Conventional wisdom about police culture rests on descriptions of a single occupational phenomenon in which the attitudes, values, and norms of members are homogenous. As police forces continue to diversify and as community policing becomes part of the philosophy of policing, it should be expected that there would be more cultural variation. Because there is a foundation for expecting and accounting for differences in culture related to organizations, rank, and individual style, research should incorporate these different levels in theories and empirical investigations of police culture. Police culture is best understood as a filtering process that is mediated through the organization and within by rank and style. Police officers may still view crime fighting as their core role orientation, but the extent to which officers and citizens are isolated from one another has waned in the community era. Differences in culture across different department (large, small, bureaucratic) of varying locales and clientele are expected. Larger bureaucratic organizations usually have more administrative policies in an attempt to control officer discretion. Patrol divisions are where most accounts of police culture focus. Socialization forces are the most powerful at the patrol level. Patrol is where officers develop their “style” of policing. Researchers should focus on gathering baseline data on officers in assessing potential individual changes in cultural identification and commitment over time. Loyalty is still a major outcome of culture. Even with changes in terms of police diversity and philosophy, there are some boundaries to fragmentation. The forces of fragmentation should be working to erode a single monolithic police culture, but not totally dissolve it. Research should establish a threshold of officer agreement in assessing cultural and subcultural attitudes, values, and norms. Differences in officers’ perceptions of citizens and supervisors might be directly related to where and when they work. Studies of police culture should employ a variety of methodological approaches, as well as utilize longitudinal data collection designs. 2 figures, 17 notes, 84 references