The position of the lunate sulcus in fossil endocasts (when it can be determined) may serve as a potential marker of cognitive development in extinct hominid species. While the lunate sulcus is reliably present in the brains of great apes and forms the anterolateral boundary of the primary visual cortex, in humans its presentation is much more variable, and even if present, it does not correspond to a functional region. Grafton Elliot Smith, who named the lunate sulcus, claimed that it was homologous in humans and the great apes. Using high-resolution MRI, we assessed the presence/absence and course of the lunate sulcus in 110 adult subjects. We found that in the vast majority of cases, lunate sulci identified on the surface of the occipital lobe are actually composed of smaller sulcal segments that converge into an apparently continuous composite lunate sulcus. We found only 3 examples in 220 hemispheres (1.4%) of continuous lunate sulci that resembled ape lunates in form (albeit in a more posterior position). Composite lunate sulci were found in 32.7% of left hemispheres and 26.4% of right hemispheres. These results, combined with those from histological and functional imaging studies, indicate that human and ape lunate sulci are not homologous structures. We suggest that the extent of functional reorganization of the occipital region during hominid evolution has been underestimated, and that changes in this region were not just passively shaped by expansion of parietal association cortex.