Patterns and Mechanisms of the Exploitation of Mutualisms Ellen Mcgrail Reid Louisiana State University and Agricultural and Mechanical College

Louisiana State University LSU Digital Commons

LSU Master's Theses Graduate School

2010 Patterns and mechanisms of the exploitation of mutualisms Ellen McGrail Reid Louisiana State University and Agricultural and Mechanical College

Follow this and additional works at: https://digitalcommons.lsu.edu/gradschool_theses

Recommended Citation Reid, Ellen McGrail, "Patterns and mechanisms of the exploitation of mutualisms" (2010). LSU Master's Theses. 2246. https://digitalcommons.lsu.edu/gradschool_theses/2246

This Thesis is brought to you for free and open access by the Graduate School at LSU Digital Commons. It has been accepted for inclusion in LSU Master's Theses by an authorized graduate school editor of LSU Digital Commons. For more information, please contact [email protected].

PATTERNS AND MECHANISMS OF THE EXPLOITATION OF MUTUALISMS

A Thesis

Submitted to the Graduate Faculty of the Louisiana State University and Agricultural and Mechanical College in partial fulfillment of the requirements for the degree of Master of Science

The Department of Biological Sciences

by Ellen McGrail Reid B.A. Boston University, 2005 August 2010 ACKNOWLEDGEMENTS

I would like to thank the members of my advisory committee: Kyle Harms, Jim Cronin, and Richard Stevens for their guidance and advice. I would especially like to thank Kyle Harms, my advisor, for his encouragement, patience, and assistance over the somewhat difficult course of my studies. I would like to thank my lab mates: Adriana Bravo and Jonathan Myers, two wonderful scientists and great people who were always ready to listen and help. I am indebted to the Lowy Foundation at Louisiana State University for funding my participation in a course through the Organization for Tropical Studies, Tropical Biology: An Ecological Approach in 2008. My work during that course provided inspiration for this thesis. I would also like to thank the Organization for Tropical Studies for funding my proposal for the grant that allowed me to carry out much of this work in Palo Verde National Park in Costa Rica. I would not have been able to complete my thesis without the help and support of several friends and colleagues: Joe Sapp, Susan Whitehead, Andre Kessler, Katja Poveda, Amanda Accamando, Natalia Aristizabal, Jessie Deichmann, Meche Gavilanez, Noah Reid, Maria Sagot, and many others. I owe many thanks to Prissy Milligan for helping me navigate the logistics of finishing my thesis. I thank Dr. C. Kelly Swing, the inspirational teacher, biologist, and conservationist who first introduced me to fieldwork in the tropics. Thanks to my parents, Robert Reid and Kathleen McGrail, who had the confidence that I could chart my path and supported me along my way. Finally, I would like to thank David Johnson for his encouragement and support.

ii TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...... ii

ABSTRACT...... iv

CHAPTER 1: INTRODUCTION ...... 1 CHAPTER 2: EXPLOITATION OF AN ANT-ACACIA MUTUALISM BY MOZENA SP. (HEMIPTERA: COREIDAE)...... 3 INTRODUCTION ...... 3 METHODS...... 5 Study site and collection of study organisms ...... 5 Specificity to ant colony of origin...... 5 Species-specificity to Vachellia host...... 5 Species-specificity to ant host species...... 6 Ant response to glandular compounds ...... 6 Ant response to lanolin-dissolved cuticular and glandular compounds...... 6 Removal of coreid cuticular hydrocarbons ...... 7 Ant response to methyl salicylate...... 7 Chemical analyses ...... 7 RESULTS...... 8 Specificity to ant colony of origin...... 8 Species-specificity to Vachellia host...... 8 Species-specificity to ant host species...... 8 Ant response to glandular compounds ...... 8 Ant response to lanolin-dissolved cuticular and glandular compounds...... 11 Removal of coreid cuticular hydrocarbons ...... 12 Ant response to methyl salicylate...... 12 Chemical analyses ...... 12 DISCUSSION...... 13 Species-specificity to ant- and plant-hosts ...... 13 The role of coreid glandular and cuticular compounds...... 14 The role of methyl salicylate in ant-coreid interactions...... 15 Chemical analyses ...... 15 Conclusions ...... 15

CHAPTER 3: THE COSTS OF EXPLOITATION OF MUTUALISMS ...... 17 INTRODUCTION ...... 17 METHODS...... 18 RESULTS...... 19 DISCUSSION...... 22

CHAPTER 4: CONCLUSIONS ...... 24

LITERATURE CITED ...... 25

VITA ...... 31

iii ABSTRACT

Mutualisms are reciprocally exploitative interactions providing net benefits to both partners. These interactions can be exploited, in turn, by individuals that take advantage of benefits offered by one or both partners in a mutualism, while offering no benefits in return. For many mutualist-exploiter interactions the mechanisms allowing exploitation, and the maintenance of mutualisms in the face of exploitation, are still poorly understood. Here I describe manipulative field and laboratory experiments to investigate the mechanisms used by an exploiter to invade an ant-plant mutualism. I tested two non-mutually exclusive hypotheses for how a coreid (Mozena sp., Hemiptera: Coreidae) feeds on mymecophytic acacia trees (Vachellia spp.) while avoiding attack by resident ants: chemical defense and chemical mimicry. I found that chemical compounds produced by Mozena sp. in both the metathoracic gland and the cuticle reduced the number of ant attacks and cuticular compounds appeared to be essential in escaping recognition on ant-occupied Vachellia spp. trees. The compounds were effective on multiple colonies and for multiple ant species, thus they are not strictly host- or species-specific. In addition, gas chromatography and mass spectrometry analyses of cuticular compounds revealed a close match between chemical profiles of Mozena sp. and Pseudomyrmex spinicola ants, suggesting chemical mimicry is the primary mechanism by which Mozena sp. exploits the ant- acacia mutualism. To examine the prevalence of a cost of exploitation for plant partners in exploited mutualisms, I conducted a meta-analysis of studies from the published literature. I found that exploitation has a weak, negative, but insignificant impact on the reproductive success of mutualistic plants. Collectively, these analyses illuminated methods by which exploiters may succeed in infiltrating mutualisms and suggested that the relatively low costs of exploitation may account for the lack of destabilization and degradation by exploiters of some mutualistic interactions.

iv CHAPTER 1

INTRODUCTION

Mutualisms are reciprocally exploitative interactions that provide net benefits to both partners based on the exchange of commodities or services (Thompson 1982; Janzen 1985; Conner 1995). Mutualisms are widespread across ecosystems and taxa and every living organism may be involved in at least one mutualistic interaction (Janzen 1985). The availability of rewards or services provides incentives for non-mutualistic species or individuals to exploit one or both partners of mutualisms without reciprocation. These organisms or individuals are referred to by a variety of terms: cheaters (Pellmyr et al. 1996); thieves (Inouye 1983); aprovechados (Soberon and Martinez del Rio 1985); robbers (Inouye 1983); parasites of mutualisms (Janzen 1975); or exploiters of mutualisms (Bronstein 2001). Like mutualists, exploiters of mutualisms have been described from a variety of taxa and habitats, and exploit a variety of commodities or services available to mutualist partners (Soberon and Martinez del Rio 1985; Bronstein 2001; Yu 2001). Examples of mutualism exploiters include bees, ants, and hummingbirds that feed on floral nectar but do not pollinate plants (Maloof and Inouye 2000), non-pollinating seed parasites of yucca and figs (West and Herre 1994; West et al. 1996; Bronstein and Ziv 1997), ants that consume plant-produced food rewards but castrate or do not defend host plants (Yu and Pierce 1998; Stanton et al. 1999), Rhizobium bacteria that do not fix nitrogen (Kiers et al. 2003), cleaner fishes that do not remove ectoparasites (Wickler 1968; Grutter and Lester 2002), and non-rewarding plants that mimic species that reward visitors (Sabat and Ackerman 1996; Moret et al. 1998). Exploiters employ a variety of strategies to invade mutualisms. Exploiters may coexist with mutualist partners through chemical mimicry or corruption of the host communication systems (Höllodobler and Wilson 1990; Akino et al 1999; Barbero 2009) or through their behavior (Inouye 1983) or morphology (Styrsky and Eubanks 2007; Reiskind 1977; Grutter and Lester 2002; Anderson 2005). However, for many mutualist-exploiter interactions these mechanisms are not known or are poorly understood. Exploitation may destabilize mutualistic relationships by leading to a net imbalance in the exchange of commodities or services between mutualist partners, rendering it more costly and less beneficial to invest in reciprocation, especially for high-cost exploitative behaviors, such as killing and consuming a partner in a mutualism (Bronstein 2001; Sakata 1994). Phylogentic evidence suggests that in some cases, exploiters have coexisted with mutualists over millions of generations (Pellmyr and Leebens-Mack 1999; Morris et al. 2003), therefore conditions favoring the long-term and short-term maintenance of mutualisms exist. Some mechanisms that may be important to the maintenance of mutualisms in the face of exploitation include sanctions or punishment for exploitation (Kiers et al. 2003; Izzo and Vasconcelos 2002), competition between mutualists and exploiters (Morris et al. 2003), spatial and temporal heterogeneity (Yu et al. 2001; Wilson et al. 2003), and low costs of exploitation (Bronstein 2001). Theory predicts that, in order for mutualisms to persist, most exploitation of mutualisms must have relatively low costs to the mutualists (Bronstein 2001). The ecological and evolutionary importance – at the individual, population, and community levels – of mutualisms and exploitation of mutualisms is recognized, but the mechanisms allowing exploitation and the maintenance of mutualisms in the face of exploitation are still poorly understood in most cases. This thesis seeks to better understand the mechanisms

1 used by an exploiter of an ant-plant mutualism and the general consequences of exploitation of mutualisms through a meta-analysis of the costs of exploitation for plant partners in exploited mutualisms.

