Feeding behavior in Orchard s piders Govardhan 1 The effect of prey danger level on the feeding behavior of Orchard spiders, Leucauge sp. Chirag Govardhan Department of Ecology and Evolutionary Biology University of California, Los Angeles EAP Tropical Biology and Conservation Program, Fall 2016 16 December 2016 ABSTRACT Predatory animal s often need to consider the costs and benefits of an attack, so as to avoid getting bitten or stung. Orchard spiders, Leucauge sp. manipulate their orb webs to detect prey that lands on their web and attack them accordingly Generally, they bite prey to inject venom and may also wrap them in silk This study investigated the behavioral sequences of orchard spiders when they were presented with h ighly dangerous prey ( Atta sp. minima ants) versus low danger prey ( Drosophila sp. fruit flies ) I tested 60 spiders, which I classified by body size (to account for size of the spider) and abdomen size (to control for their hunger level). I split them into two groups based on which prey I fed them, and observed their behavior, which I categorized into predatory and non predatory. The results showed that on average, orchard spiders showed predatory behavior far more often tow ards Drosophila sp. than to Atta sp Additionally, spider size was shown to not be a factor in the likelihood of predatory behavior. The results also showed that hunger level does not play a large part in whether or not pred atory behavior was displayed. I l patterns, which showed that individual spider behavioral sequences were far more random and unique with Atta sp. than with Drosophila sp. This was either due to the novelty of the ants as prey or the un certainty about how to react to dangerous prey. These results ultimately support the idea that orchard spiders take a more offensive approach with non dangerous prey than with dangerous prey Efecto del nivel de peligro de presa en el comportamiento de ataque de las araas de jardn, Leucauge sp. RESUMEN Los depredadores deben considerar los costos y beneficios de un ataque a su presa, para evitar daos mayores por ejemplo mediante mordeduras o picaduras. Las araas de jardn Leucauge sp., utilizan sus telas orbiculares para detectar las presas que caen en su tela desde el meollo y evaluar la situacin de ataque. Generalmente, muerden la presa para inyectar veneno y tambin pueden envolverlos en seda. Este estudio investig las secuencias de comportamiento de estas araas cuando se les present presas altamente peligrosas (hormigas Atta sp., minima) versus presas de bajo riesgo (moscas de la fruta Drosophila sp.). Utilic 60 araas, que clasifiqu segn el tamao del cuerpo (para tener en cue nta el tamao de la araa) y el tamao del abdomen (para controlar su nivel de hambre). Observ su comportamiento con base en el tipo de presa suministrado, el cual clasifiqu en depredador y no depredador. Los resultados mostraron que en promedio las ara as mostraron comportamiento depredador mucho ms a menudo hacia Drosophila sp. que a Atta sp. Adems, se demostr que el tamao de la araa no era un factor en la probabilidad de comportamiento depredador. Los resultados tambin mostraron que el nivel de l lenura no tena un papel importante en el comportamiento depredador.
Feeding behavior in Orchard s piders Govardhan 2 Constru etogramas para detallar los patrones de comportamiento en ambas categoras, con lo cual demuestro que las secuencias de comportamiento de las araas eran mucho ms aleatorias y nicas con Atta sp. que con Drosophila sp. Esto fue debido posiblemente a la novedad de las hormigas como presa o a la incertidumbre de reaccin hacia las presas peligrosas. Estos resultados apoyan la idea de que estas araas se comportan de manera ms ofen siva con presas no peligrosa que con las presas peligrosas Predation, a biological interaction where one animal feeds on another, is not one that is always risk free. Many types of prey possess physical, chemical, or behavioral defenses against predation, allowing them to a void detection, fight back, or avoid being eaten in other ways (Jeschke and Tollrian, 2000; Broom and Ruxton, 2005). Forbes (1989) proposed the dangerous prey hypothesis, where he suggested that more dangerous prey must be handled more carefully, and as a result, predators will attempt to attack dangerous prey less often While he suggested this hypothesis by considering a fish bird interaction, it can be universally applied across the animal kingdom. Due to v ariations in prey defense that have developed over time in response to predation (Creel, 2008), many different approaches to predation have come into existence. Two of the biggest strategies to predation are pursuit predation and ambush predation. Pursuit pre dation usually requires the predator being fast enough to chase after and catch the prey as it runs away (Janis and Wilhelm, 1993). Ambush predation, on the other hand, involves the predator remaining at and rey to come by before snatching them. Orb weaver spiders use a variation of this strategy, through the use of an orb web that they build and wait on. These spiders do not do the capturing themselves they wait for prey to pass O rb webs generally consist of 12 to 20 radial threads that extend from a central hub to the exterior, connected