A sample of 107 sons of alcoholics aged 22-33 is compared to a representative sample of 1,274 Prague men of the same age range with respect to alcohol abuse and its psychosocial antecedents and correlates. Data sources were cumulative records of alcohol abuse in health care files and interviews. The estimated relative risk of alcohol abuse for sons of alcoholics versus other men is about 3.5 with both registration and self-report criteria if a broad definition of abuse is adopted. If, however, serious registered abuse and/or medical treatment of alcoholism at or before age 22 is the criterion of abuse, the estimated relative risk is about 10. In both compared samples, father's low education, discontinued family socialization and undisciplined behavior in childhood were antecedents of abuse registered at or before age 22. Self-reported abuse in the 6 months before interview had similar psychosocial correlates in both samples (heavily drinking friends, a positive attitude to heavy drinking, etc.). In both samples about 40% began to drink regularly at 17 or earlier. However, early start of drinking led very frequently to early registered abuse in sons of alcoholics whereas no such contingency was observed in the general male population.