S U M M A R YBetween weeks 3 and 16 of lactation 191 British Friesians received a basal ration ad libitum of either grass silage (GS), grass and maize silage (GS/MS) blended in a ratio of 1:2 or grass silage and 4% NaOH-treated barley straw (GS/TS) blended in a 4:1 ratio. The grass silage, which was not well preserved, had an estimated metabolizable energy (ME) content of 9-9 MJ/kg D.M., while corresponding values for the GS/MS and GS/TS mixtures were 106 and 9-5 MJ/kg D.M., respectively.A conventional concentrate (CC) based on barley and soyabean meal was fed at three levels (9, 6 and 3 kg fresh weight/day). Although the 9 kg level was only used in conjunction with GS, the two lower levels were fed with all three forage types. Concentrate composition was also varied at the 6 and 3 kg levels by replacing, in each case, 2 kg CC with a special concentrate (SC) which, in addition to barley, wheat and soyabean meal contained, on an air-dry basis, 25 % fishmeal and 25 % protected fat in the form of Megalac* The concentrations of crude protein (CP) and ME in the CC and SC were 186 g/kgD.M. and 12-9 MJ/kg D.M. and 285 g/kg D.M. and 17-2 MJ/kg D.M., respectively.At an early stage the treatments in which 3 kg SC were fed in conjunction with either GS or GS/ TS were abandoned due to an extremely high incidence of ketosis. The problem appeared to be linked to the use of sources of undegraded protein and protected energy in cases where intake was low and energy deficit high. In the remaining treatments the use of SC did not significantly affect milk yield, milk composition or live-weight change.The D.M. intakes and milk yields from the GS/MS mixture were consistently higher, at comparable concentrate inputs, than those obtained from either GS or GS/TS. Forage intakes decreased by 0-21, 0-26 and 0-23 kg D.M./kg concentrate D.M. for GS, GS/MS and GS/TS treatments, respectively.Averaged over concentrate type the response to increasing concentrates from 3 to 6 kg fresh weight/day with GS, GS/MS and GS/TS forage treatments was 1-36, 100 and 1-38 kg milk/kg additional concentrate D.M., respectively. Although no very clear pattern emerged for milk fat concentration, the yield of milk fat declined as concentrate input decreased, but was higher for GS/ MS (816g/day) when compared with the GS (710g/day) and GS/TS (696 g/day) treatments. At both the 3 and 6 kg concentrate levels the concentration and yield of milk protein were higher for the GS/MS treatments when compared with GS or GS/TS treatments.