This article investigates economic bases for individuals' selection of marriage partners. Its central focus is matching of spouse pairs based on unobservable components of hourly earnings. Using a data extract from the Panel Study of Income Dynamics, we have observations of individuals' wages before and immediately after marriage, along with their spouses' earnings after marriage. We exploit the data to estimate the relation between spouses' wage residuals. The evidence supports positive marital matching on the basis of earnings. The results are robust with respect to variations in the statistical framework of the model.