2 CHAPTER 2

EXPLOITATION OF AN ANT-ACACIA MUTUALISM BY MOZENA SP. (HEMIPTERA: COREIDAE)

INTRODUCTION

Mutualisms, reciprocally exploitative interactions that provide net benefits to both partners (Thompson 1982; Janzen 1985; Conner 1995), are common and widespread across ecosystems and taxa. Many organisms use commodities or services provided by other species to enhance survival, growth, and reproduction, reciprocating by producing commodities or services (e.g., transportation, protection, or nutrition) available to their mutualistic partners (Noë and Hammerstein 1994; Schwartz and Hoeksema 1998). The availability of these commodities or services provides an opportunity for non-mutualists to exploit them, taking these commodities or services while providing nothing in return to either mutualist partner. Such individuals are often called exploiters of mutualisms (Bronstein 2001), parasites of mutualisms (Janzen 1975), aprovechados (Soberon and Martinez del Rio 1985), cheaters (Pellmyr et al. 1996), robbers or thieves (Inouye 1983). Ant-plant mutualisms, especially those in which ants inhabit and defend their host trees and in which trees provide resources to ants, provide an excellent opportunity for access to nutrional and defense resources for organisms that are able to exploit them. The bull thorn acacia-ant mutualism is one such interaction. Many Vachellia spp. (formerly known as Acacia spp.), bull thorn acacias, form obligate mutualisms with Pseudomyrmex spp. or (less commonly) Crematogaster spp. ants. These myrmecophytes (plants that live in association with ants) provide nectar via extrafloral nectaries to ants, Beltian bodies (nutrient-rich protein-lipid structures produced on young leaves and fed to ant larvae), and shelter in the form of domatia, in which ants rear brood (Janzen 1983). Ants defend host trees from herbivores and pathogens, remove debris, and often clear other vegetation from the area surrounding the base of a host tree (Janzen 1966). Species able to infiltrate ant-acacia mutualisms may gain benefits such as enemy-free space, reduced competition, and access to food sources. Furthermore, the leaf tissue of myrmecophytic acacia trees has lower concentrations of plant chitinases, cyanogenic glycosides, and alkaloids compared to non-myrmecophitic acacias (Heil et al 2000; Rehr et al 1973); thus leaves are relatively chemically undefended and relatively palatable to herbivores that circumvent ant defenses. Ants patrol acacia trees and remove most herbivores that are not able to withstand or circumvent their defenses and herbivores that can persist on these trees may gain enemy-free space and reduced competition for food resources. Exploiters infiltrate ant-plant mutualisms through chemical mimicry or corruption of the host communication systems (Höllodobler and Wilson 1990; Akino et al 1999; Barbero 2009) or through their behavior or morphology (Styrsky and Eubanks 2007; Reiskind 1977). Exploiters that persist within the acacia-ant mutualism have various adaptations in each of these categories that allow them to escape the aggressive behavior of the ants to survive on host trees, including morphological traits: toughened exocuticles in ruteline scarab beetles; behavioral traits: shelter building by shoot-tip-eating moth larvae (Aristotelia corallina (Gelechiidae)) and avoidance behavior of Bagheera kiplingi, a spider that feeds on Beltian bodies (Meehan et al 2009); and chemical traits: absorbance of colony odor by orioles and kiskadees that nest in acacia trees

3 (Janzen 1983; Eubanks et al 1997). Even so, few other exploiters of the acacia-ant mutualism – or the means by which they exploit it – are known. Ant societies have complex systems of communication (Höllodobler and Wilson 1990; Jackson and Ratnieks 2006) based on the exhange of semiochemicals important for nestmate recognition, initiation of defense, navigation, and to determine the outcome of intraspecific interactions (Höllodobler and Wilson 1990; Hölldobler 1999). Exploiters of the acacia-ant mutualism may use this reliance on communication systems to circumvent normal ant defenses. Corruption of chemical communication systems can be accomplished by acquiring cuticular hydrocarbons of their host through direct physical contact (trophyllaxis or allogrooming), biosynthesis, or through the diet (Dettner and Leipert 1994; Lenoir et al 2001; Elgar and Allan 2004). For some exploiters or social parasites this mode of mimicry carries the high cost of host specificity (Elgar and Allan 2004), so that the individual or species is confined living with one species of ant or even just a single colony. In general, the strategies exploiters use to circumvent colony defense are known, however, for any given interaction between exploiters and mutualists the mechanisms are often not described and are poorly understood. In a neotropical deciduous lowland forest (Palo Verde National Park, Guanacaste, Costa Rica) a hemipteran (Hemiptera: Coreidae) was found (personal observation) feeding on Vachellia spp. trees, undisturbed by resident ants. This as-yet identified Mozena sp. appears to feed only on the juices of young leaves (personal observation), destroying important resources for both trees and resident ants. Adults of this coreid feed on young Vachellia spp. (bull thorn acacia) leaves with a piercing-sucking proboscis, causing the leaflet to wilt, resulting in the loss of leaves, Beltian bodies and extra-floral nectaries. No recruitment of ant defenders occurs when these coreids are feeding or walking on Vachellia spp. trees (personal observation). If this hemipteran is encountered, ant defense is not initiated (personal observation). Behavioral and morphological mechanisms for avoiding ant defenses have been ruled out by previous research (Whitehead et al., unpublished data). Therefore, the mechanism allowing Mozena sp. to escape ant detection and attack is unknown, but assumed to be due to chemical signaling. Mozena sp. is known to produce volatile organic compounds in glands on the metathorax, which are sprayed in defense (Aldrich et al. 1982). These chemical cues, perhaps in addition to cuticular hydrocarbons which corrupt ant communication systems, may be used to invade myrmecophytic acacias and allow the successful exploitation of acacia ants and Vachellia spp. trees. I aimed to elucidate the mechanisms by which the herbivore, Mozena sp. (Hemiptera: Coreidae), is able to exploit the mutualism between Vachellia spp. trees and Pseudomyrmex spp. ants in Palo Verde National Park, Costa Rica. I used a combination of behavioral bioassays and chemical analyses. First, in a series of behavioral bioassays I determined the host specificity of Mozena sp. individuals. Next I tested two non-mutually exclusive hypotheses for how Mozena sp. feeds on V. collinsii while avoiding ant attack: chemical defense and chemical mimicry. In addition, behavioral bioassays were used to quantify ant responses to Mozena sp. using various treatments that isolated the effects of chemical cues from methathoracic glands and cuticle. Preliminary chemical analyses of defensive sprays of coreids revealed that one major chemical component was methyl salicylate, therefore ant response to this chemical was tested in the field. Finally, I used chemical analyses to compare and contrast cuticular hydrocarbons and defensive spray of coreids with the cuticular hydrocarbons and defensive spray of ants.

4 METHODS

Study site and collection of study organisms

I conducted this study in Palo Verde National Park (a neotropical deciduous lowland forest), Guanacaste Province, Costa Rica from January 27 to February 3 and from March 24 to April 3, 2008. In this forest, Vachellia collinsii (obligate myrmecophyte), V. cornigera (obligate myrmecophyte), and V. farnesiana (facultative myrmecophyte) co-occur. The trees form a mutualism with one of three ant species, in order of most to least common at the study site (personal observation): Pseudomyrmex spinicola, P. flavicornis, and Crematogaster brevispinosa. I conducted field experiments from 8 am until 12 pm, to standardize for level of ant activity, which appears to peak during the morning and decline in the afternoon (personal observation). I conducted opportunistic searches for Mozena sp. feeding on V. collinsii with colonies of P. spinicola mutualist ants in forests bordering the park road near the Organization for Tropical Studies’ Palo Verde Biological Station. I held Mozena sp. (Hemiptera: Coreidae) individuals in vials for no more than 48 hours before completion of both field and laboratory experiments.

Specificity to ant colony of origin

I collected ten adult coreids and held them in vials for one hour. Five were then translocated to new trees, each with a different colony of P. spinicola and five were replaced on their trees of origin with the original colony of P. spinicola. I placed each individual on a tree with a colony of P. spinicola, on a young leaf containing Beltian bodies. I observed and recorded ant behavior in thirty second trials, recording the number of attacks (defined as obvious stings or bites), the number of passes (defined as ants that encountered the coreid, but either passed over or around it without biting or stinging), and the number of avoidances (defined as ants that encountered the coreid, then backed away or reversed direction). The experiment was blind in that the researcher was not aware of the treatment (replacement or translocation) while scoring ant response. I arcsine square-root transformed proportions of attacks and compared ant attacks on trees of origin vs. new trees using at t-test in R (R Development Core Team 2009).

Species-specificity to Vachellia host

To determine whether Mozena sp. feed preferentially on one of the three acacia species present in Palo Verde National Park, I conducted choice trials. Twenty trials were conducted using undamaged branches with young leaves, collected opportunistically from similarly sized individuals (3-5 m tall) of three acacia species: V. collinsii, V. farnesiana, and V. cornigera, irrespective of species of ant resident. I removed ants from branches and domatia and trimmed branches to a length 30 cm. I placed one branch from each acacia species and one coreid in a mesh enclosure. Acacia branches were placed in water to prevent wilting and coreids remained in enclosures for 12 hours (overnight). At 0800 hours, I surveyed enclosures and noted the location of the coreid. Feeding on an acacia branch was scored as a preference. Coreids sitting on the mesh were considered non-participants and were not included in subsequent analyses. To determine whether Mozena sp. preferred one species of acacia, I performed a Chi- squared test. A Chi-squared test was also performed to assess preference of obligate ant-acacias

5 (V. collinsii and V. cornigera) over a facultative ant-acacia (V. farnesiana). All statistical analyses were performed in JMP (SAS Institute 2004).

