SCIENTIFIC PROBLEMS IN RADIOTELEGRAPHY 365 



Yet the length of these visible light waves bears the same re- 

 lation to the diameter of such a small sphere that radiotelegraphic 

 waves one kilometre in length bear to the diameter of the earth. 

 Hence it is by no means obvious that Transatlantic radio- 

 telegraphy is conducted in virtue of the diffraction of these long 

 waves. The matter has, however, been more carefully ex- 

 amined by mathematicians— by Lord Rayleigh, the late Prof. 

 Henri Poincare, Prof. H. M. Macdonald and Dr. Nicholson. 

 They have all come to the conclusion that radiotelegraphic 

 waves of even a kilometre or more in length cannot be 

 diffracted round the earth to an extent sufficient to account for 

 Mr. Marconi's long-distance wireless telegraphy. Hence our 

 fundamental problem is to find a valid reason for the pro- 

 pagation of these long electric waves in spite of the curvature 

 of the earth across the Atlantic or even, as achieved by 

 Mr. Marconi, from Ireland to South America, a distance of 

 6,000 miles or one-quarter of the way round the earth. A 

 mathematical discussion of the problem has made it tolerably 

 clear that if the earth were a ball of copper immersed only in 

 aether, no long-distance radiotelegraphy would be possible on 

 it. The waves generated at any place would soon glide off it 

 and be lost in space. The fact that we can conduct such 

 telegraphy on it over long distances is only due either to the 

 imperfect conductivity of the earth or to its possessing an 

 atmosphere of such a nature that electric waves created on it 

 are prevented by some means from rushing off it tangentially 

 into space. One explanation has been formulated by Prof. 

 A. Sommerfeld of Munich as the result of an elaborate 

 mathematical discussion of the problem of wave generation by 

 an oscillator at the boundary of two dielectrics of different kinds. 

 His conclusion is that, in the case of an oscillator at the 

 bounding surface of earth and air, there is not only an electro- 

 magnetic space wave radiated through the air but a surface 

 wave which travels along the bounding surface and is limited 

 to a small region on either side. There is a certain analogy 

 with a similar effect in the case of earthquakes, in which, as 

 investigation has shown, there are space waves travelling 

 through the earth and surface waves more or less confined to 

 the surface crust. Sommerfeld shows that these surface waves 

 would degrade in amplitude much less fast than the space 

 waves and would follow round the surface of the earth in spite 



