Three pairs of Indiana University rat lines (inbred alcohol-preferring and nonpreferring rat lines [P/NPs], high- and low-alcohol-drinking rat lines [HAD/LAD1s and HAD/LAD2s]) were bred in the School of Medicine colony to drink high versus low daily amounts of a 10% vol/vol alcohol test solution (>5.0 g/kg body weight vs. <1.5 g/kg body weight), and a high versus low proportion of alcohol to water (>2:1 vs. <0.5:1) by the end of a 3-week alcohol-water choice condition. This choice phase was always preceded by four days of a forcing procedure with alcohol as the only fluid. The present study examined the contribution of the forcing procedure to the alcohol intake of animals in each pair of lines by comparing daily alcohol intake of rats housed in experimental chambers in a forced group (4 days with only alcohol solution to drink followed by 22 choice days) versus a choice group (both alcohol and water available all 26 days). As expected, under the initial alcohol exposure, high-drinking line rats drank more alcohol than low-drinking line rats, and all forced groups drank more alcohol than choice groups. At the start of the choice phase, all low-drinking line forced groups immediately dropped their alcohol intake to the level of their choice groups. In contrast, all high-drinking line forced groups maintained a high level of alcohol intake under choice, whereas all high-drinking line choice groups slowly increased average alcohol intake across the 22-day choice phase, ending near the average intake of their forced groups. However, a small subset of each high-drinking line choice animals failed to increase alcohol intake until subsequently forced with alcohol for 4 days and tested again in choice. These results indicate that the alcohol-forcing procedure used in deriving these lines resulted in the selection of more than one pathway to a high-drinking phenotype. In addition, high-drinking line animals appeared more sensitive to the differences between laboratory- and colony-testing environments than low-drinking line animals. These data suggest that these high-drinking lines may represent an unexpectedly appropriate complex model of how multiple factors may contribute to the genesis of human alcoholism.