The Evolution of Caste-Specific Morphology and Colony Structure in Ants

THE EVOLUTION OF CASTE-SPECIFIC MORPHOLOGY AND COLONY STRUCTURE IN ANTS

JO-ANNE C. HOLLEY

DISSERTATION

Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Entomology in the Graduate College of the University of Illinois at Urbana-Champaign, 2015

Urbana, Illinois

Doctoral Committee:

Professor Andrew V. Suarez, Chair, Director of Research Research Assistant Professor Jonathan D. Marcot Professor Gene E. Robinson Professor James B. Whitfield

ABSTRACT

An organism can adapt to its environment physiologically, behaviorally, and morphologically, but biological functions may conflict with each other and therefore constrain their evolution.

Eusocial organisms live in colonies that possess morphologically distinct castes that divide biological roles among individuals. Males and reproductive females fill the reproductive role of the colony which includes mating, dispersal and egg laying. The typically sterile worker caste performs other tasks for the colony such as food collection, defense, and brood care. Workers are therefore released from the morphological constraints of reproduction and can physically adapt to other roles. In addition to exhibiting large degrees of morphological variation, eusocial insect colonies can also vary in colony structure. For example, some species have a single reproductive queen while others can vary in queen number that in turn influences the genetic relatedness of colony members. Furthermore, colonies can be centralized in one nest, or dispersed over several spatially separated nests. Variation in morphology and colony structure within and among species reflects a combination of adaptation to ecological pressures and historical constraints.

My research goals were to investigate intra and interspecific variation in caste specific morphology and colony structure in a comparative framework to understand how traits evolve relative to species’ ecology. In chapter 2, I studied the species rich ant genus Pheidole to understand how worker morphology can adapt to environmental factors. Workers of Pheidole are almost always dimorphic, comprising a small minor subcaste and a major or soldier subcaste with a distinctively large head. Pheidole species vary in their diet, but many specialize on seeds that are milled by majors using large jaws powered by mandible closer muscles that fill the head cavity. I hypothesized that seed harvesting species will have majors with larger heads compared

ii to non-seed-harvesting species to accommodate more powerful mandibular muscles required to mill seeds. I found that majors of seed-associated Pheidole do not have larger heads than majors of non-seed-harvesting species, but there is a positive relationship between seed-harvesting and minor head length. Additionally, there was a difference in the size of minors relative to majors within species; seed-harvesting Pheidole species have smaller minors and larger majors than their non-seed-harvesting congeners. This finding suggests that head size of major and minor subcastes can evolve for diet specialization, but they do so relative to the morphology of other worker castes.

In chapter 3, I used the ant genus Linepithema to examine intra and interspecific variation in morphology and colony structure. Despite our general understanding of caste differentiation, there are few studies quantifying morphological variation within and among reproductive females, workers, and males. I hypothesized males and queens would vary less than workers that are not constrained by dispersal and reproduction. Workers and reproductive females were more similar to each other than they were to males, which had smaller antennal scapes and heads.

Reproductive females were larger overall and males exhibited the largest amount of variation among species. Males of some species also possessed unique morphological features that may reflect adaptations for reproductive strategies suggesting the genus Linepithema deserves further research to understand what factors are driving male specific variation in morphology.

Finally, in chapter 4 I examined the evolution of colony structure in the genus

Linepithema. I estimated within and among population variation in nest number and dispersion

(polydomy), and queen number for colonies of eight species in the genus from Argentina,

Ecuador and Brazil. These were compared to published data for native and introduced populations of L. humile, an invasive species. My research revealed significant variation in nest

iii number among the nine species of Linepithema, however, L. humile was the only species to have colonies with nests dispersed over 250 meters. Polydomy occurs at larger distances in a stepwise fashion within the ‘humile’ clade, suggesting species have incrementally increased the spatial extent of their colonies sequentially rather than via an abrupt change in colony structure in L. humile. Queen number estimates from microsatellite genotyping reveal a similar pattern; species within the ‘humile’ clade typically exhibit multiple queens but vary from an average of just over one per nest for L. gallardoi to over nine per nest in L. oblongum and none possess as many egg- laying queens as seen in L. humile colonies. Together, these findings suggest the evolution of unicoloniality in L. humile involved a shift from limited polydomy and polygyny throughout the genus, to extreme polydomy and polygyny in L. humile via a gradual change in the ‘humile’ species group. Additionally, I compared the data on colony structure in Linepithema to the frequency in which they have been detected in quarantine at ports of entry into the United States

(a measure analogous to propagule pressure). Six species of Lineptihema have been intercepted in commerce arriving to the United States, yet only L. humile has become established outdoors.

Of the remaining intercepted species, none exhibit large degrees of polygyny and polydomy suggesting that opportunity alone is not sufficient for establishment. We did not find records of the two species that exhibit high degrees of polygyny and polydomy L. oblongum and L. micans, suggesting they may pose invasion risks based on their biology, but may not have been commonly transported. Collectively, these results highlight how both species-level characters and opportunity influence species’ ability to invade new habitats.

ACKNOWLEDGEMENTS

I would like to thank my PhD advisor, Andrew Suarez, for his guidance, mentorship and support of my research and professional development. Many thanks as well to my committee members Jonathan Marcot, Gene Robinson, and Jim Whitfield for guiding my dissertation research and providing invaluable feedback. Alex Wild also provided essential support by assisting field collections, fight trials, photographing ants, identifying

Linepithema, and in so many other ways.

Thanks to all my labmates past and present for advice, inspiration, assistance, feedback, and friendship. Ordered by when we overlapped in the lab, thank you to: Fred

Larabee, Bill Wills, Moni Berg-Binder, Dietrich Gotzek, Adrian Smith, Andrea Walker, Selina

Ruzi, Rafael Achury, Josh Gibson, and Kim Drager. Thanks as well to Corrie Moreau and Joe

Laird for their collaboration on the Pheidole research.

Field research in South America depended on the help of many people. Collecting ants in Argentina and Ecuador was assisted by Carolina Paris and Juan Vieira. Many thanks to all the people who helped in Brazil. Thanks to Marcio Pie for sending ants and connecting me with his lab. Thanks for Stella for taking us collecting on the Atlantic Forest, and to Felipe Neves for devoting so much time to my collecting. I am very grateful to Aline

Nondillo for sharing her research and assistance in the field and lab, also thanks to Marcus

Alamança and the Nondillo family for generously hosting us. A big thanks to Ricardo Solar for hosting us in Viçosa, and to him and Júlio Chaul for taking the time to assist us in the field and lab. Thank you in particular to Júlio for providing valuable specimens. Thanks to

Rodrigo Feitosa for hosting us at the Museu de Zoologia da Universidade de São Paulo.

Thanks also goes to Newton Hood for his invaluable assistance crushing many, many ants. Also thanks to Stephanie Tham and Catherine Choi for their help with the molecular component of my research. Thanks also Kim Walden, Michelle Duennes, and Jeff

Lozier for their advice regarding molecular techniques.

I feel fortunate to have had an opportunity to join the University of Illinois

Department of Entomology. Special thanks goes to May Berenbaum for being so generous with her support and time. Thanks also to Hugh Robertson for inviting me to work on the

Linepithema humile genome; to Brian Allan for his teaching mentorship; and to Larry Hanks for challenging teaching opportunities. I am also very fortunate to have strong support from my family, family-in-law, many good friends, and one coach. I am very grateful to every one of them.

TABLE OF CONTENTS

CHAPTER 1: INTRODUCTION.…..…………………………………...………………….…….1

CHAPTER 2: CASTE-SPECIFIC EVOLUTION OF HEAD SIZE IN THE ANT GENUS PHEIDOLE…………………………………………………………………………………….….6

CHAPTER 3: CASTE-SPECIFIC MORPHOLOGICAL VARIATION IN THE ANT GENUS LINEPITHEMA……………………………………………………………………..…....………30

CHAPTER 4: COMPARISON OF COLONY STRUCTURE IN THE ANT GENUS LINEPITHEMA REVEALS THE EVOLUTION OF TRAITS RELATED TO INVASION SUCCESS IN THE ARGENTINE ANT…….……………………………………………..……63

REFERENCES..……………………………………………………………….…………….…..91

vii

CHAPTER 1

INTRODUCTION

Morphological and social variation in ants

Eusocial organisms are characterized by cooperative brood care, overlapping generations, and a reproductive division of labor whereby some individuals mate and produce offspring while others are sterile and forego reproduction entirely (Wilson, 2000). This degree of group organization rarely occurs in animals but it has arisen at least 11 times in the Hymenoptera

(Andersson, 1984; Wilson and Hölldobler, 2005). Eusocial hymenopterans typically possess at least one reproductive female that lays eggs, males that mate with reproductive females, and often have a sterile worker caste that foregoes reproduction to assist the colony (Wilson, 2000).

My research is centered on the strictly eusocial insect family, Formicidae, the ants. In this group all non-parasitic species possess three distinct morphological castes specialized on different roles; males, reproductive females (or gynes), and workers. Reproductive female morphology is adapted for their two primary functions in a colony - dispersal and reproduction

(Wheeler 1910; Hölldobler and Wilson, 1990). As a result they usually have wings and well developed wing muscles for finding mates and dispersal, and developed ovaries for egg production (Peeters, 1997). Male ants participate in mating, but serve no other social role in the colony and are cared for by workers (Hölldobler and Wilson, 1990). Males, therefore, also have well developed sensory organs, wings, and genitalia for the function of finding a reproductive female and mating (Wheeler, 1910; Hölldobler and Wilson, 1990). Finally, workers do not have wings or wing muscles, are sterile or have severely reduced reproductive capabilities and are responsible for most tasks including colony maintenance, brood care and foraging. Castes can also further specialize into morphologically distinct subcastes, which is most commonly seen in

1 workers. Approximately 13% of ant species (16% of ant genera) have a polymorphic worker caste, where worker morphology is associated with task specialization (Wilson, 1953; Oster and

Wilson, 1978). For example, worker morphology can be adapted for foraging and prey capture

(Wilson, 1984; Hölldobler and Wilson, 1990; Ferster et al., 2006; Powell and Franks, 2005;

Powell and Franks, 2006); colony defense (Rettenmeyer, 1963; Hölldobler and Wilson, 1990;

Powell, 2008); food storage (Wilson, 1974; Tsuji, 1990; Lachaud et al., 1992), and task efficiency (Oster and Wilson, 1978, but see Dornhaus, 2008). The incredible variation seen in castes within and among species of ants makes them a particularly good group for evolutionary studies of morphological form and function.

The social organization of colonies also varies among eusocial species. The ancestral condition of all eusocial hymenopteran lineages, including ants, is mostly likely one singly mated queen (Hughes et al. 2008; Boomsma, 2009). This ensures colony members are closely related, which is considered essential for the evolution of eusociality (Hamilton, 1964; Hughes et al., 2008; Boomsma, 2009). Queens, however, can also cofound a colony with other reproductive females or join their natal nest to produce a polygynous colony (Holldobler and Wilson, 1977).

The advantages of polygyny are numerous. Establishing new nests alone is dangerous and the worker force of polygynous colonies can grow more quickly than in monogynous colonies

(Hölldobler and Wilson, 1977, Tschinkel and Howard, 1983). One cost of polygyny however, is reduced relatedness among workers that can result in increased conflict over reproduction

(Trivers and Hare, 1976; Herbers, 1984, Heinze et al., 1994, Keller, 1995). Additionally, social structure among ants can vary depending on the distribution of queens and workers throughout the territory. Ant colonies can live in a single nest (monodomy), or in multiple interconnected nests (polydomy). Polydomy in turn can be related to polygyny (Holldobler and Wilson, 1977;

Keller, 1991; Debout et al., 2007). Some species exhibit substantial variation in colony structure, for example, overwintering in single nests and then dispersing into multiple nests in the summer

(reviewed in Debout et al., 2007). Polydomous colonies may have arisen as a secondary step to accommodate multiple queens (Keller, 1995); or as a safeguard against colony destruction in ecologically unstable environments (Debout et al., 2007). Many abundant and ecologically dominant ant species are often polydomous including most highly invasive ant species (Holway et al., 2002; Debout et al., 2007).

Ants as invasive species

Invasive species cause dramatic ecological changes worldwide. They alter communities, cause species extinctions, and are economically costly (Mack et al., 2000; Gurevitch and Padila, 2004).

Ants are numerically abundant and dominant in many communities (Hölldobler and Wilson,

1990). Five ant species are including in the list of the 100 world’s worst invasive species (Lowe,

2000), more than one third the total number of terrestrial invertebrates included in the list.

Several hundred ant species have been transported globally and have established populations outside their native range (McGlynn 1999a)

Much research into invasive species attempts to identify traits associated with their success (van Keunen et al., 2010). Invasive ants possess a suite of social traits that are thought to help them establish populations in new environments including polygyny, polydomy, and having a large number of small workers (Holway et al., 2002). These traits may allow invasive ants to reach high densities where introduced, making them successful competitors (Holway, 1999) and ecologically destructive (O’dowd et al., 2003). Furthermore invasive ants can exist in unicolonial populations, where all conspecific members of a population recognize each other as nestmates

3 from one colony (Bourke and Franks, 1995). There are thousands of queens in these colonies, worker mix freely between nests and display no aggression (Helenterä et al., 2009). Invasive ants are often at the extreme end of the polygyny and polydomy spectra, and therefore the subjects of many studies on these traits. The argentine ant, Linpithema humile, is an example of a species exhibiting unicoloniality within its introduce range (Tsutsui et al., 2000; Giraud et al. 2002; Van

Wilgenburg et al., 2010). However, the colonization of new habitats by invasive species depends not only the traits than allow them to establish, but also on having an opportunity to reach a new area. As the frequency of introduction events (called propagule pressure) increases, the likelihood that a species will become established also increases (Lockwood et al. 2009;

Simberloff 2009).

The comparative method

Research comparing species requires an understanding of evolutionary history underlying the relationships between them. Species are not statistically independent units; closely related species are more likely to share characters compared to more distantly related species

(Felsenstein, 1985). I incorporate this complication into my analyses when making comparisons among species by using species level phylogenies and employing the comparative method

(detailed further in each chapter) (Felsenstein, 1985; Harvey and Pagel, 1991).

Comparisons between closely realted invasive and non-invasive species can reveal which traits are associated with invasive success while controlling for many other aspects of biology that related species might share (Fisher and Owens, 2004; van Kleunen et al. 2010). This approach, however, has rarely been taken to investigate traits associated with invasive success in ants. One study compared the invasive garden ant, Lasius neglectus, with its sister species L.

4 turcius (Cremer et al. 2008). This comparison highlighted a difference in character traits between the invasive lineage but it did not reveal broader patterns of trait evolution. The research presented in Chapter 4 is the first time the globally invasive Argentine ant, Linepithema humile, has been compared to its non-invasive congeners. This was done to investigate which traits found in L. humile are present throughout the species, and which are isolated in the invasive lineage. This is an important and novel step towards understanding the evolution of traits that confer invasive success in ants.

Overview

In the following chapters I address how ants adapt morphologically and behaviorally to environmental conditions and their roles in the colony. In chapter 2 I examined how worker subcastes of the genus Pheidole are morphologically adapted to diet. This dimorphic species possesses a major and minor sub-caste, so I investigated variation in head morphology of majors and minors varies in relation to diet and in relation to each other. In chapter 3 I compared differences in caste-specific morphology among males, reproductive females and workers of the genus Linepithema. Caste morphology varies based on the biological role of the castes, but it also varies as species adapt to their environment. I also separately compared the morphology of each caste across the genus to quantify morphological diversity in the genus. Finally, in chapter

4, I examine the evolution of polydomy and polygyny in the genus Linepithema. I place this research in the context of the invasive success of the Argentine ant that forms unicolonial populations.

CHAPTER 2

CASTE-SPECIFIC EVOLUTION OF HEAD SIZE IN THE ANT GENUS PHEIDOLE

Abstract

An organism’s morphology is constrained by evolutionary history and the need of the organism to meet a variety of, and potentially competing, functions. The ant genus Pheidole is presently the most species-rich of the 324 described ant genera. Workers of Pheidole are almost always dimorphic, comprising a small minor subcaste and a major or soldier subcaste with a distinctively large head. This separation of workers into two distinct subcastes makes Pheidole an ideal genus to address questions on the evolution of morphology in relation to ecological specialization. Major workers can perform a variety of tasks, but they are usually specialized for defense and the retrieval and processing of food. Pheidole species vary in their diet but many species gather seeds and the majors are known to mill the seeds using the large jaws powered by mandible closer muscles that fill the head cavity. I examined the relationship between seed- harvesting and head size, hypothesizing that seed harvesters have majors with larger heads compared to non-seed-harvesting species to accommodate more powerful mandibular muscles.

By taking a phylogenetically controlled comparative approach, I found that majors of seed- associated Pheidole do not have larger heads (width and length) than do majors of non-seed- harvesting species, but there is a positive relationship between seed-harvesting and minor head length. Additionally, there is a significant relationship of head width and length to seed- harvesting when examining the difference in size of minors relative to majors within species.

Seed-harvesting Pheidole species have smaller minors and larger majors than their non-seed- harvesting congeners. This finding suggests that head size of major and minor subcastes can evolve for diet specialization, but they do so relative to the morphology of other worker castes.

Introduction

An organism’s morphology dictates how it interacts with the abiotic and biotic environment. The size and shape of the head, for example, and its associated feeding structures can have a strong influence on diet (Miles and Ricklefs, 1984; Clifton and Motta, 1998; Wainwright and Richard,

1995; Grant, 1999; Aguirre et al., 2002; Hulsey and Wainwright, 2002; Marshall et al., 2012), and subsequently influence community structure (Karr and James, 1975; Grant, 1999).

Comparative studies on morphology often reveal a strong phylogenetic signal suggestive of conservation of form and function among related taxa. Historical constraints can limit both the amount of variation and the tempo of morphological evolution in lineages (Harvey and Pagel,

1991; Garland et al., 1992) and in turn drive patterns of coexistence and diversification among taxa (Ricklefs and Travis, 1980; Wainwright and Reilly, 1994).

In most animals, morphological variation is constrained by the need to meet a variety of functions, including feeding, mating, and dispersal. In holometabous insects, however, complete metamorphosis allows selection to optimize two separate morphologies. For example, larvae may be specialized for feeding while the adult form may be specialized for mating and dispersal.

Similarly, a reproductive division of labor makes social insects a compelling group to use for testing hypotheses about the evolution of eco-morphology (Powell, 2009; Keller et al., 2014).

Having distinct reproductive and non-reproductive castes may eliminate some morphological constraints or tradeoffs faced by solitary organisms. In many ants, the worker caste is further divided into subcastes based on size and shape. The evolution of worker polymorphism in ants has been hypothesized to be associated with specialization related to foraging and prey capture

(Wilson, 1984; Hölldobler and Wilson, 1990; Ferster et al., 2006; Powell and Franks, 2005;

Powell and Franks, 2006), defense (Rettenmeyer, 1963; Hölldobler and Wilson, 1990; Powell,

2008), food storage (Wilson, 1974; Tsuji, 1990; Lachaud et al., 1992), or task efficiency (Oster and Wilson, 1978).

