1. This study was designed to determine prospectively whether changing alcohol consumption influenced the proportion of plasma linoleic acid independently of diet or smoking habits, and to evaluate changes in the plasma linoleic acid concentration as a potential marker of changes in alcohol consumption. 2. Fasting plasma fatty acid profiles were investigated in 72 male drinkers who were randomly assigned to drink low-alcohol beer or to maintain their usual drinking habits for a period of 4 weeks. 3. At entry to the study, a higher alcohol intake was associated with lower proportions of plasma linoleic acid, arachidonic acid and dihomolinolenic acid and higher proportions of plasma palmitoleic acid, independently of changes in body mass index. Smoking habits were unchanged and there were no major changes in diet during the period of the intervention. 4. Because the plasma palmitoleic acid concentration has been suggested as a possible marker of 'at risk' drinking, we investigated plasma fatty acid concentrations as indicators of alcohol intake. The plasma palmitoleic acid concentration was not a useful discriminator. Indices determined using logistic regression and combining plasma apolipoprotein A-II and linoleic acid concentrations gave better discrimination than either variable alone.