Forced Vibrations of Electromagnetic Systems. 205 



but in really good telephones of the best kind, with very small 

 time-constants, it is not great. We see therefore that tele- 

 phony, so far as the electrical part of the matter is concerned, 

 can be made as nearly perfect as possible on lines of thousands 

 of miles in length, But the distortion that is left, due to 

 imperfect translation of sound-waves into electromagnetic 

 waves at the sending-end, and the reproduction of sound- 

 waves at the receiving-end, is still very great ; though, prac- 

 tically, any fairly good telephonic speech is a sufficiently 

 good imitation of the human voice. 



There is one other way of increasing the inductance which 

 I have described, viz. in the case of covered wires to use a 

 dielectric impregnated with iron dust. I have proved expe- 

 rimentally that L can be multiplied several times in this way 

 without any increase of resistance ; and the figures I have 

 given above (in Note B) prove what a wonderful difference 

 the self-induction makes, even in a cable, if the frequency is 

 great. Hence, if this method could be made practical, it 

 would greatly increase the distance of telephony through 

 cables. 



Now, passing to iron wires, the case is entirely different, 

 on account of the great increase in resistance that the sub- 

 stitution of iron for copper of the same size causes, which 

 increases P and the attenuation. Taking for simplicity the 

 very extreme case of such an excessive frequency as to make 

 the formula 



nearly true, R being the steady and R' the actual resistance, 

 we see that increasing either Ror /i increases R/ and there- 

 fore P, because LV tends to the value L v. Thus the carrying- 

 power of iron is not greatly above, but greatly below that of 

 copper of the same size. 



I have, however, pointed out a possible way of utilizing 

 iron (other than that above mentioned), viz. to cover a bundle 

 of fine iron wires with a copper sheath. The sheath is to 

 secure plenty of conductance ; the division of the iron to faci- 

 litate the penetration of current, and so lower the resistance 

 still more, to the greatest extent, whilst at the same time 

 increasing the inductance. But the theory is difficult, and it 

 is doubtful whether this method is even theoretically legiti- 

 mate. First class results were obtained by Van Rysselberghe 

 on a 1000-mile circuit in America (2000 miles of wire), using 

 copper-covered steel wire. Here the resistance was very low, 

 on account of the copper, and the inductance considerable, on 

 account of the dielectric alone ; so that there is no certain 



