The role of the placenta in controlling the supply of fatty acids to the fetus was investigated in term placentas from non-smokers (n 5), smokers (>ten cigarettes/d; n 5) and after addition of ethanol at 2 mg/ml (n 4). The maternal side was of the placenta was perfused ex vivo for 90 min with a physiological mixture of fatty acids and fatty acid:human albumin ratio. There was no effect of smoking on the transfer of linoleic (LA, 18 : 2 n-6), α-linolenic (αLN, 18 : 3 n-3), arachidonic (AA, 20 : 4 n-6) or docosahexaenoic acid (DHA, 22 : 6 n-3), expressed per perfused area (calculated from H218O exchange). However, the presence of ethanol in the perfusate at a concentration of 2 mg/ml significantly reduced (P<0·01) the absolute rate of transfer of the two n-3 polyunsaturated fatty acids, αLN and DHA. This specific effect of ethanol on αLN and DHA also resulted in an altered selectivity for transfer of individual fatty acids. In the non-smoking control group the placenta selectively transferred polyunsaturated fatty acids to the fetus in the order DHA>AA>αLN>LA. The order of selectivity was unaltered in placentas from smokers, but the addition of ethanol to the perfusion medium altered the order of selectivity to AA>αLN>LA>DHA. The presence of ethanol in the perfusate was also associated with a significant reduction (P<0·05) in the clearance of H218O. These results suggest that the presence of ethanol at a concentration of 2 mg/ml may reduce the availability of polyunsaturated fatty acids to the developing fetus.