Twenty-four male Ile-de-France lambs (six blocks of homologous lambs) were used to study the effect of four feeding systems on muscle triglyceride (TG) and phospholipid (PL) fatty acids (FA) from the longissimus thoracis (LT): raised and finished on cool season grasses (G), raised on the same grasses and stall-finished, indoors. on concentrates and hay, respectively, for 22 (GSS) and 41 days (GSL), and stall-feeding, indoors, on concentrate and bay during both growing and finishing periods (S). In TG, similar decreases (P < 0.05) of proportions of linolenic acid were observed after changing from grass feeding to stall feeding (GSS and GSL), and a decrease (P < 0.05) in proportions of conjugated C18:2 cis9, trans11 (CLA cis9, trans11) was obtained after a long period of concentrate feeding (GSL). In PL, C22:5 n - 3 achieved a significantly (P < 0.05) lower level in GSL lambs compared both G and S lambs. A similar non-significant tendency was observed in the case of the other very long chain n - 3 polyunsaturated FA. The separate analysis of fatty acids of TG and PL from the LT muscle underlined that TG afforded a more significant lowering effect than PL on the overall ratio between C18:2 n - 6 and C18:2 n - 3 in muscle lipids and on the health potential of meat for the consumer. A PCA analysis combining FA composition of TG and PL, and growth performances of the lambs allowed an efficient discrimination between the four feeding systems. (c) 2007 Elsevier Ltd. All rights reserved.