Urinary excretion of purine derivatives (PD) was used to estimate the microbial N (MN) supply to sheep in three experiments designed to examine the effects of DMI and BW on the efficiency of microbial N supply (EMNS) to the host animal. In Exp. 1, four sheep of about 45 kg BW were given 328, 656, 984, and 1313 g of DM/d of a hay/concentrate diet in a Latin square design. Excretion of PD per kilogram of digestible organic matter intake (DOMI) increased with intake, and EMNS increased from 12.0 to 28.3 g of MN/kg of OM digested in the rumen (DOMR). In Exp. 2, 19 sheep ranging from 22 to 73 kg BW were all offered 820 g of DM/d of the same diet as that fed in Exp. 1. Although DM digestibility was relatively constant, PD excretion varied from 4.5 to 13.5 mmol/d and EMNS from 8 to 36 g of MN/kg of DOMR, both inversely related to animal BW. In Exp. 3, five sheep of 48 to 57 kg BW were given a different diet at 702, 966, or 1,237 g of DM/d. Purine derivative excretion per kilogram of DOMI increased with the DMI:BW ratio. Calculated EMNS ranged from 23 to 35 g of MN/kg of DOMR. Pooled data from all experiments showed EMNS to be related to the DMI:BW ratio. It is suggested that the DMI:BW ratio defines the ruminal digesta passage rate and hence outflow of microbial protein. The results imply that the EMNS for a given diet is not constant, but changes with intake.