AS a source of potential for the production of supervoltage x-rays, a Van de Graaf generator has been in use at the Collis P. Huntington Memorial Hospital for some months, and the present paper is a report of some of the measurements made on the radiation secured from a roentgen tube actuated by this generator. A description of the machine itself has been given previously (4, 12) and need not be repeated in detail. The roentgen tube is in a vertical position with the water-cooled target at ground potential. Originally, the target was a leadplated sheet of copper inclined at about 38° to the horizontal so that radiation could be examined in the vertical direction, that is, at an angle of 0° with the electron beam, or in the horizontal direction at an angle of 90° with the electron beam. This arrangement was soon modified, however, by replacing the 38° lead target by a gold-plated horizontal target so that at present only the vertical radiation is available for experimental purposes. The method of measuring the voltage, as has been described previously (12), consists in measuring the current through a high resistance in parallel with the tube. This resistance had been calibrated at low voltages and also at higher voltages with spark gap and generating voltmeter. For short runs this was quite satisfactory, but with prolonged running there was apparently an increase in voltage, though the x-ray output remained constant. This indicated a decrease in value of the high resistance in parallel with the tube. Therefore, a corona meter was installed as a check on the constancy of voltage and all voltage determinations were made with the high resistance before any prolonged running. As a further check on output, a parallel-plate ionization chamber was connected to the input of a direct current amplifier and the output read on a five micro-ampere meter. This apparatus was made by Mr. B. J. Cosman. The ionization chamber is placed permanently in the x-ray beam and serves merely as an indicator of constancy of x-ray output. By carefully maintaining the corona meter current, x-ray tube current, and x-ray output meter reading constant, one can secure a very stable and satisfactory source of radiation. Of course, maintaining any two of these readings constant should be sufficient, but three serve as a better check. Maintaining the x-ray output meter constant is by no means sufficient, since a decrease in voltage may be compensated for by a higher current. Consequently, during physical measurements the voltage was determined at the very beginning of the run by means of the current through the high resistance and then kept as nearly constant as possible by reading the corona current, tube current, and x-ray output. The voltage measurements are considered accurate to ± 2 per cent.