Ethanol use by pregnant women increased, in a dose-dependent manner, the rate of sheep erythrocyte rosette (E-rosette) formation with T lymphocytes. The time curve for E-rosette formation by T cells from nondrinking subjects was biphasic, with a rapid formation of half the E-rosettes within the first 16 min, followed by a much slower rate for E-rosette formation until the maximal T-cell percentage was reached overnight. For pregnant drinkers, greater than 85% of the E-rosettes formed during the initial rate period, with a concomitant smaller number forming during the overnight incubation. Despite the faster initial rate of E-rosette formation in the drinking subjects, the total percentage T cells was the same for both groups. Other demographic factors, like tobacco or marijuana use, or trimester, did not significantly contribute to the observed differences. An increase in the rate of E rosetting was also obtained by incubating lymphocytes from nondrinkers overnight in physiologically attainable concentrations of ethanol (less than or equal to 0.1%). These results demonstrate that drinking by pregnant women, even at relatively moderate levels (2 oz/week absolute ethanol), causes alterations in their cellular immune systems. With the ability of ethanol to cross the placental barrier and persist in utero, it is apparent that these levels of ethanol have the potential to affect the developing fetal immune system.