The objectives of the study were to determine the effect of the partial replacement of soyabean meal and rapeseed meal with feed grade urea or a slow-release urea on the performance, metabolism and whole-tract digestibility in mid-lactation dairy cows. Forty-two Holstein-Friesian dairy cows were allocated to one of three dietary treatments in each of three periods of 5 weeks duration in a Latin square design. Control (C) cows were offered a total mixed ration based on grass and maize silages and straight feeds that included 93 g/kg dry matter (DM) soyabean meal and 61 g/kg DM rapeseed meal. Cows that received either of the other two treatments were offered the same basal ration with the replacement of 28 g/kg DM soyabean and 19 g/kg DM rapeseed meal with either 5 g/kg DM feed grade urea (U) or 5.5 g/kg DM of the slow-release urea (S; Optigen R ; Alltech Inc., Kentucky, USA), with the content of maize silage increasing. There was no effect (P . 0.05) of dietary treatment on DM intake, which averaged 22.5 kg/day. Similarly, there was no effect (P . 0.05) of treatment on daily milk or milk fat yield but there was a trend (P 5 0.09) for cows offered either of the diets containing urea to have a higher milk fat content (average of 40.1 g/kg for U and S v. 38.9 g/kg for C). Milk true protein concentration and yield were not affected by treatment (P . 0.05). Milk yield from forage and N efficiency (g milk N output/g N intake) were highest (P , 0.01) in cows when offered S and lowest in C, with cows receiving U having intermediate values. Cows offered S also tended to have the highest live weight gain (0.38 kg/day) followed by U (0.23 kg/day) and C (0.01 kg/day; P 5 0.07). Plasma urea concentrations were higher (P , 0.05) at 2 and 4 h post feeding in cows when offered U and lowest in C, with animals receiving S having intermediate values. There was no effect (P . 0.05) of treatment on whole-tract digestibility. In conclusion, the partial replacement of soyabean meal and rapeseed meal with feed grade urea or a slow-release urea can be achieved without affecting milk performance or diet digestibility, with the efficiency of conversion of dietary N into milk being improved when the slow-release urea was fed.