Historically speaking divorce is a relatively recent phenomenon. Divorce has become more important over the last century and has undoubtedly had an impact on the social reproduction strategies typically linked to marriage. In this paper use is made of the concept of Assortative Mating to explore if initial and subsequent marriage partner choices differ on the basis of heterogamy. This research question is formulated in terms of two contrasting hypotheses: (1) the learning and (2) the marriage market hypotheses (Gelissen 2004). These two perspectives form the basis of the theoretical framework used in the empirical analyses reported. Here logistic and linear regression and log-linear modelling techniques are used. The data used to test the learning and marriage market hypotheses contains information on all officially recorded marriages in the Czech Republic over a ten year period (1994-2004) gathered by the Czech Statistical Office. The results presented reveal that divorce changes the socially reproductive aspects of marriage choice strategies, and these changes vary systematically on the basis of gender. Whereas first and subsequent marriage choices are very similar for men, repeated marriage choices for women are on the whole more diverse. A number of explanations explaining this gender based difference are presented. These explanations centre on dissimilarities in the context of repeated choices for men and women.