Bat assemblages on Neotropical land‐bridge islands: nested subsets and null model analyses of species co‐occurrence patterns

A fundamental goal of ecology is to understand whether ecological communities are structured according to general assembly rules or are essentially dictated by random processes. In the context of fragmentation, understanding assembly patterns and their mechanistic basis also has important implications for conservation. Using distribution data of 20 bat species collected on 11 islands in Gatún Lake, Panama, we tested for non‐randomness in presence–absence matrices with respect to nestedness and negative species co‐occurrence. We examined the causal basis for the observed patterns and conducted separate analyses for the entire assemblage and for various species submatrices reflecting differences in species’ trophic position and mobility. Furthermore, we explored the influence of weighting factors (area, isolation, abundance) on co‐occurrence analyses. Unweighted analyses revealed a significant negative co‐occurrence pattern for the entire assemblage and for phytophagous bats alone. Weighting analyses by isolation retained a pattern of species segregation for the whole assemblage but nullified the non‐random structure for phytophagous bats and suggested negative associations for animalivores and species with low mobility. Area‐ and abundance‐weighted analyses always indicated random structuring. Bat distributions followed a nested subset structure across islands, regardless of whether all species or different submatrices were analysed. Nestedness was in all cases unrelated to island area but weakly correlated with island isolation for incidence matrices of all species, phytophagous bats, and mobile species. Overall, evidence for negative interspecific interactions indicative of competitive effects was weak, corroborating previous studies based on ecomorphological analyses. Our findings indicate that bat assemblages on our study islands are most strongly shaped by isolation effects and species’ differential movement and colonization ability. From a conservation viewpoint this suggests that even in systems with high fragment–matrix contrast, a purely area‐based approach may be inadequate, and structural and functional connectivity among patches are important to consider in reserve planning.