A 23-week trial was conducted using 4 groups each of 6 ewes to study the effects on milk yield and composition of substituting non-protein nitrogen (NPN) for 75 % of the dietary nitrogen. The control group received a winter ration which included silage followed by a summer ration containing lucerne. The test groups received substitutes of urea, ammonium sulphate, or urea + (NH 4 ) 2 SO 4 . The milk yield of the test groups was lower (P < 0-01) than that of the control group. Depression of the fat content of the milk reached statistical significance (P < 0-01) in the groups receiving (NH 4 ) 2 SO 4 and urea + (NH 4 ) 2 SO 4 . In comparison with the control group, the protein content of the milk was greater in the group receiving urea and smaller in the other 2 test groups. The milks of the test groups had lower amounts of essential amino acids in the free state and higher amounts of non-essential amino acids than did the control milk. Milk-clotting time was increased in the test groups; the increase was greatest (P < 0-01) for the group receiving (NH 4 ) 2 SO 4 and least (P < 0-05) for that receiving urea.The desirability of increasing the percentage of ruminant dietary protein which can be replaced by non-protein nitrogen (NPN) has prompted a large number of investigations in this area. A number of authors (Virtanen, 1966(Virtanen, , 1967Todorov et al. 1969) have studied the influence of rations in which NPN substances, mostly urea and ammonium salts, were the only N source.Rations in which NPN replaced over 50 % of the required N were fed to animals by various workers. Thus, Hale (1955) reported that lambs fed rations in which urea replaced 92 % of the needed N showed good weight gain, and Johnson et al. (1942) found no adverse influence of 65 % urea. Sauer (1938) also found this when feeding 32 g urea/head daily to wether lambs. Trials conducted in Ohio by Woods & Tolman (1966) demonstrated that pelleting of urea with lucerne hay may allow 40 % N to be replaced in the ration of highproducing dairy cows.Results from experiments carried out in Nebraska by Conrad & Hibbs (1966) showed that dairy cows used urea to advantage when it replaced 66-5 % of dietary N. These authors considered that in such a case the amount of mineral supplementation should be increased several fold.