xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xlink http:www.w3.org1999xlink xmlns:xsi http:www.w3.org2001XMLSchema-instance
leader 00000nas 2200000Ka 4500
controlfield tag 008 000000c19749999pautr p s 0 0eng d
datafield ind1 8 ind2 024
subfield code a M39-00352
Una prueba de la teora Hubbell usando comunidades protistas en los tanques de las bromelias como un sistema modelo
A test of the Hubbell Theory using protist communities in bromeliad tanks as a model system
In order to test the Hubbell Theory, the species composition and abundances of protist communities living in bromeliad tanks was assessed, as well as dispersal distance and tank diameter measured. The model system included local communities (the bromeliad tanks), 3 meta-communities (the tree) and a meta-metacommunity (the three trees with 48 bromeliads). It was hypothesized that the protist communities in the bromeliad on a given tree behave as in a Hubbell meta-community and experience community drift and
zero-sum game conditions. Five predictions were used to test whether or not the Hubbell Theory fit this model system: (1) As the dispersal distance increases, the diversity indices S, H (Shannon-Weiner), and E (evenness) will decrease. (2) As N (the number of individuals in a community) increases, S (species richness) will increase. (3) As the bromeliad tank diameter increases, N will increase, and therefore so will the diversity indices S, H and E. (4) The diversity indices will steadily increase as the meta-community size increases. And (5) the Dominance-Density Curves for the three meta-communities will follow Hubbells predicted curves. The only prediction that the Hubbell Theory successfully described was the positive correlation between N and S, which was found for 2 of the 3 meta-communities and the metameta-community. However, this study does not conclude that the Hubbell Theory cannot fit this model
system. Instead, the investigation concludes that if future studies modify the methods in the recommended ways, it can be conclusively discovered whether or not the Hubbell Theory fits this model system.
La composicin de especies y la abundancia de las comunidades de protistas que vivan en tanques de bromelias fueron determinadas para probar la teora de Hubbell, as como tambin se midi la distancia de la dispersin y el dimetro del tanque. El sistema modelo incluy a las comunidades locales (los tanques de bromelias), 3 metacomunidades (el rbol) y una metametacomunidad (los tres rboles con 48 bromelias). Se predijo que las comunidades de protistas en las bromelias en un rbol dado se comportaran como en una metacomunidad de Hubbell y se ven afectadas por la deriva de la comunidad y condiciones de juego de suma cero. Se propusieron 5 predicciones para probar si la teora de Hubbell se adaptaba a este sistema modelo: (1) Conforme la distancia de la dispersin aumenta, los ndices S, H ' (Shannon-Weiner), y E (uniformidad) de la diversidad disminuye. (2) Conforme N (el nmero de individuos en una comunidad) aumenta, S (la riqueza de especies) aumenta. (3) Conforme el dimetro del tanque de la bromelia aumenta, N aumenta, y por lo tanto los ndices S, H ' y E de la diversidad tambin aumenta. (4) Los ndices de diversidad aumentan constantemente conforme el tamao de la metacomunidad aumenta. Y (5) las curvas de Densidad y Dominancia para las tres metacomunidades seguirn las curvas predichas por Hubbell. La nica prediccin que la teora de Hubbell describi con xito fue la relacin positiva entre N y S que se encontr para 2 de las 3 metacomunidades y para la metametacomunidad. Sin embargo, este estudio no concluy que la teora de Hubbell no puede explicar este sistema modelo. Esta investigacin concluy que si los estudios futuros modifican los mtodos de las maneras recomendadas, se puede aclarar concluyentemente si la teora de Hubbell explica este sistema modelo.
Text in English.
