Perspectives The Bipedalism of the Dmanisi Hominins: Pigeon-Toed Early Homo? Ian J. Wallace, 1 * Brigitte Demes, 2 William L. Jungers, 2 Martin Alvero, 3 and Anne Su 1 1 Interdepartmental Doctoral Program in Anthropological Sciences, Stony Brook University, Stony Brook, NY 11794 2 Department of Anatomical Sciences, Stony Brook University, Stony Brook, NY 11794 3 Department of Anthropology, Stony Brook University, Stony Brook, NY 11794 KEY WORDS talar neck angle; tibial torsion; in-toeing ABSTRACT In the recent description of the hominin postcranial material from Dmanisi, Georgia, Lordkipanidze and colleagues (Lordkipanidze et al. [2007] Nature 449: 305–310) claim that the Dmanisi hominins walked with more medially oriented feet than do modern humans. They draw this functional inference from two postcranial features: a wide talar neck angle and a slight medial tor- sion of the tibia. However, we believe that the data pro- vided by the authors fail to support their conclusions. Talar neck angle and tibial torsion values from the Dma- nisi specimens fall comfortably within the range of modern human variation. We further submit that foot orientation cannot be reliably deduced from the tibia and talus alone. Am J Phys Anthropol 136:375–378, 2008. V V C 2008 Wiley-Liss, Inc. Lordkipanidze et al. (2007) recently described new and important hominin postcranial material from Dmanisi, Georgia, one of the oldest sites documenting the pres- ence of the genus Homo outside of Africa. In their dis- cussion of the elements of the lower extremity, the authors suggest that the Dmanisi hominins walked with a different foot orientation than do modern humans, implying that they walked toed-in or pigeon-toed. They draw their functional inferences about this novel form of bipedalism from the co-occurrence of two postcranial fea- tures: a wide talar neck angle and a slight medial tor- sion of the tibia. They argue that ‘‘this combination results in a more medial orientation of the foot, and more equal load distribution on all rays than in modern humans’’ (Lordkipanidze et al., 2007:309). Such a configuration is significant and surprising be- cause it is potentially maladaptive. In children, in-toeing has been demonstrated to shift the body’s center of grav- ity to the lateral aspect of the foot causing impaired ambulation, such as an inability to run without tripping (McCrea, 1985; Donatelli, 1990). Likewise, the recon- struction of a medially oriented foot among the Dmanisi hominins would seem to run counter to the consensus characterization of Homo erectus (including H. ergaster) hominins as accomplished and efficient bipeds (Rose, 1984; Walker, 1993; Spoor et al., 1994; Bramble and Lieberman, 2004; Ruff, 2008). We believe that the data provided by Lordkipanidze et al. (2007) fail to support the authors’ conclusions in this regard, and we submit that foot orientation cannot be reliably deduced from these bony elements alone. In the morphological description of the Dmanisi talus D4110 and tibia D3901, both the talar neck angle and tibial torsion are stated to be similar to those of modern humans. However, rather curiously and in apparent con- tradiction, in their discussion of the functional morphol- ogy of the lower extremity, the talar neck angle and tib- ial torsion are interpreted in combination as indicative of a more medial foot orientation than is found in mod- ern humans. Some clarification of these data would seem to be in order. First, the talar neck angle is asserted to be wide. The talar neck angle is the angle formed by the longitudinal axes of the trochlea and the neck, and is indicative of how much the neck and head of the talus deviate medi- ally, away from the axis of flexion and extension of the talocrural joint. On the basis of the comparative samples presented by Lordkipanidze et al. (2007), the neck angle of the Dmanisi talus (268) falls within the upper range of modern human variation. If one casts a broader net, however, it is apparent that the Dmanisi angle is very similar to the published means of several prehistoric hominin samples (Table 1). Mean talar neck angles from Neandertals, Middle Paleolithic H. sapiens, and Predy- nastic Keneh are virtually identical to Dmanisi; Early Upper Paleolithic H. sapiens, Late Woodland Libben Native Americans, Medieval Serbians, and Bronze Age Palestinians are also very similar. The talar neck angles from some, but not all, modern and historic samples tend to be slightly smaller than those of the prehistoric samples, and consequently are more similar to the mean of the comparative sample used by Lordkipanidze et al. (19.48; Table 1). One speculative explanation of these dif- ferences between prehistoric and historic/recent samples *Correspondence to: Ian J. Wallace, Interdepartmental Doctoral Program in Anthropological Sciences, Stony Brook University, Stony Brook, NY 11794. E-mail: iwallace@ic.sunysb.edu Received 4 January 2008; accepted 22 January 2008 DOI 10.1002/ajpa.20827 Published online 18 March 2008 in Wiley InterScience (www.interscience.wiley.com). V V C 2008 WILEY-LISS, INC. AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 136:375–378 (2008)