Distribution of stingless bees Meliponinae at two elevations and their reaction to deforestation on the Pacific slope of Costa Rica Christina Murphy Department of Environmental, Population, and Organismic Biology, University of Colorado Boulder _____________________________________________________________________________________ Abstract The distribution and abundance of Meliponinae and Apidae, as well as the effects of deforestation on Meliponinae, at two elevations on the Pacific slope of Costa Rica were studied. All apids were collected using a pan trap method at 700 m and 1700 m in forest and pasture. The low elevation had higher diversity of Apidae and Meliponinae than the high elevation. These findings are consistent with InBIO Instituto Na cional de Biodiversidad data on overall elevational trends in Meliponinae and Apidae, yet four species of Meliponinae found in this study add new elevational ranges to InBIOÂ€s records. This study found that apids and stingless bees had more diversity in p astures than in forests. Also there was no significant difference in Meliponinae diversity between the forest and pasture at the low elevation but more diversity was found in the pasture while individuals of species were more abundant in the forest. This w as probably due to forest nesting stingless bees foraging in pasture. These results show that deforestation could change the behavior of stingless bees and thus have drastic consequences for the pollination of forest plants and ecosystems as a whole. This study found that Meliponinae and Apidae are more diverse at lower elevations and that Meliponinae are affected by deforestation. Resumen Se estudiÃ³ la distribuciÃ³n y la abundancia de Meliponinae y Apidae, asÃ como los efectos de la deforestaciÃ³n en Meliponinae, a dos elevaciones diferentes en la vertiente PacÃfica de Costa Rica. Las abejas fueron colectadas con un mÃ©todo de trampa de cazuela a dos elevaciones 700 m y 1700 m en el bosque y en pasto. La elevaciÃ³n baja tuvo una diversidad mÃ¡s alta de Ap idae y Meliponinae que la elevaciÃ³n alta. Estos res ultados estuvieron de acuerdo con los datos de InBIO Instituto Nacional de Biodiversidad en las tendencias generales de la diversidad a diferentes elevaciones en Meliponinae y Apidae, sin embargo cuatro especies de Meliponinae que se encontraron en este estudio aumentaron los rangos de elevaciones de los datos del InBIO. Este estudio descubriÃ³ que Apidae y Meliponinae fueron mÃ¡s diversos en los pastos que en los bosques. TambiÃ©n , se encontrÃ³ que no hubo u na diferencia significativa en la diversidad de Meliponinae entre el bosque y el pasto a la elevaciÃ³n baja pero si una diversidad mÃ¡s alta en el pasto mientras que individuos por especie fue mÃ¡s abundante en el bosque. Esto probablemente se debiÃ³ a que los Meliponinae que anidan en el bosque hurgan en el pasto. Estos resultados mostraron que la deforestaciÃ³n pudo cambiar el comportamiento de Meliponinae y tiene consecuencias drÃ¡sticas en la polinizaciÃ³n de plantas del bosque y ecosistemas. Este estudio enco ntrÃ³ que Meliponinae y Apidae fueron mÃ¡s diversos a elevaciones inferiores y que Meliponinae se encontrÃ³ afectada por la deforestaciÃ³n.
