Diversity of Arboreal and Ground dwelling Ant Species in a Primary Forest and Grazed Pasture Marie F. Bostrom Department of Wildlife Ecology and Conservation, University of Florida ABSTRACT Ants are very important to ecosystem function. They help to aerate soil, disperse seeds, distribute nutrients, and regulate other insect populations Holldobler and Wilson 1990. Since a non forested area will have a larger microclimate variation than a forest, it is expected to support more species Torres 1984. In addition, arboreal ants can be expected to have a lower overall diversity than terrestrial ants, since there may be a smaller number of species adapted to tolerate the elevated range in abiotic conditions that occurs in arboreal environments Nadkarni a nd Longino 1990. Ants were collected over eight days, using traps baited with tuna and honey. The ground dwelling ant species were significantly more diverse than arboreal ant species modified t test, t = 3.9172, P < 0.001, and had the highest richness and capture rates. No significant difference was found in the species diversity of pasture and forest ants modified t test, t = 0.8785, 0.5 > P > 0.2. These results may indicate that there are a lower number of ant species adapted to tolerate arboreal co nditions, or that the forestÂ€s lack of microclimatic heterogeneity contributes to a lower richness and diversity of forest species. RESUMEN Las hormigas son muy importantes a la funcin del ecosistema. Ellas se ayudan a agitar la tierra, dispersar las semi llas, distribuir los nutritivos, y regular las poblaciones de otros insectos Holldobler and Wilson 1990. Porque desde la area sin el bosque va a tener un clima ms grande variacin, se pree tener una diversidad total ms grande que el bosque primero Tor res 1984. Adems, las hormigas arbreas van a tener una diversidad total mas bajo que hormigas terrestres, porque haya un nmero mas bajo adaptado de tolerar la extensin elevada de condiciones abiticas Las hormigas fueron recogidas para ocho das, usando trampas con atn y miel. Las especies de las hormigas que viven en el suelo fueron significativo mas variadas que las especies arbreas modified t test, t = 3.9172, P < 0.001, y tuvo el rango mas alto del botn. Una diferencia importante no fue encontrada entre la diversidad total de las hormigas del potrero y el bosque modified t test, t = 0.8785, 0.5 > P > 0.2. Estos resultados demostraron que haya un mas bajo numer o de especies adaptado tolerar condiciones arbreas, o que la diversidad estructural del bosque se conceda un numero mas bajo de especies. INTRODUCTION Tropical forests are estimated to House one half of the earthÂ€s currently described 1.4 million species Wilson 1998. Insecta is the dominating class by far, with more than
750,000 described species Wilson 1992. Insects are known to be vital contributors to pollination and dispersal systems, trophic interactions, organic decomposition, and biomass regulat ion in their habitat Wilson 1992. Ants Formicidae in particular, are extremely important to ecosystem function, helping to aerate soil, disperse seeds, distribute organic matter, and regulate many insect populations Holldobler and Wilson 1990. In Co sta Rica, a 1991 1992 survey estimated the amount of tropical forestlands being converted for agricultural use to be between 5,000 and 10,000 hectares per year Lutz et al 1993. This clearing of land both destroys and creates habitats. Species that were o nly able to survive in forest conditions will migrate or perish, but species adapted to edges or open habitat will flourish. In other words, deforesting a landscape will change its species diversity. In a study of Puerto Rican ant communities in a sub trop ical wet forest, a grassland, and an agricultural crop area, Torres 1984 found the greatest ant species richness in the agricultural cropland. Additionally, he found that microhabitat is important to antsÂ€ nesting and foraging activities. He hypothesized that the greatest number of species found in the agricultural area was due to a more variable microclimate, greater food overlaps, smaller guilds, and an increased amount of aggressiveness among the present ant species Torres 1984. In addition, the fore stÂ€s comparatively great structural diversity led to a reduced variation in microclimate range, and reduced the number of species present Torres 1984. This study seeks to compare the ant species richness and diversity of primary growth lower montane wet forest with a currently grazed cattle pasture, with the expectation that the pasture will have a higher richness and diversity. In order to take an inventory of FormicidaeÂ€s richness and diversity, it is necessary to sample from both of their two principl e microhabitats, the ground and the trees. Even within these microhabitats, one can expect to see differences in species diversity Longino and Nadkarni 1990. Both arboreal and ground dwelling ants have similar functions, such as enriching the area of the ir nest with dead organic material to inadvertently promote soil or humus richness, dispersing seeds, and preying on other arthropods. In addition, many studies have illustrated that invertebrate communities are very sensitive to microclimatic differences in moisture, temperature, and sunlight Nadkarni and Longino 1990; Longino and Nadkarni 1990 and arboreal species will experience a larger range of these abiotic factors Nadkarni and Longino, 1990. Fewer species of ants may be adapted to these conditio ns, leading to a lower diversity when compared to terrestrial environments. With this in mind, arboreal ant diversity and richness is expected to be lower than terrestrial ant diversity and richness in both primary forest and grazed cattle pasture. In thi s study, the diversity and richness of arboreal ants is compared to the diversity and richness of ground dwelling ants in both a grazed pasture and lower montane primary forest.