Species-specificity to ant host species

To determine whether ant mutualist species respond differently to Mozena sp., I placed a coreid on three different V. collinsii trees in succession, with mutualist ant colonies of P. spinicola, P. flavicornis, and C. brevispinosa. Each of 10 coreids was tested with each ant species and the order of presentation was randomized for each coreid, for a total sample size of 10. Each ant colony was used only once. I released coreids on or near new leaves on a branch and ant response was scored for 30 seconds following release. Ant behavior was classified as stinging, biting, passing or avoiding upon encounter with the coreid. P. flavicornis and P. spinicola sting and bite, whereas C. brevispinosa bite only. Avoidance behavior was defined as ants approaching an Mozena sp. individual, but changing direction following the encounter, whereas passing was defined as ants continuing on their original path following an encounter with Mozena sp. To determine differences in ant responses to coreids I compared the proportions of all encounters that were attacks among the three ant species (P. spinicola, P. flavicornis, and C. brevispinosa) using ANOVA for the randomized complete block design, blocked on individual coreids, in R (R Development Core Team 2009).

Ant response to glandular compounds

To examine ant responses to Mozena sp. glandular compounds I induced coreids to spray on cotton swabs in the field. I conducted seven paired trials, measuring the responses of P. spinicola ants to sterile (control) and glandular spray (experimental) cotton swabs placed in random order on a branch, always near a young leaf with Beltian bodies and extra-floral nectaries, typical locations of feeding coreids. I recorded ant behaviors as described above. The experiment was blind in that the researcher was not aware of which treatment had been applied while scoring ant response. The numbers of ant responses were converted to proportions of total encounters per treatment and arcsine square-root transformed for analyses. I analyzed the responses (attack, pass, and avoidance) from both the cotton swab experiments using MANOVA models, a significant MANOVA was followed by individual ANOVA tests for each of the three response variables. I performed these analyses in Statistica (StatSoft, Inc. 2001).

Ant response to lanolin-dissolved cuticular and glandular compounds

To determine if there would be similar ant responses to substances produced in Mozena sp. glands versus those on cuticles, I isolated and removed these substances from Mozena sp. individuals, employing a novel method using lanolin paste. Lanolin is natural wax derived from sheep wool and composed of esters, fatty acids, and alcohols that can be used to dissolve a variety of solutes. I coated each of seven individuals with lanolin paste, first spreading lanolin over the metathoracic gland and then on the exposed cuticle surfaces, including the dorsal side of the thorax and the ventral side of the abdomen. Lanolin-coated coreids were left for eight hours to allow the absorption of chemicals, after which I transferred lanolin paste from the gland of a

6 coreid and the cuticle of a coreid to sterile cotton swabs. I conducted behavioral trials on V. collinsii trees, measuring ant responses to cotton swabs with glandular substances, cuticular substances, and a control treatment with lanolin only for two minutes each on different V. collinsii trees. Each of seven replicate trials was conducted using the same placement criteria and ant behavior definitions as described above. The experiment was blind in that the researcher was not aware of which treatment had been applied while scoring ant responses. I analyzed ant responses using a Kruskal-Wallis test for non-parametric data using Statistica (StatSoft, Inc. 2001).

Removal of coreid cuticular hydrocarbons

To assess the role of Mozena. sp. cuticular substances, I washed coreids in several solutions known to remove various cuticular chemicals in arthropods. Then I compared P. spinicola worker responses among coreids washed in the various solutions. Coreids were killed by immersion in solution and washed in one of three solutions: 100% distilled water (polar solvent), 100% methanol (a polar solvent), and 100% hexane (a non-polar solvent). Within a replicate, I allowed washed individuals to dry for an hour and then placed them on branches of a single V. collinsii tree in a random order. All experiments were conducted between 0800 and 1100 hrs. I recorded resident ant responses for two minutes as described above for each of fifteen replicates. The experiments were blind in that the researcher was not aware of the treatment that was applied when scoring ant responses. I analyzed the proportions of total encounters that were attacks using the Friedman test (non-parametric test) for the blocked design (analysis was blocked by tree on which individuals were placed, n=15) using R (R Development Core Team 2009).

Ant response to methyl salicylate

I prepared a solution of lanolin containing methyl salicylate, in similar concentration to that which was found in a preliminary chemical analysis of coreid defensive spray- 12.5!l/ml, by melting lanolin in sterile beakers over a burner, adding pure methyl salicylate, mixing, and allowing to cool until the solution returned to a semi-solid state. I treated pure lanolin used as a control with heat, without the addition of methyl salicylate. I applied these solutions to sterile cotton swabs and wrapped in aluminum foil before transport into the field for use in bioassays. For each replicate, the control or experimental cotton swabs were presented in random order on the same branch of one V. collinsii tree. I recorded ant responses to the cotton swabs according to the categories described above (bite, sting, pass, avoid) for two minute trials, for a total of 15 trials. I conducted each trial on a different tree, each at least 25 meters apart to ensure different colonies of ant-occupants. The experiment was blind in that the researcher was not aware of which treatment had been applied while scoring ant response. All experiments took place between 0800 and 1100 hrs. I compared attack rates using a paired t-test of the arcsine square-root transformed data in R (R Development Core Team 2009).

Chemical analyses

To analyze the chemical structure of coreid and ant cuticular compounds, I preserved the washes from the above experiment using water, methanol and hexane. Additionally, I collected

7 coreid defensive sprays by inducing individuals to spray onto sterile glass wool and then depositing the glass wool containing the spray into vials containing methanol, water, and hexane. I used a similar technique to obtain the alarm pheromones of P. spinicola workers. I presented ants with sterile glass wool until they released their alarm pheromones and preserved the wool in each of the same three chemicals (water, hexane, and methanol). I washed ants in the same chemicals to collect their cuticular hydrocarbons. I transported all substances obtained in this fashion to Dr. Andre Kessler at Cornell University where they were analyzed using gas chromatography mass spectrometry (GCMS) to determine the identities and relative abundances of chemicals present.

RESULTS

Specificity to ant colony of origin

There was no significant difference in attacks by P. spinicola on replaced (control) or translocated (experimental) individuals (t-test, df=8, t=-0.68, p= 0.52).

Species-specificity to Vachellia host

Of 20 Mozena sp. used in captive choice trials, I found ten on acacia branches rather than on the mesh enclosures used in feeding trials. Mozena sp. individuals showed no statistically significant preference of acacia species: five chose Vachellia collinsii; four chose V. cornigera; and one chose V. farnesiana (Chi-square test, df = 2, X2 = 4.36, p = 0.11). There was a non- significant trend for Mozena sp. to prefer the obligate ant-acacias (V. collinsii and V. cornigera) to the facultative ant acacia, V. farnesiana (Chi-square test, df = 1, X2 = 2.99, p = 0.08) (Figure 2.1).

Species-specificity to ant host species

Nine out of ten coreids remained on all branches for the duration of 30-second field experiments with different ant species. The proportion of total encounters that resulted in attack (biting and stinging) was not different among the ant species (ANOVA, F2,6 = 5.42, p = 0.39) (Figure 2.2).

Ant response to glandular compounds

P. spinicola ants responded significantly differently to cotton swabs with Mozena sp. glandular cues than to untreated cotton swabs (MANOVA, F3,21= 4.00, p < 0.001). Specifically, ants avoided cotton swabs containing glandular compounds significantly more often than controls (ANOVA, F1,13 = 48.74, p < 0.001) and attacked cotton swabs containing glandular cues significantly less often than controls (ANOVA, F1,13 = 151.17, p < 0.001), whereas there was no significant difference in passing behavior (ANOVA, F1, 13, p=0.72) (Figure 2.3).

8 6 5 4 3 2 Frequency 1 0 V. collinsii V. cornigera V. farnesiana

Acacia Species Figure 2.1. The frequency of choice of acacia host plant by Mozena sp. individuals in captive feeding experiments. There was no significant difference among the acacia species.

0.4 0.35 0.3 0.25 0.2 0.15

resulting in bites 0.1 0.05 Mean proportion of encounters 0 C. brevispinosa P. flavicornis P. spincola Ant species

Figure 2.2. Mean proportion of encounters (+/- standard error) resulting in bites for Mozena sp. placed on V. collinsii with different ant residents.

9 a. 1

0.8

0.6

0.4

Proportion Attack Proportion 0.2

0 Control Glandular Substances Substance Type b. 1.2 1 0.8 0.6 0.4

Proportion Avoid Avoid Proportion 0.2 0 Control Glandular substances Substance Type c. 0.25 0.2 0.15 0.1

Proportion Pass 0.05 0 Control Glandular Substances Substance Type

Figure 2.3. Mean proportion (+/- standard error) of encounters in which P. spinicola attacked (a), avoided (b), or passed (c) cotton swabs containing no secretions from Mozena sp. (control) and those containing glandular substances. P. spinicola attacked glandular cues significantly less often than controls (p<0.001), avoided glandular substances significantly more often than controls (p<0.001), and passed both control and glandular swabs equally (p=0.72).

10 Ant response to lanolin-dissolved cuticular and glandular compounds

Control cotton swabs were attacked significantly less often by P. spinicola than those treated with either glandular or cuticular substances (Kruskal-Wallis, df=2, H = 13.29, p = 0.0013). There were no differences among control, glandular and cuticular treatments for either pass or avoidance behaviors (Kruskal-Wallis, df=2, Pass: H = 1.13, p = 0.57, Kruskal-Wallis, df=2, Avoid: H = 2.30, p = 0.32) (Figure 2.4). a. 1 0.8 0.6 0.4 0.2 0 Control Glandular Cuticular Proportion Attack Proportion Substances Substances Substance Type b.

