Spatial patterns in species-area relationships and species distribution in a West African forest-savanna mosaic

Aim To investigate the relationship between the slope z of the species-area relationship (SAR) and the intensity of spatial patterns in species number and dissimilarity for woody plants with different modes of seed dispersal. According to island theory we expect, for any given archipelago, steeper slopes and more pronounced spatial patterns for groups of less dispersive species. Location Ivory Coast, West Africa. Methods In a West African forest-savanna mosaic we collected presence-absence data for woody plant species in 49 forest islands. The parameters of the SARs were fitted by nonlinear regressions and then compared for plant species aggregated according to their mode of seed dispersal. We used the Mantel test to calculate the intensity of spatial patterns in species number, i.e. residual deviation from SAR, and species dissimilarity. Results The z-value for bird-dispersed species was lower (0.11) than that for wind-dispersed species (0.27), with mammal-dispersed species taking an intermediate value (0.16). This result suggests that, as a group, bird-dispersed species are better colonizers. The spatial pattern in species number as well as species similarity was more pronounced for bird- compared with wind-dispersed species. Main conclusions The standard interpretation of the theory of island biogeography claims that shallow slopes in the SAR imply low isolation of islands, i.e. good dispersal abilities of species. The results of our study appear to contradict this statement. The contradiction can eventually be resolved by a more detailed account of the colonization process, i.e. by distinguishing between dispersal and consecutive establishment of populations.