THE ENVIRONMENTAL CONTROL OF INSULAR VARIATION IN BIRD SPECIES ABUNDANCE

Abstract

From multiple regression analyses of environmental correlates (area, elevation, isolation) of insular bird species abundance in the East Indies, West Indies, and East-Central Pacific, evidence is presented for the following for each archipelago, area is the major predictor of insular variation in species numbers; the species-area curve is nonlinear or curvilinear in the first 2 archipelagos, but linear in the 3rd one; when effects of environmental variables other than area are removed, the species-area slopes (z) by Model II (nonlinear variation) are, respectively, 0.280, 0.237, and 0.303; and, by Model I (linear variation) corrected species-area relations are such that partial regression coefficients are, respectively, 0.0010, 0.0014, and 0.0013. The uniformity of the z values (average = 0.273) closely approaches Preston''s theoretical exponential of 0.270 for species-area curves for isolates, and our findings support his theory insofar as bird species are concerned. Previous studies of plant species abundance in the Galapagos Islands reveal elevation, not area, as the major determinant, and the possibility exists that Preston''s formula for isolates (species number = 10 area0.27) is not applicable to all groups of plant and animal species. The suggestion is made that it will predict most accurately insular species-area relations for space-demanding animals whose individual numbers are primarily regulated by density-dependent means.