Alcohol addicts and control subjects were tested by experimental tasks to determine the effect of moderate doses of whisky on performance demanding continuous attention over relatively long periods. When working in isolation the two groups of subjects did not differ significantly in accuracy, nor did alcohol significantly affect their performance. Working under competitive instructions in a group setting, the addicts made more errors than the control subjects, and alcohol impaired accuracy in both types of subject. Questionnaire data revealed considerable uncertainty about the alcohol content of the beverages, little reliance on their taste, and a predominantly unfavorable evaluation of alcohol effects. An experiment in signal detection requiring rapid search showed a sizable drop in performance as a result of alcohol, and gradual improvement as the toxic effects wore off.