Costa Rica--Puntarenas--Monteverde Zone--San Luis
Diversidad de especies
Costa Rica--Puntarenas--Zona de Monteverde--San Luis
Tropical Ecology Spring 2005
Ecologa Tropical Primavera 2005
t Monteverde Institute : Tropical Ecology
1 A test of the Hubbell Theory using protist communities in bromeliad tanks as a model system Zoe Assaf Department of Integrative Biology, University of California Berkeley ABSTRACT In order to test the Hubbell Theory, the species composition and abun dances of protist communities living in bromeliad tanks was assessed, as well as dispersal distance and tank diameter measured. The model system included local communities (the bromeliad tanks), 3 meta communities (the tree) and a meta meta community (the three trees with 48 bromeliads). It was hypothesized that the protist communities in the bromeliad on a given tree behave as in a Hubbell meta community and experience community drift and zero sum game conditions. Five predictions were used to test whether or not the Hubbell Theory fit this model system: Weiner), and E (evenness) will decrease. (2) As N (the number of individuals in a community) increases, S (species richness) wil l increase. (3) As the bromeliad tank diameter increases, N will increase, and therefore so will community size increases. And (5) the Dominance Density Curves for the three meta communities will follow positive correlation between N and S, which was found for 2 of the 3 meta communities and the meta meta commun ity. However, this study does not conclude that the Hubbell Theory cannot fit this model system. Instead, the investigation concludes that if future studies modify the methods in the recommended ways, it can be conclusively discovered whether or not the Hu bbell Theory fits this model system. RESUMEN La composicin de especies y la abundancia de las comunidades de protistas que vivan en tanques de bromelias fueron determinados para probar la teora de Hubbell, as como tambin se midi la distancia de la dispersin y el dimetro del tanque. El sistema modelo incluy a las comunidades locales (los tanques de bromelias), 3 metacomunidades (el rbol) y una metametacomunidad (los tres rboles con 48 bromelias). Se predijo que las comunidades de protistas en las bromelias en un rbol dado se comportaran como en una metacomunidad de Hubbell y se ven afectadas por la deriva de la comunidad y condiciones de juego de suma cero. Se propusieron 5 predicciones para probar si la teora de Hubbell se adaptaba a este s istema modelo: (1) Conforme la distancia de la dispersin aumenta, los ndices S, H (Shannon Weiner), y E (uniformidad) de la diversidad disminuye. (2) Conforme N (el nmero de individuos en una comunidad) aumenta, S (la riqueza de especies) aumenta. (3 ) Conforme el dimetro del tanque de la bromelia aumenta, N aumenta, y por lo tanto los ndices S, H y E de la diversidad tambin aumenta. (4) Los ndices de diversidad aumentan constantemente comforme el tamao de la metacomunidad aumenta. Y (5) las cur vas de Densidad y Dominancia para las tres metacomunidades seguirn las curvas predichas por Hubbell. La nica prediccin que la teora de Hubbell describi con xito fue la relacin positiva entre N y S que se encontr para 2 de las 3 metacomunidades y pa ra la metametacomunidad. Sin embargo, este estudio no concluy que la teora de Hubbell no puede explicar este sistema modelo. Esta investigacin concluy que si los estudios futuros modifican los mtodos de las maneras recomendadas, se puede aclarar concl uyentemente si la teora de Hubbell explica este sistema modelo.
2 INTRODUCTION Biodiversity plays a major role in the stability of ecosystems. Ecological communities that are more diverse have been shown to display higher resistance and resilience in the face of perturbations, as well as higher constancy (Worm and Duffy 2003). This stability is most likely conferred by the increased redundancy of diverse systems, which contain more species that are functionally similar, as well as by the increased likelih ood that keystone species, crucial to community structure, will occur (Worm and Duffy 2003). Thus, knowing how biodiversity is maintained is essential to understanding and preserving ecosystems. There are two major models of biodiversity that predict whic h factors determine what species are present or absent and their relative abundances. The older and more fixed assemblage of species, determined by the partitioning of li mited resources and the functional role of the species within it, each of which has evolved to be the best model that says community drift, stochastic extinction ev ents, history, random speciation and dispersal determine the continuously changing community composition (Hubbell 2001). Or in other words, chance. This theory is a neutral model, which assumes that all individuals in a trophic level are competitively and ecologically equivalent (Hubbell 2001). The first widely accepted neutral, dispersal assembly model was the Island Biogeography Theory, which asserts that the two major forces determining species richness on islands are immigration and extinction (MacArthu r and Wilson 1967). Stephen Hubbell (2001) recently generalized this theory and provided more readily testable predictions. He proposed that there is a fixed J, or a total number of individuals in the local and meta community that never changes. Every time one individual dies, it will be replaced by one other individual. The new individual can be an offspring from the local community, an immigrant from a different local community in the meta population, or a newly evolved species. Since all individuals are competitively equivalent, the probability that the new individual will be a particular species is directly proportional to the relative abundance of that species in the meta community. Hubbell dubbed this the Zero Sum Game, because the community size never changes. When immigration and speciation are reduced or absent, this game will result in community drift. Drift leads to common species becoming more common, and rare species becoming rarer, eventually leading to monodominance. Thus, immigration between l ocal communities becomes essential for maintaining biodiversity. richness is determined by the rate of drift, richness will be higher when drift is slower. Thus, (1) local commun ities that have high connectivity (i.e. high immigration rates) will have higher species richness. Also, (2) local and meta communities that have a high fixed J will have higher species richness, both because it takes longer for monodominance to triumph, a nd because population sizes can be larger (thus more viable populations will exist). A higher fixed J can be a result of a larger physical area or of a higher productivity, as is found in the tropics (Gaston 2000, Worm and Duffy 2003). Therefore (3) richne ss should be positively correlated with the physical size of the local community.