Introduction Twenty thousand species of Apidae all bees have been described for the tropics with the majority belongi ng to the subtribe Meliponinae, the stingless bees Roubik 1989; Gri swold et al. 1995. Despite their abundance there is little known about them Roubik 1992. Meliponinae are found exclusively in the tropics and subtropics Roubik 1989. In Costa Rica the re are 12 genera and approximately 50 species described Lobo 2000. Stingless bees have very complex colonies and their nest structures are typically made of wax, resin, gum, mud, feces, plant fiber, and leaves Wilson 1971; Hogue 1993. Colonies can numb er in the tens of thousands and have division of labor, advanced chemical alarms, and recruitment Wilson 1971. Meliponinae are important for ecosystem processes. They are influential in the reduction of carrion in the tropics Breed et al. 1999. Stingle ss bees are also believed to be the most important pollinators in tropical forests Liow et al . 2001 and are the only known pollinator of some palm and Piper species Janzen 1983. The distribution of stingless bees and apids across elevation is not well studied, yet it is important to their conservation and to understanding their role in ecosystems. There are many factors that could affect Apidae elevational distribution. Lobo 2000 found that most hymenoptera are restricted to only one or two of these altitudinal zones: lowland 0 800m, intermediate 800 2000m, or montane 2000 + m. Lowland species of hymenoptera often have a larger distribution and montane species are often endemic Lobo 2000. Hymenoptera tend to be unc ommon at high elevations because there are great fluctuations in temperature, abundant cloud cover, fewer nesting sites, more parasites, and temperatures too low for foraging Lobo 2000. These same factors limiting the Hymenoptera in general could also be limiting the range of Meliponinae Roubik 1989. A study by Baumgartner and Roubik 1989 found the highest diversity of stingless bees at 200m. Hilario et al. 2001 found that one species of Meliponinae Plebia pugnax demonstrated consistent flight act ivity between 22 34Â°C and 30 100% humidity but that activity was also proportional to the amount of sunlight. This could limit stingless bee distribution to areas with consistent temperatures, adequate sunlight, and humidity. Understanding the factors cont rolling the distribution of stingless bees are essential to understanding the possible effects of habitat loss fragmentation on Meliponinae populations Breed et al. 1999. This study attempts to further the knowledge of Meliponinae distribution. The effe cts of deforestation on Meliponinae are uncertain and studies on the subject are conflicting. Stingless bees could be particularly vulnerable to deforestation because old queens cannot fly and new nests need to be prepared before moving Brown and Albrecht 2001. Therefore, if a stingless bee nest needs to move away from a disturbance they can only move a few hundred meters. Breed et al. 1999 found that pastures impeded stingless bee movement and failed to find any there. Hubbell and Johnson 1977 found that large sized Meliponinae species like Trigona fulviventris could be limited to living in primary forests due to their need for large nesting trees and large foraging areas. Brown and Albrecht 2001 studied the distribution of Melipona across a
gradien t of deforestation and discovered that two species of Melipona were not affected at all by deforestation and that two species were so greatly affected by deforestation that they were only found in the most intact forest. Overall Melipona species richness i ncreased with intact forest Brown and Albrecht 2001. Liow et al. 2001 studied apids across a disturbance gradient and found the highest abundance of each species in larger primary forests, but the highest species richness in disturbed forests. Baumgart ner and Roubik 1989 found higher species richness of stingless bees near humans than in pristine rainforest. Accordingly, Eltz et al. 2002 found the density of stingless bee nests to be greatest in fragments bordered by pasture or mangroves and the lea st dense in continuous forest. Those Meliponinae in the fragments foraged mostly on non forest plants Eltz et al. 2002. Some species of stingless bees strongly benefit from edge due to increased diversity of foraging habitat Eltz et al. 2002. The effec ts of deforestation on Meliponinae diversity and abundance are variable and different for each species. This study tests which of these conclusions fit best to Costa RicaÂ€s Pacific slope and tries to better understand of the effects of deforestation on Mel iponinae and their ecosystem. Materials and Methods Apidae and Meliponinae were trapped at a Lower Montane Rain Forest in Monteverde, Costa Rica 1750 to 1800m and at a Premontane Moist Forest in San Luis, Costa Rica 750 to 800m in April and May of 20 03. At each location a total of 12 traps were set up, six in a pasture and six in a forest. In Monteverde traps in the pasture were set up approximately within a one hectare area and in the forest in about a hectare and a half area. In San Luis the pasture traps were in about a Â¾ of a hectare area and within a hectare area in the forest. A pan trap method was used for collection. Twenty four bowls were painted half yellow and half blue. Each trap in a pasture was secured to the top of three poles hammered into the ground, and each trap in a forest was hung from a branch with metal wire. Meliponinae and Apidae collection and the resetting of traps took place every two to three days at each location. There were 170 trap days for each location. After collecti on each apid was cleaned, pinned, and identified to genus or species. Data from InBIO Instituto Nacional de Biodiversidad on Meliponinae and Apidae distribution from relevant Pacific slope conservation areas were also analyzed to support and strengthen my conclusions. Species richness S and Smarg, Evenness E, and Shannon Weiner HÂ€ were computed. A Modified t test was used to compare HÂ€ values Magurran 1988. To find similarity between sites SorensenÂ€s quantitative index of similarity was calculat ed. A Simple Linkage Cluster Analysis was done to represent graphically the results of the SorensenÂ€s similarity indices Krebs 1989.