1 0.8 0.6 0.4 0.2 Proportion Avoid Avoid Proportion 0 Control Glandular Cuticular Compounds Compounds Substance Type c. 0.8 0.6 0.4 0.2

Proportion Pass 0 Control Glandular Cuticular Substances Substances Substance Type

Figure 2.4. Mean proportion (+/- standard error) of behaviors of P. spinicola that were attacks (a), avoids (b), and passes (c) for lanolin-only control, glandular, and cuticular substances (both dissolved in lanolin wax). P. spinicola did not attack swabs containing glandular or cuticular substances (a) and there was no significant difference in passing or avoiding between control, glandular, or cuticular substances (b, c).

11 Removal of coreid cuticular hydrocarbons

There was no significant difference in attack rates among water-washed, methanol- 2 washed, and hexane-washed individuals (Friedman test, df=2, X r =4.98, p=0.08) (Figure 2.5). Hexane-washed individuals experienced higher proportions of attacks, though this was not a significant difference. Therefore, properties of cuticular chemistry necessary for Mozena sp. to avoid ant attack may include both polar and non-polar compounds. !#($" !#(" !#'$" !#'" !#&$" !#&" !#%$" !#%" !#!$" .*,-"541514/61-"17",//,89:" !" )*+,-*" .*/0,-12" 3,/*4" ;4*,/<*-/"

Figure 2.5. Mean proportion of total encounters (+/- standard error) that resulted in attacks by P. spinicola on hexane-, methanol-, and water-washed Mozena sp. (p=0.08).

Ant response to methyl salicylate

Pseudomyrmex spinicola ants attacked control cotton swabs containing pure lanolin significantly more than experimental cotton swabs containing methyl salicylate (paired t-test, df=14, t=-9.74, p< 0.001). In addition, there was no effect of order of presentation on ant response.

Chemical analyses

A GCMS analysis of the cuticle washes from P. spinicola ants and Mozena sp. revealed similar chemical profiles and relative abundances from the gas chromatograph data (Figure 2.6). These chemical profiles differ from the gas chromatograph profiles of other ants and coreids (Wagner et al. 2000; Aldrich et al. 1993). The mass spectrometry results allowed identification of some compounds from their fragmentation pattern into individual ions, and all were hydrocarbons. Due to contamination, the results of GCMS analysis of defensive sprays of Mozena sp. and P. spinicola ants were inconclusive.

Figure 2.6. Gas chromatography (GC) analysis of the cuticular hydrocarbon profiles of Mozena sp. and Pseudomyrmex spinicola ants.

DISCUSSION

Exploiters infiltrate ant-plant mutualisms through chemical mimicry or corruption of the host communication systems (Höllodobler and Wilson 1990; Akino et al 1999; Barbero 2009) or through their behavior or morphology (Styrsky and Eubanks 2007; Reiskind 1977). Through a series of behavioral bioassays and chemical analyses I examined and found evidence for the third approach: mimicry or corruption of a host’s chemical communication system, which allows Mozena sp. to exploit the ant-acacia mutualism.

Species-specificity to ant- and plant-hosts

First, in a series of behavioral bioassays I determined the host specificity of Mozena sp. individuals to both ant and plant hosts. I predicted that in order to successfully infiltrate the ant- acacia relationship, Mozena sp. would be highly specialized with respect to either ant, ant colony, or acacia species. Counter to my predictions, I found that Mozena sp. individuals showed no significant specificity to either ant or plant hosts. Suggesting that they are able to persist with multiple ant colonies and species and on multiple acacia tree species. This suggests a generalized chemical cue is responsible for Mozena sp. ability to invade the ant-acacia mutualism.

13 While plants with ant residents continue to produce defensive secondary compounds, they may contain lower levels than their conspecifics without mutualist ants. Piper cenocladum individuals with ant mutualists, for example, invest less in amide-loading of plant tissue than plants lacking ants (Dyer et al 2001). Myrmecophytic acacias contain reduced amounts of cyanides, chitinases, and alkaloids in comparison to non-myrmecophytic acacias (Heil et al. 1973). Plants with reduced levels of secondary compounds should be more palatable to herbivores, but in this study Mozena sp. appeared to prefer all acacia species equally, suggesting that they are able to tolerate plant secondary compounds in the diet. However, owing to small sample size the non-significant trend in preference for obligate acacias (V. collinsii and V. cornigera) should be interpreted conservatively. Given a larger sample size, preference of obligate ant-acacias likely be significant. Other species in the Mozena genus are known to use a volatile chemical glandular compound in defense (Aldrich et al. 1982). In addition to this apparently general defense, cuticular compounds have been implicated in rendering coreids “invisible” to ants as I have shown here. Chemical cues that confer such invisibility or that garner the aid of ants are often species, if not colony, specific. For example, orioles and kiskadees build nests that are thought to absorb ant colony odor, eventually causing ants to ignore these foreign objects their host trees (Janzen 1983). I found no difference in the proportion of total encounters that resulted in attacks on coreids among the three ant species, suggesting that no such colony specificity exists for Mozena sp., perhaps because the coreids use a broad-spectrum chemical cue, effective for many ant colonies and species.

The role of coreid glandular and cuticular compounds

My subsequent hypotheses addressing the specific nature of chemical cues as either deterrents or camouflages showed varying levels of support. Ants avoided glandular compounds, suggesting a deterrent effect on ants. This finding is consistent with previous studies that support a defensive role for coreid glandular compounds (Aldrich et al. 1982). However, lanolin-dissolved cuticular compounds showed no significant difference between pass and avoidance behaviors. It is clear that cuticular cues play a role in Mozena sp. persistence on V. collinsii, since no attacks occurred on cotton swabs containing cuticular compounds, but whether as a deterrent or a camouflage remains ambiguous from these results. The larvae of a lycaenid butterfly mimic the cuticular chemistry of ant broods; a larva is able to trick ants into carrying it back to the ant nest and supplying its food until it undergoes metamorphosis (Nash et al 2008). Whether the cuticular compounds of Mozena sp. have a similar mimetic function remains unclear from these results, but was elucidated by GCMS analysis of Mozena sp. and P. spinicola chemistry, discussed below. To further investigate cuticular chemistry of Mozena sp. individuals, and to make inferences about its chemical properties, I performed a series of washes of coreids in different solvents. Bioassays using individuals washed in each of three solvents (water, methanol, and hexane) showed no significant difference among the treatments. Hexane-washed individuals were attacked more frequently than water- or methanol-washed individuals, although this was not significant. Hexane is known to dissolve non-polar substances such as oils and waxes, therefore I infer that Mozena sp. cuticular chemistry includes these substances in addition to polar compounds that might be dissolved by water and methanol. Mozena sp. individuals observed naturally present on trees are not attacked by ants (personal observation), suggesting

14 that Mozena sp. individuals produce or acquire chemicals that allow them to remain undetected by ants.

The role of methyl salicylate in ant-coreid interactions

Preliminary chemical analyses of defensive sprays of coreids revealed that one major chemical component was methyl salicylate. Using bioassays, I tested ant response to this chemical in the field and found that ants attacked cotton swabs containing lanolin-dissolved methyl salicylate significantly less often than control swabs containing lanolin only. Thus, methyl salicylate is confirmed as a chemical important in Mozena sp. persistence on ant-acacias. However, methyl salicylate is a polar compound, conflicting somewhat with my solvent wash findings that showed non-polar compounds as most important in Mozena sp. ability to persist without initiating normal defenses of P. spinicola colonies. Ants are known to produce and use a wide variety of semiochemicals of varying polarity, volatility, stereochemistry, and functional groups (Keeling et al 2004), thus it is likely that there is a suite of cuticular and glandular compounds produced or absorbed by Mozena sp. individuals and many play a role in averting attacks by resident ants. Seeds, flowers, and leaves are known to contain methyl salicylate in some legume and acacia species (Brouat et al. 2000; Lamarque et al. 1998). Methyl salicylate is a known component of herbivore-induced volatile compounds (also called green leafy volatiles) released by plants that are attractive to predators and parasitoids of herbivores (Kessler and Baldwin 2001). Concentrations of chemicals often determine function and I am unsure if the concentrations of methyl salicylate indicate similar functions as methyl salicylate emitted from damaged leaves. Methyl salicylate plays diverse roles in interspecific interactions among insects: for example it plays a role in seed dispersal by some ants (Seidel et al. 1990) and acts as an attractant for pollinators in others (Knudsen and Tollsten 1993), in addition to signaling damaged tissue. The role of methyl salicylate in Psuedomyrmex spp. communications is unknown, but coreids may secrete a combination of green leafy volatiles (including methyl salicylate) when disturbed in order to initiate ant defense of the host plant, while using cuticular hydrocarbons to remain chemically undetectable from ants, gaining their protection.

Chemical analyses

Finally, I used chemical analyses to compare and contrast cuticular hydrocarbons and defensive spray of coreids with the cuticular hydrocarbons and defensive spray of ants. Coreid cuticular substances and coreid defensive spray as well as ant cuticular substances and sprays were collected for chemical analyses revealed a close match between Mozena sp. and P. spinicola cuticular chemical profiles, further supporting chemical mimicry as the proximate mechanism mediating this interaction.

