The hypothesis that indications of change in Marquesan faunal assemblages reflect changes in prehistoric subsistence practices is challenged through reanalyses of the identified faunal remains from the Hane Dune site. An alternative hypothesis is proposed: that the supposed indications of change actually reflect intersite spatial and functional variability. Using bootstrap techniques to estimate the standard error of Spearman's rank correlation coefficient, reanalyses of the Hane Dune site assemblages are shown to be flawed by a failure to consider the effects of small sample sizes. The hypothesis that indications of change reflect intersite spatial and functional variability is weakened by the results of recent excavations. Recently reported evidence for change in Marquesan faunal and artifact assemblages supports the inference that Marquesan subsistence practices changed markedly over the course of prehistory.