Interference competition is widely considered to structure ant communities. Competition’s effect, however, may be contingent upon disturbance or the abiotic environment. The interaction of temperature and competition is implicit in a wide body of ant community research; however, very few studies have experimentally manipulated these variables. To investigate the role of competition and temperature on ant communities, I (i) employed null models to investigate how species partition their spatial, temporal, and thermal environments in disturbed and undisturbed forests, (ii) used pairwise behavioral experiments to construct a Markov chain model to predict relative abundance patterns and correlated behavioral indices to species co-occurrence patterns, and (iii) conducted a shade, physiological thermal tolerance, and fully factorial shade and removal experiment to investigate the interaction of competition and temperature on ant community structure. The results of these studies are summarized below. First, I took advantage of a natural experiment, the 2002 Biscuit Fire, to investigate how species partition their temporal, thermal, and spatial environments in disturbed and undisturbed forests with null models. I found that most sites displayed a high degree of temporal niche overlap and species aggregation along the thermal axis. Half of the sites, however, had regular spacing of the temperature at which species obtain maximum activity. Species co-occurrence patterns in space modulated with diurnal temperature variations. Unburned sites had more spatial segregation of species than burned sites. Overall, it appears as though species activity is regulated, at least in part, by the thermal niche axis, and ant communities may repeatedly assemble and disassemble throughout the day. Second, I used data from pairwise behavioral experiment to generate transition probabilities for a Markov chain model. Assuming the landscape represents a large number of patches, the model predicted the relative abundance of an assemblage. I compared Markov chain predictions of relative abundance to relative abundance measurements on the local and regional scale. I used the same pairwise behavioral data to predict species co-occurrence values in three sites. Neither model accurately predicted community patterns. The only significant result was the Markov chain prediction of bait occurrence on the local scale; however, the relationship was opposite of the prediction. Finally, I conducted a shade experiment to investigate how communities respond to an altered thermal environment and associated their response to results from physiological thermal tolerance experiments. I then conducted a fully-factorial shade and Formica moki removal experiment to investigate if thermal responses were mediated by competitive effects. The addition of shade tables greatly reduced temperatures in the field, and Temnothorax nevadensis abundance was consistently lower in shade treatments. Decreased abundance at shade stations did not appear to be an indirect effect of F. moki activity. Physiological thermal tolerance was strongly associated with changes in abundance in shade treatments: the lower a species thermal tolerance, the greater its positive change in abundance after shade additions. The only species with a strong foraging response to F. moki removal was T. nevadensis, a species who was often cooccurred with F. moki on baits. I did not find evidence for the interaction of competition and temperature, and it appears as though physiological differences strongly influence the foraging activity of Siskiyou ant communities.

In this study, I used two independent perciform lineages (Neotropical Cichlidae and Nearctic Centrarchidae) to examine patterns of species richness and species coexistence a two spatial scales (e.g., macrohabitat and mesohabitat) and to examine inter-faunal patterns of ecomorphological convergence. The study was conducted during the low-water periods in four lowland rivers: the Cinaruco in Venezuela, the Tambopata in Peru, and the Neches and the Brazos rivers in Texas (USA). These rivers were chosen because of their similar characteristics, in terms of geomorphology, sediments, and water quality. The Cinaruco River and the Neches River have clear slightly-stained waters, whereas the Tambopata and the Brazos River have turbid waters with high loads of suspended sediments. I used morphological approaches as a surrogate to investigate patterns of species distribution in niche space, and predict patterns of species richness at different spatial scales. Despite high variation in the number of species in these two perciform assemblages, morphological analysis based on the means and standard deviations of nearest neighbor distance (NND) and mean distance to centroid (CD) revealed similar trends of morphological similarity in relation to species richness. Comparison of observed versus randomized data mesohabitat scale for all four rivers generally supported the niche expansion model of response to increase in species richness. At the scale of mesohabitats within rivers, most species assemblages appear to be organized by competitive interactions in accordance with the niche expansion model. The tropical species-rich Cinaruco River revealed particularly strong support for the niche expansion model. Intercontinental comparison of functional morphology and diets based on analysis of stomach contents and stable isotope ratios indicated broad morphological and dietary overlap between cichlid and centrarchid assemblages. For the most part, morphological ordinations showed that the two groups have diversified in a parallel manner within the confines of ram-suction modes of prey ingestion. This study concludes that even though differences are observed in historical and stochastic factors structuring fish assemblages in different geographic regions, consistent patterns of convergence at the species and assemblage levels results from natural selection under similar environmental conditions.

