A proper regression analysis of the data in Table II in Welsch (1996) adds strong support to the conclusion contained in Moore and Romney (1994, 1995) and Roberts, Moore, and Romney (1995) that distance and language both contribute about equally to the prediction of village assemblage similarity. We also show that the new data presented by Welsch on the identity of the Collectors contains no detectable bias with respect to kinds of artifacts collected and reinforces our original finding that both distance and language contribute to the explanation of village assemblage similarity. In addition we present a new analysis employing principal components that demonstrates that the original artifact by site frequency data are a better point of departure for the analysis of the relative effect of distance and language than the same data coded in binary presence/absence form as suggested by Welsch.