Ants As Flower Visitors and Their Effects on Pollinator Behavior and Plant Reproduction

Ants as flower visitors and their effects on pollinator behavior and plant reproduction

Adam Richard Cembrowski

A thesis submitted in conformity with the requirements for the degree of Masters of Science Ecology and Evolutionary Biology University of Toronto

Ants as flower visitors and their effects on pollinator behavior and plant reproduction

Adam Cembrowski

Masters of Science

Ecology and Evolutionary Biology University of Toronto

2013

Abstract

Ants regularly visit flowers, but they may decrease plant reproductive success by competing with pollinators or damaging reproductive structures. However, how ants may exert these costs needs further clarification. In Chapter 1, I review the literature, finding that flower-visiting ants often have neutral effects on plant fitness. In Chapter 2, using artificial flowers with male and female function, I investigate how interference competition between flower-visiting Myrmica rubra ants and Bombus impatiens bumblebees changes pollen analogue movement patterns. Ant presence and scent significantly reduced pollen analogue donation and reception because bees avoided flowers with ant cues. In Chapter 3, to assess the frequency of palynivory among ants, I conducted an acetolysis survey of 75 neo- and paleo-tropical ant species. Ants consistently contained low numbers of pollen grains, suggesting opportunistic pollen consumption may be widespread. Altogether, ant visitation may be costly, but the mechanism depends on the plant, pollinator and ant identities.

Acknowledgments

Science is never an individual undertaking. Numerous people made the completion of this thesis possible. First and foremost, Megan Frederickson gave me the freedom to pursue research outside of her realm of expertise, was extremely patient my ever-changing interests and supported the decisions I made in my degree. My committee also deserves strong thanks. James Thomson was instrumental in my studies, giving me lab space, supplies and, most importantly, time and guidance when I had questions about anything pollination, bee or flower related. Ben Gilbert gave statistics help and was helpful with coding questions. As well, thank you to my examination committee, Spencer Barrett and Helen Rodd.

The Frederickson and Thomson lab groups helped me immeasurably, giving encouragement, guidance, and reality checks when I went too far off track. Thank you to Kyle Turner, Lina Arcila Hernandez, Kirsten Prior, Jane Ogilvie, Alison Parker, Eric Youngerman, and Rebecca Batstone. Kyle especially helped me in all regards both inside and outside of school and was, without question, the best field roommate and officemate I could have had. Several volunteers, work study students and our lab tech fed ants, washed artificial flowers and counted tiny objects, saving me countless hours, particularly Harry Rusnock, Shannon Meadley Dunphy, Margaret Thompson, and Jackie Day. I would also like to highlight Marcus Guorui Tan in particular for his work collaborating with me and running trials when I was in out of town and being unwaveringly positive about our research. Our experiment would not have been nearly as successful without him.

Several people in the department and associated with the lab provided feedback, jokes and friendship, without which grad school would have been a depressing place. So thank you Natalie Jones, Rachel Germaine, Kelly Carscadden, Alex de Serrano, Jordan Pleet, Aaron Hall, Susana Wadgymar, Emily Austen, Eddie Ho, Amanda Gorton, Dorina Szuroczki, Jenn Coughlan, Jon Sanders, and Gabe Miller.

People not in the department and both inside and outside of ecology, made my time here possible and more enjoyable than otherwise. Thanks to the Tuner family for taking me in during holidays, Maureen Murray and Matt Mazowita for everything, and my old supervisor Colleen Cassady St. Clair for helping me get where I am.

iii

Lastly, my family did an astounding amount for me in letting me cultivate my love of biology, buying me books, letting me keep strange animals, and getting me outdoors. So, thank you to my brothers John and Mark, my dad George, and my mom Kay. And to my partner Katherine, for her unbelievable patience, both in regards to my academic life and in her willingness to adapt hers for mine, thank you. This could not have happened without you.

“Vor Allem sind aus der Reihe der flügellosen Insecten die weitverbreiteten flügellosen Ameisen sehr unwillkommene Gäste der Blüthen. Und dennoch sind gerade sie nach dem Nectar der Blüthen in hohem Grade lüstern…”

“Of all the wingless insects it is the widely dispersed ants that are the most unwelcome guests to flowers. And yet they are the very ones which have the greatest longing for nectar…”

A. Kerner 1878, Flowers and their Unbidden Guests

Table of Contents

Acknowledgments...... iii

List of Tables ...... vii

List of Figures ...... viii

Chapter 1 Ants as flower visitors and their effects on pollinator behavior and plant reproduction ...... 1

Abstract ...... 1

Introduction ...... 2

1. Ant interactions with flowers and floral visitors ...... 3

2. Ants as pollinators ...... 12

3. Floral defenses against ants ...... 14

4. Future directions ...... 20

References ...... 24

Tables ...... 39

Chapter 2 Ants and ant scent reduce bumblebee pollination of artificial flowers ...... 54

Abstract ...... 54

Introduction ...... 55

Methods ...... 57

Results ...... 62

Discussion ...... 63

Acknowledgements ...... 66

References ...... 68

Figures ...... 73

Chapter 3 Not just for the bees: pollen consumption is common among tropical ants ...... 75

Abstract ...... 75

Introduction ...... 75

Methods ...... 77

Results ...... 78

Discussion ...... 79

Acknowledgements ...... 82

References ...... 83

Tables ...... 86

Figures ...... 91

Concluding Remarks ...... 93

List of Tables

Table 1.1 List of studies of flower-visiting ants that consumed floral rewards or interacted with pollinators and included a measure of fitness...... 39

Table 1.2 List of studies that have either demonstrated ant pollination through exclusion experiments or whose results strongly suggest it...... 48

Table 3.1 Numbers of adult worker and larval ants examined for each species collected in Peru and numbers of pollen grains found...... 87

Table 3.2 Numbers of adult worker and larval ants examined for each species collected in Papua New Guinea and numbers of pollen grains found...... 90

vii

List of Figures

Figure 2.1 Photograph of artificial flower lids with anther and stigma...... 74

Figure 2.2 Dye transferred (mean ± 1SE) by B. impatiens a) in the presence and absence of M. rubra ants and b) with or without M. rubra scent...... 75

Figure 3.1 Photographs of pollen grains found inside ants...... 92

Figure 3.2 Relationship between trophic level and pollen presence for neotropical ant species...... 93

viii

Chapter 1 Ants as flower visitors and their effects on pollinator behavior and plant reproduction

Cembrowski, A.R., and Frederickson, M.E

Planned submission to a journal such as Oikos or Insectes Sociaux ARC wrote the manuscript with input from MEF

Abstract

Ants often provide a range of benefits to flowering plants, particularly by reducing herbivory.

However, when ants visit flowers, they can deter pollinators or damage floral structures.

Currently, it is unclear how often, and in what direction, flower-visiting ants affect plant reproductive success. A review of the literature shows that ants can help or hinder pollination processes, but overall flower-visiting ants most commonly have a non-significant net effect on plant fitness. Although ants may directly harm flowers or indirectly affect pollination by consuming floral nectar or harassing pollinators, they can also benefit plants by attacking florivores or predators of pollinators. Ants occasionally act as pollinators themselves, though ant secretions often kill pollen grains. There is growing evidence that many plants have traits that allow them to avoid or minimize the negative effects of flower-visiting ants. More research is needed to determine if these traits provide fitness benefits to plants by preventing ant visitation, or if they function as generalized repellents of floral antagonists.

1 2

Introduction

Ants and angiosperms share a long evolutionary history. Ant diversification closely followed that of flowering plants, as ants took advantage of the large prey numbers that angiosperms supported and the new habitats they provided (Moreau 2006). Subsequently, numerous angiosperm lineages evolved intimate associations with ants that function as an indirect or biotic form of plant defense. Food rewards such as extrafloral nectar or pearl bodies, or housing in the form of domatia, attract ants to plants, where they reduce herbivory and increase plant fitness (reviewed in Rosumek et al. 2009). However, the traits that make ants good plant defenders, particularly their abundance, attraction to sugar sources, and aggressiveness, can also cause problems for plant reproduction. Ants, attracted by floral nectar, can compete with pollinators for resources and defend flowers, potentially reducing plant fitness (Ness 2006, Junker et al. 2007). That ants negatively affect plant reproduction is an old and pervasive hypothesis; Kerner (1878) called ants

“most unwelcome guests” on flowers and more recent authors have similarly expressed that ant visitation to flowers reduces plant reproductive success (Ness 2006, Lach 2007, Willmer et al.

2009).

Nonetheless, studies assessing the impact of flower-visiting ants on plant reproductive success have produced mixed results. Ants can have negative impacts on pollination and flower function (Galen 1983, Puterbaugh 1998, Ness 2006, Hansen and Müller 2009). However, other studies have found no effect, or even strong benefits, of ant visits to flowers (Rico-Gray and

Thien 1989, Altshuler 1999, Schürch 2000, Ashman and King 2005). This discrepancy, between ants being “unwelcome guests” and their actual effects on plant reproduction, requires clarification. Our goal here is to highlight the ways that flower-visiting ants may affect plant reproduction, both positively and negatively, and to discern, through a review of the literature,

3 what impacts ants have on plant reproductive success. First, we discuss mechanisms by which flower-visiting ants can directly and indirectly impact plant reproduction, both through floral damage and by mediating changes in pollinator or florivore behavior. We end this section by evaluating published research that collectively shows that flower-visiting ants have predominantly neutral effects on plant fitness measures. Next, we discuss pollination by ants, focusing on the “antibiotic hypothesis,” before describing how plants may prevent ants from visiting flowers. Lastly, we discuss promising new avenues of research and highlight work that still needs to be performed to give a more complete understanding of ants as flower visitors.

1. Ant interactions with flowers and floral visitors i) Damage to flowers

Ants can directly disrupt plant reproduction by damaging floral structures, preventing either the fertilization of ovules or the presentation of pollen. Rarely, ants forage on floral tissue itself; for example, Puterbaugh (1998) found that Formica neorufibarbus ants chewed through the coronal ring of Eritrichium aretiodes flowers to get lipids. More often, but still not commonly, foraging ants chew through reproductive tissue to access floral nectar. This can cause sterilization of whole flowers (Echium plantagineum, Kirk 1984), or affect female function only if just styles are damaged, effectively creating male flowers (Polemonium viscosum, Galen 1983). Antibiotic secretions on ant integuments may also directly damage pollen grains, but the fitness effects of this are unclear (see 2, below).

Several species of Crematogaster and Allomerus ants sterilize their host plants by attacking floral buds (Yu and Pierce 1998, Stanton et al. 1999, Gaume et al. 2005, Malé et al.

2012, Table 1.1). These ants nest in “ant-plants” or “myrmecophytes”—i.e., plants that host ant

4 colonies in specialized structures (domatia). The sterilizing ants do not appear to consume floral tissues (Yu and Pierce 1998) and two hypotheses may explain why ants destroy floral buds. First, ant colony size is often limited by domatia availability on their host plant. By destroying floral buds, ants can cause plants to divert resources from reproductive to vegetative growth, increasing domatia number and allowing for larger ant colonies (Yu & Pierce 1998; Frederickson 2009).

Second, when plant-ants must defend their colony from takeovers by other colonies, they may benefit from limiting access to their host plant. By destroying buds that could develop into flowers, branches or leaves, ants can prevent their host plant from contacting nearby plants occupied by enemy ant colonies (Stanton et al. 1999).

The impact of sterilization on lifetime plant fitness is unclear. Ants may destroy all of their host plant’s floral buds (Yu and Pierce 1998), completely preventing reproduction, or only a portion of them (Malé et al. 2012), allowing some seed set. However, counter-intuitively, this sterilization behavior may provide long-term fitness benefits to plants (Frederickson 2009,

Palmer et al. 2010). Ant-plants often host several different ant species over their lifetimes, some of which are non-sterilizing. If hosting a sterilizing species at a certain life stage increases survival or growth relative to hosting a non-sterilizing species at that life stage, plants may increase lifetime fitness by largely forgoing current reproduction in favor of increasing future reproduction (Palmer et al. 2010). Why plants invest resources in producing flowers that are quickly destroyed by ants is not known, but the timing of flower destruction may be important. If ants attack flowers after anthers dehisce, plants may still realize some male fitness via pollen donation (Yu and Pierce 1998). Additionally, if the cost of flower production is relatively minor and ants sometimes “miss” flowers (Edwards and Yu 2008), attempting limited reproduction may be a better strategy than forgoing it entirely.

ii) Competition with pollinators

Ant activity on flowers can dramatically alter pollinator behavior. Ants can reduce the species diversity, frequency, or duration of pollinator visits to flowers (Norment 1988, Tsuji et al. 2004,

Ness 2006, Junker et al. 2007, Lach 2007, 2008b, Romero et al. 2011, and see Chapter 2). These changes are largely due to either exploitative competition (i.e., consumption of shared resources) or interference competition (i.e., direct aggression) between ants and pollinators.

a) Exploitative competition

Pollinators and ants frequently compete for floral nectar. Flower-visiting ants are often classified as nectar thieves (sensu Inouye 1980) because they typically enter flowers through the

“legitimate” opening and consume nectar, but transfer no pollen (Wyatt 1980, Ness 2006,

Chamberlain and Holland 2008). Ants can also act as primary nectar robbers, chewing holes through corollas to reach nectaries (Willmer and Corbet 1981, Koopowitz and Marchant 1998), although they are more commonly secondary nectar robbers and use holes made by previous robbers (Newman and Thomson 2005).

Like other flower visitors, foraging ants will often leave nectar behind in flowers. This amount can be less (Bleil et al. 2011), more (Fritz and Morse 1982), or similar to (Lach 2008a) amounts left by other visitors. Often, ants are morphologically constrained from fully exploiting nectar. Unlike bees and butterflies, ants do not have proboscises and must get close to nectar to consume it. Narrow corollas or nectar spurs can prevent ant access to some or all floral nectar

(Schemske 1980, Galen and Cuba 2001, and see below). Smaller ants may be better able to

6 access nectar (Newman and Thomson 2005, Agarwal and Rastogi 2007), but can be limited by their crop size in the amount of nectar they can remove (Lach 2005). If foraging ants leave enough nectar behind for a pollinator to recover the cost of visiting a flower, ants may have minimal effects on plant reproductive success (Maloof and Inouye 2000, Turner et al. 2012).

Differences among ant species in behavior and diet also affect their consumption of floral nectar. Species vary in how many flowers they visit per inflorescence or per plant, and in how many ants will visit a flower simultaneously (Lach 2005, Junker et al. 2007, Lach 2013).

Invasive ants—thought to be better at exploiting and defending resources than native ants

(Holway et al. 2002)—often recruit to flowers in large numbers, sometimes several dozen or more (Visser et al. 1999, Hansen and Müller 2009). Overall, many factors may influence how heavily ants exploit floral nectar on plants, including ant density, plant height, the spatial or temporal distribution of floral nectar, and the availability of alternative food resources for ants

(Cushman and Addicott 1991, Lach 2013).

Rarely, ants and pollinators may compete for pollen. We know of only a few studies that have recorded ants harvesting pollen directly from anthers (Horskins and Turner 1999, Ness

2006, Byk and Del-Claro 2010). Other studies have found pollen grains in ants (e.g., Baroni

Urbani and de Andrade 1997, and see Chapter 3), but this may not indicate competition with pollinators; ants sometimes collect stray pollen grains from the environment (Creighton 1967).

