UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Macroecological patterns of biotic interactions and their consequences in prey communities Camacho, Luis 2020

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2021_may_camacho_luis.pdf [ 3.67MB ]
Metadata
JSON: 24-1.0395376.json
JSON-LD: 24-1.0395376-ld.json
RDF/XML (Pretty): 24-1.0395376-rdf.xml
RDF/JSON: 24-1.0395376-rdf.json
Turtle: 24-1.0395376-turtle.txt
N-Triples: 24-1.0395376-rdf-ntriples.txt
Original Record: 24-1.0395376-source.json
Full Text
24-1.0395376-fulltext.txt
Citation
24-1.0395376.ris

Full Text

   Macroecological patterns of  biotic interactions and their consequences in prey communities  by  Luis Camacho  M.Sc., Pontificia Universidad Católica del Ecuador, 2014  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  Doctor of Philosophy in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Zoology)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  December 2020  © Luis Camacho, 2020 ii  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, the dissertation entitled: Macroecological patterns of biotic interactions and their consequences in prey communities  submitted by Luis Camacho in partial fulfillment of the requirements for the degree of Doctor of philosophy in Zoology  Examining Committee: Leticia Avilés, Zoology, UBC Supervisor  Jill Jankowsky, Zoology, UBC Supervisory Committee Member  Wayne Maddison, Zoology, UBC Supervisory Committee Member Mary O’Connor, Zoology, UBC University Examiner Allan Carroll, Forestry, UBC University Examiner  Additional Supervisory Committee Members: Juli Carrillo, Land and Food Systems, UBC Supervisory Committee Member    iii  Abstract  Seventy years after Dobzansky suggested that biotic interactions are more important in the tropics, ecologists are still assembling evidence and elucidating potential mechanisms behind such macroecological patterns. My thesis contributes to this field by demonstrating that predation rates and the strength of mutualistic associations decrease with elevation in the New World tropics.  I also tease apart possible mechanisms behind these patterns, which are likely ultimately linked to changes in temperature and productivity with elevation.  Using manipulative experiments and field observations across 4000-meter elevational gradients in the equatorial Andes, I show that decreasing predation rates on arthropods at higher elevations are driven by decreasing predator abundance and activity. Ants were responsible for 80% of predation in the lowlands, but were replaced by other predators above 1500m, revealing a parallel qualitative gradient of predation.   Along the same gradients, I show that the frequency of ant-hemipteran mutualistic associations also decreases with elevation, driven by a decrease in ant and hemipteran abundance. However, ant abundance limits the interaction above 1500m and hemipteran abundance below, which shaped the resource-consumer dynamics occurring in partner aggregations.  I provide evidence that, in response to pervasive ant predation in the lowlands, hemipterans may ‘bribe’ ants with honeydew primarily to dissuade them from predating upon them, rather than, as generally assumed, to obtain their defensive services. This is the first documentation of a mutualistic interaction involving prey offering a reward to predators in exchange for their lives, challenging our fundamental understanding of both mutualism and predation.  Finally, I show that anti-predator investment in hemipteran communities does not decrease at the same rate as predation rates at higher elevations, suggesting that the fitness effects of predation may decline at slower rate than predation rates. Hemipteran investment in ant-mutualism across elevations mirrored the contribution of ants to predation further supporting the idea of ant-hemipteran mutualism as an anti-predator strategy from ants themselves.  iv   Lay Summary  Interactions between species –predation, competition, and mutually beneficial –are fundamental forces shaping biodiversity and maintaining ecosystems. I studied predation and mutualism in insect communities of species across elevations in the tropical Andes where the broad range of environmental conditions can help us understand global patterns in biodiversity. I show predation is less frequent at higher elevations. Here, ants were responsible for 80% of predation in the lowlands but were replaced by other predators above 1500 m. In response to this, sap-sucking insects take advantage of the overwhelming ant presence in the lowlands and recruit them as bodyguards in exchange of honeydew; but, at the highlands, rely in maternal care to defend their offspring from predators. Interestingly, I show that sap-sucking insects offer honeydew to ants primarily to avoid being predated by them rather than obtaining their defensive services, akin to a bribe.         v  Preface  I identified the research system and questions, collected the data, and performed the analysis for all chapters of the thesis.  A version of chapter 2 has been published. Camacho, F. L. & Avilés, L. (2019) Decreasing predator density and activity explains declining predation of insect prey along elevational gradients. The American Naturalist 194: 334-343. I conducted all stages of the research. The manuscript was written jointly with Avilés, L.      vi  Table of Contents  Abstract .............................................................................................................................. iii Lay Summary ...................................................................................................................... iv Preface ................................................................................................................................. v Table of Contents ............................................................................................................... vi List of Tables ................................................................................................................... viii List of Figures .................................................................................................................... ix Acknowledgements ........................................................................................................... xi Dedication ........................................................................................................................ xii Chapter 1. General introduction .......................................................................................... 1 1.1 Macroecological gradients across elevations in the Andes ........................................................ 1 1.2 Treehoppers as a study system ....................................................................................................... 2 1.3 Overview of Chapters ..................................................................................................................... 6 Chapter 2. Decreasing predator density and activity explain declining predation rates along elevational gradients .................................................................................................. 9 2.1 Introduction ...................................................................................................................................... 9 2.2 Methods ........................................................................................................................................... 12 2.2.1 Study area ................................................................................................................................ 12 2.2.2 Data collection ........................................................................................................................ 12 2.2.3 Data analysis ............................................................................................................................ 14 2.3 Results .............................................................................................................................................. 15 2.4 Discussion ....................................................................................................................................... 17 2.4.1 Conclusion............................................................................................................................... 21 Chapter 3. Partner supply and demand mediate macroecological gradients in the outcome of ant-hemipteran mutualistic associations ...................................................................... 22 3.1 Introduction .................................................................................................................................... 22 3.2 Methods ........................................................................................................................................... 25 3.2.1 Study location ......................................................................................................................... 25 3.2.2 Data collection ........................................................................................................................ 25 3.2.3 Data analysis ............................................................................................................................ 27 3.3 Results .............................................................................................................................................. 28 3.4 Discussion ....................................................................................................................................... 32 3.4.1 Conclusion............................................................................................................................... 35 vii  Chapter 4. Consequences of macroecological gradients of predation on maternal care and mutualistic interactions with ants in Andean treehoppers ............................................... 36 4.1 Introduction .................................................................................................................................... 36 4.2 Methods ........................................................................................................................................... 37 4.2.1 Predation and mutualist-ant availability across elevations in the Andes ........................ 37 4.2.2 Behavioral composition of treehopper communities across elevations in the Andes . 39 4.2.3 Data analysis ............................................................................................................................ 41 4.3 Results .............................................................................................................................................. 42 4.4 Discussion ....................................................................................................................................... 47 4.4.1 Conclusion............................................................................................................................... 50 Chapter 5. Befriending the enemy: Can prey establish mutualistic interactions with their would-be predators? .......................................................................................................... 52 5.1 Introduction .................................................................................................................................... 52 5.2 Methods ........................................................................................................................................... 54 5.2.1 Predation by potential ant partners and occurrence of mutualist treehoppers across elevations in the Andes ................................................................................................................... 54 5.2.2 Partner recognition and defensive value of honeydew ..................................................... 56 5.3 Results .............................................................................................................................................. 59 5.3.1 Predation by potential ant partners and occurrence of mutualist treehoppers across elevations in the Andes ................................................................................................................... 59 5.3.2 Partner recognition and defensive value of honeydew ..................................................... 60 5.4 Discussion ....................................................................................................................................... 63 Chapter 6. Conclusion ....................................................................................................... 67 References.......................................................................................................................... 70 Appendices ........................................................................................................................ 86 Appendix A. Supplementary figures .................................................................................................. 86 Appendix B. Supplementary tables .................................................................................................... 89      viii  List of Tables  Table 2.1. Studies addressing macroecological patterns of predation across elevation or latitude........................................................................................................................................................ 11 Table 2.2. Test statistics of general linear mixed effects models comparing the contribution of elevation and predator type topredation rate, predator density, and predator activity rate; and the contribution of predator density, controlling for temperature and predator type, to predation rate across elevations........................................................................................................................................ 16 Table 3.1. Test statistics of generalized linear mixed models and general linear models testing the effect of elevation on various measurements of ant and treehopper abundance related to their mutualistic associations on the Andes of Ecuador................................................................................ 29 Table 4.1. Test statistics of general linear models (GLM) and non-linear models (NLS) looking at anti-predator investment of treehopper Andean communities as a function of elevation and predation rates, and their investment in ant-mutualism as opposed to maternal care as a function of the availability of mutualist ants and the contribution of mutualist ants to predation..................................................................................................................................................... 47 Table 5.1. Test statistics of Cox proportional hazards mixed-effect models (CMM) and generalized linear mixed-effect models (GLMM) testing the behavioral response of ants in relation to baits consisting of treehoppers or termites and with presence or absence of added honey........................................................................................................................................................... 60     ix    List of Figures  Figure 1.1. Sample of treehopper diversity............................................................................................. 3 Figure 1.2. Treehopper mutualism with various Hymenoptera.......................................................... 4 Figure 1.3. Treehopper maternal care..................................................................................................... 6 Figure 2.1. Map of the study area in northern Ecuador...................................................................... 12 Figure 2.2. Predation rates across the Andes elevational gradient.................................................... 15 Figure 2.3. Factors associated with predation patterns across elevation.......................................... 17 Figure 2.4. Diagram of postulated link between temperature and predation rates........................ 19 Figure 3.1. Experimental setup and sampling locations at the eastern and western slopes of the Andes of Ecuador...................................................................................................................................... 24 Figure 3.2. Abundance of treehoppers and ant workers and the proportion of individuals belonging to species known to engage in mutualism across elevations in the Andes of Ecuador...................................................................................................................................................... 30 Figure 3.3. Number of ant-treehopper aggregations across elevations in the Andes of Ecuador...................................................................................................................................................... 31 Figure 3.4. Supply and demand of ant workers and ant nests in relation to ant-treehopper mutualistic interactions............................................................................................................................. 31 Figure 3.5. Number of associated treehoppers and ant workers per aggregation and ratio of the number of associated treehoppers to ant workers per aggregation across elevations in the Andes of Ecuador.................................................................................................................................................. 32 Figure 4.1. Treehopper behavioral strategies against predation, showing associated scores of anti-predator investment in maternal care or ant-mutualism and of total anti-predator investment.... 40 Figure 4.2. Patterns across elevation of predation rates, proportion of total predation exerted by mutualist ants, and mutualist-ant recruitment rate at the western and eastern slopes of the Andes of Ecuador.................................................................................................................................................. 43 Figure 4.3. Species behavioral assemblages and anti-predator investment in treehopper communities across elevation in the Andes of Ecuador...................................................................... 44 Figure 4.4. Average anti-predator investment in treehopper communities in relation to predation rates across elevations in the Andes of Ecuador................................................................................... 45 x  Figure 4.5. Ratio of ant mutualism to maternal care investment in treehopper communities in relation to the availability of mutualist ants and the proportion of total predation exerted by the mutualist ants across elevation in the Andes of Ecuador..................................................................... 46 Fig 5.1. Setup of experiment 1 addressing predation by mutualist ants, and experiment 2 addressing mutualist recognition and value of honeydew.................................................................... 55 Fig 5.2. Patterns across elevation of predation rates by potential partner ants and non-partner predators, proportion of total predation exerted by potential partner ants, and incidence of treehoppers individuals that can potentially partner with ants at the western and eastern slopes of the Andes of Ecuador............................................................................................................................... 59 Fig 5.3. Proportion of baits predated by ants over time following discovery.................................. 61 Fig 5.4.  Ant behavior after locating different bait types in the Ecuadorian Amazon................... 62     xi  Acknowledgements  I thank Leticia for her thoughtful and generous mentoring throughout my program, and for helping me become part of a grander world of science. I thank my committee members Jill, Juli, and Wayne for their valuable and challenging insights.   I thank Jennifer Guevara, Mark Robertson, Phillipe Fernández, Sam Straus, Andrea Haberkern, Gabriel Greenberg, and all members of the Avilés lab for the wonderful opportunity of sharing the field, the lab, and the classroom with you. Special thanks to Melisa Guzmán for being a great office mate, friend, and colleague. I extend my appreciation to the marvelous community of the Biodiversity Research Centre.  I thank the Zoology department, the Biodiversity Research Centre, and the University of British Columbia for receiving me and for being a wonderful host. I extend this sentiment to Canada, Vancouver, and their people.  Special thanks to the museum of invertebrates QCAZ at the Universidad Católica del Ecuador for sharing with me their wonderful collection. I also thank Ana Ávila, Yesenia Campaña, and Washington Pruna for their support in Ecuador.  I thank the economic support of the government of Ecuador through the Secretaría Nacional de Educación Superior, Ciencia, y Tecnología (SENESCYT) and the National Geographic Society, which made this program possible.   xii  Dedication  To all the insects that sacrificed their lives involuntarily for the advancement of knowledge. 1  Chapter 1. General introduction  Biotic interactions are fundamental properties of ecosystems and are likely factors responsible for the ubiquitous gradient of increasing biodiversity at the tropics (Lawton, 1999; Brown, 2014). Biotic interactions have traditionally been studied at small scales –individual locales, lakes, and petri dishes –and while these studies have provided fundamental insights, broader ecological patterns have been difficult to identify (Lawton, 1999; Keith et al., 2012). As an alternative, the broad environmental conditions across macroecological gradients such as latitude and elevation offer the opportunity to identify general ecological patterns arising above the contingencies occurring at smaller scales (Lawton, 1999; Gaston & Blackburn, 1999; Keith et al., 2012). In fact, Dobzhansky (1950) postulated that biotic interactions should decrease in importance at higher latitudes and ecologists are still gathering evidence behind this claim. So far, studies addressing this topic have mostly focused on predation and herbivory and suggested biotic interactions are indeed less important at higher latitudes and elevations (Pennings, 2005; Schemske, 2009; Keith et al., 2012). However, other interactions such as mutualism or competition remain relatively unaddressed (Schemske, 2009). Moreover, the process driving macroecological gradient in biotic interactions as well as their consequences in communities have not been addressed (Gaston & Blackburn, 1999; Keith et al., 2012).  1.1 Macroecological gradients across elevations in the Andes  In this thesis I contribute to the field of macroecology by addressing the mechanisms and ecological outcomes behind gradients of predation and mutualism in invertebrates, across 4000-meter elevational transects at the Western and Eastern slopes of the Andes of Ecuador. The tropical Andes are an ideal and often overlooked model for macroecological studies as the range of environments it exhibits across elevations is in many ways comparable to that of an entire latitudinal gradient without the variability caused by seasonality (Sarmiento 1986; Garreaud 2009; Malhi et al. 2010). The equatorial Andes in particular –approximately between -2 and 2 degrees of latitude –have the advantage of presenting two elevational gradient replicates as the weather regimes and geological ages of the western and eastern slopes are similar and without much interference from inter-Andean valleys. In our elevational transects, ranges of mean annual temperature are similar to those occurring between 0 and 40 degrees of latitude where both slopes 2  transition from lowland rainforests (i.e. the Amazon on the east and the Chocó on the west) to cloud forests, and finally into highland paramo.  1.2 Treehoppers as a study system  This study is pertinent to all invertebrates but focuses primarily on treehoppers (Hemiptera: Membracidae) a group of sap-sucking insects occurring across the whole elevational gradient where species are commonly found engaging in mutualistic associations with ants (Wood, 1993; Godoy et al., 2006). Being primary consumers, treehopper communities are regulated primarily by predation (i.e. top-down regulated) (Holt & Lawton, 1994; Singer & Stireman, 2005; Vidal & Murphy, 2018). Moreover, sap-sucking insects are thought to be especially vulnerable to predation because of the difficulty associated to removing their stylets from plant tissue that prevents a quick escape (Tallamy & Wood, 1986). Treehopper taxonomic and trait diversity peaks in the lowlands and mountainous regions of the neotropics where approximately half of all described species and genera occur (Wood, 1993; Godoy et al., 2006). The Membracidae is a monophyletic clade including approximately 3450 described species organized in 428 genera and 9 subfamilies (Wood, 1993; Dietrich et al., 2001; Cryan et al., 2004; Lin et al., 2004; Roy et al., 2007; Deitz et al., 2010; Evangelista et al., 2016). The family originated in America where all subfamilies are represented and one subfamily (Centrotinae) dispersed to the rest of the world with the exception of New Zealand, Madagascar, and Antarctica (Wood, 1993).  