Results Regarding the Apidae family, there was higher species richness, evenness, and thus overall diversity HÂ€ at the low elevation than at the high elevation Table 1. The San Luis forest and the Monteverde pasture were significantly different in Apidae diversity, as well as the San Luis pasture and the Monteverde pasture Modified t test; Table 2. The San Luis for est and San Luis pasture were not significantly different in Apidae diversity Modified t test; Table 2. San Luis as a whole and Monteverde as a whole were also significantly different in Apidae diversity Modified t test; Table 2. The Monteverde forest was significantly different from all three of the other sites San Luis forest, San Luis pasture, and Monteverde pasture. Using just Meliponinae data, the San Luis forest and San Luis pasture were not significantly different Modified t test; Table 2. At 1750 1800m Monteverde no Meliponinae were found and at 750 800m San Luis six species of Meliponinae were found. The San Luis pasture and forest were the most similar sites for Apidae composition followed by the two pastures and then the San Luis fore st and Monteverde pasture Sorenson quantitative index of similarity and Simple Linkage Cluster Analysis; Fig. 1. The InBIO data on Meliponinae and on all Apidae have a negative relationship between species richness and elevation Fig. 2 and Fig. 3. At the same elevations as the study sites in this project InBIO data indicate nine species of Meliponinae living in the range where six species were trapped 700 800m and seven species where zero species were trapped 1700 1800m Table 3. At the 700 800 m elevation only two stingless bee species collected in this study and InBIOÂ€s studies were the same Trigona fulviventris and Partamona sp .. Four of the Meliponinae species collected Lestrimillata limao , Trigona ferricauda , Plebeia jatiformis , and Paratri gona sp . were not included on InBIOÂ€s list for this elevation Table 3. Discussion The results for all Apidae, as well as Meliponinae, showed that species richness, evenness and diversity were higher at the low elevation San Luis than the high elevat ion Monteverde. Also the two locations were significantly different from each other in Apidae diversity. This difference in diversity is consistent with InBIO data that show a negative relationship between Apidae species richness and elevation Fig. 2 a nd well as Meliponinae species richness and elevation Fig. 1. This decrease in diversity with elevation could be due to climatic factors at the higher elevation such as greater cloud cover and lower temperatures and/or physical factors at the higher elev ation such as more parasites, fewer nesting s i t es, less nesting material, and fewer food resources that could restrict some species of Apidae from living there Lobo 2000; Roubik 1989. No Meliponinae were found in the Monteverde pasture or forest 1750 1 800m. A small sample size or sampling time could be the reason for no stingless bees being trapped there. A lack of stingless bees at the high elevation site could be due to the
factors above that limit the distribution of all apids. The missing Meliponin ae could also be due to a low abundance of Meliponinae or highly dispersed stingless bee nests at high elevations. If there are not stingless bees in Monteverde or they are not in great abundance it could greatly affect the ecosystem there. Meliponinae are ecologically important and probably the most important polli nator in the tropics Liow et al . 2001. Plants at higher elevations in which stingless bees are absent might have needed to adapt to other pollinators. Further study on what high elevation plant s are pollinated by and how this affects high elevation ecosystems would be of great interest. The results for Apidae at both elevations and Meliponinae in San Luis show that pastures have higher richness than their forest counterpart. High richness in pa stures could be due to small sample size. It could also be due to the traps being more visible in the pastures or the scents dissipating further in the pastures than in the forests. High richness in pasture could have some significance to conservation. Mor e species of apids are foraging in pasture due to deforestation. This could mean drastic consequences for forest plants if species of Apidae and Meliponinae are being drawn out of the forest and are pollinating mostly pasture species. Apids and in particul ar stingless bees are very important pollinators in tropical forests and a change in their foraging behavior could mean a decrease in pollination and fitness of forest plants. The San Luis pasture and the San Luis forest 750 800m were not significantly different in Apidae or Meliponinae diversity. This similarity could be due to a small sample size. The San Luis pasture and the San Luis forest were different in some ways though. The stingless bees trapped in the forest had less species richness than thos e in the pasture, but forest