The Demography of Local Marriage Markets" is a completely revised version of a proposal first submitted to the study section in January 1989. The main objective is to examine the effects of spatially-based marriage-market conditions on male and female patterns of first marriage transitions and assortative mating in the United States. A guiding hypothesis is that local marriage-market opportunities and constraints affect (1) the extent and timing of marriage; and (2) the choice of marital partners. The choice of a geographic scale most appropriate to the analysis of local marriage markets is empirical. Consequently, the first objective is to evaluate several alternative definitions of marriage-market areas. We define these areas using labor markets identified in the 1980 PUMS-D file, and counties and multicounty units available in the 1988 Area Resource File. Indicators of marriage-market conditions such as local economic opportunities and spouse availability are then included in ecological models of local-area proportions ever-married. This analysis addresses a basic limitation of previous research, which is the implicit assumption that marriage markets are national rather than local in scope. The second objective is to fit semi-parametric hazards models of male and female transitions to first marriage. Specifically, local marriage-market indicators are merged for the first time with individual marital history data from the 1979-1988 waves of the National Longitudinal Survey of Youth (NLSY). These newly merged data provide a unique opportunity to build true "contextual" models of marital status transitions. To our knowledge, no other study has included local marriage-market indicators in behavioral models of the transition to first marriage. Finally, a third objective is to examine the effects of local marriage-market conditions on patterns of assortative mating. Marriage-market indicators are linked to married-couple records from the PUMS-D in order to evaluate their effects on patterns of in- and out-group marriage. We discuss various log-linear models that are especially appropriate to identifying patterns of association in cross-classifications of husbands' and wives' social characteristics. We also describe it "competing-risk" hazards models of marital homogamy or heterogamy, using the linked NSLY/PUMS-D database.