In this paper, the authors report on a system of indicating shaft-horsepower with a milliwattmeter to which a magneto-striction type torsionmeter and an electric tacho-meter are connected. The magneto-striction type torsionmeter has been investigated and trially manufactured since 1947, and was tested at the four vessels' sea trials by attaching it to the intermediate shafts. But the reason was not understood at that time, why the results obtained with this torsionmeter have shown the lower values for the range of low speed and higher values for the high speed than those obtained simultaneously with the torsionmeter of Togino's type and of the Hopkinson-Thring type, and the percentage errors were not negligible at the time of the progressive speed trials of the tested three vessels, while practically successful results had been already obtained in the case of the “Seikan-Maru. No. 11”. But in this paper, the authors give a solution to this problem after the careful experiments on the “Nichirei-Maru” and the “Gekko-Maru” that the above-mentioned erros were mostly due to the temperature effect on the CuO rectifier in the secondary circuit of the pick-up of the magneto-striction type. This temperature effect can be compensated thoroughly or made smaller enough to be neglected. The electric tachometer which has shown a superior linearity in its revolution-voltage characteristics, is a D. C. generator with permanent magnets. For the milli-wattmeter as the shaft-horsepower indicater, two meters have been trially manufactured, one of without core type and another of with core type, and the latter has shown better quality. Last section of this paper deals with some examples of the results of measurement, obtained with the above-mentioned instruments attached to the intermediate shafts of the “Nichirei-Maru” and the “Gekko-Maru”, at the official speed trials, turning trials and the backing trials.