Nonrandom co‐occurrence patterns of rainforest chameleons

Null‐model analysis of co‐occurrence patterns is a powerful tool to identify ‘structure’ in community ecology data sets. We evaluated the community structure of chameleons in rainforest regions of Nigeria and Cameroon using available data in the literature, including peer‐reviewed articles and unpublished environmental reports to industries. We performed Monte Carlo simulations (5000 iterations, using the sequential swap algorithm) under several model assumptions to derive co‐occurrence patterns among species. Food and spatial (habitat) segregation patterns in both lowland rainforest and montane forest were investigated. We subjected four indices of co‐occurrence patterns (C‐ratio, number of checkerboard species pairs, number of species combinations, and V ‐score) to randomization procedures. Overall, the chameleon communities do not show random organization, but instead exhibit precise deterministic patterns. In lowland rainforest, chameleon communities are assembled deterministically along the food niche resource axis, but not along the habitat niche resource axis. The opposite holds for chameleon communities in montane rainforest. We predict that these patterns can be generalized to other regions of tropical Africa, thus helping to determine the general structure of chameleon communities in tropical African forests.