Treehoppers are notorious for the great diversity and often outrageous shapes of their pronotum –the most anterior dorsal segment of the thorax. The level of modification and diversification of this structure is without parallel among insects, and the nature of the shapes some genera exhibit border with the realm of fiction (Fig 1.1) (Wood, 1993; Godoy et al., 2006). Although less recognized, treehopper behavior is also highly diverse. Species exhibit behavioral strategies which are a combination of various levels of predisposition to engage in mutualistic associations with ants (i.e. trophobiosis), degrees of investment and sophistication in maternal care (subsociality), and gregariousness (Wood, 1993; Delabie, 2001; Godoy et al., 2006). These traits create a continuum of behavioral strategies ranging from asocial to highly subsocial, where the former is represented in species occurring as solitary individuals or loose small aggregations from egg to adult and with no maternal care. The whole social spectrum of treehoppers can occur in 3  combination with ant mutualism (Wood, 1993; Godoy et al., 2006). Maternal care and ant mutualism have been acquired and lost throughout the treehopper evolutionary tree, but it remains unclear which character precedes the other (Cryan et al., 2004; Lin et al., 2004).   Figure 1.1. Sample of treehopper diversity. From left to right on first row: Alchisme ustulata, Membracis foliatafasciata, Cladonota sp., Bolbonota sp. Cladonota machinullus; second row: Notocera sp., Calloconophora sp., Acutalis sp., Todea sp., Cyphonia clavata; third row: Polyglypta sp. Cyphonia longistyla, Adippe histrio, Lycoders sp., Alobia sp.; fourth row: Heteronotus sp., Nassunia sp., Anchistrotus sp., Bocydium globulliferum, and Stegaspis fronditia.  Treehoppers are among the few clades within the Hemiptera that engage extensively in ant mutualism both in frequency and ecological strength in six out of the nine subfamilies (Delabie, 2001; Dietrich et al., 2001; Cryan et al., 2004; Cryan & Urban, 2012). Species engage in facultative mutualism with ants and, occasionally, other hymenopterans (e.g. Vespidae wasps and Meliponini bees) during the nymphal stage or adulthood as well (Delabie, 2001; Godoy etal., 2006; Oda et al., 2014) (Fig 1.2). In this interaction, ants actively protect treehoppers from predators and parasitoids 4  in exchange of treehopper honeydew, which is a byproduct of excess sugar during sap digestion (Delabie, 2001). Ant services may also include parasite removal, prevention of fungal infection, and shelter building (e.g. Azteca and Phelidole ants) (Fig 1.2A) (Delabie, 2001; Stadler et al., 2001). Ant attention has been shown to improve survival of treehopper individuals and the persistence of populations, but this likely comes at an ecological cost for treehoppers (Cushman & Whitham, 1991; Del Claro & Oliveira, 2000; Billick et al., 2007; Fagundes et al., 2013). Other ant-attended Hemiptera have shown slower individual and population growths as they relinquish nitrogen and sugars during sap digestion to make honeydew more attractive to the ants (Stadler & Dixon, 1998; Yao et al., 2000; Fischer et al., 2001; Fischer et al., 2002;  Yao & Akimoto, 2002; Kay et al., 2004; Fischer et al., 2005). Moreover, ants have also been recognized to predate on their hemipteran hosts when facing a shortage of protein or a better sugar source is present (Stadler et al., 2001; Billick, 2007; Yao, 2014).   Figure 1.2. Treehopper mutualism with various Hymenoptera. A: Endoiastus treehoppers attended by Pheidole ants, which have already begun to accumulate debris around the treehoppers to build a refuge over them. B: Vespidae wasps attending and aggregation of Membracis nymphs. C: Meliponini bees attending females of Antianthe.  Treehoppers are also especially prone within the Hemiptera to exhibit maternal care, which has independently evolved several times throughout the Membracidae evolutionary tree (Tallamy & Wood, 1986; Wood, 1993; Cryan et al., 2004; Lin, 2004; Godoy et al., 2006). In its most basic form, treehopper maternal care involves a solitary mother clumping her eggs in a single egg-mass and staying on top of it to passively protect them from the elements and predators (Fig 1.2A). 5  This is typically exhibited by species that also associate with ants where mothers can even abandon their egg masses and lay additional ones when tending ants are plentiful (Wood, 1982; Cushman & Whitham, 1991; Godoy et al. 2006). In its most complex form, treehopper maternal care extends until nymphs achieve adulthood and involves an array of behaviors to actively protect eggs and nymphs from predators as well as facilitating feeding (Godoy et al., 2006; Camacho et al., 2014). This is exhibited by species that actively reject the attention of ants. In these species, defense is coordinated by substrate-borne signaling between the mother and nymphs where the former employs offensive behaviors such as kicking, lateral movements of the body, and wing buzzing to engage parasitoids and other predators attacking eggs and nymphs (Godoy et al., 2006) (Fig 1.2B). Mother can also create feeding slits in plant tissue to facilitate nymph feeding just after hatching and herd them to feeding sites across their host plant (Godoy et al., 2006) (Fig 1.2C). Treehopper maternal care has been shown to be effective in increasing offspring survival, albeit at the cost of reduced fecundity and increased female mortality (Tallamy y Schaefer, 1997; Camacho et al., 2014; Godoy et al., 2006). While caring for their brood, females are unable to feed properly and they allocate energy in offspring care instead of in spawning the next offspring generation, which has been shown to result in fewer egg masses laid in a lifetime (Tallamy y Schaefer, 1997). Females caring for their brood are exposed to predators unable to escape when attacked and are required to engage in direct defensive encounters with them, which leads to increased risk of mortality (Fig 1.2B) (Tallamy y Schaefer, 1997).   6   Figure 1.3. Treehopper maternal care. A: A Thuris binodosus female sitting on her eggs while guarded by a Dolichoderus ant. B: Nymphs of Alchisme ustulata guarded by their mother with an injured wing, a frequent sighting in females providing brood care. C: A female Metcalfiella has already made feeding slits next to her eggs mass, upon which nymphs will feed after hatching  1.3 Overview of Chapters  In chapter 2, I aimed to confirm the decrease in predation at higher elevations in the Andes and tease the role of predator abundance and activity as potential mechanisms behind such pattern. I assessed predation across elevations as reflected on predation rates on live fly baits, and predator abundance based on visual sampling of potential predators. The identity of predators was also determined as well as their activity rates (predation rate per individual predator). I found that around eighty percent of predation events at the lower elevations were due to ants and that the decline in predation with elevation was mainly driven by a decline in the abundance of these organisms, whose importance relative to other predators also declined. I show that both predator density and activity decreased with elevation, thus ascribing specific mechanisms to known predation patterns. I suggest that changes in these two mechanisms may reflect changes in primary productivity and metabolic rate with temperature, factors of potential relevance across latitudinal and other macroecological gradients.  Along the same elevational gradients, in chapter 3, I looked for the frequency and ecological outcomes of the mutualism between ants and treehoppers across the elevations. I used sugary baits and visual surveys of treehoppers and their associations with ants to determine the local 7  availability of potential ant and treehopper partners and the frequency and size of their associations. I show lower density of ants and treehoppers as well as less treehopper species that are known to associate with the ants resulted in decreasing frequency of ant-treehopper associations at higher elevations. However, associations were limited by a shortage of treehoppers at the lowlands, and a shortage of ants above 1500 m. I suggest decreasing availability of both partners coupled with the shift in their relative availability may lead to shifts in their ecological value as resources and explain our findings of increasing number of individuals per aggregation with a bias towards treehoppers at higher elevations.  In chapter 4, I studied the extent to which anti-predator investment in treehopper Andean communities is linked to predation rates and the activity of ants either as predators or mutualist partners across elevations. Using museum collections of treehoppers and information on their natural history I determined the community composition and their degree of anti-predator investment in maternal care and ant mutualism across elevations. I show treehopper anti-predator investment decreases only above 2500 m despite the steep decrease in predation rates. This pattern may arise as the negative effects of temperature and productivity on predators and prey cancel out up to the highest elevations where primary productivity can no longer sustain a predator trophic level. The drop of ants as the main predators at higher elevations, however, where associated with a shift from ant-mutualism to maternal care as the predominant anti-predator strategy, revealing that qualitative shifts in predation can have important consequences to prey populations.  Finally, in chapter 5, I explore the possibility that as a consequence of the pervasive ant predation in the lowlands, hemipterans may ‘bribe’ ants with honeydew primarily to dissuade them from predation, rather than, as generally assumed, obtain their defensive services. I used an experimental set up were termite and treehopper baits with and without added honey to determine the behavior of ants in presence of a sugary reward and a potential mutualistic partner. I present evidence that the most likely predators of sap-sucking hemipterans are their ant-mutualist themselves, showing preventing predation from ants has greater ecological value than obtaining their defensive services against other predators. I further show that, mutualist ants refrained from attacking and provided attention to their hemipteran partners, even in the absence of an immediate sugary reward. This suggests that the ants recognize their hemipteran partners, either through an evolved or learned response. By ‘extorting’ hemipterans into delivering honeydew in exchange for 8  their lives, ants may thus have become the hypothesized ‘prudent predator,’ which exploits their victims with the lowest possible impact. This is the first conceptualization of a mutualistic interaction involving prey offering a reward to predators in exchange for their lives, challenging our fundamental understanding of both mutualism and predation.    9  Chapter 2. Decreasing predator density and activity explain declining predation rates along elevational gradients  2.1 Introduction  Predation is a fundamental biotic interaction as it regulates the flux of energy and materials between trophic levels and also shapes the life histories and behavior of prey via top-down forcing (Jeffries & Lawton, 1984; Lima & Dill, 1990; Holt & Lawton, 1994). A variety of studies have addressed macroecological patterns of predation on terrestrial and marine invertebrates, bird nests, and insects by their parasitoids, finding evidence of higher predation rates at lower latitudes and elevations (Table 2.1). The mechanisms through which these patterns arise, however, are poorly understood. Potential mechanisms include changes in the composition of predator communities, their absolute and relative abundances, and the activity rate of different predator types (Solomon, 1949; Holling, 1959; but see Sam et al., 2015). Greater predator density would increase predation rate by increasing the encounter rate between predators and prey, whereas more active predators would have shorter search times and increased attack effectiveness (Solomon, 1949; Holling, 1959; Abrams, 1992). The composition of the communities may also change as invertebrate predators appear to be more important at low latitudes and elevations, whereas vertebrates may be relatively more important at higher latitudes and elevations (Tvardikova & Novotny, 2012; Sam et al., 2015; Roslin et al., 2017). The environmental factors likely to mediate changes in the composition, abundance, and activity patterns of predators are temperature and primary productivity, which are known to co-vary with elevation and, to some extent, with latitude (Korner, 2007; Brown, 2014; Gillman et al., 2015). Primary productivity, which depends on temperature and precipitation (Lieth, 1975; Grosso et al., 2008), may have an overall positive effect on predator density (Oksanen et al., 1981; Nisbet et al., 1997; Savage et al., 2004; O’Connor, 2009; O’Connor et al., 2009; Brown, 2014). Temperature, on the other hand, has a positive effect on activity rates of ectotherms (Addo-Bediako et al., 2002; Brown et al., 2004; O’Connor, 2009; Dell et al., 2011; Buckley et al., 2012; Dell et al., 2014), but a more limited effect on endotherms, which can self-regulate their body temperatures and thus maintain activity rates (Anderson & Jetz, 2005; Boyles et al., 2011; Buckley et al., 2012). As a result, predation rates by ectotherms may decrease with elevation and latitude, whereas endotherm predators would be less affected.  Teasing apart the relative importance of the various mechanisms behind macroecological patterns of predation 10  would allow us to assess their generality and better understand their implications for biodiversity (Lawton, 1999; Schemske et al., 2009).  Here we address macroecological patterns of predation on insect prey by taking advantage of elevational gradients in the equatorial Andes. The tropical Andes are ideal models for macroecological studies as the range of environmental variation they exhibit, from lowland rainforests to mountain-top glaciers, is in many ways comparable to that of an entire latitudinal gradient. Such broad range, however, is often contained in less than a 100 Km (Appendix A, Table S1.1) and involves conditions that are relatively constant throughout the year (Sarmiento, 1986; Garreaud, 2009). This reduces the potential confounding effects of local historical factors and seasonality. Despite their great potential, however, macroecological studies on the tropical Andes are scarce (Malhi et al., 2010).  We used live baits and visual sampling of potential predators to assess patterns of predation on arthropods along 4000-meter transects on either side of the Andes. We sought to determine the contribution of predator density and activity (predation rate per individual predator) to the observed decrease in predation rates at higher elevations. We also sought to determine whether the observed patterns are caused by declines in predation of a single predator type or of a variety of predators. We do so by taking advantage of the equatorial Andes and methods that allow us to assess the identity of predators and their predation rates over a broad gradient of elevation and temperature. By identifying potential mechanisms behind the observed patterns, we provide a basis for the generalization of our results to other macroecological gradients (Louthan et al., 2015; Schemske et al., 2009). 11  Table 2.1. Studies addressing macroecological patterns of predation across elevation or latitude.  Study Gradient Range Region N Effect Methods Measurement type Terrestrial invertebrates Jeanne 1979 Latitude -2 - 42° America 5 - Live baits Ant predation rates Purcell & Avilés 2008 Elevation 400 - 2100 m Tropical Andes 4 - Live baits Survival rates Tvardikova & Novotny 2012 Elevation 120 -1700 m Papua New Guinea 4 - Clay baits Attack rates Sam et al. 2015 Elevation 200 - 3700 m Papua New Guinea 8 - Clay baits Attack rates Hoffman & Avilés 2017 Elevation 400 - 2100 m Tropical Andes 4 - Predator exclusion Survival rates Lovei & Ferrante 2017 Latitude 5 - 60° Worldwide  0 Literature review Attack rates Roslin et al. 2017 Latitude, elevation -30 - 74°, 0 - 2110 m Worldwide 32 -, - Clay baits Attack rates Insect parasitism Stireman et al. 2005 Latitude 0 - 45° America 15 0 Literature review Parasitism incidence Moya-Raygoza et al. 2012 Elevation 870 - 1650 m America 4 - Egg baiting Parasitism incidence Maunsell et al. 2015 Elevation 500 - 1100 m Australia 12 - Wild lepidoptera larvae survey Parasitism incidence Morris et al. 2015 Elevation 300 - 1100 m Australia 5 - Wild wasp nest baits Parasitism incidence Bird nest Soderstrom 1999 Latitude 0 - 53° America  0 Literature review Predation rate Boyle 2008 Elevation 30 - 2910 m Costa Rica 8 - Artificial nest baits Predation rate Remes et al. 2012 Latitude - 10 - -30° Australia  - Literature review Predation and survival rates Marine invertebrates Bertness et al. 1981 Latitude 9 - 42° America 2 - Natural baits Predation rate Freestone et al. 2011 Latitude 9 - 41° America 4 - Predator exclusion Prey abundance and composition Any system        Sih et al. 1985 Latitude  Worldwide  0 Literature review  Schemske et al. 2009 Latitude  Worldwide  - Literature review   12  2.2 Methods  2.2.1 Study area We sampled transects encompassing ~4000-m elevation gradients on the western and eastern slopes of the Andes in northern Ecuador (Fig. 2.1). These gradients encompass ranges in mean annual temperature between ~24 and ~6 °C (Appendix B, Table S1), which are similar to those occurring between ~0° and ~±40° of latitude (Hijimans et al., 2005). The two sides of the show similar climate regimes and transition from lowland rainforest (i.e. Amazonia on the east and Chocó on the west) to cloud forest, and finally into highland paramo.Andes.    Figure 2.1. Map of the study area in northern Ecuador. Both the eastern and western slopes of the Andes were sampled from 0 to 4000 m.a.s.l. at 500 m intervals, three sites per interval for a total of 48 sites (24 per hillside).  2.2.2 Data collection We collected data on three sites at each of 500-meter intervals from 0 to 4000 m.a.s.l. (n = 48 sites, 24 sites at each side of the Andes) (Fig. 2.1, Appendix A, Table S1.1). At each site, we set two transects separated by at least 100 m. In order to equalize the ecological context of transects across sites, all transects were set up on rain-less days between 8 am and 1 pm along forest edge next to secondary roads, farmland, or trails. We collected data during May and June 2017. The order in which sites were visited was scrambled within and across the east and west sides of the Andes, to prevent any possible, although minimal, effects of seasonality in the observed patterns.   13  We assessed predation rate by setting up live baits of lab-bred adults of the fly Lucilia sericata (Diptera: Calliphoridae). These baits were intended to represent actual live prey while minimizing intrinsic anti-predator adaptations such as escape behavior and physical or chemical defenses. L. sericata flies, which are soft-bodied organisms lacking chemical defenses (Pasteels & Gregoire, 1983; Witz, 1990), were affixed to vegetation by their back using sticky insect-trap paper (Appendix A, Fig S1).  The flies were thus prevented from escaping, but their body movements would have provided visual clues to predators searching for prey. We placed the baits on stems and leaf petioles of overhanging vegetation (Figure A1), being careful not to contaminate them with sweat, which would have affected their luring effect (Kaspari et al., 2008). Each transect consisted of 15 baits total separated by at least 5 m from one another.   We monitored the baits at 30-minute intervals for three hours, recording at this time whether they had disappeared or were being attacked. Bearing in mind that observer presence could interfere with potential predators, we limited our presence near the baits to brief moments while checking if baits had been predated upon. We calculated the proportion per hour of baits missing or being attacked, and averaged these proportions across the three hours sampled per transect. This represents the rate of predation per hour per transect. When predators were observed consuming baits, they were identified to family, and later classified as ant or non-ant. We assumed missing baits as consumed by non-ant predators because ant predation is a slow swarming process that is unlikely to be missed. We calculated the ant, non-ant, and total rate of predation per transect and averaged across the two transects at a site. Each data point per site thus encompassed 90 bait-hours, allowing us to detect predation rates of 0.01 per hour.  We assessed the density of predators by conducting 30-second visual samplings of potential predators found within a ~50-cm radius of each bait. We considered vertebrates such as small birds, lizards, and frogs, and predatory arthropods including ants, wasps, bugs, various beetles, spiders, and others. As above, we identified the predators to family level and then classified them as either ant or non-ant. Scans for predator presence were made every 30 minutes for 3 hrs at a time.  The counts of ant or non-ant predators were then averaged per bait, across baits, and across the two transects per site. These numbers represent average predator density per ~0.5 m3 per 30 s.   14  We also assessed predator activity rate—the predation rate exerted by individual predators—for ant and non-ant predators.  We obtained this estimate by dividing total predation rate at each transect by predator density for ant and non-ant predators and averaging across the two transects per site.  2.2.3 Data analysis Using sites as the units of analyses, we first tested for differences on predation rate as a function of elevation between the eastern and western slopes of the Andes using a generalized linear model (GLM) with binomial error structure and logit link function. Elevation, hillside (i.e. western and eastern slopes), and their interaction were included as fixed effects and the overall rate of predation was the response variable. Using the same model construction, we ran an additional GLM, which tested, instead of elevation, the effect of mean annual temperature (Hijimans et al., 2005), which correlates tightly with elevation (5.5°K for every 1000 m; Korner, 2007).  We used a series of general mixed effect models (GLMM) to compare the contributions of ant and non-ant predators to the predation gradient and test whether these were associated with their densities and efficiencies. These models included elevation, predator type (i.e. ant and non-ant), and their interaction as fixed effects, and individual sites, nested within the western and eastern slopes, as random effects. The response variables (average of the two transects per site) were predation rate (Model A), predator density (Model B), and predator activity (Model C). We assessed the effect of predator density on predation rates in a fourth model that included as fixed effects temperature, predator type, and their interactions (Model D). We applied an arcsine transformation to predation rate data and a log10+1 transformation to density and activity rate data prior to analyses.  All statistical analyses were carried out in R v3.2.2 software (R core team, 2019) using packages lme4 (Bates et al., 2015) for model construction and car (Fox & Weisberg, 2011) to calculate P-values based on type III sum of squares. Data deposited in the Dryad Digital Repository: https://doi.org/10.5061/dryad.5k586p0 (Camacho & Avilés, 2019).    15  2.3 Results  Overall predation rate on our insect baits decreased with increasing elevation or decreasing temperature, being practically nil above 3000 m (Fig. 2.2A), with no difference between eastern and western slopes and no interaction between hillside and elevation or temperature (for elevation: GLM, slope per 1000 m (log-odds) = -2.927 ± 1.608, chi-squared = 9.6, P = 0.002*; for temperature: GLM, slope (log-odds) = 0.676 ± 0.379, chi-squared = 9.7, P = 0.002*).    Figure 2.2. Predation rates across the Andes elevational gradient. (A) Probability of predation per hour across elevations and mean annual temperatures (n=48 sites). (B) Percentage of total predation exerted by ants in sites where predation was detected (n=27); bars represent average predation exerted by ants and other predators within an elevation range. Circles and solid line represent site data from the eastern and triangles and dashed line from the western slope.  There was a significant effect of predator type, elevation, and their interaction on predation rates across the elevation gradient (Table 2.2, Model A). Ants exerted much higher predation rates than other predators, but their contribution as elevation increased decreased at a higher rate, essentially disappearing above 1500 m.a.s.l. (Fig. 2.2B).  Among baits consumed by non-ant predators, we observed the predation event taking place in 31.9% of occasions, which included Salticidae spiders, Vespidae wasps, Pentatomidae and Reduviidae bugs, Asilidae flies, and Staphylinidae beetles, and birds. Among all predated baits, ants were responsible for 80% of predation events,  16  other arthropods 6%, birds 0.5%, and 13% were unknown predators. The vast majority of predators (86%) were, thus, arthropods.   Table 2.2. Test statistics of general linear mixed effects models comparing the contribution of elevation and predator type to (A) predation rate, (B) predator density, and (C) predator activity rate; and (D) the contribution of predator density, controlling for temperature and predator type, to predation rate across elevations. In all models, sites are nested within the eastern and western side of the Andes as random effects. Terms with significant P-values marked with an asterisk; (a) slope per 1000-meter change; (b) difference in mean; and (c) slope.  