One New World genus, Cephalotes, might be specially adapted for pollen consumption. They possess “proventricular shields”, structures that may restrict solid food to the crop, allowing nestmates to exchange pollen grains via trophallaxis. It is unclear if this structure evolved for this purpose, but Cephalotes are the most frequently documented palynivorous ants (e.g., Creighton

1967, Baroni Urbani and de Andrade 1997, Byk and Del-Claro 2010). How efficiently ants

7 digest pollen, and whether ants, like bees, feed it to larvae as a protein source, is unknown. At least two studies have found that ant larvae consume pollen, providing some support for this idea

(Wheeler and Bailey 1920, and Chapter 3).

b) Interference competition

Ants are relatively unique among flower visitors in defending floral resources from competitors

(but see Willmer and Corbett 1981, Roubik 1982). This defense ranges from chance encounters, with ants attacking visitors when flower occupancy overlaps (Ashman and King 2005, Junker et al. 2007), to actively patrolling inflorescences or constructing protective galleries around nectaries (Horvitz and Schemske 1984, Gaume et al. 2005, Hansen and Müller 2009).

Interference competition, even in the absence of exploitative competition, is sufficient to alter pollinator behavior (Chapter 2).

The strength and outcome of interference competition depends on the identities of the ant and the pollinator. Pollinators might not alter their visitation patterns in response to non- aggressive ants (Junker et al. 2007), but more antagonistic ants may elicit avoidance behavior before or soon after the visitor lands (Willmer and Corbet 1982, Altshuler 1999, Schürch et al.

2000, Table 1.1). Pollinator size may help predict their reaction to ants, because ants may deter only small-bodied pollinators (Gonzálvez et al. 2012). It is not clear whether flower visitors view ants as aggressive competitors or as potential predators; rarely, ants prey on pollinators (Schatz and Wcislo 1999), but in most cases ants merely cause pollinators to relocate.

How well pollinators recognize and avoid ant-visited flowers is unresolved. Collectively, pollinators have lower visitation rates to flowers visited by ants (Romero et al. 2011). This is

8 dependent on pollinator identity, however; dipterans do not appear to discriminate between ant- visited and ant-free flowers, while bees can identify and avoid ant-visited flowers. Bees likely use a combination of visual and olfactory cues to recognize ant-visited flowers. They can identify previous conspecific and some heterospecific insect visitors by the hydrocarbon

“footprints” they leave during foraging (Stout and Goulson 2001). Bumblebees can detect ant hydrocarbons, although this has been shown only in the laboratory (Ballantyne and Willmer

2012, Chapter 2). No study has yet examined if bees can use sight alone to detect ants, but several suggest it (Junker et al. 2007, Gonzálvez et al. 2012). Some bees may discriminate between ant species when visiting flowers (e.g., Lach 2008b), which follows if not all ant- pollinator interactions are equally costly. The ability of bees to do so will affect how generalized their avoidance behaviors will be. If they can discriminate between species, bees may avoid specific flowers visited by more “costly” ants. However, if they cannot, bees might either avoid all ant-visited flowers or none, depending on how negative the average interaction with an ant is.

iii) Interactions with flower and pollinator antagonists

Ants can defend the flowers they visit. Numerous plant species have circumfloral or bracteal nectaries (i.e., extrafloral nectaries around flowers), and ants attracted to these nectaries can reduce bud, ovule or flower predation, increasing plant fitness (Inouye and Taylor 1979, Rico-

Gray and Thien 1989, Sugiura et al. 2006). Although the presence of EFNs (extrafloral nectaries) so close to floral nectaries (sometimes less than 1 cm, Ness 2006) might increase discovery and exploitation of floral nectar by ants, only one study has investigated this hypothesis, finding no relationship between the proximity of EFNs to floral nectaries and ants’ use of floral nectar

(Galen 2005). Holes made by nectar robbers are functionally analogous to circumfloral nectaries

9 and can also attract ants that defend flowers (Newman and Thomson 2005), and nectar-thieving ants can similarly prevent florivory in some systems (Lach 2007, Bleil et al. 2011).

Ants can also reduce the abundance of flower-dwelling predators or parasites of pollinators. Ambush predators such as crab spiders often forage at flowers, where they may reduce pollinator visitation rates and plant fitness (Dukas 2001, Gonçalvez-Souza et al. 2008).

Ants can decrease spider abundance and richness on plants (Rudgers 2004, Renault et al. 2005,

Nahas et al. 2012), and some evidence suggests that they can do the same on flowers (Lach

2008). In select cases, ants may also reduce the abundance of pollinator parasitoids (Holland et al. 2011). Whether ants increase plant reproductive success by removing pollinator antagonists has not yet been examined, but should be. If ants create ‘enemy-free’ flowers, pollinators might prefer to land on flowers with ants.

iv) Net effects

Table 1.1 compiles studies in which the authors recorded ants consuming floral rewards or interacting with pollinators and measured some aspect of plant fitness (e.g., fruit or seed set, pollen deposition). We included only a single study per plant species, unless subsequent studies found conflicting results or measured the fitness of another plant sex. Records of ant pollination are included in a separate table (Table 1.2).

Table 1.1 reveals relatively few examples of flower-visiting ants negatively affecting plant fitness given how costly ant visitation is assumed to be. The net effect of ants on plant reproductive success is actually quite variable, with similar numbers of studies finding positive, negative or neutral effects of ant presence. Surprisingly, when ants do have negative effects, it is

10 usually because they directly damage floral structures and not because they prevent pollinators from visiting flowers and successfully transferring pollen. Why are ants not more often negative for plant reproduction when this is commonly assumed to be true?

First, ants may compensate for reducing pollinator visitation rates by improving pollination quality. As pollinators move between flowers on a plant, they often deposit self- pollen on stigmas (i.e., geitonogamy). In some plants with late-acting self-incompatibility, ovules fertilized by self-pollen are subsequently aborted, reducing the number of seeds that can be produced (Barrett 2002 and references therein). Geitonogamy can also reduce offspring viability or vigor (de Jong et al. 1993) and cause pollen discounting, a reduction in the amount of pollen available for outcrossing because of self-pollination (Barrett 2002). By competing with pollinators, ants might decrease rates of geitonogamous pollination if pollinators then visit fewer sequential flowers within a plant. Ants might also cause pollinators to fly further between plants, potentially increasing the genetic diversity of pollen arriving at stigmas. Although such effects have been reported for other floral larcenists (Maloof 2001, Irwin 2003), to our knowledge, these hypotheses have not been investigated in relation to ants.

Second, ants and pollinators do not always forage for the same resources, possibly allowing them to partition floral rewards. If ants forage for nectar but not pollen, and flowers can be pollinated efficiently by visitors looking to gather mostly pollen and little or no nectar, then plant reproduction will not be affected by ants consuming floral nectar. Junker et al. (2010a) found that honeybees foraged for both pollen and nectar in the absence of ants, but reduced their foraging on nectar but not pollen when ants could access flowers, resulting in no difference in fruit set between treatments. Whether pollinators commonly adjust their foraging behaviors in response to ants warrants further research.

Third, ants are somewhat unique among floral larcenists in that they often provide direct benefits to the plants they visit. In the studies listed in Table 1.1, ants sometimes decreased herbivory on plants (Chamberlain and Holland 2008, Bleil et al. 2011) or otherwise aided the plant (e.g., ant presence repelling smaller, less effective pollinators, Gonzálvez et al. 2012). If the positive effects of ants cancel or outweigh the negative effects of flower-visiting ants on pollinators, then ant visitation may be neutral or even positive for plant fitness.

Fourth, ants may impose fitness costs to plants not captured in the studies in Table 1.1.

Nectar robbers have stronger negative effects when plants are pollen-limited (Burkle et al. 2007), but only one study in Table 1.1 assessed whether flowers would set more seed if they received more pollen, finding no evidence of pollen limitation (Lach 2008a, Table 1.1). Even if ants decrease pollinator visitation rates, there is limited scope for ants to decrease plant reproductive success when plants are not pollen-limited. Furthermore, nearly all the studies in Table 1.1 measured only female fitness. Only three quantified aspects of male fitness (Wyatt 1980, Fritz and Morse 1981, Galen and Geib 2007) and we know of only one other study, in a system in which ants occasionally pollinate, that has also measured male fitness (Ashman 2000). All four studies measured pollen or pollinia removal rates, and none found any effect of flower-visiting ants on male fitness. However, whether other measures of male fitness (e.g., seeds sired, pollen deposition patterns) would reveal similar or different effects needs investigation.

Finally, although floral defenses against ants (see 3, below) may prevent us from observing the negative effects of ants on plant reproductive success, we consider this unlikely.

Several studies have examined interactions between ants and plants that share little or no coevolutionary history, usually in the context of ant invasions. Although invasive ants often recruit to flowers in larger numbers than native ants, they do not appear to reduce plant

12 reproductive success (Blancafort and Gómez 2005, Lach 2007, Lach 2008a, but see Hansen and

Müller 2009). And in Hawaii, where angiosperms evolved for millions of years in the absence of ants and are thought to have secondarily lost floral defenses against ants (Junker et al. 2011a), there is no evidence that the preponderance of recently introduced flower-visiting ants negatively affects plant fitness (Junker et al. 2010a, Bleil et al. 2011).

2. Ants as pollinators

Ants may directly benefit plant reproduction by acting as pollinators, but this is a relatively rare occurrence. Ants have long been considered poor pollinators (Kerner 1878). Initially, this was based on morphological and behavioral differences between ants and recognized pollinators.

Worker ants are wingless and often relatively hairless, making them poor pollen dispersers. As well, because many ants exhibit Ortstreue (literally, “place fidelity”), ants were thought to return to the same branch or flower repeatedly, effecting no cross pollination (Hölldobler and Wilson

1990). Though largely true, these traits do not necessarily preclude ants from acting as pollinators (Table 1.2). Although they commonly have restricted foraging ranges (Hölldobler and

Wilson 1990), some ants can effect outcrossing (e.g., Ashman and King 2005, Carvalheiro 2008) and many ants are as hairy, if not hairier, than their pollinating counterparts.

In the late 1970s, the “antibiotic hypothesis” was put forward to explain the rarity of ant pollination. Ants use antibiotic secretions to protect themselves against the fungal and bacterial pathogens common in their nests and environs. Iwanami and Iwadare (1978) found that one of these compounds, myrmicacin, also greatly reduces pollen germination rates. Subsequently,

Beattie et al. (1984, 1985) found that short periods of contact (≤ 30 minutes) with an ant’s integument caused pollen grains to fail to germinate. Some of these pollen-killing compounds

13 were traced back to the ant metapleural gland, which produces hygienic secretions (Hölldobler and Wilson 1990), although contact with ants lacking metapleural glands can also reduce pollen viability (Beattie et al. 1985, Dutton and Frederickson 2012). Ant mandibular or anal gland secretions, and potentially ant venom, may likewise defend ants against pathogens and have similar effects on pollen germination (Beattie et al. 1985, Graystock and Hughes 2011).

Nonetheless, ant pollination has been documented repeatedly. There are at least 18 cases in which ant pollination has been shown experimentally and five others in which observations strongly suggest it (Table 1.2). The majority of these plants fit the “ant pollination syndrome” originally proposed by Hickman (1974): dense stands of short plants with small flowers and easily accessible nectaries. Typically, ant-pollinated plants are associated with hot and dry climates (principally in the Mediterranean), but dry alpine environments may also favor the evolution of ant pollination (e.g., Svensson 1985, Puterbaugh 1998, Schiestl and Glaser 2011). In most cases, ants appear to be the sole or major pollinator of the plant in question (Hickman 1974,

Peakall et al. 1990, de Vega et al. 2009), although ants sometimes make only supplementary contributions to pollination (Schürch et al. 2000, Ashman and King 2005). This suggests that coevolution between ants and plants may not be necessary for ants to be useful pollinators—which is surprising, given the potential negative effects of ants on plant reproduction.

Ant-pollinated plants cope with ant secretions in a variety of ways. Several plant species may simply tolerate pollen damage (Gόmez and Zamora 1992, de Vega et al. 2009), apparently possessing no overt adaptations to prevent it. These plants might receive enough pollen to overcome reduced pollen viability. Other plants may place pollen on ant body parts thought to have fewer secretions, thereby potentially escaping damage (e.g., on ant legs, Ramsey 1995), or attach pollen to ants with specialized structures, avoiding direct contact with the ant integument

(Peakall et al. 1990). It is not known whether pollen evolves resistance to ant secretions, or whether plants deposit pollen on ants which have less disruptive secretions, although these possibilities have been suggested elsewhere (e.g., García et al. 1995, Ramsey 1995).

One key question regarding the effects of ant secretions on pollen grains urgently needs to be resolved. Several studies have suggested that the pollen-killing secretions of ants may have contributed to the evolution of floral defenses against ants (e.g., Wagner 2000, Galen and Geib

2007). For this to be true, ants must decrease male fitness by damaging pollen grains before they leave flowers and are deposited on receptive stigmas. Only two studies have investigated the effects of ant secretions on pollen still in anthers (Wagner 2000, Galen and Butchard 2003) and both constrained ants unnaturally to flowers for significantly longer than ants would normally stay on flowers. Although these studies found strong reductions in pollen vigor, it is unclear if ant secretions decrease male fitness under natural ant visitation conditions.

3. Floral defenses against ants

Plants are not passive recipients of flower visitors and have some ability to “choose” which organisms can access floral rewards. Flowers resist ants using both structural and chemical traits.

More thorough investigation is needed to determine whether these traits evolved to prevent or minimize the negative effects of ants on plant reproductive success (i.e., whether these traits can truly be considered defenses against ants, see 4ia, below). It is unlikely, however, that ants experience reciprocal selection to circumvent floral resistance traits because the fitness benefits to ants of accessing flowers are probably small. Thus, there may be little scope for coevolutionary arms races between floral defenses against ants and ant counter-adaptations.

i) Structural resistance traits

a) Restricting ant access

Numerous plant species have physical structures preventing ant access to flowers. Thin pedicels can limit ants’ gripping abilities and make ants easier to dislodge when flowers are moved by wind (Gaume et al. 2005), and waxy secretions can render stems too slippery for ants to climb

(Harley 1991). Structures surrounding flowers, such as water moats or dense hairs, may similarly keep ants from accessing floral rewards (Willmer et al. 2009). Many such traits function as broad filters and prevent not only ants, but most crawling insects from reaching flowers. Thus, it is difficult to determine whether ants, florivores, or both acted as selective agents favoring the evolution of physical barriers to flowers.

Flower shape may also reduce ant-pollinator conflict. Long, narrow corollas or nectar spurs can prevent all but the smallest ants from accessing floral nectaries (Blüthgen et al. 2004,

Agarwal and Rastogi 2007), while still allowing long-tongued pollinators access to rewards.

Selection to exclude nectar-thieving ants in such a manner, however, may be counteracted by pollinator preferences for wider corollas that allow easier flower handling (Galen and Cuba

2001). Additionally, these types of flowers may be more prone to robbery (Irwin and Maloof

2002), limiting their usefulness in preventing nectar larceny.

b) Separation in space or time

Plants may separate floral rewards and ants temporally or spatially to minimize ant impacts on plant reproduction. Several authors have proposed that myrmecophytes have adapted to the

16 sterilization or otherwise detrimental activities of ants by locating domatia or food rewards on different branches than flowers (Izzo and Vasconcelos 2002, Raine et al. 2002, Edwards and Yu

2008). Temporal separation of ant and pollinator activities could likewise reduce conflict, though evidence is limited. Some acacia flowers dehisce during periods of low ant activity (Willmer and

Stone 1997, Nicklen and Wagner 2006), but it is unclear if this is in response to ants, temperature, or pollinator activity. Timing of nectar secretion may similarly reduce ant-pollinator competition. If ants are active prior to pollinators, they may deplete floral nectar and reduce pollinator visitation rates (Schaffer et al. 1984). However, if nectar is secreted when ants are inactive, pollination may occur unimpeded (Norment 1988). This temporal separation could favor pollinators that are active when ants are not (e.g., Wyatt and Morse 1981), but this is poorly documented. Further studies, particularly focusing on flowers with the potential for nocturnal and diurnal pollination, may provide more examples.

c) Extrafloral nectaries (EFNs)

One hypothesis put forward to explain the evolution of EFNs proposes that EFNs “distract” ants from consuming floral nectar (Kerner 1878, Wagner and Kay 2002). Wagner and Kay (2002) tested this hypothesis by constructing model plants using artificial nectaries, assigned at random to represent either floral or extrafloral nectaries. They found that the presence of EFNs could reduce ant exploitation of floral nectar. However, as highlighted by Galen (2005), this design is closer to a plant making a surplus of flowers, some of which will avoid ant visitation if visitation rates increase sublinearly with increasing flower number. Producing extra flowers to escape ant visitation appears to be a viable strategy in some systems (Ashman and King 2005, Galen 2005).

