366 SCIENCE PROGRESS 



of curvature and irregularities. His suggestion is that long- 

 distance wireless telegraphy is chiefly effected by such surface 

 waves. Although his analysis is no doubt valid, yet neverthe- 

 less the trend of experimental evidence seems to be against it. 



If Sommerfeld's explanation were the true one, it is hard 

 to see why long-distance wireless telegraphy should either be so 

 much affected by daylight and by direction or exhibit the abnor- 

 malities with regard to wave length which it actually does. 



There remains then one other assumption for which the 

 evidence is far from complete but which has been the ultimate 

 refuge of all those who have found it impossible to account for 

 the facts either on the basis of diffraction or on the hypothesis of 

 a surface wave. This assumption is that the upper levels of the 

 earth's atmosphere, say at a height of 60 to 100 miles, are 

 perpetually in a state of ionisation to such a degree as to render 

 it a fairly good conductor, possibly as good as dilute sulphuric 

 acid, at any rate sufficiently good to enable it to act as a 

 reflector for long electric waves. 



Since pure wave diffraction is excluded as a possible 

 explanation of long-distance wireless telegraphy, we have to 

 fall back on one or other of two hypotheses, one of which 

 requires the production of surface waves in the crust of the 

 earth and the other refraction or reflection of waves in or by 

 the earth's atmosphere. As will be seen by what follows, the 

 abnormalities and irregularities of long-distance radiotelegraphy 

 seem to point to the determining cause of the bending of the 

 waves round the earth being something in the atmosphere 

 rather than in the earth. Yet, on the other hand, the existence 

 of surface waves is not disproved and there are some facts 

 which rather strongly support this latter supposition. The 

 experimental achievements and chief practical experience with 

 long electromagnetic waves which have to be explained are 

 therefore as follows : electric waves from 1 to 4 miles in 

 wave length can travel at least one quarter of the way round 

 the earth. This suggests at once the questions — Could they 

 travel half-way round ? Is wireless telegraphy between London 

 and New Zealand within the possibility of practice ? In the 

 next place, there are great differences in the reduction of 

 amplitude experienced by such a wave when travelling by day 

 and by night over long distances ; and certain extraordinary 

 variations in the strength of signals near sunrise and sunset. 