3 Furthermore, higher species richness will cause there to be greater diversity, and thus (4) diversity indices should steadily increase as the community gets bigger. Also, th ere are predicted, observable trends than can be easily tested. Larger meta communities will have higher species richness. But, as species richness increases, each individual species will decrease in relative abundance, and thus species evenness will incre ase. This will Density Curves (See Fig. 1): the largest meta community is dominated by several species and the smallest meta community is dominated by a only a few species. These predict ions along with other tests of the Hubbell Theory are only beginning to be investigated. assembly is relatively new (2001), neutral theories in general are by no means new or widely unaccepted. The neutralist theory of evolutionary genetics has been around for well over a decade, and it is remarkable how easily a theory based in molecular genetics can be slightly modified to fit an ecological scale. This neutralist theory asserts that the majority of evolution ary changes at the DNA level are not caused by Darwinian selection, but rather by random fixation of selectively neutral or nearly neutral mutations (Kimura 1986). This genetic drift is functionally identical to community drift, in that chance fixes compet itively equivalent units into the population, and all units are selectively neutral (Kimura 1991). This theory has been supported by a range of studies, one of the most convincing being a study of viral genes (an excellent system because of the rapid rate of evolution) that found the patterns of gene substitution were readily explained by the neutralist theory (Gojobori et al. 1990). Although it is a new addition to ecology, the neutral theory has been tested several times on an ecological scale, and has r eturned mixed results. Hubbell himself showed the remarkable predictive power of his models when he fit them amazingly well to empirical data on species abundances within communities, often using his own massive dataset from years of work in the tropical f orests of Panama (Brown 2001, Hubbell 2001). There also exists an interesting study that combines the molecular and ecological scale by studying the genetic variance of island populations (Lande 1992). The study found that chance processes and connectivity as measured by mutation and migration, were very important in maintaining genetic variance in small, local populations. However, there are also studies that have rejected or devalued the Hubbell Theory. Zhang and Lin (1997) tested the robustness of the c ommunity drift model by slightly varying the assumption of competitive equivalency, or in other words giving one species a little competitive advantage over another. The study found that slightly unequal competitive abilities drastically impacted the lifes pan of the species, and concluded that the drift hypothesis is rather inapplicable to the real world where slight variances most likely exist. Another study conducted in floodplain forests found that the model fails to account for the high levels of compos itional similarity in disjunctive sample areas (Terborgh et al. 1996). Tuomisto et al. (2003) recently performed a study evaluating the relative importances of environmental and geographical distances on floristic differences in western Amazonian forests. The study showed that although the geographic distance (and thus dispersal) accounted for much of the variation, environmental distances were a better explanatory variable. Lastly, a study performed by a CIEE student in 2002 (Kurton 2002) used insect biodi versity to test the Hubbell predictions, but her results did not
4 conclusively support or disprove the model. Overall, these conflicting studies demonstrate that the Hubbell Theory needs to be more extensively investigated. A possible system for testing th e Hubbell Theory is protist communities living in bromeliad tanks. Bromeliads are epiphytic plants (Bromeliaceae) that contain tanks that harbor communities of organisms. These bromeliads are living on trees, and thus provide easily definable local and met a communities. In fact, during an investigation of the impact of exotics on diversity, it was found that a community of exotic bromeliads in close proximity to one another had much greater diversity and species richness than a community of native bromeliad s more sparsely dispersed (Rhoney 2004). Although the spatial composition (and thus connectivity) was not conclusively responsible for the differences in diversity, it was suggested as a possible factor. Furthermore, it has been suggested that responses to environmental or biological factors rather than predator prey relationships, such as an increase in a major food resource (i.e. productivity), are responsible for increases in