Whereas most ants are generalist omnivores and scavengers, there are many examples of dietary specialization in ants, including preying on specific insect taxa (Peeters and Crewe, 1987,

Hölldobler and Wilson, 1990; Brandao et al., 1991; Leal and Oliveira, 1995; Dejean and

Evraerts, 1997; Larabee and Suarez, 2014), rearing fungus as food (Mueller et al., 2001), and harvesting seeds (Wheeler, 1910; Davidson, 1977; Brown et al., 1979). Seed harvesting is commonly seen in desert ecosystems where food availability is seasonal (Wheeler, 1910; Brown et al., 1979). Seeds are collected by foraging workers and taken to the nest where seed chaff is removed and the seeds are stored in dedicated nest galleries called granaries. Many seed- harvesting ants have polymorphic workers and worker size is positively correlated with the size of seeds collected and dispersed by ants (Davidson, 1977; Wilson, 1978; Traniello and Beshers,

1991; Kaspari, 1996; Ness et al. 2004; but see Rissing, 1981; Willott et al., 2000). Worker dimorphism in particular often occurs in seed-harvesting species (Hölldobler and Wilson, 1990) and it has been suggested that species with bimodal distributions of head shape have evolved a specialized major subcaste to process seeds for food (e.g. the genus Pheidole, some

Pogonomyrmex, and Solenopsis geminata; Wilson, 1978; Hölldobler and Wilson, 1990;

Tschinkel, 1998; Ferster et al., 2006).

The genus Pheidole is ideal for examining questions related to the evolution of eco- morphology for a number of reasons. First, Pheidole ranks among the most species-rich of the

324 described ant genera, with over 1500 described species (Wilson, 2003, http://antcat.org).

Second, whereas most Pheidole are generalist scavengers and insectivores, seed harvesting has evolved multiple times in this genus (Moreau, 2008) and many are important seed predators. For

8 example, in a Chihuahuan desert ecosystem, Pheidole species remove an estimated 108 seeds/ha/year, ten times what was removed by other seed-harvesting ants (Whitford et al., 1981).

Finally, the genus is characterized by a dimorphic worker caste consisting of small “minors” and larger “majors”, often referred to as “soldiers”. Major and minor workers vary markedly in the size and shape of the head (Figure 2.1). Major workers have enlarged heads compared to minor workers, a trait that has earned this genus the common name of “big-headed” ants. Pheidole majors often specialize in defense, food-handling (including seed husking and milling), and food storage (Whitford et al., 1981; Wilson, 1984; Tsuji, 1990; Detrain and Pasteels, 1992; Lachaud et al., 1992; Dejean et al., 2005; Mertl et al., 2010; Huang, 2010). In contrast, minor workers have a more extensive behavioral repertoire that includes brood care, food collection, storage and distribution, and nest maintenance (Wilson 1976a, Wilson 1976b, Wilson 1984, Brown and

Traniello, 1998). Major task repertoire varies depending on the ratio of minors to majors, with majors performing most tasks occasionally (especially when minors are experimentally removed), and repertoires vary extensively among species (Wilson, 1984; Mertl et al., 2009;

Sempo and Detrain, 2010; McGlynn et al., 2012).

In this study, I examined the eco-morphology of minor and major workers of New World species in the genus Pheidole. I took advantage of a recently published phylogeny of 140

Pheidole species (Moreau, 2008) to ask how major and minor worker head size and shape have evolved in relation to each other and to diet. Given that worker polymorphism has been associated with seed-harvesting in some genera, I hypothesized that seed-harvesting species of

Pheidole will have a more pronounced size difference between major and minor workers.

Furthermore, if the major subcaste is specialized for seed milling, I predicted greater differences in the shapes of the head between majors and minors of seed harvesting species. Mandible force

9 is governed largely by the mandible closer muscle the size and function of which correlates with external head morphology, specifically the width of the head (Paul and Gronenberg, 1999; Paul,

2001). I therefore predicted that a dedicated major subcaste specialized to mill seeds should have musculature optimized for force production and should have correspondingly broader heads.

Finally, because a switch in diet to seed harvesting is considered adaptive in habitats where food availability is unpredictable (Wheeler, 1910). I hypothesized that seed-harvesting lineages will display faster rates of evolutionary change and increased diversification permitted by leveraging a new and abundant food source.

Methods

Species selection, seed harvesting, and morphometrics

I mapped head width and head length measurements of majors and minors onto a phylogeny for

60 species of Pheidole and examined patterns of change within and between castes, and between castes and foraging preferences. Species were chosen because they met three criteria: 1) they were included in the phylogeny of Moreau (2008) so their evolutionary relationships to other species were known; 2) data were available on dietary preferences (seed harvesting presence/absence); and 3) morphometric data were available from Wilson’s (2003) revision of

New World Pheidole. I excluded the two species with a tri-morphic worker because the presence of a super-major may influence the morphology of majors and minors in ways unrelated to diet.

Seed-harvesting data were taken from Moreau (2008), by S. P. Cover (pers. comm.), Hölldobler and Wilson (1990), Johnson (2000), and Wilson (2003).

I used three linear morphometric measurements, two that provide information on head size and shape (head width – HW and head length - HL), and pronotal width (PW) which is the

10 best single predictor of total mass for ants in the subfamily Myrmecinae (r2=0.96, Kaspari and

Weiser 1999). HW, HL, and PW measurements were log-transformed for all size-based analyses.

The difference between major and minor HW and HL (DiffHW, DiffHL) was calculated by subtracting the minor head size from the major head size and log-transforming the result. I also calculated a head shape index by dividing un-transformed HL by HW. Ants with a head shape index of 0 have square-shaped heads, <0 indicates heads that are longer than broad, and >0 indicates ant heads that are broader than long.

Phylogeny

The phylogenetic tree used in the comparative analyses is a maximum likelihood chronogram adapted from Moreau (2008). The molecular phylogeny was inferred in GARLI v0.94 (Zwickl,

2006) and pruned of all species without seed-harvesting data. This pruned maximum likelihood chronogram contains 60 species.

Analyses

I hypothesized that seed-harvesting Pheidole species have larger heads and that seed-harvesting majors have broader heads (larger HW) than generalist species. Associations between HL, HW, and their difference were tested in a phylogenetic framework using analysis of variance

(ANOVA), implemented in R using the phylANOVA function from the phytools package

(Revell, 2012). I also investigated the relationship between PW and seed-harvesting. If head size variation between seed harvesters and non-seed harvesters is representative of overall body size, then PW variation should display the same patterns as HW and HL for majors, minors, and the difference between majors and minors. If, however, the patterns of PW variation differ from HW

11 and HL then it is likely that head morphology may be evolving relative to diet and not tightly correlated to overall body size. Phylogenetic ANOVAs were also conducted on major and minor

PW and the difference in PW between castes relative to seed-harvesting.

To further test whether differences in head size are associated with seed-harvesting, I calculated phylogenetic independent contrasts (PICs) (Felsenstein, 1985; Harvey and Pagel,

1991). In all cases, head size and the head shape index were compared with respect to whether species do or do not harvest seeds. These data met the requirements of the PICs; the Brownian motion model of evolution best fits the data and the contrasts were standardized using the branch length method of Nee (Garland et al., 1992; Purvis, 1995). This was a conservative transformation; no PICs produced unique statistically significant results with this branch transformation, and these results were replicated using alternative branch transformations (not shown). I hypothesized that a change in diet is associated with a larger change in HW, HL, Diff

HW, and Diff HL. Using Mann-Whitney U test performed in R, I compared the contrasts where there was a switch in diets to contrasts when species did not switch diets to test this hypothesis.

I used Pagel’s lambda (Pagel, 1999) to infer the amount of phylogenetic signal in all head size characters. If there is phylogenetic signal on these characters then a change in head size associated with diet is an indication of selection acting on head size. If there is no phylogenetic signal then changes in head size cannot be attributed to adaptation to diet. The lambda parameter

(λ) ranges from 0 to 1 with values close to 0 representing no phylogenetic signal of trait distributions and increasing phylogenetic signal in traits as λ approaches 1. I performed three λ transformations of the phylogenetic tree (λ=0, 0.5, 1) and then fit the character data to the three trees. The best model was inferred by a difference in AICc of 4 or more (Burnham and

Anderson, 2002), and is demonstrated by the highest Akaike weight of the three tests. The

12 evolution of head size was explored by modeling rate shifts across the phylogeny using all six measurements for head size (major and minor HW, major and minor HL, Diff HW, and Diff

HL). This hypothesis were modeled using the transformPhylo.ML function of the R package

MotMot (Thomas and Freckleton, 2011). I analyzed the data with both rate shift models (tm1 and tm2), both identified eight identical relative rate estimates. All statistical analyses were performed using R (version 2.15.2; R Development Core Team), with the exception of the PICs and branch transformations, which were conducted in Mesquite v 2.75 (Madison and Madison

2011).

Results

Phylogenetic ANOVAs

I used Phylogenetic ANOVAs to determine whether large head size was associated with seed- harvesting when controlling for shared ancestry (Figure 2.2, measurements in Table 2.1). Head size of major workers in seed-harvesting species did not differ from head size in non-seed- harvesting species (majorHW: F1,58=0.003, p=0.92, majorHL: F1,58=0.61, p=0.33). Contrary to my predictions, minor head width and head length were both larger in non-seed-harvesting species (minorHW: F1,58=3.35, p=0.029, minorHL: F1,58=9.995, p=0.0001). Therefore, seed- harvesting minors have smaller, broader heads relative to non-harvesting species. The difference between major and minor head length was also significant between seed harvesters and non- harvesters (DiffHL: F1,58=8.39, p=0.002). The calculated DiffHW did not differ statistically between groups (F1,58=2.039, p=0.082).

I investigated pronotal width, a correlate of body mass, to identify whether differences in head size reflect changes in overall body size (Figure 2.3). There was no difference in majorPW

13 between seed harvesters and non-seed harvesters (ANOVA, F1,58=2.045, p=0.096). MinorPWs of seed-harvesting species were smaller than non seed-harvesters (ANOVA, F1,58=2.434, p=0.047).

Finally, the difference between major worker and minor worker PW did not differ statistically

(ANOVA, F1,58=2.5252, p=0.0624).

Phylogenetic Independent Contrasts

I conducted phylogenetic independent contrasts (PIC) to determine if head size varied more greatly between related species that exhibited a diet shift to seed harvesting. There was not a significant relationship between size and diet for HW or HL for either caste (majorHW r2=

0.0151, p=0.342; majorHL: r2=0.042, p=0.111;,minorHW r2=0.0128, p=0.931; minorHL r2=0.0011, p=0.794) (Figure 2.4), indicating that related species do not undergo a rapid change in head morphology with a diet shift. In contrast, the differences between major and minor head width and head length did show a relationship with change in diet. DiffHW (r2=0.0139, p=0.013) and DiffHL (r2=0.0858, p=0.021) were both more pronounced among seed-harvesting species.

I conducted Mann-Whitney U-tests to determine if independent contrasts between related species that had undergone a diet switch statistically differed from related species that shared diet preferences. These results supported the PICs (Table 2.2). A change in diet is associated with a larger difference in head size between castes, but major and minor head size individually did not change size when species change diets.

I also predicted that seed harvesters should have broader heads, as manifested by a size index >1. A PIC of majors revealed no relationship between head shape and diet (r2=0.0021, p=0.72), and the same applied to minor worker head shape (r2=0.0344, p=0.15) (Figure 2.5).

Head Size Rate Shifts

All Pheidole head size characters best fit the λ=1 transformed tree (Table 2.3). This indicates the inferred phylogeny predicts trait distributions and there is strong phylogenetic signal in the head morphology. Of the seven estimated rate shifts, all were faster than the modeled background rate of change with one exception, P. nitella (node 6). Node 1 therefore indicates a rate shift at a node including almost all taxa, likely to account for the slow rate of change of P. nitella. Several species showed a slower rate of change than most species (but faster than the baseline rate of change); the P. tristicula and P. rhinoceros clade (node 3). Two individual species showed accelerated rates of change: the seed harvester P. pilifera (node 5), and the non-seed harvesting

P. astur (node 2), and two clades stand out as evolving faster. These are denoted by node 4, a group of seed harvesters and non-harvesters, and node 6 a clade of three desert seed harvesters.

The clade indicated at node 6 includes P. barbata, P. militicida, and P. psamophilla, all possess workers with a particularly large difference in the head size of majors and minors. I repeated the phylogenetic ANOVA and PICs without this clade and obtained the same results, so this clade contributes to our main finding but does not drive it. There was no overall relationship between diet and rate of evolution; rate shifts did not occur as Pheidole species switched to or from seed harvesting.

Discussion

Understanding how size and shape evolves among related taxa, and how those changes may be driven by foraging ecology, has been a longstanding goal of evolutionary ecologists (Grant,

1999; Richmond and Price, 2002; Wainwright and Richard, 1995; Wainwright and Reilly, 1996;

Dumont et al., 2012). Eusocial insects are a particularly interesting group for such analysis

15 because individuals in the group are divided into morphologically distinct castes. In approximately 13% of ant species workers are further subdivided into morphologically variable subcastes (Wilson, 1953; Oster and Wilson, 1978), although this is rare among other eusocial

Hymenoptera. The evolution of distinct subcastes allows form and function to diverge among sisters in the same colony (Powell, 2009).

I mapped characters related to head size, head shape, pronotal width (a surrogate for body size), and diet onto a phylogeny of the dimorphic ant genus Pheidole to understand how shape evolves in relation to specialization for seed harvesting in a genus with two distinct worker castes. I found that head width and length of major and minor workers show a strong phylogenetic signal (Table 2.3), and that the difference in head size between the subcastes is related to diet. Specifically, seed-harvesting species have a more pronounced size difference in head size between the major and minor workers. In contrast to my predictions, however, head sizes of majors did not increase in seed-harvesting lineages, and minor workers of seed- harvesters have smaller, broader heads than minors of non-seed-harvesters. I also examined if related species that differed in diet had a larger difference in head-size compared to related species that do not differ in dietary habits. I did not see a positive slope in the PICs as predicted if seed harvesting species had larger head sizes then related non-seed harvesting species, but there was a positive relationship between diet shifts and the difference in head size between castes (Figure 2.4).

I expected seed-harvesting species to have majors with broader heads than non-seed harvesting species. This prediction arose from two observations. First, majors have been reported to be a specialized seed milling caste in some Pheidole (Wheeler, 1910). Second, mandibular force production is controlled by parallel muscle fibers attached to the apodeme by filaments in

16 the head capsule, and is correlated with the broadness of the head (Paul and Gronenberg, 1999).

In contrast, Wheeler (1910) argued that the head morphology of seed-harvesting ants is similar to insectivorous species, because the musculature required for crushing seeds was also needed for carving up insect exoskeletons. The finding of no relationship between head shape of majors and seed harvesting (Figure 2.2 and Figure 2.4) concurs with Wheeler’s conclusion. I only investigated head width and length, however, so it remains possible that differences between seed-harvesters and generalists lay in other properties such as mandible size and shape. For example, older seed-milling individuals often have blunt mandibles largely free of teeth, worn from heavy use.

A more detailed examination of head shape could reveal that seed-milling majors have a different head shape compared to related non-milling species in more subtle ways (Ferster et al.,

2006). The unrelated seed-harvesting ants Pogonomyrmex badius and Solenopsis geminata both have larger occipital regions of the head (above the eyes), and larger head width, when compared allometrically to seven other head and body measurements (Ferster et al., 2006; Tschinkel,

2013). There may be further differences between seed harvesting and non-harvesting Pheidole species if more head measurements are examined, or if geometric morphometrics to model head shape were conducted in a phylogenetic context (Zelditch et al., 2012). Additionally, the seed- harvesting/non-harvesting dichotomy may by oversimplified. Pheidole species may put seeds to different uses; perhaps not all supposed “seed-harvesting” Pheidole majors are milling seeds instead workers collect them for the lipid rich elaiosomes exploited by many ant species

(Sernander, 1906). Seed-harvesting species, however, are concentrated in regions without rich diversity of eliaosome producing trees (Lengyel et al., 2009), suggesting the eliaosomes are not driving seed collection (Beattie 1985), but it is unknown how Pheidole process eliaosome-

17 bearing seeds they collect (Hughes et al., 1994; Cuautle et al. 2005). Thus, further studies on seed use in Pheidole will benefit greatly from direct observation of seed milling.

The difference in size between major and minor workers was greater in the seed harvesting species of Pheidole (Figure 2.2 and Figure 2.4). Dimorphic worker subcastes have been associated with a dietary switch to seed harvesting in P. badius and S. geminata, as their non-harvesting relatives are monomorphic (Ferster et al., 2006). In these species the major subcaste specializes on seed milling to supplement other food sources collected by the minor workers (Ferster et al., 2006; Holldobler and Wilson, 1990). There was also a faster relative rate of head size evolution in a clade of seed harvesting, desert-dwelling Pheidole, all of which possess particularly large majors and small minors (Figure 2.6: clade 6: P. barbata, P. militicida, and P. psammophila). These findings may indicate a more pronounced division of labor between seed-milling majors and minors. A large difference in body size between majors and minors was also observed in ground dwelling Pheidole species in the tropics (Mertl et al., 2010), which was suggested to be associated with defense. An alternative explanation is there is a relationship between seed-harvesting and living in the soil. Seed-harvesting may be difficult in twig-nesting and canopy species; it could be difficult for these ants to produce chambers for storage and to keep them dry.

Among the ants, several genera display subcaste variation as an adaptation to food processing. Attine leaf-cutting ants have continuously polymorphic workers that cut or masticate leaf fragments based on size (Wilson, 1980), and Powell (2006) demonstrated the submajor worker subcaste was present in three army ant species that handle “awkwardly shaped” prey.

Dimorphic worker subcastes are present in seven ant genera (Oster and Wilson, 1978) and are often associated with nest defense (eg. Colobopsis and Cephalotes, Powell, 2008). Using a

18 phylogeny spanning nearly all ant genera (Moreau et al., 2006), Pie and Tschá (2013) found that size evolves much more quickly than shape among ant genera. In contrast, within the genus

Pheidole, they concluded there was a decelerating rate of change in body size in major workers of the seed-harvesting desert-dwelling clade (Pie and Tschá, 2013). In the study, relative rate of head size change was faster than the baseline rate of change in all taxa but one (P. nitella).

The rate of head size evolution was not associated with diet; head size of seed harvesting species had the same estimated rate of change as non-harvesting species in most cases. This suggests diet switching does not lead to an increased rate of morphological change and refutes the hypothesis that switching to seed-harvesting allowed Pheidole species to diversify rapidly with access to a new food source. Therefore changes in head morphology associated with seed- harvesting are possibly the result of selection. Pie and Traniello (2006) found morphological variation among Pheidole species was driven largely by differences in size (size explained up to

83% and 78% percent of the variation among majors and minors respectively in a PC analysis) and not by large changes in shape. Within species, the strongest differences between castes were associated with variation in head width, head length and thorax length (Pie and Traniello, 2007).

