If the claim that predation risk is the main factor favouring group living in diurnal non-human primates is correct, a reduction in the risk of predation should lead to: (1) a reduction in the size of a primate group and of the parties into which it fissions, (2) a change in the relationship between the size of a party and its composition, and (3) (in South-East Asia) to a change in the relationship between party size and height in the canopy, for small parties. These predictions were tested in a comparison between a population of long-tailed macaques (Macaca fascicularis) in a Sumatran site where a variety of felids occurs, and one on the island of Simeulue, where no cats occur. The first two predictions were strongly supported, whereas the third was weakly supported by the data. The Simeulue monkeys were also different in morphology (blackish coats, reduced sexual dimorphism), lived at high densities in small home ranges, and were territorial. All these differences between the populations could be plausibly ascribed to the difference in predation risk from felids. It is concluded that the presence of felids has a profound influence on the social organization of this arboreal primate species.