In a series of experiments, Klebsiella pneumoniae was inoculated intratracheally into rats and mice, and the temperature of the animals was recorded twice daily using microchip transponders. Transponders are interrogated by radio frequencies and were implanted either subcutaneously or intraperitoneally. The microchip temperatures were compared with rectal temperatures taken at the same time. The purpose of the experiments was (a) to investigate the practicability and reliability of the ELAMS for temperature recording; (b) to compare values given by subcutaneously or intraperitoneally implanted transponders with rectal temperatures; and (c) to determine a 'temperature-cut-off point' as an alternative for 'death of the animal' as an end point for the experiment. The results showed that the ELAMS was easy to operate and no important drawbacks in the use of the system were observed. The temperatures generated by the transponders implanted subcutaneously and intraperitoneally did not differ significantly from rectal temperatures. In two out of three experiments on rats, it was shown that when the temperatures reached values below 36 degrees C, the median survival time of the animals was 24 h. In the one experiment on mice the same median survival time was observed at 36 degrees C. In one experiment using rats, however, the disease was so acute that death occurred before any temperature drop was seen. The results show that when a 36 degrees C cut-off point is used instead of the time of death in this particular animal model, the statistical analysis was not altered, but that it would spare animals further suffering for approximately 24 h. The argument that measuring body temperature is a laborious job and stressful to the animals is overcome when the ELAMS system is used.