The analysis was informed by routine activity theory, which views crime as a function of three main elements: motivated offenders, suitable targets, and absence of guardians. Based on this approach and on the differences between homicide and robbery with regard to motivation, victim-offender relationships, and planning, the study hypothesized that robbery would show a seasonal trend and would peak during the winter (November through March) due to the increase in the cost of living and the facilitating environmental conditions during these months. In contrast, the study expected homicide to be most prevalent in August, when social interaction is at its highest level, and that it would not follow a distinct seasonal pattern, because victim availability and suitability, as well as the variety of motives involved in homicide, make this crime much less dependent on climatic conditions. The data were analyzed with three statistical methods: a stochastic model (SARIMA), a X-11 seasonal adjustment, and a rank order of the three highest months. The results generally support the hypotheses, thus indicating the utility of the routine activity theory as a useful framework for analyzing the seasonality of crime. The methodological implications of using different definitions and measures of seasonality are discussed, and suggestions are offered for further study in this field. 6 tables, 2 notes, and 50 references