For the purposes of the survey, a prison riot was defined as an incident involving 15 or more inmates and resulting in damage to property and/or physical injury. Data were collected from January 1, 1984, to August 31, 1984. All States provided information but Delaware, Massachusetts, and Vermont. Twelve of the responding States had no riots between 1971 and 1983. For each riot, the data solicited were date, location, hostages taken, injuries to inmates and guards, property damages, riot duration, and how the riot ended. Based on the data obtained, this report concludes that the costs of riots are likely to increase. The causes of prison riots are better understood and revolve around identifiable ills found in most prison environments. The techniques of riot prevention -- grievance procedures, inmate advisory councils, and attitudinal surveys -- are likely to gain acceptance as their potential benefits are recognized. Appended letters accompanying the survey form, data from a similar Canadian survey, and 10 references.