Amazon diversity in light of the paleoecological record

Disjunct distributions of Amazonian species have been explained previously by a refugial theory which postulates that Amazonian rain forest was preserved in large highland regions throughout the Pleistocene. No direct, radiocarbon dated evidence exists for the last glacial maximum with which to test this theory. The only radiocarbon dates of Pleistocene age from the Amazon basin are of fossiliferous deposits in the proposed Napo refugium of the West, where both pollen assemblages and wood samples indicate that forest with cool Andean elements existed there at two intervals in the last cycle of northern hemisphere glaciation, implying a temperature depression of at least 4°C in the Amazon lowlands.Under modern climatic conditions, lateral erosion by river meander, together with surface erosion, serves as a rejuvenating mechanism for the rain forests of Peru and Ecuador. The instability of late Holocene Amazonian climates is demonstrated by documenting a precipitation event in the eastern Andean cordillera that caused widespread flooding of western Amazonian forests 800–1300 BP. Late Holocene pollen histories from widely dispersed parts of central Amazonia distinguish between vegetation histories in the drainage of northern, south-western and western watersheds, but all show histories of fluctuating intensities of dry seasons. Radiocarbon dating of charcoal layers in soils of Venezuelan Amazonia demonstrates the apparently random incidence of wild fires at wide intervals over at least the last 6 ka.The high species richness of Amazonia is a result of numerous opportunities for vicariance because of a very large total area, wide variety of habitats and intermediate levels of disturbance, particularly by hydrological processes, that has varied on timescales from years to millennia. Amazonian disjunct distributions probably reflect regional environmental discontinuities in both interglacial and glacial times.