MATERIALS AND METHODS The first study site was a currently grazed dairy catt le pasture in Santa Elena, and the second was the forest behind the Estacin Biolgica Monteverde, Puntarenas, Costa Rica. Data were collected over eight non consecutive days in early November, 2000. Ants were baited to the small plastic vial traps with a mixture of tuna and honey. In the pasture, six isolated trees were selected, ranging from 46.3 cm to 85.4 cm DBH diameter at breast height. Four species were represented: Cedrella tonduzii Meliaceae, Sapium Anacardiaceae To set the arboreal traps, a s ling shot was used to shoot a rope over the highest point that the trunk branched. Ant traps were taped to the rope at three different heights, top where the rope passed the highest point, midway between the highest point and the ground, and low one t o two meters from the ground. In all cases the rope touched the trunk. Next, three ground traps were set at approximately two meters from the trunk, on three sides of the tree. The same was repeated in the forest. The forest represented were: Quercus brun essii Fagaceae, Schefflera rodriguezinna Araliaceae, Gymnoiporia haberiana Celastraceae, Ocotea tonduzii Lauraceae, Ficus sp. Moraceae, and one unidentified tree. DBH ranges were from 64 cm to 301 cm. Pasture arboreal traps were set at heights ran ging from one to 11 meters. Forest arboreal traps were set ranging from two to 14 meters. The traps were checked about every 24 hours, and when ants were present they were collected in alcohol and taken to the lab for identification. The ants were identif ied to genus using a Longino 1995 key. A Shannon Weiner diversity index was used to determine ant diversity between the sites, and a modified t test Zarr 1984 was used to determine significant differences. A Morisita similarity index was also calculate d between sites RESULTS In the pasture, four subfamilies were represented, including 10 genera and 22 morphospecies Table 1. Eleven species were only ground dwelling, seven species were arboreal, and four species were in both microhabitats Table 2. Ground dwelling pasture species diversity was significantly higher modified t test, t = 2.796, 0.01 > P > 0.005 than arboreal pasture species diversity Shannon Wiener diversity index, HÂ€ = 1.096; HÂ€ = 0.876. The similarity index between pasture ground dwelling and arboreal species was low MorisitaÂ€s index of similarity, CH = 0.0569. In the forest, three subfamilies were represented, including five genera and 14 morphospecies Table 1. Nine species were found only on the ground, three were strictly a rboreal, and two species were in both microhabitats Table 2. Ground dwelling forest species diversity was significantly higher modified t test, t = 6.192, P < 0.001 than arboreal species diversity Shannon Wiener diversity index, HÂ€ = 0.778; HÂ€ = 1.038 The similarity of arboreal and ground dwelling forest species was also low MorisitaÂ€s index of similarity, CH = 0.150.
