In May 1980, Liberty City Miami, a black area, suffered a humanmade disaster: a riot. The effects of the riot on the utilization pattern of the only psychiatric emergency service (PES) for the Liberty City catchment area were studied in two ways: autocorrelation and cross-correlation analyses were used to determine whether PES presentations demonstrated a 7-day cycle and whether this cycle was disturbed by the riot; and analyses of variance (ANOVAs) were used to compare numerically the riot week to the preceding week, the riot week to the following 12 weeks, and the 12 pre-riot weeks to the 12 post-riot weeks. Black PES presentations did not demonstrate a statistically discernable cycle, and none of the ANOVAs were statistically significant. It was concluded that the riot did not have a significant psychiatric morbidity effect as measured by PES presentation data. These results are discussed in reference to previous disaster studies.