Part I arrest data for the 12 cities were obtained from the Uniform Crime Reports. Population data for each of the cities were obtained from U.S. censuses for 1960, 1970, and 1980. Arrest rates for specific age-sex-race categories were computed for 1960 and 1980. The 1960 rate was then used with the 1980 population data to compute the expected number of 1980 arrests. This determined how many arrests to expect in 1980 if the arrest rates for specific population categories did not change but the number of people in those categories did. Changing racial composition in the cities produced a substantial portion of the increased number of arrests since 1960. Changes in age composition were important, but in most central cities they apparently resulted from racial composition changes. Changes in the conduct of females also explained some of the arrest increases. These changes, however, were heavily influenced by the racial composition of the cities. The most plausible interpretation of the data is that the limited income of young nonwhite men and women in central cities and their changing family arrangements contribute to the criminalization process. 4 tables and 9 footnotes.