3i0 ATMOSPHERIC ELECTKICITY. 



the snow surface while the return was the sea. During the winter the ' speaking ' over the 

 wire was perfect, showing that the insulation was sufficient, but after the naiddle of November 

 no electrical signals could be sent through the wire showing that then the snow had lost 

 its insulating power. If the snow could not insulate for the low voltages and high current 

 density used in the telephone circuit during the summer, it certainly could not insulate for 

 the high potentials and small charge produced by friction. As the positive potential gradient 

 was as high during blizzards in the summer as in the winter, it is obvious that the charge 

 concerned is not on the ground. The only remaining place to look for the charge is the air 

 itself, and I beheve that the charge carried by the air is the solution of our problem. 



There can be no doubt from Rudge's work quoted above and from experiments made 

 by the writer that the cause of the electrification of dust when blowoi into the air is not 

 friction electricity as commonly understood, but an effect similar to that found when water 

 is splashed or water drops broken, in which the solid or liquid particles retain one kind of 

 electricity while the opposite electricity is given to the air, probably in the form of slowly 

 moving ions. Let us imagine an isolated cloud of ice particles which on account of the 

 turbulent motion of the air are constantly colliding with one another. The result would be 

 that the ice particles would be charged with one kind of electricity (the sign of which will 

 be discussed later) and the air would receive the opposite charge. There would, however, 

 be no exterior field produced, as the two charges would neutralise one another at an appre- 

 ciable distance from the cloud. If, however, the snow particles slowly settled in the air, the 

 charge on the snow would become concentrated in the lower half of the cloud and that on 

 the air in the upper haU:. This would produce an electrical field in the cloud itself. Now 

 imagine that the cloud comes into contact with, the ground, then every time an ice crystal 

 touches the ground its charge is lost, but the opposite charge remains in the air above. 

 The cloud in this way would soon have an excess of the electricity associated with the air 

 and in course of time might become very highly charged. It is now only necessary to 

 assume that the snow in the process becomes negatively charged and the air positively charged 

 to have a complete explanation of the normal relationship foimd between the drift and the 

 potential gradient. 



Let us consider first the case of sui-face drift. The term surface drift is never used when 

 the driven snow rises more than a few feet above the ground. The ice particles in this drift are 

 constantly colliding with one another and ^vith the fixed snow on the ground. According to 

 the theory every colUsion is accompanied by a separation of electricity, the ice particles 

 becoming negatively charged and the aii- receiving a positive charge. The snow, however, is 

 constantly coming into contact with the gromid when it gives up any charge it has, it 

 will probably become charged again at the instant of separation, but every contact ^\'ith the 

 ground while adding more positive electricity to the air adds no further charge to the snow. 

 Thus the air in which the drift is carried along becomes more and more highly charged 

 with positive electricity, without the snow retaining the corresponding quantity of negative 

 electricity. Owing to the irregularities of the air motion the charged air near the groimd 

 mixes with the air above and in a short time the whole of the air above the ground pro- 

 bably to several hundred feet becomes more or less highly charged with positive electricity. 

 Between this positively charged air and the ground an intense electrical field may be set up. 

 The direction of the field is the same as the normal field of the atmosphere and therefore 

 the efiect of the surface drift is to produce a high positive potential gradient. 



Let us now consider that the whole of the lower atmosphere becomes full of cbifting 

 snow to a great height. The collisions" of the snow particles produce the same separation of 

 electricity, but owing to the constant downward motion of the snow relatively to the air 



