A GENETIC PERSPECTIVE OF MAMMALIAN VARIATION AND EVOLUTION IN THE INDONESIAN ARCHIPELAGO - BIOGEOGRAPHIC CORRELATES IN THE FRUIT BAT GENUS CYNOPTERUS

This study investigated allozyme and morphometric variability within t he genus Cynopterus, with particular emphasis on C. nusatenggara, whic h is endemic to Wallacea, the area encompassing the Oriental-Australia n biogeographic interface. The genetic distances between Cynopterus sp ecies are small by mammalian standards and suggest that this genus has undergone a recent series of speciation events. The genetic distance between populations of C. nusatenggara is strongly correlated with bot h the contemporary sea-crossing distance between islands and the estim ated sea crossing at the time of the last Pleistocene glacial maximum, 18,000 B.P. This observation, together with low levels of population substructure within islands as shown by F-statistics, indicates that t he sea is a primary and formidable barrier to gene exchange. The genet ic distance and the great-circle populations of C. nusatenggara are no t correlated, although a principal-coordinates analysis of genetic dis tance reveals relationships between the populations that are similar t o their geographical arrangement. A strong negative correlation exists between the level of heterozygosity within island populations of C. n usatenggara and the minimum sea-crossing distance to the nearest large source population. This is interpreted as reflecting an isolation eff ect of the sea, leading to reduced heterozygosity in populations that have larger sea barriers between them and the large source islands. In dependently of this, heterozygosity is negatively associated with long itude, which in turn is associated with systematic changes in the envi ronment such as a gradual decline in rainfall from west to east. The a ssociation between heterozygosity and longitude is interpreted as refl ecting an association between genetic and environmental variance and s upports the niche-width theory of genetic variance. Morphometric varia bility did not show any of the main effects demonstrated in the geneti c data. Furthermore, there was no evidence that, at the level of indiv iduals, genetic and morphometric variability were associated.