protozoan abundances (Carrias et al. 2001). It has also been shown that species richness, abundance, and the amount of organic matter within the tank are all directly related to plant size (Richardson 1999). These suggested relationships of connectivity and community size (J) with protist diversity demand further examination. Assessi ng the protist community composition in relation to other variables would provide an ideal system for testing the Hubbell predictions. I predict that protist communities in bromeliad tanks on a given tree behave as in a Hubbell meta community and thus exp erience community drift and zero sum game conditions. Several predictions that follow from the Hubbell Theory will be tested by recording the dispersal distance and the physical size of the community, both crucial factors in determining species richness an d abundance. Those predictions are: (1) As the Weiner), and E (evenness) will decrease; (2) As N (the number of individuals in a community) increases, S (species richness) will increase; (3 ) As the bromeliad tank diameter increases, N will will steadily increase as the meta community size increases; and (5) the Dominance Density Curves for the three meta (See Fig. 1) MATERIALS AND METHODS This investigation was conducted at the San Luis Ecolodge and Research Station located in the greater Monteverde area of Costa Rica during the dry season from Apri l 16 until May 7 of 2005. The Hubbell Theory was tested on a model system of tree meta communities hosting bromeliads, each of which was considered a local community. Bromeliads are epiphytes in the family Bromeliaceae, most recognized for their unique ros ette structure, which forms a tank in the center of the plant. This tank contains a wide array of organisms including the microscopic, unicellular, eukaryotic protists that live in the tank water. The study evaluated three tree meta communities (n = 8, 12, and 28) that contained 48 bromeliads (the meta meta community). Bromeliads that had a base diameter > 2.5 cm, contained tank water, and that were accessible (using a ladder) were sampled. The meta communities with respect to the number of bromeliads maint ained their relative sizes; however, the actual number of individuals on each tree was greater
5 than the number sampled (GET THIS NUMBER). The bromeliad/tank diameter and the nearest neighbor/community distance for each bromeliad were recorded. The local co mmunities of protists were evaluated in this study for species richness and abundance Each bromeliad was assessed for its location in the community, and water contents sampled. All the water contained in the tank was collected with an eyedropper and p laced in a vial. The nearest neighbor distance (NND) and the nearest sub community distance (NCD) were measured as the distance between bromeliad bases. Only bromeliads with a base diameter > 2.5 cm were used, although the neighbor did not need to contain water. NND is the distance to the nearest bromeliad. Bromeliads growing directly next to each other or connected were given a NND of zero. Sub communities, defined as all bromeliads in contact with each other, were included because of the potentially confo unding factor of vegetative reproduction frequently yielding an NND value of zero. NCD was measured as the distance from the individual bromeliad to the nearest different community. Tank diameter and bromeliad diameter (the widest point, leaf tip to leaf tip) were also measured. One drop of water was used from each bromeliad for the protist analysis. The vial was shaken before withdrawing the drop, and aloe was added to the slide in order to slow down the protists. For each slide, five locations were exami ned using a Leitz Wetzlar dialux microscope at 100x and 40x for morphospecies and their abundance. Twenty six morphospecies were observed. Common characteristics used to identify protists included body shape, size, shading, and patterns of movement. RES ULTS The diversity indices S (species richness), S marg (Shannon Weiner diversity index) were assessed for each bromeliad local community and each tree meta community (See Fig. 2). Simple regression analyses were used to evaluate whether there was a significant positive or negative trend in the diversity indices and N (number of protists) as the meta community size increased. There was a slight positive trend for N, S, and S marg (p = 0.89, 0.23, 0.29 respectively; R 2 = 3.56E 4, 0.03, 0.03 2 = 0.02, 1.25E 5 respectively), but none of the trends were significant. A modified t test was used to evaluate whether there was a significant community. A significant difference was found between the Tree Meta Community of 8 = 1.80)(t = 3.86, p < 5), as well as between the Tr ee Meta Community of 8 Bromeliads and the Tree Meta although there was no significant trend when using a simple regression, the modified t test shows that T8 has significantly higher dive rsity than T12 or T28. Simple regression analyses were used to evaluate the relationship between the community and the meta meta community. A multiple regression analysi s was not used because a significant correlation was found between NND and NCD (R 2 = 0.11, p = 0.025). Of the 24 analyses, only 1 returned a significant correlation. This was a negative trend found between NND and S for T12 (R 2 = 0.048, p = 0.14)(See Fig. 3). As the dispersal distance increased, species richness decreased. But overall, the relationship is
