Taxonomic diversity of island biotas.

TLDR: In this paper, the authors studied the distribution of the mean number of species per genus (S/G) on an island is usually lower than S/G for its presumed source area (MacArthur and Wilson, 1967).

Abstract: Students of biogeography since Darwin have focused disproportionately on oceanic islands. The prime bases for this interest have been the distinct forms which have evolved in the genetic isolation provided by islands and the ecological situation pertaining because the species successfully colonizing any island are but a small subset of the mainland species pool. One aspect of the latter effect which has received attention is that, within any higher taxon, the mean number of species per genus (S/G) on an island is usually lower than S/G for its presumed source area (MacArthur and Wilson, 1967). If it is assumed that congeneric species tend to resemble one another more in any measurable biological characteristic than do less closely related species, then the lower S/G on an island implies a more "diverse" biota on the island than on the source area. Although Williams (1964) pointed out that a random subset of any species pool has an expected S/G lower than that of the entire pool, Moreau (1966) and Grant (1966) attach significance to the lower S/G per se on islands, without regard for whether this S/G is lower or higher than expected, and attribute the lower insular value to ecological and/or evolutionary phenomena. In this paper I will first treat qualitatively the general distribution of the S/G ratio for random subsets of any species pool, then analyze the data for a series of well-studied island groups, and finally reassess the ecological and evolutionary ideas formulated on this subject in the light of the statistical treatment. THEORETICAL CONSIDERATIONS