Response variable Model factor Estimate Std. error χ2 P Model A (n=96):      Predation rates Elevation (E) -0.223a 0.021 41.3 ˂0.01* (arcsine transformed) Predator type (PT) 0.442b 0.068 114.3 ˂0.01*  E * PT -0.153a 0.029 27.1 ˂0.01* Model B (n=96):      Predator density Elevation (E) -0.219a 0.012 304.8 ˂0.01* (log10+1 transformed) Predator type (PT) 0.417b 0.037 124.7 ˂0.01*  E * PT 0.147a 0.016 84.02 ˂0.01* Model C (n=71):      Predator activity rate Elevation (E) -0.022a 0.006 11.8 ˂0.01* (log10+1 transformed) Predator type (PT) 0.010b 0.013 0.6 0.4  E * PT 0.008a 0.008 1.1 0.3 Model D (n=96):      Predation rates Temperature (T) 0.004c 0.006 0.5 0.5 (arcsine transformed) Predator type (PT) -0.008b 0.109 0.005 0.9  Predator density (PD) -0.485c 0.61 0.6 0.4  T * PT 0.005c 0.009 0.3 0.6  T * PD 0.061c 0.025 5.8 0.02*  PT * PD -0.741c 0.416 3.1 0.07  Predator density had a positive association with predation rates (Table 2.2, Model D) (Fig. 2.3A), but did so to a lesser extent at lower temperatures (Table 2.2, Model D) (Fig. 2.3A). The density of ants was generally higher than that of other predators, but whereas the density of both ant and non-ant predators decreased with elevation, it did more so for ants (Table 2.2, Model B) (Fig. 2.3B). Predator activity rates, on the other hand, decreased similarly with elevation for both predator types (Table 2.2, Model C) (Fig. 2.3B).      17     Figure 2.3. Factors associated with predation patterns across elevation. (A) Predation rate as a function of predator density across elevation in the Andes; points colored according to local annual mean temperatures of 0°-10° (blue), 10°-20° (yellow), and 20°-30° (red); lines represent predicted effect of density at representative temperatures of 10°, 20°, and 30° based on results of a GLMM with significant interaction term (n = 96 data points, one for ants and one for non-ants at each of 48 sites). (B) Predator density and activity rate across the Andes elevational gradient; lines represent the best fit of GLM (n = 96 density estimates, one for ants and one for non-ants at each of 48 sites; and n = 71 activity rates estimates at the 40 sites where either ant or non-ant individuals were detected). All models included sites nested within the eastern and western slopes as random effects; circles represent data from the East and triangles from the West. Models used arcsine transformed rate of predation per hour; predator density as the log10+1 transform number of predators per m3 per 30s; and predator activity rate as the log10+1 transformed of the predation rate divided by predator density.   2.4 Discussion  By examining predation rates on live insect baits along 4,000 m elevation gradients on both sides of the equatorial Andes, we confirm a sharp decline in predation rates with elevation and show that the vast majority of predators, 86%, were arthropods, mainly ants. The decline in predation rates was mostly due to a decrease in the abundance of predatory ants, which essentially disappeared above 1500 m.a.s.l. (Fig. 2.2B). Predation rate of non-ant predators also decreased with elevation, but at a lower rate than that of ants, resulting in most predation being by non-ant predators at the higher elevations. Both predator density and predator activity for both ant and non-ant predators decreased with elevation (Fig. 2.3B). This was reflected in predation rates increasing with predator density, but at a lower rate at the higher elevations where temperatures were lower (Fig. 2.3A).  18   Here we have shown that a drastic shift in the abundance of ants may be the main factor shaping predation rates along the Andes elevational gradient. The negative correlation between ant abundance and elevation has been relatively well documented in a variety of earlier studies (Samson, 1997; Bruhl et al., 1999; Kaspari et al., 2000; Longino & Colwell, 2011; Longino et al., 2014; Gillete et al., 2015). Decreasing ant predation with elevation has also been documented in a handful of other studies (Purcell & Avilés, 2008; Tvardikova & Novotny, 2012; Sam et al., 2015; Hoffman & Avilés, 2017). Our study, however, is the first to establish a link between ant abundance and their predation patterns across elevations, while teasing apart the separate contributions of abundance and activity by ants and other predators on predation on insect prey along temperature gradients. In terms of predation levels, our study, as well as others also relying on live prey (Olson, 1992; Novotny et al., 1999; Purcell & Avilés, 2008; Hoffman & Avilés, 2017), found overall higher predation rates than studies relying on clay baits in similar environments (Tvardikova & Novotny, 2012; Sam et al., 2015; Lovei & Ferrante, 2017; Roslin et al., 2017). These differences may be attributed to clay baits experiencing proportionately less predation than live baits (Lovei & Ferrante, 2017). Nevertheless, studies relying on either clay or live baits agree on the general pattern of decreasing predation rate with increasing elevation. The benefits and drawbacks of using clay vs. live baits have been recently addressed by Low et al. (2014) and Lovei & Ferrante (2017).  The observed decrease in predation rates with elevation likely reflects the decline in temperature with elevation (5.5°K for every 1000 m; Korner, 2007), which, by affecting primary productivity and metabolic rate, should influence the density and activity rates of predatory arthropods. It is well known that temperature has a positive effect on primary productivity (Brown, 2014), which, in turn, has been shown to have positive effects on consumer density across trophic levels (Oksanen et al., 1981; Nisbet et al., 1997; Savage et al., 2004; O’Connor, 2009; O’Connor et al., 2009; Brown, 2014) (Fig. 4). Temperature also increases whole body metabolic rate of ectotherms (Addo-Bediako et al., 2002; Brown et al., 2004; O’Connor, 2009; Dell et al., 2011; Buckley et al., 2012; Dell et al., 2014), which should increase the activity level of arthropod predators (Fig. 2.4). Changes in temperature (productivity and metabolic rate) may also explain the parallel decline in predation rates associated with latitude (see Table 1 for examples), as elevation and latitude are ultimately gradients of temperature. Common mechanisms may thus underlie patterns of  19  predation along macroecological gradients, at least for ectotherm predators. The situation may differ for endothermic predators (birds and mammals), who can maintain high levels of activity in cold environments if sufficient energy is available to meet their elevated metabolic needs. Although their abundance may still be affected by lower environmental productivity in colder habitats (Buckley et al., 2012), in the balance the relative incidence of endotherm predators may increase with elevation and latitude as they may be less affected overall by temperature than their ectotherm counterparts (e.g., Tvardikova & Novotny, 2012; Sam et al., 2015; Roslin et al., 2017).    Figure 2.4. Diagram of postulated link between temperature and predation rates. Temperature would affect predator activity through its effect on metabolic rate (Brown 2004; Dell et al. 2011; Dell et al. 2014) and, by acting on primary productivity (O’Connor 2009), would also affect predator density (O’Connor et al. 2009), both of which would ultimately affect predation rates.  A separate question concerns the effect of predation, and of predation gradients, in particular, on prey communities. There is evidence, for instance, that species exhibit stronger anti-predator adaptations at lower latitudes (Schemske et al., 2009; Santos et al., 2017) and have their ranges limited by predators at low elevations and latitudes (Hodkinson, 2005). The types of antipredator adaptation prey develop may also reflect changes in the composition of predator communities. Higher ant predation in lowland tropical rainforests, for instance, has been suggested to promote more motile invertebrate communities compared to temperate forests (Floren et al., 2002). Predation by ants may also be one of the factors favoring group living in certain groups of organisms, as in certain spider genera in lowland tropical habitats (e.g., Henschel, 1988; Guevara & Avilés, 2015; Hoffman & Avilés, 2017).   20   There may not be, however, a 1:1 relationship between predation rates and actual ecological impact on prey (Louthan et al., 2015; Schemske et al., 2009), as some prey characteristics may also change along gradients. Ectotherm species, for instance, tend to exhibit lower fecundity and longer developmental rates at colder temperatures (Brown et al., 2004; Savage et al., 2004; Frazier et al., 2006), which may result in a given level of predation having a greater impact on their populations. Indeed, metabolic theory of ecology predicts both prey intrinsic population growth (rmax) and predation rates should increase with temperature at the same rate, suggesting the two factors may cancel each other out (Brown et al., 2004; Savage et al., 2004). Evidence shows, however, that the intrinsic population growth (rmax) of insect species at their physiological optimum decreases at a rate of 8% to 12% for every 1°C drop in temperature (Frazier et al., 2006), which is proportionally smaller than our results showing a decrease in predation of 33% per 1°C temperature drop. This suggests that with increasing elevations the rate at which prey replenish outpaces the rate at which they are being consumed. Additionally, food web length, which decreases with decreasing ecosystem productivity, can determine the relative contribution of top-down and bottom-up control, which would need to be considered when assessing fitness impacts of predation on prey (Fretwell, 1977; Oksanen et al., 1981; Fretwell, 1987; Power, 1992).  Decreasing ant abundance along macroecological gradients should also release niches for non-ant predators to exploit and reduce predation on non-ant predators themselves, which would explain the preponderance of other types of predators at higher elevations in ours and other studies (Tvardikova & Novotny, 2012; Sam et al., 2015; Rodríguez-Castañeda et al., 2016; Roslin et al., 2017). Nevertheless, the trophic status of ants, which may change across ecological gradients, may determine their impact on prey and other predators, and would need to be considered when assessing their ecological relevance (Fretwell, 1977; Oksanen et al., 1981; Fretwell, 1987; Power, 1992). Predator exclusion experiments may prove useful to test the relative importance of different predator groups (Mooney, 2007; Piñol et al., 2010; Hoffman & Avilés, 2017). Finally, as the ecological relevance of ants is not limited to predation (Del Toro et al., 2012), understanding their activity patterns along macroecological gradients should provide major insights into ecosystem functioning and community structure.   21  2.4.1 Conclusion We show that the decrease in insect predation rate with elevation is mainly driven by a decline in predation by ants closely associated with a decrease in their density and activity. The finding that ants represent 80% of all predation at the lower elevations suggests that these organisms likely play a significant role in shaping prey communities in the lowland tropics. The association of predation rates with predator density and activity point to temperature, acting through its effect on primary productivity and metabolic rate, as a primary driver of macroecological patterns of predation. Finally, predation rate measurements may not fully reflect the ecological impact of predation in ecosystems as this depends on prey life histories and on the relative impact of bottom-up and top-down control, all of which may also vary along gradients. Future studies should thus aim at assessing the ecological impact of predation in the context of prey life histories and other biotic and abiotic interactions to which prey may be subject.   22  Chapter 3. Partner supply and demand mediate macroecological gradients in the outcome of ant-hemipteran mutualistic associations  3.1 Introduction  One of the key challenges of macroecology is to understand the mechanisms shaping broad patterns of biodiversity and ecosystem function (Keith et al., 2012; Brown, 2014). Such patterns have long been recognized to be partly driven by gradients in biotic interactions, with research on the topic still ongoing (Dobzansky, 1950; Lawton, 1999; Schemske et al, 2009; Brown, 2014). Several studies have provided important insights on the geography of predation and herbivory, but biotic interactions such as mutualism have been greatly understudied (Schemske et al, 2009; Moles et al., 2011; Roslyn et al., 2017; Camacho & Avilés, 2019; Hargreaves et al., 2019). Mutualism is a pervasive natural phenomenon that plays a major role in ecosystem processes such as pollination, seed dispersal, predation, energy/material cycling, among others (Bronstein, 1994; Stachowicz, 2001; Bastolla et al., 2009; Bascompte, 2019). Studying mutualistic interactions across macroecological gradients can not only provide a better understanding of mutualisms, per se, but also of the factors shaping biodiversity and ecosystem function (Bronstein, 1994; Hoeksema & Bruna, 2000; Schemske et al, 2009).  Along macroecological gradients, the occurrence and ecological outcomes of mutualistic interactions may be driven by shifts in partner abundance, in turn a result of changes in environmental factors (Bronstein, 1994; Holland et al., 2002; Stanton, 2003; Holland & Deangelis, 2010). For instance, decreasing temperature and productivity with increasing latitude and elevation may drive a decline in the abundance of one or both mutualistic partners (Berggren et al., 2009; Brown, 2014). Lower partner availability should limit the frequency of the interactions, with the disappearance of either of them limiting the geographical range of the interaction (Geib & Galen, 2012; Plowman et al., 2017; Chomicki & Renner, 2017).  The relative abundance of the partners may also set the economic parameters mediating the interaction. These may be understood in the context of consumer-resource dynamics, where both partners act as consumers and resources of each other, with the outcome of their interaction mediated by their relative supply and demand (Jhonston & Bshary, 2008; Holland & Deangelis,  23  2010; Schoner et al., 2013). For instance, declining abundance should make a partner less predictable in space, which, in turn, may require greater investment by their other partner in order to retain them (Schoner et al., 2013). Asymmetries in the abundance of partners may, thus, become a mechanism of control, as the partner in lowest supply is more difficult to replace (Noe & Hammerstein, 1994; Jhonston & Bshary, 2008; Hoeksema & Bruna, 2015; Silknetter et al., 2018). Here, the least abundant partner gains an ecological advantage and is able to derive a benefit from the interaction while incurring in lower cost to provide a service in return.  Taking advantage of the broad elevational range of the equatorial Andes, here we ask how changes in the relative abundance of ants and sap-sucking hemipterans, specifically treehoppers (Hemiptera: Membracidae) (Fig 3.1A), influence the presence and pattern of their mutualistic associations. Many treehopper species offer honeydew to the ants in exchange of the ants’ protection from predators as well as variety of other services. Many treehoppers species form aggregations that are attended by ants, which protect treehoppers from predators as well as offering a variety of other services (Delabie, 2001; Stadler & Dixon, 2005). In exchange, treehoppers offer a honeydew reward, which, while being a waste product, is enriched with nutrients excreted during digestion to increase its attractiveness (Bluthgen et al., 2000; Delabie, 2001; Yao & Akimoto, 2002; Bluthgen et al., 2006; Stadler & Dixon, 2005). In the mutualistic interaction, ant colonies are the fitness units and workers the service providers, whereas treehoppers are both fitness units and service providers. This prompts ants and treehoppers to interact at two scales, one within individual aggregations, mediated by the abundance of ant workers, another across aggregations, mediated by the occurrence of ant colonies (Cushman & Whitham, 1991; Del-Claro & Oliveira, 1996; Del-Claro & Oliveira, 2000; Morales, 2000; Gove & Rico-Gray,  2006). Here, ant colonies and treehoppers incur a higher ecological investment by increasing aggregation size as ants commit more workers to attend treehoppers and treehoppers  in larger aggregations compete for ant-worker attention as well as feeding sites (Cushman & Whitham, 1991; Morales, 2000; Holland et al., 2002; Reithel & Campbell, 2008; Palmer & Brody, 2013). Thus, ants and treehoppers should aim to minimize their own numbers while maximizing their partners’, when ecologically possible.   24   Fig 3.1. (A) Mutualistic association between Componotus ants and Dioclophara treehoppers in the eastern lowlands of Ecuador. (B) Experimental setup and sampling locations at the eastern and western slopes of the Andes of Ecuador  In the lowlands, ants comprise more than 50% of all arthropods and are commonly found attending treehopper aggregations (Davidson & Patrell-Kim, 1996). Ants, however, rapidly decline in abundance with elevation, virtually disappearing above 2000 m.a.s.l. (Kaspari et al., 2000;  25  Sanders et al., 2007; Dunn et al., 2009; Plowman et al., 2017; Camacho & Aviles, 2019). Whether such a decline is matched by treehoppers is unknown, although treehopper communities can be seen well above 2000 m.a.s.l. (Olmstead & Wood, 1990; Wood, 1993; Godoy, 2006). It does appear, therefore, that the relative abundance of ants and treehoppers is neither symmetrical nor constant across elevation, potentially exhibiting reversals in partner supply and demand.  Here, we tested the hypothesis that patterns in partner abundance, supply, and demand across elevations mediate the ecological investment ants and treehoppers while associating. To do this, we first measured the abundance of both ant and treehopper partners along a 4000-meter elevation gradient on either side of the Andes. We took into account that not all species of ants and treehoppers may engage in the mutualism. We then determined the frequency of their associations and the size of the treehopper aggregations and of the ant worker force involved in the associations. If partners incur a higher ecological cost in response to a shortage of partners at a given elevation, we expect that the size of the treehopper aggregations, or the number of ant workers interacting with treehoppers, will increase.  Likewise, if the partner in lowest supply gains an ecological advantage and incurs in lower ecological cost relative to the other partner, we expect that aggregations will consist of proportionally more individuals of the partner at a disadvantage.  3.2 Methods  3.2.1 Study location We sampled across two 4000-meter elevational gradients on the western and eastern slopes of the Andes in northern Ecuador (Fig 3.1B). Both slopes can be considered as biogeographically independent replicates of similar ecological conditions. Ecosystems ranged from lowland rainforest at the lowest elevations, transitioning into cloud forests above 1500 m, and finally into highland paramo above 3000 m. Along the elevational gradient temperatures range from ~24 and ~6 °C with minimal seasonality (Appendix B, Table S1).  3.2.2 Data collection We collected data from three sites located at each of 500-meter intervals from 0 to 4000 m.a.s.l. (n = 48 sites, 24 sites on each slope of the Andes) (Fig 3.1B). At each site, we set two ~70 m length transects separated by at least 100 m. To equalize their ecological context, all transects were  26  set up along the edge outside the forest next to secondary roads, farmland, or trails. We collected data on rainless days between 8am and 1pm. The order in which sites were sampled was scrambled during May and June 2017 to prevent any seasonality effects influencing observed geographical patterns.  At each transect, we estimated the composition and abundance of ant species at sugar baits. Transects consisted of 15 baits separated by at least 5 m from one another. Each bait was an Eppendorf tube with cotton soaked in a 0.1 g/ml sucrose solution (as in Fowler et al., 2014). Baits were tied to apical stems of overhanging vegetation where treehoppers usually dwell, being careful not to contaminate the baits with sweat as this could affect their luring effect (Kaspari et al., 2008). We monitored the baits at 30-minute intervals for three hours and recorded whether ants were consuming the sugar, the number of ants doing so, and collected specimens for identification. Here, we estimated the abundance of colonies of each species in each transect by counting the number of tree crown at which ants occurred in baits. This approach is similar to that used in other ant studies, wherein occurrence at a bait or pitfall trap or leaf-litter sample is used as an estimate of abundance (Delabie et al., 2000; Fowler et al., 2014).  Since only one ant species ever occurred in a treehopper aggregation, we assumed that an occurrence represent the presence of an ant colony (Plowman et al., 2017), although we are aware that one colony could contain multiple nests and occur on multiple trees, in which case we may be overestimating the number of colony present along a transect.  During a 3-hour monitoring period, we estimated the composition and abundance of treehopper species occurring along each transect. When prodded with a pencil, treehoppers and their attending ants did not react beyond a 25 cm radius. We, thus, used this distance from the center of a group to define clusters of treehoppers as discrete aggregations. We recorded the number of treehopper individuals in the aggregation, as well as the identity of treehopper species and their ant associates. We assumed that all ant workers of the same species in a treehopper aggregation belonged to the same colony (Bluthgen et al., 2000; Bluthgen et al, 2006; Dejean et al, 2007).  We assessed whether the species of ants attracted to sugar baits were mutualist partners of treehoppers based on separate observations of ant-treehopper associations in the field.  Species not separately observed in association with treehoppers were considered partners if congeneric  27  species had been deemed so in the literature or through our own observations (Appendix B, Table S2) (Godoy et al., 2006; Brandao et al., 2012). We used a similar approach to assess which treehopper species associate with ants (Appendix B, Table S3) (Godoy et al., 2006).  Although wasps may also interact with treehoppers (Wood, 1993; Godoy et al., 2006), we did not consider wasps that visited baits in the analysis, as we could not determine whether these wasps were actual partners. Moreover, wasps occurred at only a small fraction of the baits (9 out of 480 baits, 1.9%) and are therefore negligible as interaction partners compared to the ants.   3.2.3 Data analysis We assessed the decline in the overall abundance of ants and treehoppers across elevation using three linear mixed-effect models (GLMM) testing the effects of elevation on each the overall abundance of ant colonies, workers, and treehoppers (Table 3.1 models A-C). Based on the overall abundance of ants and treehoppers and the proportion corresponding to mutualist species, we calculated the abundance of actual partners across elevations and the effect of partner abundance in the frequency of associations at each transect. We used a GLMM testing the effect of elevation on the number of ant-treehopper aggregations (Table 3.1 models D), and another GLMM testing the relation between the number of ant colonies, ant workers, and treehopper partners on the number of aggregation (Table 3.1 models E).  We assessed the relative abundance of partners along the gradient using GLMMs testing the effect of elevation on the supply and demand of ant workers and colonies per transect. Using the population of mutualist ants and treehoppers in each transect, we calculated the potential local supply of ants as the ratio between the total number of ant workers and treehoppers. This could be interpreted as the number of ant workers potentially available per treehopper individual in the environment. We also calculated the potential local demand for ants as the ratio between the number ant workers and treehoppers found in association. This could be interpreted as the number of ant workers required per treehopper individual in associations. We applied the same calculations for the supply and demand for ant colonies. We avoided abundances of zero in our ratio calculations by replacing these with abundances predicted by elevation based on GLMMs. Ratios were log10-transformed and used as response variables in two GLMMs comparing the supply and demand for ants across elevations, one focusing in ant workers and the other in ant colonies (Table 3.1 models F, G).  28   Finally, we assessed the dynamics occurring within aggregations of ant-treehopper associations by their assessing the size and partner composition across elevations. For each ant-treehopper aggregation (n=113), we used a GLMM testing the number of attending ant-workers as a function of the number of aggregated treehoppers and elevation (Table 3.1 model H), and another GLMM comparing the number of associated ant workers and treehoppers in each aggregation across elevations (Table 3.1 model I).  All GLMMs used locations (n=48) nested within Andean slope (n=2) as random effects, except for model L that also nested individual ant-treehopper aggregations within locations as a random effect. GLMMs with number of treehoppers, ants, or aggregations as response variables used poisson or negative binomial error distributions with a log link. In GLMMs where the response count data included a large proportion of zeros, a zero-inflation factor was included in the model. GLMMs with proportion data as response variable used a binomial distribution with a logit link. Models including interaction terms used sum of squares type 3 instead of type 2.   