Twenty-nine corpses were subdivided into three groups. Normally from the 3rd h post mortem on, they were suspended undressed in a tub holding 1,000 l in nearly still water of temperatures approximately 20 degrees, 10 degrees and 0 degrees C. The rectal temperature was measured, normally until the 33rd h post mortem. Time of death was calculated by means of the mathematical analytical two-exponential formula suggested by Marshall and Hoare (1962), in the version used by Brown and Marshall (1974). The adapting parameters of the formula were standardized according to the principle of Henssge (1979, 1981) and related to standardization by adjusting factors to body weight stated for standard values of cooling, i.e., undressed corpses in calm air. After termination of the post mortem temperature plateau, it was found that undressed corpses suspended in water of temperatures of approximately 20 degrees and 10 degrees C cool as quickly as undressed corpses of half the body mass in calm air of the same temperatures. As to the duration of the post mortem temperature plateau in water suspension time from the time of death, it may only be indirectly concluded that it is linked to the subsequent speed of cooling in the same way which is well known in the case of air cooling. Statistical standard values are given concerning the differences between the computed and the real times of death. Unexpectedly, the experiments in water at approximately 0 degrees C yielded distinctly slighter temperature which were especially marked at rectal temperatures up to approximately 11 degrees C in corpses of great body mass and small body surface in proportion to and equally, without regard to body mass. As an explanation of this, a decrease in the thermal conductivity of the subcutaneous adipose tissue in connection with a decrease in tissue temperature is discussed.