A group of 40 mice involved in bacterial LD50 estimations was monitored over a period of 4 days by two teams of observers. One team measured the colonic temperature of the mice, the other assessed them clinically and killed those they judged terminally ill. The temperatures of all mice considered terminally ill dropped before this judgement was made. Measurement of such infection-induced hypothermia provides a more objective, reproducible, and earlier endpoint than that used in conventional LD50 estimation.