A total of 115 male residents, 14 to 18 years old, were assigned to 2 dormitories of a large correctional institution in which density levels were systematically varied over 6-week periods. The design allowed exploration of social and spatial aspects of density. The effects of high and low density were assessed over a 6-month period using Moos' Correctional Institutions Environment Scale, a sociometric measure of the dorm, ratings by teachers, and biweekly measures of blood pressure. The multivariate analysis of covariance results indicated that there were significant effects of social density, spatial density, and the interaction of spatial and social density. No effects for blood pressure were found. The results were congruent with the control model of Baum, Harpin, and Valins and with Altman's stimulus reduction model. Tables and about 30 references are provided. (Author abstract modified)