Cohen and Felson argue that structural changes in routine activities affect crime rates by affecting the convergence in space and time of three minimal elements of direct-contact predatory violations: motivated offenders, suitable targets (opportunities for crime), and an absence of capable guardians against a violation. In this study, these variables were measured with respect to the incidence of arson. Criminal motivation was determined by the percent poor, the percent unemployed, the percent black, and the percent youth. Criminal opportunities were measured by percent old housing, percent multiple housing, and the number of commercial/industrial establishments. Data on guardianship were derived from police employment, police expenditure, and female labor force participation. A fourth variable -- ecological niche -- was determined by employment concentration and median house value. Arson rates for the study communities were obtained for 1980. Consistent with the routine activities model, a multiplicative effect of the preconditions for crime on crime rates was also estimated. Motivation, opportunities, and guardianship were found to have direct and/or indirect additive effects on arson rates, thus supporting the 'routine activities' model. A multiplicative effect of the preconditions of crime was also found. 2 figures, 1 table, 44 references.