The role of the placenta in controlling the supply of fatty acids to the fetus was investigated in term placentas (n = 5) from normal pregnancies. The maternal side was perfused ex vivo for 90 min with a modified Krebs Ringer solution containing a physiological mixture of fatty acids - designed to mimic the composition of non-esterified fatty acids (NEFA) measured in the last trimester of pregnancy (n = 10) - and ratio of fatty acid to human albumin. The selectivity for alpha-linolenic acid (alphaLN) transfer to the fetal circulation was not significantly different from that observed when using the triglyceride (TG) composition (1.21 +/- 0.04), but significantly different for AA (1.43 +/- 0.12; p < 0.001) and docosahexaenoic acid (DHA; 2.02 +/- 0.09; p = 0.048). The absolute rate of transfer (nmol. ml-1) compared to that using the TG maternal perfusate composition was significantly different for LA (0.562 +/- 0.038; p = 0.50), alphaLN (0.130 +/- 0.009; p < 0.001), arachidonic acid (AA; 0.218 +/- 0.022; p = 0.001) and DHA (0.383 +/- 0.04; p < 0.001). Thus, placental selectivity for alphaLN and DHA appears to be relatively unresponsive to changes in the mixture of fatty acids in the maternal circulation but the selectivity for AA increased with the increase in the maternal AA concentration. For an 8-fold increase in the concentration of DHA in the maternal circulation there was a 13-fold increase in the transfer of DHA to the fetal circulation. For a 2-fold increase in the concentration of AA, transfer was increased 8-fold. For a 1.3-fold increase in the concentration of alphaLN, transfer was increased 2.1-fold. These results suggest that the maternal concentration of individual fatty acids, and hence the composition of the maternal diet, can have large effects on polyunsaturated/long-chain polyunsaturated fatty acids delivery to the fetus.