Local vs. regional control on species richness: a new approach to test for competitive exclusion at the community level

Abstract

ABSTRACTAim The degree to which ecosystem properties (niche availability) and interspecific competition limit local species richness continues to be debated. We present a new method that takes advantage of special biogeographical conditions where habitat islands are embedded in a moderately dissimilar matrix. Small, isolated islands lack habitat specialists to varying degrees, allowing matrix species to occupy island habitat. We examined the relative influence of local vs. biogeographical factors on bird richness of habitat islands by evaluating three predictions each for three hypotheses (competitive exclusion, mass effect, species‐area effects).Location Five naturally isolated dry forest islands and eight extensive dry forest sites in Bolivia.Methods Linear and multiple regressions were calculated between standardized species richness, the number of typical dry forest, humid forest and habitat generalist species vs. canopy height, elevation, precipitation, habitat area and distance from evergreen forest. The same analyses were performed at the guild level.Results Our data are consistent with one prediction under each hypothesis, thus partly supporting and partly rejecting all three hypotheses. Bird richness of dry forest islands was primarily determined by local ecosystem properties (vegetation complexity, productivity), with a minor effect of habitat area (inverse) and distance to evergreen forest. The number of dry forest specialists was influenced by habitat area and isolation from source habitats, with elevation and precipitation adding moderate explanatory power. The number of humid forest specialists was determined by precipitation, habitat area (inversely) and canopy height. These patterns were also evident at the guild level, demonstrating that the loss of dry forest species in a given guild was correlated with a gain of humid forest species in the same guild, rather than random species turnover.Main conclusionsTotal local species richness of semi‐isolated habitat islands is determined primarily by local (niche availability) but also by regional (biogeographical) factors, exemplifying MacArthur's paradox. Small dry forest islands historically lost part of their habitat specialists. Vacant niches were filled by species from adjacent humid forest that are competitively excluded by habitat specialists from large dry forest areas, ideally until the community is saturated, although this pattern is to some degree modified by biogeographical factors. This ‘semipermeable island’ approach can be applied to other taxa and regions and is a promising, widely applicable tool for gaining new insights into the regulation of diversity.