Based on correlations between the cranial base angle (CBA) and the index of relative encephalization (IRE, calculated as the cubed root of brain volume divided by basicranial length), several recent studies have identified relative brain size as the factor most responsible for determining basicranial flexion in primates. IRE, however, scales with positive allometry relative to body mass, unlike the negatively allometric relationship between brain volume and body mass. This poses new questions concerning the factors underlying the correlation between IRE and CBA. Specifically, if basicranial flexion represents a spatial solution to the problem of housing a large brain within a neurocranium of limited size, then why is it that the problem is greatest in those species whose brains are smallest relative to body mass? To address this question, the scaling relationships of IRE and the measurements used to calculate it were examined in 87 primate species. It was found that the positive allometry of IRE is due to the fact that its denominator, basicranial length (BL), scales with very strong negative allometry relative to body mass. The scaling relationship of BL may reflect the fact that the noncortical components of the brain (i.e., diencephalon, mesencephalon, medulla) also scale with strong negative allometry relative to body mass, perhaps because of energetic constraints. Importantly, BL and these three brain components scale isometrically against each other. Thus, although cranial base flexion may be an adaptation to accommodate the size of the brain relative to basicranial length, the reason why that adaptation is necessary is not the evolution of a large brain, but rather the evolution of a short cranial base. In so far as basicranial length is affected by the strong negative allometry of the diencephalon, mesencephalon and medulla, the scaling relationships of these brain components are therefore indirectly responsible for the evolution of basicranial flexion. (C) 1999 Academic Press.