Kui 1 Differential Maternal Defense Behavior in Three Species of Treehoppers (Hemiptera: Membracidae) Mackenzie Kui Department of Molecular, Cellular, and Developmental Biology University of California, Santa Barbara EAP Tropical Biology Program, Fall 2017 15 December 2017 ABSTRACT Maternal care in the animal kingdom can vary immensely from species to species, but the reason for its ubiquity underlies an evolutionary impulse to pass on favorable genes to the next generation. In treehoppers, maternal care increases the rate of offspring survival through species specific levels of protection from both predation and parasitism. To investigate the factors that influence maternal care, I introduced a variety of animate and inanimate stimuli to both brooding and nonbrooding treehopper s in three species: Antianthe expansa, Alchisme grossa, and Umbonia crassicornis By quantifying their responses as proxies for energy output, I concluded that maternal defense is a speciesspecific phenomenon that is distinct from personal defense, relies primarily on kinesthetics rather than visual input, and is stimulus dependent. This type of species specific ity is likely governed by different types of environmental pressures that necessitate the evolution of direct defense behavior s or inter species mutualisms. Cuido Maternal en Tres Especies de membr cidos (Hemiptera: Membracidae) RESUMEN El cuido maternal en el reino animal puede variar enormemente de una especie a otra, pero la razn de su ubicuidad subyace a un factor evolutivo de transmitir genes favorables a la siguiente generacin. En membrcidos (Membracidae), el cuido maternal aumenta la tasa de supervivencia de las cras a travs de niveles especficos de proteccin contra la depredacin y el parasitismo. Para investigar los factores que influyen en el cuido maternal, introduje directamente una variedad de estmulos animados e inanimados a las hembras con y sin huevos o ninfas de tres especies: Antianthe expansa, Alchisme grossa y Umbonia crassicornis. Al cuantificar sus respuestas como aproximaciones de consumo de energa, conclu que la defensa maternal es un fenmeno especfico de la especie que es distinto de la defensa personal. Adems, se basa principalmente en la cinestti ca en lugar de la entrada visual, y es dependiente del estmulo. La especificidad de cuido maternal segn la especie probablemente se rige por diferentes tipos de presiones ambientales que requieren la evolucin de comportamientos de defensa directa o mutualismos con otras especies.
Maternal Defense Behavior of Treehoppers Kui 2 The biological foundation underlying both parasitoidhost and predator prey relationships stems from an innate struggle for survival characterized by adaptive evolution, natural selection, and specialization over time. Treehoppers (Membracidae) in particular, experience parasitism of their eggs from many species of wasps; some of which can be as small or even smaller than the eggs themselves (Godoy et al. 2006). Predation by comparably large invertebrates such as spiders, flies, beetles, and mantids is also common for many treehoppers despite their hard exterior, cryptic coloration, and ant mutualists ( Godoy et al 2006). As a result, some species of treehoppers have evolved varying levels of subsocial maternal behavior in an effort to protect their brood from both parasites and predators (Godoy et al. 2006). This type of behavior, in contrast to its solitary and gregarious counterparts, distinguishes itself by having the mother facilitate the survival, growth, and development of her offspring (Eickwort 1981). For some species, maternal care may be limited to the passive use of their bodies as shields for their eggs until they hatch, although abandonment prior to hatching may still occur (Haviland 1925). For others, maternal care may extend beyond passive protection through kicking, shaking, and wing vibrations, in response to potential threats (Godoy et al. 2006). In either case, the survival of the offspring often times hinges on the presence of the mother (Cocroft 2002). However subsocial behavior of hemipterans also increases the risk of predation and decreases fecundity of egg guarding mother s (Tallamy and Schaefer 1997). Some researchers suggest that the advent of ant mutualisms and adult aggregations were responses to the heightened risks associated with subsocial behaviors (Gadelha et al. 2016). In exchange for honeydew produced by nymphal or adult excretion the ants provide protection against potential predators (Gadelha 2016). This mutualism with ants and their propensity to aggregate en masse suggests that there is some visual, kinesthetic, chemosensory or acoustic component relevant to the distinction of friend or enemy. F or many treehoppers, the presence or absence of these ants may dictate the nature and extent of their maternal defense. In my study, I addressed the following questions : (1) does defense behavior of brooding treehoppers depend on the type of stimulus and (2) how does this defense behavior differ among treehopper species. I used Alchisme grossa, Antianthe expansa, and Umbonia crassicornis to study the d ifferential maternal defense behaviors across this diverse family of insects. METHODS Subject Selection All experimental subjects for this study were found in the Monteverde area between 13 November and 25 November Four brooding mothers with egg masses and one nonbrooding mother of Alchisme grossa were extracted from the leaves of Acnistus arborescens (Guitite trees) No two subjects were ever found occupying the same branch. Each branch containing A. grossa was then cut, transplanted into a jar containing water, sealed with parafilm, and placed in a large glass housing for the remainder of the experiments. The same procedure was carried out for two Antianthe expansa with egg masses and three A. expansa with nymphs. In all cases, adult A. expansa were observed to be in groups of three or more with ants and nymphs always close by. One Umbonia crassicornis with nymphs was found on an unknown Fabaceae tree and experiments were done on site for this species. Three types of stimuli were used for this study: a stink bug, caterpillar, and a Guitite leaf.
