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



constitution of the tether, as compared with common matter in 

 bulk. There is observational reason to believo that the sun some- 

 times causes magnetic disturbances here of the ordinary kind. It 

 is impossible to attribute this to any amount of increased activity 

 of emission of the sun so long as we only think of common radia- 

 tion. But, bearing in mind the long waves of electroraagnetism, 

 and the constant speed, we see that disturbances from the sun 

 may be hundreds or thousands of miles long of one kind (i. e. 

 without alternation), and such waves, in passing the earth, would 

 cause magnetic " storms," by inducing currents in the earth's 

 crust and in telegraph-wires. Since common radiation is ascribed 

 to molecules, we must ascribe the great disturbances to move- 

 ments of large masses of matter. 



There is nothing in the abstract electromagnetic theory to indi- 

 cate whether the electric or the magnetic force is in the plane of 

 polarization, or rather, surface of polarization. But by taking a 

 concrete example, as the reflexion of light at the boundary of trans- 

 parent dielectrics, we get Fresnel's formula for the ratio of reflected 

 to incident wave, on the assumption that his " displacement " 

 coincides with the electric displacement ; and so prove that it is the 

 magnetic flux that is in the plane of polarization. 



Note B. 



I give these numerical examples : — 



Take a circuit 100 kilom. long, of 4 ohms and | microf. per 

 kilom. and no inductance in the first place, and also no leakage in 

 any case. Short-circuit at beginning A and end B. Introduce at 

 A a sinusoidal impressed force, and calculate the amplitude of the 

 current at B by the electrostatic theory. Let the ratio of the full 

 steady current to the amplitude of the sinusoidal current be p, 

 and let the frequency range through 4 octaves, from n=1250 to 

 n =20,000 ; the frequency being n/2ir. The values of p are 



1-723, 3-431, 10-49, 58-87, 778. 



It is barely credible that any kind of speaking would be possible, 

 owing to the extraordinarily rapid increase of attenuation with the 

 frequency. Little more than murmuring would result. 



Now make L=2 (very low indeed), L being inductance per 

 centim. Calculate by the combined electrostatic and magnetic 

 formula?. The corresponding figures are 



1-567, 2-649, 5-587, 10-496, 16-607. 



The change is marvellous. It is only by the preservation of the 

 currents of great frequency that good articulation is possible, and 

 we see that even a very little self-induction immensely improves 

 matters. There is no " dominant " frequency in telephony. What 



