Estimates of average daily energy expenditure and minimal observed oxygen consumption are commonly used to characterize the energy metabolism of neonates. Yet, the errors inherent in these estimates have not been defined. Using measurements of oxygen consumption and carbon dioxide production made in healthy growing low birth weight infants during eight consecutive 3-h interfeeding epochs, we have determined the variability in the mean oxygen consumption, carbon dioxide production, respiratory quotient, total daily energy expenditure, and the minimal observed oxygen consumption among the feeding epochs. The coefficient of variation for oxygen consumption ranged from 3.1 to 9.1%, for minimal observed oxygen consumption from 3.7 to 16.7%, for carbon dioxide production from 3.3 to 7.4%, and for total daily energy expenditure from 2.9 to 7.6%. The SDs for respiratory quotient ranged from 0.008 to 0.066. From these 24-h data we have calculated the error in predicting daily estimates of the mean values for these variables if observations are made for less than 24 h. As expected, this error decreases with increasing duration of observation. These data should prove useful in the design and interpretation of investigations of neonatal energy expenditure.