MODERN VIEWS WITH RESPECT TO ELECTRIC CURRENTS 663 



attention to the application of these results to lightning rods. Almost 

 since the time of Franklin there have been those who advocated the 

 making of lightning rods hollow, to increase the surface for a given 

 amount of copper. We now know that these persons had no reason for 

 their belief, as they simply drew the inference from the fact that elec- 

 tricity at best is on the surface. Neither were the advocates of the solid 

 rods quite correct, for they reasoned from the fact that electricity in a 

 state of steady flow occupies the whole area of the conductor equally. 

 The true theory, we now know, indicates that neither party was entirely 

 correct and that the surface is a very important factor in the case of a 

 current of electricity so sudden as that from a lightning discharge. But 

 increase of surface can best be obtained by multiplying the number of 

 conductors, rather than making them flat or hollow; and, at the same 

 time, Maxwell's principle of enclosing the building within a cage can. be 

 carried out. Theory indicates that the current penetrates only one- 

 tenth the distance into iron that it does into copper. As the iron has 

 seven times the resistance of copper, we should need 70 times the sur- 

 face of iron that we should of copper. Hence I prefer copper wire 

 about a quarter of an inch diameter and nailed directly to the house 

 without insulators, and passing down the four corners, around the eaves 

 and over the roof, for giving protection from lightning in all cases where 

 a metal roof and metal down spouts do not accomplish the same purpose. 

 Whether the discharge of lightning is oscillatory or not does not enter 

 into the question, provided it is only sufficiently sudden. I have re- 

 cently solved the mathematical problem of the electric oscillations along 

 a perfectly conducting wire joining two infinite and perfectly conducting 

 planes parallel to each other, and find that there is no definite time of 

 oscillation, but that the system is capable of vibrating in 'any time in 

 which it is originally started. The case of lightning between a cloud of 

 limited extent and the earth along a path through the air of great re- 

 sistance is a very different problem. Both the cloud and the path of the 

 electricity are poor conductors, which tends to lengthen the time. If I 

 were called on to estimate as nearly as possible what took place in a flash 

 of lightning, I would say that I did not believe that the discharge was 

 always oscillating, but more often consisted of one or more streams of 

 electricity at intervals of a small fraction of a second, each one continu- 

 ing for not less than 1 0*0 o o second. An oscillating current with 100,000 

 reversals per second would pentetrate about ^ inch into copper and ^-J-g- 

 inch into iron. The depth for copper would constitute a considerable 

 portion of a wire inch diameter, and, as there are other considerations 