b y a seri es of spirals starting from the area around the hub and going further and further away ( Levi and Levi, 1990; Milne and Milne 1980 ; Opell, Bond, and Warner, 2006 ). Orb w ebs in particular have a very small amount of space between their spirals, and have a great amount of stickiness relative to their volume. In addition, their threads are coated in an aqueous solution that collects into translucent droplets that reflect very little UV light, reducing their visibility among insects (Opell, Bond, and Wa rner, 2006). These webs are also strong yet flexible, adding to their efficiency at capture (Swanson, Blackledge, and Hayashi, 2007). They are crucial survival, as they use them t o catch prey, collect water, and carry out courtship behavior (Foelix, 1996). Orb weaver spider s sit on the hub of their web and wait for a n i nsect to land on their web (Milne and Milne, 1980) Once the spider detects vibra tion, the spider often runs to the prey to attack. It may grab the prey from the web with its chelicerae (which subdues them immediately), bite the prey to inject a paralyzing venom, or wrap them up in silk threads while the prey struggles around (Olive, 1980) Some types of prey make their presence known more e asily than others, such as flies. They emit high frequency energy through the flapping of their wings. These attention (Masters and Markl, 1986). Having said that they generally will attack any type of small insect that lands on thei r web, and have abdomens that can expand to accommodate prey of different sizes (Jakob, Marshall, and Uetz, 1996).
Feeding behavior in Orchard s piders Govardhan 3 Orchard spiders, Leucauge sp. are a species of orb weaver spiders from the family Tetragnathidae. They are abundant in Monteverde, Costa Rica, and can be found sitting on their webs in many gardens in the area. Various types of prey can land on their webs. While some types of prey pose little to no threat, there are othe rs that may fight back when approached. If a potentially dangerous prey such as one that bites or stings, gets caught on the web, it may be in of the we b to reduce the chance that it attacks the spider (Escalante, 2015) My question addresses th is : d oes the potential danger level of prey affect whether or not orchard spiders, Leucauge sp. choose to attack? For this study, I will assume low danger prey to be fruit flies ( Drosophila sp. ) and high danger prey to be leaf cutter minima ants ( Atta sp. ). I predict that spiders will take a very predatory approach when faced with low danger prey, but will take a more non predatory, defensive approach when faced wi th a high danger prey. MATERIALS AND METHODS Study Site This research project was carried out from 14 November 2016 to 27 November 2016. I found sixty individuals of Leucauge sp. in the garden of the Monteverde Institute and the garden of Maria Estela Coghi, both in Monteverde, Costa Rica. I selected spiders which were sitting on the hubs of well formed and maintained webs. To offer the spiders a non dangerous prey, I chose fruit flies ( Drosophila sp. ) as they have no form of defense against the spide rs. They were around 2 mm in length and I collected them by opening up the organic waste can at the Monteverde Institute and swiping the top with a butterfly net. Once in the net, I transferred them to a plastic bag to store them until the trials began. I chose leaf cutter ants ( Atta sp. ) as dangerous prey, and collected them from a large ant nest on the around 4 5 mm in length, and were riding on the leaves carried by worker ants. They were chosen as dangerous prey due to their sharp mandibles which they could use to att ack the spider. However, they were the smallest caste of leaf cutter ant, and did not pose an unbeatable threat to the spiders. I used a pair of soft tweezers to collect them and I stored them in small plastic containers while I transported them to the gar dens in which I carried out the trials Feeding Behavior T rials To avoid error based on different size and age of the spiders, I classified them based on their size. I measured the length of the spiders from the top of the cephalothorax to the bottom of the abdomen, and classed them as small (1 2 mm), medium (3 5 mm), and large (6 8 mm). Additionally, I considered the fact that the relative length of time since their las t meal may have been a confounding variable with respect to whether or not they would show predatory behavior, so I attemp ted to control for this by comparing the size of their abdomen which can expand after meals (Jakob, Marshall, and Uetz, 1996), to the size of their cephalothorax. I cl assified these observations into two groups : (abdomen slightly larger than or equal in size to cephalothorax full (abdomen significantly larger than cephalothorax). I split all the spiders found into two experimental groups: low danger prey (those fed with flies ) and high danger prey (those fed with ants ). Before beginning each trial, I ensured that the