Ground dwelling pasture species had a higher capture frequency Fig. 1, as well as a higher richness than ground dwelling forest species Table 1. Also, the ground dwelling pasture species showed a slightly higher, but insignificant diversity modified t test, t= 1.156, 0.50 > P > 0.20 than the ground dwelling forest species Shannon Wiener diversity index, HÂ€ = 1.096; HÂ€ = 1.03 7. Their similarity was very low MorisitaÂ€s index of similarity, CH = 0.0347. Additionally, arboreal pasture species were captured more frequently Fig. 1, and had a higher richness than arboreal forest species Table 1. Arboreal pasture species diver sity was insignificantly higher modified t test, t = 1.755, 0.10 > P > 0.05 than arboreal forest species diversity Shannon Wiener diversity index, HÂ€ = 0.0876; HÂ€ = 0.778. Their similarity was higher than ground dwelling pasture and forest species, but still low MorisitaÂ€s index of similarity CH = 0.140. Ground dwelling species over both pasture and forest were found to have a significantly higher diversity modified t test, t = 3.917, P < 0.001 than arboreal species over both habitats Shannon Wiener diversity index HÂ€ = 1.288; HÂ€ = 1.023. Pasture species diversity was not significantly different modified t test t = 0.879, 0.5 > P > 0.2 than forest species diversity Shannon Wiener diversity index, HÂ€ = 1.157; HÂ€ = 1.209. However th e pasture and forest species were not similar MorisitaÂ€s index of similarity, CH = 0.0411. Both ground dwelling and arboreal species were captured more frequently in the pasture Fig. 1. DISCUSSION As expected, ground dwelling species diversity was gre ater than arboreal species diversity in the pasture. Many studies have documented that invertebrates are very sensitive to abiotic, microclimatic changes in moisture, temperature, sunlight, and wind Nadkarni and Longino 1990. Arboreal ants are generally subject to a greater range of these abiotic conditions than are ground dwelling ants Nadkarni and Longino 1990. Therefore, it may be that fewer species are able to survive in these conditions. The low index of similarity between these habitats is in acco rdance with Longino and NadkarniÂ€s 1990 finding that ground nesting species are typically of a different genera than tree nesting species. Ground dwelling species diversity was significantly higher than arboreal species diversity in the forest. The diffe rence in diversity may be explained by a lesser amount of species adapted to tolerate arboreal conditions, as in the pasture. An ant diversity study in the Monteverde Cloud Forest Preserve demonstrated that canopy ant richness was lower than ground dwellin g ant richness Longino and Nadkarni 1990, which is in accordance with these results. The low similarity of ground dwelling to arboreal forest species may be explained by the different chemical and physical composition of ground soil and canopy soil Long ino and Nadkarni 1990. This being where ants nest, if follows that different species would be inclined to nest only in a certain habitat. This agrees with Longino and HansonÂ€s 1995 finding that there is very low overlap in species between the forest flo or and the canopy.
Ground dwelling pasture species were more frequently captured Fig. 1, and had a higher richness than ground dwelling forest species Table 1. The fact that pasture ground dwellers had a higher capture rate may be explained by Carroll and Janzen 1973. They found that the more unpredictable a habitat is, the greater the competition for food resources. During the study, three out of six pasture areas were disturbed by cattle. This disturbance changed the study area by reducing vegetati on and turning up soil considerably, creating a very unstable environment. Furthermore, the ground dwelling pasture species showed a slightly higher, but not significantly different diversity than the ground dwelling forest species. Torres 1984 found tha t non forested areas had a high number of species, partially due to more food resource overlaps, and a greater number of microclimatic habitats. It is a possibility that these reasons contributed to the higher pasture diversity in this study. The similarit y was extremely the two habitats which may be explained by the lack of leaf litter and understory vegetation in the pasture, allowing a different suite of species to exist in each habitat. Arboreal pasture species were slightly more diverse than arboreal forest species. However, this difference was not significant. In the forest there are many epiphytes, as well as vines and lianas, which may actually support a higher arboreal ant community diversity by providing habitat and additional food and water reso urces, but further studies are necessary to determine this. Additionally, the highest ant traps were located ranging from 12 to 14 meters against the trunk, which may have been only about 2/3 of total tree height. It is possible that these large trees har bored ant colonies that did not forage below the canopy. Furthermore, all of the baited pasture trees had epiphyte loads of at least 10% personal observation. These epiphyte mats may have contributed to the number of species the trees supported, since it is known that at least two species captured here i.e. Solenopsis zeteki and Pachycondyla sp nest on branch surfaces under epiphyte mats Longino 2000. The similarity of arboreal pasture and arboreal forest species was still low, but slightly higher th an the pasture and forest ground dwelling species. The most dominant genus over all study sites was Pheidole spp., member of the Myrmicinae subfamily Table 1. Pheidole spp. are generalized scavengers with 27 species in