All statistical analyses were carried out in R v3.2.2 software (R Core Team, 2019). We used packages glmmTMB (Brooks et al., 2017) for GLMM construction, car (Fox & Weisberg, 2019) for P-value calculations; and ggplot2 (Wickham, 2016), ggpubr (Kassambara, 2019), and lemon (McKinnon, 2019) for figure design.   3.3 Results  The abundance of ants and treehoppers decreased with elevation, with ants being virtually absent above 2000 m, whereas treehoppers continued to occur up to 4000 m (Fig 3.2; Table 3.1 model A-C). The majority of ants recorded across elevations belonged to genera known to associate with treehoppers, whereas treehoppers at higher elevations belonged to genera known not to associate with ants (Fig 3.2; see Appendix B Tables S2 and S3 for details). These patterns resulted in a decline in the abundance of ant-treehopper associations with elevation, which disappeared above 2100 m (Fig 3.3; Table 3.1 model D). Such decline was a function of the local number of treehoppers and ant colonies declining, but not of ant workers (Table 3.1 model E).   29  Table 3.1. Test statistics of generalized linear mixed models and general linear models testing the effect of elevation on various measurements of ant and treehopper abundance related to their mutualistic associations on the Andes of Ecuador. Error distributions in models denoted as a: negative binomial with log link, b: zero-inflated poisson with log link, c: zero-inflated negative binomial with log link, d: poisson with log link, e: linear model. All models included locations (n = 48) nested within Andean slope (n = 2) as random effects, and model O also nested individual aggregations within locations (n = 113).  Model, response variable Model factor Estimate ± std. error P Model Aa (n = 96)      Number of treehopper Elevation (km) -0.78 ± 0.12 <0.01 Model Bb (n = 96)      Number of ant colonies Elevation (km) -0.99 ± 0.18 <0.01  Zero-inflation factor 4.09 ± 1.09 <0.01 Model Cc (n = 96)      Number of ant workers Elevation (km) -1.81 ± 0.27 <0.01  Zero-inflation factor 4.33 ± 1.09 <0.01 Model Dc (n = 96)    Number of ant-treehopper aggregations Elevation (km) -1.31 ± 0.46 <0.01  Zero-inflation factor 5.18 ± 4.92 0.29 Model Ed (n = 96)    Number of ant-treehopper aggregations Number of treehoppers 0.21 ± 0.32 <0.01  Number of ant colonies 0.15 ± 0.44 <0.01  Number of ant workers 0.19 ± 0.17 0.99 Model Fe (n = 108)      Ratio of ant colonies to treehoppers (E) Elevation (km) -0.19 ± 0.12 0.1  (R) Ratio type (supply) 0.61 ± 0.12 <0.01  E * R (supply) -0.17 ± 0.09 0.07 Model Ge (n = 108)      Ratio of ant workers to treehoppers (E) Elevation (km) -0.29 ± 0.13 0.03  (R) Ratio type (supply) -1.45 ± 0.19 <0.01  E * R (supply) -0.94 ± 0.14 <0.01 Model Ha (n = 113)      Number of ant workers per aggregation (T) Number of treehoppers 0.16 ± 0.02 <0.01  (E) Elevation (km) 0.36 ± 0.26 0.15  T * E -0.07 ± 0.02 <0.01 Model Ia (n = 226)      Number of partners per aggregation (P) Partner type (treehoppers) -0.44 ± 0.14 <0.01  (E) Elevation (km) 0.51 ± 0.24 0.04  E*P (treehoppers) 0.44 ± 0.16 <0.01    30     Fig 3.2. (top) Total abundance of treehoppers and ants, and (bottom) the proportion of individuals belonging to species known to engage in mutualism across elevations in the Andes of Ecuador. Lines in the top graphs represent the best fit of generalized mixed effect models.  Changes in the abundance of ants and treehoppers across elevations resulted in a reversal in the relative abundance of the partners.  Thus, at lower elevations there was an excess in the supply of ant-worker partners relative to their demand (i.e., treehopper aggregations), whereas at elevations between 1500 and 2100 the opposite was true (Fig. 3.3; Table 3.1 model F, G). On the other hand, treehoppers never experienced a shortage of ant colonies relative to their demand by treehoppers across elevations.   31   Fig 3.3. Number of ant-treehopper aggregations across elevations in the Andes of Ecuador.    Fig 3.4. Ant-worker to treehopper ratio, reflecting the supply and demand of (A) ant workers and (B) ant colonies, in relation to ant-treehopper mutualistic associations along the elevational gradient in the Andes of Ecuador. Supply is measured as the ratio of ants to treehoppers present in the environment and the demand as the ratio of ants to treehoppers that are engaged in mutualism  Within individual aggregations, a single ant colony tended a treehopper aggregation independent of treehopper numbers, but with a greater number of workers in larger treehopper aggregations (Table 3.1 model H). The number of ants and treehoppers within aggregations increased with elevation, but treehopper numbers increased at a faster rate, resulting in a deficit of ant workers  32  relative to treehoppers at the highest elevations where associations occurred (Fig 3.5; Table 1 model I).   Fig 3.5. (A) Number of associated treehoppers and ant workers per aggregation and (B) their ratio across elevations in the Andes of Ecuador.  3.4 Discussion  By surveying ants, treehoppers, and their mutualistic associations across 4000-meter elevational gradients in the tropical Andes, we show that their associations become less frequent at higher elevations and disappear above 2100 m (Fig 3.3). This pattern was mostly driven by a decline in ant and treehopper abundance, as well as a steady decline in the number of treehoppers belonging to species that engage in mutualism with ants (Fig 3.2). The relative availability of partners across elevations, however, was neither constant nor symmetrical. Whereas the ratio of ant colonies to treehoppers remained relatively constant across elevations, there was a surplus of ant workers at the lowlands and a shortage above 1500 m (Fig 3.4). These patterns should mediate shifts in the ecological cost partners incur while associating across elevations, as reflected in the size of their aggregations and relative composition of partners. Despite decreasing abundances of both ants and treehoppers with elevation, the size of the aggregations, both in terms of ants and treehoppers, increased, albeit with treehopper numbers increasing at a faster pace (Fig 3.5).   33  We show the effects of declines in partner abundance were not limited to a decrease in the frequency of ant-treehopper associations, but also on their apparent investment on the interaction. We show the number of ant workers and treehoppers within aggregations are positively correlated, likely a response to the positive feedback caused by increased honeydew production and increased ant attention (Cushman & Whitham, 1991; Morales, 2000). However, the benefits both partners experience with increasing numbers likely exhibit diminishing returns, which should prompt each partner to minimize their numbers while maximizing their partners’, when ecologically possible (Cushman & Whitham, 1991; Morales, 2000; Holland et al., 2002; Reithel & Campbell, 2008; Palmer & Brody, 2013). Thus, in the lowlands, the high availability of partners is likely responsible for the formation of many, but small aggregations. As the abundance of ant colonies, ant workers, and treehoppers decrease at higher elevations, however, the availability of partners likely becomes less spatially predictable. This would cause the relative value of partners as a resource to increases, leading to fewer but larger aggregations. When colony density is low, treehoppers should refrain to leave their aggregation after ant partners have been procured. At the same time, ant colonies may increase the number of workers involved in the associations to reduce the chances the treehoppers they are tending leave, thus increasing their hold on the limited local supply of treehoppers (i.e. resource monopolization). Additionally, a larger group of ant workers may better defend the precious sugar resources provided by treehoppers (Palmer, 2004).  A separate question concerns why the size of treehoppers’ aggregations increase at higher elevations while tended by fewer ant workers, in what appears a lower-quality deal for the treehoppers. The smaller ant workforce available at higher elevations would explain their limited numbers in aggregations, but not the large number of treehoppers aggregating around them. Treehoppers, however, may be forced to accept a suboptimal deal by the scarcity of ant-workers relative to treehoppers, which would make ants less predictable and more difficult to replace (Noe & Hammerstein, 1994; Jhonston & Bshary, 2008; Hoeksema & Bruna, 2015). This could be aggravated if low ant-colony densities give ants a further advantage over treehoppers, prompting them to become exploitative (Bronstein, 2001). Ant exploitation is especially dangerous for treehoppers as ants could turn into direct predators of the treehoppers if allowed to become exploitative (Stadler & Dixon, 2005). This may explain why, rather than relying on ants to provide protection from predators, most treehoppers at higher elevations have evolved alternative strategies to counteract predation such as maternal care, which is predominant in high elevation  34  treehopper communities (Olmstead & Wood, 1993). These species do not associate with ants even though they also produce honeydew, which would allow them to do so.  The proportion of ant and treehopper individuals relying on the mutualism across elevations may reflect the degree to which the interaction contributes to their fitness, thus reflecting the interaction’s strength (Schemske et al., 2009). Our results thus suggest that the mutualistic interaction becomes less important for treehopper communities at higher elevations, which is potentially associated to lower ant abundance, their propensity to become exploitative, and the low necessity for their protection services against low predation rates (Roslyn et al., 2017; Camacho & Avilés, 2019). Conversely, the sustained interest of ants to engage in mutualism across elevations suggests the interaction remains favorable to them. In fact, ants may derive higher benefits from honeydew produced by treehoppers or other hemipterans at higher elevations where ecosystem productivity is lower and thus sugar-based resources scarcer (Kaspari et al., 2020).  Declines in ant abundance affected associations at two scales, one within individual aggregations, mediated by the abundance of ant workers, another across aggregations, mediated by the occurrence of ant colonies. Our results are consistent with the availability of ant colonies, rather than ant workers, being that factor that imposes a limit on the abundance of ant-treehopper aggregations.  The number of treehoppers within aggregations, however, was positively related to the number of ant workers. Treehopper aggregations are attended by a single ant colony, which involves higher numbers of ant workers when the number of aggregated treehoppers increases and vice-versa (Cushman & Whitham, 1991; Del-Claro & Oliveira, 1996; Del-Claro & Oliveira, 2000; Morales, 2000; Gove & Rico-Gray, 2006). The number of ants available to provide attention, however, is dependent on the distance to the nest (Morales, 2000; Gove & Rico-Gray, 2006). Thus, by tracking the spatial occurrence of ant workers, treehoppers my indirectly track the spatial distribution of ant nests and their mutualistic aggregations to match that distribution. Thus, the spatial distribution of ant colonies likely influences the availability of the ‘ant resource,’ whereas ant worker abundance influences the quality of the resource for which treehoppers compete in the vicinity of a colony (Cushman & Whitham, 1991; Stadler et al., 2001). Conversely, treehopper abundance influences both the predictability and quality of the resource for the ants (Pringle et al., 2011; Pringle et al., 2014), where individual nests compete to lure as many treehoppers as possible and thus potentially monopolize the entire local treehopper pool (Vincent et al., 1996;  35  Bluthgen et al., 2000; Bluthgen et al., 2006). A potential consequence of this pattern is that the benefit treehoppers derive from the ants scales with individual biomass (size of the ant colony), whereas the benefit ants derive from treehoppers scales with their abundance.  3.4.1 Conclusion We show the effects of declining abundance of ant and treehopper partners at higher elevations was not limited to a reduction in the frequency of their associations, but also influenced the size and composition of the aggregations. Aggregations at higher elevations grew larger with treehoppers increasing at a faster pace, despite declining abundance of both partners. We suggest this pattern arises from consumer resource-dynamics, where lower relative abundance of partners increases their ecological value, and thus their aggregative power towards the other partner. Treehoppers may experience a reduction in the ecological benefits they derive from the interaction, as well as being forced to compete with each other for a reduced ant workforce at higher elevations. This may explain why most treehoppers at higher elevations do not associate with ants and had exhibit alternative strategies. Our results thus point towards possible consequences of ecological shifts in mutualistic interactions mediated by partner abundance on the composition of ecological communities, either through assembly or evolutionary routes. The diversity of the partner species and their network configuration should prove fruitful avenues for future studies addressing the ecological dynamics in the ant-treehopper mutualism, as well as other mutualistic systems (Stanton, 2003; Bascompte, 2019).     36  Chapter 4. Consequences of macroecological gradients of predation on maternal care and mutualistic interactions with ants in Andean treehoppers  4.1 Introduction  Ever since Dobzhansky (1950) proposed the idea that biotic interactions are stronger in the tropics, the search for evidence to that claim has been ongoing. The shared characteristics between latitudinal and elevational gradients have led macroecologists to expect predation to also be stronger at lower elevations. Thus, some studies have focused on predation and have suggested that predation is more important at lower elevations and latitudes (Schemske et al., 2009; Roslyn et al., 2017; Camacho & Avilés, 2019). These studies, however, have largely based their conclusions on measurements of predation rates across time, which may not necessarily reflect the impact of predation on prey populations (Schemske et al., 2009; Louthan et al., 2015; Camacho & Avilés, 2019). Decreasing predation rates with elevation and latitude have been proposed to be ultimately linked to lower temperatures which, by affecting primary productivity and metabolic rates, would reduce the abundance and activity of predators (Brown, 2014; Camacho & Avilés, 2019). However, reduced productivity and metabolic rates should also reduce prey abundance, fecundity, and ability to escape predators, thus potentially increasing the impact on prey populations of a given predation rate (Munch et al., 2005; Camacho & Avilés, 2019). As a result, there is the distinct possibility that, whereas predation rates decrease with latitude and elevation, the actual effect of predation on prey populations remains constant or even increases (Louthan et al., 2015). In addition to gradients in predation rates, associated shifts in predator composition and the means available to counteract them may add a qualitative aspect to the effect of predation gradients on prey populations. For instance, decreasing abundance of ants with elevation and latitude (Roslyn et al., 2017; Camacho & Avilés, 2019), reduce their contribution to predation as well as their availability as mutualist partners to provide protection against other predators (Chapter 2, Chapter 3). This should have important consequences in how prey invertebrate prey communities are affected by predation across latitudes and elevation. A potential approach to address both the quantitative and qualitative aspects of the impact of predation on prey is to study anti-predator adaptations in prey communities across such gradients (Abrams & Matsuda, 1993; Schemske et al., 2009).   37  Here we focus on communities of treehoppers (Hemiptera: Membracidae), a group of sap sucking insects whose adaptations against predation are closely associated to each species’ behavioral strategies (Wood, 1984). Treehoppers can rely on different combinations of investment in maternal care and in mutualistic associations with ants, with both potentially exhibited during different life stages (Wood, 1993; Godoy et al., 2006) (Fig 1). These strategies, however, come at a cost for treehoppers, as mothers caring for offspring suffer from increased mortality and reduced fecundity (Tallamy & Wood, 1986, Tallamy & Schaefer, 1997, Zink, 2003) and treehoppers give up nutrients from sap to make honeydew more attractive to ants (Stadler & Dixon, 1998; Fischer et al., 2002; Yao & Akimoto, 2002; Billick et al., 2007).  We looked for changes in anti-predator investment in treehopper communities in response to quantitative and qualitative gradients of predation across elevations in the equatorial Andes. Treehoppers occur from lowland rainforest up to 4000 m of elevation, being subject to decreased predation rates at higher elevations due primarily to a decrease in the abundance of ants (Camacho & Avilés, 2019). Mediated by the presence and abundance of ants, treehoppers are thus subject to two parallel gradients of biotic interactions –one of predation and another of mutualism. In response to these gradients, we would expect that if predation rates have a 1:1 correspondence with predation impact, treehoppers would decrease their overall anti-predator investment in parallel to that of predation rates. We further expect that whether communities invest in ant-mutualism or in maternal care would depend on the presence or absence of ants across elevations, either as partners that can provide defense in exchange of honeydew or as predators to be dissuaded from predation by means of a honeydew offering (Delabie, 2001).  4.2 Methods  4.2.1 Predation and mutualist-ant availability across elevations in the Andes We collected data on the east and west slopes of the Andes in Ecuador at 500-meter intervals from 0 to 4000 m.a.s.l., for a total of 48 sites (24 sites on each slope) and three sampling locations per site (Appendix B, Table S1). At each sampling location, we set two transects separated by at least 100 m. All transects were set up on rain-free days between 8 am and 1 pm along forest edge next to secondary roads, farmland, or trails, which are places where treehoppers are frequently found. We collected data during May and June 2017. The order in which sites were visited was  38  scrambled within and across the east and west sides of the Andes to reduce temporal effects on the observed patterns.  We used live insect baits to measure overall predation rates on local insect communities and the proportion of this predation exerted by mutualist ants. For this purpose, we set up baits of live lab-bred adults of the fly Lucilia sericata (Diptera: Calliphoridae). These baits were intended to represent actual live prey while minimizing intrinsic anti-predator adaptations such as escape behavior and physical or chemical defenses. Additionally, we used sugary baits to determine the local availability of mutualist ants to associate with treehoppers. The latter baits consisted of Eppendorf tubes filled with cotton soaked in a 30% V/V solution of sucrose that approximates the sugar concentration, albeit not composition, of hemipteran honeydew (Fischer et al., 2002, Fischer et al., 2005).  Fly and sugary baits were attached to apical stems and leaf petioles of overhanging vegetation where treehoppers usually dwell. Flies were affixed to vegetation by their backs using sticky insect-trap paper, whereas sugary baits were tied to vegetation with plastic string. We were careful not to contaminate baits with sweat, which would have affected their luring effect (Kaspari et al., 2008). Each transect consisted of 15 baits separated by at least 5 m from each another.  We monitored fly baits every 30 minutes for three hours and recorded whether they had disappeared or were being attacked. We calculated the proportion per hour of baits missing or being attacked and averaged these proportions across the three hours sampled per transect. This represents the rate of predation per hour per transect. The same procedure was used for sugary baits, where we calculated the proportion per hour of baits that had attracted ants across the three hours sampled per transect. This represents the rate of mutualist-ant recruitment per hour.  We collected the ants seen attacking the fly baits and assumed that missing baits had been consumed by non-ant predators, as ant predation is a slow swarming process that is unlikely to be missed. We separately collected ants attending hemipterans in the transects’ surroundings. We compared these ants to those found consuming the fly baits to identify overlapping species, which allowed us to calculate the proportion of all predation exerted by mutualist ants at each transect (Appendix B, Table S4).  39   4.2.2 Behavioral composition of treehopper communities across elevations in the Andes We used the invertebrate collection at the QCAZi Museum at the Pontificia Universidad Católica del Ecuador to determine the composition of treehopper communities across elevations in the western and eastern slopes of the Andes of Ecuador. This collection houses over 9000 treehopper specimens, encompassing more than 500 species collected over 400 locations in Ecuador. We classified specimens into morpho-species and identified them as close to genus as possible. We determined the elevational ranges at the western and eastern slopes for each morpho-species based on their minimum and maximum elevations in the collection.  We characterized each morpho-species along axes of maternal care (no care, only to eggs, or care extended to nymphs) and ant mutualism (no mutualism, only during nymphal stage, or mutualism extended to adulthood), which yielded six observed categories (A-F) (Fig 4.1). Based on these categories we scored each species investment in maternal care and ant mutualism using an index that assigns one point for each life stage exhibiting the behavior, and then added these scores to obtain an overall anti-predator investment score (Fig 4.1). Behavioral strategies were assigned based on direct observations of Ecuadorian treehopper fauna as well as literature on species, genera, or tribes (Wood, 1982; Olmstead & Wood, 1990; Godoy et al., 2006) (Appendix B, Table S5).    40    Figure 4.1. Treehopper behavioral strategies against predation, showing associated scores of anti-predator investment in maternal care or ant-mutualism (numbers in parenthesis) and of total anti-predator investment (bottom right number in each cell). Most treehoppers do not show any investment in brood care or ant-mutualism (A, 166 species). In other species, females care for their brood during the egg stage (B, 37 species) or until offspring reach adulthood (C, 48 species). Species can also offer honeydew to associate with ants during the nymphal stage (D, 75 species) or throughout the adult stage as well (E, 57 species). In other species, females care for their egg masses while ants attend both nymphs and adults (F, 123 species). A: a Lycoderes individual minds his own business at ~1900 m.a.s.l. in eastern Ecuadorian Andes; B: a female Polyglypta guards her egg mass at ~2900 m.a.s.l. in the western Ecuadorian Andes; C: a female Alchisme grossa guards her nymph brood at their final instar at ~1400 m.a.s.l. in the western Ecuadorian Andes; D: two nymphs of Heteronotus vandamei are attended by Camponotus ants at ~450 m.a.s.l. in the eastern Ecuadorian Amazon; E: a group of adults and nymphs of a membracini species is attended by Camponotus ants at ~250 m.a.s.l. in the eastern Ecuadorian Amazon; Horiola picta females are attended by Camponotus ants at ~800 m.a.s.l. in the eastern Ecuadorian Andean foothills.  Based on the elevational range of each morphospecies, we determined the species composition of treehopper communities at elevational intervals of 500 m in the western and eastern Andean slopes. For each community we calculated the proportion of species exhibiting the various behavioral strategies and calculated the average anti-predator index, ant mutualism investment index, and maternal care investment index of each community. We repeated this procedure  41  separately for the phylogenetically independent subfamilies Membracinae and Smiliinae to provide a phylogenetic assessment in our analyses (Dietrich et al., 2001; Cryan et al., 2004; Lin et al., 2004; Evangelista et al., 2016).  4.2.3 Data analysis We compared the behavioral assemblage of treehopper communities across elevations and between slopes using a permutational analysis of variance (PERMANOVA).  We used generalized linear models (GLM) with binomial error distribution and a logit link function to generate predictive regression models of predation rate, proportion of predation exerted by mutualist-ants, and mutualist-ant recruitment rate as a function of elevation and slope. Predictions made by these models were then used in subsequent analyses testing the relationship of these variables with anti-predator investment in treehopper communities.  We tested whether the overall anti-predator investment of treehopper communities decreased with elevation either at a constant rate or nonlinearly with a drop above an elevational threshold.  