The reason for Pheidole’s hyperdiversity is unknown, but Wilson (2003) suggested it may result from having a dimorphic worker subcaste. In particular, the ubiquitous big-headed major allowed Pheidole species to become ecologically dominant; their head morphology enables species to fill a variety of specialized roles such as defense, guarding, and food-gathering while the minor caste retains a generalized function, resisting morphological change (Wilson,

2003; Pie and Traniello, 2007; Mertl and Traniello, 2009; Mertl et al., 2010). Alternately,

Pheidole may have adapted behaviorally rather than morphologically leading to niche differentiation and subsequent speciation with little morphological change (Mertl et al., 2010).

Furthermore, the proportion of major workers to minors varies from species to species, seasonally, and in response to resource availability, adding yet another means of restructuring the worker force (McGlynn et al., 2012). Essentially, Pheidole may meet new ecological challenges by drawing from existing castes, or altering their relative representation. These characteristics allow the genus to maintain a flexible worker force that can adjust to different habitats and food sources quickly, without relying on conspicuous morphological evolution. These results indicate that Pheidole morphology evolves more readily by size, but not shape. It also highlights the importance of size differences between worker subcastes, which may vary according to diet and potentially other factors. While this does not explain Pheidole’s hyperdiversity, it lends further support for why a major subcaste may contribute to the ecological success of this genus.

Tables

Table 2.1: Mean head size (mm) ± standard error of the mean (SEM) for each species of Pheidole based on diet. Head measurements for each species were log transformed prior to calculating the mean.

Table 2.2: Results of the Mann-Whitney U test comparing independent contrasts of related species that do have a diet switch compared to those that do not have a diet switch (data from Figure 2.4). Asterisks denote a significant difference between the two diet groups.

Table 2.3: Fit of Lambda (λ) transformed phylogenetic trees to Pheidole head size characters and the difference between majors and minors (Pagel 1999). The lower the AICc score and the higher Akaike weight indicate the best fit of the three models.

Figures

Figure 2.1: Pheidole metallescens, a non-seed-harvesting species, major worker (left) and minor worker (right). Images © Alex Wild.

Figure 2.2: Mean head width (HW) and head length (HL) of major workers and minor workers and the difference between major and minor head width (difference HW) and head length (Difference HL). White represent seed harvesting species, grey represents non-seed harvesting species, bars represent the SEM. Results of the phylogenetic ANOVA are indicated by asterisks denoting statistical significance. *p=0.05 **p=0.01, ***p<0.001.

Figure 2.3: Mean pronotal width (PW) of major workers and minor workers and the difference between major and minor pronotal width (difference PW). White represent seed-harvesting species, grey represents non seed-harvesting species, bars represent the SEM. Asterisks denote statistical significance derived from phylogenetic ANOVA. *p=0.05.

Figure 2.4: Phylogenetic independent contrast (PIC) regressions of worker head size and diet (seed harvesting or not). When change in diet is >0 it indicates a switch to or from seed harvesting, when 0 indicates related species remained as seed harvesters or non-seed harvesters. (a) Major head width, (b) minor head width, (c) major head length, (d) minor head length, (e) the difference between major and minor head width, (f) the difference between major and minor head length. White markers indicate contrasts where related species changed diet, gray markers indicate no diet change. Triangles represent head width contrasts, circles represent head length contrasts.

Figure 2.5: Phylogenetic independent contrast (PIC) regressions of worker head shape and diet (seed harvesting or not). White markers indicate contrasts where related species changed diet, gray markers indicate no diet change. (a) Major head shape, (b) minor head shape.

Figure 2.6: Head size rate shifts in Pheidole. Branch lengths are scaled to the maximum likelihood relative rate estimates for change in head size (rates listed in the inset table). Node numbers denote where the seven rate shifts were observed. Red names indicate seed harvesting species.

CHAPTER 3

CASTE-SPECIFIC MORPHOLOGICAL VARIATION IN THE ANT GENUS LINEPITHEMA

Abstract

The morphology of an organism is shaped by biological functions such as foraging, food processing, and mating, as well as by ecological factors including habitat structure and climate.

Obligately social organisms may divide labor among members of the group enabling individuals to specialize behaviorally or morphologically to suit their tasks. Notably, eusocial animals possess a pronounced reproductive division of labor, where males and reproductive females mate, and the worker caste is nearly always sterile. Workers are freed from the morphological constraint of reproduction and can physically adapt to other roles in the colony such as food collection, defense, and brood care. The reproductive castes, in contrast, possess morphological adaptations related to dispersal and mating. Despite a general understanding of morphological differentiation among castes, there are few studies detailing how shared morphological features vary among all three castes in an evolutionary context. I use the recently revised Neotropical genus Linepithema, to quantify morphological variation within and among the three castes for each species in the genus. I tested the hypothesis that males and queens will exhibit less morphological variation in non-reproductive characters than workers that are not constrained by reproduction and dispersal. Contrary to my prediction, variation was greatest among males.

Workers and reproductive females are more similar to each other proportionately than they are to males, which have smaller antennal scapes and heads. Some species possess unique caste- specific morphological features that may reflect adaptations for reproductive strategies. Future

30 work on Linepithema is needed to elucidate how sexual selection and natural selection have acted to shape variation in morphology among males in this genus.

Introduction

Understanding the ecological and behavioral significance of morphological variation is a cornerstone of comparative evolutionary biology (Schluter, 2000). Morphological diversification is predicted to promote speciation, allow co-existence among related species, and is generally assumed to result from natural selection in response to ecological variation (Brown and Wilson,

1956; Pfennig and Pfennig, 2009; Burbrink and Pyron, 2010; Mahler et al., 2010; Martin and

Wainwright, 2011). Molecular phylogenetic approaches coupled with multivariate morphometric analyses, have been essential for illuminating patterns of morphological variation among species

(Adams et al., 2009). However, less attention has been paid in the role of intraspecific diversification how it relates to variation among species (e.g. Bolnik et al., 2003; Fjerdingstad and Crozier, 2006).

Most examinations of intraspecific variation in morphology concentrate on differences related to sex or age (Bolnick et al., 2003). Sex-specific variation in particular has received considerable attention in comparative studies (Shine, 1989). Consistent morphological differences between the sexes can result form sexual selection and the elaboration of secondary sexual characteristics (Darwin, 1874), can reduce intra-specific competition for resources

(Selander, 1966; Schoener, 1967; Shine, 1989), and may be driven by sex specific reproductive or dispersal strategies (Greenwood, 1980; Harrison, 1980; Pusey, 1987).

In obligately social species, the occurrence of non-reproductive individuals in groups can further select for morphological variation within sexes. This is most extreme in the eusocial

31 insects where females are separated into reproductive and non-reproductive castes. Each caste performs different tasks in a colony and their morphology reflects these roles (Oster and Wilson,

1978). Reproductive females (alates and queens, also called gynes) are responsible for mating, establishing a nest (often through winged dispersal), and laying eggs (Oster and Wilson, 1978).

They may start new colonies independently or with the aid or workers and independent colony founders may or may not forage to provision their first batch of larvae (Keller and Passera,

1989). Queens that actively forage have enlarged first thoracic segments and corresponding muscles to facilitate prey capture (Keller et al., 2014). Workers, in contrast, perform a broad range of tasks associated with nest building and maintenance, brood and queen care, food collection and distribution, and defense.

The majority of research on morphological diversity in ants focuses on workers, particularly in species that exhibit caste diversity and worker polymorphism (Oster and Wilson,

1978; Fjerdingstad and Crozier, 2006; Powell, 2009). In contrast, studies of variation in queen and male morphology are rare outside the taxonomic literature. Reproductive features such as male genitalia are often highly specialized, giving them strong value for differentiating closely related species in many taxa (Hosken and Stockley, 2004). Reproductive female morphology represents their two primary roles in a colony: 1) developed wings and wing muscles to disperse, the muscles are later metabolized to provision the first brood; 2) developed ovaries and correspondingly large abdomens for egg production (Peeters, 1997). Male ants have well developed sensory organs, wings, and genitalia for the function of finding a reproductive female and mating (Wheeler, 1910; Hölldobler and Wilson, 1990). Males are produced seasonally and are not easily collected; they spend little time outside of the nest, males exit the nest to mate during flights, then die following copulation. Furthermore, males are collected away from the

32 nest are often unidentifiable (Shik et al. 2013), further hindering our understanding of this caste.

Interspecific morphological variation of males can be interpreted two ways. Wheeler (1910) argued that male morphology is the most stable of the three castes. However, male genitalia is often highly specialized and species-specific making it an important taxonomic tool (Boudinot,

2013). Research on males of many animal species indicates that phenotypic variation in males can result from alternative reproductive strategies (Gross, 1996). Surprisingly for such a common group, variation among male ants remains largely ignored. Male ant morphology has been studied extensively only in a handful of cases (reviewed in Shik et al., 2013), including the fighting male subcastes of Cardiocondyla and Hypoponera (Heinze and Hölldobler, 1993;

Heinze et al. 2005; Kinomura and Yamauchi, 1987) and the queen-like males of Dorylus army ants (Franks and Hölldobler, 1987).

The genus Linepithema is an interesting group for investigating caste differences because, at least superficially, workers from all species are monomorphic and look similar enough to have confounded accurate identification for decades (Wild, 2004). Linepithema workers range from 2-

3 mm in length, and vary in color from yellow to brown/black. Even if subtle, morphological differences between species undoubtedly reflect ecological differences. For example, while many

Linepithema are habitat generalists, some are primarily arboreal (L. iniquum and L. leucomelas), while others are thought to be largely subterranean (L. cryptobioticum) (Wild, 2009). Species also vary in their reproductive mode and colony structure (Chapter 4). For example, in L. humile queens are flightless, colonies are extremely polydomous and each nest can contain a large number (>35) reproductive queens (Suarez et. al, 1999; Krieger and Keller, 2000; Pedersen et al., 2006). In contrast, colonies of L. gallardoi are nearly always monodomous and single queened (Wild, 2007, Chapter 4). Here we describe intra- and inter-specific differences in

33 morphology among the three castes of 18 Linepithema species and consider these changes in an evolutionary context.

This research has two goals: first, to describe and quantify variation within and among the three castes for each species in the genus Linepithema. Comparisons between reproductive castes and workers typically focus on genitalia and thorax morphology due to the presence of wings and associated structures (Wheeler, 1910; Hölldobler and Wilson, 1990). I focus on other characters such as limbs, antennae, head size, eyes and limbs, which are shared among the three castes. Second, I describe differences in caste morphology among Linepithema species, accounting for evolutionary history, while assessing differences in light of species’ ecology. The

18 species included in this study live in a variety of habitats and niches throughout the

Neotropics. Given constraints related to dispersal and reproductive biology, I test the hypothesis that morphology will vary less among males and queens in the genus, relative to workers that are not constrained by these factors.

Methods

Study organisms

The genus Linepithema contains 19 described species in 6 well defined clades endemic to the

New World that range from northern Mexico and the Caribbean south to central Argentina

(Figure 3.1)(Wild 2007). Linepithema spp. inhabit a variety of ecosystems and niches within these different habitats. For example L. pulex nests in rotting wood and litter in the moist subtropical Atlantic forests of Brazil, while its likely sister species L. aztecoides inhabits the much drier Cerrado. Linepithema iniquum and L. leucomelas are exclusively arboreal.

Linepithema spp. also vary in their abundance. Many species are quite common where they occur

(e.g. the invasive Argentine ant, L. humile), while others are rare or occur at low densities throughout their range. Additionally, not all castes are easily accessible; ephemeral males and alates are produced only at certain times of the year and queens can be difficult to sample when nests are deep.

Morphometrics

Body measurements were taken on workers, reproductive females (queens and alates), and males using a MicroDRO three axis stage micrometer and a Leica 10x microscope. A diagram of the measurements taken for each caste is displayed in Figure 3.2 and the names of the measurements and their abbreviations are listed in Table 3.1. All measurements were log transformed to base

10, and the mean value for measurement for each species and number of specimens is presented in Table 3.2. Most measurements included in this study were taken from the taxonomic revision of the genus (Wild, 2007). Additional specimens were collected by locating and excavating nests during a series of expeditions to Argentina, Brazil, and Ecuador in 2009-2012.

Phylogeny

The ultrametric chronogram used for phylogenetic least squares was inferred using Bayesian analysis in BEAST 1.0.2 (Drummond et al., 2012). Molecular data from four loci were taken from a recent molecular phylogeny (Wild, 2009) and analyzed in a four partition scheme: ITS-2, nuclear protein-coding genes (Wingless, Long-wave Rhodopsin), mitochondrial COI codon positions 1 and 2, and mitochondrial COI codon position 3. Additionally, two species from the

Dolichoderine genera Technomyrmex and Tapinoma were included as outgroups. A recent subfamilial phylogeny by Ward et al. (2010) estimated the Linepithema crown clade at 8-14

35 myo, allowing rough calibration of the phylogeny. The consensus chronogram was calculated from a 40 million generation MCMC run, sampled every 100,000 generations, with a pre- convergence burn-in of 1 million generations removed. The topology was congruent with earlier analyses by Wild (2009), and the consensus chronogram was pruned so that only one tip remained for each species included in the morphometric data set.

Analyses

I conducted linear discriminant analysis to identify how the three castes differ from one another morphologically. The data were grouped by caste and produced two linear discriminants. This was conducted on the six morphological characters common to the three castes (HL, HW, SL,

EL, FL, LHT). This was done using the Mass package in R (Venables and Ripley, 2002).

Bartlett’s test for homogeneity of variances was used to compare the morphological variation among castes using the car package (Fox and Weisberg, 2011). Statistical comparisons between the castes were conducted using ANOVAs between each pair of morphological variables with caste as an additional predictor variable. This test was used even though the variances are not homogenious, Howell (2007) advises that ANOVAs are robust to this as long as the variances do not vary by more than four, which they did not. Tukey’s HSD test was used to identify differences among the castes. P-values were adjusted using the sequential Bonferroni approach

(Holm, 1979) to address the problem family-wise errors when conducting multiple comparisons.

Principal Component Analyses (PCA) were used to examine the variance among species for each caste separately, and to identify which characters are responsible for these differences. I included the six measurements in common for all castes with the data grouped by caste. Further

PCAs were conducted to compare how morphological differences among species differed for

36 each caste separately with the ants grouped by species. For this analysis, the morphometric measurements used differed for each caste (Table 3.2). All PCAs were conducted in R using the prcomp function in the basic R statistics (Venagbles and Ripley, 2002). The worker PCA included HL, HW, SL, MFC, EL, EW, FL, LHT, PW, and PDH; the reproductive female PCA included HL, HW, SL, EL, EW, FL, LHT, and Weber; and the male PCA included HL, HW, SL,

EL, FL, LHT, MML, MMW, and PH (Table 3.2).

I investigated the relationships between each morphometric measurement for each of the three castes using phylogenetic general least squares (PGLS) implemented in R using the Ape package (Paradis et al., 2004; Revell, 2010). This statistical test accounts for species non- independence by incorporating a parameter of phylogenetic signal, Lambda (Pagel, 1999), into the model and an estimate of Lambda is provided with the PGLS output (Grafen, 1989; Revell,

2010). Lambda (λ) has a range of 0 to 1 with values of 0 indicating a character has no phylogenetic signal; therefore the trait is supposed to be evolving independent of evolutionary relationships. Values approaching 1 indicate a strong phylogenetic signal. In some instances the

PGLS returned a very high Lambda value (>1), which led to unrealistically high t-statistics and resulting p values (e.g., t-values > 500). To provide realistic results, I constrained Lambda to 1.

To ensure the results were appropriate I also employed simple linear models as a comparison, these models always yielded very similar results. I only reported the PGLS values, because incorporating phylogenetic signal into regressions is considered more reliable when comparing closely related species (Revell, 2010). In some instances the Lambda values were very low, which also resulted in very high t-statistics. In this case I present both the PGLS and general linear model results. Multiple comparisons were conducted, so the sequential Bonferroni

37 adjustment was applied to the P values (Holm, 1979). All statistical analyses were performed using R (version 2.15.2; R Development Core Team).

Results

Differences among castes

The results of the LDA support our hypothesis that castes vary morphologically (Figure 3.3).

Workers and reproductive females cluster together in the first linear discriminant (LD1), and the males group together with the exception of three individuals of one species that are more similar to the female castes. LD2 separates the reproductive females from the males and workers. Males differ from both female castes as they have much shorter antennae and much larger eyes; the means of each measurement for each caste are presented in Table 3.3. The three male individuals clustering closer to the females along LD1 belong to L. dispertitum and are addressed below.

Reproductive females are larger than males and workers for all characters, and workers have shorter legs than males but with larger heads and antennae. Furthermore males display more morphological variability than reproductive females and workers for each of the six morphometric measurements (Figure 3.4). Bartlett’s tests confirmed there was a statistically significant difference in variance of each measurement among the three castes (Table 3.4).

Comparisons of pairs of morphological characters reveal further differences when grouped by caste (Figure 3.5). All relationships between pairs of characters are highly significant

(p<0.0001 for each pair of measurements, results in Table 3.5), indicating that allometric relationships are somewhat consistent among species. Tukey’s HSD tests for each pairwise correlation revealed most relationships between characters differed among the castes. There were a few exceptions; reproductive females and workers had similar relationships between head

38 length (HL) and head width (HW), and head width with femur length (FL), eye length (EL), and scape length (SL). Males and workers had similar head length (HL) and scape length (SL) measurements, while males and reproductive females had similar eye (EL) and femur lengths

(FL). Overall, workers and reproductive females had similar relationships between morphological characters, but the workers are smaller. Males differ from the female castes by having disproportionately small scapes, even relative to their small head size and large eyes which are more similar in size to reproductive females (but still variable among species). Three male L. dispertitum individuals sit as outliers for their unusually long, worker-like scapes.

Differences within caste

Principal component analyses reveal within caste morphological differences among the species groups (Figure 3.6). PC1 was comprised of all characters in similar amounts for each caste and separated individuals by general body size. Among the workers, for example, the small L. pulex is separated to the left of the PC1 axis. PC2 composition differed for each caste. The worker PC2 was comprised mostly of eye size (EW and EL), so the ‘humile’ group, for example, separates because they have large eyes and slightly large body size compared to other Linepithema species.

Species in the ‘fuscum’ group have smaller eyes for their body size, particularly L. keiteli, which sits at the lower edge of PC2. Body size along the PC axis is reversed in the reproductive female figure; small individuals fall to the right (for example, the ‘iniquum’ group and lone L. pulex point), and the larger ‘fuscum’ group queens sit to the left. PC2 is comprised largely of eye characters (EW and EL), leg characters (FL and LHT), and scape length (SL), which further separate the ‘iniquum’ group. Overall, however, few morphological differences separate the reproductive females of each species. Finally, the PCA of male morphology shows that the males

39 fall into several discrete groups. PC1 is overall body size, running small to large from left to right along the x axis, which separates the ‘fuscum’ group which has generally large males, the

‘humile’ group is also drawn to the right by L. humile which has unusually proportioned males

(see below). Scape length (SL) and mesosoma width (MMW) are the major contributors to PC2.

This again separates L. dispertitum (long scapes), and therefore the ‘iniquum’ group. The

‘humile’ group is drawn further along PC2 because L. humile males have disproportionately broad mesosomas (MMW). The variation within the male caste is partially represented in Figure

3.1, with diagrams of L. humile. L. micans, L. dispertitum, and L. tsachila males (it is worth noting that the L. disperitum male does represent individuals with elongate antennae). These are mapped on the phylogeny and color-coded the same as the clades in Figure 3.6.

