Early theories explaining the determinants of postmarital residence connected it with the sexual division of labor. However, to date, cross-cultural tests of this hypothesis using worldwide samples have failed to find any significant relationship between these two variables. Our tests show that the female contribution to subsistence does correlate significantly with matrilocal residence in general; however, this correlation is masked by a general polygyny factor. Although an increase in the female contribution to subsistence tends to lead to matrilocal residence, it also tends simultaneously to lead to general non-sororal polygyny, which effectively destroys matrilocality. If this polygyny factor is controlled (e.g., through a multiple regression model), division of labor turns out to be a significant predictor of postmarital residence. Thus, Murdock's hypotheses regarding the relationships between the sexual division of labor and postmarital residence were basically correct, though the actual relationships between those two groups of variables are more complicated than he expected.