While studies of canine dimorphism in primates are common, only a few have examined canine tooth size independently within each sex. Recently, Greenfield and Washburn (Am. J. Phys. Anthropol. 84:17-34, 1991) proposed that there are two types of male canines which reflect different allometric scaling patterns of canine crown height against canine occlusal dimensions. They also suggest that proportional canine shape, measured as canine crown height (or projection) relative to the occlusal dimensions, provides an estimate of the canine's function as a weapon, though they provide no test of this hypothesis. This analysis critically examines the claim that there are two types of male canines among anthropoids. It then tests the hypothesis that relative male canine size (measured against body weight) and proportional canine shape are related to estimates of intermale competition, diet, and substrate (used as a surrogate measure of predation pressure). While there is strong taxonomic variation in canine size and shape among male anthropoids, no evidence is found for two discrete canine types. Rather, within families and subfamilies, canine dimensions scale isometrically against body weight and against each other, with a continuum of canine shapes among different taxa. While variation in male canine size is associated with intermale competition and substrate, even when taxonomic variation is controlled, variation in proportional canine shape is not. Neither canine size nor shape are generally associated with variation in diet.