However, Galen (2005) found that applying honey (simulating extrafloral nectar) to P. viscosum

17 leaves had the opposite effect and increased ant visits to flowers. Only two studies, both in

Pachycereus schottii, have found evidence for the distraction hypothesis under field conditions

(Chamberlain and Holland 2008, Holland et al. 2011).

There is limited empirical support for the distraction hypothesis, but abundant evidence that EFNs attract ants that defend plants against herbivores (reviewed in Rosumek et al. 2009).

To function as attractants, the presence of EFNs must increase ant abundance on plants.

However, because of the increased numbers attracted, EFNs would not be expected to reduce the abundance of ants on nectar rewards or be able to satiate the ants recruited. Only if ant abundance stops increasing at a reward density much lower than that available on the plant would EFNs serve to distract ants. In this case, the amount of rewards may overwhelm ants, leaving some flowers underexploited. However, even in P. schottii, ants still visited flowers

(Chamberlain and Holland 2008). In cases where ant abundance would continue to increase with additional rewards, it is unlikely that EFNs function as distractions.

ii) Chemical resistance traits

a) Toxic nectar

One of the earliest proposed floral defenses against ants was toxic nectar (Janzen 1977).

However, studies testing Janzen’s (1977) hypothesis uncovered only a very few isolated examples (e.g., Catalpa speciosa, Stephenson 1981; Crinum erubescens, Guerrant and Fiedler

1981); in short, most floral nectar is not toxic or repellent to ants (Baker and Baker 1978,

Schubart and Anderson 1978). In fact, ants sometimes prefer floral nectar over solutions with equivalent sugar concentrations (Koptur and Truong 1998).

In contrast, floral nectar can repel other flower visitors (e.g., birds, Johnson et al. 2006; lepidopterans, Kessler and Baldwin 2007). Repellency is often caused by plant secondary metabolites, which may occur in nectar as a by-product of their production elsewhere in the plant, where they function as defenses against herbivory (Adler 2001). Alternatively, toxic or repellent nectar may be adaptive; repellent nectar may promote effective pollination by reducing the amount of reward any one pollinator takes, spreading nectar over several pollinator visits

(Kessler and Baldwin 2007) or it may function as a filter, selecting for specific pollinators

(Johnson et al. 2006). The latter may explain why repellent nectar is not more common; legitimate pollinators, in addition to floral larcenists, may be deterred from visiting flowers with repellent nectar. This may be particularly true of nectars that would repel ants, which are closely related to bees; nectar distasteful to ants may often also be distasteful to bees, and those nectars acting as a filter for bees may still be palatable to ants. For example, Prŷs-Jones and Willmer

(1992) found that the ammonia-containing nectar of Lathraea clandestina is readily consumed by both bees and ants, even though it deters birds that act as nectar thieves. Although floral nectar itself rarely repels ants, Guerrant and Fielder (1981) found that the presence of macerated petals within nectar can often keep ants away. Alkaloids in petals may be responsible and might explain why ants seldom act as primary nectar robbers; these secondary compounds may prevent ants from chewing through corollas.

b) Floral volatiles

Repellent floral volatiles appear to be a more widespread form of chemical resistance against ants. These were reviewed recently by Willmer et al. (2009) and will be discussed here only briefly. Numerous plant taxa produce floral scents repellent to ants, including myrmecophilous,

19 myrmecophytic and non-ant associated plants in both temperate and tropical ecosystems

(Willmer et al. 2009, Junker et al. 2011a, Ballantyne and Willmer 2012). In three large studies, floral volatiles that repel ants were detected in about a third of all plants tested (Willmer et al.

2009, Junker et al. 2011a, Ballantyne and Willmer 2012). Pollen is often the source of ant- repellent volatile compounds (Willmer et al. 2009, Ballantyne and Willmer 2012), but they can be emitted from other floral parts, including petals. That pollen is often the source of repellence could suggest that preventing ant visitation may be more important in protecting male than female function. Additionally, because repellence would decrease as pollen is removed from anthers, ants may return and defend flowers more quickly than if repellent chemicals were in longer-lasting structures (e.g., petals). In some cases, floral volatiles are thought to mimic ant alarm pheromones (Willmer et al. 2009). However, floral scents elicit inconsistent reactions across ant species, ranging from repellence to attraction (Agarwal and Rastogi 2007, Ballantyne and Willmer 2012).

Producing structural or chemical resistance traits may be costly for plants. Native

Hawaiian plants, which have little recent coevolutionary history with ants, appear to have secondarily lost floral ant repellents (Junker et al. 2011a). Both Willmer et al. (2009) and Junker et al. (2011a) found that highly chemical repellent flowers did not have strong physical defenses and vice versa, suggesting trade-offs between the two (but see Ballantyne and Willmer 2012).

Like toxic nectar, floral ant repellents might sometimes deter both pollinators and ants. For example, Galen et al. (2011) found that 2-phenylethanol in P. viscosum floral scent repels both flower-damaging F. neorufibarbis ants and the plant’s main pollinator, Bombus kirbyellus.

Interestingly, four recent studies have found that floral scents can also attract ants

(Agarwal and Rastogi 2007, Edwards and Yu 2008, Schiestl and Glaser 2011, Gonzálvez et al.

2012). One study (Schiestl and Glaser 2011) has shown that floral scents attract ants to a potentially ant-pollinated plant, and such systems may provide many more examples.

4. Future directions

We have highlighted areas in need of further investigation throughout the review. Here, we focus on several questions we consider high priorities for future research.

i) Floral defenses against ants

a) Do floral resistance traits increase plant fitness?

The existence ant-repellent floral traits is often taken as evidence that ants negatively affect plant reproductive success. However, despite the abundance and variety of mechanisms excluding ants from flowers, few studies show that ants reduce plant fitness when they visit flowers (Table 1.1).

This absence is particularly apparent with repellent floral volatiles, which may be the most widespread floral resistance trait (Willmer et al. 2009, Junker et al. 2011a, Ballantyne and

Willmer 2012). To our knowledge, only one study has shown that floral ant repellents benefit plants, with repellent P. viscosum flowers avoiding damaging ant visits (Galen et al. 2011).

Research is urgently needed to investigate if these traits increase plant fitness by excluding ants, or if ant visitation is inconsequential. More studies taking advantage of standing variation in traits (Galen et al. 2011), or creating it artificially (e.g., Junker et al. 2010b) will allow us to distinguish between the two.

b) If ant visitation is often not costly, what caused the evolution of repellent traits and

what maintains them? Similarly, are these repellent traits truly defenses against ant

visitation or are they generalized defenses against a range of floral antagonists?

Ant-repellent floral traits are widespread and their production may be costly (Willmer et al.

2009, Junker et al. 2011a). The studies in Table 1.1, however, indicate that flower-visiting ants often do not disrupt plant reproduction, particularly outside of myrmecophytes. What, then, selected for these traits? One possibility is that these repellents are not directed solely at ants.

Instead, a community of antagonistic flower visitors may exert diffuse selection on plants, thereby selecting for and maintaining these traits. Structural modifications may repel crawling insects, while chemical repellents such as floral volatiles may repel both flying and non-flying insects (Junker et al 2011b). How much overlap there is in floral volatiles that repel both ants and other antagonists needs investigation. Until we examine how specific these defenses are to ants, we risk overstating the importance of ants in their evolution.

ii) Do ants alter pollen movement patterns or reduce male fitness?

We are aware of only four studies that have measured if ant visitation impacts male fitness

(Wyatt 1980, Fritz and Morse 1981, Ashman 2000, Galen and Geib 2007). While none of these studies found an effect, all measured only amounts of pollen or pollinia removed. However, this measure does not adequately encompass male fitness; alterations to pollen movement patterns and eventual deposition location may be more important to male function than changes in removal rates. If, as suggested by some studies (Altshuler 1999, Lach 2007), ants increase pollinator relocation rates, they may decrease pollen discounting, increasing male fitness.

However, if this increased relocation reduces overall pollen export rates, ant competition with

22 pollinators may depress male function. Studies using pollen-tracking methods such as fluorescent dyes will allow us to see how ants change pollen movement patterns, and methods like paternity analysis will reveal how competition between ants and pollinators alters male fitness.

iii) Coevolution between ants and plants: do myrmecophytes exhibit stronger, more

specific defenses against ants?

The floral traits of myrmecophytes are most likely to coevolve with ants, because all plant reproduction occurs in the presence of ants. Their floral defenses would thus be expected to be both stronger, and more specific, to their resident ants than those other plants visited by an array of more opportunistic ant species. Some evidence supports this; Willmer et al. (2009) found that resident ants of acacias were repelled by flowers, with more aggressive ant species showing stronger repellence than less aggressive ants. Additionally, ants not normally associated with acacias did not respond to acacia floral volatiles. More studies are needed in other systems, and should take advantage of phylogenetic approaches, examining floral traits in closely related myrmecophytic and non-myrmecophytic plant species to determine whether ants commonly select for floral defenses against ants in myrmecophytes.

Acknowledgements

For comments and discussion, we thank the Frederickson lab, particularly Eric Youngerman,

Lina Arcila Hernandez, and Kyle Turner, as well as Alison Parker and Jordan Pleet. MEF acknowledges financial support from an NSERC Discovery Grant, a Connaught New Researcher

Award, an Ontario Ministry of Economic Development and Innovation Early Researcher Award, and the University of Toronto; ARC was supported by an Ontario Graduate Scholarship and

Sigma Xi.

References

Adler, L. S. 2001. The ecological significance of toxic nectar. Oikos 91:409–420.

Agarwal, V. M., and N. Rastogi. 2007. Role of floral repellents in the regulation of flower visits

of extrafloral nectary-visiting ants in an Indian crop plant. Ecological Entomology 33:59–

65.

Altshuler, D. L. 1999. Novel interactions of non-pollinating ants with pollinators and fruit

consumers in a tropical forest. Oecologia 119:600–606.

Ashman, T.L 2000. Pollinator selectivity and its implications for the evolution of dioecy and

sexual dimorphism. Ecology 81:2577-2591.

Ashman, T.L., and E. A. King. 2005. Are flower-visiting ants mutualists or antagonists? A study

in a gynodioecious wild strawberry. American Journal of Botany 92:891–5.

Ballantyne, G. and P. Willmer. 2012. Floral visitors and ant scent marks: noticed but not used?

Ecological Entomology 37:402-409.

Ballantyne, G., and P. Willmer. 2012. Nectar theft and floral ant-repellence: a link between

nectar volume and ant-repellent traits? PloS One 7:e43869.

Baker. H. G., and I. Baker. 1978. Ants and flowers. Biotropica 10:80.

Barrett, S. C. H. 2002. Sexual interference of the floral kind. Heredity 88:154–159.

Baroni Urbani, C., and M. L. de Andrade. 1997. Pollen eating, storing, and spitting by ants.

Naturwissenschaften 84:256–258.

Beattie, A. J. ., C. Turnbull, R. B. Knox, and E. G. Williams. 1984. Ant inhibition of pollen

function: a possible reason why ant pollination is rare. American Journal of Botany

71:421–426.

Beattie, A. J., C. Turnbull, T. Hough, S. Jobson, and R. B. Knox. 1985. The vulnerability of

pollen and fungal spores to ant secretions: evidence and some evolutionary implication.

American Journal of Botany 72:606–614.

Blancafort, X., and C. Gómez. 2005. Consequences of the Argentine ant, Linepithema humile

(Mayr), invasion on pollination of Euphorbia characias (L.) (Euphorbiaceae). Acta

Oecologica 28:49–55.

Bleil, R., N. Blüthgen, and R. R. Junker. 2011. Ant-plant mutualism in Hawai‘i? Invasive ants

reduce flower parasitism but also exploit floral nectar of the endemic shrub Vaccinium

reticulatum (Ericaceae). Pacific Science 65:291–300.

Blüthgen, N., G. Gottsberger, and K. Fiedler. 2004. Sugar and amino acid composition of ant-

attended nectar and honeydew sources from an Australian rainforest. Austral Ecology

29:418–429.

Burkle, L. A., R. E. Irwin, and D. A. Newman. 2007. Predicting the effects of nectar robbing on

plant reproduction: implications of pollen limitation and plant mating system. American

Journal of Botany 94:1935–1943.

Byk, J., and K. Del-Claro. 2010. Nectar- and pollen-gathering Cephalotes ants provide no

protection against herbivory: a new manipulative experiment to test ant protective

capabilities. Acta Ethologica 13:33–38.

Caballero, P., C. G. Ossa, W. L. Gonzáles, C. González-Browne, G. Astorga, M. M. Murúa, and

R. Medel. 2013. Testing non-additive effects of nectar-robbing ants and hummingbird

pollination on the reproductive success of a parasitic plant. Plant Ecology 214:633–640.

Carvalheiro, L. G., E. R. M. Barbosa, and J. Memmott. 2008. Pollinator networks, alien species

and the conservation of rare plants: Trinia glauca as a case study. Journal of Applied

Ecology 45:1419–1427.

Chamberlain, S. A., and J. N. Holland. 2008. Density-mediated , context-dependent consumer-

resource interactions between ants and extrafloral nectar plants. Ecology 89:1364–1374.

Creighton, W. S. 1967. Studies on free colonies of Cryptocerus texanus Satschi (Hymenoptera:

Formicidae). Psyche 74:34–42.

Cursach, J., and J. Rita. 2011. Implications of the reproductive biology of the narrow endemic

Naufraga balearica (Apiaceae) for its conservation status. Plant Systematics and Evolution

298:581–596.

Cushman, J.H. and Addicott, J.F. In: Huxley, C.R., and D.F. Cutler (eds). 1st edition. Ant-plant

interactions. Oxford University Press, Oxford. de Jong, T. J., N. M. Waser, and P. G. Klinkhamer. 1993. Geitonogamy: the neglected side of

selfing. Trends in Ecology and Evolution 8:321–325. de Vega, C., M. Arista, P. L. Oritz, C. M. Herrera, and S. Talavera. 2009. The ant-pollination

system of Cytinus hypocistis (Cytinaceae), a Mediterranean root holoparasite. Annals of

Botany 4:517–8.

Dukas, R. 2001. Effects of perceived danger on flower choice by bees. Ecology Letters 4:327–

333.

Dutton, E. M., and M. E. Frederickson. 2012. Why ant pollination is rare: new evidence and

implications of the antibiotic hypothesis. Arthropod-Plant Interactions 6:561–569.

Edwards, D. P., and D. W. Yu. 2008. Tolerating castration by hiding flowers in plain sight.

Behavioral Ecology and Sociobiology 63:95–102.

Frederickson, M. E. 2009. Conflict over reproduction in an ant-plant symbiosis: why Allomerus

octoarticulatus ants sterilize Cordia nodosa trees. The American Naturalist 173:675–81.

Fritz, R. S., and D. H. Morse. 1981. Nectar parasitism of Asclepias syriaca by ants: effect on

nectar levels, pollinia insertion, pollinaria removal and pod production. Oecologia 50:316–

319.

Galen, C. 1983. The effects of nectar thieving ants on seedset in floral scent morphs of

Polemonium viscosum. Oikos 41:245–249.