Maternal Defense Behavior of Treehoppers Kui 3 Threat Introduction Animate predators (stink bug or caterpillar) were introduced to brooding mothers using a leaf vector The leaf was used to mimic a quasi natural seting to prevent the stinkbug or caterpillar from being prematurely startled prior to each trial. In order to ensure that the predator would make contact with the mother, vinegar was applied with a pincel along the edges and underside of the leaf to deter the predator from moving elsewhere. Physical contact between the predator and the mother or her egg sac and thirty seconds thereafter constituted one trial. At least two minutes were allotted before the initiat ion of the next trial. For trials in which a treehopper was not on an egg mass, I waited until the subje ct was stationary to introduce the stimulus Fresh Guitite leaves were used as inanimate stimuli for A. grossa and A expansa while leaves from the associated Fabaceae plant were used for U. crassicornis. One leaf was held to the pronotum of each treehopper for either 5 seconds or 10 seconds. The response during the application of the stimulus and 30 seconds thereafter constituted one trial. Response Scoring Determination In order to provide a quantitative measure for the various defens ive responses of the treehoppers, I assigned scores to each type of response. Each score was determined based on response frequencies during preliminary observations an d relative projected energy output s For each trial, the responses were recorded, quantified, and summed as follows: 0: stationary passive defense, 1: movement away from the egg mass, 2: lateral shaking, 3: kicking, 4: wing vibrations, and 5: flying away (Table 1). Statistical analyses of differential maternal defense for the various stimuli were performed using twosample, two tailed t tests of the mean and standard deviation values. Table 1. Response Score Quantification Score Name Behavior in response to stimulus 0 Stay Remaining stationary; p assive defense 1 Movement away Walkin g away 2 Shake Shaking of the body laterally 3 Kick Kicking of the stimulus using its hindlegs 4 Wing vibration Flapping its wings for intimidation 5 Fliy Flying away RESULTS Maternal defense response is speciesspecific To understand the specieslevel variability of maternal defense behavior I assessed the differential response frequencies of A. expansa, U. crassicornis and A. grossa. Because I was only interested in the types of defenses present in each species, I disregarded the nature of the stimulus for simplification. For A. expansa, only three types of responses were observed: passive defens e shaking, and walking away; for U. crassicornis, all except flying away was observed; for A. grossa, all responses were observed (Figure 1).
Maternal Defense Behavior of Treehoppers Kui 4 Figure 1. The range of maternal defense responses to any type of threat is species -specific. Passive maternal defense throughout brooding is variable and species specific To further understand the extent of passive maternal care, I recorded the percentage of mothers who were present on an egg sac every morning for 14 days U. crassicornis remained with her nymphs every day of the study while A. grossa and A. expansa both showed varying levels of passive maternal care (Figure 2). However, because experiments for U crassicornis were done on sit e as compared to the other two species in the lab I cannot omit the possibility that this type of behavior was environmentally dependent Figure 2. Maternal presence of U. crassicornis is persistent while A. grossa and A. expansa demonstrate inconsistency of presence throughout the brooding process.
