper1y, M. W. Bruford, D. M. Glen, C. W. Wiltshire and J. D. Harwood. 2006. Biodiversity vs. biocontrol: positive and negative effects of alternative prey on control of slugs by carabid beetles. Bulletin of Entomological Research. 96: 637–645.

Thomas, R. S., D. M. Glen and W. O. C. Symondson. 2008. Prey detection through olfaction by the soil-dwelling larvae of the carabid predator Pterostichus melanarius Soil biology and biochemistry. 40: 207–216.

Thiele, H. U. 1977. Carabid beetles in their environments. Springer, Berlin.

Tieszen, L. L., T. W. Boutton, K. G. Tesdahl and N. A. Slade. 1983. Fractionation and turnover of stable carbon isotopes in animal tissues: implications for δC13 analysis of diet. Oecologia, 57: 32-37.

Toft S. and T. Bilde. 2002. Carabid diets and food value. In J. M. Holland (Ed.), The agroecology of carabid beetles (pp. 81-104). Andover, UK: Intercept.

Tooker, J. F. and L. M. Hanks. 2004. Trophic position of the endophytic beetle, Mordellistena aethiops Smith (Coleoptera: Mordellidae). Environmental Entomology, 33(2): 291-296.

Tooley, J. and G. E. Brust. 2002. Weed seed predation by carabid beetles. In: J. M. Holland (Ed.), The Agroecology of Carabid Beetles. Intercept Ltd., Andover, pp. 215- 230. Triltsch, H. 1997. Gut contents in field sampled adults of Coccinella septempunctata (Coleoptera: Coccinellidae). Entomophaga 42: 125-131.

USDA. 2010. Oregon Agriculture and Fisheries Statistics. United States Department of Agriculture National Agricultural Statistics Service, Oregon Field Office.

USDA NRCS web soil survey http://websoilsurvey.nrcs.usda.gov/app/WebSoilSurvey.aspx. accessed April 29, 2010.

Vander Wall, S. B., K. M. Kuhn and M. J. Beck. 2005. Seed removal, seed predation and secondary dispersal. Ecology, 86(3): 801-806. 57

Vanderklift, M. A. and S. Ponsard. 2003. Sources of variation in consumer-diet d15N enrichment: a meta-analysis. Oecologia, 136: 169-182.

Wallace, S. K. 2004. Molecular gut analysis of carabids (Coleoptera: Carabidae) using aphid primers. Master‘s Thesis. Montana State University, Department of Entomology.

Wallin, H., P. A. Chiverton, B. S. Ekbom and A. Borg. 1992. Diet, fecundity and egg size in some polyphagous predatory carabid beetles. Entomologia Experimentalis et Applicata, 65(2): 129-140.

Webb, S. C., R. E. Hedges and S. J. Simpson. 1998. Diet quality influences the δ13C and δ13N of locusts and their biochemical components. The Journal of Experimental Biology 201: 2903–2911.

Westerman, P. R., A. Hofman, L. M. Vet, W. van der Werf. 2003. Relative importance of vertebrates and invertebrates in epigeaic weed seed predation in organic cereal fields. Agriculture, ecosystems and environment, 93: 417-425.

Wheater, C. P. 1989. Prey detection by some predatory Coleoptera (Carabidae and Staphylinidae). Journal of zoology, 218(2): 171 – 185. Blackwell Publishing, Ltd. Winder, L., C. J. Alexander, J. M. Holland, W. O. C. Symondson, J. N. Perry and C. Woolley. 2005. Predatory activity and spatial pattern: the response of generalist carabids to their aphid prey Journal of Animal Ecology. 74, 443–454.

Zaidi R. H, Jaal Z, Hawkes N. J., Hemingway J., Symondson W. O. C. 1999. Can the detection of prey DNA amongst the gut contents of invertebrate predators provide a new technique for quantifying predation in the field?

Zwahlen, C., A. Hilbeck, R. Howald and W. Nentwig. 2003. Effect of transgenic Bt corn litter on the earthworm, Lumbricus terrestris. Molecular Ecology, 12: 1077-1086.