The DSU population for the study consisted of all inmates classified DSU and residing in the DSU on January 1, 1983, in addition to those inmates classified for DSU placement between January 1, 1983, and December 31, 1983. Variables address inmate personal background data, criminal history, present offense, incarceration information, the DSU offense, and information surrounding DSU classification and the resultant segregation placement. Data were obtained from the departmental computerized data base, inmate master cards, central office inmate folders, and the classification and movement files at Walpole. All information was analyzed using the Regents Computer Network. More than two-thirds of the DSU inmates were white, and most were single and in their mid-twenties when they received their DSU commitment. Most had sporadic employment and did not graduate from high school. Most were incarcerated for person offenses, and just over one-tenth had been placed in protective custody prior to their DSU classification; over half (51 percent) had appeared previously before the DSU board. When the DSU offense occurred, almost all of the inmates were at Walpole. Staff assault, riot participation, inmate assaults, and weapons possession were the most prevalent DSU offenses. Two-thirds of the inmates had been released from DSU status by July 1, 1984, and the average time in segregation was July 1, 1984, and the average time in segregation was 8.6 months. The placements given inmates immediately after DSU release are reported. 40 tables and 10 footnotes.