Maternal Defense Behavior of Treehoppers Kui 5 The extent of maternal care is stimulusdependent Maternal defense behavior is kinesthetic, not visual T o isolate a possible visual component to maternal defense, I compared the responses of one mother to leaves held in front of her and leaves that made physical contact with her pronotum. I then repeated these experiments in the dark and observed their responses using red light Because I only wanted to isolate the possibility of visual detection of potential threats, I limited my trials to one individual. Regardless of the presence of light physical contact was necessary for the elicitation of a defensive response (Figure 3). Figure 3. Physical contact with the mother is necessary to elicit a defensive response. (N=1) Maternal defense behavior could be temporally dependent In order to assess whether the duration of a stimulus is important in maternal defense, I subjected each treehopper to leaf stimulation for five or ten seconds. For all three species, the average response score was higher when the stimulus was prolonged (F igure 4). However, only U. crassicornis showed a statistically significant difference in aver age response scores with regard to the duration of the stimulus. This result is reported with the caveat that I only observed the behavior of one individual of U. crassicornis. Treehopper mothers display a heightened defens ive response for animate stimuli In order to determine whether treehopper mothers can distinguish between animate and inanimate stimuli, I provoked a defensive response through the introduction of either an insect or a leaf For all three species, the average response score was greater for animate stimuli, but only A. grossa and U. crassicornis displayed statistical difference s between the stimuli (Figure 4). Next, because treehoppers must grapple with both predation from larger invertebrates and parasitism of their eggs from smaller wasps, I investigated how the size of animate threats affects maternal defense response. Thus, I used stink bugs of comparable si ze to the treehopper and first instar caterpillars as my predation and parasitoid size surrogates respectively. For all three
Maternal Defense Behavior of Treehoppers Kui 6 species, the treehopper mothers do not significantly alter their behavior in response to size. However, because caterpillars and stink bugs are different in more ways than just size, this conjecture may be unfounded. Figure 4. A verage response score of maternal defense is higher for larger, animate stimuli and for prolonged inanimate stimuli as compared to their associated cou nterparts. **p<0.01, ***p<0.001; ns, not significant. A. grossa can recognize individuals of its own species Because adult A. grossa have been observed to aggregate in nature, I sought to understand their behavior when faced with a familiar stimulus. On eight separate, unabated occasions, I observed more than one A. grossa mother on the same leaf under laboratory settings These aggregations normally involved more than one female within five c entimeters of each other and with at least one female occupying the space atop an egg mass. The particular individual atop an egg mass varied by occasion. Du ring physical interactions among individuals in these situations, I frequently noted only passive defens ive behaviors, such as remaining stationary or walking away. Typical responses to animate stimuli such as kicking or wing vibrations were rarely observed. Personal defense is distinct from maternal defense in A. grossa To distinguish between maternal defense and personal defense, I conducted threat mimicry experiments on A. grossa who were not guarding an egg sac. Regardless of the threat type (animate or inanimate), treehopper mothers who were actively guarding an egg sac displayed a heightened response score as compared to astray individuals. This experiment was l imited to A. grossa because they showed the greatest variability in response types and the highest frequency of egg mass abandonment.
Maternal Defense Behavior of Treehoppers Kui 7 Figure 5. Average response score of maternal defense in A. grossa is higher when the mother is guarding her egg sac than when she is not. ** *p<0.001 DISCUSSION V ariation in defensive responses with regard to the type of stimulus revealed an acute ability to kinesthetically, not visually distinguish between different types of threats (Figure 3). Anatomically, this provides some evidence to previous notions that the pr onotum has a sensory function (Wood 1984). This decreased reliance on vision likely stems from the evolutionary adaptation for substrate borne acoustic communication as the primary form of threat detection (Cocroft 1999). The elevated average response score displayed across all three species for animate stimuli may be (1) because of an innate energy allocation mechanism that responds to more genuine threats or (2) because the leaf stimulus, unlike the caterpillar and stinkbug, did not induce substrate borne vibrations before making to physical contact; thus the difference in response scores may also rely on the mothers ability to detect the threat prior to its arrival. Although there was no difference in the response scores with regard to the type of animate stimulus (caterpillar or stink bug), the treehopper mothers demonstrated the ability to distinguish between conspecifics and hetero conspecifics. This ability likely stems from the integration of substrate vibrations and pheromone communication in maternal defense (Cocroft 1999). Furthermore, the duration of a stimulus seemed to play a significant role in the level of responses of the treehoppers. This determination may also be the result of yet another energy allocation mechanism that prevents the waste of unnecessary energy on ephemeral stimuli. Additionally, there were always undefined refractory periods between successive kicks, further suggesting that the duration of the stimulus may play a role in the determination of a defense response. Taken together, I propose that the stimulus dependent response scores stems from a kinesthetically dependent reaction optimized to energy conservation. Phylogenetic patterns of trait divergence among Membracidae tribes reveal a potential tradeoff between ant mutualisms and extended maternal care (Olmstead and Wood 1990b) (Figure S1) With regard to the species in my study, A. grossa and U. crassicornis occupy the Hoplophorioni ni tribe while A. expans a occupies the Smili ini tribe. Moreover, neither U. crassicornis nor A. grossa were ever found in the presence of ant mutualists ; while A. expansa
Maternal Defense Behavior of Treehoppers Kui 8 was always found in the presence of ant mutualists T he extended range of behaviors and increased average response scores among U. crassicornis and A. grossa, as compared A. expansa, may suggest that the advent of ant mutualisms contributed to this evolutionary divergence (Figure 1 and 4). It thus stands to reason that individuals like A. expansa who benefit from the protection of mutualists would not require the energetically costly adaptations designed for the personal or maternal defense as seen in A. grossa and U. crassicornis (Figure 1). However, this then begs the question: why dont A. grossa and U. crassicornis have ant mutualists ? One hypothesis for the evolution of this proposed tradeoff between ant mutualisms and extended maternal care involves elevational gradients (Wood 1984) Because higher altitudes preclude the development of robust ant populations, treehoppers who occupy such areas likely evolved advanced degrees of personal and maternal defense, like A. grossa and U. crassicornis (Wood 1984, Janzen et al 1976). This finding thus gave rise to the notion that treehoppers with ant mutualists are ancestral to those without and that hoplophorioninines like A. grossa and U. crassicornis secondarily lost their assocation with ant mutualists ( Costa 2006) Howe ver, because A. expansa, its ant mutualists and A. grossa are all often found occupying Acnistus arborescens trees, A. grossa likely began its evolutionary history occupying higher elevation trees, but through time, returned to occupy those in lower elevations Thus, the absence of ant mutualisms for A. grossa is not limited by ant abundance; but perhaps ( 1) because ants can distinguish between different sources of honeydew Like most Membracids, A. grossa excretes honeydew as part of their plant sap diet Therefore there may be some difference in honeydew composition that dissuades ants from approaching A. grossa. O r ( 2) because ants have learned over time that interactions with non mutualists may result in bodily harm Correspondingly, the evolution of ant mutualism s likely preclude s the development of kicking behavior to prevent the inadvertent kicking of mutualist ants. However, f urther research into the precise sensory and genetic variables of ant mutualisms is required to completely understand this ant treehopper relationship The large bodies and sharp lateral pronotal horns of U. crassicornis and A. grossa, also suggests a deviation from a reliance on hymenopteran mutualists. For U. crassicornis, these pronotal adaptations are essential to their defense against anuran predation (Wood 1977). In fact, removal of their scelero tized pronotum greatl y increases their rate of predation (Wood 1993). These tribal distinctions in protective behavior and physical anatomy thus give credence to the proposed phylogenetic relationship that U. crassicornis and A. grossa are more closely related to each other than they are to A. expansa (Dietrich 2001). Additionally extended maternal behavior has been closely linked to semelparous individuals who are limited to only one brood throughout its lifetime (Wood 1984). Evolutionarily speaking, abandonment of