same laws as have been observed in these experiments the buildings must be largely 

 protected. 



But, you say, in a real thunderstunn the clouds and earth are much larger, 

 compared with the real buildings, than the cloud and earth in this experiment, and 

 the electric charge is infinitely greater than that produced by this small machine. 

 Quite so, but the real cloud and earth are 20,000 times as far apart as the cloud 

 and earth of this machine, and this great distance requires the entire lightning 

 charge to break through the atmosphere. Hence, anything that would relieve the 

 electric pressure just a little bit would prevent a flash. Lightning rods must relieve 

 the pressure if lightning is produced by electricity. Consequently, we expect that 

 real lightning rods in a real storm would protect the buildings just as well as these 

 little ones do in these experiments. 



Why Stkokes Sometimes Occur in Spite of Rods. 



But, one is disposed to ask, if rods in many cases prevent strokes, why not in 

 all ? In the foregoing experiment? the cloud and earth have in all cases been 



Fig. 32. — But if the upper end of the ground wire projects above the fence in a 

 point, no flash occurs, which is even better. 



charged gradually, and when so charged flashes between them can be prevented. 

 However, the machine can be modified to unload its entire positive and negative 

 charges to the cloud- and earth-plates in possibly one-thousandth or one-millionth 

 part of a second. When this occurs the points cannot let the charge leak off fast 

 enough and consequently a flash follows. This shows that it is possible for a rodded 

 building in a real thunderstorm to be struck. Generally, real thunderclouds become 

 r-harged gradually and lightning strokes can be prevented, but in rare cases a cloud 

 receives an immense cliarge in a very small fraction of a second. This is some- 

 times broiight al)out in tlio following wav : Two clouds lie near each other. The 

 buildings under the first are rodded. but not under the second. A flash occurs 

 between the second and the ground, followed by a flash between the first cloud and 

 the second. This impulsive rush changes the pressure in the first cloud so suddenly 

 that the leakage does not neutralize the charge as fast as the difference in pressure 

 increases, hence a flash occurs. But that flash is probably much weaker than if the 

 rods had not been there, and in most cases of this kind tlie rods will conduct the 

 current to the ground without damage. 



