Milk from forage (MF) is an estimation of the milk produced solely from forage intake. It is calculated by subtracting milk production theoretically allowed by concentrates from total milk production, assuming that maintenance requirements are covered by the forage portion of the diet. Eight multiparous Holstein cows in early lactation were used in a replicated 4 x 4 Latin square design to evaluate the impact on MF of different sources of carbohydrate with forage that was high in RDP. Diets were alfalfa-based total mixed rations that were formulated to provide similar concentrations of NEL and CP while differing in rumen degradability of concentrate carbohydrates. Treatments were 1) cracked corn (control), 2) ground corn (GC), 3) GC plus wheat starch (GC+S), and 4) GC plus dried whey permeate (GC+W). The GC and the GC+S treatments increased MF as calculated on a protein basis (14.8 vs. 10.5 kg) and increased average MF production (8.6 vs. 5.5 kg) compared with the control. Protein of forage was used more efficiently with GC and with GC+S, as shown by the lower differences between allowable MF, which estimates the potential for milk production from forage, and MF on a protein basis for these 2 treatments when compared with the control. Compared with the control, DMI increased with GC and GC+S; GC+W yielded the highest DMI. Milk production with GC+W (35.8 kg/d) was lower than with GC and GC+S (37.5 kg/d) but was higher than the control (34.0 kg/d). Milk fat concentration was higher with GC+W and lower with GC+S; GC and the control had intermediate values. Milk urea was higher with the control diet compared with the other 3 treatments. Results emphasize the advantage of using concentrates of higher degradability in the rumen to improve MF and milk production when feeding silage with high rumen-degradable protein.