For the former we used a general linear model (GLM) and the following parabolic function (NLM), for the latter: 𝑦 = 𝑏 − (𝑥𝑚)𝑛  where b determines the intercept, m the amplitude of the curve, and n the steepness of the decrease. We compared these models using Akaike information criteria scores corrected for small sample sizes (AICc). We then tested whether anti-predator investment decreased with elevation at a rate more or less closely associated with decreasing predation rates by comparing the AICc scores of the GLM described above and a GLM testing the average anti-predator index of communities as a function of predation rate.  We also tested if investment in communities in ant-mutualism, as opposed to maternal care, covaried more closely with the local availability of mutualist ants or with the contribution of mutualist-ants to total predation. For this we compared the AICc values generated from two GLMs, testing the proportion of total anti-predator investment invested in ant-mutualism as a  42  function of mutualist-ant recruitment rate in the first model, and proportion of total predation exerted by mutualist ants in the second model.   All models controlled for the Andean slope by including it as a fixed factor. NLMs allowed only the intercept (b) in the parabolic function to vary with elevation.  To address possible phylogenetic correlation in our patterns, and in the absence of a detailed treehopper phylogeny, we replicated our community analyses using the phylogenetically independent subfamilies Membracinae and Smiliinae (Dietrich et al., 2001; Cryan et al., 2004; Lin et al., 2004; Evangelista et al., 2016). If the observed patterns are replicated in the two independent evolutionary lineages, this would be consistent with the hypothesis that they are driven by ecological, rather than being a unique historical coincidence. Membracinae and Smiliinae represent respectively 36% and 39% of the 509 species included in our analyses and encompass all the anti-predator traits considered (Wood, 1993).  All statistical analyses were carried out in R v3.2.2 software (R Core Team, 2019). We used packages dplyr for data manipulation (Wickham et al., 2020), nlme for NLM (Bates et al., 2018), MuMIn for AICc calculations (Barton, 2020), sjstats for effect size calculations (Lüdecke, 2020), and ggplot2 (Wickham, 2016), ggpubr (Kassambara, 2019), and lemon (McKinnon, 2019) for figure design.  4.3 Results  Higher elevations saw a decrease in predation rates (GLM, slope per 1000-meters (log-odds) = -2.24 ± 0.92, χ2 = 14.39, P < 0.001), proportion of total predation exerted by mutualist-ants (GLM, slope per 1000-meters (log-odds) = -2.05 ± 0.82, χ2 = 9.72, P = 0.002), and the availability of mutualist ants (GLM, slope per 1000-meters (log-odds) = -2.31 ± 0.97, χ2 = 14.1, P < 0.001), with no difference between the western and eastern slopes (Fig 4.2).   43   Figure 4.2. Patterns across elevation of (A) predation rates, (B) proportion of total predation exerted by mutualist ants, and (C) mutualist-ant recruitment rate at the western and eastern slopes of the Andes of Ecuador. Lines represent the best fit of GLM with a binomial error distribution (n = 48).  The communities were significantly different across elevations in terms of the behavioral strategies they contained, either when all groups were considered together (PERMANOVA, F = 13.02, R2 = 0.47, P < 0.001) (Fig. 4.3A) or when considering separately the subfamilies Membracinae or Smiliinae (PERMANOVA,  F = 37.94, R2 = 0.76, P < 0.001, for the former; F = 14.18, R2 = 0.45, P < 0.001, for the latter) (Appendix A, Fig S2). In either case, there was no significant difference between the western and eastern slopes, apart from western Smiliinae communities tending to have an overall 10% less representation of mutualism and maternal care (PERMANOVA, F = 3.99, R2 = 0.13, P = 0.02) (Appendix A, Fig S2). On both sides of the Andes, the abundance of mutualist treehoppers peaked below 1000 m, decreasing thereafter to eventually disappear above 2500 m (Fig 4.3A). The abundance of treehoppers with maternal care increased with elevation, peaking between 2000 and 3500 m and decreasing only at the highest elevations (Fig. 4.3A). The abundance of species exhibiting neither ant-mutualism nor maternal care remained relatively constant up to 2500 m, but increased thereafter to become the prevalent system above 3500 m (Fig 4.3A). These patterns were also exhibited by Membracinae and Smiliinae communities independently, with the only difference that the Membracinae treehoppers that do not exhibit ant-mutualism nor maternal care do not increase in prevalence with elevation (Appendix A, Fig S2).   44    Figure 4.3. (A) Species behavioral assemblages and (B) anti-predator investment in treehopper communities across elevation in the Andes of Ecuador. Lines represent the best fit of NLM with a parabolic function (n = 16). All assemblages are presented as the average between western and eastern slopes.    45  The average anti-predator investment in treehopper communities decreased with elevation only above 2500 m (NLM, AICc = -6.77) rather than linearly across elevations (GLM, AICc = 15.75) (Fig 4.3B, Table 4.1). When considering individual subfamilies, Membracinae did not show differences in anti-predator investment across elevations (GLM, AICc = 9.83), but Smiliinae did. Specifically, the Smiliinae anti-predator investment decreases with elevation (GLM, AICc = 9.06) and shows a sharper decline above 2000m (NLE = 8.93, Table 4.1). Anti-predator investment did not closely parallel decreasing predation rates across elevations, either overall (GLM, AICc = 25.19) (Fig 4.4, Table 4.1) or when considering separately the subfamilies Membracinae (GLM, AICc = 9.29) or Smiliinae (GLM, AICc = 21.29) (Table4. 1).     Figure 4.4. Average anti-predator investment in treehopper communities in relation to predation rates across elevations in the Andes of Ecuador. Predation rates lines represent the best fit of a GLM with a binomial error distribution (n = 48), and anti-predator investment lines represent the best fit of NLM with a parabolic function (n = 16).  Instead, at higher elevations, communities shifted their anti-predator investment from ant-mutualism to maternal care at a rate that was almost perfectly correlated with the decrease in the  46  contribution of mutualist ants to predation (GLM, AICc = -43.27) (Fig 4.5A,B; Table 4.1), rather than to the availability of ants as mutualist partners (GLM, AICc = -4.61) (Fig 4.5A,C; Table 4.1). This pattern was also present in the Membracinae and the Smiliinae (Table 4.1).    Figure 4.5. Ratio of ant mutualism to maternal care investment in treehopper communities in relation to the availability of mutualist ants and the proportion of total predation exerted by the mutualist ants across elevation in the Andes of Ecuador. Lines represent the best fit of GLMs (n = 16)     47  Table 4.1. Test statistics of general linear models (GLM) and non-linear models (NLS) looking at anti-predator investment of treehopper Andean communities as a function of elevation and predation rates, and their investment in ant-mutualism as opposed to maternal care as a function of the availability of mutualist ants and the contribution of mutualist ants to predation. NLS fitted a parabolic function. All models controlled for the effects of the Western and Eastern Andean slopes by including them as a fixed factor. Models performance based on Akaike’s information criteria for low sample sizes (AICc).  Response variable / test community Model type Predictor Slope ± std. error R2 AICc Effect size (ω2) Anti-predator investment index     All Membracidae NLM Elevation (Km)   -6.77   GLM Elevation (Km) -0.25 ± 0.07 0.59 15.75 0.39  GLM Predation rate 0.73 ± 0.51 0.27 25.19 0.06        Membracinae NLM Elevation (Km)   9.91   GLM Elevation (Km) 0.004 ± 0.07 0.03 9.83 0  GLM Predation rate -0.21 ± 0.13 0.07 9.29 0        Smiliinae NLM Elevation (Km)   8.93   GLM Elevation (Km) -0.33 ± 0.05 0.8 9.06 0.54  GLM Predation rate 1.45 ± 0.44 0.58 21.29 0.38 Proportion of anti-predator investment in ant-mutualism     All Membracidae GLM Mutualist-ant recruitment rate 1.3 ± 0.19 0.78 -4.61 0.72  GLM Proportion of total predation exerted by mutualist ants 0.87 ± 0.03 0.98 -43.27 0.96        Membracinae GLM Mutualist-ant recruitment rate 1.26 ± 0.19 0.79 -4.77 0.73  GLM Proportion of total predation exerted by mutualist ants 0.84 ± 0.05 0.96 -29.94 0.93        Smiliinae GLM Mutualist-ant recruitment rate 1.29 ± 0.21 0.75 -2.06 0.67  GLM Proportion of total predation exerted by mutualist ants 0.87 ± 0.05 0.96 -27.75 0.92  4.4 Discussion  An analysis of how antipredator strategies change with elevation in 48 treehopper communities on the Andes (509 treehopper species belonging to 105 genera and 8 families) reveals a relatively constant anti-predator investment up to 2500 m and a steep decline, thereafter (Fig 3B, Fig A1B). These patterns of change in investment did not parallel changes in predation rate, which, instead declined rapidly with elevation, reaching near zero values near 2500 m (Fig 4). The lack of a 1:1 correspondence with predation rates suggests that the impact of predation on prey populations remains even at low predation rates. We also found that the pattern by which anti-predator adaptations changed with elevation—from ant-mutualism to maternal care—closely matched the declining contribution of ants to predation, rather than to the availability of ants as mutualist partners (Fig 5). The latter observation suggests that treehopper mutualism with ants may be a strategy against predation by the ants themselves, rather than a strategy to recruit the ants’ services  48  against other predators. Similar patterns were present when two phylogenetically independent subfamilies were analyzed separately (Fig A1, Table A3), supporting the hypothesis that it is ecological, rather than historical evolutionary processes, that shape the pattern of treehopper anti-predator strategies across elevation.   To understand the antipredator investment of treehoppers in its most fundamental form, we need to determine the degree to which species sacrifice their reproductive potential to counteract predation (Jeffries & Lawton, 1984; Stearns, 1989; Lind & Cresswell, 2005; Beckerman et al., 2010). Reproductive potential, in turn, should be related to the rate at which ectotherms extract energy from the environment and invest it into growth and reproduction, which must be limited by temperature at higher elevations (Brown, 2004; Munch et al., 2005; Gripenberg & Roslin, 2007; Brown, 2014; Schramski et al., 2015). This should result in treehoppers being vulnerable even to low predation rates at higher elevations (or lower temperatures) and being required to maintain high levels of anti-predator investment as in the lowlands. In fact, the metabolic theory of ecology predicts that the rate at which organisms extract energy from the environment, their rate of energetic expenditure in growth and reproduction, and the rate of predation they experience all scale at the same rate with temperature. This suggests that as temperature decreases along the elevation gradient, the negative effects on prey population growth cancels out with the positive effects of reduced predation rates; thus, causing the impact of predation to remain constant across elevations (Brown, 2004; Brown, 2014; Schramski et al., 2015; Brown et al., 2018). Treehoppers should thus be affected by predation as long as local primary productivity supports a predator trophic level (Fretwell 1977; Oksanen et al. 1981; Fretwell 1987; Hunter & Price, 1992; Power 1992), which may no longer be the case at the highest elevations, potentially explaining the sudden decrease in treehopper anti-predator investment above 2500 m.  The mechanism by which predation influences treehopper assemblages across elevations can be understood in the context of how species establish their ecological niche and attain enemy-free space. Treehopper species presumably establish their ecological niches based on resources, while minimizing their exposure to predation (i.e. enemy-free space) (Hutchinson, 1957; May, 1974; Jeffries & Lawton, 1984; Holt & Lawton, 1994; Singer & Stireman, 2005; Holt, 2009). Treehoppers may avoid resource use overlap by adapting to different host plants (Singer & Stireman, 2005; Vidal & Murphy, 2018). Enemy-free space, on the other hand, may be attained by a choice of  49  different environments or by species adopting alternative anti-predator strategies within a given environment (Singer & Stireman, 2005). The combined effect of these processes should influence the makeup of treehopper communities. For instance, enemy-free space could be intrinsically available in environments with low predation, favoring species with low anti-predator investment, as may be the case for most treehoppers occurring above 3500 m. Other environments could favor species able to exploit habitat properties to find enemy-free space, such as mutualist treehoppers taking advantage of the ubiquity of ants in the lowlands. Finally, some environments could favor species that create their own enemy-free space by evolving adaptations that enable them to engage predators, such as treehoppers with maternal care at high elevations. Thus, the elevational gradient in the strength of predation is a composite of quantitative and qualitative factors affecting prey and can be understood as an elevational gradient in enemy-free space upon which treehopper communities assemble.  As strategies, treehopper-ant mutualism and maternal care are likely tightly correlated with a suit of other behavioral and morphological traits that further influence the degree of anti-predator investment in communities across elevations. While a formal analysis of such traits is beyond the scope of this paper, natural history of different taxa provides insights on the nature of such adaptation. Treehoppers engaging in ant mutualism, for instance, are likely largely protected from predation by the mutualist-ants themselves as well as other predators and parasitoids repelled by the ants. Thus, mutualist treehoppers likely concentrate their anti-predator investment in ant recruitment through gregariousness and the offering of honeydew to ants. Species exhibiting maternal care, on the other hand, are required to engage predators to defend their brood and likely complement their investment in maternal care with other anti-predator adaptations such as armor or toxin sequestration. For instance, the treehopper Alchisme grossa (Fig 1C), in addition to sophisticated brood care, exhibits armor-like morphology and sequesters toxins from pungent Solanum host plants (Camacho et al., 2014; Torrico-Bazoberry et al., 2014; Pinto et al., 2016).  The latter is a combination shared by many species occurring between 1000 m and 3500 m in species in the phylogenetically independent Membracinae and the Smilliinae (e.g. tribes Hoplophorionini (Fig 1C) and Polyglyptini (Fig 1B)). Thus, treehoppers in the lowlands may be largely protected from predation by their association with mutualist ants. The fact that at higher elevations they combine maternal care with other anti-predator adaptations is further consistent with the  50  argument that they suffer important impacts of predation, despite documented lower predation rates in those areas.  Parasitoid wasps, which attack treehopper eggs, could also be a relevant threat shaping patterns of treehopper anti-predator investment across elevations (Hawkins et al., 1997; Godoy et al., 2006). Here, the same adaptations that treehoppers use as protection from predation—maternal egg-care and protection by ant-mutualists—are also used to repel parasitoid wasps (Godoy et al., 2006; Camacho et al., 2013; Fagundes et al., 2013). The frequency of such adaptations in communities across elevations suggests parasitism is an important threat to treehoppers throughout most of the elevational gradient (Fig 3A, Fig A1A). However, the pattern of how these strategies are distributed across elevation cannot be attributed solely to parasitism because egg maternal care is also effective against other egg predators and nymph maternal care is only effective against non-parasitoid predators (Olmstead & Wood, 1990b; Wood, 1993; Godoy et al., 2006).  4.4.1 Conclusion We show treehopper communities decrease their anti-predator investment only above 2500 m, despite a steep and constant decline in predation rates throughout the elevational range and it becoming virtually nil above 1500 m. We suggest the reason for this pattern is that changes in temperature and productivity with elevation not only affect predator but also prey populations; thus, causing the positive effects of reduced predation and the negative effects of reduced growth to cancel each other out. The impact of predation on prey populations would thus remain constant up until the highest elevations beyond which primary productivity can no longer sustain a predator trophic level. Similar effects may apply to other antagonistic interactions, such as competition and parasitism, as well as to mutualisms, and both along elevation and latitudinal gradients, particularly for ectotherms. In contrast to the cancelling effects of temperature and productivity, qualitative changes in partner composition, on the other hand, should play a key role in the outcome of biotic interactions. Our results show, for instance, that the shift from ant-mutualism to maternal care as the predominant anti-predator strategy was associated with a decline of ants as the main predators at higher elevations. This is consistent with the idea that qualitative changes in community composition may play a more significant role in shaping life histories and biodiversity than the energetic constraints imposed by temperature (Brown et al., 2018). The patterns described here,  51  which have also been documented in other regions of the Andes (Olmstead & Wood, 1990), may apply to other geographic gradients, as well, as ant abundance is known to decline with elevation and latitude globally (Camacho & Avilés, 2019, and references within). Future studies in this and other systems should aim to determine the degree to which reproductive potential is sacrificed to counteract predation, which may include a comprehensive assessment of anti-predator traits, foraging strategy, fecundity, and life history (Lind & Cresswell, 2005). Knowledge of the natural history of the system is key, as qualitative changes in the interactions of a group of organisms with their environment may determine the outcome of predation and the mechanisms at play for avoiding it.   52  Chapter 5. Befriending the enemy: Can prey establish mutualistic interactions with their would-be predators?  5.1 Introduction  Animals may employ an array of behavioral strategies to avoid being preyed upon (Lima & Dill, 1990; Sih, 1994; Preisser et al., 2005). In addition to simple strategies, such as escaping or hiding, animals may develop more complex behaviors such as forging mutualistic interactions, whereby prey may offer resources or services to other species in exchange for protection from predators (Boucher et al., 1982; Bronstein, 2001; Leigh, 2010). An interesting and often overlooked possibility, in particular in general treatments of mutualism, is that prey may offer resources or services directly to the predators themselves –a ‘bribe’ – to avoid becoming their meal. Befriending a would-be predator could hold great ecological benefits to prey, not only by reducing predation pressure, but also by potentially deriving a benefit from an otherwise antagonistic interaction. Such mutualistic systems have the special property of being mediated at the level of the predatory interaction, in addition to the exchange of resources or services (Pierce et al., 2002). Thus, mutualistic interactions between prey and predators challenge our understanding of both predation and mutualism.  Sap-sucking Hemiptera (e.g. Aphididae, Coccidae, Aetalionidae, and Membracidae) are widely recognized to offer honeydew to ants in exchange of their defensive services against a variety of predators (Delabie, 2001; Stadler & Dixon, 2005). Ants, however, are major predators in arthropod communities and may also predate upon their hemipteran partners (Novotny, 1999; Delabie, 2001; Floren et al., 2002; Tvardikova & Novotny, 2012; Roslyn et al., 2017; Camacho & Aviles, 2019). This poses a challenge for hemipterans as, in order to engage in mutualism with ants, they need to first dissuade the ants from preying upon them. Hemipterans are known to appease ants with honeydew, which, despite being a waste product, it is typically enriched with carbohydrates and amino acids relinquished from sap to make the honeydew more attractive to the ants (Stadler & Dixon, 1998; Yao et al., 2000; Fischer & Shingleton, 2001; Fischer et al., 2002; Yao & Akimoto, 2002; Kay et al., 2004; Fischer et al., 2005). As a result, hemipterans may be able to turn ants from predators into mutualistic partners that provide enemy-free space through their defensive services or, at least, their reluctance to attack.  53   Two non-mutually exclusive benefits may, thus, prompt hemipterans to engage in mutualistic associations with ants: preventing predation by the ants themselves (the appeasement hypothesis) or obtaining their defensive services against other predators (the protection hypothesis) (Pierce et al., 2002). Whether mutualist hemipterans benefit more from the appeasement or protection aspects should depend on the relative importance in their local habitat of predation by potential ant partners versus other non-mutualist predators. For instance, in environments where most predation is exerted by the community of potential ant predators, most of the ecological benefit of associating with the ants would come from ant appeasement. The opposite would be true in environments where most predation is exerted by other types of predators, against which the ants would provide protection.  From the perspective of the ants, the nutritional value derived from honeydew may be the primary proximate mechanism appeasing the ants. In this case, ant appeasement would be dependent on a constant supply of honeydew regardless of identity of the producer. Alternatively, ants may refrain from predating on their hemipteran partners, even in the absence of a honeydew offering, if they recognize their partners through partner recognition mechanisms. Whether appeasement depends on a honeydew offering, partner recognition, or both may reflect the extent to which the predatory aspect of the interaction is mediated at the ecological time scale (honeydew only) or the evolutionary time scale, as well (both honeydew and partner recognition involved).  Here we use a two-pronged approach to study the main function and the mechanisms mediating the associations between ants and hemipterans, particularly treehoppers (Hemiptera: Membracidae). First, we teased the importance of the appeasement and protection hypotheses along a 3000-meter elevational range in the equatorial Andes by assessing the relative importance of predation from potential ant partners versus other non-partner predators (ant or otherwise) along the gradient. The extent to which treehoppers derive a benefit from the appeasement versus the protection aspects of their association should depend on the ants’ contribution to predation across elevations. In a second experiment, we tested the extent to which a honeydew offering vs. partner recognition may be the proximate mechanisms deterring ants from predating upon their hemipteran partners. We did so by performing choice experiments with treehoppers and other prey in the presence or absence of honey as a potential ‘bribe’ against predation. The ants  54  refraining from predating on treehopper baits in the absence of honeydew would suggest partner recognition through either learned or evolved responses. Ants refraining from predating on non-treehopper baits laced with honey, on the other hand, would be indicative of a protective effect of honeydew above and beyond partner recognition.  5.2 Methods  5.2.1 Predation by potential ant partners and occurrence of mutualist treehoppers across elevations in the Andes We collected data on the east and west slopes of the Andes in Ecuador at 500-meter intervals from 0 to 3000 m.a.s.l., for a total of 36 sites (18 sites on each slope) (Fig 5.1A, B; Appendix B, Table S1). Both slopes, which have similar ecological conditions transitioning from lowland rainforests, to cloud forests, to highland paramo, were considered as biogeographically independent replicates (Appendix B, Table S1). At each sampling location, we set two transects separated by at least 100 m (Fig 5.1B). All transects were set up on rain-less days between 8 am and 1 pm along forest edge next to secondary roads, farmland, or trails, which are places where treehoppers are frequently found. We collected data during May and June 2017. The order in which sites were visited was scrambled within and across the east and west sides of the Andes to reduce temporal effects on the observed patterns.  We used live insect baits to measure overall predation rates on local insect communities and the proportion of this predation exerted by ant species that are potential partners of hemipterans. For this purpose, we set up baits of lab-bred adults of the fly Lucilia sericata (Diptera: Calliphoridae), which we deployed live to ensure these prompted predation rather than scavenging (Fig 5.1C). These baits were intended to represent actual live prey while minimizing intrinsic anti-predator adaptations such as escape behavior and physical or chemical defenses. Using sticky insect-trap paper, we attached fly baits by their backs to apical stems and leaf petioles of overhanging vegetation where treehoppers usually dwell (Fig. 5.1C). We were careful not to contaminate baits with sweat, which would have affected their luring effect (Kaspari et al. 2008). Each transect consisted of 15 baits separated by at least 5 m from one another.   