The checkered history of checkerboard distributions.

To address the idea that the process of interspecific competition can be inferred from data on geographical distribution alone and that evidence from geographical distribution implies an important role for interspecific competition in shaping ecological communities, we reexamine the occurrence of "true checkerboard" distributions among the land and freshwater birds in three Melanesian archipelagoes: Vanuatu, the Bismarck Archipelago, and the Solomon Islands. We use the most recently published distributional records and explicitly include the geography of the distributions of species within each archipelago. We use the overlap of convex hulls to estimate the overlap in the geographic range for each pair of species in each of these archipelagoes. We define a "true checkerboard" to consist of a pair of species with exclusive island-by-island distributions, but that have overlapping geographical ranges. To avoid the "dilution effect," we follow Diamond and Gilpin in focusing only on congeneric and within-guild species pairs as potential competitors. Few, if any, "true checkerboards" exist in these archipelagoes that could possibly have been influenced by competitive interactions, and even "true checkerboards" can arise for reasons other than interspecific competition. The similarity between related species pairs (congeneric and within-guild pairs) and unrelated species pairs in their deviation from expectation of the number of islands shared and the overlap of their geographic ranges indicates that these are not distinct statistical populations, but rather a single population of species pairs. Our result, which is based on an examination of the distributional data alone, is consistent with the interpretation that, in these avifaunas, the distributions of congeneric, within-guild, and unrelated species pairs are shaped by a common set of biological and physical environmental processes.