WHITEHEAD— HIGH VOLTAGE CORONA IN AIR. 255 



upon by several observers in recent years, is usually stated in the 

 form 



in which E is the critical or corona forming voltage gradent at the 

 surface of the wire expressed in kilovolts per centimeter, r is the 

 radius of the wire in centimeters, and 8 the relative density of the 

 air, having: the value 



p being the pressure is centimeters of mercury, and t the tempera- 

 ture in degrees Centigrade. 



A more convenient form of the relation for our purpose is 



E B' 



which gives a linear relation of E8 and i/V^^'. 



The above relatively simple relations have now been corroborated 

 by a number of observers with fairly close agreement as to the value 

 of A and B. The form of the law is the same for both continuous 

 voltages and crest values of alternating voltages. With continuous 

 voltage, however, there are appreciable differences in the values 

 of the constants A and B as between positive and negative corona 

 forming wire, the form of the law in each case remaining the same. 



No attempt will be made in this place to give a complete review 

 of the large number of observations which have been taken on ten 

 different sizes of corona forming rod under wide variations of rela- 

 tive density 8. Table i, however, gives several sets of readings and 

 indicates, particularly in columns 7 and 8, the accuracy with which 

 the observations repeat themselves. Column g gives the readings of 

 galvanometer G^ measuring the condenser charging current, and 

 column 14 is the reading of an ordinary alternating voltmeter con- 

 nected to the low voltage terminals of the high voltage transformer. 

 The readings of this voltmeter were not used in the calculations, but 

 its indications provided at all times a convenient means for de- 



