The three explanations ascribe the higher rates of urban crime to (1) the degree of urbanization and populated density, (2) the greater rates of migration and population growth in urban populations, and (3) the differences in demographic structures between urban and rural areas, urban areas having greater proportions of young people. Eight indicators are used to test six hypotheses for six crimes against persons and property, with the 46 Japanese perfectures as units of analysis. Measures are taken to constrain the degree of multicollinearity that exists among the independent variables. While the structural indicators measured at an ecological level are not powerful in explaining inter-perfectural variations in rates of homicide, rape and arson, they are significantly associated with larceny, robbery and assault rates. Multiple regression analysis ascribes greatest weight in explaining the variation in rates of property crime to the higher proportion of young adults in urban populations, and to the higher rates of migration in these populations. The differences between Japanese and Western levels and patterns of crime are attributable to the operation of powerful mechanisms of family and community social control and to methods of policing. Trends comparable with those in the West are emerging through youth crime in the context of urban conditions. (Author abstract)