6 not strong between dispersal distance and the diversity indices because of the few number of NND points. Simple regression analyses were used to evaluate whether there was a significant correlation between N and S for each meta community and for the meta meta community (See Fig. 4). Three of the four analyses found a significant positive correlation: T8, T12 and the meta meta community (R 2 = 0.68, 0.35 and 0.10 respectively; p = 0.01, 0,04, and 0.026 respectively). So, as the number of individuals in the population increased, species richness also increased. There was no significant correlation found between N and S for T28 (R 2 = .0062, p = 0.69). Simple re gression analyses were used for each meta community and the meta meta diameter. Of the 16 analyses, significant negative correlations were found on T12 diameter (R 2 = 0.37, p = 0.03) and between S and tank diameter (R 2 = 0.36, p = 0.04) (See Fig. 5). So, as tank diameter increased, the species richness and diversity decreased. A Dominance Diversity Graph (See Fig. 6) was produced to show the relative ab undances of each species within the three tree meta communities and whether a community is dominated by one or a few species. The curves fit to the data points were not linear relations because the line of best fit was an exponential relationship. The T12 curve had the most strongly negative slope, thus this meta community species composition was less even and more dominated by a single species than the other meta communities. This meta community also had the shortest line, and thus it was the least species rich. The T28 curve had a slope very similar to the T8 curve, and thus these two meta community species compositions were very similar with regards to evenness. Both were more even than the T12 meta community, and thus were more dominated by a few species than by one species. The T28 curve had the longest line, and thus was the most species rich. DISCUSSION This investigation sought to determine whether protist communities in bromeliad tanks on a given tree behave as in a Hubbell meta community and thus experience community drift and Zero Sum Game conditions. Several predictions were tested, and most of them either were not consistent with Hubbell predictions, or were inconclusive. But, rather than indicate that the Hubbell Theory can never be applied to protist communities in bromeliads, this investigation demonstrates that new and improved tests can and should be used on this model system. The first prediction was that as the dispersal distance increased, as measured by nearest neighbor distance (NND) and nearest sub community distance (NCD), the diversity found, the simple regression was based only on three different NND measurements (T12 had only two actual distances and the rest were zeros). This is a very small sample size, and thus no strong relationship was considered found between dispersal distance and the
7 diversity indices. This prediction was based on Hubbell is a strong determinant in the species composition and abundances of local and meta communities when drift is slower, the species richness and therefore the diversity indices will be higher. This assertion leads to the p rediction that when local communities have high immigration rates and thus are more connected, community drift will be slower. The lack of a correlation between connectedness and diversity in this investigation suggests that this system is not subject to c ommunity drift. This means that Zero Sum Game conditions, in which the relative abundances of competitively equivalent species determine the community richness, are not being experienced by this system. This is somewhat consistent with a recent study done by Tuomisto et al.