Spider Webs in Strangler Figs Daniel 1 Vertical stratification of w eb-building spiders in strangler fig trees? Justin N. Daniel Department of Ecology and Evolutionary Biology University of California, Santa Cruz EAP Tropical Biology and Conservation Program, Fall 2017 15 December 2017 Abstract Web building spiders rely on their silk structures to capture prey. The morphologies of the webs produced by spiders vary greatly and they provide insight into spiders behavior and ecology. Strangler fig trees, due to a life history that results in trunks with complicated structures, host a great diversity of web-building spiders. I studied the vertical distribution of the web -building spiders that inhabit strangler figs I used a Single Rope method of canopy access to climb five strangler f ig trees to record webs and webbuilding spiders. Whether a vertical stratification of these spiders exists remains unclear. Apparent patterns, which were observed on four trees, among 12 morphospecies and 7 families, here are analyzed and discussed. Web types, families and species were not stratified with the exceptions of the species Tengella radiata and the family Agelenidae. Existe estratificacin vertical de araas tejedoras de tela en higuerones estraguladores? Resumen Las araas que construyen telas dependen de sus estructuras de seda para capturar presas. La morfologa de las telaraas varan mucho y pueden proporcionar informacin sobre su comportamiento y ecologa. Los higuerones estranguladores (Ficus spp.), debido a una historia de vida que resulta en troncos con estructuras complicadas, albergan una gran diversidad de araas que construyen telas. Estudi la distribucin vertical de araas tejedoras de tela que habitan higuerones estranguladores. Utilic el mtodo de Cuerda nica de acceso a dosel para trepar los higuerones estranguladores para examinar las araas tejedoras de tela presents y sus telaraas. Los resultados demuestran una que no hay una clara estratificacin vertical de las 12 morfoespecies de araas encontradas. Las familias de araas, las morfoespecies, ni las telaraas mostraron una estratificacin vertical, con excepcin de Tengella radiata y la familia Agelenidae. Sin embargo, se observaron posibles patrones en cuatro rboles, los cuales analizo y discuto en este studio. Spiders are abundant generalist predators with a dominant presence in almost every ecosystem, and silken prey-capture webs are a key characteristic contributing to the ecological and evolutionary success of this group (Garrison et al. 2016). Although web morphology (particularly at the level of species ) remains understudied (Eberhard 1990), it is accepted that morphologies of webs are unique to the spiders by which they are produced (Foelix 2011). Although an individual spider may prey on a great diversity of arthropods, different types of webs employ different mechanisms for prey capture, and are thus more efficient in capturing certain types of prey (Vollrath, Fritz, and Paul Selden
! Spider Webs in Strangler Figs Daniel 2 2007). For instance, orb webs, whi ch are most efficient in the capture of flying insects, have threads with sticky gum to ensnare prey. Funnel webs, which lack sticky gum, function by catching prey that fall or walk onto their surfaces, signaling the spider to come out of its retreat to capture prey (Foelix 2011). T he structure of a web represents an intimate interaction between morphology and behavior, and h ence, silk structures can be used as indicators for a spider's specific ecology and behavior ( Vollrath, Fritz, and Paul Selden 2007 ) How web building spiders distribute themselv es with respect to the ground could be the result of a number of important ec ological and biological factors. Perhaps the vertical stratification of prey arthropod species (Dial et al. 2006 ) influence s the distribution of spiders and their respective web morphologies. Epiphyte abundances and compositions at different heights also may play a critical role. E nvironmental differences such as sunlight, wind, humidity and temperature at dif ferent heights fro m the ground may be determinants of spider species distributions. D ifferences in spider species' modes of long and short distance travel and/or dispersal (Foelix 2011) are also extremely infl uential biological factors that determine a spider's access to a given surface Therefore, an understanding of the vertical distribution (a stratification) of web building spider communities, or a lack thereof, in tropical forests would lend itself to a better understanding of the behavior and ecology of these spiders. Strangler figs, which belong to the genus Ficus (Moraceae) germinate in the canopies of the r ainforest, and grow by dropping aerial roots to the ground that strangle and eventually k ill their host trees ( Dobzhansky, Theodosius, and MurÂa Pires 1954 ). The life history of strangler fig trees results in some very tall trees with complicated structures such as large buttresses, and hollow trunks with many holes and crevices. These structures, according to my observations, provide an excellent structural surface for spider webs. This makes these trees a habitat for diverse spider communities I attempted to answer the following question: Is there a link between web building spiders' web morphology and the locations of their webs along the vertical strata of strangler fig trees? I hypothesize some degree of stratification exists I expected to see sheet and funnel webs (and their respective families) closer to the bases of these trees because these webs are most efficient at catching falling prey and I exp ected to see orb and tangle webs higher in the trees because they are more efficient at catching flying prey (Foelix 