Galen, C. 2005. Catching ants with honey: an experimental test of distraction and satiation as

alternative modes of escape from flower-damaging ants. Oecologia 144:80–87.

Galen, C., and B. Butchart. 2003. Ants in your plants: effects of nectar-thieves on pollen fertility

and seed-siring capacity in the alpine wildflower, Polemonium viscosum. Oikos 101:521–

528.

Galen, C., and J. Cuba. 2001. Down the tube: pollinators, predators, and the evolution of flower

shape in the alpine skypilot, Polemonium viscosum. Evolution 55:1963-1971.

Galen, C., and J. C. Geib. 2007. Density-dependent effects of ants on selection for bumble bee

pollination in Polemonium viscosum. Ecology 88:1202–1209.

Galen, C., R. Kaczorowski, S. L. Todd, J. Geib, and R. A. Raguso. 2011. Dosage-dependent

impacts of a floral volatile compound on pollinators, larcenists, and the potential for floral

evolution in the alpine skypilot Polemonium viscosum. The American Naturalist 177:258–

272.

García, M. B., R. J. Antor, and X. Espadaler. 1995. Ant pollination of the palaeoendemic

dioecious Borderea pyrenaica (Dioscoreaceae). Plant Systematics and Evolution 198:17–

27.

Gaume, L., M. Zacharias, and R. M. Borges. 2005. Ant – plant conflicts and a novel case of

castration parasitism in a myrmecophyte. Evolutionary Ecology Research 7:435–452.

Graystock, P., and W. O. H. Hughes. 2011. Disease resistance in a weaver ant, Polyrhachis

dives, and the role of antibiotic-producing glands. Behavioral Ecology and Sociobiology

65:2319–2327.

Gόmez, J. M. 2000. Effectiveness of ants as pollinators of Lobularia maritima: effects on main

sequential fitness components of the host plant. Oecologia 122:90–97.

Gόmez, J. M., and R. Zamora. 1992. Pollination by ants: consequences of the quantitative effects

on a mutualistic system. Oecologia 91:410–418.

Gόmez, J. M., R. Zamora, J. A. Hόdar, and D. García. 1996. Experimental study of pollination

by ants in Mediterranean high mountain and arid habitats. Oecologia 105:236–242.

Gonçalves-Souza, T., P. M. Omena, J. C. Souza, and G. Q. Romero. 2008. Trait-mediated effects

on flowers: artificial spiders deceive pollinators and decrease plant fitness. Ecology

89:2407–2413.

Gonzálvez, F. G., L. Santamaría, R. T. Corlett, and M. A. Rodríguez-Gironés. 2012. Flowers

attract weaver ants that deter less effective pollinators. Journal of Ecology 101:78–85.

Guerrant, E. O., and P. L. Fiedler. 1981. Flower defenses against nectar-pilferage by ants.

Biotropica 13:25–33.

Harley, R. 1991. The greasy pole syndrome. In: Huxley, C.R., and D.F. Cutler (eds). 1st edition.

Ant-plant interactions. Oxford University Press, Oxford.

Hansen, D. M., and C. B. Müller. 2009. Invasive ants disrupt gecko pollination and seed

dispersal of the endangered plant Roussea simplex in Mauritius. Biotropica 41:202–208.

Hickman, J. C. 1974. Pollination by ants: a low-energy system. Science 184:1290–1292.

Holland, J. N., S. a. Chamberlain, and T. E. X. Miller. 2011. Consequences of ants and

extrafloral nectar for a pollinating seed-consuming mutualism: ant satiation, floral

distraction or plant defense? Oikos 120:381–388.

Hölldobler, B. and Wilson, E. O. 1990. The Ants, 1st Edition. Harvard University Press,

Cambridge.

Holway, D.A., L. Lach, A.V. Suarez, N.D. Tsutsui, and T.J. Case. 2002. The causes and

consequences of ant invasions. Annual Review of Ecology and Systematics 33:181-233.

Horskins, K. and Turner, V.B. 1999. Resource use and foraging patterns of honeybees, Apis

mellifera, and native insects on flowers of Eucalyptus costata. Australian Journal of

Ecology 24:221-227.

Horvitz, C. C., and D. W. Schemske. 1984. Effects of ants and an ant-tended herbivore on seed

production of a neotropical herb. Ecology 65:1369–1378.

Inouye, D.W. 1980. The terminology of floral larceny. Ecology 61:1251-1253.

Inouye, D.W., and O. R. J. Taylor. 1979. A temperate region plant-ant-seed predator system :

consequences of extra floral nectar secretion by Helianthella quinquenervis. Ecology 60:1–

Irwin, R.E. 2003. Impact of nectar robbing on estimates of pollen flow: conceptual predictions

and empirical outcomes. Ecology 84:485-495.

Irwin, R.E. and Maloof, J.E. 2002. Variation in nectar robbing over time, space and species.

Oecologia 133:525-533.

Iwanami, Y., and T. Iwadare. 1978. Inhibiting effects of myrmicacin on pollen growth and pollen

tube mitosis. Botanical Gazette 139:42–45.

Izzo, T.J. and Vasconcelos H.L. 2002. Cheating the cheater: domatia loss minimizes the effects

of ant castration in an Amazonian ant plant. Oecologia 133:200-205.

Janzen, D. H. 1977. Why don’t ants visit flowers? Biotropica 9:252.

Johnson, S.D., Hargreaves, A.L., and Brown, M. 2006. Dark, bitter-tasting nectar functions as a

filter of flower visitors in a bird-pollinated plant. Ecology 87:2709-2716.

Junker, R., A. Y. C. Chung, and N. Blüthgen. 2007. Interaction between flowers, ants and

pollinators: additional evidence for floral repellence against ants. Ecological Research

22:665–670.

Junker, R. R., R. Bleil, C. C. Daehler, and N. Blüthgen. 2010a. Intra-floral resource partitioning

between endemic and invasive flower visitors: consequences for pollinator effectiveness.

Ecological Entomology 35:760–767.

Junker, R. R., and N. Blüthgen. 2010b. Floral scents repel facultative flower visitors, but attract

obligate ones. Annals of botany 105:777–82.

Junker, R. R., C. C. Daehler, S. Doetteri, A. Keller, and N. Blüthgen. 2011a. Hawaiian ant –

flower networks : nectar-thieving ants prefer undefended native over introduced plants with

floral defenses. Ecological Monographs 81:295–311.

Junker, R. R., J. Gershenzon, and S. B. Unsicker. 2011b. Floral odor bouquet loses its ant

repellent properties after inhibition of terpene biosynthesis. Journal of Chemical Ecology

37:1323–31.

Kawakita, A., and M. Kato. 2002. Floral biology and unique pollination system of root

holoparasites, Balanophora kuroiwai and B. tobiracola (Balanophraceae). American

Journal of Botany 89:1164–1170.

Koopowitz, H., and T.A. Marchant. 1998. Postpollination nectar reabsorption in the African

epiphyte Aerangis verdickii (Orchidaceae). American Journal of Botany 85:508-512.

Kirk, W. D. J. 1984. Ecological studies on Thrips imaginis Bagnall (Thysanoptera) in flowers of

Echium plantagineum L. in Australia. Australian Journal of Ecology 9:9–18.

Kerner, A. 1878. Flowers and their unbidden guests. C. Kegan Paul, London.

Kessler, D., and I. T. Baldwin. 2007. Making sense of nectar scents: the effects of nectar

secondary metabolites on floral visitors of Nicotiana attenuata. The Plant Journal 49:840–

54.

Koptur, S., and N. Truong. 1998. Facultative ant-plant interactions: nectar sugar preferences of

introduced pest species in South Florida. Biotropica 30:179-189.

Lach, L. 2005. Interference and exploitation competition of three nectar-thieving invasive ant

species. Insectes Sociaux 52:257–262.

Lach, L. 2007. A mutualism with a native membracid facilitates pollinator displacement by

Argentine ants. Ecology 88:1994–2004.

Lach, L. 2008a. Argentine ants displace floral arthropods in a biodiversity hotspot. Diversity and

Distributions 14:281–290.

Lach, L. 2008b. Floral visitation patterns of two invasive ant species and their effects on other

hymenopteran visitors. Ecological Entomology 33:155–160.

Lach, L. 2013. A comparison of floral resource exploitation by native and invasive Argentine

ants. Arthropod-Plant Interactions 7:177-190.

Luo, C. W., K. Li, X. M. Chen, Z. Y. Huang, and K. U. N. Li. 2012. Ants contribute significantly

to the pollination of a biodiesel plant, Jatropha curcas. Environmental Entomology

41:1163–1168.

Malé, P.J. G., C. Leroy, A. Dejean, A. Quilichini, and J. Orivel. 2011. An ant symbiont directly

and indirectly limits its host plant’s reproductive success. Evolutionary Ecology 26:55–63.

Maloof, J. E. 2001. The effects of a bumble bee nectar robber on plant reproductive success and

pollinator behavior. American Journal of Botany 88:1960–1965.

Maloof, J. E., and D. W. Inouye. 2000. Are nectar robbers cheaters or mutualists ? Ecology

81:2651–2661.

Moreau, C.S., C.D. Bell, R. Vila, S.B. Archibald, and N.E. Pierce. 2006. Phylogeny of the ants:

diversification in the age of the angiosperms. Science 312:101-104.

Nahas, L., M. O. Gonzaga, and K. Del-Claro. 2012. Emergent impacts of ant and spider

interactions: herbivory reduction in a tropical savanna tree. Biotropica 44:498–505.

Ness, J. H. 2006. A mutualism’s indirect costs : the most aggressive plant bodyguards also deter

pollinators. Oikos 113:506–514.

Newman, D. A., and J. D. Thomson. 2005. Interactions among nectar robbing, floral herbivory,

and ant protection in Linaria vulgaris. Oikos 110:497–506.

Nicklen, E. F., and D. Wagner. 2006. Conflict resolution in an ant-plant interaction: Acacia

constricta traits reduce ant costs to reproduction. Oecologia 148:81–7.

Norment, C. J. 1988. The effect of nectar-thieving ants on the reproductive success of Frasera

speciosa (Gentianaceae). American Midland Naturalist 120:331–336.

Orivel, J., L. Lambs, P.-J. G. Malé, C. Leroy, J. Grangier, T. Otto, A. Quilichini, et al. 2011.

Dynamics of the association between a long-lived understory myrmecophyte and its

specific associated ants. Oecologia 165:369–76.

Palmer, T. M., D. F. Doak, M. L. Stanton, J. L. Bronstein, E. T. Kiers, T. P. Young, J. R.

Goheen, et al. 2010. Synergy of multiple partners, including freeloaders, increases host

fitness in a multispecies mutualism. Proceedings of the National Academy of Sciences

107:17234–9.

Peakall, R., A. J. Beattie, and S. H. James. 1987. Pseudocopulation of an orchid by male ants : a

test of two hypotheses accounting for the rarity of ant pollination. Oecologia 73:522–524.

Peakall, R., and A. J. Beattie. 1989. Pollination of the orchid Microtis parviflora R . Br . by

flightless worker ants. Functional Ecology 3:515–522.

Peakall, R., C. J. Angus, A. J. Beattie. 1990. The significance of ant and plant traits for ant

pollination in Leporella fimbriata pollination. Oecologia 84:457–460.

Puterbaugh, M. N. 1998. The roles of ants as flower visitors: experimental analysis in three

alpine plant species. Oikos 83:36–46.

Prŷs-Jones, O.E., and Willmer, P.G. 1992. The biology of alkaline nectar in purple toothwort

(Lathraea clandenstina): ground level defenses. Biological Journal of the Linnean Society

45:373-388.

Raine, N.E., Willmer, P., and Stone, G.N. 2002. Spatial structuring and floral avoidance

behaviour present in ant-pollinator conflict in a Mexican ant-acacia. Ecology 83:3086-

3096.

Ramsey, M. 1995. Ant pollination of the perennial herb Blandfordia grandiflora (Liliaceae).

Oikos 74:265–272.

Renault, C.K., L.M. Buffa, and Delfino, M.A. 2005. An aphid-ant interaction: effects on

different trophic levels. Ecological Research 20:71-74.

Rico-Gray, V., and L. B. Thien. 1989. Effect of different ant species on reproductive fitness of

Schomburgkia tibicinis (Orchidaceae). Oecologia 81:487–489.

Romero, G. Q., P. A. P. Antiqueira, and J. Koricheva. 2011. A meta-analysis of predation risk

effects on pollinator behaviour. PLoS One 6:e20689.

Rosumek, F.B., F.A.O. Silveira, F. de S Neves, N.P. de U Barbosa, L. Diniz, Y. Oki, F. Pezzini,

et al. 2009. Ants on plants: a meta-analysis of the role of ants as plant biotic defenses.

Oecologia 160:537–549.

Roubik, D. W. 1982. The ecological impact of nectar-robbing bees and pollinating

hummingbirds on a tropical shrub. Ecology 63:354–360.

Rudgers, J. A. 2004. Enemies of herbivores can shape plant traits: selection in a facultative ant-

plant mutualism. Ecology 85:192–205.

Schatz, B., and W.T. Wcislo. 1999. Ambush predation by the ponerine ant Ectatomma ruidum

Roger (Formicidae) on a sweat bee Lasioglossum umbripenne (Halictidae), in Panama.

Journal of Insect Behavior 12:641-663.

Schaffer, W. M., D. W. Zeh, S. L. Buchmann, S. Kleinhaus, M. V. Schaffer, and J. Antrim.

1983. Competition for nectar between introduced honey bees and native North American

bees and ants. Ecology 64:564–577.

Schemske, D. W. 1980. The evolutionary significance of extrafloral nectar production by Costus

woodsonii (Zingiberaceae): an experimental analysis of ant protection. Journal of Ecology

68:959–967.

Schiestl, F. P., and F. Glaser. 2011. Specific ant-pollination in an alpine orchid and the role of

floral scent in attracting pollinating ants. Alpine Botany 122:1–9.

Schubart, H.O.R., A.B. Anderson. 1978. Why don’t ants visit flowers? A response to D. H.

Janzen. Biotropica 10:310-311.

Schürch, S., M. Pfunder, and B. A. Roy. 2000. Effects of ants on the reproductive success of

Euphorbia cyparissias and associated pathogenic rust fungi. Oikos 1:6–12.

Stanton, M.L., Palmer, T.M., Young, T.P., Evans, A., and Turner, M.L. 1999. Sterilization and

canopy modification of swollen thorn acacia tree by a plant-ant. Nature 401: 578-581.

Stephenson, A.G. 1981. Toxic nectar deters nectar thieves of Catalpa speciosa. American

Midland Naturalist 105:381-383.

Stout, J. C., and D. Goulson. 2001. The use of conspecific and interspecific scent marks by

foraging bumblebees and honeybees. Animal Behaviour 62:183–189.

Sugiura, S., T. Abe, and S. Makino. 2006. Loss of extrafloral nectary on an oceanic island plant

and its consequences for herbivory. American Journal of Botany 93:491–495.

Svensson, L. 1985. An estimate of pollen carryover by ants in a natural population of

Scleranthus perennis L. (Caryophyllaceae). Oecologia 66:373–377.

Tsuji, K., A. Hasyim, and K. Nakamura. 2004. Asian weaver ants, Oecophylla smaragdina, and

their repelling of pollinators. Ecological Research 19:669–673.

Turner, R. C., J. J. Midgley, P. Barnard, R. E. Simmons, and S. D. Johnson. 2012. Experimental

evidence for bird pollination and corolla damage by ants in the short-tubed flowers of

Erica halicacaba (Ericaceae). South African Journal of Botany 79:25–31.

Visser, D., M. G. Wright, and J. H. Giliomee. 1996. The effect of the Argentine ant ,

Linepithema humile (Mayr) (Hymenoptera: Formicidae), on flower-visiting insects of

Protea nitida Mill. (Proteaceae). African Entomology 4:285–287.

