The empirical implementation of theoretical contributions in the social sciences usually rests on variables defined and measured by persons other than the theorist. Such is the case with the theory of marriage. For the United States, the only relevant demographic quantity measured with accuracy is two-person, male-female legal marriage. Thus, my decision to study legal marriage was imposed by the Bureau of the Census and the inherent difficulties of measuring other forms of communal living. This does not seem to be restrictive for the United States, where consensual marriage is often a prelude to legal marriage, but it may be important in countries such as those in Latin America, where legal and other social institutions (i.e., divorce laws) differ.1 That a model of marriage based on rational choice has substantial explanatory power is demonstrated in this study. Three factors-the ratio of the sexes, the potential returns to marriage, and the cost of divorceexplain many of the areal differences in marital behavior in the United States. It is also evident that the marriage market has important indirect effects on the reproduction rate.