Capillary blood alcohol concentrations (BAL) measured in 22 young adult male volunteers each of whom received three different treatments of alcohol (total N = 66) have been related quantitatively to the dose of alcohol ingested. Linear regression with reasonably homogeneous variances have been found when the BAL at 2, 2.5 and 3 hr are divided by the person's body weight and plotted versus the g/kg (D/W) dose. Error analysis indicated that the least error in the predicted dose (D/W) was obtained for BAL measured at 2 hr post dosing. In the latter case the mean absolute error was 6.23%, 59% of the errors were within +/- 5% and 88% of the errors were within +/- 10%.