Nestedness and discontinuities in species-range-size distributions.

Gaston's1 discussion of the patterns and mechanisms shaping distribution curves of species-range-sizes leaves out two important issues related to scale and hierarchical structuring of biota. Several provide evidence that species distributions are nested, that is, overlapping in a particular manner in which species with broad ranges often contain much smaller ranges of other species. In nested systems, range-size distributions should be positively skewed because of the structural limitations on the number of large size-ranges. 'Nestedness' may give an even richer insight into ties among patterns of distribution, ecological determinants and speciation processes that were hinted at by Gastonl. In particular, the niche breakage metaphor for species ranges would be easier to make if those ranges were seen as being influenced by ecological factors such as hierarchical habitat structure and community assembly, which involves a generalist-specialist gradient of habitat use. Habitat specialists have ecological ranges that are often nested within those of generalists5. It is possible that this community-level observation is but a hint of a more general mechanism responsible for patterns that span several spatial scales and that range size is correlated with ecological range of species. If this is the case, the wealth of ecological information could be used when interpreting biogeographical patterns, including prediction. Another observation adds to this suggested link between the habitat level and geographical level pattern. Nested use of habitat necessarily implies discontinuities in ecological range distributions6. Such discontinuities were found in a variety of taxa, from marine foraminiferans to coral reef fishes and desert rodents. Notably, discontinuities in geographical ranges have also been found in North American birds of prey6 (data in Brown7). The discontinuities, while corroborating the hierarchical perspective, imply a potential change of focus away from the preoccupation with the log-normal statistics of distribution to other modes of description and analysis. For example, i t may be useful to determine how nestedness and discontinuities correlate with habitat heterogeneity and fragmentation, both factors being fundamental to conservation, biodiversity maintenance and speciation dynamics.