To assess the impact of habitat fragmentation on tropical avian communities, we sampled lowland forest birds on six land-bridge islands and two mainland forest sites in Lake Kenyir, Peninsular Malaysia using timed point counts, hypothesizing that insectivorous birds are the worst affected guild. We used an information-theoretic approach to evaluate the effects of area, isolation, primary dietary guild (omnivore, frugivore and insectivore) and their interactions in predicting species richness, abundance and diversity. Our analysis showed that a model that considered the effects of area, dietary guild and their interaction best explained observed patterns of species richness. But a model considering both area and dietary guild best explained the variation in abundance. Notably, insectivorous birds were singled out as the dietary guild most sensitive to fragmentation, followed by frugivorous and omnivorous birds and hence provide support for our hypothesis. Assemblages of insectivorous birds were clearly depauperate on anthropogenic forest islands in Lake Kenyir and are consistent with forest fragmentation studies in the Neotropics. Given their specialized foraging ecology and diversity, conservation of intact communities of insectivorous bird guilds in Malaysia will be critical for maintaining predator–prey interactions in lowland tropical forests.

The results are reported of a survey of rodents on 10 forested land-bridge islands ranging in size from 0.2 to 350 ha in the state of Bolívar, Venezuela. The islands were contained within a lake formed c. 12 years before the study by the damming of the Caroni River for hydroelectric power. Rodents were sampled on each island by live-trapping along transects that sampled all available habitat types on each island, and microhabitat structure was measured at each trap station. A total of 674 captures of 359 individuals of six species of rodents was recorded. Species composition changed from the largest to the smallest islands, and small and medium islands (0.2–11 ha) displayed the typical effects of insularity, with fewer species and increased abundances and biomass. The largest island (350 ha) seemed to function more like a mainland. Most species were associated with a suite of microhabitat variables. It is suggested that release from top-down control by predators was responsible for higher abundances and biomass on the smaller islands and that predators moving between large islands and other nearby landmasses help maintain a mainland community structure on large islands. However, changes in species composition on smaller islands may be the result of patchy occurrences of some species before isolation, changes in microhabitat structure following isolation, and species-specific microhabitat requirements.