Practice makes perfect? Inter‐analyst variation in the identification of fish remains from archaeological sites

Identification of faunal specimens based on a morphological comparison with known-identity reference specimens is the standard methodology used in zooarchaeological analysis. However, the accuracy of identifications is rarely considered. In this paper, we report results of an experiment in which 13 analysts were asked to identify 50 fish skeletal elements from a reference collection and 50 fish skeletal elements from an archaeological collection in southern Ontario. The type and level of experience of the analysts and the amount of time they invested in the identification were controlled. The archaeological specimens were subsequently identified taxonomically using ZooMS. Our findings demonstrate that taxonomic and element identifications are far from perfect, both in the reference collection set and in the archaeological collection set. Probable contributing factors include the richness of taxonomic groups; distinctiveness of skeletal morphology; experience level of the analyst; and size of the individual specimens and whether the analyst had access to comprehensive, well-labeled reference collections. We recommend emphasis be placed in training on the importance, for most species, of not making a taxonomic identification unless the element identification is certain; conservatism in identification of species in groups with many members; clear knowledge of the range of species possible within a region; and active involvement by the instructor or mentor to ensure that neophyte analysts are corrected