Over 40 years ago, Clifford Jolly noted different ways in which Hadropithecus stenognathus converged in its craniodental anatomy with basal hominins and with geladas. The Malagasy subfossil lemur Hadropithecus departs from its sister taxon, Archaeolemur, in that it displays comparatively large molars, reduced incisors and canines, a shortened rostrum, and thickened mandibular corpus. Its molars, however, look nothing like those of basal hominins; rather, they much more closely resemble molars of grazers such as Theropithecus. A number of tools have been used to interpret these traits, including dental microwear and texture analysis, molar internal and external morphology, and finite element analysis of crania. These tools, however, have failed to provide support for a simple dietary interpretation; whereas there is some consistency in the inferences they support, dietary inferences (e.g., that it was graminivorous, or that it specialized on hard objects) have been downright contradictory. Cranial shape may correlate poorly with diet. But a fundamental question remains unresolved: why do the various cranial and dental convergences exemplified by Hadropithecus, basal hominins, and Theropithecus exist? In this paper we review prior hypotheses regarding the diet of Hadropithecus. We then use stable carbon and nitrogen isotope data to elucidate this species' diet, summarizing earlier stable isotope analyses and presenting new data for lemurs from the central highlands of Madagascar, where Hadropithecus exhibits an isotopic signature strikingly different from that seen in other parts of the island. We offer a dietary explanation for these differences. Hadropithecus likely specialized neither on grasses nor hard objects; its staples were probably the succulent leaves of CAM plants. Nevertheless, aspects of prior hypotheses regarding the ecological significance of its morphology can be supported. Am. J. Primatol. 78:1098-1112, 2016. © 2015 Wiley Periodicals, Inc.