Abstract
Observations of the holotype of Syndyoceras cooki and comparative material indicate that protoceratid basicrania share no striking resemblances to those of camelids, a group purported to be closely related to the protoceratids. The protoceratid petrosal has a massive posterodorsal (canalicular) part with a strong dorsal crest separating the cerebral and cerebellar surfaces, like the petrosals of ruminants, and its endocranial surface lacks a fossa mastoidea, a feature of uncertain polarity found in camelid petrosals dating back at least as far as the Oligocene genus Poebrotherium. Furthermore, the subarcuate fossa of the petrosal is reduced in protoceratids (a derived condition found in pecoran ruminants). These results, while they cannot be analyzed completely with currently available data, indicate: 1) a pressing need for the reappraisal of interrelationships of "tylopod" artiodactyl taxa and their relationships with the ruminants, and 2) the potential that protoceratids themselves may be more closely related to ruminants than previously thought, hearkening back to conclusions originally reached by Scott in the late 19th century. Comprehensive study of all Eurasian and North American neoselenodont artiodactyl taxa with preserved basicrania will be necessary to clarify these issues.