In this cross-sectional study, 96 randomly selected female detainees were interviewed by one of two trained interviewers, using the Structured Clinical Interview for DSM (SCID)-IV, the Addiction Severity Index, the Global Assessment of Functioning Scale (GAF), and the Medical Outcomes Study 36-Item Short-Form Health Survey. Criminal offense categories were derived from the FBI's Uniform Crime Reporting Program. Forty-eight percent of the detainees had been charged with committing at least one Part I crime; 27 percent were charged with committing a violent crime against persons. Functional impairment was the best indicator of arrest for more serious or violent crimes. According to the multivariate analysis, women who had functioned more poorly in the month preceding incarceration, as indicated by a lower GAF score, had a higher likelihood of being arrested for more serious crimes. Previous conviction for a serious crime was related to current arrest for a serious crime, although not specifically for a violent offense. Clinical, criminal, and sociodemographic factors were otherwise insignificant in distinguishing the subgroup charged with more serious offenses. Clinicians must consider the potential for criminal behavior, especially in mentally ill persons with a history of arrest and conviction and in those with more severe functional impairment, regardless of the presence of psychosis. Mental health services in the community, at least equal to those available to men, must be made available for these women. 3 tables and 53 references