Two levels of analysis were conducted. The first examined a sample of clients of the Los Angeles County Department of Mental Health who were arrested for felony and misdemeanor offenses. The sample consisted of all adult mental health patients admitted as clients during 1978. In a second analysis, only first admissions as mental health clients in 1978 were used. With this cohort, an identifiable date of entry into the mental health system was established. This in turn provided a way of establishing arrest rates during the 24-month period before and the same period after admission. The study population included approximately 285,000 persons who generated over 500,000 admissions to the Department of Mental Health over the 1976-79 period. Study findings suggest that the rates of arrest for a population cohort under treatment in community mental health programs are consistently higher than those for the general population. Other findings regarding arrest rates were a follows: females in the mental health cohort were arrested at higher rates than females in the general population; arrest rates fell less rapidly with advancing age for the mental cohort than for the general population; and arrest rates for the Hispanic and black members of the cohort were substantially lower than for these groups in the general population, but arrest rates for Anglo clients were higher than for the general Anglo population. Despite higher arrest rates, the mental health cohort contributes a very small increment to the total volume of arrests. Regarding the analysis of patterns of felony arrest dispositions, apparently formal identification as a mental health client has a mitigating effect on final disposition for some types of offenses; however, assessment of dangerousness is a more important determinant of disposition than being a mental health client. Tabular data and five references are provided.