Comparability of fossil data and its significance for the interpretation of hominin environments

Unraveling the context in which the evolution and diversification of early hominins occurred has become one of the core and highly debated subjects in paleoanthropology. Over the past three decades substantial progress has been made due to the proliferation of fieldwork and a consequently expanding fossil record, and development of new methods of analysis. The present study uses data of fossil mammals from the Shungura Formation of Ethiopia, with specimens collected semi-independently by French and American research teams who worked in the southern and northern parts of the Shungura area respectively. We compare these two samples in terms of collection methods, taxonomy, taphonomy, and local environmental differences. The following results were obtained: (1) No major taphonomic differences were observed between the two collections. The effect of a major taphonomic shift that occurred in the middle of Member G (G-13) is observed in both samples and is caused by the important change in the depositional environment from fluvial to lacustrine conditions. (2) The French team collected more specimens than the American team, in part because it had a larger area of exposures, and it spent two extra seasons in the field. Additionally, the French team collected more large-sized taxa including their postcranial elements, while the American team recovered a restricted set of postcranial bones. In contrast, the American team collected more primates and carnivores than the French team. (3) Despite these differences, comparable taxonomic composition and number of species are observed in both collections. (4) A study of changes in relative abundance in bovid tribes indicates that similar patterns of variation through time are observed in both samples. This is considered to be evidence for the prevalence of generally similar habitats (and habitat change through time) in the north and south of the Shungura area. (5) However, habitat differences may have occurred locally, as inferred by differences in taxonomic abundances at the species level. For example, the bovid Menelikia lyrocera was more common in the southern parts of the Shungura exposures, while Kobus sigmoidalis was more common in the north. (6) Finally, the present study underscores the importance of the quality of data in unraveling past environments and patterns of faunal changes through time. Well-controlled and standardized collecting methods and systematic documentation procedures are critical for future fieldwork activities. This will improve the quality of our data, facilitate comparisons across regions, and lead to more robust hypotheses.