The effect on liver fatty acids of two typical beef production systems. steers fed grass (grazing) and bulls fed concentrates, was investigated. Liver fatty acids were also studied in lambs grazing grass. Total fatty acid content of liver in the beef animals was not affected by production system, being 3.5% for grass and 3.7% for concentrates although carcasses of the latter animals were leaner. The percentages of the major non polyunsaturated fatty acids (PUFA) were also similar. (concentrates in parentheses) 18:0, 25.2% (25.1%); 16:0, 13.3% (14.5%); 18:1 12.6% (14.0%). Total liver PUFA contents were similar for the two production systems and much higher than in muscle fr om the same animals. All n-3 PUFA were present in greater amounts in liver from grass fed animals and n-6 PUFA were higher in concentrate fed animals: mg/100 g liver, 18:3n-3 92 (32); 20:5 n-3, 151 (17); 22:5 n-3, 283 (108); 22:6 n-3, 83 (32); 18:2 n-6, 172 (444); 20:4 n-6, 194 (270). The P:S ratios were 0.20 (grass) and 0.32 (concentrates) and the n-6:n-3 ratios were 0.71 and 4.8, respectively. Livers from grass-fed lamb had a higher fat content, 4.9%, than the beef livers and a lower percentage of 18:0, (21%); but more 16:0, (16%) and 18:1 (21%). Total PUFA content of lamb's liver resembled that of beef liver and the composition was similar to that of the steers fed grass. However the concentrations of 18:3 n-3 and 22:6 n-3 were higher in lamb and contributed to a lower n-6. n-3 ratio of 0.46 although the P:S ratio of 0 18 was similar to that in beef liver. Ruminant liver is potentially a good source of C20 and C22 PUFA in the human diet particularly from grass fed animals, with a highly desirable n-6:n-3 ratio and this may be more important nutritionally than the low P:S ratio since people can offset this elsewhere in the diet. (C) 1998 Elsevier Science Ltd. All rights reserved.