Limited data is available on the application of post-mortem temperature methods to non-standard conditions, especially in problematic real life cases in which the body of the victim is found in cold water environment. Here we present our experience on two cases with known post-mortem times. A 14-year-old girl (rectal temperature 15.5 degrees C) was found assaulted and drowned after a rainy cold night (+5 degrees C) in wet clothing (four layers) at the bottom of a shallow ditch, lying in non-flowing water. The post-mortem time turned out to be 15-16 h. Four days later, at the same time in the morning, after a cold (+/- 0 degrees C) night, a young man (rectal temperature 10.8 degrees C) was found drowned in a shallow cold drain (+4 degrees C) wearing similar clothing (four layers) and being exposed to almost similar environmental and weather conditions, except of flow (7.7 l/s or 0.3 m/s) in the drain. The post-mortem time was deduced to be 10-12 h. We tested the applicability of five practical methods to estimate time of death. Henssge's temperature-time of death nomogram method with correction factors was the most versatile and gave also most accurate results, although there is limited data on choosing of correction factors. In the first case, the right correction factor was close to 1.0 (recommended 1.1-1.2), suggesting that wet clothing acted like dry clothing in slowing down body cooling. In the second case, the right correction factor was between 0.3 and 0.5, similar to the recommended 0.35 for naked bodies in flowing water.