We performed an in vitro study to determine the thermal safety of a domestic microwave to warm intravenous crystalloid solutions. Five-hundred-millilitre bags of crystalloid, randomly allocated to groups which differed in power setting, timer setting and whether or not agitation was performed after warming, were heated in a microwave oven to a calculated temperature of 39 degrees C. Timer accuracy was checked by stopwatch. Bag temperature was measured using an infrared tympanic temperature probe and fluid temperature was measured with an in-line thermocouple. Mean times measured by stopwatch were higher than set. No in-line temperatures reached 40 degrees C. Wider overall ranges and a higher mean were found with the tympanic probe compared with in-line temperature measurement. There were significant differences between the in-line temperatures of shaken and unshaken bags at each power setting, but not when groups were added together. There was no change in colour or odour of bags or fluid. One bag developed a pinhole leak when the packaging was removed.