public use sample of the 1970 U.S. Census. The two theoretical predictions are: (1) a group's relative size is inversely related to the proportion of its members who are outmarried; and (2) an SMSA's heterogeneity is directly related to the rate of intermarriage in it. The underlying assumption is that the structural constraints of size distributions affect marriage notwithstanding cultural values promoting ingroup marriages. The data confirm the two predictions (corroborating the underlying assumption) for most size differences and most forms of heterogeneity examined. Thus, heterogeneity in national origins, mother tongue, birth region, industry, and occupation raise intermarriage rates in these respects. Although racial heterogeneity does not have this predicted effect, the reason is that the great socioeconomic differences between races consolidate racial boundaries and thereby counteract the influence of heterogeneity on intermarriage. Empirical evidence supports this explanation: when racial income differences are controlled, the predicted positive relationship between racial heterogeneity and intermarriage becomes apparent.