Abstract

The butterfly assemblages of pairs of forest habitats, differing in disturbance level, within the Victoria Mayaro reserve of South-East Trinidad, are described using walk-and-count transects and canopy and understorey fruit traps. The concurrent use of these two butterfly censusing techniques, revealed major but conflicting differences in species accumulation rates under different disturbance conditions. The disturbed evergreen habitat had the significantly highest accumulation rate from walk-and-count data but the significantly lowest from fruit trap data. This reflects the specificity of much of the fruit-feeding guild for closed canopy forest. Disturbed habitats were found to lack a distinct canopy fauna. These results are discussed in light of the intermediate disturbance hypothesis. Within a region of forest, butterflies were found to be more characteristic of a disturbance level than of a particular forest type, lending weight to the belief that butterfly faunas can be used as bioindicators of forest disturbance. Several restricted geographic range species were not adversely affected by forest disturbance, at these levels of disturbance. The butterfly censuses in this study suggest that the optimal strategy for safeguarding butterfly species richness under natural forest management regimes would be to maintain a mosaic of habitats that included areas of undisturbed primary forest and a network of other forest patches, that varied in management regime and level of disturbance.