A competição interespecífica pode representar uma força importante na determinação da distribuição e da abundância de organismos na natureza. Comunidades naturais compostas por espécies que utilizam o mesmo recurso de forma semelhante supostamente são estruturadas pela competição interespecífica. Tal força interativa no decorrer da evolução pode ter levado à diferenciação de nicho entre os competidores no passado, sendo esse fator relevante para explicar a coexistência de espécies ecologicamente similares. As comunidades reprodutivas de anuros tropicais são caracterizadas por alta diversidade e grande sobreposição espacial de espécies, por isso representam um ótimo modelo de estudo para avaliar o papel da competição em escala local. Nessas comunidades reprodutivas a alta densidade de machos de diferentes espécies em atividade de vocalização poderia causar problemas de comunicação intraespecífica por interferência acústica. Por essa razão, considerando a hipótese de que o ambiente acústico pudesse ser partilhado, os pesquisadores vêm avaliando, na grande maioria das vezes de forma empírica, estratégias que as espécies de anuros podem empregar para reduzir a competição por canais acústicos no interior da comunidade, como a separação espectral (uso de diferentes faixas de frequência) e a diferenciação de uso espacial e/ou temporal. Nesse contexto, no presente trabalho estudei agregações de anuros formadas em 16 ambientes reprodutivos de seis localidades de Mata Atlântica do estado de São Paulo, com o principal objetivo de testar, por meio de modelos nulos, se ocorre partilha de nicho acústico e espacial entre as espécies. Foram gravados os cantos de anúncio dos machos de cada espécie presente nos ambientes reprodutivos, e seus sítios de vocalização foram caracterizados quanto à natureza e altura do substrato. Os testes por modelos nulos não evidenciaram qualquer padrão, tanto na ocupação do ambiente acústico como na distribuição no gradiente vertical. A propriedade acústica temporal \"duração de canto\" não reduziu a sobreposição espectral das espécies no ambiente reprodutivo, e é possível que as fêmeas, da grande maioria das espécies, localizem seus parceiros em agregações heteroespecíficas com base na frequência dominante e não sofram prejuízo com a interferência acústica. Contudo, essa é uma hipótese que ainda deve ser testada dentro das agregações reprodutivas de anuros da Mata Atlântica. Os resultados, de maneira geral, corroboraram a ideia de que fatores abióticos, heterogeneidade ambiental e diversidade filogenética podem ser mais importantes para explicar a ocorrência de espécies nas ricas agregações de anuros na Mata Atlântica. / Interspecific competition may represent a relevant force determining the distribution and abundance of organisms in nature. Natural communities composed by species that use the same resource in a similar fashion are, supposedly, structured by interspecific competition. This interactive force in the course of evolution may have led to niche differentiation among competitors in the past, and this is a relevant factor to explain the coexistence of ecologically similar species in the same habitat. Reproductive aggregations of tropical frogs are characterized by high diversity and large spatial species overlap, therefore they represent an excellent model to evaluate the importance of competition on local scale. In these communities the high density of breeding males of different species in calling activity could cause a masking effect on intraspecific acoustic communication. Therefore, considering the hypothesis that the acoustic environment could be shared, researchers have been evaluating, in most cases empirically, strategies employed by frogs to reduce competition for acoustic channels within the community, as the spectral separation (using different frequency bands) and spatial and/or temporal segregation. In this context, this paper studied the formation of anuran reproductive aggregations in 16 aquatic breeding sites belonging to six localities in the Atlantic Forest in the state of São Paulo, with the main objective of testing, using null models, whether niche partitioning occurs between species, considering spectral and spatial occupancy. It were recorded the advertisement calls of males from each species present in breeding sites, and their calling sites were characterized by their nature and height of the calling substrate. Tests for null models showed absence of significant patterns both on acoustic domain and the distribution on vertical gradient. Call duration did not reduce the species spectral overlap in the breeding sites and it is possible that reproductive females locate their specific males in heterospecific aggregations based on dominant frequency and not suffer with masking effect on acoustic communication. However, this is a hypothesis that should be tested within the anuran breeding aggregations in the Atlantic forest. All results support the idea that other factors, such as abiotic conditions, environmental heterogeneity and phylogenetic diversity, may be more decisive to explain the occurrence of species in the rich aggregations of breeding frogs in the Atlantic forest.

