Social cohesion theory is tested using data on ethnic intermarriage in former Yugoslavia. Before the disintegration of Yugoslavia, the proportion of marriages outside the own ethnic group was generally low, but in this respect large differences among the groups existed. The proportion of mixed marriages with a Serbian partner was much higher among the Montenegrins and Hungarians than among the Muslims, Slovenes, or Albanians. The findings are largely in line with the predictions of social cohesion theory that intermarriage reduces the probability of violent conflict among social groups. Besides proportions of mixed marriages, loglinear parameters are presented. These parameters show that percentages not always give a good indication of the social distances among groups. The boundaries of the largest ethnic groups, the Serbians and Croatians, were less closed than their rather low intermarriage rates suggested. The social distance between the Hungarians and the Serbians, on the other hand, was larger than expected on the basis of their intermarriage rates. The findings stress the importance of including information on ethnically mixed marriages into models of ethnological monitoring and early warning systems for ethnic conflicts. (Contains 2 tables and 4 figures.)