Five cocktail parties, lasting 60-70 min., were attended by a total of 91 male college students. Scores were obtained on the Park problem-drinking scale and on Heilbrun need scales from the Gough and Heilbrun Adjective Check List (ACL). The ACL was given before the parties, after Ss [subjects] had drunk 4 oz of liquor (choice of gin, scotch or bourbon), and at the end of the parties, during which the Ss had drunk an average of 11 oz of liquor (range, 4-28 oz). Preparty personality variables on which those with high scores on the problem-drinking scale (high PDS) scored low were deference, order, affiliation, nurturance, intra-ception and endurance. High PDS scored high on aggression, anatomy, change and exhibition. Thus high PDS tended to characterize themselves as autonomous, self-sufficient, aggressive, relatively unconcerned with and uninterested in others, lacking in perseverance, impulsive and disorganized, and as seeking novelty and variety in experience. High scores on the problem-drinking scale were also positively associated with amount of liquor consumed at the parties. Predictions of what changes would occur under alcohol were made on the basis of descriptions of drinking self on the ACL made while in a sober condition by a previous sample of male college students. Results indicated that in the present sample autonomy, aggression, exchibition and hetero-sexuality increased under alcohol; deference, order, affiliation, succorance, abasement, nurturance and endurance decreased. In general, inhibition vanished from drinking self descriptions and Ss tended to describe themselves in terms of activity rather than thought. Descriptions became increasingly self-centered as social characteristics tended to drop out of self-descriptions. This pattern of becoming less socialized under alcohol occurred particularly at levels of consumption above 4 oz. Hiph PDS did not increase or decrease on any of the variables under alcohol more than low PDS and in a sober condition, however, high PDS had high scores on variables which increased under alcohol, and low scores on variables which decreased under alcohol. It is suggested that this may provide high PDS a reward from drinking which may, in turn, help to account for their heavy and frequent drinking and possible future alcoholism.