From the individual to species level, it is common for animals to have connections with one another. These connections can exist in a variety of forms; from the social relationships within an animal society, to hybridisation between species. The structure of these connections in animal systems can be depicted using networks, often revealing non-trivial structure which can be biologically informative. Understanding the factors which drive the structure of animal networks can help us understand the costs and benefits of forming and maintaining relationships. Multivariate modelling provides a means to evaluate the relative contributions of a set of explanatory factors to a response variable. However, conventional modelling approaches use statistical tests which are unsuitable for the dependencies inherent in network and relational data. A solution to this problem is to use specialised models developed in the social sciences, which have a long history in modelling human social networks. Taking predictive multivariate models from the social sciences and applying them to animal networks is attractive given that current analytical approaches are predominantly descriptive. However, these models were developed for human social networks, where participants can self-identify relationships. In contrast, relationships between animals have to be inferred through observations of associations or interactions, which can introduce sampling bias and uncertainty to the data. Without appropriate care, these issues could lead us to make incorrect or overconfident conclusions about our data. In this thesis, we use an established network model, the multiple regression quadratic assignment procedure (MRQAP), and propose approaches to facilitate the application of this model in animal network studies. Through demonstrating these approaches on three animal systems, we make new biological findings and highlight the importance of considering data-sampling issues when analysing networks. Additionally, our approaches have wider applications to animal network studies where relationships are inferred through observing dyadic interactions.

A competição interespecífica pode representar uma força importante na determinação da distribuição e da abundância de organismos na natureza. Comunidades naturais compostas por espécies que utilizam o mesmo recurso de forma semelhante supostamente são estruturadas pela competição interespecífica. Tal força interativa no decorrer da evolução pode ter levado à diferenciação de nicho entre os competidores no passado, sendo esse fator relevante para explicar a coexistência de espécies ecologicamente similares. As comunidades reprodutivas de anuros tropicais são caracterizadas por alta diversidade e grande sobreposição espacial de espécies, por isso representam um ótimo modelo de estudo para avaliar o papel da competição em escala local. Nessas comunidades reprodutivas a alta densidade de machos de diferentes espécies em atividade de vocalização poderia causar problemas de comunicação intraespecífica por interferência acústica. Por essa razão, considerando a hipótese de que o ambiente acústico pudesse ser partilhado, os pesquisadores vêm avaliando, na grande maioria das vezes de forma empírica, estratégias que as espécies de anuros podem empregar para reduzir a competição por canais acústicos no interior da comunidade, como a separação espectral (uso de diferentes faixas de frequência) e a diferenciação de uso espacial e/ou temporal. Nesse contexto, no presente trabalho estudei agregações de anuros formadas em 16 ambientes reprodutivos de seis localidades de Mata Atlântica do estado de São Paulo, com o principal objetivo de testar, por meio de modelos nulos, se ocorre partilha de nicho acústico e espacial entre as espécies. Foram gravados os cantos de anúncio dos machos de cada espécie presente nos ambientes reprodutivos, e seus sítios de vocalização foram caracterizados quanto à natureza e altura do substrato. Os testes por modelos nulos não evidenciaram qualquer padrão, tanto na ocupação do ambiente acústico como na distribuição no gradiente vertical. A propriedade acústica temporal \"duração de canto\" não reduziu a sobreposição espectral das espécies no ambiente reprodutivo, e é possível que as fêmeas, da grande maioria das espécies, localizem seus parceiros em agregações heteroespecíficas com base na frequência dominante e não sofram prejuízo com a interferência acústica. Contudo, essa é uma hipótese que ainda deve ser testada dentro das agregações reprodutivas de anuros da Mata Atlântica. Os resultados, de maneira geral, corroboraram a ideia de que fatores abióticos, heterogeneidade ambiental e diversidade filogenética podem ser mais importantes para explicar a ocorrência de espécies nas ricas agregações de anuros na Mata Atlântica. / Interspecific competition may represent a relevant force determining the distribution and abundance of organisms in nature. Natural communities composed by species that use the same resource in a similar fashion are, supposedly, structured by interspecific competition. This interactive force in the course of evolution may have led to niche differentiation among competitors in the past, and this is a relevant factor to explain the coexistence of ecologically similar species in the same habitat. Reproductive aggregations of tropical frogs are characterized by high diversity and large spatial species overlap, therefore they represent an excellent model to evaluate the importance of competition on local scale. In these communities the high density of breeding males of different species in calling activity could cause a masking effect on intraspecific acoustic communication. Therefore, considering the hypothesis that the acoustic environment could be shared, researchers have been evaluating, in most cases empirically, strategies employed by frogs to reduce competition for acoustic channels within the community, as the spectral separation (using different frequency bands) and spatial and/or temporal segregation. In this context, this paper studied the formation of anuran reproductive aggregations in 16 aquatic breeding sites belonging to six localities in the Atlantic Forest in the state of São Paulo, with the main objective of testing, using null models, whether niche partitioning occurs between species, considering spectral and spatial occupancy. It were recorded the advertisement calls of males from each species present in breeding sites, and their calling sites were characterized by their nature and height of the calling substrate. Tests for null models showed absence of significant patterns both on acoustic domain and the distribution on vertical gradient. Call duration did not reduce the species spectral overlap in the breeding sites and it is possible that reproductive females locate their specific males in heterospecific aggregations based on dominant frequency and not suffer with masking effect on acoustic communication. However, this is a hypothesis that should be tested within the anuran breeding aggregations in the Atlantic forest. All results support the idea that other factors, such as abiotic conditions, environmental heterogeneity and phylogenetic diversity, may be more decisive to explain the occurrence of species in the rich aggregations of breeding frogs in the Atlantic forest.

