IN the Elektrotechnische Zeitschrift for Aug. 13, L. Pungs and H. Vogler describe a new electro-optical method of measuring voltages of very high frequencies. In practice, some type of electrometer is generally employed, but with very high frequencies its readings are untrustworthy. The authors made a test on a good electrometer of a type usually employed, and found that up to one and a half million cycles per second it gave accurate and consistent results, but at twelve million cycles its readings were incorrect by errors lying between thirty and forty per cent. As these instruments are largely used in radio work for making investigations on the dielectric behaviour of insulating materials, it was necessary to find the cause of the inaccuracies and either correct their design or try to invent a new method. The errors of the instrument were found to be mainly due to the fact that its capacitance varied with its deflexion, for this led to resonance effects which made the readings quite irregular.