Flood research has used a variety of methods, yielding inconsistent findings. Universal definitions of illness are paramount to the science of psychiatric epidemiology of disasters. St. Louis area survivors (N = 162) of the Great Midwestern Floods of 1993 received a structured diagnostic assessment at 4 and 16 months postdisaster, with 88% follow-up. The purpose of the assessment was to examine predisaster and postdisaster rates of disorders and symptoms. Flood-related posttraumatic stress disorder was diagnosed in 22% and 16% at index and follow-up, respectively. Comorbidity with major depression determined whether the posttraumatic stress disorder would have remitted by 1 year later. Nearly one half of the men in the sample had a pre-existing alcohol use disorder. Virtually no new substance abuse followed the floods, and hence, substance abuse did not develop in response to the disaster or as part of coping with its aftermath. Somatization disorder was not observed; new somatoform symptoms represented a fraction of postflood somatic complaints. Findings are inconsistent with causal attribution of floods in the etiology of alcohol abuse and somatization. Methodological differences may account for much of the apparent discrepancy of these findings, with recent reports of increased alcohol use and somatic symptoms observed after other disasters.