Each morphological character was compared to all others for workers, reproductive females and males independently. I used phylogenetic general least squares (PGLS) and incorporated phylogenetic signal into the model by estimating lambda (Table 3.6). Paired PGLS produced a lot of statistically significant relationships between characters in workers (Figure

3.7). Propodeal height (PDH) does not form statistically significant correlations with any character, while the minimum frontal carina (MFC) only formed correlations with head length

(HL) and head width (HW). Additionally, pronotal width (PW) did not correlate with scape length (SL) or the two eye characters (EW and EL). It is worth noting that these characters are not easy to measure, so this may reflect error in the data. Length characters, however, have strong associations with each other; head length (HL), scape length (SL), femur length (FL), and hind tibia length (LHT). Measurements on the same part of the body, such as the two head or two leg measurements, have the tightest allometric relationships.

The paired PGLS of the reproductive females was unique because all regressions were significant, even after the conservative adjustment of the p-value (Figure 3.8). There are the same relationships between body parts as for the workers: length characters (HL, SL, EL, FL, and LHT) tend to have strong correlations, along with characters on the same body part (HL and

HW, EW and EL, FL and LTH)(Table 3.6). Eye size characters (EW and WL) varied most compared to other morphological traits. Male morphology was more variable than that of reproductive females (Figure 3.9). Again I found tight associations between length related characters (HL, SL, FL, and LHT)(Table 3.6). Male characters show more significant correlations than in workers, however, the correlations are not as tight as those found in the reproductive females, and more are outliers. The outliers are based on the previous LDA and

PCAs: the mean length of L. dispertitum first antennal segment (SL) is particularly long, and the mesosoma of L. humile males is notably broad and long.

Discussion

The size and shape of morphological structures often reflect adaptation to environmental factors such as climate, habitat structure, and nesting sites, but also behavioral factors such as mate choice and reproduction (Wainwright, 1996). Highly eusocial insects are distinct because colonies possess a sterile worker caste, so their morphology is driven by the tasks they perform for the colony and are unconstrained by reproduction (Oster and Wilson, 1978). I compared six shared morphological features of males, reproductive females, and workers to each other across the genus Linepithema and hypothesized reproductive females and males would show lower amounts of morphological variation relative to workers. In contrast to my prediction, male ants

41 displayed the most amount of morphological diversity, while reproductive females and workers displayed less variation among species.

Of the morphological features shared among the three castes, some were proportionally similar among castes, but varied in overall size (such as head and leg characteristics). Strong correlations like these suggest morphological integration; the developmental pathways of some features are likely preserved in the three castes (Klingenberg, 2008; Tschinkel, 2013). Other traits were caste-specific, suggesting these traits are associated with their roles in the colony.

Males possessed smaller antennal scapes perhaps because they do not use their antennae extensively while in the nest. They have large eyes, as do reproductive females, which are like used to navigate during mating flights. Gynes had large bodies overall, while workers had similar body proportions but were smaller.

Male ants were the most morphologically divergent caste among Linepithema species.

This pattern was unexpected given how difficult it is to identify species based on males alone

(Wheeler, 1910; Hölldobler and Wilson, 1990). Within-caste variation in size can be attributed to environmental conditions (e.g. larval nutrition), which changes over the lifespan of a colony

(Wilson, 1985). For example, Pogonomyrmex males vary in size depending on the age and size of the colony; larger colonies produce larger males (Smith and Tschinkel, 2006). Solenopsis invicta males also vary in size seasonally, with the smallest males produced in March and the largest in June (Tschinkel, 1993). Sociogenesis may therefore account for some of the variation we see among males of Linepithema, however, we would expect that similar changes would occur in females castes as well (Tschinkel 1993).

The interspecific comparison of males revealed Linepithema vary in three distinct ways:

1) males of the ‘fuscum’ clade are distinctly larger and possess external genitalia and female

42 wing venation (Wild, 2007). I hypothesize that male morphology in this group may be adapted for direct competition for mates, as has been observed in solitary insects. For example, large body size has been correlated with greater sperm transfer in Pogonomyrmex ants (Wiernasz et al., 2001), and selection for size has been observed in yellow dung flies (Borgia, 1980), and for elaborate genital morphology in dung beetles (House and Simmons, 2003). 2) L. dispertitum males show striking intraspecific variation in antennal length and head morphology that, in some populations, more closely resembles that of workers. In this instance, the variation in morphology is abrupt, with males displaying a variety of worker traits, even while retaining wings. This suggests expression changes in genes of large effect, perhaps associated with the worker developmental pathway. As male behavior has never been observed in L. dispertitum, the functional or ecological significance of such variation is unknown. Finally, 3) L. humile males deviate from isometric relationships because they have particularly large mesosomas compared to the wing length and profemur length congeners. The enlarged mesosoma of the L. humile male may result from the attachment of enlarged wing musculature, which in turn may enhance flight capabilities. Males of L. humile may have evolved a morphological adaptation corresponding to queen flightlessness, a character thought to be unique to this species (Wild, 2007). Linepithema humile females mate within the nest (Passera et al. 1988) and establish new nests by budding

(Holway et al. 2002). Males breeding outside their natal nest must fly to a new location where workers will drag them into the nest to mate with receptive alates (unmated reproductive females) (Passera and Keller, 1994). Increased flight capabilities in L. humile males would encourage outbreeding between colonies. On the other hand, genetic data suggest that there is little to no male mediated gene flow among separate supercolonies in L. humile (Pedersen et al.

2006, Thomas et al. 2006). Overall the findings of morphological diversity suggest research on

43 male social insects will be a rich area for discovery, as has been noted by other authors (Beani et al., 2014).

We found little variation in queen morphology among Linepithema species. Only L. pulex appeared to be different, and that was because of its overall small body size. If morphology is constrained by the life history, which likely changes little among Linepithema species, then there is little reason to expect variation with speciation Previous research has demonstrated that queen morphology varies according to the mode of colony founding; queens that establish nests alone or stay in the colony without foraging are larger overall, but particularly large wing muscles that are metabolized to provision for their first brood (Peeters and Ito, 2001; Keller et al., 2014). The invasive red imported fire ant, Solenopsis invicta, has both monogyne and polygyne forms. There is a negative association with queen number and queen weight- the more queens in a colony the smaller and less fecund they are (Vargo and Fletcher, 1989). Linepithema humile queens remain within the nest, or start a new nest by budding, and there can be hundreds of queens in a colony

(Pedersen, et al., 2006), yet surprisingly they are not smaller than their congeners who do not have as many queens (Chapter 4). Additionally, there are no differences among ‘fuscum’ group queens, even though the males are morphologically distinct in ways that suggest a change in reproductive strategies. Queen morphology is independent of mating strategy, but is driven by other factors such as nest founding and egg-laying, adding further support to findings by researchers (Peeters and Ito, 2001; Keller et al., 2014).

Ant workers are the primary caste used for species identification and their characters are likely shaped by the ecological environment they inhabit. Workers of Linepithema exhibited little morphological variation relative to other castes. However, one species was an exception based on the outliers along the bottom row of pairwise PGLS regressions (Figure 3.7): L.

44 aztecoides. Workers from this species often hold their gasters above their bodies, similar to

Azteca, and this behavioral shift appears to have been accompanied by morphological changes such as a flattened petiole and lowered proprodeum (Wild, 2007). Linepithema cryptobioticum also stands out as a unique species based on the PCA results, clustered with the rest of the

‘fuscum’ clade (Figure 3.6). This species has particularly small eyes, but is small over all, and workers are yellow, suggesting they are largely subterranean (Wild, 2007). This species has only been collected in litter samples, so little is known of its biology.

Reproductive female and worker morphology may be constrained by polygyny, as selection does not exert a strong influence on caste phenotypes when multiple queens (and multiple genotypes) make up a colony (Oster and Wilson, 1978; Frumhoff and Ward, 1992). As a result there is a strong relationship between polygyny and a monomorphic worker force

(Frumhoff and Ward, 1992; Fjerdingstad and Crozier, 2006). Specifically, polygyny may reduce selection on workers constraining their morphology (Oster and Wilson, 1978; Frumhoff and

Ward, 1992; Fjerdingstad and Crozier, 2006). Alternatively, it has been shown that phenotypic change does not actually require strong selection pressures (Kingsolver and Pfennig, 2007), so the lack of variation in female castes may indicate similar biology among Linepithema species.

Most variation among worker morphology in the 18 Linepithema species could be attributed to relative body size. Polygyny is also correlated with body size variation where polygynous species tend to produce smaller workers than monogynous species (need citation). Most

Linepithema species appear to be polygynous; L. humile is highly polygynous (Tsutsui and Case,

2001; Ingram, 2002; Pedersen et al., 2006), while other species are only weakly polygynous

(Chapter 4). Future work examining how worker morphological variation correlated with the degree of polygyny is warranted.

Morphological diversification and innovation are correlated with increased speciation rates in many taxa (Burbrink and Pyron, 2010, but see Adams et al. 2009). Given the diversity of morphological variation in the Formicidae, coupled with recent phylogenetic hypothesis for the entire family (Brady et al., 2006; Moreau, 2006) ants may provide an unparalleled opportunity to test how morphology promotes diversification. Of particular interest are diversification rates among lineages that do or do not exhibit strong variation between reproductive and non- reproductive castes, or among lineages that have monomorphic workers (like Linepithema in this study) and those that have polymorphic workers to varying degrees.

Unfortunately, little is known about the ecology and behavior of most Linepithema species. The single pest species L. humile is the only member of the genus with a significant body of literature; knowledge of the remaining species does not extend much beyond the taxonomic descriptions. Thus, the underlying mechanisms driving the patterns of Linepithema morphology inferred here are completely unknown.

The extensive morphological variation among Linepithema males provides a variety of research avenues. Recent morphological evolution of ‘fuscum’ group and L. humile males may represent selection for mating strategies (Emlen and Oring, 1977), while the highly variable males of L. dispertitum may serve as an interesting model for the evolutionary ecology of male form and function. However, further inferences will require targeted studies on the ecology and behavior of these species.

Tables

Table 3.1: The morphological characters used for this study for each caste; see Figure 3.2 for an explanatory diagram.

Code Morphological Character Workers Reproductive females Males HW Head Width X X X HL Head Length X X X SL Scape length X X X EL Eye Length X X X FL Femur Length X X X LHT Length Hind Tibia X X X MFC Min Frontal Carinal Distance X PW Pronotal Width X PDH Propodeal Height X EW Eye Width X X Weber Weber's length X MML Maximum Mesosoma Length X MMW Maximum Mesosoma Width X PH Petiole Height X WL Wing length X X

Table 3.2: The number of specimens of each caste used in this study and the mean log measurement for each morphometric character. Not all characters were measured for each caste.

Caste Species Number HL HW SL EL FL LHT EW MFC PW PDH WL Weber MML MMW L. anathema 6 -0.202 -0.279 -0.196 -0.899 -0.265 -0.216 -0.896 -0.864 -0.443 -0.648 L. angulatum 33 -0.186 -0.219 -0.241 -0.851 -0.29 -0.279 -0.967 -0.777 -0.407 -0.61 - - - - L. aztecoides 17 -0.183 -0.24 -0.211 -0.844 -0.295 -0.243 -0.948 -0.774 -0.433 -0.768 - - - - L. cerradense 17 -0.253 -0.31 -0.311 -0.88 -0.381 -0.347 -0.995 -0.885 -0.488 -0.712 - - - - L. cryptobioticum 6 -0.245 -0.279 -0.387 -1.009 -0.397 -0.377 -1.15 -0.775 -0.447 -0.608 - - - - L. dispertitum 45 -0.157 -0.206 -0.183 -0.827 -0.255 -0.214 -0.937 -0.739 -0.403 -0.578 - - - - L. gallardoi 9 -0.202 -0.25 -0.253 -0.809 -0.325 -0.311 -0.931 -0.836 -0.443 -0.657 - - - - L. humile 9 -0.161 -0.212 -0.146 -0.74 -0.228 -0.185 -0.84 -0.803 -0.392 -0.621 - - - - Workers L. iniquum 81 -0.204 -0.257 -0.205 -0.872 -0.29 -0.252 -0.972 -0.816 -0.461 -0.663 - - - - L. keiteli 31 -0.175 -0.193 -0.23 -0.895 -0.279 -0.255 -1 -0.758 -0.378 -0.591 - -

L. leucomelas 22 -0.162 -0.216 -0.147 -0.845 -0.231 -0.192 -0.947 -0.793 -0.409 -0.647 - -

L. micans 41 -0.174 -0.216 -0.209 -0.804 -0.271 -0.246 -0.906 -0.799 -0.406 -0.611 - - - - L. neotropicum 37 -0.211 -0.254 -0.241 -0.83 -0.31 -0.279 -0.944 -0.827 -0.446 -0.687 - - - - L. oblongum 13 -0.158 -0.234 -0.122 -0.781 -0.201 -0.154 -0.872 -0.815 -0.407 -0.602 - - - - L. piliferum 20 -0.145 -0.186 -0.169 -0.802 -0.223 -0.189 -0.917 -0.736 -0.352 -0.573 - - - - L. pulex 14 -0.246 -0.291 -0.324 -0.918 -0.399 -0.415 -1.03 -0.828 -0.5 -0.696 - - - - L. tsachila 13 -0.185 -0.202 -0.238 -0.849 -0.275 -0.261 -0.954 -0.771 -0.4 -0.591 - - - -

Table 3.2 (cont.): The number of specimens of each caste used in this study and the mean log measurement for each morphometric character. Not all characters were measured for each caste.

Caste Species Number HL HW SL EL FL LHT EW MFC PW PDH WL Weber MML MMW L. angulatum 2 -0.072 -0.089 -0.136 -0.545 -0.133 -0.084 -0.623 - - - 0.674 0.193 - - L. gallardoi 5 -0.1 -0.12 -0.177 -0.555 -0.171 -0.134 -0.654 - - - 0.663 0.183 - - L. humile 13 -0.056 -0.068 -0.073 -0.486 -0.086 -0.041 -0.592 - - - 0.648 0.267 - - L. iniquum 2 -0.102 -0.13 -0.141 -0.619 -0.144 -0.103 -0.748 - - - 0.625 0.185 - -

Reproductive L. keiteli 2 -0.012 -0.012 -0.06 -0.513 -0.033 0.034 -0.594 - - - 0.813 0.303 - - females L. micans 4 -0.036 -0.042 -0.083 -0.494 -0.07 -0.021 -0.569 - - - 0.727 0.293 - - L. neotropicum 5 -0.095 -0.144 -0.146 -0.553 -0.148 -0.115 -0.645 - - - 0.704 0.216 - - L. oblongum 2 -0.075 -0.118 -0.072 -0.557 -0.089 -0.042 -0.676 - - - 0.69 0.195 - - L. piliferum 1 0.029 0.004 -0.032 -0.385 0.031 0.07 -0.517 - - - 0.926 0.397 - - L. pulex 1 -0.144 -0.172 -0.248 -0.64 -0.246 -0.204 -0.757 - - - - 0.137 - - L. angulatum 4 -0.13 -0.16 -0.678 -0.475 0.075 0.07 - - - - 0.655 - 0.221 -0.067 L. cerradense 1 -0.385 -0.416 -0.975 -0.71 -0.397 -0.449 ------0.133 -0.327 L. dispertitum 6 -0.231 -0.276 -0.515 -0.682 -0.212 -0.214 - - - - 0.448 - -0.01 -0.275 L. gallardoi 9 -0.287 -0.307 -0.881 -0.621 -0.283 -0.346 - - - - 0.377 - 0.005 -0.234 L. humile 12 -0.214 -0.203 -0.818 -0.495 -0.17 -0.222 - - - - 0.47 - 0.22 -0.031 L. iniquum 5 -0.313 -0.338 -0.907 -0.687 -0.282 -0.335 - - - - 0.351 - -0.056 -0.284 Males L. keiteli 4 -0.182 -0.212 -0.67 -0.584 -0.015 -0.016 - - - - 0.615 - 0.159 -0.093 L. micans 9 -0.257 -0.265 -0.82 -0.588 -0.23 -0.297 - - - - 0.395 - 0.043 -0.19 L. neotropicum 5 -0.314 -0.325 -0.903 -0.668 -0.287 -0.337 - - - - 0.379 - -0.022 -0.282 L. oblongum 4 -0.284 -0.317 -0.851 -0.658 -0.231 -0.286 - - - - 0.389 - -0.001 -0.208 L. piliferum 4 -0.163 -0.198 -0.636 -0.453 -0.017 -0.042 - - - - 0.656 - 0.176 -0.074 L. pulex 4 -0.336 -0.34 -0.943 -0.744 -0.349 -0.415 - - - - 0.348 - -0.102 -0.335 L. tsachila 4 -0.135 -0.156 -0.647 -0.383 0.053 0.042 - - - - 0.649 - 0.202 -0.071

Table 3.3: Mean values of morphometric measurements for each caste.

Table 3.4: Results of the Bartlett’s tests for homogeneity of variance for each of the morphometric measurements comparing workers, reproductive females, and males (Figure 3.4). K2 is the Bartlett’s K2 statistic with significance values.

Morphometric K2 p value measurement Log HL 50.73 <0.0001 Log HW 22.74 <0.0001 Log SL 146.5 <0.0001 Log EL 29.43 <0.0001 Log FL 137.43 <0.0001 Log LHT 137.54 <0.0001

Table 3.5: ANOVAs of the relationship between each body measurement for each caste, and Tukey’s HSD test for differences between the three castes. The categories indicate the relationship between the x variable and the y variable; the relationship between the two variables for reproductive females contrasted with males; the two variables for workers contrasted against males; and the two variables for workers contrasted against reproductive females. Figures in bold indicate characters that are not different between the two castes.

Independent Males vs Reproductive females Variables Males vs Workers variable (y) Reproductive females vs Workers x y F value p value difference p value difference p value t value p value HL HW 8132.5 <0.0001 0.02807 <0.0001 0.03254 <0.0001 0.00447 0.14709 HL SL 8.9 <0.0001 0.01304 0.01667 0.00239 0.71059 0.01543 0.00032 HL EL 10.3 <0.0001 0.04255 <0.0001 0.02763 <0.0001 0.01492 0.00220 HL FL 21.8 <0.0001 0.04050 <0.0001 0.04091 <0.0001 0.00041 0.98541 HL LHT 3301.0 <0.0001 0.09472 <0.0001 0.06917 <0.0001 0.02555 <0.0001 HW SL 5.1 <0.0001 0.01684 0.03443 0.00293 0.00144 0.01977 0.77632 HW EL 7.4 <0.0001 0.03840 0.00000 0.02420 0.00000 0.01420 0.37381 HW FL 9.1 <0.0001 0.03041 <0.0001 0.02957 <0.0001 0.00084 0.98020 HW LHT 1472.7 <0.0001 0.08693 <0.0001 0.04164 <0.0001 0.04529 <0.0001 SL EL 708.2 <0.0001 0.70665 <0.0001 0.44430 <0.0001 0.26235 <0.0001 SL FL 64.7 <0.0001 0.68625 <0.0001 0.54961 <0.0001 0.13665 <0.0001 SL LHT 154.5 <0.0001 0.62397 <0.0001 0.58865 <0.0001 0.03532 0.00153 EL FL 2621.6 <0.0001 0.04274 0.11588 0.12751 <0.0001 0.10792 <0.0001 EL LHT 1697.0 <0.0001 0.07918 <0.0001 0.22041 <0.0001 0.14122 <0.0001 FL LHT 14846.1 <0.0001 0.06470 <0.0001 0.07399 <0.0001 0.00929 0.00175

Table 3.6: List of significance values from the pairwise PGLS for the three castes. The x and y variables are morphometric measurements, the t values and p values are calculated from the PGLS, and their significance scores, and the λ values indicate the calculated lambda value ranging from 0-1. The asterisks denote the when λ was fixed at 1. Several GLS tests yielded unusual results, these are indicated by the † symbol (the linear model of these relationships give the same results).

