An archival study of 110 inmates in a maximum-security provincial institution. Found that more than 39% of inmates were transient, with no fixed address upon release. Significant differences were found between homeless and non-homeless inmates in terms of: institutional behaviour, occurrence of psychopathology including significantly greater observed symptoms of major mental illness among the homeless, use of medical services, prior involvement with mental health services, prior criminal history, as well as incidence of violence and parasuicidal behaviour. No significant differences were found in terms of age, level of education, or length of sentence. Methodological limitations, the need for an increased role of conununity-based programs to deal wilh homeless intnates, and suggested direction for future research are discussed. Of 110 inmates in an institution established that over 39% were transient, with no fixed address upon release. Significant differences were found between homeless and non-homeless inmates in terms of: institutional behaviour, occurrence of psychopathology including significantly greater observed symptoms of major mental illness among the homeless, use of medical services, prior involvement with mental health services, prior criminal history, as well as incidence of violence and parasuicidal behaviour. (Original abstract-amended)