Coral reefs are the most diverse marine ecosystem, but are increasingly threatened by local and global anthropogenic changes. In this thesis, I examine the impact of local stressors on the spatial variability of coral reef fish community composition by modeling the !-diversity of 35 islands across the Pacific Ocean that are characterized by either low or high human disturbance. By examining !-diversity across three spatial scales (within island, within island group, and across island group), and using null models to control for differences in alpha-diversity or abundance, I reveal previously undocumented effects of human disturbance on coral reef fish assemblages. At all scales, human disturbances alter !-diversity. At the largest-scale, islands with high human disturbance have lower incidence- and abundance-based !-diversity, consistent with biotic homogenization. This pattern was driven by both species with high and low abundances that differed across islands. At the smaller two scales (within islands or island groups), the presence of low abundance species is more variable on islands with high human disturbance (manifest in greater incidence-based !-diversity), but these islands have lower abundance-based !- diversity driven by moderately abundant and widespread species. Multivariate techniques show that islands with high human disturbance have a weaker species-environment relationship, and as such, I suggest that homogenization of coral reef fish assemblages by human disturbances is resulting in greater stochasticity of species composition. / Graduate / 0329 / wiwchar@uvic.ca

A estrutura de comunidades em poças tem sido explicada por competição de recursos, predação e permanência da água. No presente estudo, eu avaliei o efeito do gradiente predação-hidroperíodo sobre a riqueza de espécies e composição de guildas de girinos em poças, e também a co-ocorrência das espécies, testando a existência de padrões segregados que podem ser atribuídos as interações competitivas. O estudo foi realizado no município de Caçapava do Sul, Rio Grande do Sul, sul do Brasil. Eu amostrei girinos e predadores invertebrados em 38 poças, com diferentes tamanhos e graus de permanência. As coletas foram realizadas em duas estações: na primavera de 2007 (novembro) e no verão de 2008 (janeiro). Eu medi a área e a profundidade das poças em cada evento de coleta e registrei a riqueza da vegetação aquática e das margens, assim como a abundância e riqueza de predadores. Eu classifiquei as espécies em guildas baseadas em oito características eco-morfológicas. As espécies foram agrupadas através de ligação completa, usando o coeficiente de Gower. Eu analisei a relação da riqueza de girinos com as variáveis ambientais através de regressão múltipla, e usei Análise de Correspondência Canônica para relacionar a composição das guildas com os descritores ambientais. A co-ocorrência dos girinos foi analisada através de três algoritmos de modelo nulo: linhas e colunas fixas, linhas fixas e colunas proporcionais descartando-se matrizes degeneradas, linhas fixas e colunas proporcionais incluindo matrizes degeneradas. Nove matrizes de presença ausência foram testadas: ocorrências totais (primavera e verão) do conjunto total de espécies, e distinguindo as espécies em duas guildas (bentônicos e nectônicos), e ocorrências de primavera e verão para o conjunto total e as duas guildas. Das 38 poças com água na primavera, somente 22 permaneceram com água na amostragem de verão. Ao todo 10852 indivíduos de 21 espécies foram coletados. As espécies foram classificadas em sete guildas: suspensor-raspador (Sr), nectônicos (N), suspensor-filtrador (Sf), bentônicos com ninho de espuma, com e sem agregação (BI e BII), e bentônicos sem ninho de espuma com olhos dorsais ou laterais (BIII e BIV). A riqueza de espécies cresceu com a abundância de predadores, que foi a única variável explanatória relacionada a mesma. A composição das guildas foi explicada pela abundância de predadores e a profundidade média registrada na primavera. A co-ocorrência dos girinos mostrou padrões segregados para 14 dos 25 modelos testados. A maioria dos modelos com matrizes degeneradas rejeitou a hipótese nula. Os modelos com linhas e colunas fixas, e com as linhas fixas, colunas proporcionais e sem matrizes degeneradas mostraram muitos resultados similares. A segregação também foi influenciada pela combinação de espécies e estação consideradas. Em geral, para o conjunto total de espécies a coocorrência foi aleatória, para os bentônicos foi segregada para as ocorrências totais e de primavera, para os nectônicos apenas no verão. Os resultados das análises de gradiente e co-ocorrência sugerem que girinos são regulados pelos predadores que filtram a composição das guildas e podem reduzir a competição. Para girinos bentônicos, que ocupam poças com baixa abundância de predadores, competição pode ser um fator de estruturação