M.Phil. (Anthropology) / The primary aim of this study was to undertake the description and analysis of the manual and pedal hominin phalanges from Drimolen, South Africa. The specimens were contextualized by examining their counterparts in a series of extant and extinct specimens. As Drimolen preserves the remains of both Paranthropus robustus and early Homo, it provides an opportunity to contribute to our understanding of morphological variation and differences between these taxa, as well as providing an important first step in conceptualizing their evolutionary history and palaeobiology. The first section focused on the manual and pedal phalanges, while the second focused exclusively on the first metatarsal. Numerous studies have analyzed Plio-Pleistocene manual and pedal phalanges to evaluate tool-related behaviours and locomotion in early hominins. In this dissertation nine phalanges were described and analyzed for the first time. The manual specimens consist of one proximal (DNH 9), two proximal pollical (DNH 116 and DNH 119), four intermediates (DNH 48, DNH 63, DNH 65 and DNH 66) and one distal pollical phalanx (DNH 13). There is currently only one pedal proximal phalanx (DNH 117). As at Swartkrans, assignment of postcranial elements was hampered by the fact that two hominin species have been found at the site, namely Paranthropus robustus and early Homo. To analyze the specimens, multivariate statistics (PCA and CVA) were conducted on the specimens using raw and size-adjusted data. This was done within the context of a large extinct and extant dataset to compare size and shape data. In general, PCA and CVA yielded similar results but only the raw data was useful in discriminating between taxa. There was poor discrimination in the analyses based upon size-adjusted data because of the huge degree of overlap between fossil and extant taxa. Consequently, results were not diagnostic and it wasn’t possible to assign specimens confidently to either Paranthropus or Homo. This was exacerbated by the isolated nature of the remains, the presence of two hominin taxa and taxonomic uncertainties of several Swartkrans phalanges. Despite these difficulties, these specimens do provide insight into the degree of morphological variation present in the phalanges from the Plio-Pleistocene of South Africa. The second section described the fragmentary first metatarsal and quantified it by using laser scanning and calculating the curvature of the proximal articular surface. Univariate statistics were then calculated in order to examine the degree of curvature between each of the extant taxa and whether these differences were statistically significant. Finally, a simple bivariate plot was generated using the curvature results which showed that this joint surface was most curved in chimpanzees and western gorillas, moderately curved in eastern gorillas, and flattened in modern humans and baboons. Highly curved joint surfaces in this region suggest a more mobile hallux, whereas flattened surfaces suggest reduced joint mobility. The fossil specimens SK 1813, STW 573, STW 595 and OH-8 all grouped closely with modern humans; A.L. 333-54 and SKX 5017 grouped with chimpanzees, while DNH 115 grouped with eastern gorillas. Overall, these results indicate that SKX 5017 was the most curved of the South African specimens and is attributed to Paranthropus (Susman, 1988). It grouped closely with DNH 115 which suggested that climbing activities may have continued to play an important role in the behavioral repertoire of these individuals during the Plio-Pleistocene.