(2003) that found that environmental differences had a stronger impact on community composition than geographical distances. But, she also found that geographical distances could predict some of the variation in community composition. Ho wever, the lack of a correlation found in this investigation could also be due to the impact of random dispersal not being accurately represented by the measurements taken, i.e. NCD and NND. The bromeliad local communities do not have linear spatial relati onships with other bromeliads on the tree meta community. For any given bromeliad, there will be local communities above, below and right next to it. Thus, the actual amount of immigration a bromeliad is experiencing may be determined by a more complicated relationship than simply by the NND or NCD. This possibility is strengthened by a past study that found that a tightly knit community of exotic bromeliads harbored a more diverse protist composition than a more loosely dispersed community of native brome liads (Rhoney 2004). All in all, this analysis of the impact of NND and NCD on diversity does not provide a system readily explained by the Hubbell Model. Species richness in relation to the number of individuals in the community A second prediction tha t was tested here was as the number of the individuals in the community (N) increases, the species richness (S) would also increase. This study did find a strong positive relationship between N and S for two of the three meta communities, and the meta meta assertion that a fixed J (number of individuals in the community) exists, but what species are present or absent is continuously changing (as opposed to the niche assembly model which predicts that the com munity assemblage is fixed). Thus, when the fixed J increases, the S will increase, both because the trend towards monodominance caused by community drift is slowed down, and because there is room for more populations of viable sizes (large enough to not b e threatened by population stochasticity). Thus, the strong correlation found in this study suggests that this system is subject to a fixed J, and thus Zero Sum Game conditions and the effects of community drift. However, this correlation does not conclusi vely reject the niche assembly model because according to that model, a community with more resources can support a larger population and more species. But, a parallel study using the same dataset as this one found no correlation between the diversity indi ces or the population sizes and the amount of resources (Spaulding 2005). Overall, this analysis of the relationship between N and S suggests that the Hubbell Model may be able to describe this system.
8 E The third prediction that was tested here was that as the bromeliad tank diameter prediction is based on the Hubbell assertion that a fixed J exists, and the larg er it is, the higher the species richness (an assertion that this study found to be applicable to this model system). The size of the fixed J can be a result of the physical area occupied by the community, or by a higher productivity (Gaston 2000, Worm and Duffy 2003). So, as the tank diameter increases, N will increase, and therefore so should the diversity indices. However, this investigation found that for T12, when the tank diameter increased, a trend in opposition to the prediction (Fig. 5). But, when one notes that N is not correlated with tank diameter, the observed trend does not necessarily reject the Hubbell Model. The fixed J can be a result of the physical area or the productivity of the c ommunity. So the lack of a correlation between N and the physical size of the community suggests that in this system the J is a result of productivity rather than of the physical size of the community. The trend ely is due to another variable, not the size of J. This is a perplexing finding when one notes a past study that found tank diameter of protists to be related to species richness and abundance (Richardson 1999). A potentially confounding variable in this study is that it was conducted during the dry season in the San Luis Valley, a relatively dry area already. Most of the time the bromeliads had very little water, and when it did rain the increase in water volume had evaporated within 2 3 days. This result ed in often variable water volumes, or physical community sizes, to which protist communities most likely did not have enough time to respond. The samples used to evaluate protist diversity were therefore dependent on whether or not it had rained recently and consequently diluted the sample. Future studies are strongly advised to conduct sampling during the rainy season so that the physical size of the protist communities (i.e. water volume) will be constant. Overall, the lack of a correlation between N and tank diameter signifies that this analysis was inconclusive as to whether or not the Hubbell Model can readily describe this system. The effect of meta community size on N, S, S marg A fourth prediction that was tested was that the diversity indices will steadily increase as the meta community size increases. The trends in N, S, S marg insignificant, but the diversity of T8 was found to be significantly higher than the diversity of T12 or T28, contrary to the Hubbell predictions The prediction was based on the Hubbell assertion that community drift is slower in meta communities that are larger and E as well. This study assumed that a more den sely packed tree of bromeliads will have both a larger fixed J, and higher connectivity. But, the relationship of distance between bromeliads and the fixed J has already been suggested to be more complicated than simply NND or NCD by this study. Furthermor e, the lack of a significant positive trend in N as the meta community size increased suggests that a meta community that consists of more local communities may not necessarily contain more individuals. Since