importante. / The community structure of pond systems has been shown to be affected by resource competition, predation and permanency. Here, I evaluated the effect of the permanency-predation gradient in ponds on species richness and guild composition of tadpoles, and accessed the co-occurrence, searching segregated patterns that could be attributed to competitive interactions. The study was carried out Caçapava do Sul, Rio Grande do Sul state, southern Brazil. I sampled tadpoles and invertebrate predators in 38 ponds, with distinct sizes and permanency degrees. The collections were carried out twice: in spring 2007 (November) and summer, 2008 (January). I measured area and depth at each pond in each collection event, and recorded richness of aquatic and bank vegetation, and abundance and richness of predators. I conducted a guild classification of the species pool based on eight variables related to morphological and ecological features. Species were grouped through complete linkage agglomerative clustering method, using Gower coefficient. I evaluated the relation of tadpole richness with environmental variables using multiple regression, and Canonical Correspondence Analysis for guild composition. The tadpole co-occurrence was analyzed through three null models algorithm: fixed rows and columns; fixed rows, proportional columns, discharging degenerate matrices; and fixed rows, proportional columns, including degenerate matrices. Nine presence/absence matrices were tested: total occurrences (spring and summer) of the complete species pool, and distinguishing species of two tadpole guild (benthonic and nektonic), and spring and summer occurrence of total species pool and two guilds. Of the 38 ponds with water in the spring sampling, only 22 remained with water in the following summer sampling. A total of 10,852 individuals and 21 species of tadpoles were collected in the two sampling events. The species were classified into seven guilds: suspension-rasper (Sr), nektonics (N), suspension-feeder (Sf), benthonics with foam nests with and without school behavior (BI and BII), and benthonics without foam nests with lateral and dorsal eyes (BIII and BIV). Species richness increased with abundance of predators, which was the only explanatory variable related to it. Guild composition was explained by abundance of predators and mean depth recorded in spring, which was correlated with maximal depth. The tadpole co-occurrence showed a segregated pattern for 14 of the 25 tested models. Most models with degenerate matrices rejected the null hypothesis. The models with row and columns fixed, and row fixed, column proportional without degenerate matrices showed more similar results. Segregation was also influenced by the combination of species dataset and the season considered. In general, for the total species pool the co-occurrence was random, for benthonic species was segregated for the total and spring occurrences, and for nektonic in summer. The results of gradient and co-occurrence analysis suggest that tadpoles are regulated by predators that filter guild composition and could reduce competition. For benthonic species, which inhabit ponds with lower abundance of predators, competition can be a major structuring constraint.

A estrutura de comunidades em poças tem sido explicada por competição de recursos, predação e permanência da água. No presente estudo, eu avaliei o efeito do gradiente predação-hidroperíodo sobre a riqueza de espécies e composição de guildas de girinos em poças, e também a co-ocorrência das espécies, testando a existência de padrões segregados que podem ser atribuídos as interações competitivas. O estudo foi realizado no município de Caçapava do Sul, Rio Grande do Sul, sul do Brasil. Eu amostrei girinos e predadores invertebrados em 38 poças, com diferentes tamanhos e graus de permanência. As coletas foram realizadas em duas estações: na primavera de 2007 (novembro) e no verão de 2008 (janeiro). Eu medi a área e a profundidade das poças em cada evento de coleta e registrei a riqueza da vegetação aquática e das margens, assim como a abundância e riqueza de predadores. Eu classifiquei as espécies em guildas baseadas em oito características eco-morfológicas. As espécies foram agrupadas através de ligação completa, usando o coeficiente de Gower. Eu analisei a relação da riqueza de girinos com as variáveis ambientais através de regressão múltipla, e usei Análise de Correspondência Canônica para relacionar a composição das guildas com os descritores ambientais. A co-ocorrência dos girinos foi analisada através de três algoritmos de modelo nulo: linhas e colunas fixas, linhas fixas e colunas proporcionais descartando-se matrizes degeneradas, linhas fixas e colunas proporcionais incluindo matrizes degeneradas. Nove matrizes de presença ausência foram testadas: ocorrências totais (primavera e verão) do conjunto total de espécies, e distinguindo