55   Fig 5.1. Setup of experiment 1 addressing the availability and predation by partner ants (A, B, C), and experiment 2 addressing partner recognition and value of honeydew (D, E, F). A: sampling locations at the western and eastern sides of the equatorial Andes. B: layout of experimental units for experiment 1. C: live bait of the fly Lucilia sericata. D: detail of the Jatun Sacha Biological reserve, where choice experiments were conducted along 12 forest-edge transect replicates (transect 12, 6 km to the west, not shown for better depiction of study area). E: sampling layout of experimental units for experiment 2. F: choice experiment of frozen Membracini treehopper with added honey on the right and without honey on the left.  We monitored fly baits every 30 minutes for three hours and recorded whether they had disappeared or were being attacked. We calculated the proportion per hour of baits missing or  56  being attacked and averaged these proportions across the three hours sampled per transect. This represents the rate of predation per hour per transect. We collected the ants seen attacking the fly baits and assumed that missing baits had been consumed by non-partner ants or other predators, as partner ants predate in a slow swarming process that is unlikely to be missed. We separately collected ants involved in ant-hemipteran associations in areas surrounding the transects. We compared these ants to those found consuming the fly baits to identify overlapping species (Appendix B, Table S4), which allowed us to calculate the proportion of all predation exerted by the community of potential ant partners at each transect.  During the 3-hour monitoring period, we estimated the composition and abundance of treehopper species occurring along each transect. We assessed whether treehoppers are potential partners of ants based on our observations of ant-treehopper associations in the area. Species that were not found partnering with the ants were also considered as potential partners if we found evidence for other species of the same genus directly in the field or in the literature (Appendix B, Table S3) (Godoy et al., 2006). For each transect, we calculated the proportion of treehoppers corresponding to species that associate with ants.  We used general linear models (GLM) with a binomial error distribution and a logit link to assess how predation rates by ant-partners and non-partner predators, the proportion of total predation exerted by ant-partners, and the proportion of treehoppers corresponding to partner species at each location changed with elevation. All models controlled for the Andean slope by including it as a fixed factor. We averaged measurements between transects in each location rather than including locations as random effects to avoid issues with model convergence.  5.2.2 Partner recognition and defensive value of honeydew At the lowest eastern elevation, we additionally performed choice experiments to assess the extent to which local ants may refrain from attacking their treehopper mutualist partners in the presence or absence of honey or of an alternative prey item. We carried out our choice experiments during June and July of 2018 at the Jatun Sacha Biological Reserve (1° 3' 58.33" S; 77° 36' 59.69" W) (Fig. 5.1D). We used as baits frozen treehoppers (Hemiptera: Membracidae and Aetalionidae) and termites (Nasutitermes sp., possibly N. corniger) in treatments combinations involving either the same or a different prey type, with and without honey. We chose treehoppers as the ant-mutualist baits,  57  as treehoppers are frequent ant associates (Delabie, 2001; Godoy et al., 2006) (Appendix B, Table S5).  We used termite workers as the alternative, non-mutualist bait as they are of comparable size and armor to treehoppers, but have a different chemical identity. In both cases, baits were locally collected and deployed frozen in order to immobilize potential prey while maintaining the integrity of their cuticular chemistry, which has been shown to play a role in the ants’ decision to attack or refrain from attacking (Hayashi et al., 2015; Endo & Itino, 2013; Hojo et al., 2014). A survival analysis showed that frozen termites and live fly baits had a similar risk of being predated upon once encountered (Cox proportional hazards mixed-effect model: hazard = 0.18 ± 0.16, chi-squared = 1.27, P = 0.26), suggesting that frozen baits were not treated differently from live ones. We added a ~1mm droplet of honeybee honey on the tip of the prey’s abdomen, which roughly approximates a honeydew offering as both contain protein, in addition to sugars (Bluthgen et al., 2004). Although the sugar composition of honey (mostly disaccharides) differs from that of honeydew (mostly trisaccharides), ants are known to accept a variety of sugar types (Volkl et al., 1999; Bluthgen & Fiedler, 2004).  In the experiment, we established six replicates of two transect types comparing: either same specimen type (treehopper or termite), one with honey, the other without; or different specimen types, both with or both without honey (Fig. 5.1E). This set up allowed us to separate the effects of specimen type versus those of presence or absence of honey. Bait alternatives in each transect were setup in couplets, with each specimen, treehopper or termite, attached with PVA glue to opposing stems of a bifurcation (Fig. 5.1F).  Each transect contained 10 couplets of each of the alternative bait types, for a total of 20 couplets per transect deployed in alternating arrangement and separated by at least 5 m from one another. Transects of the two types were placed in comparable locations on the landscape (Fig. 5.1D) and surveyed on rainless days between 9 am and 2 pm, alternating transect types across days to minimize possible effects of weather. We monitored baits at 30-minute intervals for three hours and recorded whether ants were attacking, probing, or absent. Attacks were recorded when ants were actively biting the baits, dismembering them, or carrying them away. Probing was recorded when ants remained with the baits touching them with their antennae and legs but not biting.  We tested for differences between bait treatments on the risk of predation by ants using a series of Cox proportional hazards mixed-effect models (CMM), which control for the time dependence  58  of predation events while testing the effects of each fixed factor simultaneously and independently. We first used two CMM, one focusing on transects with bait couplets comparing specimen type (treehopper and termite, both with or both without honey), and the other focusing on transects with bait couplets comparing the presence of honey (treehoppers or termites, one with honey and the other without) (Table 5.1, model A, B). In both models, specimen type and presence or absence of honey served as fixed factors. In a third CMM, we included both transect types and added experiment type as another fixed factor to test the effect of bait choice in ant predation independently of context (Table 5.1, model C). All three CMM included bait couplets (n=120 per transect type) nested within transect as random effects (n=6 per transect type). Baits predated by unknown predators (likely non-ant) (2.7% of all baits) were not included in the analyses, and those not registering an attack within the 3-hour monitoring period were coded as right-censored (10.2% of all baits).  We used a general linear mixed-effect model (GLMM) with a logit function and binomial error distribution to test whether at each 30-min time interval baits encountered by ants, but not predated upon, were being probed or were abandoned by the ants. Whether baits were being probed (coded as 1) or had been abandoned (coded as 0) served as the response variable. Transect type, time interval, bait type, and presence or absence of honey served as fixed factors; and each bait (n=240 per transect type), nested within bait couplet (n=120 per transect type), nested within transect (n=6 per type), served as random effects (Table 5.1, model D).  All statistical analyses were carried out in R v3.2.2 software (R Core Team, 2019). We used packages dplyr for data manipulation (Wickham et al., 2020), survival (Therneau, 2015) for construction of Cox models and coxme (Therneau, 2019) to incorporate random effects; glmmTMB (Brooks et al., 2017) for GLMM construction and car (Fox & Weisberg, 2019) for P-value calculations; and ggplot2 (Wickham, 2016), ggpubr (Kassambara, 2019), and lemon (McKinnon, 2019) for figure design.    59   5.3 Results  5.3.1 Predation by potential ant partners and occurrence of mutualist treehoppers across elevations in the Andes The rate of predation per hour declined steeply with elevation, with the vast majority of predation being exerted by potential ant-partners (50 to 100% of total predation) below 1500 m and by non-partner predators above this elevation (GLM: slope (log-odds) = -3.29 ± 1.31, chi-squared = 15.51, P < 0.01) (Figs. 5.2A and B). The switch in the identity of predator type along the gradient was the result of predation rates by the community of potential ant partners declining much more steeply with elevation (GLM: slope (log-odds) = -2.39 ± 1.07, chi-squared = 10.01, P < 0.01) than that by non-partner predators (GLM: slope (log-odds) = -0.73 ± 1.65, chi-squared = 0.22, P = 0.64) (Fig 5.2A). The proportion of treehoppers individuals that could potentially engage in ant mutualism also decreased steeply with elevation (GLM: slope (log-odds) = -2.39 ± 0.77, chi-squared = 18.33, P < 0.01), paralleling the decline seen in the proportion of total predation exerted by potential ant partners (Fig 5.2C).    Fig 5.2. Patterns across elevation of (A) predation rates by potential partner ants and non-partner predators, (B) proportion of total predation exerted by potential partner ants, and (C) incidence of treehoppers individuals that can potentially partner with ants at the western and eastern slopes of the Andes of Ecuador. Lines represent the best fit of GLM with a binomial error distribution and a logit link.   60  5.3.2 Partner recognition and defensive value of honeydew In the choice experiments, following discovery, ants were 0.31 as likely to predate upon treehoppers compared to termite baits when encountering a choice between the two (Table 5.1, model A; top panels in Appendix A, Fig S3). Ants encountering a choice of baits with honey or without were 0.27 as likely to predate on treehoppers or termites with added honey compared to their honey-less counterparts (Table 5.1, model B; bottom panels in Appendix A, Fig S3). There were no significant differences in the predation risk of baits of a particular type across the two choice-type experiments, which suggests that the ants’ predation choices were for the most part independent of the alternatives available in the immediate vicinity (Table 5.1, model C). Overall, ants were 0.30 times as likely to predate on baits consisting of treehoppers than termites and 0.32 as likely to predate on baits with honey than on baits without honey (Fig 5.3; Table 1, model C). Predation on termites and treehoppers without honey occurred for the most part within the first 60 minutes since discovery, albeit reaching a higher overall level for termites (Fig. 5.3). Predation on treehoppers with honey, on the other hand, occurred at a small but sustained rate across time. After ants depleted the honey from treehopper and termite baits, predation risk did not increase to match that of baits without honey.  Table 5.1. Test statistics of Cox proportional hazards mixed-effect models (CMM) and generalized linear mixed-effect models (GLMM) testing the behavioral response of ants in relation to baits consisting of treehoppers or termites and with presence or absence of added honey.  Model / response variable Fixed factor Odds ratio Log odds χ2 P Model A (CMM): Transects with bait couplets comparing specimen type Proportional predation risk Within couplets: Specimen type (treehopper) 0.31 -1.18 ± 0.17 48.79 <0.01  Between couplets: Honey on baits 0.25 -1.37 ± 0.17 63.63 <0.01 Model B (CMM): Transects with bait couplets comparing presence of honey Proportional predation risk Within couplets: Honey on baits 0.39 -0.94 ± 0.17 29.17 <0.01  Between couplets: Specimen type (treehopper) 0.27 -1.32 ± 0.19 45.07 <0.01       Model C (CMM): All transects Proportional predation risk Specimen type (treehopper) 0.30 -1.24 ± 0.12 100.06 <0.01 Honey on baits 0.32 -1.13 ± 0.12 87.76 <0.01  Transect type (specimen couplets) 1.29 0.26 ± 0.15 2.87 0.09 Model D (GLMM): Ant behavior towards baits, if not predated Probability of probing rather than abandoning Specimen type (treehopper) 1.19 0.18 ± 0.50 0.12 0.73 Honey on baits 3.89 1.35 ± 0.48 8.08 <0.01  Minutes after discovery 0.97 -0.03 ± 0.003 92.09 <0.01  Transect type (specimen couplets) 1.15 0.14 ± 0.51 0.08 0.78  61      Fig 5.3. Proportion of baits predated by ants over time following discovery. Lines represent means from 6 transects each with 10 couplets of each type.  When considering the behavior of ants that failed to predate on baits (Fig. 5.4), we found that ants were 3.9 times more likely to remain probing rather than abandon baits that originally had honey compared to those without (Fig 5.4; Table 5.1, model D). After reaching different maximum levels of predation, ants showed no difference in their behavior towards treehopper or termite baits (Fig 5.4; Table 5.1, model D). Also, as time passed, ants were less likely to stop probing, rather than abandon, baits with honey (Fig 5.4; Table 5.1, model D).   62   Fig 5.4. Ant behavior after locating different bait types in the Ecuadorian Amazon. Stacked bars represent the proportion of baits experiencing different outcomes from the ants’ behaviors. Predated baits are represented as an accumulated proportion over time (baits cannot become un-predated), whereas probed or abandoned baits are represented as the actual proportion experiencing each behavior at each time interval (baits were often abandoned or revisited over time). Sixty minutes after finding the baits, ants had already depleted the added honey. Proportions shown are averages of 12 transects, each consisting of 10 baits per bait type and honey or no-honey combination.     63  5.4 Discussion  By examining predation on live insect baits along an elevational gradient, we show that ant species known to participate in mutualistic associations with hemipterans were by far the most likely predators of insects below 1500 m (Fig. 5.2A, B). Here, potential ant partners were responsible for 50% to 100% of all recorded predation but got replaced by non-partner predators above 1500 m, albeit predation rates above this elevation were low (Fig. 5.2A, B). Interestingly, the decline in the contribution of ant partners to predation at higher elevations closely matched the decline in the incidence of mutualist treehoppers (Fig. 5.2B, C). We also found that ants were much less likely to predate upon baits consisting of treehoppers, or termite baits with added honey, than on control termite baits with no honey (Fig 5.3), with predation on the latter being comparable to that on live-fly baits. Moreover, even after depleting the honey, ants refrained from predating upon treehopper baits with or without honey or on termite baits that originally contained honey (Fig 5.3). Instead, ants tended to continue probing these baits rather than attack or abandon them (Fig 5.4).  The overwhelming contribution of ant partners to predation in the lowlands suggest mutualist treehoppers derive higher benefit from ant appeasement rather than their defensive services against the minority of predation exerted by non-ant predators (Fig. 5.2A, B). Treehoppers may derive higher benefits from the ants’ protection against the non-partner predation predominant at the highlands, although predation at those elevations was notably lower than in the lowlands. Nonetheless, these results suggest that whether hemipterans benefit from the appeasement or protection aspects of their association with ants is dependent on the ecological context of predation across elevations (Fig. 5.2A, B). Alternatively, ant appeasement can be considered the primary benefit throughout the entire elevational gradient given that low predation rates at high elevations would render the benefit obtained from the ants’ protection services negligible (Fig. 5.2A). The ant’s protective services could be further undermined by the fact that ants are not infallible defenders and, in some cases, can even exploit honeydew without providing protection (Yao, 2014; Zachariades et al., 2009). Finally, regardless of elevation, any protection services the ants may provide is contingent upon them not attacking hemipterans in the first place. Dissuading ants from predation may thus be the primary ecological challenge prompting hemipterans to  64  associate with the ants and may explain the close match between the incidence of mutualist treehoppers and the ants’ contribution to predation (Fig. 5.2B, C).  Our choice experiments, carried out at the lowest elevation, provided evidence of a protective effect of honeydew, as well as of partner recognition on the part of the ants. Thus, ants refrained from predating upon treehopper or termite baits laced with a sugary reward in 85% and 50% of cases, respectively. These results are comparable to observations on mutualism between ants and aphids (Hemiptera, Aphididae). Ants attending known partner aphid species have been shown to be dissuaded from predation 100% of the time, or 50% of the time if offered honeydew by a novel aphid partner (Sakata, 1994; Offenberg, 2001; Hayashi et al., 2015; Fischer et al., 2001).   In terms of partner recognition, given the opportunistic predatory nature of ants it is not surprising that they fed readily on termites. What is intriguing is that they avoided feeding on treehoppers even when these lacked a sugary reward. This suggests some level of recognition of the treehopper partners above and beyond their association with honeydew (Sakata, 1994; Offenberg, 2001; Yao, 2014; Endo. & Itino, 2013; Hojo et al., 2014). Here, hemipteran cuticular hydrocarbons may play a role by providing chemical crypsis and/or partner signalling, as shown in reports of treehoppers mimicking the chemical composition of their hostplants (Silveira et al., 2010), or aphids mimicking their ant partner’s chemical signaling (Endo. & Itino, 2013). Possible partner recognition through cuticular hydrocarbons or other means is suggestive of a co-evolved behavioral adaptation between the mutualistic partners or it may also be a learned response based on past interactions (Sakata, 1994; Yao, 2014; Hayashi et al., 2015; Endo. & Itino, 2013). Ants may also learn to recognize novel sources of honeydew (Choe & Rust, 2006; Dupuy et al., 2006; Hayashi et al., 2015; Endo. & Itino, 2013), as suggested by the ants’ reduced aggression towards termites laced with honey, even after honey had been depleted (Fig. 5.3, 5.4). Still, as honey was depleted, ants deciding to predate upon termites did so almost immediately, whereas predation on treehoppers occurred at a consistently lower rate over time, even after the honey had been depleted. The latter pattern may reflect an expectation of a sugary reward from treehoppers mediated by partner recognition, as ants have been found to be less aggressive towards partners they recognize (Sakata, 1994). Studying the behavior of mutualist ants towards baits consisting of various hemipteran taxa with various degrees of ant mutualism could further our insights on ant partner recognition and aggression reduction (Yao, 2014).  65   Just as in any other mutualistic interactions, however, ants should also be susceptible to turn into exploiters of their hemipteran partners either failing to provide any services in exchange for honeydew or even extorting the hemipterans for a richer reward (Ewald, 1987; Bronstein, 2001; Yao, 2014; Zachariades et al., 2009). Since ants are largely in control of the hemipteran’s enemy-free space, ants are in a position to demand a higher ecological investment from their partners in exchange for sparing their lives, in a sort of ecological extortion (Bronstein, 2001; Hoeksema & Bruna, 2000). Ants can cease to protect and even predate on hemipterans if a better sugar source is available and the demand for protein in the nest increases (Offenberg, 2001; Hayashi et al., 2015; Fischer et al., 2001; Billick et al., 2007). As a result, hemipterans may have their individual and population growth restricted as they may be forced to relinquish carbohydrate and amino acid absorption from sap to increase the attractiveness of honeydew (Stadler & Dixon, 1998; Yao et al., 2000; Fischer & Shingleton, 2001; Fischer et al., 2002; Yao & Akimoto, 2002; Kay et al., 2004; Fischer et al., 2005). For the ant-hemipteran mutualism to be stable, therefore, hemipterans need mechanisms to punish over-exploitative partners, which they may do by refusing to aggregate around non-cooperative ants (Bronstein, 2001; Hoeksema & Bruna, 2000; Johnston & Bshary; 2002). Thus, cooperative ants may be rewarded with increasing number of aggregating hemipterans, with populations of both partners entering a positive growth feedback as they benefit from each other’s services. The outcome is ants still predating upon their hemipteran partners, albeit with restraint, as ants consume nutrients that would otherwise be absorbed by hemipterans. In this process ants may have become the hypothesized ‘prudent predator’, a concept proposed by Slobodkin (1968) to imply a predator that exploits prey with the lowest possible impact to the growth or their populations to maximize their own long-term benefit (Slobodkin, 1968; Mertz & Wade, 1976; Van Baalen & Sabelis, 1995; Goodnight et al., 2008).  Whereas the existence of a mutualistic interaction between prey and their predator may appear counterintuitive, virulence theory has long recognized the potential for mutualistic interactions arising from antagonistic ones (Van Beneden, 1875; Hoeksema & Bruna, 2000; Lion & Boots, 2010). Applying the logic of virulence theory, the prediction would be that ants should be selected to behave mutualistically towards their hemipteran partners when their reproductive success depends on the persistence of the hemipteran populations and the services ants provide carry little cost (Matsuda & Shimada, 1993; Genkai-Kato & Yamamura, 1999; Hoeksema & Bruna, 2000).  66  Ants appear to meet these criteria. Ants are known to increase the persistence of tended hemipteran populations (Cushman & Whitham, 1991; Muller & Godfray, 1999; Flatt & Weisser, 2000; Quental et al., 2005; Gove & Rico-Gray, 2006), whose supply of honeydew constitutes a large portion of the nutritional demands of ants, especially in tropical canopies (Davidson et al., 2003; Bluthgen et al., 2000). Moreover, ants likely incur little cost while tending and defending their hemipteran partners, as the activities they perform are an extension of their regular foraging and territorial defense behaviors (Bluthgen et al., 2000; Dejean et al., 2007).   Ant-hemipteran associations illustrate the interesting possibility of a mutualistic association where prey offer a reward to predators –a ‘bribe’ – in exchange for their lives. Other documented cases of victim-exploiter mutualism include plants being pollinated or having their seeds dispersed by herbivores or cleaner fish feeding on ecto-parasites of potential predators (Poulin & Grutter, 1996; Soares et al., 2007; Bronstein et al., 2009; Irwin, 2010; Altermatt & Pearse, 2011; Revilla & Encinas-Viso, 2015). Ant-hemipteran mutualism, however, differs fundamentally from these others in that the main benefit obtained by prey is reduced aggression from the predator –the exploiter –itself, rather than other commodities such as dispersal or food.     67  Chapter 6. Conclusion  My thesis contributes to the field of macroecology by demonstrating that predation rates and the frequency of mutualistic associations decrease with elevation in the New World tropics.  I also tease apart possible mechanisms behind these patterns, which are likely ultimately linked to changes in temperature and productivity with elevation. Using manipulative experiments and field observations across 4000-meter elevational gradients in the equatorial Andes, I show that decreasing predation rates on arthropods at higher elevations are driven by decreasing predator abundance and activity. Ants were responsible for 80% of predation in the lowlands, but were replaced by other predators above 1500m, revealing a parallel qualitative gradient of predation. Along the same gradients, I show that the frequency of ant-hemipteran mutualistic associations also decreases with elevation, driven by a decrease in ant and hemipteran abundance. However, a surplus of ants at the lowlands and a subsequent shortage at the highlands sets an elevational limit to the mutualism and a gradient in the control dynamics between partners.  