The effect of the proportion of clover in the diet (200, 500 or 800 g/kg total dry matter (DM)) on milk production of cows housed indoors and fed on a mixture of perennial rye-grass and white clover was measured in mid (Expt I) and late (Expt II) lactation. Higher clover contents increased the nutritive value of the diets, resulting in increased energy and protein intakes. DM intakes of cows offered 500 or 800 g clover/kg DM diets ad lib. (Expt I and Expt II, Period 1) were not significantly different but were 11–17% greater (P<0·05) than intakes of cows fed on 200 g clover/kg total DM diets. Cows offered restricted allowances (Expt II, Period 2) had similar intakes irrespective of diet. In Expt I cows fed on 500 or 800 g clover/kg DM diets ad lib. produced 30 or 33% respectively more milk (P<0·05) than cows fed on 200 g clover/kg total DM diets. During Expt II, Period 1, cows fed on 500 or 800 g clover/kg DM diets ad lib. produced 18 or 16% more milk (P<0·05) respectively than cows given 200 g clover/kg total DM diets. In both these experiments the increased milk yields were due to increased intake and the higher nutritive value of the high clover diets. There was no difference in the feed conversion efficiencies of cows if maintenance energy requirements were taken into account. However, cows on restricted allowances (Expt II, Period 2) showed no significant difference in milk yield, indicating that the effect of increased nutritive value was very slight. There were no consistent effects on milk fat, protein or lactose concentrations. Concentrations of blood and milk urea increased as the clover content of the diet increased (Expt 1 only), and this was associated with increased milk non-protein N and a decreased ratio of casein N[ratio ]total N. Both trials indicated an optimum clover content in the diet for milk production of 600–700 g/kg total DM.