The model takes into account the relative opportunities for a criminal act and the relative efficiency of the police per area of the city. An empirical test on arson and avandalism in Philadelphia, Pa., resulted in the model explaining 58 percent of the variance in the actual pattern of the offenses. The 58 percent result assumed equal awareness of arsonists and vandals of the desirability of specific regions of the city for the commission of the crimes. When the analysis assumed that aggregate familiarity with and accessibility to the city decreases from the central business district, the model's explanation increased to 73 percent of the variance in the actual distribution. The aggregate model describes that pattern that would result if the assumptions about the individual decisionmaking of each offender hold. Still, it cannot be said that the individual decisions about where to commit the crimes caused the aggreagate pattern. The resulting spatial pattern is thus necessary support, but no sufficient proof, of the individual decisionmaking process. What is needed is further evidence of the search behavior of criminals as they select a location for a specific crime. Not only must the model be tested using other crimes representing varying criminal objectives, but data at the individual level are required to show the actual travel behavior of a criminal from his/her residence to the crime location. Tabular and graphic data and seven notes are provided. For related papers are NCJ 74011.