The characteristics of rabies epizootics among raccoons were investigated in 11 eastern states along a North-South gradient from New York to North Carolina. Epizootics were defined as discrete intervals of time of at least 5 months in duration, when reported cases of raccoon rabies from an individual county exceeded the median value of raccoon rabies cases reported by that county over the entire period rabies was present among raccoons in the county. Over the approximately 20-year study period, 35,000 cases of raccoon rabies were reported, and epizootics were detected from 251 (64.4%) of 390 counties. The median annual incidence was 0.14 epizootics per year. During the first defined epizootic in a county, the median total number of raccoons reported rabid was 47, with a median monthly incidence of rabies in raccoons of 3.1. The median lag time from the first report of a rabid raccoon in a county to the beginning of the first epizootic was 4 months. Significant differences in the annual incidence of epizootics and monthly incidence of rabid raccoons during epizootics were observed among different states. Although human population density and per capita health spending within counties were positively associated with increasing magnitude of epizootics, a significant difference in the characteristics of rabies epizootics in northern and southern states was apparent. We hypothesize that environmental conditions and perhaps human influence resulted in rabies epizootics in southern states that were smaller, less-frequent, and lacking in well-defined temporal structure compared with those in northern states.