Wagner, D. 2000. Pollen viability reduction as a potential cost of ant association for Acacia

constricta (Fabaceae). American Journal of Botany 87:711–715.

Wagner, D., and A. Kay. 2002. Do extrafloral nectaries distract ants from visiting flowers? An

experimental test of an overlooked hypothesis. Evolutionary Ecology Research 4:293–305.

Wang, C., Y. Luo, Y. Tai, D. An, and Y. Kou. 2008. Ants pollinate Neottia listeroides

(Orchidaceae) in Sichuan, China 46:836–846.

Wyatt, R. 1980. The impact of nectar-robbing ants on the pollination system of Asclepias

curassavica. Bulletin of the Torrey Botanical Club 107:24–28.

Wheeler, W.M., and I.W. Bailey. 1920. The feeding habits of Pseudomymine and other ants.

Transactions of the American Philosophical Society 22:235-279.

Willmer, P. G., and S. A. Corbet. 1981. Temporal and microclimatic partitioning of the floral

resources of Justicia aurea amongst a concourse of pollen vectors and nectar robbers.

Oecologia 51:67–78.

Willmer, P. G., and G. N. Stone. 1997. How aggressive ant-guards assist seed-set in Acacia

flowers. Nature 388:165–167.

Willmer, P. G., C. V. Nuttman, N. E. Raine, G. N. Stone, J. G. Pattrick, K. Henson, P. Stillman,

et al. 2009. Floral volatiles controlling ant behaviour. Functional Ecology 23:888–900.

Young, T. P., C. H. Stubblefield, and L. a. Isbell. 1997. Ants on swollen-thorn acacias: species

coexistence in a simple system. Oecologia 109:98–107.

Yu, D. W., and N. E. Pierce. 1998. A castration parasite of an ant-plant mutualism. Proceedings

of the Royal Society B: Biological Sciences 265:375–382.

Tables

Table 1.1. Studies of ant-flower or ant-pollinator interactions that included a measure of plant fitness. Each study system is listed only once, except when results varied between years or when different studies examined male and female fitness. Significant effects of ants on plant fitness are either ‘positive’ or ‘negative’; non-significant effects are ‘neutral’. When ants significantly affected only one fitness component but several were measured (e.g., fruit set and seed production), we still consider the result significant.

Plant species Ant species Fitness measure Direction of effect Proposed mechanism Reference

Acacia Crematogaster Proportion Negative Destruction of floral buds Young et al. 1997 drepanolobium nigriceps fruiting *

Asclepias exaltata Several Pollinia inserted/ Negative for pollinia Exploitative competition Wyatt 1980 T

removed insertion; neutral for

pollinia removal

Plant species Ant species Fitness measure Direction of effect Proposed mechanism Reference

Asclepias syriaca Several Fruit set/ pollinia Negative for fruit Exploitative competition Fritz and Morse

removed set; neutral for 1981 T

pollinia removal

Cordia nodosa Allomerus cf. Fruit set Negative Destruction of floral buds Yu and Pierce

demerarae 1998*

Costus woodsonii Wasmannia Seed set Positive Unknown, ant pollination Schemske 1980

auropunctata suggested (but unlikely?)

Eritrichum aretioides F. neorufibarbis Seed set Negative Ants damaged flowers Puterbaugh 1998

gelida

Plant species Ant species Fitness measure Direction of effect Proposed mechanism Reference

Euphorbia characias Several Seed set Neutral Plants in Argentine ant- Blancafort and

invaded area set less seed Gόmez 2005

than those in un-invaded

sites, but an ant-exclusion

experiment did not produce

significant results

Ferocactus wislizeni Solenopsis xyloni Seed set Unknown/ negative Interference and exploitative Ness 2006

and others competition; S. xyloni

reduced seed set relative to

other ants, but the study did

not include a no-ant control

Plant species Ant species Fitness measure Direction of effect Proposed mechanism Reference

Frasera speciosa F. neorufibarbis Seed set Neutral Nocturnal pollination, Norment 1988

and Formica sp. occurring when ants were

less active, may have

compensated for reduced

diurnal pollinator visitation

Gypsophila struthium Several Seeds per Neutral No effect, despite large Gόmez et al.

inflorescence numbers of ants on flowers 1996

Hirtella physophora Allomerus Fruit set Negative Destruction of flowers Orivel et al.

decemarticulatus 2011*

Humboldtia brunonis Crematogaster Fruit set Negative Destruction of flowers Gaume et al.

dohrni 2005 *

Plant species Ant species Fitness measure Direction of effect Proposed mechanism Reference

Lepidium subulatum Several Percentage of Positive Unknown Gόmez et al.

ovules setting 1996

seed

Leucospermum Several Seed set Neutral Site not pollen limited Lach 2008 conocarpodendron

Linaria vulgaris Several “Female fitness Positive Reduction in florivory and Newman and

index” seed predation; nectar Thomson 2005

(combination of robbing holes increased ant

number of seeds, visitation

fruits, flowers,

and seed weight)

Plant species Ant species Fitness measure Direction of effect Proposed mechanism Reference

Melastoma Oecophylla Seed set Positive Ants deterred less efficient Gonzálvez et al. malabathricum smaragdina pollinators, increasing 2012

visitation by the more

efficient pollinator

Metrosideros Several Fruit set Neutral Ants reduced nectar Junker et al. polymorpha collection rates by foraging 2010a

bees but had no effect on

pollen collection rates or

fruit set

Orexis alpine F. neorufibarbis Seed set Neutral Autogamy or not pollen Puterbaugh 1988

gelida limited

Plant species Ant species Fitness measure Direction of effect Proposed mechanism Reference

Ouratea spectabilis Cephalotes Seeds per fruit Neutral/ positive C. pusillus had no effect on Byk and Del

pusillus and others seed number, but other ants Claro 2010

collectively increased seed

set

Pachycereus schottii Several Seed set Neutral/ positive In one year, ants showed a Chamberlain and

positive effect (seed Holland 2008

number, 2008), but no effect

with increased sample sizes

Polemonium F. neorufibarbis Seed set Negative Ants damaged styles Galen 1983 viscosum gelida

Plant species Ant species Fitness measure Direction of effect Proposed mechanism Reference

Polemonium F. neorufibarbis Pollen removal Neutral Ants did not affect pollen Galen and Geib viscosum gelida removal, even at high ant 2007 T

densities

Protea nitida Linepithema Seed set Neutral/ positive L. humile had no effect on Lach 2007

humile and others P. nitida seed set, but native

ants increased it in one year;

ants may cause pollinators

to relocate

Psychotria Ectatomma spp. Fruit set Positive Suggested increased Altshuler 1999 limonensis pollinator relocation

Plant species Ant species Fitness measure Direction of effect Proposed mechanism Reference

Roussea simplex Technomyrmex Seed set Negative Interference and exploitative Hansen and

albipes competition Müller 2009

Tristerix aphyllus Several Fruit set Neutral Nectar robbing ants Caballero et al.

minimally reduced nectar 2013

levels and had no effect on

hummingbird visitation

Vaccinium Several Fruit set Neutral Ants reduced florivory, but Bleil et al. 2011 reticulatum did not affect fruit set

T Study measured male fitness

* Study of sterilizing ants showed short-term decrease in female fitness; long-term effects are unclear (see text)

Table 1.2. Studies demonstrating either i) ant pollination through an exclusion experiment or ii) that ants transfer pollen (or pollen analogue) between flowers, strongly suggesting ant pollination.

Plant species Ant species Ant pollination? Evidence Reference

Alyssum purpureum Several Yes Exclusion experiment Gόmez et al. 1996

Arenaria tetraquetra Several Yes Exclusion experiment Gόmez et al. 1996

Balanophora kuroiwai Leptothorax sp. Suggested Exclusion experiment and Kawakita and Kato 2002

observed pollen transfer

and pollen tube formation,

but did not see if plants set

seed

Blandfordia grandiflora Iridomyrmex sp. Yes Exclusion experiment Ramsey 1995

Borderea pyrenaica Several Yes Exclusion experiment García et al. 1995

Plant species Ant species Ant pollination? Evidence Reference

Cytinus hypocistis Several Yes Exclusion experiment de Vega et al. 2009

Epipactis thunbergii Camponotus Suggested Observed transfer of Sugiura et al. 2006

japonicas pollinia from ants to

orchids

Euphorbia cyparissias Several Yes Exclusion experiment Schürch et al. 2000

Fragaria virginiana Several Very likely Ant-only treatments set Ashman and King 2005

comparable seed to flying

pollinator-only treatment,

seed set increased with

increasing ant visitation

Frasera speciosa Several Yes Exclusion experiment Norment 1988

Frankenia thymifolia Several Yes Exclusion experiment Gόmez et al. 1996

Plant species Ant species Ant pollination? Evidence Reference

Hormathophylla spinosa Proformica Yes Exclusion experiment Gόmez and Zamora 1992

longiseta

Jatropha curcas Several, Yes Exclusion experiment Luo et al. 2012

Tapinoma

melanocephalum

most common

Leporella fimbriata Myrmecia urens Yes Observed pollinia transfer Peakall et al. 1987

and subsequently

confirmed as ant

pollination

Plant species Ant species Ant pollination? Evidence Reference

Lobularia maritima Several, Yes Exclusion experiment Gόmez 2000

Camponotus

micans most

common

Microtis parviflora Iridomyrmex Yes Observation, used Peakall and Beattie 1989

gracilis laboratory colonies to

replicate observations

Naufraga balearica Several, Suggested Observation, over three Cursach and Rita 2011

especially years of study, only ants

Plagiolepis were observed visiting;

pygmaea and ants carried pollen

Lasius grandis

Plant species Ant species Ant pollination? Evidence Reference

Neottia listeroides Leptothoras sp., Suggested Observed transfer of Wang et al. 2008

Paratrechina sp. pollinia from ants to

orchids

Paronychia pulvinata Formica Yes Exclusion experiment Puterbaugh 1998

neorufibarbis

gelida

Polygonum cascadense Formica argentea Very likely Grew plants in greenhouse Hickman 1974

open to flying visitors with

minimal seed set, only

plants with access to ants

set significant seed

Plant species Ant species Ant pollination? Evidence Reference

Retama sphaerocarpa Several Yes Exclusion experiment Gόmez et al. 1996

Scleranthus perennis Several, Suggested Ants transferred pollen Svensson 1985

especially analogue to stigmas and

Formica carried pollen on bodies

rufibarbis

Sedum anglicum Proformica Yes Exclusion experiment Gόmez et al. 1996

longiseta

Trinia glauca Several Yes Exclusion experiment Carvalheiro et al. 2008

Chapter 2 Ants and ant scent reduce bumblebee pollination of artificial flowers

Cembrowski, A.R. Tan, M.G., Thomson, J.D., and Frederickson, M.E.

In press, The American Naturalist All authors designed the experiment, ARC and MGT performed the experiment, ARC and MEF performed statistical analyse

s and ARC wrote the manuscript with input from all authors

Abstract

Ants on flowers can disrupt pollination by consuming rewards or harassing pollinators, but it is difficult to disentangle the effects of these exploitative and interference forms of competition on pollinator behavior. Using highly rewarding and quickly replenishing artificial flowers that simulate male or female function, we allowed bumblebees (Bombus impatiens) to forage (1) on flowers with or without ants (Myrmica rubra), and (2) on flowers with or without ant scent cues.

Bumblebees transferred significantly more pollen analogue both to and from ant free flowers, demonstrating that interference competition with ants is sufficient to modify pollinator foraging behavior. Bees also removed significantly less pollen analogue from ant scented flowers than from controls, making this the first study to show that bees can use ant scent to avoid harassment at flowers. Ant effects on pollinator behavior, possibly in addition to their effects on pollen viability, may contribute to the evolution of floral traits minimizing ant visitation.

54 55

Introduction

Trait-mediated indirect interactions arise when a focal species causes phenotypic changes, including behavioral modifications, in a second species, and these effects cascade to still other species (Werner and Peacor 2003). Although trait-mediated indirect interactions are often studied in food webs (e.g., Werner and Peacor 2003; Preisser et al. 2005), they are not limited to trophic interactions. For example, the threat of predation can change the behavior of a mutualist, thus affecting its partners (Suttle 2003). In animal-pollinated plants, predators can disrupt pollination directly through density-mediated indirect interactions (Dukas 2005) or via trait-mediated indirect interactions by changing pollinator behavior. When predators are present, pollinators may switch to visiting less rewarding flowers or they may avoid flowers altogether, potentially reducing the amount of pollen donated or received by a flower and thus decreasing plant fitness

(Gonçalves-Souza 2008, Ings and Chittka 2009). Similarly, competition for floral rewards with other flower visitors may cause behavioral changes in pollinators (Maloof and Inouye 2000,

Ohashi et al. 2008). Previous research has shown that bees may spend less time visiting flowers depleted of rewards (Thomson 1986), and may avoid flowers bearing scent marks that indicate recent visits by other individuals of the same or different bee species, as these flowers are likely to be depleted (Stout and Goulson 2002).

Whereas competition between pollinators may not be particularly costly for the plant because both competitors are likely to provide pollination services, competition between pollinators and organisms that do not usually provide a pollination service, such as ants, has a greater potential to reduce plant fitness. Ants are common floral visitors; they are attracted to flowers for nectar (Lach 2007) and, in some cases, pollen (Byk and Del Claro 2010), and will often defend these resources against other flower visitors (Altshuler 1999). However, ants rarely

56 contribute much useful pollination; antibiotic secretions present on their cuticles kill pollen grains (Beattie et al. 1985, Dutton and Frederickson 2012), and when foraging in flowers, ants sometimes impede female function by damaging stigmas (Galen and Cuba 2001). We will henceforth consider ants to be non-pollinators, and will contrast them to animals like bees that are pollinators. Ants are capable of competing with pollinators in two ways: (1) by consuming floral rewards and reducing their availability to pollinators (i.e., exploitative competition) and (2) by directly antagonizing and excluding pollinators from flowers (i.e., interference competition).

Many plants benefit from the presence of ants; numerous plant species recruit ants using extrafloral nectar, food bodies, or domatia because ants provide protection against folivores (Heil and McKey 2003). Similarly, some plants have extrafloral nectaries on or near reproductive tissues that recruit ants that deter florivores (e.g., Inouye and Taylor, 1979). However, plants may incur reproductive costs as a result of their association with ants. Some ant species that nest in myrmecophytes sterilize flowers to shunt plant resources from reproduction to vegetative growth, allowing for greater ant colony growth (Frederickson 2009). Also, ants visiting extrafloral nectaries may be attracted to floral nectaries, where they can harass pollinators (Ness

2006). There is growing evidence that many angiosperms have evolved floral traits that prevent ants from accessing flowers during anthesis (Willmer and Stone 1997, Ballantyne and Willmer

2012). The effects of flower-visiting ants can be substantial. Ants can decrease the frequency, duration, or species diversity of pollinator visits to flowers (Lach 2008, Hansen and Müller 2009,

Junker et al. 2010, Gonzálvez et al. 2012), all of which potentially impact pollen donation and receipt. In some studies, changes to pollinator behavior induced by ants have also been linked to seed set.

It is difficult, however, to disentangle the intertwined effects of exploitative and interference competition in nature. If ants disrupt pollination, plants may often evolve traits to

57 deter ants from flowers (Willmer et al. 2009). From a practical standpoint, this would then limit our ability to study the effects of ants on pollinator visitation, because ants would not visit flowers. Here, we used artificial flowers to explore ant-bee interactions in the absence of floral defenses against ants. We used highly rewarding, quickly replenishing, artificial flowers to examine how direct harassment (i.e., interference competition expressed through behavior) changes bee foraging behavior, and how this affects donation and receipt of a pollen analogue

(powdered food dye), while minimizing the effects of exploitative competition. We predicted that flowers visited by ants would both donate and receive less pollen analogue than flowers without ants because bees would avoid flowers with ants, or leave them sooner. Because scent plays a large role in bumblebee communication and flower choice (Stout and Goulson 2002), we also tested whether the presence or absence of ant scent on artificial flowers would affect pollen analogue donation because of changes in the foraging behavior of bumblebees.

