Sixty-four multiparous Friesian cows were given one of eight diets from weeks 4–20 of lactation at the former National Institute for Research in Dairying, Shinfield, Reading in 1983–84. The diets consisted of 13·5 kg/day of experimental concentrates plus hay at 4·5 kg/day (fixed) or ad libitum. Concentrates were high-starch (S) based on cereal grains and cassava, high-fibre (F) based on highquality by-products, and two intermediate types containing 2:1 (SF) and 1:2 (FS) starch:fibre, all with a crude protein concentration of c. 180 g/kg DM. The diets were designed to provide similar intakes of digestible energy and crude protein at fixed hay intakes.
With ad libitum feeding, hay intake was lowest on S concentrates. With increasing fibre in the concentrates, the yields of milk, protein and lactose fell but fat concentration and yield were increased. These changes were linear and were unaffected by hay intake. Increasing hay intake to ad libitum had no effect on milk yield but increased the concentrations and yields of fat and protein. It also changed a net liveweight loss on fixed intakes to a gain on ad libitum intakes.
Diet digestibility was measured using sheep at about maintenance and cows at production intakes. Digestibility was higher in the sheep, probably because of the difference in level of intake. In the cows, increasing fibre in the concentrates reduced the digestibility of dry matter, organic matter, energy and nitrogen but increased the digestibility of fibre fractions and ether extract. Level of hay intake had variable effects on mean digestibility and on the relationship with concentrate type.
Increasing the fibre in the concentrates increased the molar proportions of acetate and n-butyrate in the rumen and reduced the proportions of propionate, n-valerate and n-caproate.
It is concluded that, with mixed diets of hay and relatively high proportions of concentrates, changes in the starch:fibre ratio of the concentrates by substitution of high quality by-products for starchy ingredients can induce linear changes in the yields of milk and the principal solids constituents and in milk fat concentration. Since changes in yields of fat and protein went in opposite directions, the optimal concentrate composition depends on the milk composition required by the market. These conclusions may not apply if poorer quality by-products are used.