On Predicting Insular Variation in Endemism and Sympatry for the Darwin Finches in the Galapagos Archipelago

For interisland variation in number of Darwin Finches in the Galapagos Archipelago, endemism is predicted by nearest-neighbor isolation; and species abundance or sympatry is predicted by average isolation. Nearest-neighbor isolation is measured by distance from the nearest island, and average isolation is the average distance to all other islands in the archipelago. The two measures of isolation are of little predictive value when tested for the avifaunas of six other archipelagos or oceanic island groupings. In these situations, area is a better predictor of species abundance or endemism; and the role of isolation appears only when measured as distance from the major avifaunal source region (e.g., New Guinea for islands of the East-central Pacific; African mainland for islands in the Gulf of Guinea). Insular isolation in these instances is, however, a relatively small contributor to variance of species number when contrasted with the greater contributions made by insular area. That numbers of insular species and endemics are respectively predicted by average and nearest-neighbor isolation, and not by area, only in the Darwin Finches demonstrates emphatically the importance of isolation in regulating endemism and species abundance (= sympatry) in the adaptive radiation of monophyletic bird groups within archipelagos. This generalization appears valid only when the intra-archipelagic speciations are mostly between islands, and not intraisland in site of origin from parental forms. It would appear to be less valid during the postspeciation, phyletic- specialization phase of radiation. It is presumed that the natural regulations of endemism and species abundance for insular avifaunas are stochastic in process. By this hypothesis, chance elements associated with isolation are of major importance early in the adaptive radiation of monophyletic bird groups, and less important later in this radiation when deterministic and chance elements associated with ecology (area, habitat, niche) predominate.