WHITEHEAD— HIGH VOLTAGE CORONA IN AIR. 247 



send has proposed an interesting explanation of the fundamental 

 law of corona formation in terms of the theory of ionization by col- 

 lision, but this theory is based entirely on experiments at very low 

 pressures, quite outside the range in which the first definitely marked 

 corona phenomena lie. So far as the writer is aware, Townsend's 

 is the only attempt to coordinate corona phenomena with modern 

 physical theory. 



In two particular aspects the corona presents difficulties for the 

 electrcial engineer : first, the presence of corona is always accom- 

 panied by a loss or waste of energy ; and second, in the presence of 

 corona insulation deteriorates rapidly and the insulating properties 

 of the neighboring air are lowered. In either case the probability of 

 sparkover and shortcircuit is greatly increased. 



The energy loss attendant upon corona introduces important 

 limitations in the design of long distance electric transmission lines. 

 The voltage of such lines is made high in order to reduce the 

 magnitude of the current to be carried, and therefore the size of the 

 conductors, the cost of the conductors being a principal item in the 

 total cost of the transmission line. The tendency therefore is to 

 higher and higher values of voltage. As the voltage is increased, 

 however, the corona forming point is reached, and in order to pre- 

 vent the presence of corona the distance between the transmission 

 wires must be increased. 



It is especially important to keep the voltage well below the 

 corona forming value, for above it the loss increases very rapidly. 

 According to F. W. Peek, from measurements made on a section of 

 a modern transmission line, the power loss may be expressed 



tv = kf{e~e,y (i) 



where iv is in watts, / is the frequency, e is the maximum value of 

 the voltage and e^ the voltage at which corona first begins ; that is 

 to say, the loss increases with the square of the excess of voltage 

 above the corona forming value. (See Fig. 2.) 



The above formula applies to alternating voltage. No complete 

 study has been made of the loss at continuous voltage, probably 

 because little use is made of high continuous voltage in the field of 