9 T28 was not much larger than T12 with regards t o the fixed J (or the N), perhaps this study did not have enough variation in meta community size to accurately determine a trend. The low variation in meta community size could mean that other stochastic factors that were not measured had a strong impact on meta community composition and abundance. For example, how long the protist community has been in existence, as measured through the age of the bromeliad or the tree, may impact protist diversity. If the community has not had enough time to reach equil ibrium size, then assessments of the species richness and abundance would not be useful in determining if the Hubbell Model could potentially describe the system. So, in order to control for other such stochastic variables, it is suggested that future stud ies more carefully define a meta community as the number of individuals present, and accordingly sample a range of meta community sizes. The lack of trends found, and the higher diversity in the smallest meta community, imply that either the Hubbell Model is not a good model for this system, or the sample size of meta communities was not large enough to show a trend. Dominance density curves of the three meta communities The fifth and final prediction was that the Dominance Density Curves for the three m eta community curves observed did not fit the Hubbell predictions. The prediction was based on the Hubbell assertion that as meta community size increases, the species ri chness will increase while each individual species will decrease in relative abundance, thus increasing evenness. This would give larger meta communities that are dominated by several species rather than by one or two. But, the prediction was also based on the assumption that the size of the meta community is a consequence of the number of bromeliads in it, which has been questioned by precious results of this study. Thus, either the Hubbell Theory does not provide a good model for this system, or a larger sample size of meta communities needs to be used for accurately determining trends. Conclusion This study initially sought to test whether the Hubbell Model fits protist communities in bromeliad tanks and determines the effects of community drift and Ze ro Sum Game conditions. Although the results of the study do not conclusively support or reject the Hubbell Model, they do readily invite further investigation of this system and offer ways of improving the methods of testing the Hubbell Model. These inclu de conducting the study during the rainy season, defining a meta community by the number of protists present (not the number of bromeliads), taking into account other stochastic factors such as colonization history, and refining the technique of measuring dispersal ability in the system. It is also important to bear in mind when examining biodiversity models that the Hubbell Theory does not necessarily render niche differentiation meaningless, and observations of niche partitioning are not necessarily rejec tions of the Hubbell Model. The dispersal assembly model affirms the role of chance and not codependence or adaptations in community composition. But, as Hubbell wrote in The Unified Neutral Theory of Biodiversity and Biogeography what that means is that rather than being a
10 changing biotic and abiotic selective environments to which the species ancestral lineages were exposed during their long, individualistic geog raphic wanderings, the ghost of ecology will be a fascinating and illuminating process of investigation. ACKNOWLEDGEMENTS Thank you Jessy for being an awesome, hilari ous partner. I would also like to thank Karen Masters for her immense help throughout this project, as well as the San Luis Ecolodge for the use of their facilities and bromeliads. Also, thank you to Matt and Ollie for providing all the materials we needed and for being there to help with the stats. LITERATURE CITED Brown, J. 2001. Toward a general theory of biodiversity. Evolution 55(10): 2137 2138. Carrias, J.F., M.E. Cussac, and B. Corbara. 