as espécies em duas guildas (bentônicos e nectônicos), e ocorrências de primavera e verão para o conjunto total e as duas guildas. Das 38 poças com água na primavera, somente 22 permaneceram com água na amostragem de verão. Ao todo 10852 indivíduos de 21 espécies foram coletados. As espécies foram classificadas em sete guildas: suspensor-raspador (Sr), nectônicos (N), suspensor-filtrador (Sf), bentônicos com ninho de espuma, com e sem agregação (BI e BII), e bentônicos sem ninho de espuma com olhos dorsais ou laterais (BIII e BIV). A riqueza de espécies cresceu com a abundância de predadores, que foi a única variável explanatória relacionada a mesma. A composição das guildas foi explicada pela abundância de predadores e a profundidade média registrada na primavera. A co-ocorrência dos girinos mostrou padrões segregados para 14 dos 25 modelos testados. A maioria dos modelos com matrizes degeneradas rejeitou a hipótese nula. Os modelos com linhas e colunas fixas, e com as linhas fixas, colunas proporcionais e sem matrizes degeneradas mostraram muitos resultados similares. A segregação também foi influenciada pela combinação de espécies e estação consideradas. Em geral, para o conjunto total de espécies a coocorrência foi aleatória, para os bentônicos foi segregada para as ocorrências totais e de primavera, para os nectônicos apenas no verão. Os resultados das análises de gradiente e co-ocorrência sugerem que girinos são regulados pelos predadores que filtram a composição das guildas e podem reduzir a competição. Para girinos bentônicos, que ocupam poças com baixa abundância de predadores, competição pode ser um fator de estruturação importante. / The community structure of pond systems has been shown to be affected by resource competition, predation and permanency. Here, I evaluated the effect of the permanency-predation gradient in ponds on species richness and guild composition of tadpoles, and accessed the co-occurrence, searching segregated patterns that could be attributed to competitive interactions. The study was carried out Caçapava do Sul, Rio Grande do Sul state, southern Brazil. I sampled tadpoles and invertebrate predators in 38 ponds, with distinct sizes and permanency degrees. The collections were carried out twice: in spring 2007 (November) and summer, 2008 (January). I measured area and depth at each pond in each collection event, and recorded richness of aquatic and bank vegetation, and abundance and richness of predators. I conducted a guild classification of the species pool based on eight variables related to morphological and ecological features. Species were grouped through complete linkage agglomerative clustering method, using Gower coefficient. I evaluated the relation of tadpole richness with environmental variables using multiple regression, and Canonical Correspondence Analysis for guild composition. The tadpole co-occurrence was analyzed through three null models algorithm: fixed rows and columns; fixed rows, proportional columns, discharging degenerate matrices; and fixed rows, proportional columns, including degenerate matrices. Nine presence/absence matrices were tested: total occurrences (spring and summer) of the complete species pool, and distinguishing species of two tadpole guild (benthonic and nektonic), and spring and summer occurrence of total species pool and two guilds. Of the 38 ponds with water in the spring sampling, only 22 remained with water in the following summer sampling. A total of 10,852 individuals and 21 species of tadpoles were collected in the two sampling events. The species were classified into seven guilds: suspension-rasper (Sr), nektonics (N), suspension-feeder (Sf), benthonics with foam nests with and without school behavior (BI and BII), and benthonics without foam nests with lateral and dorsal eyes (BIII and BIV). Species richness increased with abundance of predators, which was the only explanatory variable related to it. Guild composition was explained by abundance of predators and mean depth recorded in spring, which was correlated with maximal depth. The tadpole co-occurrence showed a segregated pattern for 14 of the 25 tested models. Most models with degenerate matrices rejected the null hypothesis. The models with row and columns fixed, and row fixed, column proportional without degenerate matrices showed more similar results. Segregation was also influenced by the combination of species dataset and the season considered. In general, for the total species pool the co-occurrence was random, for benthonic species was segregated for the total and spring occurrences, and for nektonic in summer. The results of gradient and co-occurrence analysis suggest that tadpoles are regulated by predators that filter guild composition and could reduce competition. For benthonic species, which inhabit ponds with lower abundance of predators, competition can be a major structuring constraint.

