1. In an experiment twenty-four cross-bred does were used, six in each of four physiological states: non-pregnant (NP), pregnant (P), lactating (L) or concurrently pregnant and lactating (CPL). They were offered a diet of high nutrient density (crude protein (nitrogen x 6.25) 272 g/kg dry matter (DM), metabolizable energy 15.5 MJ/kg DM) at one of two intakes on a scale based on metabolic body-weight (kg body-weight 0.75). The study lasted for 24 d and was divided into three consecutive 8 d periods: 'early', 'mid-' and 'late' pregnancy or lactation, or concurrent pregnancy and lactation. 2. Around the mid-point of each period the does were placed for 24 h in a gradient-layer calorimeter, and measurements of energy exchange were made. N balance was also measured throughout each period of the study. 3. Milk output in L and CPL does was measured by weighing the does immediately before and after suckling once each day. Milk samples were taken from a parallel group of does and the estimates of milk composition applied to the experimental group. 4. Milk yields and composition were similar for L and CPL does in early and mid-lactation, but in late lactation the milk output from CPL does declined rapidly. This was the period of maximal fetal growth rates. Energy and N retention increased in P and CPL does in late pregnancy. 5. Regression analysis was carried out to summarize the relations between metabolizable energy intake (MEI) and energy retention (ER; MEI-heat loss) in each physiological state. The equations were: NP: ER = 0.67 (SE 0.057) MEI-208, P: ER = 0.67 (SE 0.095) MEI-224, L: ER = 0.90 (SE 0.022) MEI-441, CPL: ER = 0.85 (SE 0.036) MEI-387, where all variables are expressed in kJ/kg body-weight 0.75 per d.