We conducted a prospective study of a smoking ban on a general inpatient psychiatry service in response to staff concerns about the feasibility of a proposed hospital-wide ban. Demographic information, smoking history, and DSM III-R diagnoses were obtained for consecutively admitted patients during two study conditions: smoking and nonsmoking. A log of p.r.n. medication, seclusion, restraint, elopement, incident reports, and smoking-related discharges was kept for each patient. Chi-square analysis of 232 patients for whom demographic, smoking, diagnostic, and log data were complete showed no significant differences between study conditions for demographic or diagnostic variables. Two-tailed t-test analysis of the log data for these 232 patients showed no significant difference in disruptive incidents during smoking and nonsmoking conditions (p = 0.183). Fifty staff members answered pre- and post-ban questionnaires. Paired t-test analysis demonstrated a significant change in staff attitude toward supporting the ban. These data indicate that smoking can be stopped on inpatient psychiatry units without increases in unit disruption or adverse effects on staff morale.