Conclusions

In conclusion, I found that Mozena sp. is an exploiter of the acacia-ant mutualism that has been able to successfully invade the relationship between ants and plants though the use of non- species, non-host-specific chemical properties. Mozena sp. does not appear to prefer to feed on one acacia species over another, appears to be able to persist on multiple colonies, and is

15 defended equally well against Pseudomyrmex spp. ant mutualists and Crematogaster brevispinosa ant mutualists. Whether Mozena sp. is a specialist on the acacia genus over other plants was not addressed in this study, but in general Mozena spp. are associated with the Fabaceae and acacias in particular (Schaefer and O’Shea 1979; Brailovsky and Barrera 2001). Chemical cues have a fundamental role in both insect-plant and insect-insect interactions. The exploitation of the ant-acacia mutualism by Mozena sp. provides an exemplary mutualist- exploiter interaction in which to address the mechanisms by which these cues function and evolve. Future experiments should seek to determine the origin of this hemipteran’s chemical compounds: whether it sequesters them from the diet, synthesizes them, or gains them by physical contact with host ants. Furthermore, identifying more chemical compounds that play a role in evading or deterring ant attack and investigating whether they act broadly against Formicidae or are specific to ants involved in mutualims with acacias, would add to our knowledge of the mechanisms by which Mozena sp. is able to persist on acacias.

16 CHAPTER 3

THE COSTS OF EXPLOITATION OF MUTUALISMS

INTRODUCTION

Many mutualistic relationships have been invaded by exploiters, from diverse taxonomic groups and habitats, that benefit from commodities or services provided by mutualistic species while giving nothing in return to either mutualist partner (Soberon and Martinez del Rio 1985, Bronstein 2001). Examples of exploitation include non-nitrogen fixing bacteria that inhabit root nodules of legumes (Kiers et al. 2003), cleaner wrasse mimics that do not clean client reef fish (Wickler 1968), non-pollinating seed parasites (Bronstein and Ziv 1997; West and Herre 1994), nectar robbing or thieving insects or vertebrates (Wyatt 1980; Deng et al. 2004; Traveset et al. 1998), non-rewarding plants that mimic species that provide rewards (Anderson et al. 2005), spiders that consume Beltian bodies produced on ant-defended plants (Meehan et al. 2009), and non-defending or castrating ants inhabiting mymecophytic plants (Yu and Pierce 1998). The net benefit of mutualism for each partner is a balance between benefits obtained from the interaction, less the costs of reciprocation to obtain them. The costs of participating in a mutualism vary, from relatively cheap, like production of extra-floral nectar to attract and reward pollinators (O’Dowd 1980), to relatively costly, like seed dispersers that destroy a large proportion of seeds (Vander Wall 1994). Individuals that can obtain benefits without investing in reciprocation increase their own net benefit, leading to exploitation replacing cooperation over evolutionary time, barring mechanisms to prevent this (Bronstein 2001). The conditions under which mutualisms can persist despite the “temptation to defect” faced by each partner are often somewhat restricted (Axelrod and Hamilton 1981; Bronstein 2001). The temptation to defect faced by each mutualist may be compounded by exploitation by individuals and species that are not partners in the mutualism. Exploiters obtain benefits, without investing in reciprocation; adding to the cost of the mutualism for one or both partners and decreasing the net benefit. The decrease in net benefits due to exploitation may destabilize mutualistic relationships by leading to a net imbalance in the exchange of commodities or services between mutualist partners, making it more costly and less beneficial to invest in reciprocation. This is especially important for high-cost exploitative behaviors, for example killing and consuming a partner in a mutualism (Bronstein 2001; Sakata 1994). Theory predicts that, in order for mutualisms to persist, exploitation of mutualisms must have relatively low costs (Bronstein 2001). Previous studies have quantified the costs of exploitation in some mutualist-exploiter interactions. Studies have focused on a variety of response variables, for example: alteration of mutualist behavior (Irwin and Brody 1999); reduction in the availability of rewards (Ordano and Ornelas 2004); impact of exploitation on growth (Clement et al. 2008); or reproduction (Yu and Pierce 1998). Maloof and Inouye (2000) reviewed nectar robbery, when flowers are pierced to gain access to floral nectar, and concluded that robbers are equally likely to increase or decrease plant fitness, but relied on a qualitative approach. Irwin et al. (2001) performed a meta-analysis of floral larceny, a term used to refer to both nectar robbing, when flowers are pierced to gain access to nectar, and thieving, when flowers are entered as pollinators would but little or no pollen is transferred. Their analysis found that floral larceny had weak negative effects on mutualists. However, the scope of their analysis was limited to floral larceny, just one category of exploitation of mutualisms. Irwin et al. (2001) also included studies that experimentally

17 manipulated nectar levels and found that artificial manipulations of larceny had larger effect sizes than natural larceny. I performed a meta-analysis of the exploitation of mutualisms across several categories of exploitation and across different ecosystems to answer the question: what is the cost of being exploited? High costs of exploitation may destabilize mutualisms, whereas low costs may allow the coexistence of mutualists and exploiters over long ecological time scales. This meta-analysis included naturally exploited mutualisms (in experimental and observational studies) and included a variety of categories of exploitation. The scope of this meta-analysis was limited to terrestrial mutualisms in which at least one mutualist partner is a plant. I chose studies that have quantified the reproductive success of a mutualist partner in the presence and absence of exploitation. The exploitation of mutualisms, especially the costs of being exploited, provides fertile ground for the study of how mutualisms are maintained in the face of potentially destabilizing exploitation.

METHODS

To quantitatively examine the cost of exploitation of mutualisms, a meta-analysis was performed using the methods of Hedges et al. (1999). The effect size of the cost of exploitation of mutualism was measured as the log response ratio of reproductive success (measured as fruit or seed set of plants) in the presence and absence of exploitation. The biological relevance of the log response ratio is clear; it is the proportional change in reproductive success (the response variable) of exploited versus unexploited individuals. A positive effect size indicates exploitation led to an increase in reproductive success, whereas a negative effect size indicates exploitation led to decreased reproductive success. The meta-analysis was based on published studies found via searches on BIOSIS, Google Scholar, and JSTOR from (1950-2009), using the search terms “aprovechados,” “cheater,” “exploiter,” “non-defending mutualist,” “non-pollinating seed parasite,” “parasite of mutualism,” “robber,” and “thief,”. The list was also supplemented with studies cited in the reference lists of articles identified, as well as by searching studies that cited one or more articles already identified. Once studies were identified, to be included in the meta-analysis they had to meet the following criteria: (1) published in English; (2) studies that quantified the effect of exploitation of a mutualism in terms of reproductive success of one or both partners in a mutualism; and (3) studies that reported costs of exploitation (as reproductive success) under field conditions, thus artificial simulation of exploitation was not included. Studies that did not meet all of these criteria were excluded from the meta-analysis. Data sets that quantified the effect of exploitation on a mutualist’s behavior, but not reproductive success, were excluded from the analysis. No studies included estimates of reproductive success of animal mutualist partners, thus fruit or seed set of plants was the only suitable response variable. When relevant data were available only as a figure, a digital copy of the figure was used to lift the data of interest using UTHSCSA Image Tool software (University of Texas, USA). I identified sixteen studies and 19 datasets acceptable for the meta-analysis (Table 3.1). Two studies (Arizmendi et al. 1996 and Zhang et al. 2009a) had multiple datasets that were suitable for the analysis since they studied interactions between mutualists and exploiters of different species. All studies included at least one mutualist partner that was a plant. Of the 19 datasets, the types of exploitation represented were floral robbing, floral thievery, non-

18 pollinating seed parasitism, non-rewarding plant mimics, and plant castration by resident ants (Table 3.1) and all quantified the reproductive success of plants, not their mutualist partners. Many relevant studies did not report estimates of variance, but these were included because their elimination would have drastically reduced the sample size of the meta-analysis. This may decrease the power to detect differences between groups (increased Type II error), but should not bias the estimate of overall effect size (Englund et al. 1999; Hedges et al. 1999). An increased probability of making a Type II error was preferable to the reduction in sample size. To determine if exploitation had a significant effect on reproductive success, I compared the overall effect size to zero in a one sample t-test using R (R Development Core Team 2009). Differences among pre-planned comparisons were investigated to compare the effect sizes among studies. Effect sizes were compared in studies with different types of exploiter-mutualist combinations (vertebrate-vertebrate, insect-vertebrate, and insect-insect) as well as across two different levels of specialization of mutualisms: those that were highly specialized (one-partner) and those that were generalized (more than one partner). When there was more than one type of mutualist or exploiter (e.g., insect and vertebrate as plant pollinators or nectar robbers) I chose the most common or ecologically important (as identified by the relevant study) for the analysis. I compared effect sizes using analysis of variance in R (R Development Core Team 2009).

RESULTS

There was no significant effect of exploitation on the reproductive success of exploited species (t=-1.77, df=18, p=0.09). The cumulative mean effect size (+/- standard deviation) was -1.03 (+/-2.53) with a range of -11.2 to 0.32. The mean effect size would indicate a very large effect, primarily due to one outlier log ratio at -11.2. When excluded from the analysis, the mean effect size (+/- standard deviation) was -0.47 (+/-0.65). A qualitative view of the effect sizes, excluding the outlier, shows that most exploitation of mutualisms had a negative, but weak effect on reproductive success, as an absolute value of greater than 1.0 is considered a very large effect, 0.8 a large effect, 0.5 a medium effect and 0.2 a small effect (Cohen 1969). Reproductive success was not significantly different among mutualisms exploited by vertebrate, insect, or plant exploiters (F=0.17, df=2, P=0.84) (Figure 3.1). Reproductive success was not significantly different between exploited vertebrate-plant mutualisms or insect-plant mutualisms (F=0.39, df=1, p=0.54) (Figure 3.1). The degree of specialization of the mutualism (number of partners) had no significant impact on reproductive success in the presence of exploitation (F=3.52, df=1, p=0.08), plants with only one mutualist partner that were exploited were as reproductively successful as those with more than one mutualist partner (Figure 3.1). Similarly, the number of exploiters had no significant impact on reproductive success; reproductive success was the same for mutualisms with one exploiter and those with more than one (F=0.16, df=1, p=0.69) (Figure 3.1). Exploitation was a relatively common occurrence: for the 12 studies that gave estimates of rates of exploitation, nine described rates of exploitation greater than 50%. Study sites represented a variety of terrestrial biomes with locations in alpine and boreal habitats, deserts, temperate regions, and tropical forests.

