Data from a 1988 survey on United States drinking practices and related problems was used to derive the proposed DSM-IV definitions of alcohol abuse and dependence. The prevalence of DSM-IV alcohol abuse and dependence combined, incorporating the DSM-III-R duration criterion, was 6.00% in this general population sample. The majority of respondents were classified as alcohol dependent (5.93%), dependent without abuse (5.24%), and dependent with physiological dependence (5.06%). The rate for DSM-IV alcohol abuse was negligible (0.06%) while elimination of the duration criterion had little impact on the prevalence of DSM-IV alcohol use disorders. Reasons for the extremely low prevalence of DSM-IV alcohol abuse and the slight increase in the prevalence of DSM-IV alcohol use disorders as the result of eliminating the duration criterion are discussed in terms of the content of the abuse category and its relationship to the dependence definition.