egg masses is likely to be low and maternal defense energy output is likely to be high in semelparous individuals as compared to their iteroparous counterparts (individuals who can lay eggs more than once in their lifetime). Empirical ovipositional data suggests that U. crassicornis is obligatorily semelparous (Wood and Dowell 1984). Despite my low sample size for this species, its persistent presence with the nymphs throughout my study and its consistently high energy output further highlights this suggestion (Figure 2 & Figure 4). This type of parental behavior is in contrast to those of A. grossa and A. expansa who demonstrated a strong propensity for eggmassabandonment and lower levels of average response scores for any given stimulus (Figure 2). Although members of the Hoplophorionini tribe were originally believed to be exclusively semelparous, one research
Maternal Defense Behavior of Treehoppers Kui 9 group observed two individuals of A. grossa tending to two distinct egg masses during different periods of the year (TorricoBazoberry et al 1984). A. grossa, thus may be either facultatively or moderately iteroparous. And despite the lack of evidence regarding ovipositional behavior in A. expansa, their observed rate of abandonment may also suggest either facultative or moderate iteroparity. There are also hypotheses that suggest that brooding iteroparous individuals have a tendency to aggregate in an effort to increase the probability that ants will tend to them (Torrico Bazoberry et al 1984). During my initial subject collection, adult individuals of A. expansa were always found aggregated in groups of three or more with ants and nymphs present, thus further strengthening the case for iteroparity in A. expansa. The observed frequency of brooding aggregations of A. grossa suggests a more complex type of extended maternal care than originally proposed (Wood 1993). However, this type of behavior is not limited to A. grossa, as fellow membracids Publilia concava and Polyglypta dispar have also demonstrated similar egg swapping behavior (Zink 2003a and Eberhard 1986). Here, I propose two hypotheses for this unusually high incidence of brooding ag gregations. My first hypothesis considers the possibility of brood parasitism in which one female deposits eggs into the existing clutch of a conspecific (Zink 2003a). This type of behavior has been observed in A. grossa but the benefits for both the parasite and recipient are not fully understood (Torrico Bazoberry et. al 1984). In the case of Publilia concava, brood parasitism does not reduce the hatching success of host eggs, increases the lifetime fecundity of the par asitic mother, and increases lifetime clutch count (Zink 2003a). However, unlike what I observed in A. grossa, parasitism by Publilia concava does not involve any type of maternal egg guarding post oviposition. For example, in most of my observed cases, a mother who had adopted a new egg mass displayed typical maternal defense behavior as if it were her own. This distinction between A. grossa and P. concava in post parasitism defense behavior may stem from regional and temporal differences of brooding that for P. concava, result s in maternal death prior to egg hatching (Wood 1993). My second hypothesis for the high incidence of brooding aggregations that I observed in A. grossa considers the possibility of cooperative care. Although largely undocumented in treehoppers i ndividuals that engage in cooperative care have been shown to benefit from aggregations through the creation of a nutrient sink in the host plant that provides food for both the adults and the nymphs ( Lin 2006). To exclude the possibility that they were sharing leaves due to some nutritional deficiency, I introduced a fresh leaf after the initial aggregation ( 19 November) Throughout the remainder of the study, I never observed any treehoppers occupying this alternate food source. Interestingly, for the last three days of my study, one A. expansa individual was consistently found atop an abandoned A. grossa egg mass. Although these events were only observed under laboratory settings, they may suggest that either the egg mass itself or the modifications made by the original mother provides some sort of additional nutrient source that fresh leaves cannot supply. F emale adult aggregations of A. grossa have also been observed to decrease the rate of parasitism of eggs through a potential cooperation among eggguarding mothers (Camacho et. al 2014). Although in my study, a maximum of three A. grossa mothers aggregated at one time, these hypotheses may provide a via ble explanation for the high incidence of maternal aggregations In sum, by quantifying their responses as a proxies for energy output, I concluded that maternal defense is a speciesspecific phenomenon that is distinct from personal defense, relies primarily on kinesthetics rather than visual input, and is stimulus dependent. This type of species specificity is likely governed by different types of environmental pressures that necessitate the evolution of direct defense behavior or inter species mutualisms. However, t hese