19 Table 3.1. Studies included in the meta-analysis and their relevant characteristics. Under the study type column, O represents observational study, E represents experimental. Study Exploiter Number of Mutualist Number of Plant Effect of Study Type Study Topic Exploiter Type Exploiters Mutualist Type Mutualists Mutualist Exploitation Non-nectar Prosoeca Zaluzianskya Anderson 2005 O reward mimic Disa nivea Plant 1 ganglbaueri Fly 1 microsiphon Negative Arizmendi et al. Multiple Salvia 1996 O Nectar robber Diglossa baritula Bird 1 hummingbirds Bird 5 mexicana Positive Arizmendi et al. Hummingbirds, Bird and Fuchsia 1996 O Nectar robber Diglossa baritula Bird 1 bumble bee Bee 4 microphylla Positive Bronstein & Ziv Non-pollinating Tegeticula 1997 O seed parasite Carpophilus longus Beetle 2 yuccasella Moth 1 Yucca schottii Negative Castro et al. Polygala 2008 E Nectar robber Multiple bees Bee 6 Multiple bees Bee 4 vayredae Negative Tamiops swinhoei Alpinia Deng et al. 2004 O Nectar robber hainanus Squirrel 1 Multiple bees Bee 3 kwangsiensis Negative Lara and Ornelas Lampornis Moussonia 2001 E Nectar robber Tropicoseius sp. Mite 1 amethystinus Bird 1 deppeana Negative Mertensia Morris 1996 E Nectar robber Bombus spp. Bee 2 Bombus spp. Bee 5 paniculata Negative Multiple Multiple Macleania Navarro 2001 O Nectar robber hummingbirds Bird 2 hummingbirds Bird 2 bullata Negative Chilopsis Richardson 1994 O Nectar robber Xylocopa californica Bee 1 Multiple bees Bee 2 linearis Negative Phaethornis Pavonia Roubik 1982 E Nectar robber Trigona ferricauda Bee 1 superciliosus Bird 1 dasypetala Negative Sephanoides Traveset et al. galeritus, bird, Bird and Fuchsia 1998 O Nectar robber Phrygilus patagonicus Bird 1 and bee bee 3 magellanica Negative Asclepias Wyatt 1980 E Nectar robber Ant Ant 1 Multiple insects Insect 5 curassavica Negative Yu & Pierce Cordia 1998 O Castrating ant Allomerus demerarae Ant 1 Azteca sp. Ant 1 nodosa Negative Zhang et al. Glechoma 2007 O Nectar robber Multiple bees Bee 2 Multiple bees Bee 7 longituba Negative Zhang et al. Corydalis 2009a O Nectar robber Bombus pyrosoma Bee 1 Multiple bees Bee 3 tomentella Negative Zhang et al. Corydalis 2009a O Nectar robber Bombus pyrosoma Bee 1 Multiple bees Bee 3 incisa Positive Zhang et al. Corydalis 2009a O Nectar robber Bombus pyrosoma Bee 1 Multiple bees Bee 3 ternatifolia Negative Zhang et al. Non-pollinating Ficus 2009b E seed parasite Diaziella bizarrea Wasp 1 Eupristina sp. Wasp 1 glaberrima Negative

20 a. c.

Insect Plant Vertebrate Insect Vertebrate 0 0

-0.5 -0.5 -1

size -1 -1.5size -1.5 -2 Cumulative mean effect Cumulative mean effect -2 Cumulative mean effect Cumulative mean effect -2.5 Type of mutualist Type of exploiter b. d. Greater than one One Greater than one One 0 0

-1 -0.5 size -2 -1 size -3 Cumulative mean effect Cumulative mean effect -1.5 -4 Number of mutualists Cumulative mean effect Cumulative mean effect -2 Number of exploiters

Figure 3.1. Cumulative mean effect sizes (loge ratio) (+/- standard error) for comparisons between (a) exploiter type, (b) number of exploiters, (c) mutualist type, and (d) number of mutualists.

21 DISCUSSION

This meta-analysis was performed to answer the question: What is the cost of exploitation of mutualisms? I found that there was a weak negative cost of exploitation, though it was not significant. Overall, rates of exploitation were relatively high, but exploitation did not result in a significant difference in reproductive success for exploited versus unexploited individuals or populations. Exploiter type or number and mutualist type or number had no significant impact on the reproductive success of exploited individuals. This result supports the hypothesis that exploitation does not inflict high costs that would destabilize mutualisms, causing them to degrade over time. The studies included in this meta-analysis were arrayed along a continuum of effects of exploitation from very high costs in one study in which exploited individuals had a reproductive success close to zero (Zhang et al. 2009b), to those that found higher reproductive success in individuals that had been “exploited” (Zhang et al. 2009a; Arizmendi et al. 1996). This pattern of variation in the effect of exploitation may result from a number of underlying mechanisms. I investigated a number of comparisons among effect sizes to find evidence of these mechanisms. First, I investigated the impact of exploiter type and number of exploiters on the cost of exploitation. I found no significant difference in effect sizes among plant, insect, or vertebrate exploiters and no significant difference in effect sizes for systems that were exploited by one species versus those exploited by more than one. While exploiters are consuming commodities or using services provided for mutualists, they are inflicting no harm on the mutualistic species, perhaps because they consume commodities or use services that are relatively inexpensive to produce, they cause no structural damage, they do not alter the behavior of mutualist partners, or a combination of these. Whether there is one exploiter species cheating the mutualism or many, there is no difference in effect sizes. Highly specialized mutualisms (one partner) and generalized mutualisms (more than one partner) fared equally well in the presence of exploitation, as did mutualistic relationships between plants and insects or plants and vertebrates. This suggests that highly specialized mutualisms are no more vulnerable to exploitation than generalized mutualisms, and that for vertebrate-plant or insect-plant mutualisms, the cost of exploitation is negligible. A few studies had a positive effect size, indicating that reproductive success was higher in “exploited” individuals than those that were not exploited. This result suggests that some cases of exploitation may be better described as low-efficiency mutualisms or as commensalisms (Arizmendi et al. 1996). Some exploiters may provide direct benefits, like low efficiency pollination (Arizmendi et al. 1996) or may stimulate vegetative growth (Yu and Pierce 1998) or may stimulate replenishment of commodities (like floral nectar) after removal, which allows plants to continue to attract mutualists (Ordano and Ornelas 2004). The indirect costs or benefits of exploitation, for example the alteration of mutualist partner behavior following exploitation, on plant or partner reproductive success was not included in the analysis, because no studies quantified this. Many of the relevant studies came from the nectar larceny literature, since studies of other types of exploitation have not yet quantified the impact of exploitation as a function of reproductive success. No studies assessed the reproductive success of both plants and their mutualist partners, thus costs for non-plant mutualist partners were not included in the analysis. It is possible that organisms at different trophic levels experience varying costs of exploitation.

22 The negative result of this meta-analysis may be due to a number of factors. First, many mutualist-exploiter interactions are neglected in studies and the literature because they are difficult to detect and the costs of exploitation are difficult to quantify. Exploiters follow a number of different strategies from pure exploitation to facultative exploitation and their interactions with many host organisms are transient, making it difficult to detect interactions with host mutualists or the duration of the interaction, thus very low costs of exploitation are often not recognized or studied. On the other hand, very high costs of exploitation that destabilize and degrade mutualisms may have already forced the elimination of the interaction, thus extant mutualisms do not reflect high cost exploitation. The studies used in this meta-analysis all quantified fitness as reproductive success measured by fruit or seed set. However, if more studies incorporated other proxies for fitness such as measures of growth, survival, and other measures of reproduction, more of the true costs of exploitation might have been apparent. Seed mass, germination rates of seeds, vegetative growth, decreases in rewards for mutualists, and investment in defense against exploitation are a few other ways that some of these studies could have measured the costs of exploitation. If the full costs had been measured it might have led to a positive result in this meta-analysis. Five out of 19 data sets contained no significant difference in the reproductive success of exploited versus unexploited individuals. The lack of statistical power in these data sets, due to small sample sizes or high variability, may have contributed to the negative result in this meta- analysis. This meta-analysis found a continuum of effects of exploitation from very small decreases in reproductive success, to very large, to positive effects of exploitation. Given the variation across studies, it seems likely that over evolutionary time some of these mutualist- exploiter interactions may convert to purely antagonistic interactions like predation, parasitism, and herbivory for those with large negative effects, remain static for those with negligible costs, or convert into mutualisms for those with positive effects. It is possible that such changes in the nature of mutualist-exploiter interactions have occurred or are occurring in some of cases. Other mechanisms that were not investigated here, but that may be important to the maintenance of mutualisms in the face of exploitation include sanctions or punishment for exploitation (Kiers et al. 2003; Izzo and Vasconcelos 2002), competition between mutualists and exploiters (Morris et al. 2003), and spatial and temporal heterogeneity (Yu et al. 2001; Wilson et al. 2003). Phylogentic evidence suggests that exploiters have coexisted with mutualists over long periods of time (Pellmyr and Leebens-Mack 1999; Morris et al. 2003), therefore conditions favoring the long-term and short-term maintenance of mutualisms do exist. This meta-analysis suggests that one of these conditions is the relatively low cost of exploitation on mutualistic relationships.