A estrutura de comunidades em poças tem sido explicada por competição de recursos, predação e permanência da água. No presente estudo, eu avaliei o efeito do gradiente predação-hidroperíodo sobre a riqueza de espécies e composição de guildas de girinos em poças, e também a co-ocorrência das espécies, testando a existência de padrões segregados que podem ser atribuídos as interações competitivas. O estudo foi realizado no município de Caçapava do Sul, Rio Grande do Sul, sul do Brasil. Eu amostrei girinos e predadores invertebrados em 38 poças, com diferentes tamanhos e graus de permanência. As coletas foram realizadas em duas estações: na primavera de 2007 (novembro) e no verão de 2008 (janeiro). Eu medi a área e a profundidade das poças em cada evento de coleta e registrei a riqueza da vegetação aquática e das margens, assim como a abundância e riqueza de predadores. Eu classifiquei as espécies em guildas baseadas em oito características eco-morfológicas. As espécies foram agrupadas através de ligação completa, usando o coeficiente de Gower. Eu analisei a relação da riqueza de girinos com as variáveis ambientais através de regressão múltipla, e usei Análise de Correspondência Canônica para relacionar a composição das guildas com os descritores ambientais. A co-ocorrência dos girinos foi analisada através de três algoritmos de modelo nulo: linhas e colunas fixas, linhas fixas e colunas proporcionais descartando-se matrizes degeneradas, linhas fixas e colunas proporcionais incluindo matrizes degeneradas. Nove matrizes de presença ausência foram testadas: ocorrências totais (primavera e verão) do conjunto total de espécies, e distinguindo as espécies em duas guildas (bentônicos e nectônicos), e ocorrências de primavera e verão para o conjunto total e as duas guildas. Das 38 poças com água na primavera, somente 22 permaneceram com água na amostragem de verão. Ao todo 10852 indivíduos de 21 espécies foram coletados. As espécies foram classificadas em sete guildas: suspensor-raspador (Sr), nectônicos (N), suspensor-filtrador (Sf), bentônicos com ninho de espuma, com e sem agregação (BI e BII), e bentônicos sem ninho de espuma com olhos dorsais ou laterais (BIII e BIV). A riqueza de espécies cresceu com a abundância de predadores, que foi a única variável explanatória relacionada a mesma. A composição das guildas foi explicada pela abundância de predadores e a profundidade média registrada na primavera. A co-ocorrência dos girinos mostrou padrões segregados para 14 dos 25 modelos testados. A maioria dos modelos com matrizes degeneradas rejeitou a hipótese nula. Os modelos com linhas e colunas fixas, e com as linhas fixas, colunas proporcionais e sem matrizes degeneradas mostraram muitos resultados similares. A segregação também foi influenciada pela combinação de espécies e estação consideradas. Em geral, para o conjunto total de espécies a coocorrência foi aleatória, para os bentônicos foi segregada para as ocorrências totais e de primavera, para os nectônicos apenas no verão. Os resultados das análises de gradiente e co-ocorrência sugerem que girinos são regulados pelos predadores que filtram a composição das guildas e podem reduzir a competição. Para girinos bentônicos, que ocupam poças com baixa abundância de predadores, competição pode ser um fator de estruturação importante. / The community structure of pond systems has been shown to be affected by resource competition, predation and permanency. Here, I evaluated the effect of the permanency-predation gradient in ponds on species richness and guild composition of tadpoles, and accessed the co-occurrence, searching segregated patterns that could be attributed to competitive interactions. The study was carried out Caçapava do Sul, Rio Grande do Sul state, southern Brazil. I sampled tadpoles and invertebrate predators in 38 ponds, with distinct sizes and permanency degrees. The collections were carried out twice: in spring 2007 (November) and summer, 2008 (January). I measured area and depth at each pond in each collection event, and recorded richness of aquatic and bank vegetation, and abundance and richness of predators. I conducted a guild classification of the species pool based on eight variables related to morphological and ecological features. Species were grouped through complete linkage agglomerative clustering method, using Gower coefficient. I evaluated the relation of tadpole richness with environmental variables using multiple regression, and Canonical Correspondence Analysis for guild composition. The tadpole co-occurrence was analyzed through three null models algorithm: fixed rows and columns; fixed rows, proportional columns, discharging degenerate matrices; and fixed rows, proportional columns, including degenerate matrices. Nine presence/absence matrices were tested: total occurrences (spring and summer) of the complete species pool, and distinguishing species of two tadpole guild (benthonic and nektonic), and spring and summer occurrence of total species pool and two guilds. Of the 38 ponds with water in the spring sampling, only 22 remained with water in the following summer sampling. A total of 10,852 individuals and 21 species of tadpoles were collected in the two sampling events. The species were classified into seven guilds: suspension-rasper (Sr), nektonics (N), suspension-feeder (Sf), benthonics with foam nests with and without school behavior (BI and BII), and benthonics without foam nests with lateral and dorsal eyes (BIII and BIV). Species richness increased with abundance of predators, which was the only explanatory variable related to it. Guild composition was explained by abundance of predators and mean depth recorded in spring, which was correlated with maximal depth. The tadpole co-occurrence showed a segregated pattern for 14 of the 25 tested models. Most models with degenerate matrices rejected the null hypothesis. The models with row and columns fixed, and row fixed, column proportional without degenerate matrices showed more similar results. Segregation was also influenced by the combination of species dataset and the season considered. In general, for the total species pool the co-occurrence was random, for benthonic species was segregated for the total and spring occurrences, and for nektonic in summer. The results of gradient and co-occurrence analysis suggest that tadpoles are regulated by predators that filter guild composition and could reduce competition. For benthonic species, which inhabit ponds with lower abundance of predators, competition can be a major structuring constraint.