Impact of analytic protocols on archaeofish abundance, richness, and similarity: a Caribbean-Pacific crossover study

Standing in notable contrast to the practice of other regions, past archaeological fish studies in the tropical Pacific and New Zealand have commonly restricted analysis to a set of five paired cranial elements and “special bones”. This method has been critiqued for reducing taxonomic richness and altering relative abundances, with consequences for fishing, dietary, and paleoecological reconstructions. Here, I assess whether these issues hold true for a Caribbean fish assemblage from the prehistoric site of Sabazan (AD 400-1400), Carriacou, Grenada. Analytic results obtained using the restricted element method are compared to those based on analysis of all potentially diagnostic bones. Building on previous critiques, I investigate the significance of observed discrepancies in relation to sample size effects. Tests of statistical significance and a similarity index are employed to evaluate and quantify the impact of the restricted analytic method for different screen-size fractions (6.4 and 1.6 mm) and quantitative units (NISP and MNI). Results show varying effects in the interaction of these variables, but indicate overall that the restricted element method does not provide taxonomic richness and abundance information comparable to that of an expanded analytic approach. I consider the larger implications of these findings for the identification of certain taxa, such as tunas (Scombridae), herrings (Clupeidae), and jacks and scads (Carangidae), the detection of net fishing, and the exploitation of pelagic fisheries. Despite the unreliability of this method, I conclude that there is a role for expedient methods, provided these are carefully evaluated prior to implementation