Feeding behavior in Orchard s piders Govardhan 4 spider was seated on its hub I pulled one prey individual out of their container with a pair of sof t tweezers, and then dropped the prey on the web so that it got stuck in the threads. I always made sure to drop the prey directly in front of the spider, between five and eight spir als away from the hub (this distance varied depending on the size of the w eb). Next, I observed the spiders for 25 minutes, or until the prey escaped from the web. I recorded the behaviors that the spiders displayed during this time period. Data A nalysis After collecting my data, I classified the behavior either into predatory non predatory, or other. I ran chi squared tests to check for significant differences in the frequency of predatory behavior between all fly fed and ant fed spiders, and did three more tests for small, medium, and large fly fed and ant fed spiders respec tively I also ran a chi squared test to check for significant RESULTS In the low danger prey group (N=30), there were seven small spiders, thirteen medium spiders, and ten large spiders. In the high danger prey group (N=30), there were two small spiders, sixteen medium spiders, and twelve large spiders. There were 29 spiders that were I sorted the behav ior that I observed into predatory non predatory Table 1 ) consisted of any behavior that involved the spider actively avoiding the prey, allowing it to escape, or removing it from the web. All other behaviors that the spider displayed which did not Table 1 Descriptions of b ehaviors displayed by orchard spiders. Behavior Class Behavior Description Predatory Wrapping Signs of spider wrapping the prey in silk Biting Signs of spider eating the prey/injecting venom Non predatory Prey escape d Prey either fell out of web, or moved around and/or cut webs to escape Cut prey out Spider cut threads around prey so that it dropped out Ignore Spider ignored prey altogether Other Ran to prey Spider ran towards the prey Ran away from prey Spider ran away from the prey Return to hub Spider returned to hub alone Pulling radii Spider pulled on radii to detect prey on its web Move to hub Moved prey closer to hub and hung it from a nearby thread
Feeding behavior in Orchard s piders Govardhan 5 With these observations, I constructed ethograms for fly fed spiders (Figure 1) and for ant fed spiders (Figure 2). E thograms are flowcharts that depict the behaviors that orchard spiders displayed as a sequence. Arrows connect behaviors, and they represent the frequency of that s equence. The ethograms are split by group (fly fed vs ant fed) but are not split by spider size or by relative length of time since last meal Thus, they depict the general sequences observed for each prey group. Figure 1. Ethogram of Fly fed spiders behavior sequence s after prey is dropped in web
Feeding behavior in Orchard s piders Govardhan 6 Figure 1 Ethogram of Ant fed spiders The (Table 2 ) was shown to not be a significant 2 = 1.73, df = 1 p>0.05 ) Table 2 Predatory behavior vs. Non predatory behavior between not full and full spiders. Predatory Behavior Non Predatory Behavior Not full 12 1 7 Full 18 13 When considering all spiders in the study (not taking i nto account the different sizes), m ost of the fly fed spiders showed predatory behavior, whereas most of the ant fed spiders showed non predatory behavior (Table 3 ) 2 = 17.1, df = 1, p<0.01 ). Table 3 Predatory behavior vs. Non predatory behavior betw een ant fed and fly fed spiders. Predatory Behavior Non Predatory Behavior Fly Fed 23 7 Ant Fed 7 23
Feeding behavior in Orchard s piders Govardhan 7 Next, I analyz ed t he spiders behavior relative to their body size (Table 4) Of the small sized spider trials, most of the fly fed spiders showed predatory behavior, while none of the ant 2 = 5.67, df = 1, p <0.025 ). Of the medium sized spiders a majority of fly fed spiders showed predatory behavior Interestingly, very few of the medium sized ant fed spiders showed any predatory behavior, with an overwhelming majority showing non 2 = 8.93, df = 1, p<0.01 ). Finally, of the large sized spiders, almost all of the fly fed spiders showed predatory behavior, with only one showing a non predatory action. On the other hand, of the large sized spiders that were fed ants, only a few showed predatory behavior, and most showed non 2 = 7.45, df = 1, p<0. 01 ). Table 4. Predatory vs. Non predatory behavior between ant fed and fly fed spiders, separated by spider size. Size Group Predatory Behavior Non Predatory Behavior Small Fly Fed 5 2 Ant Fed 0 2 Medium Fly Fed 9 4 Ant Fed 3 13 Large Fly Fed 9 1 Ant Fed 4 8 DISCUSSION Aside from the fact that most orchard spiders seem to run towards the prey as soon as they are dropped on the web, the bulk of their feeding behavior sequences are drastically different from each other (Fig 1, Fig 2) While the ethogram for the fly fed spiders had a more consistent progression of behavior, the ethogram for ant fed spiders was far more variable with many unique behavior patterns that were only displayed by one or two spiders. This lack of consistency ma y be because the spider s were uncertain about how to deal with this high danger prey causing them to take different approaches from each other Another reason for this type of behavior with different prey could simply be because Atta sp. are novel prey to the spiders. Generally, flying insects are a far more common type of prey since the spider webs are not often placed on ground level (Opell, Bond, and Warner, 2006) As