Additive partitioning of spider diversity in a fragmented tropical dry forest (Valle del Cauca, Colombia)

Understanding the variation of diversity patterns requires analysis at multiple spatial scales. In this study we estimated the diversity components (alpha, beta and gamma) of the spider community at El Vnculo Natural Regional Park, using the additive partitioning of diversity (species richness, Shannon's diversity index and Simpson's index) for the first time on this taxon in Colombia. We collected the specimens following a nested sampling design that consisted of two spatial scales. At the local scale, we quantified additive diversity components in 238 sampling units, and at the regional scale in five vegetation types. Total observed regional diversity (γ) was partitioned into its additive components: within sampling units (α1), among sampling units (βl) and among vegetation types (β2). We used the same approach to compare common and infrequent spider species and to compare sampling methods. A total of 1565 adult spiders and 72 identifiable juveniles, including 193 morphospecies from 36 families, was sampled during the study. In all cases (entire community, infrequent species, common species and four different sampling methods) we found that a significant percentage, relative to that of randomization tests, of the diversity measurements used was attributed to beta diversity among vegetation types. The relative contributions of alpha and beta diversity to total observed regional diversity depended on the diversity measurement used. The contribution of beta diversity with respect to alpha diversity was low using Simpson's index (less than 20), whereas with species richness and Shannon's index the contribution was high (up to 90 and up to 66, respectively). Our results suggest that beta diversity is the main component of diversity in the natural park. We concluded that the maintenance of a large variety of vegetation types can be an important tool for the conservation of spider richness at the natural park.