Data have been obtained on the sparkover voltage of the two‐centimeter sphere gap set at a gap of 4 mm (10.3‐kv peak) showing that: (1) The effect of irradiation is to reduce the sparkover voltage by approximately 2.5 percent from its unirradiated value and to decrease the scattering of individual observations by a factor of about four. This irradiation effect is readily saturated by an open, coredcarbon, arc at 50 cm from the gap. (2) The effect of humidity, which is apparently independent of the sphere material for the five metals used here, is to increase the sparkover voltage by +0.13 percent per mm (of mercury) increase in vapor pressure of the water in the atmosphere. (3) There seems to be no choice between the metals used for spheres (aluminum, brass, chromium, nickel and steel) so far as repeatability of results is concerned, the probable error of a day's results averaging ±0.28 percent. This probable error can be largely if not wholly accounted for in terms of known sources of uncertainty. (4) The final sparkover voltages for the various metals, even when corrected to the same humidity, differ from each other by far more than can be accounted for by any definitely recognized source of uncertainty.