Two models of desegregation change between 1968 and 1974 for a number of U.S. urban school districts are tested using a block-recursive technique incorporating the effects of community environment, the school system, and federal influence. The models can explain a considerable amount of change in the North but much less in the South. In both regions, federal intervention is a dominant influence, although for one model in the North, the earlier year level of desegregation is the most powerful effect.