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



then determined by micrometer measurements of ordinates taken 

 every 7.5 degrees over two half waves. The several values of this 

 ratio so obtained revealed a minimum at 55 cycles thus explaining 

 the rise in the lower curves of Fig. 5 at that frequency. In the 

 upper curves the points indicated are the voltage of the lower curve 

 multiplied by the ratio of maximum to effective value as calculated 

 from measurements of the oscillograms for the corresponding 

 frequencies. 



The upper corrected curves of Fig. 5 show a lowering of the 

 critical voltage with increasing frequency. The result leaves some- 

 thing to be desired in the accuracy of location of the points upon the 

 curve. It should be noted however that owing to the magnifi- 

 cation of the scale, the error of the points off the upper curves and 

 the 25-cycle portion of the lower curves is only about i per cent. 

 Several other sets of observations for different sizes of wire reveal 

 curves of the same general characteristics. The measurement of the 

 ratio of maximum to effective value from an oscillogram is subject 

 to considerable error. The maximum at 55 cycles, however, on 

 the lower curves is brought below the values for lower frequencies 

 when the correcting factor is introduced, and particularly, the lower- 

 ing at 91 cycles is far too great to be questioned on the score of a 

 possible error of this nature. The curves therefore show with a 

 fair accuracy the nature of the variation of the critical voltage with 

 the frequency. This variation within the range of the present com- 

 mercial frequencies 25 to 60 cycles per second, is only about 2 

 per cent. 



Influncc of Pressure. — The influence of pressure on the various 

 forms of spark discharge has been closely studied. Paschen's^ law 

 states that the sparking potential for a given spark length is directly 

 proportional to the pressure ; his investigations covered the range 

 of pressure between 10 and 75 cm. of mercury. Carr* has shown 

 that this linear relation extends down to pressures of a few milli- 

 meters if the spark lengths are not greater than i cm. but does not 

 obtain for lower pressures. Townsend^ has shown that the potential 



'Paschen, Wied. Ann., XXXVII., 79, 1889. 

 * Carr, Proc. Roy. Soc, LXXI., 374, 1903. 

 °Townsend, Phil. Mag., VI., i, 198, 1901. 



