Primiparous (n = 8) and multiparous (n = 24) Holstein cows averaging 591kg of BW (SE = 12) were allotted at 2 wk postpartum to eight blocks of 4 cows each on the basis of calving date and parity. Cows were fed timothy silage for ad libitum intake. The experiment was carried out between wk 4 and 15 of lactation. Cows within each block were assigned randomly to a concentrate containing soybean meal fed with corn, beet pulp, or a mixture of oats and barley (50:50 on a DM basis) or fish meal fed with beet pulp. Concentrates were fed at about .8% of BW. The four treatments were designed to give similar CP and NEL intakes from the concentrate. Starch degradability and energy source (nonstructural vs. structural carbohydrates) had no effect on total DMI of cows supplemented with soybean meal. Total DMI was similar for cows fed beet pulp with soybean meal or fish meal. Production of 4% FCM was on average 3.5 kg/d higher for cows fed beet pulp than starch. All cows gained BW. In low concentrate diets, beet pulp compared with starch improved N utilization of high moisture timothy silage by increasing production of 4% FCM without affecting total DMI; however, protein degradability does not seem to be important in low concentrate diets.