Several polyunsaturated fatty acids (C18-C22 acids) have been compared in their uptake by human platelets and their acylation into glycerophospholipid subclasses. This was also studied in the presence of linoleic and/or arachidonic acids, the main fatty acids of plasma free fatty acid pool. Amongst C20 fatty acids, dihomogamma linolenic acid (20:3(n-6)), 5,8,11-icosatrienoic acid (20:3(n-9)) and arachidonic acid (20:4(n-6)) were better incorporated. The uptake of 5,8,11,14,17-icosapentaenoic acid (20:5(n-3)) was significantly lower and comparable to that of C22 fatty acids (7,10,13,16-docosatetraenoic acid (22:4(n-6)) and 4,7,10,13,16,19-docosahexaenoic acid (22:6(n-3)) and linoleic acid (18:2(n-6)). In this respect, linolenic acid (18:3(n-3)) appeared the poorest substrate. The bulk of each acid was acylated into glycerophospholipids although the presence of linoleic and/or arachidonic acids diverted a part towards neutral lipids. This was prominent for 18:3(n-3) and C22 fatty acids. The glycerophospholipid distribution of each acid differed substantially and was not affected by the presence of linoleic and or arachidonic acids, except for 18:3(n-3) and 22:6(n-3) that were strongly diverted towards phosphatidylethanolamine (PE) at the expense of phosphatidylcholine (PC). The main features were an efficient acylation of 20:3(n-9) into phosphatidylinositol (PI) followed by 20:3(n-6) and 20:4(n-6), then by 20:5(n-3) and 22:4(n-6), and finally 22:6(n-3) and C18 fatty acids. This was reciprocal to the acylation into PE and to a lesser extent into PC which remained the main storage species in all cases. We conclude that human platelets may exhibit a certain specificity for taking up polyunsaturated fatty acids both in terms of total uptake and glycerophospholipid subclass distribution. Also the presence of polyunsaturated fatty acids of normal plasma, like linoleic and arachidonic acids, may interact specifically with such an uptake and distribution.