The evolution of the brain in mammals has been accompanied by a progressive enlargement of the cerebral cortex. Allometric analysis of the volume, surface and convolutedness of this cortex shows that among mammals two major groups can be distinguished: (1) species with lissencephalic brains, where surface-volume relationships are determined by linear geometric laws, and (2) species with gyrencephalic brains, where strict geometric similarity no longer applies. Different mathematical models are required to describe the brain morphology in these groups. It has been shown, moreover, that among species with convoluted brains, marine mammals (Cetacea) form a subgroup in that the cerebrocortical surface in these animals is more folded than in terrestrial mammals of similar brain size. Thus it appears that the use of a single allometric relation in studying the effect of size increase on the geometry of the brain - irrespective of convolutedness or ecological strategy - as has been done in previous studies, is unjustified and obscures the differences in the structural organization of mammalian brains. Finally, a dimensionless index of cortical folding is proposed, which indicates that the degree of cortical folding depends not only on the volume of the brain or cortex but on cortical thickness as well.