a result, they may not have ever encountered ants before. An interestin g note is that while their behavioral sequences seem to be more irregular, the behaviors themselves are, for the most part, not unique to these circumstances (Escalante, 2015). This might raise the question: is attacking prey in an orb web a set of behavio rs that are based on instinct alone, or is it a learning process? This could be tested with further trials on the same spider individuals, recording how their behaviors against ants c hange over time. Al though these ants are dangerous due to th e ir sharp mandibles, none of them brought any physical harm to the spiders. If spiders and begin to express more regularity in their behavior sequences. on P holcidae predatory behavior supported the idea that nave spiderlings have plasticity in their behavior patterns, showing that they can adapt to novel prey. It would be interesting to see if orb
Feeding behavior in Orchard s piders Govardhan 8 weavers learned how to deal with ants as well. In addition, future studies could also compare the orchard spider behavioral traits seen with Atta sp. prey with other high danger flying insects to understand whether the random behavior is based more on the danger level or the novelty of t he prey. An interesting sequence that was highlighted by the ethograms is that the orchard spiders often moved their silk wrapped prey closer to the hub, in order to hang them from one of the first spirals that surrounded it Since orb weavers spend a lot of their time on the hub ( Milne and Milne, 1980; Biere, 1981), moving prey closer to the hub allows the spider to easily access their food while they manage and monitor their web from this position. The Cut behavior, which was the act of cutti ng the threads around the prey to release it from the web (Escalante, 2015), is another interesting behavior derived from the ethograms. Spiders only displayed this behavior with ants and did not display it with fruit flies. This may be another result of t heir uncertainty about how to respond to ants, choosing to cut them out of the web rather than leaving them in and risk getting bitten. The results of the experiment showed that across all trials, orchard spiders were more likely to show predatory behavio r when f ruit f lies were on their web, and show non predatory behavior when ants were on their web (Table 3) This provides support to the idea that orchard spiders prefer not to attack high danger prey (prey that bite). This may be because it is costly to the spiders to attack them either in energy or risk, as proposed i hypothesis. The ethograms also support this hypothesis as they depict the haphazard behavior that is displayed with ant fed spiders, which suggests that they take more time (and thus energy) figuring out how to approach this prey (Forbes, 1989; Escalante, 2015). Initially, I had suspected that the spider size influenced feeding behavior which is why I had split the data by spider size. However, all spiders, regardless of size, showed non predatory behavior more often than predatory behavior against dangerous prey (Table 4) While the large spiders were the most likely to attack ants among all the spider sizes, they too on average still showed non pr edatory behavior. Some possible reasons that the spiders attacked the ants may include that those spiders either mistook the ants for a less dangerous prey or they did not feel that the threat that the ants posed was great enough to offset the benefit of c onsuming them. These reasons may be related to a variation in temperament, where the spiders that attacked are more likely to take risks than the others (Pruitt and Riechert, 2012). A future study could try feeding the orchard spiders large, non dangerous prey, such as stingless bees, to verify if size truly is not involved. Future research could also attempt to take this further and study different prey types of varying danger levels, to see if such results are consistent throughout. On the other hand, a n overwhelming majority of orchard spiders of all sizes attacked the low danger prey (fruit flies). It is possible that this is because fruit flies do not pose a gre at threat to the spiders so there is little reason for them to pass up the opportunity of an easy food source The results also correlate with the findings of Masters and Markl (1981), showing that the high respond and less likely to ignore. The relative time since last meal also did not significantly influence feeding behavior, giving evidence to the idea that their likelihood of attacking prey has little to do with this factor. This implies that orchard spiders are opportunistic hunters, attacking any feasible