Monteverde alone Longino, J. 2000 personal communication. They are common in trees and on the ground, across many habitats Longino and Hanson 1995, and this is the likely reason for their frequent capture. Solenopsis spp. was found in trees and on the ground in both habitats as well. They are known to be widespread in leaf litter, and a common genus under epiphyte mats Longino and Hanson 1995. The capture of three Ponerinae genera Ectatomma, Leptogenys Pachycondyla could be explained by the fact that they are predacious ants, and were either eating the tuna or preying on other occupants of the traps at the time of capture. One genus Nomamyrmex sp. in the subfamily Ecitoninae was captured in the pasture. Ecitoninae is a subfamily of generalized scavengers, whose foraging behavior is important to many. Flies, beetles, silverfish, antbirds, and others depend on foraging columns and raids to either flush out or capture their food Longino and Hanson 1995. EcitoninaeÂ€s presence in the pasture may have important ecological implications for this reason. Two genera of fungus growing ants subfamily Myrmicinae, tribe Attini, Acromyrmex sp., Atta sp. were also
captured, which was unexpected, considering these genera feed on fungus, not tuna. In Monteverde, Camponotus sp., subfamily Formi cinae are strictly arboreal scavengers Longino, J. 2000, personal communication, however, they were captured foraging on the pasture ground. This makes sense considering the pasture trees are isolated, and its inhabitants are not able to forage from tre e to tree. Ground dwelling species diversity was significantly higher than arboreal species diversity over both habitats, which supported the hypothesis. Even though pasture species diversity was not significantly higher than forest diversity, a trend was present which supported the idea that pasture diversity is higher. The forest is structurally complex, containing a large amount of ant suitable habitats and food sources Carroll and Janzen 1973. However, this complexity could act as a buffer to sharp c hanges in microclimate habitats, which in turn would reduce the number of these habitats Torres 1984. A smaller number of microclimatic habitats would lead to smaller number of species, according to Torres 1984. This may explain the trend towards highe r pasture species diversity. ACKNOWLEDGEMENTS I thank Mauricio Garca for his guidance in this study, for his patience, and for the extra time he spent helping me in every step of this project. I thank Andrew Rodstrom for his sling shot aim, his patience, and valuable ability to find whatever you need. And thank you Tim Kuh man for helping us through the semester. Thank you Karen and Alan Masters for being so inspirational, for giving us your knowledge and for being so down to earth. Many thanks to Julia Crown for her friendship, study sessions, and reading this paper a milli on times. Thanks to Carolyn Weaver for giving me good suggestions. I want to thank the Torres Leitn family for opening their home to me and for their unselfish kindness. Thank you Jos Manuel Torres for translating, teaching me Spanish and identifying tre es for me. And I especially thank Elsida Torres Leitn for being so incredibly beautiful.
LITERATURE CITED Carroll, C.R. and D.H. Janzen. 1973. Ecology of foraging ants. Journal not known in data base. Holldobler, B. and E.O. Wilson. 1990. The ants. Page 374. Harvard University Press, Cambridge, Massachusetts, USA. Longino, J.T. 2000. The ants of Monteverde. Pages 132 136 in N.M. Nadkarni and N.T. Wheelwright, editors. Monteverde: ecology and conservation of a tropical cloud forest. Oxford Univ ersity Press, New York, New York, USA. Longino, J.T., and P.E. Hanson. 1995. The ants Formicidae. Pages 588 620 in P.E. Hanson and I.D. Gauld, editors. The hymenoptera of Costa Rica Oxford University Press Inc., New York, USA. Longino, J.T. and N.M. Nadkarni. 1990. A comparison of ground and canopy leaf litter ants Hymenoptera: Formicidae in a neotropical montane forest. PSYCHE 97: 81 93. Lutz. E., M.W. Vedova, H. Martnez, L.S. Romn, R.L. Vazquez, A. Alvarando, L. Merino, R. Celis, J. Huising 1993. The World bank Environmental Department. Paper No. 61. Nadkarni, N.M. and J. T. Longino. 1990. Invertebrates in canopy and ground organic matter in a neotropical montane forest, Costa Rica. Biotropica 22: 286 28. Torres, J.A. 1984. Diversity and abundance of ant communities in Puerto Rico. Biotropica 16: 296 303. Wilson, E.O. 1992. The diversity of life W.W. Norton and Company, inc. New York New York, USA. P. 131 137. Wilson, E.R. and F.B. Baird. 1988. The current state of biological di versity. Pages 3 18 in E.O. Wilson and F.M. Peter, editors. Biodiversity. National Academy Press, Washington, D.C., USA. Zarr, J.H. 1984. Biostatistical Analysis. Prentice Hall Inc. Englewood Cliffs, New Jersey, USA.
Table 1. Number of different morphological species of each genus captured in each habitat. The forest has an overlap of two species, and the pasture has an overlap of four. Subfamily Genus Forest Ground Forest Tree Pasture ground Pasture tree Myrmicinae Pheidole 6 2 7 7 Myrmicinae Solenopsis 2 2 1 1 Myrmicinae Acromyrmex 0 0 1 1 Myrmicinae Atta 0 0 1 0 Ponerinae Ectatomma 1 0 1 0 Ponerinae Leptogenys 0 0 0 1 Ponerinae Pachycondyla 1 1 1 0 Ecitoninae Nomamyrmex 0 0 2 0 Formicinae Componotus 0 0 1 0 Formicinae Paratechina 1 0 0 1 Totals 11 5 15 11 Table 2. Microhabitat of each genera found in either the pasture or forest collection site. Genera Ground Dwelling Arboreal Pheidole X X Solenopsis X X Acromyrmex X X Atta X Ectatomma X Leptogenys X Pachycondyla X X Nomamyrmex X Componotus X Paratechina X X