Meliponinae species were collected in greater abundances. Most stingless bees nest in trees Griswold et al. 1995. This could explain why stingless bees were more abundant the forest; they had more trees suitable for nesting. Only one of the stingless bees caught in the pasture does not nest in trees Partamona sp . Griswold et al. 1995 but this species was also caught in the forest. There are few trees and fewer large trees in the San Luis pasture so there were probably not many available nesting sites for most Meliponinae in the pasture. This could suggest that the stingless bees trapped in the pasture are nesting in the forest and just foraging in the pasture. This is consistent with the Eltz et al. study 2002 that found the highest density of stingless bee nests in fragments bordered by pasture and these Meliponinae mostly foraged in the pasture. The increased diversity of foraging habitat for these edge nesting stingless bees could be the reason for the increased diversi ty of Meliponinae found in the San Luis pasture. Finding more species richness in the pasture and more abundance in the forest is also consistent with the Liow et al. study 2001 that found for all Apidae the highest abundance of each species was in large r primary forests yet the highest richness was in disturbed forests. Even though the two sites in San Luis were not significantly different in Meliponinae diversity, they do have differences in species richness and abundance. These differences, like more s pecies richness being found in the pasture, could greatly affect the pollination of forest plants and the ecosystem as a whole. The composition of Meliponinae differs between the San Luis Pasture and the San Luis forest. There were two species in common b etween the two habitats. One species
was unique to the forest Paratrigona sp . but it is found in all habitat types so it being only in the forest is probably due to a small sample size. Three species were unique to the pasture Trigona ferricauda , Trigon a fulviventris , and Lestrimillata limao . Yet, all three species nest in trees Roubik, 1992. One species, Lestrimillata limao , is parasitic on other stingless bees so its presents is probably due to its hostÂ€s presents in the pasture. InBIOÂ€s data on Me liponinae species between 700 and 800m on Costa RicaÂ€s Pacific slope are different from the species found in this study at the same elevation. Only two stingless bee species are the same from InBIOÂ€s data and this studyÂ€s data. The four species missing fro m InBIOÂ€s list for this elevation are missing for most other elevations as well. They each have wide gaps between elevations where they were collected. This study has served to fill in some holes in the elevational distribution of these four Meliponinae sp ecies. The information on the elevational distribution of Apidae and Meliponinae is still incomplete. This studyÂ€s data as well as InBIOÂ€s data show that diversity of both Apidae and Meliponinae decreases as elevation increases. More research needs to be done on species distribution and the reasons for these distributions. No Meliponinae were found at the high elevation. More research is needed to understand why stingless bees are not common at high elevations and what adaptations high elevation plants hav e in order to cope without this pollinator. This study did not find a significant difference in the composition of Apidae or Meliponinae between forest and pasture yet deforestation did affect stingless bee abundance and richness. Other research on how dis turbance, fragmentation, and deforestation affect Apidae and Meliponinae is contradictory. More studies need to be done to see which species are the most sensitive to habitat change and the implications of changing stingless bee ecology on the entire ecosy stem. Much is still unknown about Meliponinae and their distribution. These pollinators are important to the health and function of tropical forests and stingless bee conservation depends on knowledge of their distribution and their reaction to deforestati on. Acknowledgements I would like to thank Karen Masters for all her guidance, help, and patience. Thanks also to the Monteverde Cloud Forest Reserve for allowing me to do my research on their property. A big thank you to Andrew for being my water boy, n ot killing me on the motorcycle, and saving me countless hours of walking. Thanks also to Rick for putting up with all my questions and reviewing this paper. Thank you to Paul Hansen for identifying one of my stingless bee species. To all the CIEE staff an d students, thank you so much for an awesome semester. Most of all, thank you God, for bringing me to Costa Rica, putting all of the people from this program into my life, and getting me through all the tough times. Literature Cited Baumgartner, D.L. and D.W. Roubik. 1989. Ecology of necrophilous and filth gathering stingless bees of Peru. Journal of Kansas Entomological Society 61: 11 22.
