PRN psychotropic drug use is a frequent practice on psychiatric units. This study examined medical records for patient demographics and prescribing patterns of drugs. The omission of the indication and frequency of dosing for the drug were common prescribing errors. There was no significant association between diagnosis and whether a patient received PRN drugs. Anxious patients were significantly more likely to receive PRN anxiolytics whereas psychotic patients were as likely to receive PRN anxiolytics as antipsychotics. Anxiolytics also were significantly more likely to be judged effective by the nursing staff than antipsychotics. A perplexing finding was that manic and schizophrenic patients were significantly less likely to receive PRN hypnotics than were patients with Major Depression. The clinical implications of the findings are discussed and recommendations are made to improve standards of practice for this commonplace treatment.