2011). MATERIALS AND METHODS I sampled Ficus Trees within the limits of the town of Monteverde Costa Rica, within 1 kilometer of the Monteverde Institute. I collected data for two weeks in late September I selected trees that were tall (lowest branch above 10 meters) straight and relatively vertical and could be climbed safely using the Single Rop e Method for Canopy Access (Dial, Roman, and S. Carl Tobin 1994 ). I sampled from the ground up along a hal f meter wide transect. I worked within plots, which had an area of approximately 0 .32m 2 at intervals of approximately 1.5 meters, which were established by sampling the webs that inte rsected the transect from my waist to my head (~.75m) before climbing until my feet rested whe re the top of my head had been to continue sampling I repeated this until I reached the branch that held my weight I did not include webs that were found
! Spider Webs in Strangler Figs Daniel 3 on the inside of the tree to avoid a bias that could result from the difference in the likelihood of these webs to be overlooked With e ach web that I encountered in the plots, I measured its he ight from the base of the tree and the width of the web. I recorded the type of web (tangle, sheet etc.), the spider's family, the spider 's species or morphospecies identity if possible and whe ther a spider was present or absent. I then took photos of the webs and collected spe cimens to be identified in the lab when possible if they could not be identified in the field Many of the spiders were too small to be iden tified accurately in the field. In the lab, I photographed and identified collected specimens. I preserved adult spe cimens in ethanol as voucher specimens O ne way ANOVA analyse s were employed using JMP to analyze data and calculate F and p values. I divided web types into five categories for the purpose of this study: (1) Funnel webs, characterized by a horizontal shee t like capture surface with a retreat or tunnel at one end, are built by spiders in the families Tengellidae and Agelenidae, (2) Orb webs, characterized by radial threads and sticky silk spiral, are woven by spiders belonging to the Araneidae, Ulorbidae and Tetrignathidae families, (3) tangle webs which often contain a retreat of dirt, moss, or leaves suspended by silk, are constructed by spiders in the family Theridiidae, (4) sheet webs which consist of a sheet suspended by tangles above and below, a re produced by spiders of the families Pholcidae and Lyniphiidae (Levi et al. 2002), and (5) curtain webs, which are webs constructed of ec ribellate silk in crevices by spiders in the family Dipluridae ( Figure 6 ). RESULTS I recorded a total of 82 webs of 12 morphospecies, and 32 individual spiders along the trunks of five Ficus trees This means that the majority of webs were vacant. I excluded the data from one of the trees from my models for reasons explained in my discussion. I did not find any statistically significant vertical distribution patterns of web types (F=0.74, p=0.57) web building spider families (F=0.68, p=0.67) or species (F=1.13, p=0.39) Webs types are vertically distributed evenly among the 4 height bins (Fi gure 1; F =0.74, p=0.57). The orb webs do not appear to be vertically distributed evenly, but this was a categ ory with a total count of only two Most families of web building spiders seem to be stratified evenly, however, I observed that the families Agele nidae and Tengellidae which are in the same web type category, did not inhabit the same strata; I only recorded Tengellidae webs at the tops and bottoms of the trees, while the greatest number of Agelenidae webs were found from 4 meters to 12 meters (Figure 2 F=0.68, p=0.67 ) The only species that were found in one height bin were those that were encountered only once (Figure 3 F=1.13, p=0.39 ) Theridiidae D could be found in three of the four height bins (n=5, figure 1), but I observed the greate st densities at a height of 6 meters.
! Spider Webs in Strangler Figs Daniel 4 Figure 1: V ertical distribution of relative abundances web types on four strangler fig trees ( n=82 F=0.74, p=0.57 ) Figure 2: Vertical distribution of relative abundances of web building spider families on four Ficus trees (n=82 F=0.68, p=0.67 ) "#$% "#$& "#$' "#&( "#$% "#$& "#$' "#&(
! Spider Webs in Strangler Figs Daniel 5 Figure 3: V ertical distribution of species and morphospecies on four strangler fig trees (n=82 F=1.13, p=0.39 ) D ISCUSSION Spiders do not appear to be site specialists with regard to the vertical strata of strangler figs, with the possible exceptions of the families Tengellidae and Agelenidae It is possible that spider distribu tions are not partial to vertical strata, but are instead dependent mostly upon spiders' abili ti es to access surfaces. This is determined by the interaction between their aerial dispersal potential and forest patch connectiv ity (Bonte et al. 2004). Although my data alone does not allow for conclusions regarding the hypothesis that web building spiders are vertical ly stratified on Ficus Trees, it does reveal patterns that demand further research t o understand them. The segregation of Agelenidae and Tengellidae webs is a particularly interesting avenue for investigation, especially because Tengella radiata is a spider species that is endemic to Costa Rica, and its distribution among the vertical strata of the forest h as not been investigated prior to my study. It is unknown weather patterns similar to the one that I observed exist in other trees, or other types of forest. Tengellidae and Agelenidae webs share a similar funnel web design (Foelix 2011 ) Perhaps these sp iders are pa rtitioned to avoid competition, while selecting sites that support their web design I postulate that Tengellidae webs, with a mean width of 29.8 cm (much larger than Agelenidae), are partial to sites with greater structural complexity (buttresses and branches) greater number of epiphytes (at higher strata) and closer "#($
