140 



NATURE 



[February 2, 1922 



Some Problems of Long-distance Radio-telegraphy.^ 



By Dr. J. A. Fleming, F.R.S. 



I. 



THE achievement of transatlantic radio-telegraphy 

 in 1 90 1 and 1902 was of interest to physicists 

 chiefly by reason of the fact that they did not see 

 clearly why it should have been possible at all. 

 The mystery of it was increased when at later dates 

 radio signals were transmitted a quarter of the way 

 round the world, and finally, with high-power 

 stations and thermionic valves detected even at the 

 Antipodes. 



The. wave-length of the waves used in the earliest 

 work at Poldhu was about 3000 ft. The earth, 

 roughly speaking, is a sphere 42 million feet in 

 diameter. Hence the ratio of wave-length then 

 used to earth diameter was about i : 14,000. 



In the case of light there is a small bending or 

 diffraction of the wave round an opaque obstacle. 

 In other words, there is some small amount of 

 illumination within the boundary of the geometrical 

 shadow. The average wave-length of light waves 

 is about I /2000th of a millimetre, and a sphere 

 having a diameter of 7 mm. would be 14,000 of 

 such wave-lengths. Now if an exceedingly small 

 source of light were placed on the pole of a sphere 

 7 mm. in diameter in a dark region, it is certain 

 that there would be no illumination at the equator of 

 the sphere. In other words, there would not be 

 any sensible diffraction at an angular distance of 

 90°. Modern long-distance radio-telegraphy con- 

 ducted with waves of wave-length approximating to 

 10 miles or so can communicate even with the 

 Antipodes. 



The mathematical treatment of the problem of 

 the diffraction round a conducting sphere of electric 

 waves which are radiated from a transmitter at its 

 pole consists in expressing the magnetic and electric 

 forces at any angular distance e in the form of a 

 series of harmonic terms. It is in the summation 

 of this series to obtain the integral effect at the 

 receiver that the chief difficulties and differences of 

 opinion occur, and most analysts have employed only 

 an approximation. In 191 8 Prof. G. N.' Watson 

 effected a new and complete summation which en- 

 ables the value of the forces to be calculated for 

 any point on an imperfectly conducting sphere. 



The result of eighteen years' Avork on this problem 

 by mathematicians of the highest rank has been 

 to give us a formula for the current in a receiving 

 aerial of given resistance determined in terms of 

 wave-length, aerial heights, and distance which repre- 

 sents the result of pure diffraction acting round a 

 spherical earth of perfect conductivity. On the 

 other hand, when we come to compare the results 

 of this diffraction formula with actual observations 

 in practice we find an enormous discrepancy. The 

 actual received currents in the case of long-distance 



1 Abridged from two section<; of the Trueman Wood lecture on "The 

 Coming of 'Vge of Long-distance Radio-telegraphy and Some of its Scientific 

 Problems," delivered before the Royal Society of Arts on November 23, 

 iq2i. The complete lecture is published in the Journal of the Society for 

 December 9 and 16, 1921. 



NO. 2727, VOL. 109] 



stations for the usual sending aerial currents are 

 hundreds of thousands, or even millions, of times 

 greater than the received current predicted by the 

 theoretical formula. Thus, to take a case quoted 

 by Dr. Van der Pol from observations made at 

 Darien Radio Station, on the Panama Canal, on 

 radio signals sent from Nauen, near Berlin, Dr. 

 L. W. Austin gives the following figures : 

 Ii=i50 amperes, A = 9-4 km., a-Ji'^—\2o metres, 

 ^2/^2- 146 metres, R2=29 ohms, and ^ = 9400 km. 

 Now the actual received current was l2=i'3 

 microamperes, but the value predetermined by 

 the formula is only o-6 of one millionth of a 

 microampere. In other words, the actual received 

 current in this case is two million times greater than 

 the predicted current. 



The upshot of the whole matter then is this : 

 Long-distance radio-telegraphy, say, round one- 

 quarter of the circumference of the earth, would 

 certainly be quite impossible but for some cause, 

 other than diffraction, operating to compel the waves 

 to follow round the earth's curvature and not 

 quickly glide off it. 



Oliver Heaviside in 1900 suggested that an upper 

 conductive layer on the atmosphere might act as a 

 guide to the waves, radio-telegraphy being, in fact, 

 conducted in a thin spherical shell of non-conductive 

 air bounded by a conductive earth and a conductive 

 upper air. He did not furnish any valid reasons 

 to explain why this upper air conducts and how its 

 conductivity is preserved, and although the sug- 

 gestion has been very generally accepted by radio 

 engiineers, it has been taken without sufficient 

 criticism of its difficulties and details. There has 

 been in the intervening twenty -one years an immense 

 accumulation of facts, all showing, however, that 

 long-distance radio-telegraphy is conditioned by the 

 physical constitution of our atmosphere and is very 

 far removed indeed from being simple electro- 

 magnetic wave propagation in empty space. An 

 important epoch in this connection is the year 1902, 

 when Senatore Marconi discovered during one of 

 his early voyages across the Atlantic in the 

 s.s. Philadelfhia in February, 1902, that radio 

 signals from Poldhu could be received at night 

 about thrice the distance they could be read in day- 

 time, being detectable only up to 700 miles by day, 

 but readable up to 2099 miles by night. 



It was at once surmised that the difference was 

 due to ionisation of the air by sunlight, which, by 

 liberating electrons from atoms, gives to the air con- 

 ductivity. It was some years before this vague 

 suggestion was converted by Dr. W. H. Eccles into 

 a more definite 'scientific theory, many speculations 

 in the meantime being found wanting in adequate 

 basis, such as that which regarded the sunlit air as 

 having an absorption for the energy of electro- 

 magnetic waves similar to that of foggy or mi.sty 

 air for visible light. 