Figures

Figure 3.1: An ultrametric tree displaying the phlogenetic relationships among the Linepithema species. The colors represent the distinct clades oulined in Wild (2007). Diagrams represent the male morphology, from top to bottom, of L. humile, L. micans, L, dispertitum, and L. tsachila.

Figure 3.2: The morphometric measurments listed in Table 3.1 on a worker and male. SL: scape length; HL: Head Length; HW: Head Width; EL: Eye Length; EW: eye width; MFC: Maximum Frontal Carinae; PW: Pronotal Width; PDH: Propodeal Height; FL: Femur Length; LHT: Length of Hind Tibia; MML: Maxium Mesosoma Length.

Figure 3.3: Results of the linear discriminant analysis showing the two dimensions grouped by caste. The contribution of LD1 is primarily from scape length, femur length, and head length. Femur length, eye length, and head width, contribute most to LD2.

Figure 3.4: Boxplots of the variance of the six morphometric measurements common to workers, reproductive females, and males, plotted separately for each caste.

Figure 3.5: Scatterplot matrix showing the relationships between of the six morphometric measurements common to males, reproductive females, and workers. Males are represented by red triangles, reproductive females by green circles, and workers by blue plus signs.

Figure 3.6: Results of the principal component analysis (PCA) of each caste grouped by species group. (a) workers, PC 1 accounts for 74.3% of the variance, PC 2 accounts for 14.0%; (b) reproductive females, PC 1 accounts for 86.9% of the variance, PC 2 accounts for 7,1%; (c) males, PC 1 accounts for 67.5% of the variance, PC 2 accounts for 26.1%. Ellipses represent 95% confidence intervals.

Figure 3.7: Worker pairwise phylogenetic general least squares (PGLS) of morphological characters. Lines indicate significant relationships between the x and y variables (table 3.7).

Figure 3.8: Reproductive female pairwise phylogenetic general least squares (PGLS) of morphological characters. Lines indicate significant relationships between the x and y variables (table 3.7).

Figure 3.9: Male pairwise phylogenetic general least squares (PGLS) of morphological characters. Lines indicate significant relationships between the x and y variables (table 3.7).

CHAPTER 4

COMPARISON OF COLONY STRUCTURE IN THE ANT GENUS LINEPITHEMA REVEALS THE EVOLUTION OF TRAITS RELATED TO INVASION SUCCESS IN THE ARGENTINE ANT

Abstract

Biological invasions are a leading threat to biodiversity yet a predictive understanding of why certain organisms are successful at establishing in new environments remains elusive. An under- utilized approach for identifying traits that may convey invasion success is to compare characteristics between related species that differ in invasiveness in a phylogenetic context. I employ this approach with the genus Linepithema that includes the Argentine ant, L. humile, one of the most damaging invasive insects worldwide. The success of this invader has been attributed to specific biological characteristics including its expansive colony structure and extreme polygyny; individual nests can contain hundreds of queens and colonies of many interconnected nests can extend across landscapes. While the Argentine ant itself is well studied, it is unknown whether these same colony traits occur in other Linepithema species. To examine the evolution of colony characteristics in the genus Linepithema, I examined nest number and nest dispersion

(i.e. degree of polydomy), and queen number for colonies of eight congeners from Argentina,

Ecuador and Brazil. These characteristics were compared to native and introduced populations of

L. humile. The observations of colony structure revealed variation in the degree of polydomy in all nine species, however, L. humile was the only species to have colonies with nests spanning more than a few hundred meters. The spatial distribution of nests increases in a stepwise fashion within the ‘humile’ clade, suggesting incremental rather than abrupt changes in colony structure

among species. Queen number estimates from microsatellite genotyping reveal a similar pattern; most species within the ‘humile’ clade have nests with multiple queens, but none possess as many egg-laying queens as L. humile. Together, these findings suggest the evolution of unicoloniality in L. humile involved a significant alteration in the degree of polydomy and polygyny, but not necessarily a fundamental change in colony structure, compared to close relatives. Finally, I compared the data on colony structure among Linepithema to the frequency in which they have been detected in quarantine at ports of entry into the United States (a measure analogous to propagule pressure). Using U.S. Department of Agriculture interception records from specimens stored at the U.S. National Museum from throughout the 20th century, I found six species of Lineptihema have been transported to the United States, yet only L. humile has become established outdoors. Of the remaining intercepted species, none exhibit large degrees of polygyny and polydomy suggesting that opportunity alone is not sufficient for establishment. I did not find records of the two species that do exhibit high degrees of polygyny and polydomy,

L. oblongum and L. micans, suggesting they may pose invasion risks based on their biology, however, may not have had the opportunity to become transported. Collectively, these results highlight how both species-level characters and opportunity influence species’ ability to invade new habitats.

Introduction

Biological invasions are widely recognized as a leading form of global change. Invasive species threaten biodiversity and carry significant economic consequences (Mack et al., 2000). A primary goal of invasion biology is to identify potential invaders and prevent their entry into novel environments (Kolar and Lodge, 2001; van Kleunen et al., 2010). It is therefore imperative

to recognize traits that predict species at risk of becoming invasive. Considerable effort has gone towards studying the biology of introduced species, but there is still little known about the evolutionary processes that permit some species to become invasive while others remain innocuous (Kolar and Lodge, 2001; van Kleunen et al., 2010). One approach that has provided valuable insights is to compare traits between introduced and non-introduced species. For example, a phylogenetic examination of 29 pine species revealed that high seedling growth rate, small seed mass, and short generation times were characteristic of successful invaders (Grotkopp et al., 2002). A meta-analysis of plants generally concluded that invasive species have more extreme values for traits associated with performance (e.g. growth rate, shoot allocation) compared to their non-invasive relatives (van Kleunen et al., 2010). Phylogenetically corrected comparative approaches have been restricted to a small number of studies, so broader patterns of trait evolution in relation to invasion success remain elusive (van Kleunen et al., 2010).

Ants are among the most dominant and conspicuous animals in many communities. Their activities have profound effects on other species and on their physical environment (Hölldobler and Wilson, 1990). Hundreds of ant species have established populations outside their native range (McGlynn, 1999a) and five are included in the ISSG listing of the world’s 100 worst invasive species (Lowe, 2000). Understanding the biology of invasive ants is therefore a priority for biodiversity conservation. A number of life history traits may enable ants to become successful invaders. Invasive ant species typically have multiple queens permitting rapid colony growth (Passera and Keller, 1994), small workers allowing mass recruitment (McGlynn, 1999b), and colonies occupying many nests (polydomy) in introduced populations (Holway et al., 2002).

These characteristics may result from atypical colony structures including unicoloniality (Tsutsui et al., 2000), high polygyny and tolerance of queen adoption (Krieger and Ross, 2002), and

clonal reproduction (Fournier et al., 2005). Collectively, these traits can lead to populations characterized by high genetic relatedness within colonies and a lack of intraspecific aggression among ants from spatially separate nests (Suarez et al., 2008).

The traits associated with the “invasive ant syndrome” may favor rapid colony growth and reduced risk of mortality of incipient colonies. Starting new nests is dangerous for newly mated queens, and survival during nest founding is typically low (Tschinkel, 1992). Queens can avoid independent founding by establishing nests with other newly mated females (pleometrosis) or by rejoining their natal colony (Hölldobler and Wilson, 1977). Polygynous colonies can avoid dangerous mating flights by relocating a subset of queens into new nests with a retinue of workers, a mode of colony founding called budding (Vargo and Porter, 1989). Indeed, queen survivorship in L. humile is enhanced when founding occurs with at least ten workers (Hee et al.,

1999). The number of egg-laying queens in a colony is a key life history trait for social organization. Having many egg-laying individuals allows colonies to grow faster and increases their chance of survival (Tschinkel and Howard, 1983). Queen number and multiple mating can also influence colony structure, including levels of intra-colony genetic variation and patterns of relatedness. Similar to variation in queen number, many ant species vary in the number of nests they construct for a single colony to inhabit. Occupying many physical nest structures may allow colonies to act as “dispersed central place foragers” and both locate and recruit to resources more quickly (Holway and Case, 2000). Such ant species can also create a network of interrelated nests that can collectively defend a larger territory and monopolize more resources. Taken together, these characters, polydomy and polygyny, have been observed in nearly all highly invasive ant species (Holway et. al, 2002).

While many invasive species are well studied, less research has examined how invasive

species differ from their non-invasive congeners. This lack of information is surprising given that many hypotheses for invasion success assume that organisms are successful due to a unique combination of species-specific traits (Kolar and Lodge, 2001; Pyšek and Richardson, 2007). For example, a comparative study of the ant genus Lasius found that characters typically associated with invasion success were concentrated in the invasive ant Lasius neglectus but did not occur in its sister species (Cremer et al., 2008). In this study, I expand this approach by examining biological characters across the Neotropical Dolichoderine ant genus Linepithema which includes the highly invasive Argentine ant, L. humile. Life history data suggests ants in this group vary in key colony characters including nest organization, colony size, and queen number

(Wild 2007). Furthermore, recent studies that clarify species boundaries (Wild, 2007) and interspecific relationships (Wild, 2009) provide a foundation for exploring the evolution of traits in a phylogenetic context.

I focus on two important components of the invasion process to elucidate possible mechanisms of success of Linepithema humile relative to its non-invasive congeners. First, I examine the distribution of traits thought to promote invasion success in ants across the genus. I hypothesize that ants that occupy a network of nests (polydomy) are more likely to become invasive. I estimated colony spatial spread in eight species of Linepithema and compare these species to the invasive L. humile. Furthermore, invasive species have many egg-laying queens in a colony (polygyny), which likely increases invasive success. Using genetic markers developed for research on L. humile (Krieger and Keller, 1999) I calculated worker relatedness within nests and estimated the number of egg-laying queens. Here, I focus on the three species most closely related to L. humile, the sister species L. oblongum, and two others L. micans, and L. gallardoi.

These species comprise the ‘humile’ clade’ (Wild, 2007), with the exception of the rarely

collected, basal species L. anathema. The comparative approach can determine whether the exaggerated polydomy and polygyny are a unique to L. humile and the pattern these traits arise.

Second, I ask how many species of Linepithema have had the opportunity to become established outside their native range. Opportunity is an important, yet often neglected, factor when considering the success of invasive species and comparisons of traits between invasive and non-invasive species need to be conducted in the context of how frequently they are moved (van

Kleunan et al. 2014). Opportunity is assessed by propagule pressure, the total number of individuals introduced to a new environment (Williamson, 1996), and is often considered the most significant factor governing the success of an invader (Kolar and Lodge, 2001; Cassey et al., 2005; Colautti et al., 2006; Lockwood et al., 2009; Simberloff, 2009; but see Nuñez et al.,

2011). The more frequent the introduction events (propagule number) the more likely a species will establish in a new environment (Lockwood et al., 2009; Simberloff, 2009). At least four

Linepithema species have been intercepted at quarantine at U.S. ports (Suarez et al., 2005). I use additional interceptions records from the U.S. National Museum to test the hypothesis that species frequently intercepted in quarantine, but that have never become established in the U.S., will not exhibit characters associated with invasion success (high degrees of polydomy and polygyny).

Methods

Study areas and specimen collection

I collected data on 35 populations from eight Linepithema species in three countries in South

America: Argentina in March 2009, Ecuador in December 2010 and January 2011, and Brazil in

June and July 2012 (Table 4.1). Study sites were found by using collection records on AntWeb

(www.antweb.org) and by opportunistically searching during field surveys in appropriate habitats. Most Linepithema species are associated with disturbed habitat (Wild, 2007), so many collecting localities were along roadsides, deforested open habitats, and agricultural land. I did gain access to several fragments of undisturbed Atlantic Forest and dry forest in Brazil, but did not find Linepithema spp. at these sites. Collecting at each population was conducted with the goal of locating colony boundaries to estimate the degree of polydomy for each species I encountered. In each population, the first was to map all Linepithema nests found in an area approximately 100m2. Second, I extended the search out to locate nests further away at 20-50m intervals over a range of approximately 500m2 to locate the boundary of the colony. If nests were difficult to locate and only a single nest was located, I laid baits (see below) or searched directly for approximately 30 minutes over a range of 200-500m depending on how accessible the habitat was (for example, until I reached a step decline or river). Searching for nests was done directly

(usually by turning over stones, or for arboreal species breaking twigs and lifting bark) and by baiting and following workers if sparsely populated or for species that are difficult to find. Nests were marked with plastic flags on wire posts and the distances between them were recorded. Five nests from each population were selected for the study, often that was all that I located, but if I found more then I collected from three nests that were close together and one nest 50-100 meters away, and another nest about 200m away from the first cluster. I excavated nests to collect workers, mated and unmated queens, males and brood. Worker ants were either placed into

Sarstedt tubes filled with 100% ethanol for genetic work or collected live into 50mL tubes with water-moistened paper for the aggression assays.

Intraspecific aggression

Ants are often highly territorial and workers from many species readily fight with workers from other colonies (Hölldobler and Wilson, 1990). Aggression between nests can therefore give an indication of colony boundaries, and in polydomous colonies, can indicates the spatial arrangement of discrete nests. Workers from different colonies do not recognize each other as nestmates and will attack each other (Suarez et al., 1999). I therefore used aggression assays to determine the degree of polydomy and spatial arrangement of nests in colonies for eight

Linepithema species and to compare these results with published data on introduced and native populations of L. humile. I conducted aggression assays between workers from different nests in eight species of Linepithema spanning the 35 populations in Argentina, Ecuador and Brazil. The assays were conducted in 3 cm diameter plastic Petri dishes, the rims of which were coated in

Fluon to prevent ant escape. Ants were selected randomly for the aggression assays from the pool live workers, although light-colored callow workers were not used if other workers were available. A single worker from a nest was placed with an aspirator in the dish and left for one minute to habituate. A second ant from a different colony was placed in the dish for three minutes of observation, after the highest degree of aggression was recorded. This was repeated three times per nest pair, and once for a control pair of nestmates per nest. For each population three workers from each nest were observed interacting with three workers from the each of the remaining four nests, and once with a nestmate. Aggression was measured using the protocol of

Suarez et al. (1999) and was marked according to the scale: 1 – Neutral, Antennal touching without repellency. 2 - Avoid, antennal touching followed by rapid running. 3 - Aggression, lunging at other ant, or biting. 4 – Fight, both ants wrestling, often to the death. I further divided this metric into general “Aggressive” (average values less than 2) and “Not Aggressive” (average values greater than 2) for L. oblongum, L. micans and L. gallardoi and used logistic regression to

estimate the average spatial extent among nests from different colonies. These results were compared across the genus to data for L. humile from Tsutsui et al. (2000), Tsutsui and Case

(2001), Heller (2004), Pedersen et al. (2006), and Vogel et al. (2009).

Genetic relatedness and effective queen number

Aggression between workers from different nest is not always a reliable indicator of colony identity (Roulston et al. 2003), so I complimented the behavioral data with genetic analyses to estimate if gene flow occurred among nests in each population. The relatedness of workers

(within between nests) and the number of egg-laying queens per nest (i.e, effective queen number) were estimated using microsatellite markers in three Argentinian and Brazilian species of Linepithema within the ‘humile’ clade: L. oblongum, L. micans, and L. gallardoi.

Microsatellites did not amplify reliably for the other Linepithema species. Sixteen individual workers from each nest were sampled, with five nests from each population, leading to 80 workers per population (Table 4.2). DNA was extracted from ant samples from each nest using

Chelex chelating resin from BioRad. Individual ants were placed in the wells of 96-well plates and crushed with pinheads attached to a pin vise. I added 150µL of 5% chelex slurry (by weight) to each well and heated in a themal cycler according to the protocol of Walsh et al. (1991). I used eleven microsatellite loci developed for Linepithema humile : Lhum-3, Lhum-11, Lhum-13,

Lhum-14, Lhum-19, Lhum-28, Lhum-33, Lhum-35, Lhum-39, and Lhum-52, Lhum-62 (Krieger and Keller, 1999). The fragment analysis trace files were scored using the microsatellite plugin for Geneious (version 8.0.4) created by Biomatters (http://www.geneious.com/).

Genetic structure analyses

FSTAT version 2.9.3.2 was used to calculate F-statistics, worker relatedness, and microsatellite compliance with Hardy-Weinberg equilibrium (Goudet, 1995). All microsatellite loci complied with Hardy- Weinberg equilibrium in L. micans and L. gallardoi. Two loci were (Lhum-35 and

Lhum-62) were not in HW equilibrium in the L. oblongum populations, however tests conducted with and without these loci yielded the same results. FSTAT provides a relatedness value for workers in a colony along with a corrected value that accounts for inbreeding in a population

(Pamilo, 1985). I used the corrected relatedness to estimate queen number because in all populations the inbreeding coefficient (FIT) was greater than 0.4. This correction also allowed us to compare queen number estimate directly to those calculated for L. humile (Ingram, 2002;

Pedersen et al., 2006). The effective queen number was calculated according to Queller (1993; equation in Ingram, 2002). The adegenet package in R was used to calculate pairwise FST between nests nest (Jombart, 2008).

A Kruskal-Wallis test indicated whether Linepithema species differed in the maximum distance between nests. Pairwise FST values (see below) from nests were compared for aggressive interactions and non-aggressive interactions for L. oblongum and L. micans. This was done using Mann-Whitney U tests to account for varying sample sizes. Linepithema gallardoi was not included in these tests as all pairwise aggression assays resulted in aggression. Linear regression models were fit to the Pairwise Fst values and distance between nests for L. oblongum, L. micans, and L. gallardoi. The estimated effective number of queens were calculated for L. oblongum, L. micans, and L. gallardoi and compared to queen estimates from introduced populations of L. humile published in Ingram (2002) and Pedersen et. al (2006).

Effective queen number was estimated to be infinite in native populations (Pedersen et al.,

2006). A Kruskal-Wallis test compared estimated queen number in the four Linepithema species.

Unless otherwise mentioned, all statistical analyses were performed using R (version 3.1.1; R

Development Core Team).

Assessment of propagule pressure

The Smithsonian houses the U. S. National Entomological Collection (NMNH) and contains a large collection of Port of Entry (POE) material intercepted throughout the 20th Century in quarantine by the United States Department of Agriculture. I examined all pinned material in the sorted and unsorted drawers for the genera Linepithema and Iridomyrmex (the former name of

Linepithema prior to the Shattuck 1992 revision of the subfamily Dolichoderinae) and recorded data from any specimens intercepted in quarantine.

