Beware the Springbok in Sheep’s Clothing: How Secure Are the Faunal Identifications upon Which We Build Our Models?

Abstract

Lively debate surrounds the introduction of non-indigenous domestic livestock to southern Africa. Despite disagreements regarding process, the archaeological community agrees, with unusual unanimity, on the broad chronology. Indeed, the certainty with which the timing is known (admittedly within the limits of radiocarbon dating) has been celebrated, because with these underpinning data in hand, issues of process can be explored in a serious and empirically grounded manner. Recently published ancient DNA (aDNA) research in southern Africa now calls into question the reliability of many faunal identifications upon which this debate rests. These data build on earlier ecological data, suggesting that some faunal identifications at sites crucial to the debate may be unreliable. A number of morphologically identified domesticate bones were chosen for aDNA sequencing to explore the relationships among southern Africa’s early domestic stock. Unfortunately, a large proportion yielded DNA sequences indicating a wild origin. This led us to consider the potential scale of the problem and the implications for existing models regarding the introduction of herding to the subcontinent. The issue may originate largely from the optimistic identification of specimens retaining too few key morphological markers. We acknowledge that reconstructions of the past are likely to be biased by discarding potential zooarchaeological data through overly conservative identification. We argue, however, that the potential ramifications of building models on unreliable data are far greater than those of being forced to acknowledge gaps in our data and are calling for further research.