Historical biogeography of small mammals in the Sunda region of Southeast Asia.

Abstract

The Sunda region of Southeast Asia was transformed during Pleistocene glaciations when land bridges connected Borneo, Java, and Sumatra to the mainland. Layered onto these topographic transformations were fluctuations in vegetational zonation. How did these changes affect small mammals of the region? First, a phylogeographic test was conducted of the hypothesis that Pleistocene land bridges enabled three rainforest rodents (<italic>Maxomys surifer, Leopoldamys sabanus</italic>, and <italic>Maxomys whiteheadi</italic>) to disperse throughout the Sunda region. The results for each taxon suggest deep divergences among island populations, a refutation of the hypothesis tested. An alternative hypothesis is proposed involving an early vicariant diversification of lineages following the Pliocene breakup of the Sunda Block. Second, a population genetic approach was used to test the hypothesis that a savanna corridor that formed in the Sunda region during the Pleistocene---a potential explanation for continued isolation among island populations despite the emergence of land bridges---extinguished populations of <italic>M. surifer, L. sabanus</italic>, and <italic>M. whiteheadi</italic>. I tested for bottlenecks and subsequent expansion of populations in southwestern Borneo, which fell within the corridor, and for patterns of population stability in northern Borneo, which remained forested throughout the Pleistocene. The signatures of bottlenecks and expansions are identified in some populations, but no pattern is found that suggests populations were uniformly adversely affected in the southwest but not in the north. Third, a test of the hypothesis that the richness of montane endemic mammals declines from northern to southern Borneo was conducted. The results of trapping studies on three mountains of southwestern Borneo are reported, extending southward the limits of systematic elevational small mammal work. Range extensions are reported for three montane species, but the gradient hypothesis is not refuted. Historical and ecological explanations for the gradient are discussed. Finally, a brief review of Pliocene and Pleistocene rodent fossils from Indochina and the Sunda region was completed to test the hypothesis that savanna rodent communities replaced forest communities during arid periods of glacials. The data demonstrate large historical shifts in the distribution of habitats, but suggest a heterogeneous matrix of forest and grassland rather than the complete elimination of forest.