The effect of housing density was examined in three different jails. High levels of density were associated with various negative psychological reactions. Social density (number of people in a housing unit) was a more important predictor of these effects than spatial density (space per person). Housing type did not affect illness complaint rates but some evidence was obtained for elevated blood pressure of inmates in highly dense housing when they were confined for large parts of the day. Tables, 5 notes, and 25 references are provided. (Author abstract modified)