Results

Spatial colony range

The aggression assays conducted on nine Linepithema species indicated nearly all species form multi-nest colonies (Figure 4.1, Table 4.1). While no species had colonies as spatially extensive as those of L. humile, L. oblongum and L. micans the two species most closely related to humile, had colonies that extended at least over 250m. In populations of Linepithema gallardoi, also from the humile clade, workers always behaved aggressively towards workers from other nests suggesting it is monodomous. I also found a population of L. micans that had small, monodomous colonies, and L. neotropicum exhibited both monodomous and weakly polydomous colony structures among populations. Polydomy was also common in the remaining

5 species outside of the ‘humile’ clade; workers from these species showed no aggression between nests spaced as far as 30-100m apart.

The spatial extent of physical aggression between conspecific non-nestmates was further

examined in the three species most closely related to L. humile; L. oblongum, L. micans, and L. gallardoi (Figure 4.2). Workers from L. oblongum are often not aggressive towards workers in other nests up to a distance of 100m, but ants collected other sites displayed more aggression indicating the degree of polydomy is variable. As a result there was no statistical relationship between intraspecific aggression and distance between nests within the study sites of approximately 200m area (F1,39=3.098. p= 0.086). In contrast, L. gallardoi workers always displayed aggression towards non-nestmates regardless of distance (F1,14=0.093. p= 0.765).

Linepithema micans workers were more likely to display aggression from nests further away

(F1,101=4.288. p= 0.0409), indicating their colonies are polydomous. Linepithema oblongum may display similar results with more extensive sampling over greater distances.

Genetic distance and aggression

Aggression between workers from different nests is a useful indicator of colony boundaries, but genetic relatedness of workers is more informative. I calculated the FST values of nests that showed aggression towards nests that did not (Figure 4.3). All species displayed more aggression towards ants with higher FST values, indicating that the greater the genetic difference the more likely workers will be aggressive towards conspecifics. This pattern was particularly clear for L. micans (W = 0, p < 0.001) and L. humile (W = 3199, p < 0.001; data from Vogel et al. 2009).

The relationship between aggression and genetic distance was not very strong for L. oblongum

(W = 35.5, p = 0.034) because workers from one population (Infernillos) showed no aggression towards conspecifics despite relatively large genetic differences. There was no comparison between FST values for L. gallardoi because workers were always aggressive towards non- nestmates.

Effective queen number

The number of egg-laying queens per colony was calculated for populations of L. oblongum, L. micans, and L. gallardoi and compared to published data for L. humile (Ingram, 2002; Pedersen et al., 2006; Figure 4.4, listed in Table 4.2). Linepithema humile displays the highest degree of polygyny with a conservative average of 74 (95% CI: 36-∞) queens per colony in the introduced range. The 95% confidence intervals indicate the number of calculated queens is infinite

(Pedersen et al., 2006), no doubt an exaggeration, but certainly many more queens than found in the other species of Linepithema. The sister species to the Argentine ant, L. oblongum, had many fewer egg-laying queens in each colony (8.8; 95% CI: 4.45-13.3), L. micans had fewer still (2.9;

95% CI: 0.57-5.4), and L. galladoi had just over 1 queen on average (95% CI 1.36-1.38). The number of egg laying queens was significantly different depending on species (X2 = 11.63, p =

0.009). This pattern suggests there is a stepwise increase in effective queen number in the

‘humile’ clade species leading to the extreme numbers of queens in the invasive L. humile.

Role of opportunity

The specimens from the Smithsonian Institute indicate that at least six species of Linepithema have been intercepted at U.S. ports (Table 4.3). Most of these specimens were likely intercepted after transport from their native range, as the origin points of most shipments were consistent with known species distributions. All specimens of L. humile were imported from areas outside its native range and in areas where the species has to be established, and several interceptions of

L. iniquum are from areas where it is not native (Table 4.3).

Discussion

Biological invasions act as natural experiments that can provide general insight into evolution, population genetics, behavior, and community ecology. These results indicate that polydomy and polygyny occur throughout the genus Linepithema. Interestingly, the highest variation in colony structure was seen among the species of the ‘humile’ clade which ranged from the singled nest and nearly single queened L. gallardoi to L humile that itself remains exceptional in the extent to which nest and queen number are exaggerated. Linepithema gallardoi seems to represent a secondary reversal to monodomy, and may be worth further study for that reason. Linepithema oblongum exhibited the highest variation in the spatial extent of colonies among populations with a few populations not showing nay aggression among nests. However, the genetic data suggests that gene flow was reduced among nests more than 50 meters apart compared to closer nests. This data suggest that the behavioral assays may not always reliably separate colony boundaries for this species. Taken together, these findings suggest the extreme degree of polydomy and polygyny may facilitate the invasive success of L. humile. While many other

Linepithema are polydomous and polygynous, none exhibit these characters to the same extent as

L. humile, However, we did not find evidence that all species have been intercepted in commerce so some may not have the opportunity to invade. Notably, the two species from the ‘humile’ clade with moderate to high level so of polydomy and polygyny, L. oblongum and L. micans, were absent from our survey of interception material suggestion that they may still be successful invaders should they have an opportunity to do so.

One-on-one worker aggression assays were used to establish if spatially separate nests

belonged to the same colonies. This approach has been widely used in studies of ant behavior and colony structure (Suarez et al., 1999; Pedersen et al., 2006; Vogel et al., 2009; Van

Wilgenberg et al., 2010), however, approaches that use more individuals or place them in the context of their colony may give more reliable results (Roulston et al., 2003). Furthermore,

Linepithema species vary in aggression levels. Linepithema iniquum workers, for example, were slow to fight and never wrestled, so caution is required when applying this aggression assay to other species and additional genotyping is important. The genetic data was largely consistent with the behavior results with the exception of L. oblongum. Workers from the Infernillos site displayed no aggression toward one another, but they were genetically divergent. These findings differ from previous studies on L. humile (Pederson et al. 2006) and the findings for L. micans and L. gallardoi where results from aggression assays closely matched genetic similarity. These nests were occupied by few egg-laying queens, which may have budded from a neighboring nest.

That would account for some genotypic variation, although possibly not as much variation as this study revealed. Alternatively, workers from this population may not act aggressively towards conspecifics. Research on Formica ants revealed aggression is costly and can be context dependent (Tanner and Alder, 2009). Workers of L. oblongum may choose not to enter aggressive interactions, instead cooperating with unrelated conspecifics at this site. This further supports the contention that one-on-one worker aggression assays need to be interpreted with care and substantiated with genetic studies (Roulston et al. 2003).

The purpose of the aggression assays was to assess the degree of polydomy, and find colony boundaries to assess the spatial dispersion of colonies from eight species. Linepithema humile exist in colonies that expand for hundreds of meters in their native range (Tsutsui et al.

2000; Tsutsui and Case, 2001; Heller, 2004; Pedersen et al. 2006), and thousands of meters in

their introduced range (Fornier et al., 2005; van Wilgenburg et al., 2009), which was a greater area than all other species included in this study. The ability to form colonies with expansive territories is likely very important to the invasive success of this species (Holway et al., 2002). It is possible colony spatial ranges are underestimated for some populations even though care was taken to find colony boundaries. Some populations extended into inaccessible habitat (for example, a steep ridge), and some nests may have gone unfound because they were small, isolated, or no workers were active on the surface. Conversely, most Linepithema are known to occur commonly in disturbed habitats (Wild, 2007) and that was where I focused collecting efforts. As a result I may have introduced bias by collecting samples from disturbed habitat, which may select for ants with traits adapted for disturbance.

Queen number varied within the ‘humile’ clade: L. humile colonies have extremely high queen numbers per colony (a conservative average of 74), the sister species L. oblongum has fewer (approximately 12), and the more distantly related species L. micans and L. gallardoi had even fewer still (4 and 1.4 queens/colony respectively). This suggests there are incremental changes in colony structure among the ‘humile’ clade, and not an abrupt increase in queens in the invasive lineage. Species outside the ‘humile’ clade also appear to have multiple queens

(Wild, 2007), so polygyny is not an innovation specific to this clade. It is unknown if there is a genetic influence on queen number in Linepithema so we don’t know if changes in social structure reflect genetic variation (and perhaps adaptive selection), or phenotypic plasticity within the genus. The genetic basis of polygyny has been studied in two ants: the invasive red imported fire ant, Solenopsis invicta, and the Alpine silver ant (Liebbrecht and Kronauer, 2014).

In S. invicta one allelic variant of the Gp-9 gene (considered an odorant binding protein)(Krieger and Ross, 2005, Gotzek and Ross, 2009, but see Leal and Ishida, 2008) is associated with

polygyny, colony budding, and worker tolerance of multiple queens (Ross and Keller, 1995;

Ross and Keller, 1998). However, recent addition to Gp-9 research argues this suite of traits is due to a ‘social chromosome’ (Wang et al. 2013) and this genetic architecture underlies the social structure of the Alpine silver ant, Formica sekysi (Purcell et al., 2014). Conversely,

Linepithema species may be tolerant of multiple queens and low worker relatedness, so an increase in queen number may not result from selection on genes such as Gp-9 or a social chromosome. Phenotypic plasticity has long been considered an important feature of invasive species (Baker, 1965). One meta-study comparing invasive species demonstrated they were more phenotypically plastic across multiple traits than non-invasive species (Davidson et al., 2011).

However, phenotypic plasticity did not correlate with a fitness benefit, (Davidson et al., 2011), so these findings are inconclusive. Linepithema humile does have many more queens and much more spatially expansive colonies, so whether this is the result of phenotypic plasticity within the genus is also inconclusive.

These findings are consistent with previous research conducted on L. humile; as genetic difference increases aggression between congeners also increases (Tsutsui et al., 2000; Tsutsui and Case, 2001; Pedersen et al., 2006). The results from L. micans mirror that of L. humile, individuals genetically very similar did not display mutual aggression even though they were collected from different nest, indicating this species is capable of forming polydomous colonies over a range of at least 100 meters, although that varied greatly (Figure 4.1). Linepithema oblongum also forms polydomous colonies, but these results were unusual because L. oblongum workers often did not display aggression. This occurred despite levels of genetic differentiation are similar to what was seen between colonies that exhibited aggression in L. micans and L. humile (Pederson et al., 2006). There are two suggested hypotheses for these results. First, the

workers may all belong to a single colony with a shared hydrocarbon profile, but workers showed high fidelity to particular nests with little intermixing after colonies budded. This would lead to genetic differentiation, but workers would maintain a shared colony recognition profile.

Alternatively, workers from this site may not display aggression towards non-nestmates, as mentioned above. Regardless of the cause, further population sampling of L. oblongum throughout its range is needed to address these unusual results. Further research would also be valuable given how little is known about this ant even though it is the sister species of an important invasive lineage.

Propagule pressure has been considered the most important factor influencing the success of an invasive species (Kolar and Lodge, 2001; Lockwood et al., 2009). The interception data revealed that six Linepithema species have been transported to the U.S. and found in quarantine, indicating they have had opportunity to establish populations outside their native range. Most species are carried in epiphytes such as bromeliads and orchids which accounts for the frequent transport of arboreal specimens L. iniquum and L. leucomelas (Suarez et al. 2005). However, only L. iniquum and L. humile have successfully established populations in new areas, suggesting other intercepted species of Linepithema do not possess the biological adaptations to enable the establishment and spread of introduced populations. Furthermore, Linepithema iniquum only occurs rarely in greenhouses (Wheeler, 1929; Creighton, 1950) and has not been detected outdoors, so this species survives only in heavily modified environments and does not impact native ant assemblages. My results show that most Linepithema species possess polydomous colonies, and many certainly possess multiple egg-laying queens, but this may not be enough for these species to establish populations. The invasive Argentine ant has these characters, but they are at extreme end of the spectrum; they have an order of magnitude more queens in a colony

and no aggression between workers over greater distances.

A goal of invasion biology is to predict potential invasive species and prevent their entry into new regions (Kolar and Lodge 2001; van Kleunen et al. 2010). Of the species studied in depth in the ‘humile’ clade, L. oblongum was the most similar to L. humile, although colonies possess many fewer queens have higher intra-colony relatedness. This species may pose an invasive threat should the opportunity arise, however there are no records of this species being transported via human commerce. Linepithema oblongum may not have the opportunity for transport. The range of this soil-dwelling species is at higher altitudes in the Andes (Wild, 2009), often in remote mountain valleys, and is therefore less likely to be picked up by human commerce and transported via coastal ports. The same applies to L. micans, another species of

Linepithema that may post a pest risk. This species has not been detected in “port of entry” samples, so it is unknown whether it is transported globally. Linepithema micans is a pest of the

Southern Brazilian wine industry because it forms mutualism with the ground pearl

Eurhizococcus braziliensis (Margarodidae), a soil scale that infests the roots of grapevines

(Nondillo et al., 2013). An association with an economically important crop may increase the risk of introducing L. micans into new environments; it may establish populations but also assist the establishment of a vineyard pest known to reduce grape yield and kill the plants (Sacchett et al., 2009).

The Port of Entry material notably lacked specimens intercepted from Argentina, Brazil,

Paraguay and Uruguay. These four countries are of significance to research on Linepithema as this area has the greatest species diversity (Wild, 2009). It is possible other species have been transported globally, and even intercepted, from this region South America, but they were not

included in the material housed at the National Collection. Intercepted ant specimens from this part of the world will be valuable in further assessing the invasion risk posed by other

Linepithema species and identifying the traits conferring success in L. humile.

As with most studies on invasion success, I focused on species level traits. However, we are beginning to appreciate that examining the characteristics of individuals (e.g. intraspecific variation in traits) may be just as important for understanding invasions success (Bolnick et al.,

2011; Chapple et al., 2012). Sampling species though broad parts of their range highlights the behavioral and morphological diversity within a species. This was most noticeable for L. micans, which I sampled throughout Argentina from Buenos Aires through to the Northwest along with

Entre Rios, and from southern Brazil north to Minas Gerais. Further research on this species is likely to reveal it as a cluster of several cryptic species (Wild, 2009), an inference supported by the behavioral observations. Linepithema micans workers vary slightly in body proportion, pilocity, color, and eye size. In the aggression assay some populations appeared more aggressive than others, and they may also differ in their diurnal activity levels (unpublished observations).

Furthermore, the levels of mitochondrial sequence divergence with nominal L. micans (Wild,

2009; Holley, unpublished data) are congruent with the presence of cryptic species. This level of species variation is almost certainly present other Linepithema (for example, L. iniquum, Wild,

2007). The collecting was constrained by time, so I sampled more species at the expense of sampling more populations per species. Future research on this genus will require more extensive population sampling particularly for L. oblongum and L. gallardoi.

Invasion has been associated with morphological innovation (Dumont et al. 2011), but it is unknown whether the unicolonial behavior of invasive L. humile is caused by selection for extreme numbers of queens or if it’s within the scope of native levels of polygyny. Further

research into the eco-evolutionary drivers behind L. humile’s invasive success will benefit greatly from comparisons to non-invasive Linepithema. For example, experimental comparisons of colony survival and growth with varying queen number would elucidate whether L. humile has an invasive advantage over its congeners. We can now pursue the molecular underpinning of polygyny now we see how queen number varies in the ‘humile’ clade.

Tables

Table 4.1: List of species and the collecting location. Maximum distance, the greatest distance (in meters) between two nests where workers did not display aggression; Nests, the number of nests found at each site and included in the aggression assays; Colonies, the number of estimated colonies that the nest belonged to based on aggression level between workers.

Maximum Species Location Country Nests Colonies distance (m)

L. oblongum Infernillos Argentina 94.6 5 1 Los Cobres Argentina 95 5 3 Las Lagunas Argentina 75 5 1 Termas de Reyes Argentina 141.7 5 2 L. micans Moro do minha Brazil 33 5 3 Univeridade Paraná carpark Brazil 0 5 5 Univeridade Paraná forest Brazil 47 5 4 Road Santa Catalina Brazil 45.7 5 3 Pinto Bandera Brazil 135 5 1 Caxaisas do sul Brazil 69 3 1 Enterna Carrancas Brazil 37 3 2 Passeos Pedras Brazil 45 5 2 El Mollar Argentina 0 4 4 Flores Amarillas Argentina 76 5 2 Arroyo do los Loros Argentina 0 3 3 Las Ruinas Argentina 0 2 2 Costenera sur Argentina 107 5 4 L. gallardoi Reartes Argentina 0 5 5 Lorenzo Argentina 0 5 5 L. neotropicum Zilda Brazil 50 5 2 Monteverde Ecuador 13 4 3 road to Hakunamatata Ecuador 0 5 5 L. iniquum Maquipucuna Ecuador 4.9 5 3 Mindo Ecuador 7.2 4 2 Moro do minha Brazil 35 4 3 Rio Bonita Brazil 3.9 4 3 Alluriquin Ecuador 26 4 2 L. piliferum Nanegalito Ecuador 65.4 5 1 Pasando de Cuyuja Ecuador 51.4 5 3 Road to Yanayacu Ecuador 17 5 2 Rio Sardina Ecuador 47 5 2 L. tsachila Maquipucuna Ecuador 100 5 1 Mindo Ecuador 100 5 1 Toachi Ecuador 85 5 2 L. angulatum Sinchi Ecuador 30.6 5 2

Table 4.2: Colony structure in four species of the Linepithema “humile clade”. n, the number of * genotyped workers per population; r, relatedness of nestmates; r , corrected relatedness; Ne, the effective queen number per colony. Hawaiian samples from Ingram (2002), ‡ number of samples genotyped total; † Hawaiian sample from Pedersen et al. (2006).

Species Location n r r* Ne L. humile Hawaii 0.18 3.9 Hawaii ‡ 0.13 6 392 Hawaii 0.02 37.5 Hawaii 0.02 37.5 Aminga 224 0.005 150 European Main 100 0.006 125 Australia 100 0.006 125 Hawaii† 100 0.005 150 California 100 0.028 27 L. oblongum Infernillos 50 0.748 0.065 11.5 Termas de Reyes 57 0.601 0.063 11.9 Las Lagunas 40 0.089 0.076 9.9 Campo Quijano 44 0.655 0.337 2.2 L. micans Santa Catalina 50 0.773 0.115 6.5 Monte Verde 64 0.7 0.288 2.6 Moro de Minha 39 0.756 0.498 1.5 El Mollar 63 0.723 0.609 1.2 L. gallardoi Los Reartes 49 0.697 0.548 1.4 Lorenzo 34 0.724 0.546 1.4

Table 4.3: Records of Linepithema spp. intercepted at ports entering the USA. Countries in bold are outside of the species known native range.

Number of Species Cargo country of origin Interceptions L. angulatum 6 Colombia, Venezuela L. dispertitum 18 Costa Rica, Guatamala, Mexico L. humile 18 Azores, Bermuda, Brazil, France, Italy, Mexico, Portugal, South Africa, Taiwan, Thailand L. iniquum 46 Brazil, Colombia, Costa Rica, Ecuador, England, France, Guatemala, Mexico, Peru, Venezuela L. leucomelas 3 Brazil L. piliferum 8 Colombia, Venezuela

Figure 4.1: The maximum distance (meters) between nests of workers that did not display aggression; a proxy for maximum spatial range of colonies. Tick marks on the bars represent the mean spatial range. Data are presented for nine Linepithema species; L.humile is separated into the introduced and native populations.

