Parasite virulence, i.e. the damage done to the host, may be a by-product of the parasite's effort to maximize its fitness. Accordingly, several life-history trade-offs may explain interspecific differences in virulence, but such constraints remain little tested in an evolutionary context. In this phylogenetic study of primate malarias, I investigated the relationship between virulence and other parasite life-history traits. I used peak parasitaemia as a proxy for virulence, because it reflected parasite reproductive success and parasite-induced mortality. Peak parasitaemia was higher in specialist than in generalist species, even when confounding life-history traits were controlled. While there was a significant phylogenetic relationship between the number of competitors per host and host specialization, peak parasitaemia was unrelated to within-host competition. Therefore, the key evolutionary factor that favours virulence is host specialization, and the evolutionary success of virulent parasites, such as Plasmodium falciparum, may be better understood when the trade-off in virulence between different hosts is considered. Such phylogenetic results may help us design better protection programmes against malaria.