Consistency of effects of tropical-forest disturbance on species composition and richness relative to use of indicator taxa.

Lawton et al. (1998) found, in a highly cited study, that the species richness of 8 taxa each responds differently to anthropogenic disturbance in Cameroon forests. Recent developments in conservation science suggest that net number of species is an insensitive measure of change and that understanding which species are affected by disturbance is more important. It is also recognized that all disturbance types are not equal in their effect on species and that grouping species according to function rather than taxonomy is more informative of responses of biodiversity to change. In a reanalysis of most of the original Cameroon data set (canopy and ground ants, termites, canopy beetles, nematodes, and butterflies), we focused on changes in species and functional composition rather than richness and used a more inclusive measure of forest disturbance based on 4 component drivers of change: years since disturbance, tree cover, soil compaction, and degree of tree removal. Effects of disturbance on compositional change were largely concordant between taxa. Contrary to Lawton et al.'s findings, species richness for most groups did not decline with disturbance level, providing support for the view that trends in species richness at local scales do not reflect the resilience of ecosystems to disturbance. Disturbance affected species composition more strongly than species richness for butterflies, canopy beetles, and litter ants. For these groups, disturbance caused species replacements rather than just species loss. Only termites showed effects of disturbance on species richness but not composition, indicating species loss without replacement. Although disturbance generally caused changes in composition, the strength of this relationship depended on the disturbance driver. Butterflies, litter ants, and nematodes were correlated with amount of tree cover, canopy beetles were most strongly correlated with time since disturbance, and termites were most strongly correlated with degree of soil disturbance. There were moderately divergent responses to disturbance between functional feeding groups. Disturbance was most strongly correlated with compositional differences of herbivores within beetles and nematodes and humus feeders within termites. Our results suggest that consideration of the impact of different forms of disturbance on species and functional composition, rather than on net numbers of species, is important when assessing the impacts of disturbance on biodiversity.