Theories of racial and ethnic relations have been plentiful, but the empirical testing of hypotheses has not led to a cumulative growth of knowledge. As yet, no strong paradigm of research has emerged. The growth of empirical studies of racial/ethnic inequality in the United States over the last decade suggests that formal models of the process of stratification may be a significant turning point in the development of the field. Not only do such models offer a rigorous statistical method of analysis, but a cumulative research tradition seems to be developing. The expansion of models to include social origins and the temporal ordering of the socioeconomic career required specification and interpretation of how ethnic inequality is generated and transmitted across and within generations. The two major frontiers of research on ethnic stratification within cumulative research strategy are trend and compartive analysis. Trend analysis speaks to the question of whether there is a growing convergence in ethnic stratification, while comparative analysis specifies certain attributes of the social structure of geographical areas which may affect the levels and processes of ethnic stratification. (Author/WI)