This investigation was devoted to a study of the starting potentials of gases of commercial purity for frequencies less than one million cycles per second. The gases used were Hydrogen, Nitrogen, Helium, and Argon. The discharge tube was of spherical design having spherical aluminum electrodes three-fourths inch to diameter. Five gap distances were used ranging from 10 mm to 50 mm;The results of the investigation show that: (1) The effect of impurities in the gas pointed out by J. Thomson were not observed. (2) The usual straight line portion of the curves of Vs versus pressure were observed to be concave downward. An explanation for this was not obtained. A general form for the curve was shown to be: Vs=Ax1/2+Bx 3/2 where x is a number proportional to the mean free path of an electron. (3) The three stages of the discharge as observed by J. Thomson were also observed. This phenomenon occurred in all the gases at a frequency of one million cycles per second only. (4) The relations connecting the starting potential, the pressure and the gap distance for plane parallel electrodes do not hold for spherical electrodes. At the larger gap distances the variation of the starting potential with the gap distance was quite unreliable. As the gap distance became less the reliability increased, also, giving a strong indication that the conditions were rapidly approaching those of the plane parallel electrodes. (5) The slope of the log p versus log d curve is independent of frequency and also of the gas for gap distances less than approximately 32 mm. This suggests that the expressions representing the conditions in the discharge are independent of the gas and of the frequency for plane parallel electrodes.