Methods

i) Subjects

Myrmica rubra is an invasive ant in eastern North America with a range and habitat preferences overlapping that of Bombus impatiens, a common bumblebee. Because M. rubra visits flowers

(A. Cembrowski, pers. obs.), the two species likely interact in nature. We collected 12 M. rubra colonies in the fall of 2011 and 2012 from Toronto, Ontario, Canada, and the surrounding area, and maintained them in environmental chambers on artificial diet (Dussutour and Simpson 2009) and a 14:10 L:D schedule (light 6:00-22:00). We used these colonies as sources of M. rubra workers for experiments.

Workers from commercially supplied colonies of Bombus impatiens (Biobest Canada,

Ltd.) foraged on artificial flowers in flight cages (either 2.4 x 2.4 x 2.1 m or 7.9 x 3.4 x 2.0m) at the University of Toronto. We tested a total of five bee colonies: four colonies were used in the ant presence trials, and two of these and one additional colony were used in the ant scent trials.

Flight cages had overhead fluorescent lights attached to timers. In contrast to most previous studies in which bees have been trained and tested individually, the entire worker force of a bee colony was free to forage at will in our experiments. Colonies were trained to forage on artificial flowers for at least four days before being used in trials. After being used for a trial, the colony was not used for at least two days to reduce dye carryover between trials. Between trials, bumblebees were fed pollen and given sugar water.

ii) Artificial flowers

The flowers (Fig 1, Thomson et al. 2012; see also Makino 2008) consisted of glass jars filled with 30% W/V sucrose solution. Sugar water travelled by capillary action up a sewing-thread wick to a hole in a blue-painted lid, accumulating in a knot that acts as a nectary. Flowers depleted by visitors were quickly replenished via capillary action and were non-rewarding only very briefly after visits, taking less than a minute to accumulate 0.5 µl; note that bumblebee visits to a single flower are often separated by several minutes or more (A. Cembrowski, unpublished data).

Unlike most previous artificial flower designs, these flowers allow for estimation of male and female fitness, because we can measure both the amount and type of dye received by female flowers (see (d) Dye quantification, below). In order to access nectar on a “male” flower, a bee must crawl through a circular “anther” of brush-like weather stripping material dusted in a

59 consistent manner with our pollen analogue, powdered food dye (FD&C 5 or 6), which is transferred to the bee in the process. On a “female” flower, a bee crawls through a sticky plastic reinforcement that functions as a “stigma,” receiving dye from the bee’s body. Male flowers dispensed dye particles over multiple bee visits and could still dispense dye at the end of trials.

They resemble many real flowers or flower heads in that pollen is dispensed gradually over time, but less dye is available for transfer with each subsequent bee visit (Harder and Thomson 1989).

iii) Flight cage trials

We conducted 6 ant scent trials between 22 March and 4 June 2012 and the remaining trials (16 ant presence trials and 9 ant scent trials) between 9 November 2012 and 6 January 2013. Ant presence trials lasted for eight hours and ant scent trials lasted for four hours. Artificial flowers were prepared and placed individually in small plastic containers treated with Fluon (Insect-a-

Slip, BioQuip Products, Inc.) to prevent ants from escaping. In ant presence trials we used 32 flowers (16 male and 16 female flowers), arranged in an eight-by-four array, with flowers spaced

45 cm apart between rows and 30 cm within rows. Ant scent trials used 20 flowers (10 male and

10 female flowers) in a four-by-five array, with flowers spaced 30 cm apart between and within rows. In all of the ant presence trials and 9 of the ant scent trials, we counted the number of bees actively foraging after one hour to get a measure of colony activity.

a) Ant presence trials

We examined the effects of ants on the amount of dye donated by male flowers and received by female flowers in sixteen trials, testing four bee colonies four times each. The containers of eight randomly chosen male flowers and eight randomly chosen female flowers received 15 M. rubra

60 workers each; the other 16 flowers remained free of ants. Isolated from their colonies, M. rubra workers have nowhere to deposit sugar water they collect, and therefore may become sated; nonetheless, a force of 15 ant workers was enough to maintain visitation to the nectaries where bees foraged, while still rarely having more than one or two ants visit the nectary at any time. As ants attacked bees, bees would occasionally carry or throw attacking ants off flowers, but despite this, some ants maintained their presence at the nectary throughout the trials. Because ants the size of M. rubra consume liquids slowly (~0.17 – 0.24 µl/min., Davidson et al. 2004), flowers replenished nectar more quickly than ants consumed it. We used two colors of dye (FD&C 5 and

FD&C 6) to differentiate male flowers with and without ants. In two trials for each bee colony, anthers of male flowers with ants were coated with FD&C 5 dye while brushes of male flowers lacking ants were coated with FD&C 6 dye. In the other two trials we reversed dye colors to control for effects of dye type and color. Dye color was assigned in a random order. Artificial flowers were placed in randomly assigned positions in the array before we opened the colony and allowed the bees to begin foraging.

b) Scent trials

We explored the effect of ant scent on dye donation in fifteen trials, testing three bee colonies in six, five, and four trials each. Bees were first exposed to ants by allowing the bees to forage for eight hours on twenty flowers (ten male and ten female), of which five male flowers had 15 ants and all others had none. No dye was used during this “exposure” day. We then collected all but the five ant-free male flowers in the flight cage, leaving these flowers to keep bees foraging.

Next, we individually stored five new male lids in Fluon-treated containers with ten M. rubra workers. Five control male lids were put in identical containers lacking ants. The following

61 morning, we removed the remaining flowers and set out ten new male and female flowers in a random spatial arrangement, using the lids having or lacking ant scent. Thus, the ant-scented flowers were not in the same positions as the ant-visited flowers on which the bees were trained.

Only five male flowers had ant scent and the other 15 flowers (five male and ten female) did not.

We used the same two dye colors as in i) to differentiate male flowers with and without ant scent, and randomized which color was used for ant-scented flowers between trials. We collected and replaced stigmas from female flowers after one hour, and collected the stigmas again after four hours. In the first trial performed, we also collected stigmas after two hours, but due to the small amount of dye transferred this was not repeated.

iv) Dye quantification

After each trial, we quantified the amount of dye transferred to female flowers using a spectrophotometer. We removed the stigmas from female flowers, placed them in test tubes, added 5.1 mL of distilled water to each tube, and vortexed each tube thoroughly to ensure the dye was evenly diluted. These were diluted further, as needed, if the absorbance exceeded the sensitivity range of the spectrophotometer. In the first six scent trials, each stigma was analyzed separately, and the dye amounts were summed to obtain a total amount of each dye color transferred to female flowers in each trial. In all the ant presence trials and the other nine scent trials, we opted to treat the experiment as the unit of replication. Therefore, we combined stigmas from each treatment type (ant-visited or ant-free) in a test tube and measured the total amount of each dye color transferred in each treatment in each trial.

Because we put different dye colors on male flowers with and without ants, we could use the amount of each dye color donated to all female flowers to measure the reproductive success

62 of ant-visited and ant-free male flowers. The total amount of dye (of both colors) received by female flowers with and without ants was our measure of female reproductive success. We calculated the amount of each dye color in the sample by measuring absorbance at 428 or 486 nm, and converting absorbance to micrograms following computational methods for overlapping spectra (Blanco et al. 1989).

v) Statistical analyses

In two trials (one ant presence and one ant scent), most female flowers received no dye because of low bee activity, so we excluded these trials from analyses. Because dye reception and donation values were non-normally distributed, we square-root transformed the data before examining the effects of ant presence or ant scent on dye transfer in ANCOVAs. For ant presence trials, we included ant presence on male and female flowers and their interaction as main effects, and the total amount of dye transferred in each trial as a covariate, to account for the large variation in overall dye transfer among trials. For ant scent trials, we included ant scent as the main effect, time (one or four hours) as a repeated measure, the interaction between scent and time, and the total amount of dye transferred as a covariate. Covariate by treatment (ant presence or ant scent) interactions were never significant, and so were excluded in final analyses.

All analyses were conducted in JMP (v. 10.0.0).

Results

Bees usually started foraging within minutes of the beginning of the trial and continued until flowers were collected from the flight cage or the lights were extinguished. An average of 3.4 ±

0.12 and 4.0 ± 0.15 (mean ± SE) bees were foraging after an hour in the ant presence and ant

63 scent trials, respectively. Male flowers with ants donated significantly less dye than male flowers lacking ants (Fig. 2.2a, F1,55 = 40.19, p < 0.0001). Similarly, female flowers with ants received significantly less dye than female flowers lacking ants (Fig. 2.2a, F1,55 = 4.61, p = 0.036). There was no significant interaction between ant presence on male flowers and ant presence on female flowers (F1,55 = 1.18, p = 0.28). Flowers with ant scent donated significantly less dye than flowers without ant scent (Fig. 2.2b, F1,25 = 112.16, p < 0.0001). There was no significant effect of time (F1,25 = 1.21, p = 0.28) or interaction between time and scent (F1,25 = 1.11, p = 0.30) in the model.

Discussion

In this study, ants altered bumblebee pollination behavior and exhibited trait-mediated indirect interactions with (artificial) flowers. Through interference competition, M. rubra workers significantly affected the pattern of dye transfer by B. impatiens, causing a preferential flow of dye from male flowers lacking ants to female flowers lacking ants. If manifested in nature, such effects would reduce the reproductive success of plants visited by ants through both male and female function.

Ants could have changed the attractiveness of artificial flowers in at least two ways. First, as ants interrupted or altogether prevented bumblebees from foraging at nectaries, bumblebees may have learned to avoid flowers with ants due to the relative inefficiency of foraging at these flowers, something bumblebees take into account (Heinrich 2004). Second, bees attempt to minimize their risk of being attacked during foraging bouts, which ants did to visiting bees. Ants often attacked or harassed bees by biting, grasping and appearing to sting visiting bees, preventing them from accessing the nectary or reducing their time on flowers (see online video

1), and bees sometimes avoided ant-tended flowers entirely. This harassment was sometimes physically traumatic (online video 1) and bees often appeared to have trouble flying after being attacked by ants. Previous research has demonstrated that bumblebees leave or avoid flowers where they have been harassed (Jones and Dornhaus 2011) and avoid foraging where there is visual or olfactory evidence of a predator or predation event (Abbott 2006, Goodale and Nieh

2012). Though it is unclear if bumblebees viewed ants as competitors or predators, they responded similarly to flowers having ants as they do to flowers housing predators (Gonçalves-

Souza et al. 2008).

To avoid artificial flowers with ants, bees likely used a combination of visual and olfactory cues. Previous research has shown that bees respond to conspicuous predators or predator “dummies” on flowers (Suttle 2003; Gonçalves-Souza et al. 2008), decreasing the frequency and duration of their visits to these flowers. Thus, bees may have been able to avoid flowers harboring ants by sight alone. However, even in the absence of ants, bees still preferentially visited flowers lacking ant scents (Fig. 2.2b), suggesting that they had learned to associate ant scent with harassment. Bees are adept at associative learning (Wright and Schiestl

2009) and can learn to recognize unique scents left behind by both conspecific and heterospecific flower visitors (Stout et al. 1998). Bees can use these various scents, often arising from tarsal gland deposits (Stout et al. 1998), to recognize recently visited flowers that are less likely to be profitable (Stout and Goulson 2002). Recently, Ballantyne and Willmer (2012) demonstrated that bees learn to associate ant scents with unrewarding artificial flowers and decrease their visitation to these flowers. Our results complement their findings, by showing that bees can associate ant scent with harassment at otherwise rewarding flowers.

In our study, ant scent caused bees to adjust their foraging strategy, and decreased the amount of pollen analogue that was donated by ant-scented flowers. Thus, the effects of

65 interference competition with ants on flowers can extend beyond immediate interactions and may have fitness consequences for plants, even when ants are absent. Like other olfactory cues, these effects are likely transitory (Stout et al. 1998). Although more dye was still donated by male flowers lacking ant scent than those with ant scent, the ratio of dye donated was, on average, closer to equality in hours 2-4 than in the first hour (Fig. 2.2b). This effect may be partially driven by dye depletion of male flowers without ant scent, but the lack of a strong corresponding decrease in flowers with ant scent suggests that visitation patterns became more similar. We did not test whether B. impatiens’ avoidance of ant scent was an innate or a learned behavior, but previous work has shown that Bombus terrestris does not innately avoid flowers with ant scent

(Ballantyne and Willmer 2012).

Ants may be necessary for plant survival and growth but can be costly for plant reproduction. The evolution of ant attractants such as extrafloral nectaries in some plant lineages suggests that the costs of ants can be outweighed by their protective abilities. However, in this study, flowers visited by ants received and donated significantly less dye, although they did retain some sexual function. The net benefit of having ants depends on whether ants increase plant fitness by reducing herbivory more than they decrease plant fitness by disrupting pollination. Alternatively, plants may actually benefit from the costs of ants to plant reproduction; ants that castrate flowers may be better defenders, increasing plant survival or vegetative growth when plants are young, allowing for increased reproduction later in life when the plant is colonized by less aggressive, non-castrating ants (Frederickson 2009, Palmer et al.

2010). In some plant species, pollinator harassment by ants may even be beneficial. Altshuler

(1999) reported that Ectatomma ants greatly increased fruit set of Psychotria limonensis, despite reductions in pollinator visitation rates, presumably due to increases in the rate of pollen outcrossing. Similarly, Gonzálvez et al. (2012) found that the presence of Oecophylla

66 smaragdina ants on Melastoma malabathricum flowers reduced visitation by less effective pollinators and increased visitation by more effective Xylocopa bees, increasing plant fitness.

Cases of ants being beneficial to flowers appear to be the minority, and many plant species have traits that limit ant access to flowers. Several studies have detailed floral volatiles that are thought to mimic ant alarm pheromones and thus repel ants (Junker and Blüthgen 2008,

Willmer et al. 2009). Other plants use structural modifications, such as narrow corollas or slippery stems, to limit ant access to flowers (reviewed in Willmer et al. 2009). Exploitative competition between ants and pollinators (Lach 2005; Ballantyne and Willmer 2012) as well as the lethal effects of ant antibiotic secretions on pollen grains (Beattie et al. 1985; Dutton and

Frederickson 2012) may have resulted in selection on plants to reduce ant visits to flowers. Our results suggest that trait-mediated indirect interactions resulting from interference competition between ants and bees may favor plants that defend their flowers and their pollinators against ants.

Acknowledgements

We thank H. Rusnock and S. Meadley-Dunphy for assistance, J. Dix for help with flight cage construction, and Biobest for bumblebee colonies. We are grateful to the Thomson and

Frederickson labs for guidance, particularly T. Makino, funded by JSPS, for helping to refine artificial flowers, J. Ogilvie for help with bees, K. Prior for M. rubra colony collections, and K.

Turner for study design suggestions. The comments of the editor and two anonymous reviewers prompted improvements in the article, especially with regards to study design. MEF and JDT acknowledge NSERC Discovery Grant funding. The University of Toronto’s Faculty of Arts and

Science supported MGRT’s participation.

References

Abbott, K. R. 2006. Bumblebees avoid flowers containing evidence of past predation events.

Canadian Journal of Zoology 84:1240-1247.

Altshuler, D. L. 1999. Novel interactions of non-pollinating ants with pollinators and fruit

consumers in a tropical forest. Oecologia 119:600-606.

Ballantyne, G. and Willmer, P. 2012. Floral visitors and ant scent marks: noticed but not used?

Ecological Entomology 37:402-409.