Figure 4.2: The level of aggression displayed towards non-nestmates conspecifics over distance (in meters). (a) L. oblongum, (b) L. micans, and (c) L. gallardoi. The line of fit indicates a statistically significant positive relationship between aggression and distance in L. micans.

Figure 4.3: The pairwise FST (± SEM) values between nests for aggressive interactions (dark grey) and non-aggressive interactions (light grey bars). Linepithema humile data drawn from Vogel et al. 2009.

Figure 4.4: Mean number of effective queens per colony for four species of Linepithema. Error bars denote 95% confidence intervals; the upper confidence interval for L. humile was calculated as ∞ (Pedersen et al., 2006).

REFERENCES

Adams, D. C., Berns, C. N., Kozak, K.H., and Weins. J.J. 2009. Are rates of species diversification correlated with rates of morphological evolution? Proceedings of the Royal Society of London B, 276:2729-2738. Aguirre, L. F., Herrel, A., van Damme, R. and Matthysen, E. 2002. Ecomorphological analysis of trophic niche partitioning in a tropical savannah bat community. Proceedings of the Biological Society, 269:1271-1278. Andersen, A. N. 1991. Seed-harvesting by ants in Australia. Ant–Plant Interactions. Eds C.R.Huxley & D.F.Cutler. Pages:493-503. Oxford University Press, Oxford. Andersson, M.1984. The evolution of eusociality. Annual Review of Ecology and Systematics, 15:165-189. Avise, J. C. 2004. Molecular markers, natural history, and evolution. 2nd Edition. Sinauer Associates, Inc., Sunderland. Baker, H. G. 1965. Characteristics and modes of origin of weeds. In: The Genetics of Colonizing Species (eds Baker, H.G. & Stebbins, G.L.). Academic Press, New York City. Beani, L., Dessì-Fulgheri, F., Cappa, F., and Toth, A. 2014. The trap of sex in social insects: From the female to the male perspective. Neuroscience and Biobehavioral Reviews, doi:10.1016/j.neubiorev.2014.09.014. Bolnick, D.I., Amarasekare, P., Araujo, M.S., Burger, R., Levine, J.M., Novak, M., Rudolf, V.H.W., Schreiber, S.J., Urban, M.C., and Vasseur, D.A. 2011. Why intraspecific trait variation matters in community ecology. Trends in Ecology and Evolution, 26:183-912. Bolnick, D.I., Svanback, R., Fordyce, J.A., Yang, L.H., Davis, J.M., Hulsey, C.D., and Forister M.L. 2003. The ecology of individuals: incidence and implications of individual specialization. The American Naturalist, 161:1-28. Boomsma, J. J. 2009. Lifetime monogamy and the evolution of eusociality. Philosophical Transactions of the Royal Society B: Biological Sciences, 364:3191-3207. Borgia, G. 1980. Sexual competition in Scatophaga stercoraria: size- and density- related changes in male ability to capture females. Behaviour, 75: 185-206. Boudinot, B. 2013. The male genitalia of ants: musculature, homology, and functional morphology (Hymenoptera, Aculeata, Formicidae). Journal of Hymenoptera Research, 30:29-49. Bourke, A. F. G. and Franks, N. R. 1995. Social evolution in ants. Princeton University Press, Cambridge. Brady, S. G., Schultz, T. R., Fisher, B. L., and Ward, P. S. 2006. Evaluating alternative hypotheses for the early evolution and diversification of ants. Proceedings of the National

Academy of Sciences, 103:18172-18177. Brandão, C. R. F., Diniz J. L. M. and Tomotake, E. M. 1991. Thaumatomyrmex strips millipedes for prey: a novel predatory behaviour in ants, and the first case of sympatry in the genus (Hymenoptera: Formicidae). Insectes Sociaux, 38:335-344. Brown, J. H. Reichman, O. J., and Davidson, D. W. 1979. Granivory in desert ecosystems. Annual Review of Ecology and Systematics, 10:201-227. Brown, J. J. and Traniello, J. F. A. 1998. Regulation of brood care in the dimorphic castes of the ant Pheidole morrisi (Hymenoptera:Formicidae): effects of caste ratio, colony size and colony needs. Journal of Insect Behaviour, 11:209-219. Brown, W.L., and Wilson, E.O. 1956. Character displacement. Systematic Zoology, 5:49-64. Burbrink, F. T. and Pyron, R. A. 2010. How does ecological opportunity influence rates of speciation, extinction and morphological diversification in new world rat snakes (Tribe Lampropeltini)? Evolution, 64:934-943. Burnham, K. P. and Anderson, D.R. 2002. Model selection and multimodel influence- a practical information-theoretic approach. 2nd Edition. Springer-Verlag. New York City. Cassey, P., Blackburn, T. M., Duncan, R. P., and Lockwood, J. L. 2005. Lessons from the establishment of exotic species: a meta-analytical case study using birds. Journal of Animal Ecology, 74:250-258. Chapple, D. G, Simmonds, S. M. and Wong, B. B. M. 2012. Can behavioral and personality traits influence the success of unintentional species introductions? Trends in Ecology and Evolution, 27:57-64. Clifton, K. B., and Motta, P. J. 1998. Feeding morphology, diet and ecomorphological relationships among five Caribbean Labrids (Teleosti, Labirdae). Copeia, 4:953-966. Colautti, R. I., Grigorovich, I. A., & MacIsaac, H. J. 2006. Propagule pressure: a null model for biological invasions. Biological Invasions, 8:1023-1037. Creighton, W. S. 1950. The ants of North America. Bulletin of the Museum of Comparative Zoology, 104:1-585. Cremer, S., Ugelvig, L. V., Drijfhout, F. P., Schlick-Steiner, B. C., Steiner, F. M., Seifert, B., and Boomsma, J. J. 2008. The evolution of invasiveness in garden ants. PLoS One, 3:e3838. Cuautle, M., Rico-Gray, V., Diaz-Castelazo, C. 2005. Effects of ant behaviour and presence of extrafloral nectaries on seed dispersal of the Neotropical myrmecochore Turnera ulmifolia L. (Turneraceae). Biological Journal of the Linnean Society, 86:67-77. Darwin, C. 1874. The Descent of Man, and Selection in Relation to Sex. 2nd Edition, John Murray, London. Davidson, A. M., Jennions, M. and Nicotra, A. B. 2011. Do invasive species show higher 92

phenotypic plasticity than native species and, if so, is it adaptive? A meta-analysis. Ecology Letters, 14:419-431. Davidson, D. W. 1977. Foraging ecology and community organization in desert seed-eating ants. Ecology, 58:725-737. Debout, G., Schatz, B., Elias, M. and McKey, D. 2007. Polydomy in ants: what we know, what we think we know, and what remains to be done. Biological Journal of the Linnean Society, 90:319-348. Dejean, A. and Evraerts, C. 1997. Predatory behaviour in the genus Leptogenys: A comparative study. Journal of Insect Behaviour, 10:177-191. Dejean, A., Le Breton, J., Suzzoni, J. P., Orivell, J. & Saux-Moreau, C. 2005. Influence of interspecific competition on the recruitment behaviour and liquid food transport in the tramp species Pheidole megacephala. Naturwissenschaften, 92:324–327. Detrain C. and Pasteels J. M. 1992. Caste polyethism and collective defense in the ant, Pheidole pallidula: The outcome of quantitative differences in recruitment. Behavioural Ecology and Sociobiology, 29:405-412. Dornhaus, A. 2008. Specialization does not predict individual efficiency in an ant. PLoS Biology, 6: e285. Dumont, E. R., Dávalos, L. M., Goldberg, A., Santana, S. E., Rex, K., & Voigt, C. C. 2011. Morphological innovation, diversification and invasion of a new adaptive zone. Proceedings of the Royal Society B: Biological Sciences, doi:10.1098/rspb.2011.2005. Emlen, S. T., and Oring, L. W. 1977. Ecology, sexual selection, and the evolution of mating systems. Science, 197:215-223. Felsenstein, J. 1985. Phylogenies and the comparative method. American Naturalist, 126:1-25. Ferster, B., Pie, M. R. and Traniello, J. F. A. 2006. Morphometric variation in North American Pogonomyrmex and Solenopsis ants: caste evolution through ecological release or dietary change? Ethology, Ecology and Evolution, 18:19-32. Fisher, D. O., & Owens, I. P. 2004. The comparative method in conservation biology. Trends in Ecology & Evolution, 19:391-398.

Fjerdingstad, E. J., and R. H. Crozier. 2006. The evolution of worker caste diversity in social insects. The American Naturalist, 167:390-400. Fornier, D., Estoup, A., Orivel, J., Foucaud, J., Jourdan, H., Le Breton, J., and Keller, L. 2005. Clonal reproduction by males and females in the little fire ant. Nature, 435:1230-1234. Fox, J. and Weisberg, S. 2011. An {R} Companion to Applied Regression, 2nd Edition. Thousand Oaks CA. URL:http://socserv.socsci.mcmaster.ca/jfox/Books/Companion Franks, N. and Hölldobler, B. 1987. Sexual competition during colony reproduction in army ants. 93

Biological Journal of the Linnean Society, 30:229-243. Frumhoff, P. C., and Ward, P. S. 1992. Individual-level selection, colony-level selection, and the association between polygyny and worker monomorphism in ants. American Naturalist, 139:559-590. Garland, T., Jr., Harvey, P. H. and Ives, A. R. 1992. Procedures or the analysis of comparative data using phylogenetically independent contrasts. Systematic Biology, 41:18-32. Giraud, T., Pedersen, J. S., & Keller, L. 2002. Evolution of supercolonies: the Argentine ants of southern Europe. Proceedings of the National Academy of Sciences, 99:6075-6079. Gotzek, D. and Ross, K. G. 2009. Current status of a model system: the gene Gp-9 and its association with social organization in fire ants. PLoS ONE, 4:e7713. Goudet, J. 1995. FSTAT (Version 1.2): A Computer Program to Calculate F-Statistics. Journal of Heredity, 86:485-486. Grafen, A. 1989. The phylogenetic regression. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 326:119-157. Grant, P. R. 1999. Ecology and evolution of Darwin’s finches. Princeton University Press. Greenwood, P. J. 1980. Mating systems, philopatry and dispersal in birds and mammals. Animal Behavior, 28:1140-1162. Grotkopp, E., Rejmanek, M., and Rost, T. L. 2002. Toward a causal explanation of plant invasiveness: seedling growth and the life-histoty strategies of 29 pine (Pinus) species. The American Naturalist, 159:396-419. Gurevitch, J., & Padilla, D. K. 2004. Are invasive species a major cause of extinctions? Trends in Ecology & Evolution, 19:470-474. Hamilton, W. D. 1964. The genetical evolution of social behaviour. Journal of theoretical biology, 7:1-16. Harrison, R. G. 1980. Dispersal polymorphisms in insects. Annual Review of Ecology and Systematics, 11:95-118. Harvey, P. H. and Pagel, M. D. 1991. The comparative method in evolutionary biology. Oxford University Press, Oxford. Hee, J. J., Holway, D. A., Suarez A. V., and Case, T. J. 2000. Role of propagule size in the success of incipient colonies of the invasive Argentine ant. Conservation Biology, 14:559-563. Heinze, J., and Hölldobler, B. 1993. Fighting for a harem of queens: physiology of reproduction in Cardiocondyla male ants. Proceedings of the National Academy of Sciences, 90:8412-8414. Heinze, J., Hölldobler, B., & Peeters, C. 1994. Conflict and cooperation in ant societies. Naturwissenschaften, 81:489-497.

Heinze, J., Trindl, A., Seifert, B., and Yamauchi, K. 2005. Evolution of male morphology in the ant genus Cardiocondyla. Molecular phylogenetics and evolution, 37:278-288. Helenterä, H., Strassmann, J. E., Carrillo, J. and Queller, D. C. 2009. Unicolonial ants: where do they come from, what are they and where are they going? Trends in Ecology and Evolution, 24:341-349. Heller, N. 2004. Colony structure in introduced and native populations of the invasive Argentine ant, Linepithema humile. Insectes Sociaux, 51:378-386. Herbers, J. M. 1984. Queen-worker conflict and eusocial evolution in a polygynous ant species. Evolution, 38:631-643. Hölldobler, B. and Wilson, E. O. 1977. The number of queens: an important trait in ant evolution. Naturwissenschaften, 64:8-15. Hölldobler, B. and Wilson, E. O. 1990. The ants. Belknap Press.Harvard University Press, Cambridge. Holm, S. 1979. A simple sequentially rejective multiple test procedure. Scandinavian journal of statistics, 6:65-70. Holway, D. A. 1999. Competitive mechanisms underlying the displacement of native ants by the invasive Argentine ant. Ecology, 80:238-251. Holway, D. A. and Case, T. J. 2000. Mechanisms of dispersed central-place foraging in polydomous colonies of the Argentine ant. Animal Behaviour, 59:433-441. Holway, D. A., Lach, L., Suarez, A. V., Tsutsui, N. D., & Case, T. J. 2002. The causes and consequences of ant invasions. Annual review of ecology and systematics, 33:181-233. Hosken, D. J., and P. Stockley. 2004. Sexual selection and genital evolution. Trends in Ecology and Evolution, 19:87-93. House, C. M., and Simmons, L. W. 2005. The evolution of male genitalia: patterns of genetic variation and covariation in the genital sclerites of the dung beetle Onthophagus taurus. Journal of Evolutionary Biology, 18:1281-1292. Howell, D. C. 2007. Statistical methods for psychology. Thomson Wadsworth, Belmont. Huang, M. H. 2010. Multi-phase defense by the big-headed ant, Pheidole obtusospinosa, against raiding army ants. Journal of Insect Science, 10:1-10. Hughes, L., Westoby, M., Jurado, E. 1994. Convergence of elaiosomes and insect prey: evidence from ant foraging behavior and fatty acid composition. Functional Ecology, 8:358-365. Hughes, W. O., Oldroyd, B. P., Beekman, M., & Ratnieks, F. L. 2008. Ancestral monogamy shows kin selection is key to the evolution of eusociality. Science, 320:1213-1216. Hulsey, C. D. and Wainwright, P. C. 2002. Projecting mechanics into morphospace: disparity in

the feeding system of labrid fishes. Proceedings of the Royal Society London Series B, 269:317- 326. Ingram, K. K. 2002. Plasticity in queen number and social structure in the invasive Argentine ant (Linepithema humile). Evolution, 56:2008-2016. Johnson, R. A. 2000. Seed Harvesting Ants (Hymenoptera: Formicidae) of North America: An Overview of Ecology and Biogeography. Sociobiology, 36:89-122. Jombart, T. 2008. adegenet: a R package for the multivariate analysis of genetic markers. Bioinformatics, 24:1403-1405. Karr, J. R., and James, F. C. 1975. Eco-morphological configurations and convergent evolution in species and communities. In: Ecology and Evolution of Communities. Eds. Cody, M. L. and Diamond, J. M. Pages 258-291. Harvard University Press, Cambridge. Kaspari, M. 1996. Worker size and seed size selection by harvester ants in a Neotropical forest. Oecologia, 10:397-404. Kaspari, M. and Weiser, M. 1999. Interspecific Scaling in Ants: the Size-Grain hypothesis. Functional Ecology, 13:530-538. Keller, L. 1991. Queen number, mode of colony founding, and queen reproductive success in ants (Hymenoptera Formicidae). Ethology, ecology & evolution, 3:307-316. Keller, L. 1995. Social life: the paradox of multiple-queen colonies. Trends in Ecology & Evolution, 10:355-360. Keller, L., and Passera, L. 1989. Size and fat content of gynes in relation to the mode of colony founding in ants (Hymenoptera; Formicidae). Oecologia, 80:236-240. Keller, R. A., Peeters, C. and Beldade, P. 2014. Evolution of thorax architecture in ant castes highlights trade-off between flight and ground behaviours. eLife 2014;3:e01539. Kingsolver, J. G., and Pfennig, D. W. 2007. Patterns and power of phenotypic selection in nature. Bioscience, 57:561-572.

Kinomura, K., and Yamauchi, K. 1987. Fighting and mating behaviors of dimorphic males in the ant. Journal of Ethology, 5:75-81. van Kleunen, M., Dawson, W., Maurel, M. 2014. Characteristics of successful alien plants. Molecular Ecology, DOI: 10.1111/mec.13013. van Kleunen, M., Dawson, W., Schlaepfer, D., Jeschke, J. M. and Fischer, M. 2010. Are invaders different? A conceptual framework of comparative approaches for assessing determinants of invasiveness. Ecology Letters, 13:947-958. van Kleunen, M., Weber, E. & Fischer, M. 2010. A meta-analysis of trait differences between invasive and non-invasive plant species. Ecology Letters, 13:235-245. Klingenberg, C. P. 2008. Morphological Intergration and Developmental Modularity. 96

Annual Review of Ecology, Evolution, and Systematics, 39:115-132. Kolar, C. S., Lodge, D. M. 2001. Progress in invasion biology: predicting invaders. Trends in Ecology and Evolution, 16:199-204. Krieger, M. J. B., and Keller, L. 1999. Low polymorphism at 19 microsatellite loci in a French population of Argentine ants (Linepithema humile). Molecular Ecology, 8:1078-1080. Krieger, M. J., & Ross, K. G. 2005. Molecular evolutionary analyses of the odorant-binding protein gene Gp-9 in fire ants and other Solenopsis species. Molecular biology and evolution, 22:2090-2103. Krieger, M. J., and Keller, L. 2000. Mating frequency and genetic structure of the Argentine ant Linepithema humile. Molecular Ecology, 9:119-126. Lachaud, J. P., Passera, L., Grimal, A., Detrain, C. and Beugnon, G. 1992. Lipid storage by major workers and starvation resistance in the ant Pheidole pallidula (Hymenoptera: Formicidae). In: Biology and evolution of social insects. Ed. Billen. J. Pages153-160. Leuven University Press, Leuven. Larabee, F. J. and Suarez, A. V. 2014. The evolution and functional morphology of trap-jaw ants. Myrmecological News, 20:25-36. Leal, W. S., and Ishida, Y. 2008. Gp-9s are ubiquitous proteins unlikely involved in olfactory mediation of social organization in the red imported fire ant, Solenopsis invicta. PLoS ONE, 3:e3762. Leal, I. R. and Oliveira, P. S. 1995. Behavioural ecology of the neotropical termite-hunting ant Pachycondyla marginata: colony founding, group-raiding and migratory patterns. Behavioural Ecology and Sociobiology, 37:373-383. Lengyel, S., Gove, A. D., Latimer, A. M., Majer, J. D., Dunn, R. R. 2009. Ants sow the seeds of global diversification in flowering plants. PLoS ONE, 4:e5480. Libbrecht, R., & Kronauer, D. J. 2014. Convergent Evolution: The Genetics of Queen Number in Ants. Current Biology, 24:R1083-R1085. Lockwood, J. L., Cassey, P., & Blackburn, T. M. 2009. The more you introduce the more you get: the role of colonization pressure and propagule pressure in invasion ecology. Diversity and Distributions, 15:904-910. Lowe, S., Browne, M. & Boudjelas, S. 2000. 100 of the world’s worst invasive alien species. Aliens, 12:1-12. Mack, R. N., Simberloff, D., Lonsdale, W. M., Evans, H. Clout, M. and Bazzaz, F. A. 2000. Biotic invasions: causes, epidemiology, global consequences, and control. Ecological Applications, 10:689-710. Madison, W. P. and Madison, D. R. 2011. Mesquite: a modular system for evolutionary analysis. Version 2.7. http://mesquiteproject.org. 97

