392 WHITEHEAD— HIGH VOLTAGE CORONA IN AIR. [April 21, 



arc! conditions of temperature and pressure, is a simple function of 

 the diameter of the conductor. The relation between pressure 

 within the range of the atmosphere, and critical voltage, for a given 

 size wire, is linear ; and although the slope of the linear relation 

 changes with the size of wire there is good reason to suppose that 

 a simple law connecting them can be found. Aluch the same may be 

 said of the influence of temperature; preliminary experiments show- 

 ing that the linear relation exists over a fairly wide range. The 

 effect of stranding the conductor has been studied for only one size 

 of strand as yet, but it seems a simple matter, with some further 

 investigation, to express the effect of each of these influences in 

 terms of the diameter of the conductor. 



The influence of the frequency does not offer promise of expres- 

 sion as a simple relation ; this influence is small however within the 

 limits of frequency met in practice. The state of the atmosphere 

 appears to be of small importance, for moisture does not influence 

 the critical voltage, nor does its state as regards ionization, as is 

 indicated by several considerations given in a later paragraph. Dirt 

 and impurities which on settling cause irregularities on the surface 

 of the wire, may lead to localized brush discharges; and if these 

 are sufficient in number they may cause a noticeable loss below the 

 normal critical voltage. 



It is of great interest, however, to consider the results in their 

 relation to present theories of the nature of the electric conductivity 

 and breakdown of a gas. It is assumed that the reader is familiar 

 with the general features of the theory of ionization. Under this 

 theory the neutral atoms and molecules of matter may be separated 

 into smaller charged particles, and the motion of these particles under 

 electric force constitutes an electric current. In a gas there are 

 always a small number of these free ions present ; this number may 

 be greatly augmented by Rontgen rays, ultra-violet light and other 

 well known ionizing agents. When so ionized currents of magni- 

 tudes within easy measuring range are obtained between terminals 

 subject to a difference of potential. If this difference of potential is 

 increased, a point is reached where the current increases sharply, 

 showing the presence of some new saurce of ionization. The theory 



