Data from three cross-sectional samples of inmates in the New York City correctional system (N = 299, 236, and 151) were analyzed to determine the prevalence of homelessness among detainees. One-fourth to one-third of each sample had been homeless at some time during the two months before arrest, and 20 percent of the primary sample of 299 inmates had been homeless the night before arrest. Further analysis of the primary sample showed that homelessness was strongly associated with mental illness: 50 percent of those who had ever been homeless during the past three years responded positively to at least one mental illness screening question, compared with 25 percent of the never-homeless inmates. More than a third of the ever-homeless group had received mental health treatment, compared with a fifth of those who had never been homeless.