Mahler D. L., Revell, L. J. Glor, R. E. and Losos, J.B. 2010. Ecological opportunity and the rate of morphological evolution in the diversification of greater Antillean anoles. Evolution, 64:2731- 2745. Marshall, C. D., Guzman, A., Narazaki, T., Sato, K., Kane, E. A. and Sterba-Boatwright, B. D. 2012. The ontogenetic scaling of bite force and head size in loggerhead sea turtles (Caretta caretta): implication for durophagy in neritic, benthic habitats. Journal of Experimental Biology, 215:4166-4174. Martin, C. H. and P. C. Wainwright. 2011. Trophic novelty is linked to exceptional rates of morphological diversification in two adaptive radiations of Cyprinodon pupfish. Evolution, 65:2197-2212. McGlynn, T. P. 1999a. The worldwide transfer of ants: geographical distribution and ecological invasions. Journal of Biogeography, 26:535-548. McGlynn, T. P. 1999b. Non-native ants are smaller than related native ants. The American Naturalist, 154:690-699. McGlynn, T. P., Diamond, S. E., and Dunn, R. R. 2012. Tradeoffs in the evolution of caste and body size in the hyperdiverse ant genus pheidole. PLoS ONE, 7: e48202. Mertl, A. L. and Traniello, J. F. A. 2009. Behavioural variation in the major worker subcaste of twig-nesting Pheidole (Hymenoptera: Formicidae): does morphological specialization influence task plasticity? Behavioural Ecology and Sociobiology, 63:1411-1426. Mertl, A. L., Sorenson, M. D. and Traniello, J. F. A. 2010. Community-level interactions and functional ecology of major workers in hyperdiverse ground-foraging Pheidole (Hymenoptera, Formicidae) of Amazonian Ecuador. Insect Sociaux, 57:441-452. Miles, D. B. and Ricklefs, R. E. 1984. The correlation between ecology and morphology in deciduous forest passerine birds. Ecology, 6:1629-1640. Moreau, C. S., Bell, C. D., Vila, R., Archibald, S. B., and Pierce, N. E. 2006. Phylogeny of the ants: diversification in the age of angiosperms. Science, 312:101-104. Moreau, C. S. 2008. Unraveling the evolutionary history of the hyperdiverse ant genus Pheidole (Hymenoptera: Formicidae). Molecular Phylogenetics and Evolution, 48:224-239. Mueller, U. G., Schultz, T. R, Currie, C. R., Adams, R. M. M. and Malloch, D. 2001. The orgin of the attine ant-fungus mutualism. Quarterly Review of Biology, 76:169-197. Ness J. H., Bronstein J. L., Andersen A. N. and Holland J. N. 2004. Ant body size predicts dispersal distance of ant-adapted seeds: Implications of small-ant invasions. Ecology, 85:1244- 1250. Nondillo, A., Sganzerla, V. M. A., Bueno, O. C., & Botton, M. 2013. Interaction between Linepithema micans (Hymenoptera: Formicidae) and Eurhizococcus brasiliensis (Hemiptera: Margarodidae) in Vineyards. Environmental entomology, 42:460-466.

Nuñez, M. A., Moretti, A., & Simberloff, D. 2011. Propagule pressure hypothesis not supported by an 80 year experiment on woody species invasion. Oikos, 120:1311-1316. O’Dowd, D. J., Green, P. T, and Lake, P. S. 2003. Invasional ‘meltdown’ on an oceanic island. Ecological Letters, 6:812-817. Oster, G. F. and Wilson. E. O. 1978. Caste and ecology in the social insects. Princeton University Press, Princeton. Pagel, M. 1999. The maximum likelikhood approach to reconstructing character states of discrete characters on phylogenies. Systematic Biology, 48:612-622. Pamilo, P. 1985. Effect of inbreeding on genetic relatedness. Hereditas, 103:195-200. Paradis, E., Claude, J., ant Strimmer, K. 2004. APE: analyses of phylogenetics and evolution in R language. Bioinformatics, 20:289-290. Passera, L. and Keller, L. 1994. Mate availability and male dispersal in the Argentine ant Linepithema humile (Mayr) (= Iridomyrmex humilis). Animal Behaviour, 48:361-369. Passera, L., Keller, L., and Suzzoni, J. 1988. Queen replacement in dequeened colonies of the Argentine ant Iridomyrmex humilis (Mayr). Psyche, 95:59-68. Paul, J. 2001. Mandible movements in ants. Comparative Biochemistry and Physiology Part A, 131:7-20. Paul, J. and Gronenberg, W. 1999. Optimizing force and velocity: mandible muscle fibre attachment in ants. The Journal of Experimental Biology, 202:797-808. Pedersen, J. S., Kreiger, M. J. B., Vogel, V., Girard, G., and Keller, L. 2006. Native supercolonies of unrelated individuals in the invasive Argentine ant. Evolution, 60:782-791. Peeters, C. 1997. Morphologically ‘primitive’ants: comparative review of social characters, and the importance of queen-worker dimorphism. The evolution of social behavior in insects and arachnids. Cambridge University Press, Cambridge. Peeters, C. and Crewe. R. 1987. Foraging and recruitment in Ponerine ants: solitary hunting in the queenless Ophthalopone berthoudi (Hymenoptera: Formicidae). Psyche, 94:201-214. Peeters, C., and Ito, F. 2001. Colony dispersal and the evolution of queen morphology in social Hymenoptera. Annual review of entomology, 46:601-630. Pfennig, K.S. and Pfennig, D.W. 2009. Character displacement: ecological and reproductive responses to a common evolutionary problem. Quarterly Review of Biology, 84:253-276. Pie, M.R. and Traniello, J.F.A. 2007. Morphological evolution in a hyperdiverse clade: the ant genus Pheidole. Journal of Zoology, 271:99-109. Pie, M.R. and Tschá, M.K. 2013. Size and shape in the evolution of ant worker morphology. PeerJ, 1, e205.

Powell, S. 2008. Ecological specialization and the evolution of a specialised caste in Cephalotes ants. Functional Ecology, 22:902-911. Powell, S. 2009. How ecology shapes caste evolution: linking resource use, morphology, performance and fitness in a superorganism. Journal of Evolutionary Biology, 22:1004-1013. Powell, S. and Franks, N. R. 2005. Caste evolution and ecology: A special worker for novel prey. Proceedings of the Royal Society of London Series B, 272:2173-2180. Powell S. and Franks N. R. 2006. Ecology and the evolution of worker morphological diversity: A comparative analysis with Eciton army ants. Functional Ecology, 20:1105-1114. Purcell, J., Brelsford, A., Wurm, Y., Perrin, N., & Chapuisat, M. 2014. Convergent genetic architecture underlies social organization in ants. Current Biology, 24:2728-2732. Pusey, A. E. 1987. Sex-biased dispersal and inbreeding avoidance in birds and mammals. Trends in Ecology and Evolution, 2:295-299. Queller, D. C. 1993. Genetic relatedness and its components in polygynous colonies of social insects. In: Queen Number and Sociality in Insects, pages132-152. Ed. Keller, L. Oxford University Press, Oxford. Rettenmeyer, C. W. 1963. Behavioural studies of army ants. University of Kansas Sciences Bulletin, 44:281-465. Revell, L. J. 2010. Phylogenetic signal and linear regression on species data. Methods in Ecology and Evolution, 1:319-329. Revell, L. J. 2012. phytools: An R package for phylogenetic comparative biology (and other things). Methods in Ecology and Evolution, 3:217-223. Richmond, A. D. and Price, T. 2002. Evolution of ecological differences in the Old World leaf warblers. Nature, 355:817-821. Ricklefs, R. E. and Travis, J. 1980. A morphological approach to the study of avian community organization. Auk, 97:321-338.

Rissing, S.W. 1981. Foraging specializations of individual seed-harvester ants. Behavioural Ecology and Sociobiology, 9:149-152. Ross, K. G., & Keller, L. 1995. Ecology and evolution of social organization: insights from fire ants and other highly eusocial insects. Annual Review of Ecology and Systematics, 26:631-656. Ross, K. G., & Keller, L. 1998. Genetic control of social organization in an ant. Proceedings of the National Academy of Sciences, 95:14232-14237. Roulston, T. H., Buczkowski, G., & Silverman, J. 2003. Nestmate discrimination in ants: effect of bioassay on aggressive behavior. Insectes Sociaux, 50:151-159. Sacchet, F., Botton, M. and Diehl, E. 2009. Ant species associated with the disperal of Eurhizococcus brasiliensis (Hemple in Wile) (Hemiptera:Margarodiidae) in the vineyards 100

of Serra Gaucha, Rio Grande do Brazil. Sociobiology, 54:943-954. Schluter, D. 2000. The ecology of adaptive radiation. Oxford University Press, Oxford. Schoener, T. W. 1967. The ecological significance of sexual dimorphism in size in the lizard Anolis conspersus. Science, 155:474-477. Selander, R.K. 1966. Sexual dimorphism and differential niche utilization in birds. Condor, 68:113-151. Sempo, G. and Detrain, C. 2010. Task regulation in the dimorphic ant, Pheidole pallidula: the influence of caste ratio. Journal of Insect Science, 10:3. doi: 10.1673/031.010.0301.

Sernander, R. 1906. Entwurf einer Monographie der europäischen Myrmekochoren. Almqvist & Wiksells Boktryckeri-A.-B., Uppsala & Stockholm. Shattuck, S.O. 1992. Generic revision of the ant subfamily Dolichoderinae (Hymenoptera: Formicidae). Sociobiology, 21:1-181. Shik, J. Z., Donoso, D. A., and Kaspari, M. 2013. The life history continuum hypothesis links traits of male ants with life outside the nest. Entomologia Experimentalis et Applicata, 149:99- 109. Shine, R. 1989. Ecological causes for the evolution of sexual dimorphism: a review of the evidence. Quarterly Review of Biology, 64:419-461. Simberloff, D. 2009. The role of propagule pressure in biological invasions. Annual Review of Ecology, Evolution, and Systematics, 40:81-102. Smith, C. R., & Tschinkel, W. R. 2006. The sociometry and sociogenesis of reproduction in the Florida harvester ant, Pogonomyrmex badius. Journal of Insect Science, 6:32. doi: 10.1673/2006 06 32.1.

Suarez, A. V., Holway, D. A., & Ward, P. S. 2005. The role of opportunity in the unintentional introduction of nonnative ants. Proceedings of the National Academy of Sciences of the United States of America, 102:17032-17035. Suarez, A. V., Holway, D. A. and Tsutsui. N. D. 1998. Genetics and behavior of a colonizing species: the invasive Argentine ant. The American Naturalist, 172:72-84. Suarez, A. V., Tsutsui, N. D., Holway D. A. and Case, T. J. 1999. Behavioral and genetic differentiation between native and introduced populations of the Argentine ant. Biological Invasions, 1:43-53. Thomas, G. H. and Freckleton, R. P. 2011. MOTMOT: models of trait macroevolution on tree. Methods in Ecology and Evolution, 3:145-151. Thomas, M. L., Payne-Makrisa, C. M., Suarez, A. V., Tsutsui, N. D., and Holway, D. A. 2006. When supercolonies collide: territorial aggression in an invasive and unicolonial social insect. Molecular Ecology, 15:4303-4315. 101

Traniello, J. F. A., and Beshers S. N. 1991. Polymorphism and size-pairing in the harvester ant Pogonomyrmex badius: a test of the ecological release hypothesis. Insectes Sociaux, 38:121-127. Trivers, R. L., & Hare, H. 1976. Haploidploidy and the evolution of the social insect. Science, 191:249-263. Tschinkel, W. R. 1992. Brood raiding and the population dynamics of founding and incipient colonies of the fire ant, Solenopsis invicta. Ecological Entomology, 17:179-188. Tschinkel, W. R. 1993. Sociometry and sociogenesis of colonies of the fire ant Solenopsis invicta during one annual cycle. Ecological Monographs, 63:425-457.

Tschinkel, W. R. 2013. The Morphometry of Solenopsis Fire Ants. PloS ONE, 8: e79559. Tschinkel, W. R., & Howard, D. F. 1983. Colony founding by pleometrosis in the fire ant, Solenopsis invicta. Behavioral Ecology and Sociobiology, 12:103-113. Tschinkel, W. R. 1998. Sociometry and sociogenesis of colonies of the harvester ant, Pogonomyrmex badius: Worker characteristics in relation to colony size and season. Insect Sociaux, 45:385-410. Tsuji, K., 1990. Nutrient storage in the major workers of Pheidole ryukyuensis (Hymenoptera: Formicidae). Applied Entomology and Zoology, 25:283-287. Tsutsui, N. D., & Case, T. J. 2001. Population genetics and colony structure of the Argentine ant (Linepithema humile) in its native and introduced ranges. Evolution, 55:976-985. Tsutsui, N. D., Suarez, A. V., Holway, D. A. and Case, T. J. 2001. Relationship among native and introduced populations of the Argentine ant (Linepithema humile) and the source of introduced populations. Molecular Ecology, 10:2151-2161. Tsutsui, N. D. and Suarez, A. V. 2013. The colony structure and population biology of invasive ants. Conservation Biology, 17:48-58. Tsutsui, N. D., Suarez, A. V., Holway D. A. and Case, T. J. 2000. Reduced genetic variation and the success of an invasive species. Proceedings of the National Academy of Sciences USA, 97: 5948-5953. Vargo and Porter, S. 1989. Colony reproduction by budding in the polygyne form of Solenopsis invicta (Hymenoptera: Formicidae). Annals of the entomological society of America. 82:307-313. Vargo, E. L., and Fletcher, D. J. 1989. On the relationship between queen number and fecundity in polygyne colonies of the fire ant Solenopsis invicta. Physiological Entomology, 14:223-232. Venables, W. N., and Ripley, B. D. 2002. Modern applied statistics with S. Springer, New York City. Vogel, V., Pedersen, J. S., d’Ettorre, P., Lehmann, L. and Keller, L. 2009. Dynamics and genetic structure of Argentine ant supercolonies in their native range. Evolution, 63:1627-1639. Wainwright, P. C. and Reilly, S. M. 1996. Ecological morphology: Integrative 102

Organismal Biology. University of Chicago Press, Chicago. Wainwright, P. C. and Richard, B. A. 1995. Predicting patterns of prey from morphology with fishes. Environmental Biology of Fishes, 44:97-113. Walsh, P. S., Metzger, D. A and Higuchi, R. 1991. Chelex 100 as a medium for simple extraction of DNA for PCR-based typing from forensic material. Biotechniques, 10:506-513. Wang, J., Wurm, Y., Nipitwattanaphon, M., Riba-Grognuz, O., Huang, Y. C., Shoemaker, D., & Keller, L. 2013. A Y-like social chromosome causes alternative colony organization in fire ants. Nature, 493:664-668. Ward, P. S., Brady, S. G., Fisher, B. L and Schultz, T. R. 2010. Phylogeny and Biogeography of Dolichoderine Ants: Effects of Data Partitioning and Relict Taxa on Historical Inference. Systematic Biology, 59:342-362.

Wheeler, D. E. 1986. Developmental and physiological determinants of caste in social Hymenoptera: evolutionary implications. American Naturalist, 128:13-34. Wheeler, W. M., 1929. Two Neotropical ants established in the United States. Psyche, 36:89-90. Wheeler, W. M. 1910. Ants: Their structure, development and behavior. Columbia University Press, New York City. Whitford, W. G., Depree, D. J., Hamilton, P., and Ettershank, G. 1981. Foraging ecology of seed-harvesting ants, Pheidole spp. in a Chihuahuan desert ecosystem. American Midland Naturalist, 150:159-167. Wiernasz, D. C., Sater, A. K., Abell, A. J., & Cole, B. J. 2001. Male size, sperm transfer, and colony fitness in the western harvester ant, Pogonomyrmex occidentals. Evolution, 55:324-329.

Wild, A. L. 2004. Taxonomy and distribution of the Argentine ant, Linepithema humile (Hymenoptera: Formicidae). Annals of the Entomological Society of America, 97:1204-1215.

Wild, A. L. 2007. Taxonomic revision of the ant genus Linepithema (Hymenoptera: Formicidae). UC Publications in Entomology. Volume 26. Wild, A. L. 2009. Evolution of the Neotropical ant genus Linepithema. Systematic Entomology, 34:49-62. van Wilgenburg, E. Torres, C. W., Tsutsui, N. D. 2010. The global expansion of a single ant supercolony. Evolutionary Applications, 3:136-143. Williamson, M. 1996. Biological Invasions. Springer, New York City. Willott, S. J., Compton, S. G., and Incoll, L. D., 2000. Foraging, food selection and worker size in the seed harvesting ant Messor bouvieri. Oecologia, 125:35-44. Wilson, E. O. 1953. The Origin and Evolution of Polymorphism in Ants. The Quarterly Review

103

of Biology, 28:136-156.

Wilson, E. O., 1974. The soldier of the ant Componotus (Colobopsis) flaxinicola as a trophic caste. Psyche, 81:182-188. Wilson, E. O., 1976a. The organization of colony defense in the ant Pheidole dentana Mayr (Hymenoptera: Formicidae). Behavioural Ecology and Sociobiology, 1:63-81. Wilson, E. O., 1976b. Behavioural discretization and the number of castes in an ant species. Behavioural Ecology and Sociobiology, 1:141-154. Wilson, E. O. 1978. Division of labor in fire ants based on physical castes (Hymenoptera: Formicidae: Solenopsis). Journal of the Kansas Entomological Society, 51:615-636. Wilson, E. O. 1980. Caste and division of labor in leaf-cutter ants (Hymenoptera: Formicidae: Atta). Behavioral Ecology and Sociobiology, 7:157-175. Wilson, E. O. 1984. The relation between caste ratios and division of labor in the ant genus Pheidole (Hymenoptera: Formicidae). Behavioural Ecology and Sociobiology, 16:89-98. Wilson, E. O. 1985. The sociogenesis of insect colonies. Science, 228:1489-1495.

Wilson, E. O. 2000. Sociobiology: The new synthesis. Harvard University Press, Cambridge. Wilson, E.O. 2003. Pheidole in the New World. Harvard University Press, Cambridge. Wilson, E. O., and Hölldobler, B. 2005. Eusociality: origin and consequences. Proceedings of the National Academy of Sciences of the United States of America, 102:13367-13371.

Zelditch, M. L, Swiderski, D. L.; Sheets, H. D.; Fink, W. L. 2004. Geometric Morphometrics for Biologists: A Primer. Elsevier Academic Press, San Diego. Zwickl, D.J., 2006. Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion. Ph.D. dissertation, The University of Texas at Austin.

104