Core and Satellite Species: Theory and Artefacts

A widely observed pattern in Nature is a positive correlation between the local abundances of species and their regional distribution, i.e. species which occur at high abundances on any one patch are to be found on more of the patches in a region. Brown (1984) has argued, and we agree, that this pattern is not a sampling artefact. A less widely observed pattern is that the regional distributions of species are bimodally distributed in that most species are either regionally uncommon, occurring on very few patches, or ubiquitous. Hanski (1982) had the insight that the first pattern could be easily incorporated into metapopulation theory and be used to generate the second. In the first part of this communication we will reinterpret Hanski's insight and show that metapopulation theory can easily explain the correlation between local abundance and regional distribution. When we do this, however, we no longer predict a bimodal distribution of regional abundances. We then discuss a number of artefactual reasons why these distributions may sometimes appear to be bimodal. Unless these possibilities can be excluded, it is not clear that bimodality is a phenomenon requiring an ecological explanation. The starting point of the analysis is Levin's (1969) metapopulation model