The workload for two different methods of floor mopping in 11 healthy female cleaners was evaluated by rating the perceived exertion, by recording the ECG and EMG and by video analysis of postures and movements. There was less perceived exertion and local muscle loading during work when the 'push' method was used than with the 'figure-of-eight' method. The 'push' method nevertheless still implied a level of static loading that may be harmful during long term work. The methods did not differ regarding measured heart rate and oxygen consumption. However, for two cleaners the 'figure-of-eight' method and for three cleaners the 'push' method implied a work load corresponding to more than 40% of their normal oxygen uptake. Large individual variations in work postures and movements were found for both methods.