There is at present no absolute standard method of measurement by which the high voltages now in common use may be compared. The paper gives an account of such a method in which the polarization of a metallic ellipsoid of revolution, suspended by an insulating fibre between vertical circular pole-plates, is observed. The dimensions and proportions of the ellipsoid and of the horizontal field were determined with high accuracy, and the magnitude of the electric gradient on the axis of symmetry of the instrument was calculated for given applied voltages. The working formula of the instrument isF=K(n2−n20)frac12;, whereFis the field strength,ka constant found from the dimensions,nthe number of swings per second of the ellipsoid with the field on andn0with it off. The influence of the spacing of the plates, of shielding the ellipsoid, of gas pressure and temperature, was examined, and the readings of the instrument were compared with Whitehead and Castellain's sphere-gapdeterminations of voltage. As a combined check on the accuracy of their readings and those of this instrument, a value of the ratiocof the electromagnetic to the electrostatic units of quantity is obtained which agrees with the best previous results. It is concluded that continuous measurements of voltages up to 200kV can be made with the precautions indicated in the paper, to an accuracy of at least 3 parts in 1 000, the constant of the instrument depending upon the value accepted for the ratioc