Theory of Thunderstorm Electricity. 627 



the condensation layer it would of necessity traverse the 

 supersaturated stratum. Now Barus* has shown that when- 

 ever ions are formed in dust- free air there is produced at the 

 same time a large number of nuclei, which exist for a long- 

 time after the ions have disappeared. As water vapour is 

 very readily deposited on these nuclei fourfold supersaturation 

 is prevented in any air in which they are present, unless the 

 supersaturation is caused by exceedingly rapid rarefaction. 

 Each lightning discharge which passes through the super- 

 saturated layer will produce a large number of these efficient 

 nuclei ; and any considerable amount of electrical discharge 

 through the stratum would be the cause of the total distruc- 

 tion of the stratum. Thus we see that lightning discharges 

 would not act in the way Gerdien describes ; but, on the 

 contrary, would be a hindrance rather than a help to the 

 process on which the theory is based. 



We are therefore forced to the conclusion that the lightning 

 itself does not provide a source of ions for condensation to 

 take place upon, and that we must depend upon the ascending 

 currents of air to provide all the free ions for the separation 

 of the electricity at the condensation layer. 



Now we have already seen that under these circumstances 

 we could not get lightning discharges from any given area 

 of the storm more frequently than one during each hour. 

 It must, however, be pointed out that even this frequency is 

 far too high, for in order to discuss the most favourable 

 circumstances we have assumed conditions which could not 

 possibly take place in nature. We have assumed that every 

 ion formed in the air as it passes from the lower cloud through 

 the supersaturated stratum continues to exist, and that com- 

 plete separation of the ions takes place at the condensation 

 layer, neither of which conditions could be fulfilled. Also 

 we have neglected the whole loss of electricity due to any 

 precipitation from the condensation layer, and further we 

 have not taken into account the loss of electricity from the 

 condensation layer which would take place through conduc- 

 tion in the great fields necessary to produce a lightning 

 discharge. It would not perhaps be estimating these losses 

 too high if we assumed that they would increase the intervals 

 between two flashes ten times, and so make it only possible 

 to have one flash in ten hours. 



It thus appears that the theory does not account for the 

 rapid electrical discharge observed in thunderstorms, and 

 therefore in this respect it is unsatisfactory. 



* C. Barns, Amer. Journ. Sci. xix. pp. 349-356. May 1905 ; and 

 numerous later papers. 



2 U2 



