THE exceptional species diversity of tropical rain forests is well known; of these the lowland forests of the Malay Peninsula are among the richest1. This, and the size of the trees, pose obvious difficulties for the maintenance of pan-mixis, and tempt speculation on the processes by which such diversity has evolved and may now be favoured. Fedorov2 suggested that natural selection is low and self-pollination prevalent, favouring random genetic drift. Prevalence of random drift, or apomixis—agamospermy—would contradict Ashton's3 view that, given a stable environment, evolution proceeds through ecotypic differentiation while diversity accrues through ever increasing niche specialisation; for this view to be correct, maintenance of genetic variability within breeding groups would be essential. Solid evidence has been meagre but there is evidence4 that genetic polymorphism, and hence heterozygosity, is common in the emergent tree Shorea leprosula Miq., and Xerospermum intermedium Radlk., an undeirstorey species. The pattern of spatial variation is consistent with the view that these species are outbreeders in which predominantly short-range pollen and fruit dispersal are accompanied by short-range heterogeneity in gene frequencies. However, both are common trees with minimal spatial isolation. We now report the occurrence of apomixis in some other rain forest trees.