Using data on poliomyelitis and typhoid fever mortality in the United States, 1914–69, we test competing theories for the twentieth century expansion of polio. We analyze data stratified by age, sex, and race. We show that some of the seemingly-paradoxical aspects of the data — principally, that whites had higher polio death rates than nonwhites but lower typhoid death rates — are consistent with the polio hygiene hypothesis. Data on racial differences show that the hygiene hypothesis is necessary and sufficient to explain patterns of polio mortality in the United States. Epidemiological phenomena are best understood in their social context.