2001. A preliminary study of freshwater protozoa in tank brome liads. Journal of Tropical Ecology 17: 611 617. Gaston, K. 2000. Global patterns in biodiversity. Nature 405: 220 227. Gojobori, T., E. Moriyama, and M. Kimura. et al. 1990. Molecular clock of viral evolution, and the Neutral Theory. Proceedings of the N ational Academy of Sciences of the United States of America 87(24): 10015 10018. Hubbell, S. 2001. The Unified Neutral Theory of Biodiversity and Biogeography, pp 1 30. Princeton University Press, Princeton, New Jersey. Kimura, M. 1986. DNA and the Neutra l Theory. Philosophical Transactions of the Royal Society of London, Series B, Biological Sciences 312(1154): 343 354. Kimura, M. 1991. Recent development of the Neutral Theory viewed from the Wrightian tradition of theoretical population genetics. Procee dings of the National Academy of Sciences of the United States of America 88(14): 5969 5973. Kurton, E. 2002. Insect Biodiversity in Ficus tuerckheimii (Moraceae): Which model fits best?. CIEE Spring Tropical Biology and Conservation, pp 42 52. Lande, R. 1992. Neutral theory of quantitative genetic variance in an island model with local extinction and colonization. Evolution 46(2): 381 389. MacArthur, R.H., and E.O. Wilson. 1967. The Theory of Island Biogeography. Princeton University Press, Princeton, N ew Jersey. Rhoney, M. 2004. Comparison of protist communities in native and exotic bromeliad species. CIEE Fall Tropical Biology and Conservation, pp 1 11. Richardson, B.A. 1999. The bromeliad microcosm and the assessment of faunal diversity in neotropic al forest. Biotropica 31(2): 321 336. Spaulding, J. 2005. Protist community diversity in relation to resources in bromeliads. CIEE Spring Tropical Biology and Conservation. Terborgh, J., R.B. Foster, and P. Nunez. 1996. Tropical tree communities: A test of the nonequilibrium hypothesis. Ecology 77(2): 561 567. Tuomisto, H., K. Ruokolainen, and M. Yli Halla. 2003. Dispersal, environment, and floristic variation of western Amazonian forests. Science 299: 241 244. Worm, B. and J.E. Duffy. 2003. Biodiversit y, productivity, and stability in real food webs. TREE 18(12): 628 632. Zhang, D.Y., and K. Lin. 1997. The effect of competitive asymmetry on the rate of competitive y 188(3): 361 367(7).
11 Density Curves showing the negative correlation between rank abundance and the log of proportional abundance. Large meta communities have a longer line with a less negative s lope, demonstrating the high species richness and the dominance in the meta community by several species. Small meta communities have a shorter line with a more strongly negative slope, demonstrating the lower species richness and the dominance in the met a community by one or two species. Large Medium Small RankAbundance Log proportiona l abundance
12 FIGURE 2. The trends between the 3 meta communities for (2a) the diversity indices S marg Weiner) (2b) N (the n umber of protists) and (2c) the diversity index S. Bars are for the entire meta community, and the error bars are plus and minus 1 SE. The N, S, and S marg trends as the meta community size incr eased. All trends were insignificant (p > .05). A modified t
13 FIGURE 3. Plot of nearest neighb or distance (NND) against species richness (S) for the tree meta community of 12 bromeliads. Each diamond represents one bromeliad local community. The regression line represents a significant negative correlation between the two variables (y = 0.1191x + 7.9823, R 2 = 0.4708, p = 0.14). FIGURE 4. A plot of the number of individuals (N) against the species richness (S) for the three meta communities (a) T8 (b) T12 (c) T28 and for the (d) meta meta community. Each diamond represents a s ingle bromeliad local community. The regression lines represent significant positive correlations between T8, T12, the meta meta community and S The positive correlation for T28 was not significant.( T8: y = 0.042x + 3.75, R 2 = 0.68, p = 0.011; T12: y = 0.011x + 5.14, R 2 = 0.35, p = 0.042; META META: y = 0.007x + 6.30, R 2 = 0.10, p = 0.026; T28: y = 0.0019x + 7.04, R 2 = 0.0062, p > 0.05 ).
14 FIGURE 5. The relationship between tank diameter (cm) and (a) species richness (S) and (b) diversity community. Each diamond represents a bromeliad local community. The regression lines represent a significant negative trend between the 0.2463x + 1.7138, R 2 = 0.3725, p = 0.03) (S : y = 1.6399x + 10.081, R 2 = 0.3577, p = 0.04) FIGURE 6: The Dominance Diversity curves for the three meta communities (T8, T12, T28). Species rank in abundance against the log of proportional abundance gives lines that represe nt each meta communities spread of species. T12 (y = 0.782e 0.5284x, R 2 = 0.9788)had the shortest line and the steepest slope, thus it is the least species rich and is most dominated by a couple of species with a high abundance. T8 (y = 0.3383e 0.3162x, R 2 = 0.9727) and T28 (y = 0.2861e 0.3273x, R 2 = 0.9713) had similar slopes, and so both were dominated by several species with lower abundances. T28 had the longest line, and so was the most species rich.