A portable indirect calorimeter adapted from adult use was validated for use in preterm infants. Oxygen consumption (VO2) and carbon dioxide production (VCO2) were subsequently measured in 16 preterm infants breathing spontaneously in room air (canopy mode) and in nine preterm infants receiving intermittent positive pressure ventilation (ventilator mode). Validation of the system was performed using a gas injection technique with nitrogen to simulate VO2 and carbon dioxide for VCO2. Mean errors in validation of the canopy mode were 1.4% and 0.2% for VO2 and VCO2 with limits of agreement of 0.6 (+2SD) ml/min and -1.3 (-2SD) ml/min, and 0.9 (+2SD) ml/min and -2.3 (-2SD) ml/min respectively. In validation of the ventilator mode mean errors were -1.8% and -5.05% for VO2 and VCO2 with limits of agreement of 1.02 (+2SD) ml/min and -0.74 (-2SD) ml/min, and 0.93 (+2SD) ml/min and -1.45 (-2SD) ml/min respectively. Values of VO2 and VCO2 in 16 preterm infants in the canopy mode were 6.2 ml/kg/min (0.5 1SD) and 6.7 ml/kg/min (0.6 1SD) and in nine preterm infants in the ventilator mode 4.98 ml/kg/min (1.09 1SD) and 4.74 ml/min/kg (1.08 1SD) respectively. Mean energy expenditure was 45.5 kcal (191 kJ)kg/day for infants measured in the canopy mode and 35.5 kcal (149 kJ)/kg/day for ventilated infants. This metabolic system can be adapted for use in the newborn but accuracy is reduced when it is used in those weighing less than 1000 g.