Maternal Defense Behavior of Treehoppers Kui 10 results and hypotheses are reported with the caveat that the various animate stimuli (caterpillars and stink bugs) used in this study are not known predators of these treehoppers. Further research involving their natural predators may provide a more robust depiction of the various factors involved in maternal defense recognition and behavior. ACKNOWLEDGEMENTS This project would not have been possible without the endless support from my primary advisor Emilia Triana who was always willing to share her knowledge and curiosity with me. Thank you to Frank for the helpful advice during the making of this paper. Than k you to Kara Powell for peer reviewing my paper. I would also like to thank my homestay mother Miriam and my homestay brother Andres for their love and hospitality throughout the making of this project. Thank you to the Estacion Biologica M onteverde for allowing me to use the lab for my experiments. REFERENCES Camacho, L., Keil, C. and Dangels, O. (2014), Factors influencing egg parasitism in subsocial insects: insights from the treehopper Alchisme grossa (Hemiptera, Auchenorrhyncha, Membracidae). Ecol Entomol, 39: 5865. Cocroft, R. B; Antipredator defense as a limited resource: unequal predation risk in broods of an insect with maternal care, Behavioral Ecology Volume 13, Issue 1, 1 January 2002, Pages 125133 Cocroft R. B. 1999b. Offspring parent communication in a subsocial treehopper (Hemiptera: Membracidae: Umbonia crassicornis ) Behaviour 136 : 1 21 Costa, James T. The Other Insect Societies. The Belknap Press of Harvard University Press, 2006. Dietrich, C. H., Rakitov, R. A., Holmes, J. L. and Black, W. C. 4th. (2001). Phylogeny of the major lineages of Membracoidea (Insecta: Hemiptera: Cicadomorpha) based on 28S rDNA sequences. Mol. Phylogenet. Evol. 18, 293305 Eberhard WG. 1986. Possible mutualism between females of the subsocial membracid Polyglypta dispar (Homoptera). Behavioral Ecology and Sociobiology 19:447453. Eickwort, G. C., 1981. Presocial insects. In:Social Insects, Vol. 2, H. R. Herman (ed.), pp. 199 280. Academic Press, New Yo rk, 491 pp. Gadelha, Y., Dttilo, W., Evangelista, O., & Lopes, B. (2016). Structure of mutualistic ant treehopper interactions in the Brazilian Atlantic Forest. Journal of Tropical Ecology, 32(3), 250259. Godoy, Carolina, et al. MembraCidos De La AmeRica Tropical = Treehoppers of Tropical America INBio, 2006. Hamilton, William D. "Geometry for the selfish herd." Journal of theoretical Biology 31.2 (1971): 295 311. Haviland MD. (1925) The Membracidae of Kartabo, Bartica District, British Guiana, with descriptions of new species and bionomical notes. Zoologica 6: 229290.
Maternal Defense Behavior of Treehoppers Kui 11 Janzen, D. H., M, Ataroff, M. Farinas, S. Reyes, N. Rincon, A. Soler, P. Soriano, and M. Vera. 1976. Changes in the arthropod community along an elevational transect in the Venezuelan Andes. Biotropica 8; 193 203. ;) Lin C. P. 2006. Social behaviour and life history of membracine treehoppers J. Nat. His. 40: 18871907. Olmstead, K. L., and T. K. Wood 1990b. Altitudinal patterns of treehopper (Homoptera) species richness: The role of ants and treehopper life Histories. Proceedings of the Entomological Society of Washington. 92:552560. Tallamy, Douglas W., and Carl Schaefer. "Maternal care in the Hemiptera: a ncestry, alternatives, and current adaptive value." The evolution of social behavior in insects and arachnids (1997): 94115. Torrico Bazoberry D, CaceresSanchez L, SaavedraUlloa D, Flores Prado L, Niemeyer HM, Pinto CF. Biology and Ecology of Alchisme gr ossa in a Cloud Forest of the Bolivian Yungas Journal of Insect Science. 2014;14:169. doi:10.1093/jisesa/ieu031. Wood, T. K. Defense in Umbonia crassicornis: Role of the Pronotum and Adult Aggregations (Homoptera: Membracidae), Annals of the Entomological Society of America Volume 70, Issue 4, 15 July 1977, Pages 524528 Wood, T. K. 1984. .Life history patterns of tropical membracids (Homoptera: Membracidae) Sociobiology 8 : 299 344 Wood, T. K., and Robert Dowell. Sex Ratio in Umbonia Crassicornis Amyot and Serville (Homoptera: Membracidae). The American Midland Naturalist vol. 112, no. 1, 1984, pp. 5866. JSTOR JSTOR. Wood, T K (1993) Diversity in the new world Membracidae. Annu Rev Entomol 38:409435 Zink AG. 2003a. Intraspecific brood parasitism as a conditional reproductive tactic in the treehopper Publilia concava. Behavioral Ecology and Sociobiology 54:406415. Supplemental Information S 1. Abbreviated phylogenetic relationship among A. expansa, A. grossa, and U. crassicornis. Tribal affiliations are listed to the right. Adapted from Dietric h et al 2006.