Skull 5 from Dmanisi: Descriptive anatomy, comparative studies, and evolutionary significance

A fifth hominin skull (cranium D4500 and mandible D2600) from Dmanisi is massively constructed, with a large face and a very small brain. Traits documented for the first time in a basal member of the Homo clade include the uniquely low ratio of endocranial volume to basicranial width, reduced vertex height, angular vault profile, smooth nasal sill coupled with a long and sloping maxillary clivus, elongated palate, and tall mandibular corpus. The convex clivus and receding symphysis of skull 5 produce a muzzle-like form similar to that of Australopithecus afarensis. While the Dmanisi cranium is very robust, differing from OH 13, OH 24, and KNM-ER 1813, it resembles Homo habilis specimens in the "squared off" outline of its maxilla in facial view, maxillary sulcus, rounded and receding zygomatic arch, and flexed zygomaticoalveolar pillar. These characters distinguish early Homo from species of Australopithecus and Paranthropus. Skull 5 is unlike Homo rudolfensis cranium KNM-ER 1470. Although it appears generally primitive, skull 5 possesses a bar-like supraorbital torus, elongated temporal squama, occipital transverse torus, and petrotympanic traits considered to be derived for Homo erectus. As a group, the Dmanisi crania and mandibles display substantial anatomical and metric variation. A key question is whether the fossils document age-related growth and sex dimorphism within a single population, or whether two (or more) distinct taxa may be present at the site. We use the coefficient of variation to compare Dmanisi with Paranthropus boisei, H. erectus, and recent Homo sapiens, finding few signals that the Dmanisi sample is excessively variable in comparison to these reference taxa. Using cranial measurements and principal components analysis, we explore the proposal that the Dmanisi skulls can be grouped within a regionally diverse hypodigm for H. erectus. Our results provide only weak support for this hypothesis. Finally, we consider all available morphological and paleobiological evidence in an attempt to clarify the phyletic relationship of Dmanisi to Homo species evolving >2.0 to 1.0 Ma.