! Spider Webs in Strangler Figs Daniel 6 adjacent trees (which might be found with greater ease at the b ottom of these trees). The middle strata of the Ficus which is host to a lesser number of epiphytes than the top, and which, according to my qualitative observations are sometim es less structurally complex than the top or bottom may be more conducive to Agelenidae webs with a mean width of 12.5 cm many of which, were built on mosses that protrude d from the trunk (Figure 4 ) Aggregations of these webs could allow for spiders to have an easier time finding mates. One of the trees that I had initially chosen to sample was a strangler fig whose host tree had died, and was rotting from inside the Ficus The deterioration had caused an abundance of wood dust to fall into every web, making them inhospitable for the spiders that built them. This tree, for reasons that I can only speculate, produced a multitude of outliers from the rest of my data T he most strange and interesting of these was a multitude of Tengellidae webs, many of which were abandoned, at heights where they had not been observed in other trees, and many more abandoned webs from all families. I encountered only two spiders on this tree (bot h T. radiata) For these reasons, I elected to exclude this tree from my data. Figure 5 shows a retreat or we b or both that I encountered of unknown precedence I encountered six of these, with a mean height of 12 mete rs, the lowest of which was at seven meters. This type of structure was always built in the very top of a crack, with the retreat at the top, and a sheet like structure spanning the crack, and extending down 3 or 4 centimeters from the retreat. This sheet like structure most likely serves to protect the retreat. More investigation must be done to understand this structure and its function. If the silk structure is a prey ca pture web, then this is an example of a height specific web building spider. Because my data collection period spanned on ly two weeks in September, my results only provide insight into the distributions of these spiders during that time of the year. Various temporal factors could have huge influence on these spiders' pref erred locations. For instance, s ome species of wasps are engaged in a n obligate relationship with strangler figs. These wasps aggregate in the fruiting areas (canopies) of these trees seasonally, changing the prey composition of spiders, and possibly changing the arthropod predator compositions as well. Seas onal bird migrations may also seasonally influence site selection for web building spiders. This study provides some insight into the way that web building spiders and their webs are distributed along the heights of strangler figs, but a substantial amoun t of further research is in order if we are to gain more understanding about the link between the ecology and behavior of web building spiders and their vertical distributions.
! Spider Webs in Strangler Figs Daniel 7 Figure 4 (left) : The web of Agelenidae A is small, and often supporte d by mosses that protrude from the trunk of the Ficus tree. Figure 5 (right) : This unique retreat likely belongs to a spider in the family Salticidae I t is unknown what investigation has been done thus far to understand this type of retreat design The function of the sheet like structure beneath the retreat is unknown. Figure 6 (left) : This web is built in a crevice. It is a matrix of fin e ecribellate sheets. It is unknown which family is responsible for this type of web design but it is likely D ipluridae
! Spider Webs in Strangler Figs Daniel 8 Acknowledgements Thank you to Frank Joyce, Emilia Triana, and AdrÂŽs Camacho for superb academic support, Eric McAdam Marvin Hidalgo Kaylor MuÂ–oz and FÂŽlix David Huertas Salazar for expert support in the field, Jorge Mora, Myriam Scally the Trostle family, and Mauricio Valverde for allowing me to conduct research on their land my peers for support, with a special thank you to my supurb peer reviewer, Anissa Carter, and to HaydÂŽe, Ginneth, Arturo, and Valentina for welcoming me in to their family and homes. References Garrison NL, Rodriguez J, Agnarsson I, Coddington JA, Griswold CE, Hamilton CA, Hedin M, Kocot KM, Ledford JM, Bond JE. (2016) Spider phylogenomics: untangling the Spider Tree of Life. PeerJ 4:e1719 Foelix, Rei ner F. Biology of Spiders. 3rd ed., Oxford University Press, 2011. Levi, Herbert Walter, et al. Spiders and Their Kin. St Martin's Press, 2002. Dobzhansky, Theodosius, and JoÂ‹o MurÂa Pires "Strangler Trees." Scientific American, vol. 190, no. 1, 1954, pp. 78 81., Dial Roman J., et al. "Arthropod Abundance, Canopy Structure, and Microclimate in a Bornean Lowland Tropical Rain Forest." Biotropica, vol. 38, no. 5, 2006, pp. 643 652. Dial Roman, and S. Carl Tobin. Description of arborist methods for canopy access and movement ." Selbyana, vol. 15, no. 2, 1994, pp. 24 37. JSTOR, JSTOR, Vollrath, Fritz, and Paul Selden. "The Role of Behavior in the Evolution of Spiders, Silks, and Webs." Annual Review of Ecology, Evolution, and Systematics, vol. 38, 2007, pp. 819 846 Bonte, D., Baert, L., Lens, L. and Maelfait, J. P. (2004), Effects of aerial dispersal, habitat specialisation, and landscape structure on spider distribution across fragmented grey dunes. Ecography, 27: 343 349. doi:10.1111/j.0906 7590.2004.03844.x Eberhard, William G. "Function and phylogeny of spider webs." Annual Review of Ecology and Systematics 21.1 (1990): 341 372.