Recent longitudinal analyses of total police force size, both across and within jurisdictions, raise questions about the explanatory power of rational-choice and conflict theories of crime control. Rational-choice theory assumes that broad-based consensual demands for policing services are primarily determinants of government spending and manpower levels. Conflict theories assume that society is composed of various groups and strata that seek to promote their particular interests and objectives. Generally, the previous research suggests that current levels of police-force strength are best explained by previous manpower levels and are relatively unresponsive to changes in the social and political structure of macro social units. The current study examined the macro social determinants of police-force size in Milwaukee, Wis. The data are annual, covering the years 1934 to 1987. The analyses of yearly changes in the size of total, patrol, detective, and civilian employment cannot sustain the supposition that municipal authorities reduce the size of some units to increase the size of others. Although the results from the first-difference equations reveal that the effects of short-term changes in the social structure vary across dimensions of police-force strength, the supplementary ARIMA analyses show that changes in the size of the patrol, detective, and civilian units are unrelated to each other. Thus, although changes in the structural conditions produce small but nontrivial yearly changes in the size of the specific units, each unit apparently responds independently to changes in the social environment. The findings suggest that police administrators may have more discretion to make adjustments to specific units than to total manpower levels. The implications of these results for rational-choice and conflict theories are discussed. 2 tables and 42 references