I show that anti-predator investment in hemipteran communities does not decrease at the same rate as predation rates at higher elevations, suggesting that the fitness effects of predation may decline at slower rate than predation rates. Moreover, hemipteran investment in ant-mutualism across elevations mirrored the contribution of ants to predation rather than their availability as mutualist partners supporting the idea of ant-hemipteran mutualism as an anti-predator strategy from ants themselves. Indeed, I provide evidence that, in response to pervasive ant predation in the lowlands, hemipterans may ‘bribe’ ants with honeydew primarily to dissuade them from predating upon them, rather than, as generally assumed, to obtain their defensive services. This is the first conceptualization of a mutualistic interaction based on prey offering a reward to predators in exchange for their lives, challenging our fundamental understanding of both mutualism and predation.  My analyses suggest that changes in temperature and productivity along elevational gradients regulate the frequency of biotic interactions, but not their ecological consequences on fitness. Temperature, light, water, and nutrients determine the amount of energy available in the environment and primary productivity is the basal biological consequence of these environmental variables (Clarke & Gaston, 2006; Brown et al., 2018). Here, temperature regulates the rate at  68  which biological process occur including the rate at which primary producers transform light, water and nutrients into biomass and this is transferred across trophic levels (Brown et al., 2004; Clarke & Gaston, 2006; Dell et al., 2011; Brown, 2014; Brown et al., 2018). To the degree that the effects of temperature are universal in all organisms, there is the possibility that the effects of temperature and productivity scale across trophic levels and the fitness effect of interacting species cancel out. Thus, while biotic interactions are more frequent at the tropics, Dobzhansky’s (1950) claim that they may be more important may not be the case.  Along the elevational gradient of the tropical Andes, temperature and productivity covary in a near perfect tandem, impeding the isolation of their individual effects in ecological processes (CITA). However, several gradients of precipitation with little variation in temperature occur at various points and elevations across the tropical Andes (e.g. the Tumbes-Chocó gradient, and the inter-Andean valley complex) and create ideal systems to tease apart their individual effects. Besides water, productivity is also a function of other factors such as the availability of light and nutrients, which also vary across the Andes. For instance, the topographical complexity of the Andes creates a mosaic of low-nutrient ridges and high-nutrient valleys, as well as light regimes based on slope aspect (CITAS). The large elevational gradient of the Andes combined with its broad range and heterogeneity of environmental conditions make them a valuable system upon a variety of macroecological hypotheses can be tested.  This study shows the potential in treehoppers as model systems to test hypotheses in macroecology as well as other fields of ecology and evolution. Treehoppers occur across a wide range of environmental conditions that expands beyond the elevational gradient of the Andes and includes all biomes across latitudes in America as well as many others in Africa, Australia and Eurasia. Treehoppers exhibit a high diversity of morphological and behavioral traits that occur along a continuum of social and mutualistic strategies. The range of environments and geographical regions upon which treehoppers have evolved and diversified, coupled with their range traits and adaptation grant treehoppers great potential as ecological models.  Most ecological research, however, focuses on a few vertebrate groups and rarely include other invertebrate groups such as treehoppers (Bonnet, 2002; Pawar, 2003; Stahlschmidt, 2011; Zuk et al., 2014; Rosenthal et al., 2017). This may limit our ability to assess the generality of ecological  69  understanding as it is based on a limited sample of clades and traits. Moreover, underrepresented taxa are mostly comprised by invertebrates and microorganisms, which comprise most of the biomass in an ecosystem and likely responsible of most of its function (CITA). Thus, findings from ecological research focusing on these organisms may be more informative about broad ecological processes. Ecological research in underrepresented groups is often limited by the lack of knowledge in their natural history and taxonomy (Bonnet, 2002; Pawar, 2003), emphasizing the importance of this type of research on these organisms.                70  References  Abrams, P. A. (1992). Why predation rate should not be proportional to predator density. Ecology, 74, 726–733. Abrams, P. & Matsuda, H. (1993). Effects of adaptive predatory and anti-predator behaviour in a two-prey-one-predator system. Evolutionary Ecology, 7, 312-326. Addo-Bediako, A., Chown, S. L. & Gaston, K. J. (2002). Metabolic cold adaptation in insects: a large-scale perspective. Functional Ecology, 16, 332–338. Altermatt, F. & Pearse, I. S. (2011). Similarity and specialization of the larval versus adult diet of european butterflies and moths. The American Naturalist, 178, 372–382. Anderson, K. J., & Jetz, W. (2005). The broad-scale ecology of energy expenditure of endotherms. Ecology Letters, 8, 310–318. Bascompte, J. (2019). Mutualism and biodiversity. Current Biology, 29, 425 – 473. Bates, D., Maechler, M., Bolker, B., & Walker, S. (2015). Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software, 67, 1–48. Beckerman, A. P., Ridgers, G. M. & Dennis, S. R. (2010). The reaction normof size and age at maturity under multiple predator risk. Journal of Animal Ecology, 79, 1069–1076 Bertness, M. D., Garrity, S. D., & Levings, S. C. (1981). Predation pressure and gastropod foraging: a tropical-temperate comparison. Evolution, 35, 995–1007. Billick, I., Hammer, S., Reithel, J. S., & Abbot, P. (2007). Ant–Aphid Interactions: Are Ants Friends, Enemies, or Both?. Annals of the Entomological Society of America, 100, 887-892. Bluthgen, N., Cottsberger, G. & Fiedler, K. (2004). Sugar and amino acid composition of ant-attended nectar and honeydew sources from an Australian rainforest. Austral Ecology, 29, 418–429. Bluthgen, N. & Fiedler, K. (2004). Preferences for sugars and amino acids and their conditionality in a diverse nectar-feeding ant community. Journal of Animal Ecology, 73, 155–166. Bluthgen, N., Mezger, D., & Linsenmair, K. E. (2006). Ant-hemipteran trophobioses in a Bornean rainforest – diversity, specificity and monopolization. Insectes Sociaux, 194–203. Bluthgen, N., Verhaag, M., Goitia, W., Jaffe, K., Morawetz, W. & Barthlott, W.   (2000). How plants shape the ant community in the Amazonian rainforest canopy: the key role of extrafloral nectaries and homopteran honeydew. Oecologia, 125, 229 – 240.  71  Boucher, D. H., James, S. & Keeler, K. H. (1982). The ecology of mutualism. The Annual Review of Ecology, Evolution, and Systematics, 13, 315–347. Boyle, A. W. (2008). Can variation in risk of nest predation explain altitudinal migration in tropical birds?. Oecologia, 155, 397–403. Boyles, J. G., Seebacher, F., Smit, B., & McKechnie, E. (2011). Adaptive thermoregulation in endotherms may alter responses to climate change. Integrative and Comparative Biology, 51, 676–690. Brandao, C. R. F., Silva, R. R. & Delabie, J. H. C. (2012). Neotropical ants (Hymenoptera) functional groups: nutritional and applied implications. Insect Bioecology and Nutrition for Integrated Pest Management, 213-236. Bronstein, J. L. (1994). Our current understanding of mutualism. The Quarterly review of Biology, 69, 31 – 51. Bronstein, J. L. (2001). The exploitation of mutualism. Ecology Letters, 4, 277 – 287. Bronstein, J., Huxman, T., Horvath, B., Farabee, M. & Davidowitz, G. (2009). Reproductive biology of Datura wrightii: the benefits of a herbivorous pollinator. Annals of Botany, 103, 1435–1443. Brooks, M. E., Kristensen, K., Benthem, K. J., Magnusson, A., Berg, C. W., Nielsen, A., Skaug, H. J., Maechler, M. & Bolker, B. M. (2017). glmmTMB Balances Speed and Flexibility Among Packages for Zero-inflated Generalized Linear Mixed Modeling. The R Journal, 9, 378-400. Brown, J. H. (2014). Why are there so many species in the tropics?. Journal of Biogeography, 41, 8–22. Brown, J. H., Gillooly, J. F., Allen, A. P., Savage, V. M. & West, G. B. (2004). Toward a metabolic theory of ecology. Ecological Society of America, 85, 1771–1789. Brown, J. H., Hall, C. A. S., & Sibly, R. M. (2018). Equal fitness paradigm explained by a trade-off between generation time and energy production rate. Nature Ecology & Evolution, 2, 262–268. Bruhl, C. A., Mohamed, M., & K. E. Linsenmair. (1999). Altitudinal distribution of leaf litter and along a transect in primary forests on Mount Kinabalu, Sabah, Malaysia. Journal of Tropical Ecology, 15, 265–277.  72  Buckley, L. B., Hurlbert, A. H. & Jetz, W. (2012). Broad-scale ecological implications of ectothermy and endothermy in changing environments. Global Ecology and Biogeography, 21, 873–885. Burns, K. C. (2004). Scale and macroecological patterns in seed dispersal mutualisms. Global Ecology and Biogeography, 13, 289–293. Callaway, R. M., Brookerm R. W., Choler, P., Kikvidze, Z, Lortie, C. J., Michalet, R., Paolini, L., Pugnaire, F. I., ANewingham, B. & Aschehougook, B. J. (2002). Positive interactions among alpine plants increase with stress. Nature, 417, 844 – 848. Camacho, F. L. & Avilés, L. (2019). Decreasing predator density and activity explains declining predation of insect prey along elevational gradients. The American Naturalist, 194, 334-343. Camacho, L., Keil, C. & Dangles, O. (2014). Factors influencing egg parasitism in sub-social insects: insights from the treehopper Alchisme grossa (Auchenorrhyncha, Membracidae). Ecological Entomology, 39, 58-65. Chamberlain, S. A. & Holland, J. N. (2009). Quantitative synthesis of context dependency in ant–plant protection mutualisms. Ecology, 90, 2384–2392. Choe, D. & Rust, M. K. (2006). Homoptera chemical signatures reduce aggression of tending ants. Chemoecology, 16, 175–178. Chomicki, G. & Renner, S. S. (2017). Partner abundance controls mutualism stability and the pace of morphological change over geologic time. Proceedings of National Academy of Science, 114, 3951–3956. Cryan, J. R. & Urban, J. M. (2012) Higher-level phylogeny of the insect order Hemiptera: is Auchenorrhyncha really paraphyletic?. Systematic Entomology, 37: 7–21. Cryan, J. R., Wiegmann, B. M., Deitz, L. L., Dietrich, C. H. & Whiting, M. F. (2004). Treehopper trees: phylogeny of Membracidae (Hemiptera: Cicadomorpha: Membracoidea) based on molecules and morphology. Systematic Entomology, 29, 441–454. Cushman, J. H. & Whitham, T. G. (1991). Competition mediating the outcome of a mutualism: protective services of ants as a limiting resource for membracids. The American Naturalist, 138, 851–865.  Davidson, D. W. (1997). The role of resource imbalances in the evolutionary ecology of tropical arboreal ants. Biological Journal of the Linnean Society, 61, 153-181.  73  Davidson, D. W. (1998). Resource discovery versus resource domination in ants: a functional mechanism for breaking the trade-off. Ecological Entomology, 23, 484–490. Davidson, D. W., Cook, S. C., Snelling, R. R., & Chua, T. H. (2003). Explaining the abundance of ants in lowland tropical rainforest canopies. Science, 300, 969–972. Dejean, A., Corbara, B., Orivel, J. & Leponce, M. (2007). Rainforest canopy ants: the implications of territoriality and predatory behavior. Functional Ecosystems and Communities, 1, 105–120. Delabie, J. H. C. (2001). Trophobiosis Between Formicidae and Hemiptera (Sternorrhyncha and Auchenorrhyncha): an Overview. Neotropical Entomology, 30, 501-516. Del-Claro, K. & Oliveira, P. S. (1996). Honeydew flicking by treehoppers provides cues to potential tending ants. Animal Behavior, 51, 1071–1075. Del-Claro, K. & Oliveira, P. S. (2000). Conditional outcomes in a neotropical treehopper-ant association: temporal and species-specific variation in ant protection and homopteran fecundity. Oecologia, 124, 156–165. Del Toro, I., Ribbons, R. R., and S. Pelini. (2012). The little things that run the world revisited: a review of ant-mediated ecosystem services and diservices (Hymenoptera: Formicidae). Myrmecological News, 17, 133–146. Dell, A. I., Pawar, S., & Savage, V. M. (2011). Systematic variation in the temperature dependence of physiological and ecological traits. Proceedings of the National Academy of Science, 108, 10591–10596. Dell, A. I., Pawar, S., & Savage, V. M. (2014). Temperature dependence of trophic interactions are driven by asymmetry of species responses and foraging strategy. Journal of Animal Ecology, 83, 70–84. Dietrich, C. H., McKamey, S. H., & Deitz, L. L. (2001). Morphology-based phylogeny of the treehopper family Membracidae (Hemiptera: Cicadomorpha: Membracoidea). Systematic Entomology, 26, 213-239 Dobzhansky, T. (1950). Evolution in the tropics. American Scientist, 38, 209-221. Doebeli, M. & Knowlton, N. (1998). The evolution of interspecific mutualisms. Proceedings of the National Academy of Sciences, 95, 8676–8680. Dupuy, F., Sandoz, J., Giurfa, M. & Josens, R. (2006). Individual olfactory learning in Camponotus ants. Animal Behaviour, 72, 1081–1091. Endo, S. & Itino, T. (2013). Myrmecophilous aphids produce cuticular hydrocarbons that resemble those of their tending ants. Population Ecology, 55, 27–34.  74  Evangelista, O., Sakakibara, A. M., Cryan, J. R. & Urban, J. M. (2016). A phylogeny of the treehopper subfamily Heteronotinae reveals convergent pronotal traits (Hemiptera: Auchenorrhyncha: Membracidae). Systematic Entomology, 42, 410-428. Ewald, P. W. (1987). Transmission modes and evolution of the parasitism-mutualism continuum. Annals of the New York Academy of Sciences, 503, 295–306. Fagundes, R., Sibeiro, S. P. & Del-Claro, K. (2013). Tending-ants increase survivorship and reproductive success of Calloconophora pugionata Drietch (Hemiptera, Membracidae), a trophobiont herbivore of Myrcia obovata O.Berg (Myrtales, Myrtaceae). Sociobiology, 60, 11-19. Fischer, M. K., Hoffmann, K. H. & Volkl, W. (2001). Competition for mutualists in an ant–homopteran interaction mediated by hierarchies of ant attendance. Oikos 92, 531–541. Fischer, M. K. & Shingleton, W. (2001). Host plant and ants influence the honeydew sugar composition of aphids. Functional Ecology, 15, 544–550. Fischer, M. K., Volkl, W., Schopf, R. & Hoffmann, K. H. (2002). Age-specific patterns in honeydew production and honeydew composition in the aphid Metopeurum fuscoviride: implications for ant-attendance. Journal of Insect Physiology, 48, 319–326. Fischer, M. K, Volkl, W. & Hoffmann, K. H. (2005). Honeydew production and honeydew sugar composition of polyphagous black bean aphid, Aphis fabae (Hemiptera: Aphididae) on various host plants and implications for ant-attendance. European Journal of Entomology, 102, 155-160. Flatt, T. & Weisser, W. W. (2000). The effects of mutualistic ants on aphid life history traits. Ecology, 81, 3522–3529. Floren, A., Biun, A. & Linsenmair, E. K. (2002). Arboreal ants as key predators in tropical lowland rainforest trees. Oecologia, 131, 137–144. Fox, J. & Weisberg, S. (2011). An {R} Companion to Applied Regression. Sage, Thousand Oaks, CA, USA. Fox, J. & Weisberg, S. (2019). An {R} Companion to Applied Regression. Thousand Oaks, pp. 1–608. Frazier, M. R., Huey, R. B. & Berrigan, D.  (2006). Thermodynamics constrains the evolution of insect population growth rates: “warmer is better”. The American Naturalist, 168, 512–520. Fretwell, S. D. (1977). The regulation of plant communities by the food chains exploiting them. Perspectives in Biology and Medicine, 20, 169–185.  75  Fretwell, S. D. (1987). Food chain dynamics: the central theory of ecology?. Oikos, 50, 291–301 Garreaud, R. D. (2009). The Andes climate and weather. Advances in Geosciences, 22, 3–11. Gaston, K. J. & Blackburn, T. M. (1999) A critique for macroecology. Oikos, 84: 353–-368. Geib, J. C. & Galen, C. (2012). Tracing impacts of partner abundance in facultative pollination mutualisms: from individuals to populations. Ecology, 93, 1581–1592. Genkai-Kato, M. & Yamamura, N. (1999). Evolution of mutualistic symbiosis without vertical transmission. Theoretical Population Biology, 55, 309–323.  Gillette, P. N., Ennis, K. K., Domínguez Martínez, G. & Philpott, S. M. (2015). Changes in Species Richness, Abundance, and Composition of Arboreal Twig-nesting Ants Along an Elevational Gradient in Coffee Landscapes. Biotropica, 47, 712–722. Gillman, L. N., Wright, S. D., Cusens, J., McBride, P. D., Malhi, Y. & Whittaker, R. J. (2015). Latitude, productivity and species richness. Global Ecology and Biogeography, 24, 107–117 Girardin, C. A. J., Malhi, Y., Aragao, L. E. O. C., Mamani, M., Huasco, W. H., Durand, L., Feeley, K., Rapp, J., Silva-Espejo, J. E., Silman, M., Salinas, N. & Whittaker, J. (2010). Net primary productivity allocation and cycling of carbon along a tropical forest elevational transect in the Peruvian Andes. Global Change Biology, 16, 3176–3192. Godoy, C., Miranda, X. & Nishida, K. (2006). Treehoppers of Tropical America. Instituto Nacional de Biodiversidad, San José, Costa Rica. Goodnight, C., et al. (2008). Evolution in spatial predator–prey models and the ‘‘prudent predator’’: the inadequacy of steady-state organism fitness and the concept of individual and group selection. Complexity, 13, 23–44. Gove, A. D. & Rico-Gray, V. (2006). What determines conditionality in ant – Hemiptera interactions? Hemiptera habitat preference and the role of local ant activity. Ecological Entomology, 31, 568–574. Gripenberg, S. & Roslin, T. (2007). Up or down in space? Uniting the bottom-up versus top-down paradigm and spatial ecology. Oikos, 116, 181-188 Grosso, S., Parton, W., Stohlgren, T., Zheng, D., Bachelet, D., Prince, S., Hibbard, K. & Olson, R. (2008). Global potential net primary production predicted from vegetation class, precipitation, and temperature. Ecology, 89, 2117–2126. Guevara, J. & Avilés, L.  (2015). Ecological predictors of spider sociality in the Americas.  Global Ecology and Biogeography, 24, 1181-1191.  76  Hadley, W., Romain F., Lionel H. & Müller, K. (2020). dplyr: A Grammar of Data Manipulation. R package version 0.8.5. In: https://CRAN.R-project.org/package=dplyr Hawkins, B. A., Cornell, H. V. & Hochberg, M. E. (1997). Predators, parasitoids, and pathogens as mortality agents in phytophagous insect populations. Ecology, 78, 2145–2152. Hayashi, M., Nakamuta, K. & Nomura, M. (2015). Ants learn aphid species as mutualistic partners: is the learning behavior species-specific?. Journal of Chemical Ecology, 41, 1148–1154.  Henschel, J. R. (1998). Predation on social and solitary individuals of the spider Stegodyphus dumicola (Araneae, Eresidae). Journal of Arachnology, 26, 61–69. Hijmans, R.J., Cameron, S.E., Parra, J.L., Jones, P.G. & Jarvis, A. (2005). Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology, 25, 1965-1978. Hodkinson, I. D. (2005). Terrestrial insects along elevation gradients: Species and community responses to altitude. Biological Reviews of the Cambridge Philosophical Society, 80, 489–513. Hoeksema, J. D. & Bruna, E. M. (2000). Pursuing the big questions about interspecific mutualism: a review of theoretical approaches. Oecologia, 125, 321–330. Hoffman, C.R.  & Avilés, L. (2017). Rain, predators, and spider sociality: a manipulative experiment.  Behavioural Ecology, 28, 589-596. Hojo, M. K., Yamamoto, A., Akino, T., Tsuji, K. & Yamaoka, R. (2014). Ants use partner specific odors to learn to recognize a mutualistic partner. PLoS ONE, 9, e86054, https://doi.org/10.1371/journal.pone.0086054 Holland, J. N., DeAngelis, D. L. & Bronstein, J. L. (2002). Population dynamics and mutualism: functional responses of benefits and costs. The American Naturalist, 159, 231–244. Holland, J. N. & DeAngelis, D. L. (2010). A consumer–resource approach to the density-dependent population dynamics of mutualism. Ecology, 91, 1286–1295. Holling, C. S. (1959). The components of predation as revealed by a study of small-mammal predation of the European pine sawfly. Canadian Entomology, 91, 293–320. Holt, R. D. & Lawton, J. H. (1994). The ecological consequences of shared natural enemies. Annual Review of Ecology and Systematics, 25, 495–520. Holt, R. D. (2009). Bringing the Hutchinsonian niche into the 21st century: Ecological and evolutionary perspectives. Proceedings of the National Academy of Sciences, 106, 19659–19665. Hunter, M. D. & Price, P. W. (1992). Playing chutes and ladders: heterogeneity and the relative roles of bottom-up and top-down forces in natural communities. Ecology, 73, 724-732.  77  Hutchinson, G. E. (1957). Concluding remarks. Cold Spring Harbor Symposium, 22, 415–427. Irwin, R. E. (2010). Evolutionary ecology: when pollinators are also herbivores. Current Biology, 20, 100–101. Jeanne, R. L. (1979). A latitudinal gradient in rates of ant predation. Ecology, 60, 1211–1224. Jeffries, M. J. & Lawton, J. H. (1984). Enemy free space and the structure of ecological communities. Biological Journal of the Linnean Society, 23, 269–286. Johnston, R. A. & Bshary, R. (2002). From parasitism to mutualism: partner control in asymmetric interactions. Ecology Letters, 5, 634–639. Johnston, R. A. & Bshary, R. (2008). Mutualism, market effects and partner control. Journal of Evolutionary Biology, 21, 879–888. Kamil, B. (2020). MuMIn: Multi-Model Inference. R package version 1.43.17. In: https://CRAN.R-project.org/package=MuMIn Kaspari, M., O’Donnell, S. & Kercher, J. R.  (2000). Energy, Density, and Constraints to Species Richness: Ant Assemblages along a Productivity Gradient. The American Naturalist, 155, 280–293. Kaspari, M., Yanoviak, S. P. & Dudley, R. (2008). On the biogeography of salt limitation: A study of ant communities. Proceedings of the National Academy of Science, 105, 17848–17851. Kassambara, A. (2019). ggpubr: 'ggplot2' based publication ready plots, R package version 0.2.1. The Comprehensive R Archive Network. In:  https://CRAN.R-project.org/package=ggpubr  Kay, A. D., Scott, S. E., Schade, J. D. & Hobbie, S. E. (2004) Stoichiometric relations in an ant-treehopper mutualism. Ecology Letters, 7: 1024–1028. Keith, S. A., Webb, T. J., Bohning-Gaese, K., Conolly, S. R., Dulvy, N. K., Eigenbrod, F., Jones, K. E., Price, T., Redding, D. W., Owens, I. P. F. & Isaac, N. J. B. (2012) What is macroecology?. Biology Letters, doi:10.1098/rsbl.2012.0672. Koptur, S. (1984). Experimental evidence for defense of Inga (Mimosoideae) saplings by ants. Ecology, 65, 1787-1793. Korner, C. (2007). The use of ‘altitude’ in ecological research. Trends in Ecology and Evolution, 22, 569–574. Lawton, J. H. (1999). Are There General Laws in Ecology?. Oikos, 84, 177–192. Leigh, E. G. (2010). The evolution of mutualism. Journal of Evolutionary Biology, 23, 2507–2528.  78  Lieth, H. (1975). Modeling the primary productivity of the world. In: Primary productivity of the biosphere (eds. Lieth, H. and Whittaker R. H.). Springer-Verlag, New York, New York, USA. Lima, S. L. & Dill, L. M. (1990). Behavioral decisions made under the risk of predation: a review and prospectus. Canadian Journal of Zoology, 68, 619–640. Lin, C., Danforth, D. N., & Wood, T. K. (2004). Molecular phylogenetics and evolution of maternal care in membracine treehoppers. Systematic Biology, 53, 400–421. Lind, J. & Cresswell, W. (2005). Determining the fitness consequences of antipredation behavior. Behavioral Ecology, 16, 945–956 Lion, S. & Boots, M. (2010). Are parasites “prudent” in space?. Ecology Letters, 13, 1245–1255.  Longino, J. T. & Colwell, R. K. (2011). Density compensation, species composition, and richness of ants on a neotropical elevational gradient. Ecosphere, 2, 1–20. Longino, J. T., Branstetter, M. G. & Colwell, R. K. (2014). How ants drop out: ant abundance on tropical mountains. Plos One, 9, 1–9. Louthan, A. M., Doak, D. F. & Angert, A. L. (2015). Where and when do species interactions set range limits?. Trends in Ecology and Evolution, 30, 780–792. Lövei, G. L. & Ferrante, M. (2017). A review of the sentinel prey method as a way of quantifying invertebrate predation under field conditions. Insect Science, 24, 528–542. Low, P. A., Sam, K., McArthur, C., Posa, M. R. C. & Hochuli, D. F. (2014). Determining predator identity from attack marks left in model caterpillars: Guidelines for best practice. Entomologia Experimentalis et Applicata, 152, 120–126. Lüdecke, D. (2020). _sjstats: Statistical Functions for Regression Models (Version 0.18.0)_. doi: 10.5281/zenodo.1284472 (URL:http://doi.org/10.5281/zenodo.1284472), <URL: https://CRAN.R-project.org/package=sjstats> Malhi, Y., Silman, M., Salinas, N., Bush, M., Meir, P. & Saatchi, S. (2010). Introduction: Elevation gradients in the tropics: Laboratories for ecosystem ecology and global change research. Global Change Biology, 16, 3171–3175. Matsuda, H. & Shimada, M. (1993). Cost-benefit model for the evolution of symbiosis. In: Mutualism and Community Organization (ed. Kawanabe, H., Cohen, J. E. & Iwasay, K.). Oxford University Press, Oxford, United Kingdom, pp. 228–238. Maunsell, S. C., Kitching, R. L., Burwell, C. J. & Morris, R. J. (2015). Changes in host-parasitoid food web structure with elevation. Journal of Animal Ecology, 84, 353–363.  