Beattie, A. J., Turnbull, C., Knox, R. B., and Williams, E. G. 1984. Ant inhibition of pollen

function: a possible reason why ant pollination is rare. American Journal of Botany

71:421-426.

Blanco, M., Runiaga, H., Maspoch, S., and Tarin, P. 1989. A simple method for

spectrophotometric determination of two-components with overlapped spectra. Journal of

Chemical Education 66:178-180.

Byk, J., & Del Claro, K. 2010. Nectar- and pollen-gathering Cephalotes ants provide no

protection against herbivory: a new manipulative experiment to test ant protective

capabilities. Acta Ethologica 13:33-38.

Davidson, D. W., S. C. Cook, and R. R. Snelling. 2004. Liquid-feeding performances of ants

(Formicidae): ecological and evolutionary implications. Oecologia 139:255–66.

Dukas, R. 2005. Bumble bee predators reduce pollinator density and plant fitness. Ecology

86:1401-1406.

Dussutour, A., and Simpson, S. J. 2009. Communal nutrition in ants. Current Biology 19:740-

744.

Dutton, E. M., and Frederickson, M. E. 2012. Why ant pollination is rare: new evidence and

implications of the antibiotic hypothesis. Arthropod-Plant Interactions 6:561-569.

Frederickson, M.E. 2009. Conflict over reproduction in an ant-plant symbiosis: why Allomerus

octoarticulatus ants sterilize Cordia nodosa trees. American Naturalist 173:675-681.

Galen, C. and Cuba, J. 2001 Down the tube: pollinators, predators, and the evolution of flower

shape in the alpine skypilot, Polemonium viscosum. Evolution 55:1963-1971.

Gonçalves-Souza, T., Omena, P. M., Souza, J. C., and Romero, G. Q. 2008. Trait-mediated

effects on flowers: artificial spiders deceive pollinators and decrease plant fitness. Ecology

89:2407-2413.

Gonzálvez, F. G., Santamaría, L., Corlett, R. T., and Rodríguez-Gironés, M. A. 2012. Flowers

attract weaver ants that deter less effective pollinators. Journal of Ecology 101:78–85.

Goodale, E., and Nieh, J.C. 2012. Public use of olfactory information associated with predation

in two species of social bees. Animal Behaviour 84:919-924.

Harder, L.D. and Thomson, J.D. 1989. Evolutionary options for maximizing pollen dispersal of

animal-pollinated plants. American Naturalist 133:323-344.

Heil, M., and McKey, D. 2003. Protective ant-plant interactions as model systems in ecological

and evolutionary research. Annual Review of Ecology, Evolution, and Systematics 34:425-

453.

Heinrich, B. 2004. Bumblebee economics, 2nd edn. Harvard University Press, Cambridge.

Ings. T. C., and Chittka, L. 2009. Predator crypsis enhances behaviourally mediated indirect

effects on plants by altering bumblebee foraging preferences. Proceedings of the Royal

Society B 276:2031-2036.

Inouye, D. W., and Taylor, O. R. Jr. 1979. A temperate region plant-ant-seed predator system:

consequences of extra floral nectar secretion by Helianthella quinquenervis. Ecology 60:1-

Junker, R. R., and Blüthgen, N. 2008. Floral scents repel potentially nectar-thieving ants.

Evolutionary Ecology Research 10:295–308.

Junker, R. R., Bleil, R., Daehler, C. C., and Blüthgen, N. 2010. Intra-floral resource partitioning

between endemic and invasive flower visitors: consequences for pollinator effectiveness.

Ecological Entomology 35:760–767.

Jones, E. I., and Dornhaus, A. 2011. Predation risk makes bees reject rewarding flowers and

reduce foraging activity. Behavioral Ecology and Sociobiology 65:1505-1511.

Lach, L. 2005. Interference and exploitative competition of three nectar-thieving invasive ant

species. Insect Sociaux 52:257-262.

Lach, L. 2007. A mutualism with a native membracid facilitates pollinator displacement by

Argentine ants. Ecology 88:1994-2004.

Lach, L. 2008. Argentine ants displace floral arthropods in a biodiversity hotspot. Diversity and

Distributions 14:281-290.

Makino, T. T., and Sakai, S. 2007.Experience changes pollinator responses to floral display size:

from size-based to reward-based foraging. Functional Ecology 21:854:863.

Maloof, J. E., and Inouye, D. W. 2000. Are nectar robbers cheaters or mutualists ? Ecology

81:2651–2661.

Ness, J. H. 2006. A mutualism’s indirect costs: the most aggressive plant bodyguards also deter

pollinators. Oikos 113:506-514.

Ohashi, K., Leslie, A., and Thomson, J.D. 2008. Trapline foraging by bumble bees: V. Effects of

experience and priority on competitive performance. Behavioral Ecology 19:936-948.

Palmer, T. M., Doak, D. F., Stanton, M. L., Bronstein, J. L., Kiers, E. T., Young, T. P., Goheen,

J. R., and Pringle, R. M. 2010. Synergy of multiple partners, including freeloaders,

increases host fitness in a multispecies mutualism. Proceedings of the National Academy

of Sciences 107:17234-17239.

Preisser, E. L., Bolnick, D. I., and Benard, M. E. 2005. Scared to death? The effects of

intimidation and consumption in predator-prey interactions. Ecology 86:501-509.

Stout, J. C., Goulson, D., and Allen, J. A. 1998. Repellent scent-marking of flowers by a guild of

foraging bumblebees (Bombus spp.). Behavioral Ecology and Sociobiology 43:317-326.

Stout, J., and Goulson, D. 2002. The influence of nectar secretion rates on the responses of

bumblebees (Bombus spp.) to previously visited flowers. Behavioral Ecology and

Sociobiology 52:239-246.

Suttle, K. B. 2003. Pollinators as mediators of top-down effects on plants. Ecology Letters

6:688-694.

Thomson, J.D. 1986. Pollen transport and deposition by bumble bees in Erythronium: influences

of floral nectar and bee grooming. Journal of Ecology 74:329-341.

Thomson, J.D., Ogilvie, J. E., Makino, T. T., Arisz, A., Raju, S., Rojas-Luengas, V., and Tan, M.

G. R. 2012. Estimating pollination success with novel artificial flowers: effects of nectar

concentration. Journal of Pollination Ecology 9:108-114.

Tsuji, K., Hasyim, A., and Nakamura, K. 2004. Asian weaver ants, Oecophylla smaragdina, and

their repelling of pollinators. Ecological Research 19:669-673.

Ulrich, S., and Bertsch, A. 1990. Do foraging bumblebees scent-mark food sources and does it

matter? Oecologia 82:137-144.

Werner, E. E., and Peacor, S. D. 2003. A review of trait-mediated indirect interactions in

ecological communities. Ecology 84:1083-1100.

Willmer, P. G., Nuttman, C. V., Raine, N. E., Stone, G. N., Pattrick, J. G., Henson, K., Stillman,

P., McIlroy, L., Potts, S. G., and Knudsen, J. T. 2009. Floral volatiles controlling ant

behaviour. Functional Ecology 23:888-900.

Wright, G. A., and Schiestl, F. P. 2009. The evolution of floral scent: the influence of olfactory

learning by insect pollinators on the honest signaling of floral rewards. Functional Ecology

23:841-851.

Figures

Figure 2.1 Photograph of artificial flower lids. A = anther, L = lid, N = nectary, R = region where anther or stigma is placed, S = stigma, T = thread, W = weight.

Figure 2.2 Dye transferred (mean ± 1SE) per trial by B. impatiens workers A) from artificial male and to artificial female flowers in the presence (unfilled circles) or absence (filled circles) of Myrmica rubra workers, and B) from artificial male flowers with (unfilled circles) or without

(filled circles) M. rubra scent to ant-free female flowers in the 1st or the 2nd to 4th hours of the trials.

Chapter 3 Not just for the bees: pollen consumption is common among tropical ants

Cembrowski, A.R., Reurink, G., and Frederickson, M.E.

Planned submission to Biotropica ARC and MEF designed the study, ARC and GR did laboratory work, ARC performed statistical analyses and wrote the manuscript with input from other authors

Abstract

Although palynivory, or pollen consumption, is widespread among other hymenopterans, there are few accounts of palynivory in ants. To quantify how often ants consume pollen, we subjected adult workers and larvae from 75 species of neo- and paleo-tropical ants to acetolysis, a process that destroys most organic material but leaves behind pollen grains. In over half of the species we examined, ants contained a few pollen grains (< 5) and ants of several Camponotus species contained abundant pollen grains (>50). We tested for an association between trophic level and palynivory using stable nitrogen isotope ratios (δ15N), but we did not find a significant correlation; we found pollen grains in ‘herbivorous’, omnivorous, and carnivorous ants. We suggest that our results indicate sporadic, opportunistic pollen consumption by ants in tropical forests.

Introduction

For those organisms that can digest it efficiently, pollen can be an excellent food source.

Depending on the plant species, pollen can contain up to 60% protein, as well as lipids,

75 76 carbohydrates, and numerous trace nutrients (Roulston and Cane 2000). Pollen feeding

(palynivory) occurs in many taxa, including birds, mammals, and arthropods (Roulston and Cane

2000), but it is best known among the Hymenoptera, particularly bees. Although adult bees feed mainly on plant exudates (i.e., nectar), nearly all bee species rely on pollen as the chief protein source for developing larvae (Willmer 2011). Closely related to bees, ants have similar nutritional requirements; adult workers need carbohydrates, whereas larvae are fed a diet rich in protein (Dussutour and Simpson 2009). Ants also exploit nectar heavily and regularly visit flowers (e.g., Herrera et al. 1984, Blüthgen et al. 2004), but unlike bees are not believed to widely consume pollen. Instead, ants are thought to receive most of the protein in their diet from predation or scavenging (e.g., Floren et al. 2002).

However, ants may not be as reliant on animal protein sources as commonly assumed. In tropical forests, arboreal ants can have similar δ15N ratios, a measure of trophic level, as herbivores, suggesting limited carnivory and substantial feeding on plant-derived protein sources

(Davidson et al. 2003). One potential protein source is pollen. This idea is supported by scattered accounts of pollen consumption by ants from other ecosystems, including deserts (Ness 2006), boreal forests (Czechowski et al. 2009), and tropical savannas (Byk and Del Claro 2010). One ant genus, Cephalotes, is widely recognized to consume pollen (Creighton 1967, Baroni Urbani and de Andrade 1997, Byk and Del Claro 2010). However, beyond this genus, our knowledge of pollen consumption by ants has largely been limited to direct observations (but see Czechowski et al. 2011), and it is unknown if palynivory is truly rare among ants or simply under-reported.

We borrowed a technique commonly used in palynology, acetolysis, to assess the frequency of pollen consumption by ants in 75 neo- and paleo-tropical ant species. This technique allowed us to examine many more ant species than we could have observed directly.

Methods

Adult worker ants were collected by hand, at baits, or from leaf litter using Winkler extractors, while larvae, along with adults, were collected by hand from nests. Collections were made by L.

Arcila Hernandez and J. Sanders in the primary tropical rainforest surrounding the Centro de

Investigaciόn y Capacitaciόn Rio Los Amigos (CICRA, 12°34’S, 70°05’W) in Peru in 2010-

2011 and by E. Youngerman from lowland primary and secondary rainforests in Madang

Province, Papua New Guinea (PNG) in 2011-2012. We sorted ant samples to morphospecies and then identified them to genus, and when possible to species using existing keys, which were unavailable for PNG ants. Ant samples used for acetolysis were whole adult workers, guts from adult workers, and whole larvae.

We used acetolysis to look for pollen grains inside ants. In acetolysis, organic material is dissolved in acetic anhydride and sulfuric acid, leaving behind stained pollen exines. When possible, we used adult and larval ants of the same species from different colonies or bait locations in an attempt to sample more representatively. We prepared samples by first washing them in 70% ethanol to remove any pollen present on the sample’s outer surface, before placing them individually in micro-centrifuge tubes with ~0.5mL of a 9:1 mixture of acetic anhydride: sulfuric acid. We then heated each sample at 95°C for 20min., allowing tissues to dissolve. Next, we centrifuged samples at 1000 RPM for 5min. and decanted the liquid supernatant, before adding ~0.5mL of glacial acetic acid to the precipitate, centrifuging again as above. We decanted the samples and added a drop of distilled water.

We examined a 150μL sub-sample of this solution for pollen grains under a light microscope. Larvae gave less than 150μL of solution following acetolysis, so we sampled the entire amount. We identified pollen grains by their staining and their surface architecture. For most species, we examined at least three adult workers and three larvae, but for some species fewer than three adult workers or larvae were available (Tables 3.1 and 3.2).

We investigated the relationship between trophic level, as measured by δ15N ratios, and pollen consumption by ants. Stable isotope (SI) data was available for some (n =23) of the

Peruvian species from a separate project in which collections overlapped. Samples were stored in

EtOH, and dried at 50°C overnight. We used heads, legs and thoraxes (but not gasters) of ants for SI analysis. Often, single individuals did not provide enough material for analysis, and we combined multiple individuals to achieve necessary weights, up to 10 for small ant species.

Samples were analyzed at the Boston University Stable Isotope Laboratory, using a Finnigan

Delta-S analyzer. For other Peruvian species for which we had samples but not corresponding stable isotope data (n =25), we took the median δ15N ratio value of that species’ genus and assigned it to the species in question. Using pollen presence/absence data for these 48 species, we used logistic regression to correlate pollen presence and δ15N ratio using R (v. 2.15.1).

Results

We found pollen grains in 37 of the 197 adult workers, 20 of the 102 larvae, and 38 of the 75 ant species we examined (Tables 3.1 and 3.2). The proportion of individuals containing pollen did

2 not differ significantly between larvae and adults (χ 1 = 0.02, p = 0.88). The number of pollen grains found was often low (median = 3), but highly variable (range: 1-468). Also, pollen

79 presence was inconsistent among conspecific ant samples; of species that contained pollen, 31% had pollen in only one of the larvae or adults examined. In only three species (two Cephalotes and one Myrmelachista) did all examined adults contain pollen.

Highly ‘herbivorous’, carnivorous and omnivorous ants contained pollen grains (Fig.

3.1). There was no correlation between δ15N ratio and pollen presence (χ2 = 1.50, df = 1, p =

0.22). In fact, several carnivorous species contained pollen grains (e.g., Odontomachus sp. 1,

Pachycondola sp.1).

Discussion

Pollen consumption by ants appears to be more common than suspected. The majority of species

(51%) had at least one individual that had consumed pollen, and in 20% multiple individuals contained pollen. Numbers of grains found tended to be low, however, with most individuals containing only a few grains. Ants may use pollen as only a small portion of their diet, exploiting a host of other protein-containing foods more heavily, including other low- δ15N foods (e.g., honeydew, epiphylls; Hölldobler and Wilson 1990, Davidson et al. 2003).

There are a few reasons why pollen may not be heavily exploited by ants. Pollen requires much post-consumption processing to extract its nutrients. These nutrients are located in the pollen’s cytoplasm, which is protected from direct digestion by the exines and intines of the grains. Organisms must either break through these layers or cause the pollen grains to germinate or pseudo-germinate to access and digest the cytoplasm (Roulston and Cane 2000). This can be achieved by consuming a sugar-containing solution (such as nectar) along with pollen. However, animals that are not adapted to feed on pollen may not be very efficient at this process, only

80 extracting nutrients from some of the grains (Herrera and Martínez del Rio 1998). Thus, while ants can likely get some nutrients from pollen, other forms of more easily processed protein may be more attractive.