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / We investigated the potential factors structuring anuran assemblages using three approaches: i) spatial structure of adult anurans; ii) spatial structure of tadpole assemblages, and iii) temporal structure of reproductive activity. The assemblage of adult anurans was composed by 22 species which co-occurred more frequently than expected by chance at the breeding sites (ponds). Pond area, depth, percentage of flat margin, and pH were the environmental descriptors structuring the assemblage. Additionally, closer ponds were more similar regarding assemblage structure. About 16% of the assemblage structure was explained exclusively by environmental descriptors, 13% exclusively by geographic distance, and 3.7% by the interaction between them. Five species groups were found regarding patterns of pond use, ranging from occasional occurrences to generalists. The tadpole assemblage comprised of 20 species belonging to six ecomorphological guilds, being the benthic the richest guild. We found a positive co-occurrence between species both within that guild and for all species together. Pond area, hydroperiod, and percentage of water covered by vegetation were the main descriptors structuring the assemblages for both scales (species and guilds). Considering species together, 20% of the assemblage structure was explained exclusively by pond environmental descriptors, and spatial influence (3%) was related to those descriptors. For guilds, only environmental descriptors explained the assemblage structure (11%). Groups of similar pond use pattern were segregated by hydroperiod. We detected partitioning among species during the reproductive season, with four guilds regarding temporal pattern: 1. Prolonged calling activity throughout the season; 2. More intense activity at the beginning of the season; 3. More intense at the end, and 4. Occasional occurrences. We found positive correlation of climatic variables with 44% of the species and total abundance of anurans. Our study showed that the assemblages were structured in space and time, with the reproductive activity of the species responding to the local climatic variation. Moreover, we suggest that the maintenance of ponds, which vary in a gradient regarding environmental descriptors, is important to maintain the distinct anuran strategies in this region. / Investigamos os potenciais fatores estruturadores de taxocenoses de anuros através de três abordagens: i) estruturação espacial de adultos; ii) estruturação espacial de girinos; e iii) estruturação temporal da atividade reprodutiva. A taxocenose de anuros em fase reprodutiva foi composta por 22 espécies que ocorreram em maior frequência que o esperado ao acaso nos ambientes reprodutivos (corpos d água). Área do corpo d água, profundidade, porcentagem de margem plana e pH foram os descritores ambientais estruturadores da taxocenose. Adicionalmente, corpos d água mais próximos foram mais semelhantes quanto à estrutura da taxocenose. Aproximadamente 16% da estrutura da taxocenose foram explicados exclusivamente pelos descritores ambientais, 13% exclusivamente pela distância geográfica, e 3,7% pela interação entre esses. As espécies foram agrupadas em cinco padrões de uso dos corpos d água, variando de ocorrências pontuais a generalistas. A taxocenose de girinos foi composta por 20 espécies pertencentes a seis guildas ecomorfológicas, sendo a bentônica a mais rica. A coocorrência das espécies foi maior que ao acaso tanto dentro desta guilda, como para toda a taxocenose. Área dos corpos d água, hidroperíodo e porcentagem de cobertura vegetal da água foram estruturadores da taxocenose em ambas as escalas (espécies e guildas). Considerando as espécies em conjunto, 20% da estrutura da taxocenose foram explicados unicamente pelos descritores ambientais dos corpos d água, sendo que toda contribuição espacial (3%) foi relacionada às características ambientais. Para guildas, apenas descritores ambientais explicaram a estrutura da taxocenose (11%). Grupos de espécies com uso similar dos corpos d água foram segregados em função do hidroperíodo dos mesmos. Encontramos partilha temporal da estação reprodutiva entre as espécies, com a indicação de quatro guildas: 1. Atividade de vocalização prolongada ao longo da estação; 2. Mais intensa no início da estação; 3. Mais intensa no final da estação; e 4. Espécies com ocorrências pontuais ao longo da estação chuvosa. Para 44% das espécies e para a abundância total de anuros, encontramos correlações significativas com os fatores meteorológicos. Nosso estudo mostrou que as taxocenoses estão estruturadas espacial e temporalmente, com a atividade reprodutiva das espécies respondendo à variação das condições meteorológicas vigentes. Além disso, sugerimos que a manutenção de corpos d água que variem num gradiente de descritores ambientais é importante para abranger as diferentes estratégias reprodutivas dos anuros da região.