To the Editor.— In a recent article on hypothermia in the elderly, Keilson et al 1 describe a method for measuring body temperature from the temperature of freshly voided urine. After finding a poor correlation with oral temperature, they conclude that "urine temperature does not measure core temperature values." This conclusion does not appear justified on the basis of the data provided, and perhaps the authors should have considered that the method they used to measure urinary temperature does not accurately measure the temperature of freshly voided urine. I have employed a newly developed urinary thermometer for ongoing clinical studies on temperature regulation and initially performed a number of experiments to validate the accuracy of this instrument. 2 This thermometer registers peak urinary temperature within 45 s of immersion with a minimum volume of 20 mL of urine. In contrast to Keilson and co-workers' studies, the correlation of this thermometer