The effect of body fat content on the protein metabolism of energy-restricted sheep has been studied in two experiments. In the first experiment, six Suffolk-cross wether sheep, three weighing about 39 kg and three of about 61 kg, were given progressively increasing amounts of casein-N from 0 to 3000 mg N/kg metabolic body weight (W0·75) daily with constant energy, 91 kJ/kg W0·75 daily, from a high-propionic acid mixture of volatile fatty acids (VFA). In the second experiment, two lean and two fat sheep of similar body weights were given progressively increasing amounts of casein with the same VFA mixture. All the animals attained a positive N balance when they were in negative energy balance. N balance was not affected by body fatness of the magnitude studied, although lean animals utilized increasing levels of standard VFA (acetate-propionate-butyrate 65:25:10, molar proportions) infusion for N accretion more efficiently than fat animals. Endogenous energy was utilized for protein accretion with an efficiency of 0·56. Supply of glucogenic VFA equivalent to 28 mmol glucose/kg W0·75 reduced fasting N excretion by 39%. Fasting heat production decreased from 335 to 300 kJ/kg W0·75 with the infusion of casein and glucogenic VFA. It is argued that fasting induces additional heat losses due to raised protein metabolism and is unsuited as a baseline for dietary assessment.