When animals are housed in a respiration chamber, the utilisation of high frequency gas analysers helps in understanding and quantifying variation in heat production (Q) and partitioning of metabolisable energy (ME) intake. The objective of this chapter is to describe a methodology based on mathematical modelling of gas exchanges between animals and the respiration chamber and to highlight some key-results that were obtained about partitioning of ME intake. The model assumes that O2 consumption and CO2 production of animals only result from processes associated with fasting heat production (FQ, constant during the day), thermic effect of feeding (TEF, induced by feed intake, digestion and metabolism of nutrients) and physical activity. From these gas exchanges, FQ, TEF and activity-related heat production (AQ) were calculated respectively, in growing pigs, calves, broilers and turkeys. The data obtained from the model were used to estimate the exponent to calculate metabolic body weight, which equalled 0.60 in growing pigs, 0.70 in birds and 0.85 in calves. These exponents range in the same order as specific allometric growth coefficients of viscera in these species, which contribute to a great extent to whole body Q. In ad libitum fed animals, FQ accounted for 30% of ME intake in non-ruminants but for 49% in calves. In pigs and calves, it has been shown that FQ depends on the previous level of feeding. When animals are fed a standard diet, TEF was similar between non-ruminants (from 13 to 17% of ME intake) but rather lower in calves, due to the low fibre content of the milk replacer. Except for turkeys (13% of ME intake), AQ averaged 9% of ME intake. Taking the effect of feeding level on FQ into account, the variable components of Q (TEF, AQ) were used to calculate energetic efficiencies of energy retention. In growing pigs, these were estimated at 60 and 80% for energy retained as protein and as fat, respectively.