Breed, M.D., T.P. McGlynn, M.D. Sanctuary, E.M. Stocker and R. Cruz. 1999. Distribution and abun dance of colonies of selected m eliponinae species in a Costa Rican tropical wet forest. Journal of Tropical Ecology 15: 765 777. Brown, J.C. and C. Albrecht. 2001. The effects of deforestation on stingless bees of the genus Melipona . Journal of Biogeograph y 28: 623 634. Eltz, T., C.E. Bruchli, S. Van der Karrs and E. Linsenmair. 2002. Determinants of stingless bee nest density in lowland dipterocarp forests of Sabah, Malaysia. Oecologia 131: 27 34. Griswold, T., F.T. Parker and P.E. Hanson. 1995. The bees Apidae. In: The Hymenoptera of Costa Rica . P.E. Hanson and I.D. Gauld, editors. Oxford University Press, Oxford, pp. 650 691. Hilario, S.D., V.L. Imperatriz Fonseca and AdeM.P. Kleinert. 2001. Response to climatic factors by foragers of Plebeia pugnax . Br azilian Journal of Diversity 61: 191 196. Hogue, C.L. 1993. Latin American Insects and Entomology . University of California Press, Chicago. Hubbell, S.P. and L.K. Johnson. 1977. Competition and nest spacing in a tropical stingless bee community. Ecology 58: 949 963. Janzen, D.H., editor. 1983. Costa Rican Natural History . The University of Chicago Press, Chicago, p . 636. Krebs, C.J. 1989. Ecological Methodology . Harper & Row, Publishers, New York, pp. 312 316. Liow, L.H., N.S. Sodhi and T. Elmquist. 2001 . Bee diversity along a disturbance gradient in tropical lowland forests of south east Asia. Journal of Applied Ecology 38: 180 192. Lobo, J. 2000. Stingless bees of cloud forests. In: Monteverde: Ecology and Conservation of a Tropical Cloud Forest . N.M. N adkarni, and N.T. Wheelwright, editors. Oxford University Press, New York, pp. 137 138. Magurran, A.E. Ecological Diversity and Its Measurement . Princeton University Press, Princeton, pp. 34 36. Roubik, D.W. 1989. Ecology and Natural History of Tropical Be es . Cambridge University Press, Cambridge. _______. 1992. Stingless Bees: A guide to Panamanian and Mesoamerican species and their nests. In: Insects of Panama and Mesoamerica . D. Quintero and A. Aiello, editors. Oxford University Press, Oxford, pp. 493 52 4. Wilson, E.O. 1971. The Insect Societies . The Belknap Press of Harvard University Press, Cambridge, pp. 89 93. Tables _______________________________________________________ ______________________________ TABLE 1. The species richness S, Margalef index Smarg, Shannon W iener index HÂ€, and evenness E for all the sites studied in Monteverde 1750 m and San Luis 750 m. Location S Smarg HÂ€ E Monteverde Pasture 2 0.434 0.611 0.881 Monteverde Forest 1 0 0 0 San Luis Pasture 8 2.73 1.84 0.884 San Luis Forest 5 1.56 1.52 0.947 All Monteverde 3 0.834 0.860 0.783 All San Luis 9 2.46 1.86 0.845
____________________________________________________________ _________________________ TABLE 2 . T tests showing differences between sites in Apidae diversity or Meliponinae diversity denoted with a Â§ significant difference indicated with a *. Sites Degrees of Freedom Critical Value T= P = *San Luis forest and Monteverde Pasture 1.63 4.30 7.03 <0.05 *San Luis pasture and Monteverde Pasture 7.71 2.36 3.21 <0.05 San Luis pasture and San Luis Forest 21.3 2.08 0.609 >0.05 *All of San Luis and all of Monteverde 24.3 2.06 3.32 <0.05 Â§ San Luis Pasture and San Luis Forest 0.014 12.7 3.23 >0.05 ____________________________________________________________ _________________________ TABLE 3. Species of Meliponinae found in this study and in InBIOÂ€s records between 700 800 m and 1700 1800 m on Pacific Slope of Costa Rica. Species San Luis Pasture San Luis Forest I n B IO 70 0 Monteverde InBIO 1700 750 800m 750 800m 800m 1750 1800m 1800m Lestrimillata limao x Melipona fascita x x Melipona fuliginosa x x Partamona sp. x x Partamona cupira x Partamona frontalis x Partamona grandipennis x Paratrigona sp. x Paratrigona ornaticeps x Plebeia frontalis x Plebeia jatiformis x x Scaptotrigona mexicana x x Trigona amalthea x Trigona corvina x Trigona ferricauda x Trigona fulviventris x x Trigona necrophago x
________________________________________________________________________ Figure 1. Clusters showing the similarity of three sites San Luis Pasture, San Luis Forest and Monteverde Pasture in their Apidae composition using a Simple Linkage Cluster Analysis. _____________________________________________________________________________________
________________________________________________________________________ Figure 2. Distribution of Stingless Bees on Pacific Slope of Costa Rica from InBIO and this studyÂ€s data. _____________________________________________________________________________________ _______________________________________________________________________ _ Figure 3. Distribution of all Apidae on Pacific Slope of Costa Rica from InBIO and this studyÂ€s data. _____________________________________________________________________________________