Many small jails have no uniform procedure for mental illness detection. This study evaluated the use and validity of an existing screening procedure administered by booking officers at a small Midwestern jail. These results were compared to the reliability, validity, and use of the Referral Decision Scale when administered by clinicians in the same facility. Participants were 95 male inmates selected from all new admissions to a county jail during a 1-month period. The Prisoner Intake Screening Procedure (PISP) was used at the jail to refer mentally ill detainees to the mental health unit within the facility for further evaluation and treatment. The Referral Decision Scale (RDS) was added to this procedure in an effort to increase the accuracy of the current method. Diagnoses resulting from the Schedule of Affective Disorders and Schizophrenia-Change Version (SADS-C) were used as the validating criteria for these two protocols. The SADS-C found 11.6 percent of the sample in this study suffered from some form of severe mental illness, a figure within the range of the 6 to 15 percent found in other studies. The findings suggest that, despite the limitations on time and budget, successful screening can be accomplished. These findings indicate that the PISP detected slightly fewer than half of the inmates diagnosed with a severe mental illness. The majority of the detainees not identified were diagnosed with depression, suggesting that questions of the PISP do not adequately address depressive symptomatology. Compared to the PISP, the RDS demonstrated greater acuity in the detection of mental illness, as nearly three-quarters of the mentally ill persons were identified. But the RDS had an elevated level of false positives when compared to the PISP. Both the PISP and the RDS appear to be useful as screening devices in detecting severe mental illness, but each instrument has obvious limitations. 1 table, 44 references