A sample of 112 U.S. adolescents completed daily diaries about their health behaviors, including drinking and intercourse, for eight weeks. In analyses using t-tests and hierarchical linear modeling, each respondent's rate of condom use after drinking was compared with his or her rate of use when not drinking, and predictors of condom use were examined. Rates of condom use did not differ significantly between sexual events preceded by drinking (use in 54% of events) and those not preceded by drinking (use in 52% of events). Condoms were more likely to be used during sexual events with casual partners than in those with steady partners, less likely to be used on occasions when other birth control was used and more likely to be used when the sexual encounter was expected. In the multivariate analyses, the odds of condom use were not associated either with whether a teenager had been drinking before sex or with the quantity of alcohol consumed.