Ethanol use is widespread in adolescents, yet only some transition to problematic drinking. It is important to understand why the risk for problematic drinking varies across sub-groups of adolescents. This study reports a short-term selection program to generate Wistar rat lines (high and low adolescent ethanol drinking, ADHI and ADLO lines, respectively) that significantly differ in ethanol drinking at adolescence. The S0 generation and filial generations 1 (S1), S2, and S3 of ADHI and ADLO offspring were tested for basal or stress-induced ethanol intake at adulthood, or for shelter-seeking and risk-taking in the multivariate concentric square field test (MSCF). The study generated lines with significant differences in free-choice ethanol drinking at adolescence. The effects of the selection were observed at adulthood, beyond the stage in which the selection was conducted: S1-ADHI but not S1-ADLO adult male rats exhibited stress-induced drinking. These effects were associated with significant alterations in shelter-seeking and risk-taking behaviors. ADHI rats spent significantly less time in areas of the MSCF whose exploration entails risk-taking and significantly more time in dark, sheltered areas. Some of these effects were normalized by the administration of 0.5 g/kg ethanol. There were no line differences in ethanol-induced latency to lose the righting reflex or sleep time. These findings indicate that genetic risk of enhanced ethanol intake at adolescence is still present at adulthood, long after the developmental window when the selective breeding occurred. Exposure to stress at adulthood triggers the vulnerability associated with this genetic risk, an effect associated with enhanced anxiety.