79  May, R. M. (1974). On the theory of niche overlap. Theoretical population biology, 5, 297-332. McKinnon, S. E. (2019). lemon: freshing up your 'ggplot2' plots. R package version 0.4.3. The Comprehensive R Archive Network. In: https://CRAN.R-project.org/package=lemon  Mertz, D. B. & Wade, M. J. (1976). The prudent prey and the prudent predator. The American Naturalist, 110, 489–496. Moles, A. T., Bonser, S. P. Poore, A. G. B., Wallis, I. R. & Folley, W. J. (2011). Assessing the evidence for latitudinal gradients in plant defence and herbivory. Functional Ecology, 25, 380–388. Mooney, K. A. (2007). Tritrophic effects of birds and ants on a canopy food web, tree growth, and phytochemistry. Ecology, 88, 2005–2014. Morales, M. A. (2000). Mechanisms and density dependence of benefit in an ant-membracid mutualism. Ecology, 81, 482–489. Moreau, C. S., Bell, C. D., Vila, R., Archibald, B. S. & Pierce, N. E. (2006). Phylogeny of the ants: diversification in the age of angiosperms. Science, 312, 101–103. Morris, R. J., Sinclair, F. H. & Burwell, C. J. (2015). Food web structure changes with elevation but not rainforest stratum. Ecography, 38, 792–802. Moya-Raygoza, G., Albarracin, E. L. & Virla, E. G. (2012). Diversity of Egg Parasitoids Attacking Dalbulus maidi (Hemiptera: Cicadellidae) Populations at Low and High Elevation Sites in Mexico and Argentina. Florida Entomologist, 95, 105-112. Muller, C. B. & Godfray, H. C. J. (1999). Predators and mutualists influence the exclusion of aphid species from natural communities. Oecologia, 119, 120–125. Munch, S. B., Snover, M. L., Watters, G. M. & Mangel, M. (2005). A unified treatment of top-down and bottom-up control of reproduction in populations. Ecology Letters, 8, 691–695 Nisbet, R. M., Diehl, S., Wilson, W. G. & Cooper, S. D. (1997). Primary-productivity gradients and short-term population dynamics in open systems. Ecological Monographs, 67, 535–553. Noe, R. & Hammerstein, P. (1994). Biological markets: supply and demand determine the effect of partner choice in cooperation, mutualism and mating. Behavioral Ecology and Sociobiology, 35, 1 – 11. Novotny, V., Basset, Y., Auga, J., Boen, W., Dal, C., Drozd, P., Kasbal, M., Isua, B., Kutil, R. & Manumbor, M. (1999). Predation risk for herbivorous insects on tropical vegetation: A search for enemy-free space and time. Austral Ecology, 24, 477–483.  80  O’Connor, M. I. (2009). Warming Strengthens an Herbivore: Plant Interaction. Ecology, 90, 388–398. O’Connor, M. I., Piehler, M. F., Leech, D. M., Anton, A. & Bruno, J. F. (2009). Warming and resource availability shift food web structure and metabolism. Plos Biology, 7, 3–8. Oda, F. H., Oliveira, A. F. & Aoki, C. (2014) Oxytrigona tataira (Smith) (Hymenoptera: Apidae: Meliponini) as a collector of honeydew from Erechtia carinata (Funkhouser) (Hemiptera: Membracidae) on Caryocar Brasiliense. Sociobiology, 61: 566-569 Cambessèdes (Malpighiales: Caryocaraceae) in the Brazilian Savanna Offenberg, J. (2001). Balancing between mutualism and exploitation: the symbiotic interaction between Lasius ants and aphids. Behavioral Ecology and Sociobiology, 49, 304–310  Oksanen, L., Fretwell, S. D., Arruda, J. & Niemela, P. (1981). Exploitation ecosystems in gradients of primary productivity. The American Naturalist, 118, 240–261. Oliver, T. H., Leather, S. R. & Cook, J. M. (2008). Macroevolutionary patterns in the origin of mutualisms involving ants. Journal of Evolutionary Biology, 21, 1597–1608. Olmstead, K. L. & Wood, T. K. (1990). Altitudinal patterns in species richness of neotropical treehoppers (Homoptera: Membracidae): the role of ants. Proceedings of the Entomological Society of Washington, 92, 552-560. Olmstead, K. L. & Wood, T. K. (1990b). The effect of clutch size and ant attendance on egg guarding by Entylia bactriana. Psyche, 97, 111-120. Olson D. M. (1992). Rates of predation by ants (Hymenoptera: Formicidae) in the canopy, understory, leaf litter, and edge habitats of a lowland rainforest in Southwestern Cameroon. In: Biologie D’une Canopée De Forêt Équatoriale—II. Rapport De Mission: Radeau Des Cimes Octobre Novembre 1991 (eds. Hallé, F. & Pascal, O.). Réserve De Campo, Cameroun. Fondation Elf, Paris, pp.101–109. Pasteels, J. M. & Gregoire, J. (1983). The chemical ecology of defense in arthropods. Annual Reviews of Entomology, 28, 263-289. Pemberton, R. W. (1998). The occurrence and abundance of plants with extrafloral nectaries, the basis for antiherbivore defensive mutualism, along a latitudinal gradient in east Asia. Journal of Biogeography, 25, 661 – 668. Pennings, S. C. & Silliman, B. R. (2005) Linking biogeography and community ecology: latitudinal variation in plant-herbivore interaction strength. Ecology, 86: 2310–2319.  81  Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D.  & R Core Team (2018). _nlme: Linear and Nonlinear Mixed Effects Models_. R package version 3.1-137, <URL: https://CRAN.R-project.org/package=nlme>. Pinto, C. F., Salinas, S., Flores-Prado, L., Echeverría, J. & Niemeyer, H. N. (2016). Sequestration of tropane alkaloids from Brugmansia suaveolens (Solanaceae) by the treehopper Alchisme grossa (Hemiptera: Membracidae). Biochemical Systematics and Ecology, 66, 161-165. Piñol, J., Espadaler, X., Cañellas, N., Martínez-Vilalta, J., Barrientos, J. A. & Sol, D. (2010). Ant versus bird exclusion effects on the arthropod assemblage of an organic citrus grove. Ecological Entomology, 35, 367–376. Plowman, N. S., Hood, A. S. C., Moses, J., Redmond, C., Novotny, V., Klimes, P. & Fayle, T. M. (2017). Network reorganization and breakdown of an ant–plant protection mutualism with elevation. Proceedings of the Royal Society B, 284, 1-10.  Poulin, R. & Grutter, A. (1996). Cleaning symbioses: proximate and adaptive explanations. BioScience, 46, 512–517. Power, M. E. (1992). Top-down and bottom-up forces in food webs: do plants have primacy?. Ecology, 73, 733–746. Preisser, E. L., Bolnick, D. I. & Benard, M. F. (2005). Scared to death? The effects of intimidation and consumption in predator-prey interactions. Ecology, 86, 501–509. Purcell, J. & Avilés, L. (2008). Gradients of precipitation and ant abundance may contribute to the altitudinal range limit of subsocial spiders. Proceedings of the Royal Society of London B, 275, 2617–2625. Quental, T. B., Trigo, J. R. & Oliveira, P. S. (2005). Host-plant flowering status and the concentration of sugar in phloem sap: effects on an ant-treehopper interaction. European Journal of Entomology, 102, 201–208. R Core Team (2019). R: A language and Environment for Statistical Computing. R Foundation for Statistical Computing. Reithel, J. S. & Campbell, D. R. (2008). Effects of aggregation size and host plant on the survival of an ant-tended membracid (Hemiptera: Membracidae): potential roles in selecting for generalized host plant use. Annals of the Entomological Society of America, 101, 70-78. Revilla, T. & Encinas-Viso, F. (2015). Dynamical transitions in a pollination– herbivory interaction: a conflict between mutualism and antagonism. PLoS ONE, 10, e0117964; https://doi.org/10.1371/journal.pone.0117964   82  Rodríguez-Castañeda, G., Brehm, G., Fiedler, K. & Dyer, L. A. (2016). Ant predation on herbivores through a multitrophic lens: how effects of ants on plant herbivore defense and natural enemies vary along temperature gradients. Current Opinion in Insect Science, 14, 73–80. Roslin, T., Andrew, N. R., Asmus, A., Barrio, I. C. & Basset, Y. (2017). Higher predation risk for insect prey at low latitudes and elevations. Science, 356, 742–744. Rosumek, F. B., Silveira, F. A. O., Neves, F. S., Barbosa, N. P. U., Diniz, L., Oki, Y., Pezzini, F., Fernandes, G. W. & Cornelissen, T. (2009). Ants on plants: a meta-analysis of the role of ants as plant biotic defenses. Oecologia, 160, 537–549. Sakata, H. (1994). How an ant decides to prey on or to attend aphids. Researches on Population Ecology, 36, 45 – 51. Sam, K., Koane, B. & Novotny, B. (2015). Herbivore damage increases avian and ant predation of caterpillars on trees along a complete elevational forest gradient in Papua New Guinea. Ecography, 38, 293-300. Samson, D. A. (1997). Ant diversity and abundance along an elevational gradient in the Philippines. Biotropica, 29, 349–363. Santos, E. S. A., Bueno, P. P., Gilbert, J. D. J. & Machado, G. (2017). Macroecology of parental care in arthropods: higher mortality risk leads to higher benefits of offspring protection in tropical climates. Biological Reviews, 92, 1688–1701. Sarmiento, G. (1986). Ecologically crucial features of climate in high tropical mountains. In: High altitude tropical biogeography (eds. Vuilleumier, F. & Monasterio, M.). Oxford University Press, Oxford, United Kingdom, pp. 11–45. Savage, V. M., Gillooly, J. F., Brown, J. H., West, G. B. & Charnov, E. L. (2004). Effects of body size and temperature on population growth. The American Naturalist, 163, 429–441. Schemske, D. W., Mittelbach, G. G., Cornell, H. V., Sobel, J. M. & Roy, K. (2009). Is there a latitudinal gradient in the importance of biotic interactions?. Annual Review of Ecology, Evolution, and Systematics, 40, 245–269. Schoner, C. R., Schoner, M. G., Kerth, G. & Grafe, T. U. (2013). Supply determines demand: influence of partner quality and quantity on the interactions between bats and pitcher plants. Oecologia, 173, 191–202.  83  Schramski, J. R., Dell, A. I., Grady, J. M., Sibly, R. M. & Brown, J. M. (2015). Metabolic theory predicts whole-ecosystem properties. Proceedings of the National Academy of Science, 112, 2617–2622 Sierra, R. (1999). Propuesta Preliminar de un Sistema de Clasificación de Vegetación para el Ecuador Continental. INEFAN/GEF-BIRF y EcoCiencia, pp. 1–174. Sih, A. (1994). Predation risk and the evolutionary ecology of reproductive ecology. Journal of Fish Biology, 45, 111–130. Sih, A., Crowley, P., McPeek, M., Petranka, J. & Strohmeier, K. (1985). Prey, competition, and prey communities: a review of field experiments. Annual Reviews of Ecology and Systematics, 16, 269–311. Singer, M. N. & Stireman, J. O. (2005). The tri-trophic niche concept and adaptive radiation of phytophagous insects. Ecology Letters, 8, 1247–1255. Slobodkin, L. B. (1968). How to be a predator. American Zoologist, 8, 43–51. Soderstrom B. (1999). Artificial nest predation rates in tropical and temperate forests: a review of the effects of edge and nest site. Ecography, 22, 455–63. Solomon, M. E. (1949). The natural control of animal populations. Journal of Animal Ecology, 18, 1–35. Stachowicz, J. J. (2001). Mutualism, facilitation, and the structure of ecological communities. BioScience, 51, 235 – 246. Stadler, B. & Dixon, A. F. G. (1998). Costs of Ant Attendance for Aphids. Journal of Animal Ecology, 67, 454-459 Stadler, B. & Dixon, A. F. G. (2005). Ecology and evolution of aphid-ant interactions. Annual Review of Ecology, Evolution, and Systematics, 36, 345–72. Stadler, B., Fiedler, K., Kawecki, T. J. & Weisser, W. W. (2001). Costs and benefits for phytophagous myrmecophiles: when ants are not always available. Oikos, 92, 467–478. Stanton, M. L. (2003). Interacting guilds: moving beyond the pairwise perspective on mutualisms. The American Naturalist, 162, 10-23. Stearns, S. C. (1989). Trade-offs in life-history evolution. Functional Ecology, 3, 259-268. Stireman, J. O., Dyer, L. A., Janzen, D. H., Singe,r M. S., Lill J. T., Marquis, R. J., Ricklefs, R. E., Gentry, G. L., Hallwachs, W., Coley, P. D., Barone, J. A., Greeney, H. F., Connahs, H. F., Barbosa, P., Morais, H. C. & Diniz, I. R.  (2005). Climatic unpredictability and  84  parasitism of caterpillars: implications of global warming. Proceedings of the National. Academy of Sciences, 102, 17384–17387. Soares, M., Cardoso, S. C. & Côté, I. M. (2007). Client preferences by Caribbean cleaning gobies: food, safety or something else?. Behavioral Ecology and Sociobiology, 61, 1015–1022  Tallamy, D. & Schaefer, C. (1997). Maternal care in the Hemiptera: ancestry, alternatives, and current adaptive value. In: The evolution of social behavior in insects and arachnids (eds. J. Choe, B. Crespi eds). Cambridge University Press. Cambridge, Reino Unido, pp. 94-115. Tallamy, D. W. & Wood, T. K. (1986). Convergence patterns in subsocial insects. Annual Review of Entomology, 31, 369-390. Thrall, P H., Hochberg, M. E., Burdon, J. J. & Bever, J. D. (2006). Coevolution of symbiotic mutualists and parasites in a community context. TRENDS in Ecology and Evolution, 22, 120 – 126. Therneau, T. (2015). A Package for survival analysis in S, version 2.38. The Comprehensive R Archive Network. In: https://CRAN.R-project.org/package=survival  Therneau, T. M. (2019). coxme: mixed effects cox models, R package version 2.2-14. The Comprehensive R Archive Network. In: https://CRAN.R-project.org/package=coxme  Torrico-Bazoberry, D., Cacerez-Sanchez, L., Saavedra-Ulloa, D., Flores-Prado, L., Niemeyer, H. N. & Pinto, C. F. (2014). Biology and Ecology of Alchisme grossa in a Cloud Forest of the Bolivian Yungas. Journal of Insect Science, 14, 1-4. Tvardikova, K. & Novotny, V. (2012). Predation on exposed and leaf-rolling artificial caterpillars in tropical forests of Papua New Guinea. Journal of Tropical Ecology, 28, 331–341. Van Baalen, M. & Sabelis, M. W. (1995). The milker-liller dilemma in spatially structured predator-prey interactions. Oikos, 74, 391–400. Van Beneden, P. J. (1875). Les Commensaux et les Parasites dans le Regne Animal. Bailliere, pp. 1-32. Vidal, M. C. & Murphy, S. M. (2018). Quantitative measure of fitness in tri-trophic interactions and its influence on diet breadth of insect herbivores. Ecology, 99, 2681–2691. Vincent, T. L. S., Scheel, D., Brown, J. S. & Vincent, T. L. (1996). Trade-offs and coexistence in consumer-resource models: it all depends on what and where you eat. The American Naturalist, 148, 1038–1058.  85  Volkl, W., Woodring, J., Fischer, M., Lorenz, M. W. & Hoffmann, K. H. (1999). Ant-aphid mutualisms: the impact of honeydew production and honeydew sugar composition on ant preferences. Oecologia, 118, 483–491. Ward, P. S. The phylogeny and evolution of ants. Annual Review of Ecology, Evolution and Systematics, 45, 23–43 (2014). Wickham, H. (2016). ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York. Wilson, E. O. & Holldobler, B. (2005). The rise of the ants: a phylogenetic and ecological explanation. Proceedings of the National Academy of Sciences, 102, 7411–7414. Witz, B. W. (1990). Antipredator mechanisms in arthropods: a twenty year literature survey. The Florida Entomologist, 73, 71-99. Wood, T. K. (1982). Selective factors associated with the evolution of membracid sociality. Biology of Social Insects, 175-179. Wood, T. K. (1984). Life history patterns of tropical membracids (Homoptera: Membracidae). Sociobiology, 8, 299-344. Wood, T. K. (1993) Diversity patterns in the new world Membracidae. Annual Reviews of Entomology, 38, 409-435.  Yao, I. (2014). Costs and constraints in aphid-ant mutualism. Ecological Research, 29, 383–391.  Yao, I. & Akimoto, S. (2002). Flexibility in the composition and concentration of aminoacids in honeydew of the drepanosiphid aphid Tuberculatus quercicola. Ecological Entomology, 27, 745–752. Yao, I., Shibao, H. & Akimoto, S. (2000). Costs and benefits of ant attendance to the drepanosiphid aphid Tuberculatus quercicola. Oikos, 89, 3–10. Zink, A. G. (2003). Quantifying the costs and benefits of parental care in female treehoppers. Behavioral Ecology, 14, 687–693.    86  Appendices  Appendix A. Supplementary figures    Figure S1. Live baits of the fly Lucilia sericata. From left to right and top to bottom: fly bait set up, predation by Crematogaster sp. ants, Pseudomyrmex sp. ant, and a wondering spider.     87   Figure S2. Species behavioral assemblage in treehopper communities at the western and eastern slopes of the Andes of Ecuador.  88    Fig S3. Proportion of baits predated by ants over time following discovery. Panels show different bait couplet configurations, where lines represent means from 6 transects each with 10 couplets of each type.     89  Appendix B. Supplementary tables  Table S1. Geographic and climatic information of study locations. Temperature and rainfall data from Worldclim (Hijimans et al., 2005). Temperature seasonal variation is calculated as the mean temperature of the warmest quarter minus the mean temperature of the coldest quarter.  Longitude Latitude Elevation (m.a.s.l.) Temperature Annual mean (°C) Temperature Seasonal variation (°C) Annual Rainfall (mm) Eastern slope      -77.555910 -1.092580 374 23.9 1.1 3568 -77.619978 -1.070100 394 24.0 1.1 3685 -77.629993 -1.066233 438 24.0 1.0 3697 -77.887760 -0.943950 741 22.6 1.0 4101 -77.898410 -0.980580 746 22.4 1.0 4076 -77.781500 -0.828750 808 22.0 1.0 3950 -77.814170 -0.788070 1114 21.1 1.0 3937 -77.822130 -0.690910 1348 19.5 1.0 3802 -77.643931 -0.151675 1349 19.6 0.8 3248 -77.788900 -0.657690 1646 18.9 1.1 3849 -77.864520 -0.448140 1754 17.8 0.9 2403 -77.871290 -0.475950 1851 18.0 1.0 2609 -77.875687 -0.580838 2049 17.3 1.1 2608 -77.890775 -0.600828 2110 16.1 1.1 2223 -78.027578 -0.415570 2388 14.4 0.9 1322 -78.016720 -0.428860 2630 15.0 1.0 1402 -78.101673 -0.370877 2844 9.7 1.2 1371 -78.106256 -0.370344 2879 9.7 1.2 1371 -78.146452 -0.372770 3263 6.5 1.2 1340 -78.149710 -0.379850 3368 8.3 1.1 1321 -78.151629 -0.350609 3448 5.5 1.1 1316 -78.187392 -0.376644 3549 7.6 1.1 1302 -78.192222 -0.372662 3682 6.0 1.1 1299 -78.192411 -0.362183 3792 5.6 1.1 1290 Western slope      -79.232140 0.111340 171 24.7 1.6 2929 -79.188274 0.098143 215 24.3 1.6 3210 -79.206070 0.106250 217 24.6 1.6 2991 -78.911800 0.181140 541 23.2 1.3 2562 -78.939840 0.154550 600 22.4 1.2 2516 -79.050140 0.075960 630 22.6 1.4 3548 -78.785900 -0.056360 1241 18.7 0.6 2485 -78.793810 -0.062520 1301 18.7 0.6 2485 -78.753630 -0.072900 1378 19.0 0.6 2448 -78.747400 -0.028100 1615 17.2 0.5 2462 -78.679910 0.021460 1781 17.3 0.4 2008 -78.676920 -0.003550 1876 16.9 0.4 1721 -78.697600 0.011340 2139 17.7 0.5 2032 -78.687770 -0.007520 2320 17.2 0.4 1921 -78.686190 -0.022920 2385 16.0 0.3 1593 -78.509170 0.043010 2554 16.2 0.3 1085 -78.508747 0.037057 2726 15.7 0.3 1123 -78.510060 0.031900 2927 15.4 0.3 1140 -78.591703 -0.273618 3298 11.0 0.3 1297 -78.603930 -0.278618 3347 11.5 0.3 1338 -78.612715 -0.291951 3493 9.1 0.5 1273 -78.538307 -0.201642 3552 6.9 0.6 1308 -78.540331 -0.193738 3796 10.4 0.4 1216 -78.540212 -0.187979 3897 6.2 0.6 1305    90  Table S2. Presence of ant genera and their status as mutualist partners of treehoppers (1 = partner; 0 = not partner, - = absent) across elevation intervals of 500 m between 0 and 4000 m.a.s.l. at the western and eastern slopes of the equatorial Andes.    Elevation intervals (m.a.s.l.) Subfamily Genus 500 1000 1500 2000 2500 3000 3500 4000 Dolichoderinae Azteca 1 1 1 1 - - - -  Dolichoderus 1 1 1 - - - - -  Linepithema - 1 1 1 - 1 - -  Tapinoma 1 - - - - - - - Ectatomminae Ectatomma 1 1 - - - - - - Formicinae Camponotus 1 1 1 - 1 - - -  Plagiolepis - - 1 - - - - - Myrmicinae Atta 0 - - - - - - -  Cardiocondyla 1 1 - - - - - -  Cephalotes 1 - - - - - - -  Crematogaster 1 1 1 1 - - - -  Pheidole 1 1 1 1 - - - -  Procryptocerus 1 - 1 - - - - -  Solenopsis 1 1 1 1 - - - -  Wasmannia 1 - 1 - - - - -  Megalomyrmex 1 - 1 - - - - - Ponerinae Pachycondyla 0 0 0 0 - - - - Pseudomyrmecinae Pseudomyrmex 0 - 0 - - - - -     91  Table S3. Presence of treehopper genera and their status as mutualist partners of ants (1 = partner; 0 = not partner, - = absent) across elevation intervals of 500 m between 0 and 4000 m.a.s.l. at the western and eastern slopes of the equatorial Andes.    Elevation (m.a.s.l.) Subfamily Genus 500 1000 1500 2000 2500 3000 3500 4000 Aetalionidae Aetalion 1 - - - - - - - Centrotinae Ischnocentrus - 0 - - - - - -  Unknown - - - 0 - - - - Darninae Alobia - 0 0 - - - - -  Darnis 0 - - 0 - - - -  Hyphenoe - - - 0 - - - - Heteronotinae Anchistrotus 1 - - - - - - -  Heteronotus 1 1 1 - - - - -  Nassunia 0 0 - - - - - - Membracinae Aconophora - - - 0 0 - - -  Alchisme - 0 0 0 0 0 0 -  Bolbonota 1 1 1 1 1 - - -  Calloconophora 0 - - - - - - -  Cladonota 0 0 0 0 - - - -  Enchenopa 1 1 1 - - - - -  Erechtia 1 1 - - - - - -  Guayaquila - 1 1 - - - - -  Leiscyta 1 1 - - - - - -  Membracis 1 1 - 1 1 - - -  Metcalfiella - - - - - 0 0 -  Ochropepla - - 0 0 - - - -  Phylia - - - 0 - 0 - -  Tylopelta - 0 - - - - - - Smiliinae Acutalis - - - - 0 - 0 -  Addipe 1 - - - - - - -  Amastris 1 1 - - - - - -  Anobilia cf. 1 1 - - - - - -  Antonae - - - 0 - - - -  Aphetea  1 - - - - - -  Ceresa cf. - - 0 0 - - - -  Cymbomorpha 0 - - - - - - -  Cyphonia 0 0 0 0 0 - - -  Ennya - - 0 0 0 0 0 0  Entylia - - 1 - - - - -  Heranice - - - - - 0 0 -  Horiola 1 1 - - - - - -  Ilithucia - - - - - 0 0 -  Metheysa 1 1 1 - - - - -  Micrutalis - - 0 - 0 0 - -  Notogonioides - - 1 - - - - -  Polyglypta - - - 0 - 0 - -  Poppea 0 - - 0 - - - -  Stilbophora 1 - 1 - - - - -  Todea - 1 - - - - - -  Tynelia 1 - - - - - - -  Vanduzea 1 1 - - - - - -  Unknown - - - - - - 0 - Stegaspidinae Lycoderes 0 0 - - - - - -  Stegaspis 0 - - - - - - - Smiliinae Acutalis - - - - 0 - 0 -     92  Table S4. Ant genera recorded predating upon live-fly baits across elevations at the western (W) and eastern (E) slopes of the equatorial Andes, and whether they are documented partners of hemipterans.     Elevation (m.a.s.l.) Subfamily Genus Mutualist 500 1000 1500 2000 2500 3000 Dolichoderinae Azteca yes E W - - - -  Dolichoderus yes WE WE - - - -  Linepithema yes - E WE - - -  Tapinoma no W WE - - - - Ectatomminae Ectatomma yes WE W - - - - Formicinae Camponotus yes WE WE WE W - - Myrmicinae Cephalotes yes E - - - - -  Crematogaster yes WE WE WE E - -  Pheidole yes WE WE WE - - -  Procryptocerus yes - - WE E - -  Solenopsis yes WE E - - - - Ponerinae Pachycondyla no - W E W W -  Odontomachus no - E - - - - Pseudomyrmecinae Pseudomyrmex no - E - - - -     93   Table S5. Treehopper taxa included in the analysis and their associated behavioral traits.  Subfamily Tribe Genus # of species Maternal care Ant-mutualism Centrotinae Boocerini Campylocentrus 1 no no   Ischnocentrus 6 no nymphs   Unknown 2 no no Darninae Cymbomorphini Cymbomorpha 7 no nymphs   Eumela 1 no no  Darnini Alobia 2 no no   Aspona 1 no no   Cyphotes 5 no no   Darnis 4 no no   Stictopelta 5 no no   Sundarion 8 no no  Hemikyptini Atypa 4 no no   Hemikyptha 1 no no   Hemikyptina 3 no no   Unknown 4 no no  Hyphenoini Almeone 1 no no   Hyphenoe 4 no no  Procyrtini Procyrta 2 no no  Unknown Unknown 3 no no Endoiastinae Endoiastini Endoiastus 4 eggs nymphs + adults   Scytodepsa 2 eggs nymphs + adults Heteronotinae Heteronotini Anchistrotus 2 no nymphs + adults   Dysyncritus 1 no nymphs   Heteronotus 11 no nymphs   Nassunia 3 no nymphs   Omolon 1 no nymphs   Rhexia 5 no nymphs + adults   Smilioarchis 1 no nymphs Membracinae Aconophorini Aconophora 7 eggs no   Calloconophora 5 egg + nymphs nymphs + adults   Guayaquila 10 eggs nymphs + adults  Hoplophorionini Alchisme 14 egg + nymphs no   Metcalfiella 11 egg + nymphs no   Ochropepla 7 egg + nymphs no   Potnia 5 egg + nymphs no   Sakakibarella 3 egg + nymphs no   Umbonia 7 egg + nymphs no  Hypsoprorini Cladonota 10 no no   Hypsoprora 12 no no   Notocera 6 no nymphs  Membracini Bolbonota 10 no nymphs + adults   Enchenopa 10 no nymphs   Enchenophyllum 1 no nymphs   Erechtia 26 eggs nymphs + adults   Havilandia 4 no nymphs   Leioscyta 7 no nymphs + adults   Membracis 22 no nymphs   Paragara 1 no nymphs + adults   Phyllotropis 2 no nymphs   Tritropidia 2 eggs nymphs + adults     94  Table S5. Continued.  Subfamily Tribe Genus # of species Maternal care Ant-mutualism Nicomiinae Nicomiini Nicomia 1 no no   Nodonica 1 no no   Tolania 9 no no Smiliinae Acutalini Thrasymedes 3 no no   Unknown 6 no no  Amastrini Amastris 12 no nymphs + adults   Harmonides 8 no nymphs + adults   Lallemandia 1 no nymphs + adults   Tynelia 6 no nymphs + adults   Vanduzea 4 no nymphs + adults  Ceresini Antonae 4 no no   Ceresa cf. 14 no no   Cyphonia 15 no no   Ilithucia 2 no no   Poppea 2 no no   Stictocephala 1 no no  Micrutalini Micrutalis 2 no no   Unknown 14 no no  Polyglyptini Adippe 3 eggs nymphs + adults   Aphetea 2 eggs nymphs + adults   Dioclophara 2 eggs nymphs + adults   Ennya 20 eggs + nymphs no   Entylia 1 eggs nymphs + adults   Gelastogonia 3 eggs no   Hemiptycha 2 eggs nymphs + adults   Heranice 1 egg + nymphs no   Metheisa 5 eggs nymphs + adults   Notogonioides 1 eggs nymphs + adults   Polyglypta 5 eggs no   Polyglyptodes 1 eggs nymphs + adults   Unknown 2 eggs no  Smiliini Antianthe 2 eggs nymphs + adults  Thuridini Thuris 1 eggs nymphs + adults  Tragopini Anobilia 8 eggs nymphs + adults   Chelyoidea 8 eggs nymphs + adults   Colisicostata 1 eggs nymphs + adults   Horiola 3 eggs nymphs + adults   Todea 8 eggs nymphs + adults   Tragopa 2 eggs nymphs + adults   Tropidolomia 1 eggs nymphs + adults   Unknown 24 eggs nymphs + adults  Unknown Unknown 1 no no Stegaspidinae Microcentrini Centroflexus 1 no no  Stegaspidini Bocydium 2 no no   Lycoderes 11 no no   Oeda 1 no no   Smerdalea 3 no no   Stegaspis 1 no nymphs + adults   Stylocentrus 2 no no     95  Table S6. List of treehopper genera used in the experiment and percentage of baits in the mutualist-bait treatment of each genus in each transect.  Family/Subfamily Genera T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 T12 Total Aetalionidae                Aetalion 10 15 0 5 0 0 0 0 0 0 0 0 2.5 Membracidae                 Darninae Cymbomorpha 0 5 0 0 0 0 0 0 0 5 0 0 0.8  Alobia 5 0 5 0 0 0 0 0 0 0 0 0 0.8   Heteronotinae Anchistrotus 0 0 0 5 0 0 0 0 0 0 0 0 0.4  Heteronotus 5 0 0 0 0 0 0 0 0 0 5 0 0.8   Membracinae Bolbonota 0 0 0 15 0 0 0 0 0 0 5 0 1.7  Campylenchia 0 0 0 0 0 0 0 0 0 10 0 0 0.8  Enchenopa 20 0 0 0 0 0 0 0 0 25 20 0 5.4  Erechtia 35 5 15 35 0 70 0 0 0 0 0 0 13.3  Membracis 5 5 0 0 0 0 0 70 40 0 0 0 10.0  Unk. Membracini 0 0 0 0 0 0 100 0 0 0 0 0 8.3   Smiliinae Amastris 0 15 25 10 0 0 0 25 15 10 0 5 8.8  Lallemandia 0 0 0 0 0 0 0 0 5 0 0 0 0.4  Cyphonia 0 10 0 10 0 0 0 5 0 0 0 0 2.1  Aphetea 0 0 0 0 0 0 0 0 0 0 30 15 3.8  Hemiptycha 0 5 0 0 0 0 0 0 25 0 0 0 2.5  Anobilia 0 15 10 0 0 0 0 0 0 0 15 0 3.3  Horiola 10 10 0 0 0 0 0 0 0 35 15 0 5.8  Stilbophora 0 0 0 0 100 0 0 0 0 0 0 0 8.3  Todea 0 5 5 0 0 30 0 0 0 0 0 30 5.8   Stegaspidinae Lycoderes 0 0 10 15 0 0 0 0 0 10 0 0 2.9  Stegaspis 10 10 30 5 0 0 0 0 15 5 10 50 11.3    

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0395376/manifest

Comment

Related Items