The importance of conditioning the inhaled gas for maintaining the body temperature during artificial respiration was investigated. The mean body temperature (MBT) was deduced from readings from five measuring sites, four of which were situated at the skin and the fifth in the rectum. Temperature recordings were made every 15th min. Twenty patients were admitted to the study. In 10 patients a heat and moisture exchanger was used (the HME group), and the other 10 were ventilated without an HME (the control group). The patients were normoventilated, and a non-rebreathing system was used. All operations were made in the ENT-region of the body. Great care was taken to avoid variation of external factors that may affect the MBT. We found that the MBT decreased 0.2 degrees C/h less in the HME group than in the control group. If corrections were made for differences in amounts of fluids given and age factors, a difference in heat loss of 41.6 kJ/h between the two groups could be derived from this figure. This difference was statistically significant. Our finding correlated fairly well with a predicted reduction of heat loss of 26.0 kJ/h for the type of HME used. A certain margin of error seemed to be inevitable in measuring body temperatures, and the reason for this is discussed. Our results support the fact that the investigation is adequately designed, and that the heat conserved with an HME is rather low.