Feeding behavior in Orchard s piders Govardhan 9 prey that lands on their web. Having said that studies have shown hunger to be an important factor in orb weaver prey selection, but have expressed the difficulty of controlling this factor in the field. They also mention that it is important to consider hunger when the spiders are facing a severe food shortage, but there is little work on prey selection var iation when the spiders are not malnourished (such as the spiders in this study) (Olive, 1982; Uetz and Hartsock, 1986 ). Future experimental studies (instead of purely observational ones) could more accurately control for this factor to see if the results that I obtained are significant or due to other unknown reasons. Ultimately, these ideas for further research could increase our knowledge about orchard spiders, and orb weavers in general, bettering our understanding of th e unique arachnid biological interactions in the Costa Rican ecosystem ACKNOWLEDGMENTS First and foremost, I would like to thank Emilia Triana, my primary advisor on my project, for all the help and support she has given me throughout this p rocess Her advice and efforts have been invaluable, and I could not have done it without her. She encouraged my interest in this spider genus and helped me con tinue my research on it even when my initial research idea did not succeed. In fact, it was an observation that Emilia and I had made together that inspired the very question around which this project was based, and I am eternally grateful for that. Next, I would like to thank Federico Chinchilla, my secondary advisor, for his assistance editing my proposal, and Justin Welch for his help in the preliminary data collection stage of my project. I would like to thank Amber Domako, a fellow student in the EAP program, for the extensive and invaluable help she gave me in the early stages of my project. I would also like to express my gratitude to Andres Camacho, Felix Salazar, Sofia Arce Flores, and the rest of the EAP staff for all the wisdom and advice that they have given me over the past two months Thank you to the staff of the Monteverde Institute, for allowing me to use their campus as a base of operations for my project, for access to their lab equipment and materials, and for allowing me to exper iment on the wildlife that lived there. I also wish to thank Maria Estela Coghi for allowing me to research the biodiversity that lived in her garden. I would also like to thank Frank Joyce Jr. for spearheading this program, and encouraging me to question explore, and discover the many diverse facets of tropical biology and conservation During my time in this program, I have had experiences far outside my comfort zone, I have interacted with the biodiversity that surrounds us first hand, and I have grown to better understand and appreciate the world that we live in. Thank you for that. I would like to express my gratitude to the Leiton family for housing me during the two weeks I worked on my project. T hank you so much for welcoming me into your home, fo r your immense kindness and hospitality, and for introducing me to the many aspects of Costa Rican culture and family life. I have learned a great deal from my experiences in your household, and for that I am forever grateful.
Feeding behavior in Orchard s piders Govardhan 10 Finally, I would like to tha nk all the students of the EAP program, who have been an amazing friend group and support system and have made my experience in Costa Rica one that I will never forget. LITERATURE CITED Biere, J. M., G. W. Uetz. Web orientation in the spider Micrathena g racilis Ecology 62:2, 1981. Broom, M., G. D. Ruxton. You can run or you can hide: optimal strategies for cryptic prey against pursuit predators. Behavioral Ecology 1 6:5, 2005. Caraco, T., R. G. Gillespie. Risk sensitivity: foraging mode in an ambush p redator Ecology 67:5, 1986. Creel, S., D. Christianson. Relationships between direct predation and risk effects Trends in Ecology & Evolution 23:4, 2008. Escalante, I. Predatory behavior is plastic according to prey difficulty in nave spiderlings. J Insect Behav, 2015. Foelix, R F. Biology of Spiders New York: Oxford UP, 1996. Print. Forbes, L. S. Prey defences and predator handling behavior: the dangerous prey hypothesis Oikos 55:2, 1989. Jakob, E. M., S. D. Marshall, G. W. Uetz. Estimating fitness: A comparison of body condition indices Oikos 77:1, 1996. Janis, C. M., P. B. Wilhelm. Were there mammalian pursuit predators in the tertiary? Dances with wolf avatars. J Mammal Evol 1:103, 1993. Jeschke, J., R. Tollrian. Density depenent effects of prey defences. Oecologia 123, 2000. Levi, H. W., L. R, Levi. Spiders and their Kin. New York: Golden Books 1990. Print. Masters, W. M., H. S. Markl. Vibration signal transmission in spider orb web. Science 213, 1981. Milne, L. J., M. G. Milne The Audubon Society Field Guide to North American Insects and Spiders New York: Knopf: Distributed by Random House, 1980. Print. Olive, C. W. Foraging specializations in orb weaving spiders. Ecology 61:5, 1980. Olive, C. W. Behavioral response of a sit and wait predator to spatial variation in foraging gain. Ecology 63, 1982.
Feeding behavior in Orchard s piders Govardhan 11 Opell, B. D., J. E. Bond, D. A. Warner. The effect of capture spiral composition and orb web orientation on prey interception. Z oology 109: 4, 2006. Pruitt, J. N., S. E. Riechert. The ecological consequences of temperament in spiders Current Zoology 58, 2012. Uetz, G. W., S. P. Hartsock. Prey selection in an orb weaving spider: Micrathena gracilis (Araneae:Araneidae). Psyche 94, 1987. Swanson, B. O., T. A. Blackledge, C. Y. Hayashi. Spider capture silk: performance implications of variation in an exceptional biomaterial. Journal of Experimental Zoology 307A:11, 2007.