The mating system of Shorea megistophylla, an endemic canopy tree from Sri Lanka, was quantified by allozyme analysis of progeny arrays using a mixed-mating model. Two adjacent populations were compared: one in forest that was selectively logged about 20 years ago, and the other in undisturbed primary forest. The selectively logged population had a lower multilocus outcrossing rate (t(m) = 0.71) compared to that of the undisturbed forest (t(m) = 0.87). Only the progeny from the logged population showed evidence of either biparental inbreeding or Wahlund effect, and the genotypic frequencies violated the assumptions of the mixed-mating model. Apomixis was detected in one isolated tree in the Royal Botanical Garden in Peradeniya by a multilocus test of progeny genotypic frequencies relative to the maternal genotype. However, significant levels of apomixis were not discerned in the natural (logged and unlogged) populations. These findings indicate that a reduction in population density of S. megistophylla following selective logging can significantly elevate the proportion of seeds produced through inbreeding. Adventive embryony may also increase in isolated trees that lack the chance to outcross.