The Missouri Alcohol Severity Scale (MASS), derived from the Alcohol History Form (AHF), was developed to measure both the severity of alcohol misuse and the effects of alcoholism treatment. The AHF responses of 5921 alcoholics, consecutive admissions to treatment programs at 8 public facilities in Missouri, USA, were examined. Most of these alcoholics, having a mean age of 43, were men (87%), whites (43%) and readmissions (60%). Item-total correlations were computed for responses to 36 of the 52 AHF variables having high face validity for alcohol severity and capable of showing posttreatment change. After 16 items with the lowest item-total correlation were dropped from the item pool, the Kuder-Richardson reliability coefficient, a measure of internal consistency, reached 0.84. The scores on these 20 items were distributed across all possible scores (0-64), the mean (.+-. SD) being 20.16 .+-. 13.25 and the median and quartiles being 17, 9 and 28. In 208 of the inpatients restested 3 mo. following discharge, the pretreatment and posttreatment means were 18.98 and 6.91, the correlation between means being 0.27. In a test-retest sample of 47 consecutive admissions to a detoxication program at St. Louis State Hospital, a correlation of 0.74 was obtained. The MASS appears to be capable of measuring alcoholism severity and posttreatment change.