144 Mr. 0. Heaviside on Electromagnetic Waves, and the 



packed closely together, with dielectric between them, forming 

 a composite dielectric conductor, and let the outermost sheet 

 be struck flush by a plane wave as above. The first sheet 

 transmits ^Hj, the second ^Hx, the third ^ H,, and so on. 

 This refers to the front of the wave, going into the composite 

 conductor at speed v. It is only necessary to go a very short 

 distance to attenuate the front, of the wave to nothing ; the 

 immense speed of propagation does not result in producing 

 any sensible immediate effect at a distance, which comes on 

 quite slowly as the complex result of all the internal re- 

 flexions and transmissions between and at the sheets. Observe 

 that there is an initial accumulation of H, so to speak, at the 

 boundary of the conductor, due to the reflexion. [Example : 

 the current-density may be greater at the outermost layer* 

 of a round wire when the current is started in it than the final 

 value, and the total current in the wire increases faster than 

 if it were constrained to be uniformly distributed.] 



Thus a good conductor may have very considerable per- 

 mittivity, much greater than that of air, and yet show no 

 signs of it, on account of the extraordinary attenuation pro- 

 duced by the conductivity. Now this is rather important 

 from the theoretical point of view. It is commonly assumed 

 that good conductors, e.g. metals, are not dielectrics at all. 

 This makes the speed of propagation of disturbances through 

 them infinitely great. Such a hypothesis, however, should 

 have no place in a rational theory, professing to represent 

 transmission in time by stresses in a medium occupying the 

 space between molecules of gross matter. But by admitting 

 that not only bad conductors, but all conductors, are also 

 dielectrics, we do away with the absurdity of infinitely rapid 

 action through infinite distances in no time at all, and make 

 the method of propagation, although it practically differs so 

 greatly from that in a nonconducting dielectric, be yet 

 fundamentally the same, with its characteristic features masked 

 by repeated internal reflexions with loss of energy. We need 

 not take any account of the electric displacement in actual 

 reckonings of the magnitude of the effects which can be 

 observed in the case of good conductors, but it is surely a 

 mistake to overlook it when it is the nature of the actions 

 involved that is in question. (See Note C, p. 153). 



Why conductors act as reflectors is quite another question, 

 which can only be answered speculatively. If molecules are 

 perfect conductors, they are perfect reflectors, and if they 

 were packed quite closely, we should nearly have a perfect 

 conductor in mass, impenetrable by magnetic induction ; and 

 * "On the S. I. of Wires/' Part I. Phil. Mag. August 1886. 