23 CHAPTER 4

CONCLUSIONS

My results contribute to the understanding of the methods that exploiters of a mutualism employ to take advantage of one or both partners in a mutualism and the potential cost of exploiters to mutualistic interactions. In Chapter 2, I explored the methods by which an herbivore infiltrates ant defenses to feed on a myrmecophytic acacia. Through a series of behavioral bioassays and chemical analyses I found evidence for mimicry or corruption of a host’s chemical communication system, which allows Mozena sp. to exploit an ant-acacia mutualism. Chemical analyses supported the experimental results, showing that the chemical profiles of ant and coreid cuticular compounds are very similar. After identifying one chemical component of coreid glandular secretions, methyl salicylate, I showed that ants do not attack objects impregnated with or covered in this compound through a series of behavioral bioassays and that it may contribute to mimicking or repelling ants so that coreids may feed without initiating normal ant defenses. In summary, I found that Mozena sp. is able to exploit the mutualism between ants and plants through the use of non-species, non-host specific chemical compounds. In Chapter 3, I performed a meta-analysis to answer the question: What is the cost of exploitation of mutualisms? The studies included in the meta-analysis showed a continuum of effects of exploitation from very high costs in one study in which exploited individuals had a reproductive success close to zero (Zhang et al. 2009b), to those that found higher reproductive success in individuals that had been “exploited” (Zhang et al. 2009a; Arizmendi et al. 1996). Exploitation occured at relatively high rates among studies, but it did not result in a decrease in reproductive success for exploited versus unexploited individuals or populations and that exploiter type (insect, plant, or vertebrate) or number (one or greater than one) and mutualist type (insect or vertebrate) or number (one or greater than one) had no significant impact on the reproductive success of exploited individuals. This result supports the hypothesis that exploitation does not inflict high costs that would destabilize mutualisms, or cause mutualisms to degrade over time. Phylogenetic evidence suggests that exploiters have coexisted with mutualists over long periods of time (Pellmyr and Leebens-Mack 1999; Morris et al. 2003); therefore conditions favoring the long-term and short-term maintenance of mutualisms do exist. This meta-analysis suggests that one of these conditions may be the relatively low cost of exploitation. Collectively, these analyses illuminate methods by which exploiters may succeed in infiltrating mutualisms, and show that the relatively low costs exploitation of mutualisms may explain why exploitation poses no threat to the stability of some mutualistic interactions over evolutionary time.

LITERATURE CITED

Aldrich, J.R., G.K. Waite, C. Moore, J.A. Payne, W.R. Lusby, and J.P. Kochansky. 1993. Male- specific volatile from nearctic and Australasian true bugs (Heteroptera: Coreidae and Alydidae). Journal of Chemical Ecology 19(12): 2767-2781.

Aldrich, J.R, J.P. Kochansky, W.R. Lusby, and S.R. Dutky. 1982. Volatile male-specific natural products of a coreid bug (Hemiptera: Heteroptera). Journal of Chemical Ecology 8(11): 1369- 76.

Akino, T., J.J. Knapp, J.A. Thomas, G.W. Elmes. 1999. Chemical mimicry and host specificity in the butterfly Maculinea rebeli , a social parasite of Myrmica ant colonies. Proceedings of the Royal Society of London (B) 266: 1419-1426.

Anderson, B., S.D. Johnson, C. Carbutt. 2005. Exploitation of a specialized mutualism by a deceptive orchid. American Journal of Botany 92(8): 1342-1349.

Arizmendi, M.C., C.A. Dominguez, R. Dirzo. 1996. The role of an avian nectar robber and of hummingbird pollinators in the reproduction of two plant species. Functional Ecology 10: 119- 127.

Axelrod, R. and W.D. Hamilton. 1981. The evolution of cooperation. Science 211: 1390-1396.

Barbero, F., J.A. Thomas, S. Bonelli, E. Balletto, K. Schönrogge. 2009. Queen ants make distinctive sounds that are mimicked by a butterfly social parasite. Science 323: 782-785.

Brailovsky, H. and E. Barrera. 2001. Six new species of Mozena from Mexico (Heteroptera: Coreidae: Coreinae: Nematopodini). The Florida Entomologist 84(1): 99-111.

Bronstein, J.L. 1991. The nonpollinating wasp fauna of Ficus pertusa: exploitation of a mutualism? Oikos 61: 175-186.

Bronstein, J.L. 2001. The exploitation of mutualisms. Ecology Letters 4: 277-287.

Bronstein, J.L. and Y. Ziv. 1997. Costs of two non-mutualistic species in a yucca/yucca moth mutualism. Oecologia 112(3): 379-385.

Brouat C., D. McKey, J-M. Bessiere, L. Pascal, and M. Hossaert-McKey. 2000. Leaf volatile compounds and the distribution of ant patrolling in an ant-plant protection mutualism: Preliminary results on Leonardoxa (Fabaceae: Caesalpinioideae) and Petalomyrmex (Formicidae: Formincinae). Acta Oecologica 21(6): 349-357.

Castro, S., P. Silveira, L. Navarro. 2008. Consequence of nectar robbing for the fitness of a threatened plant species. Plant Ecology 199: 201-208.

25 Clement, L.W., S.C.W. Koppen, W.A. Brand, M. Heil. 2007. Strategies of a parasite of the ant- Acacia mutualism. Behavioral Ecology and Sociobiology 10.1007 (online first).

Connor, R.C. 1995. The benefits of mutualism: a conceptual framework. Biological Reviews, 70:427-457.

Deng, X-B., P-Y. Ren, J-Y. Gao, Q-J. Li. 2004. The striped squirrel (Tamiops swinhoei hainanus) as a nectar robber of ginger (Alpinia kwangsiensis).

Dettner, K. and C. Liepert. 1994. Chemical mimicry and camouflage. Annual Review of Entomology 39: 129-154.

Dyer, L.A., C.D. Dodson, J. Beihoffer, and D.K. Letourneau. 2001. Trade-offs in antiherbivore defenses in Piper cenocladum: ant mutualists versus plant secondary metabolites. Journal of Chemical Ecology 27(3): 98-331.

Elgar, M.A. and R.A. Allan. 2004. Predatory spider mimics acquire colony-specific cuticular hydrocarbons from their ant model prey. Naturwissenschaften 91: 143-147.

Englund, G., O. Sarnelle, S.D. Cooper. 1999. The importance of data selection criteria: meta- analyses of stream predation experiments. Ecology 80: 1132-1141.

Eubanks, M.D., K.A. Nesci, M.K. Peterson, L. ZhiIi, and H.B. Sanchez. 1997. The exploitation of an ant-defended host plant by a shelter-building herbivore. Oecologia 109: 454-460.

Grutter, A.S. and R.J.G. Lester. 2002. Cleaner fish Labroides dimidiatus reduce Argothona macronema (Corallanidae) isopod infections on the coral reef fish Hemigymnus melapterus. Marine Ecology Progress Series 243: 247-255.

Hedges, L.V., J. Gurevitch, and P.S. Curtis. 1999. The meta-analysis of response ratios in experimental ecology. Ecology 80: 1150-1156.

Heil, M., C. Staehelin, and D. McKey. 2000. Low chitinase activity in Acacia myrmecophytes: a potential trade-off between biotic and chemical defences? Naturwissenschaften 87(12): 555-558.

Hölldobler, B. 1999. Multimodal signals in ant communication. Journal of Comparative Physiology A 184: 129-141.

Hölldobler, B and E.O. Wilson (1990). Pgs 513-515 in The Ants. Harvard University Press, Cambridge, MA.

Inouye, D. 1983. The ecology of nectar robbing. Pgs 153-173 in B. Bently and T. Elias, eds. The biology of nectaries. Columbia University Press, New York.

26 Irwin, R.E. and A.K. Brody. 1998. Nectar robbing in Ipomopsis aggregata: effects on pollinator behavior and plant fitness. Oecologia 116: 519-527.

Irwin, R.E., A.K. Brody, N.M. Waser. 2001. The impact of floral larceny on individuals, populations, and communities. Oecologia 129: 161-168.

Izzo, T.J. and H.L. Vasconcelos. 2002. Cheating the cheater: domatia loss minimizes the effects of ant castration in an Amazonian ant-plant. Oecologia 133(2): 200-205.

Jackson, D.E. and F.L.W. Ratnieks. Communication in ants. Current Biology 16(15): R570- R574.

Janzen, D.H. 1966. Coevolution of mutualism between ants and acacias in Central America. Evolution 20: 249-275.

Janzen, D.H. 1975. Pseudomyrmex nigropilosa: a parasite of mutualism. Science 188:936-937.

Janzen, D.H. 1983. In: Costa Rican Natural History. University of Chicago Press, Chicago, IL: 735-754.

Janzen, D.H. 1985. The natural history of mutualisms. In The Biology of Mutualism (ed. Bouher, D.H.). Oxfod University Press, New York, USA: 40-49.

Keeling, C.I., E. Plettner, K.N. Slessor. 2004. Hymenopteran semiochemicals. Pgs 133-178 in S. Schulz, ed. The chemistry of pheromones and other semiochemicals, Volume 1. Springer, Berlin, Germany.

Kessler, A. and I.T. Baldwin. 2001. Defensive function of herbivore-induced plant volatile emissions in nature. Science 291(5511): 2141-2144.

Kiers, E.T., R.A. Rousseau, S.A. West, R.F. Denison. 2003. Host sanctions and the legume- rhizobium mutualism. Nature 425: 78-81.

Knudsen, J.T. and L. Tollsten. 1993. Trends in floral scent chemistry in pollination syndromes: floral scent composition in moth-pollinated taxa. Biological Journal of the Linnean Society 113: 263-284.

Lamarque, A.L., D.M. Maestri, J.A. Zygadlo, N.R. Grosso. 1998. Volatile constituents from flowers of Acacia caven (Mol.) Mol. Var. caven, Acacia aroma Gill. Ex Hook., Erythrina crista- galli L. and Calliandra tweedii Benth. Flavor and Fragrance Journal 13(4): 266-268.