Pollen may also pose handling difficulties for ants. Despite the large number of studies that have recorded ants visiting flowers (Chapter 1), only a few papers detail ants harvesting pollen directly from anthers. Byk and Del Claro (2010) reported that Cephalotes pusillus ants removed almost all of the available pollen from Ouratea spectabilis flowers, and at least two other studies describe ants consuming pollen along with nectar from flowers (Horskins and

Turner 1999, Ness 2006). Unlike bees, which possess both morphological (e.g., specialized hairs for holding pollen) and behavioral adaptations (e.g., stereotyped pollen-packing motions), ants have no apparent mechanisms to increase the amount of pollen harvested, and are likely limited by the number of grains they can ingest or carry. Rather than collect pollen directly from flowers, it appears more likely that ants collect it haphazardly in the environment, leading to a smaller number of pollen grains consumed at any one time. It is known that at least Cephalotes ants engage in such behavior; Creighton (1967) observed Cephalotes texanus collecting pollen grains trapped by Citrus x. paradisi leaf hairs.

There is a chance that the small numbers of pollen grains we observed came from secondary pollen consumption (i.e., consuming a palynivorous organism) or consumption of liquid containing pollen. While we cannot rule this out, we believe our results represent actual, targeted pollen consumption by ants. Cephalotes ants, known palynivores (Creighton 1967,

Baroni Urbani and de Andrade 1997), contained similar numbers of pollen grains as many of the pollen-containing ant species that we examined. As well, in at least once instance, ants that we found containing pollen grains were in the process of preying on non-pollen consuming

81 organisms (e.g., a Pachycondola species that was carrying a termite in its mandibles) or were attracted by sugar baits.

Contrary to our expectations, there was no relationship between an ant species’ δ15N ratio and whether it had consumed pollen (Fig. 3.2). We expected palynivory to be more common in herbivorous ants because pollen presents a widespread source of protein, a potentially limiting resource. The lack of pattern might reflect how available pollen grains are in the environment.

While other forms of food may be preferable, ants tend to be opportunistic foragers (Carroll and

Janzen 1973) and there is likely little cost associated with the consumption of scattered pollen grains during foraging trips beyond the volume they occupy in the gut. Some genera were found to consistently consume pollen. In support of previous observations (Baroni Urbani and de

Andrade 1997, Byk and Del Claro 2010), Cephalotes species were among the most common pollen consumers. Camponotus species also consistently contained pollen grains; of the eight species examined, five contained pollen, and pollen was found in species from both Peru and

PNG. Camponotus species also contained the two largest numbers of pollen grains found (468 and 271). Taken together, these results suggest that they may be another, as yet unrecognized, commonly palynivorous genus.

In general, larvae did not contain large numbers of pollen grains (ranging from 1-38 grains). This is counter to what is seen in bees, where larvae are the main recipients of pollen

(Willmer 2011). This difference may be because ant larvae are continually fed via trophallaxis by adults throughout their development, whereas bee larval cells are provisioned with pollen and then sealed while larvae complete their development. If both adults and larvae are able to extract nutrients from pollen, we would expect to see a preferential flow of nutrients from pollen to larvae, but not necessarily the grains themselves.

Unlike palynivory by organisms that also provide pollination services, pollen consumption by ants is unlikely to benefit plants. Ants rarely act as pollinators (Beattie et al.

2004). They have pollen-killing antibiotic secretions on their cuticles (Dutton and Frederickson

2012) and they can compete with other flower visitors, reducing flower visitation rates (Ness

2006). The negative impacts of flower-visiting ants are thought to have led to the evolution of ant-repellent floral volatiles (reviewed in Willmer 2009). These volatiles can pre-empt competition between ants and pollinators, and may often be located within the pollen itself

(Willmer 2009). If floral volatiles truly are a defense against ant visitation, our results suggest an additional evolutionary pressure to not only prevent ant-pollinator interactions, but to also protect plant gametes from occasional consumption by a non-pollinating visitor.

Acknowledgements

We would like to thank Jon Sanders, Lina Arcila Hernandez and Eric Youngerman for providing samples, the Thomson lab for use of supplies, the Pierce lab for funding support during JS’ and

EY’s collections, and the staff of CICRA and the New Guinea Binatang Research Center for logistical support. The Peruvian Ministry of Agriculture provided permits (Nos. 394-2009-AG-

DGFFS-DGEFFS and 299-2011-AG-DGFFS-DGEFFS). MEF acknowledges financial support from an NSERC Discovery Grant, a Connaught New Researcher Award, an Ontario Ministry of

Economic Development and Innovation Early Researcher Award, and the University of Toronto;

ARC was supported by an Ontario Graduate Scholarship and Sigma Xi.

References

Baroni Urbani, C., and M. L. de Andrade. 1997. Pollen eating, storing, and spitting by ants.

Naturwissenschaften 84:256–258.

Beattie, A. J., C. Turnbull, R. B. Knox, and E. G. Williams. 1984. Ant inhibition of pollen

function: a possible reason why ant pollination is rare. American Journal of Botany 71:421–

426.

Blüthgen, N., G. Gottsberger, and K. Fiedler. 2004. Sugar and amino acid composition of ant-

attended nectar and honeydew sources from an Australian rainforest. Austral Ecology

29:418–429.

Byk, J., and K. Del-Claro. 2010. Nectar- and pollen-gathering Cephalotes ants provide no

protection against herbivory: a new manipulative experiment to test ant protective

capabilities. Acta Ethologica 13:33–38.

Carroll, C. R., and D. H. Janzen. 1973. Ecology of foraging ants. Annual Review of Ecology and

Systematics 4:231–257.

Creighton, W. S. 1967. Studies on free colonies of Cryptocerus texanus Satschi (Hymenoptera:

Formicidae). Psyche 74:34–42.

Czechowski, W., B. Markό, and A. Radchenko. 2009. Rubbish dumps reveal the diet of ant

colonies: Myrmica schencki Em. and Myrmica rubra (L.) (Hymenoptera: Formicidae) as

facultative pollen-eaters. Polish Journal of Ecology 56:737–741.

Czechowski, W.,B. Markó, K. Erős, and E. Csata. 2011. Pollenivory in ants (Hymenoptera:

Formicidae) seems to be much more common than it was thought. Annales Zoologici

61:519–525.

Dussutour, A., and S. J. Simpson. 2009. Communal nutrition in ants. Current Biology 19:740–4.

Dutton, E. M., and M. E. Frederickson. 2012. Why ant pollination is rare: new evidence and

implications of the antibiotic hypothesis. Arthropod-Plant Interactions 6:561–569.

Floren, A., A. Biun, and K. E. Linsenmair. 2002. Arboreal ants as key predators in tropical

lowland rainforest trees. Oecologia 131:137–144.

Grant, B. R. 1996. Pollen digestion by Darwin’s finches and its importance for early breeding.

Ecology 77:489-499.

Herrera, L. G., and C. Martínez del Rio. 1998. Pollen digestion by new world bats: effects of

processing time and feeding habits. Ecology 79:2828–2838.

Hölldobler, B. and Wilson, E. O. 1990. The Ants, 1st Edition. Harvard University Press,

Cambridge.

Horskins, K. and Turner, V. B. 1999. Resource use and foraging patterns of honeybees, Apis

mellifera, and native insects on flowers of Eucalyptus costata. Australian Journal of

Ecology 24:221-227.

Ness, J. H. 2006. A mutualism’s indirect costs : the most aggressive plant bodyguards also deter

pollinators. Oikos 113:506–514.

Roulston, T. H. I., and J. H. Cane. 2000. Pollen nutritional content and digestibility for animals.

Plant Systematics 222:187–209.

Van Tets, I. G. 1997. Extraction of nutrients from Protea pollen by African rodents. Belgian

Journal of Zoology 127:59-65.

Willmer, P. G. 2011. Pollination and Floral Ecology, 1st Edition. Princeton University Press,

New Jersey.

Willmer, P. G., C. V. Nuttman, N. E. Raine, G. N. Stone, J. G. Pattrick, K. Henson, P. Stillman,

et al. 2009. Floral volatiles controlling ant behaviour. Functional Ecology 23:888–900.

Tables

Table 3.1 Numbers of adult worker and larval ants examined for each species collected in Peru.

Numbers of pollen grains found in individual samples are given in parentheses after the sample size; when there were no grains, no numbers are given.

Genus Species Number of adults analyzed Number of larvae analyzed (numbers of grains found) (numbers of grains found)

Dolichoderinae

Azteca sp. 1 3 0

sp. 2 2 0

sp. 3 0 2

sp. 4 3 (1) 0

sp. 5 3 (2) 0

Dolichoderus decollatus 3 (66) 0

sp. 1 3 0

sp. 2 3 1 (2)

sp. 3 3 3

sp. 4 0 2

Ecitoninae

Labidus sp. 1 3 0

Formicinae

Camponotus sp. 1 3 3

sp. 2 3 (82) 1 (2)

sp. 3 3 (3) 0

sp. 4 0 2

sp. 5 8 (1,1) 0

sp. 6 4 (7, 16, 468) 0

sp. 7 5 (12) 0

sp. 8 2 (1) 0

Paratrechina sp. 1 4 (1,1) 0

Myrmelachista sp. 1 3 (5, 11, 18) 0

sp. 2 0 3

Trachymyrmex sp. 1 3 (21) 0

sp. 2 3 0

Myrmicinae

Atta sp. 1 3 0

Cephalotes atratus 3 (2, 2, 21) 0

placidus 3 (1, 4, 21) 0

sp. 1 3 0

sp. 2 1 0

sp. 3 2 0

Crematogaster sp. 1 3 1

sp. 2 3 (35) 0

sp. 3 9 (7) 0

sp. 4 2 0

Megalomyrmex sp. 1 3 0

sp. 2 0 2

sp. 3 4 (7) 0

sp. 4 3 0

Monomorium sp. 1 0 3 (2)

Pheidole sp. 1 0 3

sp. 2 3 0

sp. 3 0 3

Procryptocerus sp. 1 3 0

Solenopsis sp. 1 0 3

sp. 2 3 (3, 5) 0

Wasmannia sp. 1 3 0

Ponerinae

Pachycondola sp. 1 3 (3) 0

Odontomachus sp. 1 3 (34) 0

Pseudomyrmicinae

Pseudomyrmex sp. 1 3 0

sp. 2 3 0

sp. 3 3 (9) 0

sp. 4 0 3

sp. 5 0 3

Table 3.2 Numbers of adult worker and larval ants examined for each species collected in Papua

New Guinea. Numbers of pollen grains found in individual samples are given in parentheses after the sample size; when there were no grains, no numbers are given.

Genus Species Number adults analyzed Number larvae analyzed (grains found) (grains found)

Aenictinae

Aenictus sp. 1 3 3 (1, 1)

Dolichoderinae

Philidris sp. 1 3 1 (4)

Tapinoma sp. 1 2 3 (38)

Technomyrmex sp. 1 0 3 (12, 12)

Ectatomminae

Rhytidoponera sp. 1 1 3 (2)

Formicinae

Acropyga sp. 1 3 3 (2)

sp. 2 3 3

sp. 3 3 3

Anoplolepsis sp. 1 3 (1) 3 (1)

Calomyrmex sp. 1 3 3 (1)

Camponotus sp. 1 3 (1, 271) 3

sp. 2 1 3

sp. 3 2 0

Nylanderia sp. 1 3 3

Opisthopsis sp. 1 3 3

Polyrachis sp. 1 1 3 (4)

Myrmicinae

Crematogaster sp. 1 3 3

sp. 2 3 (32) 3

Monomorium sp. 1 3 (1) 3 (1)

sp. 2 3 3 (1, 1, 1)

Pheidole sp. 1 3 3

Pristomyrmex sp. 1 3 3 (1,1)

Pyramica sp. 1 3 0

Ponerinae

Odontomachus sp. 1 3 3 (4)

Figures

Figure 3.1 Photographs of pollen grains found in two ant species: a) Dolichoderus sp. 2 and b) Camponotus sp. 6. Scale bars = 50 µm.

Figure 3.2 Relationship between trophic level, as measured by δ15N ratio, and pollen presence for 48 neotropical ant species: 1 Atta sp., 5 Azteca spp., 8 Camponotus spp., 5 Cephalotes spp., 4 Crematogaster spp., 5 Dolichoderus spp., 1 Labidus sp., 3 Megalomyrmex spp., 1 Monomorium sp., 2 Myrmelachista spp., 1 Odontomachus sp., 1 Pachycondola sp., 1 Paratrechina sp., 3 Pheidole sp., 4 Pseudomyrmex spp., 2 Solenopsis spp., and 1 Wasmannia sp.

Concluding Remarks

Ants have traditionally been considered to be unwelcome flower visitors; they consume floral rewards ostensibly meant for pollinators while rarely providing pollination services themselves.

In this thesis, I investigate the validity of this assumption and use an experimental study and an observational study to examine two ways that ants may impact plant reproduction.

In the first chapter, I review the literature on ants as flower visitors. I show that the common assumption of ants always being categorically negative for plant reproduction is incorrect; while ants may reduce floral attractiveness to pollinators, the net effect of ants on plant fitness measures is variable, ranging from positive to negative to neutral. I also discuss ants as pollinators and ways that plants may minimize ant visitation to flowers, before finishing with a section highlighting questions that still need investigation.

In the second chapter, I explore how interference competition with ants can structure bumblebee foraging behavior and how this may alter pollen movement patterns. Using artificial flowers with male or female function, I allowed Bombus impatiens bumblebees to forage on flowers that either did or did not have Myrmica rubra ants on them. Ant presence significantly reduced the amount of pollen analogue donated and received by flowers. Building on these results, I found that the presence of ant scent alone elicited similar changes to bumblebee behavior, reducing the amount of pollen analogue donated by male flowers. My results show that interference competition is sufficient to alter bumblebee foraging behavior and that the effects of ants on pollinators may extend after ant visitation to flowers. Further study is needed to determine if bees are able to differentiate between ant species by visual or olfactory cues. Their ability to do so

94 could affect how generalized their response to ant presence is and may help to better predict how ants will affect plant reproduction.

In the final chapter, I investigate how common palynivory is amongst tropical ant species. I used acetolysis and light microscopy to examine whole ants, larvae and ant guts from Papua New

Guinea and Peru. I found that tropical ants consistently contain pollen grains, albeit often in small numbers (<5). However, some genera, notably Camponotus and Cephalotes consistently contained higher numbers, providing the first evidence that Camponotus may be a palynivorous genus. The small numbers of pollen grains present may indicate that ants preferentially exploit other, more easily digestible high-protein foods over pollen. Future studies should investigate exactly what role pollen plays in ant diets by supplementing the diet of captive ant colonies with pollen. This would allow us to examine how pollen consumption affects colony demography, as well as how efficiently ants digest pollen and whether ants exhibit pollen preferences.

I bring up several pressing and unanswered questions within this thesis and highlight three here.

First, numerous plant species possess chemical or physical traits assumed to increase plant fitness by reducing ant visitation. However, this has rarely been tested and instead research has focused predominately on documenting ant repellence. We should move away from this and begin examining plant fitness in the presence or absence of these repellent traits to see how beneficial these traits truly are. Second, how flower visiting ants impact male fitness has been almost completely ignored in favor of female fitness. Future studies should investigate how ant visitation affects both pollen donation and pollen movement patterns. Lastly, research involving antibiotics present on ant cuticles and their effect on pollen viability has been too small in scope, often limited to demonstrating the pollen killing effects of ant secretions. Investigation into

95 whether plants have evolved any resistance to the secretions of the local ant community or why some plants appear unaffected by their ant partners may yield exciting results.

In conclusion, ants have the capacity to strongly affect plant reproduction, both positively and negatively. The fact that these diametrically opposing forces exist means we cannot approach flower-visiting ants as either positive or negative a priori. Instead, we should strive to work towards a better understanding of what is occurring in the system of interest and what role ants play within it.

Chapter 2, “Ants and ant scent reduce bumblebee pollination of artificial flowers” is current in press in The American Naturalist. Request for its inclusion here was sought and approved.