Lara, C. and J.F. Ornelas. 2001. Nectar ‘theft’ by hummingbird flower mites and its consequences for seed set in Moussonia deppeana. Functional Ecology 15(1): 78-84.

Lenoir, A., P. D’Ettorre, C. Errard, A. Heftetz. 2001. Chemical ecology and social parasitism in ants. Annual Review of Entomology 46: 573-599.

Maloof, J.E. and D.W. Inouye. 2000. Are nectar robbers cheaters or mutualists? Ecology 81: 2651-2661.

Martin, S.J., E.A. Jenner and F.P. Drijftout (2007). Chemical deterrent enables a socially parasitic ant to invade multiple hosts. Proceedings of the Royal Society: 1-5

Meehan, C.J., E.J. Olson, M.W. Reudink, T.K. Kyser, and R.L. Curry. 2009. Herbivory in a spider through exploitation of an ant-plant mutualism. Current Biology 19(19): 892-893.

Moret, J., J-B. Ferdy, P-H. Gouyon, B. Godelle. 1998. Pollinator behavior and deceptive pollination: learning process and floral evolution. American Naturalist 152: 696-705.

Morris, W.F. 1996. Mutualism denied? Nectar-robbing bumble bees do not reduce female or male success of bluebells. Ecology 77(5): 1451-1462.

Morris, W.F., J.L. Bronstein, W.G. Wilson. 2003. Three-way coexistence in obligate mutualist- exploiter interactions: the potential role of competition. The American Naturalist 161(6): 860- 875.

Nash, D.R., T.D. Als, R. Maile, G.R. Jones, J.J. Boomsma. 2008. A mosaic of chemical coevolution in a large blue butterfly. Science 319: 88-90.

Navarro, L. 2001. Reproductive biology and effect of nectar robbing on fruit production in Macleania bullata (Ericaceae). Plant Ecology 152: 59-65.

Noë, R. and P. Hammerstein. 1994. Biological markets: supply and demand determine the effect of partner choince in cooperation, mutualism, and mating. Behavioral Ecology and Sociobiology 35: 1-11.

O’Dowd, D.J. 1980. Pearl bodies of a neotropical tree, Ochroma pyramidale: ecological implications. American Journal of Botany 67: 543-549.

Ordano, M. and J.F. Ornelas. 2004. Generous-like flowers: nectar production in two epiphytic bromeliads and a meta-analysis of removal effects. Oecologia 140: 495-505.

Pellmyr, O., J. Leebens-Mack, and C.J. Huth. 1996. Non-mutualist yucca moths and their evolutionary consequences. Nature 380: 155-156.

Pellmyr, O. and J. Leebens-Mack. 1999. Forty million years of mutualism: evidence for Eocene origin of the yucca-yucca moth association. Proceedings of the National Academy of Sciences of the USA 96: 9178-9183.

R Development Core Team. 2009. R Foundation for Statistical Computing. Vienna, Austria.

28 Rehr, S.S., P.P. Feeny, D.H. Janzen. 1973. Chemical defense in Central American non-ant Acacias. Journal of Animal Ecology 42: 405-416.

Reiskind, J. 1977. Ant-mimicry in a Panamanian clubionid and salticid spiders (Araneae: Clubionidae, Salticidae). Biotropica 9(1): 1-8.

Richardson, S.C. 2004. Are nectar-robbers mutualists or antagonists. Oecologia 139(2): 246- 254.

Roubik, D.W., N.M. Holbrook, G. Parra V. 1985. Roles of nectar robbers in reproduction of the tropical treelet Quassia amara (Simaroubaceae). Oecologia 66(2): 161-167.

Roubik, D.W. 1982. The ecological impact of nectar-robbing bees and pollinating hummingbirds on a tropical shrub. Ecology 63(2): 354-360.

SAS Institute. 2004. JMP 5.1.2. SAS Institute Inc., Cary, North Carolina, USA.

Sakata, H. 1994. How an ant decides to prey on or to attend aphids. Research on Population Ecology 36: 45-51.

Schwartz, M.W. and J.D. Hoeksema. 1998. Specialization and resource trade: biological market as a model of mutualisms. Ecology 79:1029-1038.

Schaefer, C.W. and R. O’Shea. 1979. Host plants of three Coreine tribes (Hemiptera: Heteroptera: Coreidae). Annals of the Entomological Society of America 72(4): 519-523.

Seidel, J.L., W.W. Epstein, D.W. Davidson. 1990. Neotropical ant gardens I. Chemical constituents. Journal of Chemical Ecology 16: 1791-1816.

Soberon, J. and C. Martinez del Rio. 1985. Cheating and taking advantage in mutualistic associations. Pgs 192-26 in D.H. Boucher, ed. The biology of mutualism: ecology and evolution. Oxford University Press, Oxford.

StatSoft, Inc. 2001. Statistica. StatSoft, Inc., Tulsa, Oklahoma, USA.

Styrsky, J.D. and M.D. Eubanks. 2007. Ecological consequences of interactions between ants and honeydew-producing insects. Proceedings of the Biological Society 274(1607): 151-164.

Suarez, A.V., C. De Moraes, and A. Ippolito. 1998. Defense of Acacia collinsii by and obligate and nonobligate ant species: the significance of approaching vegetation. Biotropica 30(3): 480- 482.

Thomas, J.A. and J. Settele 2004. Evolutionary biology: butterfly mimics of ants. Nature 432: 283.

Thompson, J.N. 1982. Interaction and Coevolution. Wiley and Sons, New York, USA.

Traveset, A., M.F. Willson, C. Sabag. 1998. Effect of nectar-robbing birds on fruit set of Fuchsia magellanica in Tierra del Fuego: a disrupted mutualism. Functional Ecology 12(3): 459-464.

Vander Wall, S.B. 1994. Seed fate pathways of antelope bitterbrush: dispersal by seed-caching yellow pine chipmunks. Ecology 75: 1911-1926.

Wagner, D., M. Tissot, W. Cuevas, D.M. Gordon. 2000. Harvester ants utilize cuticular hydrocarbons in nestmate recognition. Journal of Chemical Ecology 26(10): 2245-2257.

West, S.A. and E.A. Herre. 1994. The ecology of the New World fig-parasitizing wasps Idarnes and implications for the evolution of the fig-pollinator mutualism. Proceedings: Biological Sciences 258(1351): 67-72.

West, S.A., E.A. Herre, D.M. Windsor, P.R.S. Green. 1996. The ecology and evolution of non- pollinating fig wasp communities. Journal of Biogeography 23:447-458.

Whitehead, S., A. Poste, and J. Sapp. 2008. Mozena sp. (Hemiptera: Coreidae) uses chemical cues to exploit ant-acacia mutualism. OTS Course Book: 2008-1.

Wilcox, C.D., S.B. Dove, W.D. McDavid, D.B. Greer. 2002. UTHSCSA Image Tool, version 3.0. University of Texas, San Antonio.

Wickler, W. 1968. Mimicry in plants and animals. McGraw-Hill, New York.

Wyatt, R. 1980. The impact of nectar-robbing ants on the pollination system of Asclepias curassavica. Bulletin of the Torrey Botanical Club 107(1): 24-28.

Yu, D.W. and N.E. Pierce. 1998. A castration parasite of an ant-plant mutualism. Proceedings: Biological Sciences 265(1394): 375-382.

Yu, D.W., H.B. Wilson, N.E. Pierce. 2001. An empirical model of species coexcistence in a spatially structured environment. Ecology 82(6): 1761-1771.

Zhang, Y-W, G.W. Robert, Y. Wang, Y-H Guo. 2007. Nectar robbing of a carpenter bee and its effects on the reproductive fitness of Glechoma longituba (Lamiaceae). Plant Ecology 193: 1- 13.

Zhang, Y-W., Q. Yu, J-M, Zhao, and Y-H Guo. 2009a. Differential effects of nectar robbing by the same bumble-bee species on three sympatric Corydalis species with varied mating systems. Annals of Botany 104:33-39.

Zhang, F., Y. Peng, S.G. Compton, Y. Zhao, and D. Yang. 2009b. Host pollination mode and mutualist pollinator presence: net effect of internally ovipositing parasite in the fig-wasp mutualism. Naturwissenschaften 96: 543-549.

30 VITA

Ellen McGrail Reid, born in 1982 to Kathleen McGrail and Robert Reid, was raised in Rochester, New York, along with her siblings Clare, Noah, and Katherine. When she was three her family moved to a new home with a wilderness as the backyard, igniting a lifelong love of the natural world and a passion for protecting it. Ellen left that wilderness for the urban wilds of Boston to study biology with a specialization in ecology and conservation biology and Spanish at Boston University. The highlight of her studies was participating in the Tropical Ecology Program in Ecuador where she got hooked on rainforests, ecology, and studying abroad. After graduating from Boston University, she spent a summer studying flowering phenology in the forests, fields, and wetlands of Massachusetts with Dr. Richard Primack’s lab group before returning to Ecuador as a teaching assistant for the Tropical Ecology Program in 2005. From there she joined the team at The School for Field Studies’ Center for Rainforest Studies in the Wet Tropics of Australia to work as an intern and student affairs manager. In the late summer of 2007, she moved to Baton Rouge, Louisiana, to pursue her degree in biological sciences at Louisiana State University, joining Dr. Kyle Harms and the Harms Lab crew. In 2008 she studied with the Organization for Tropical Studies in Costa Rica, where she conducted the fieldwork for this thesis. In the future, she plans to bring together her interests in science and international education in her professional life, while continuing to admire and work to protect the natural world.