486 BELL SYSTEM TECHNICAL JOURNAL 



wave-swept aether as though there were nothing else in the world 

 but itself and the waves. If, however, the electron were occasionally 

 to strike and bounce off a molecule it would leave some (even though 

 but a small part) of its kinetic energy behind, and this would have 

 to be replenished by the waves. Indeed as soon as collisions are 

 taken into account, the algebra tells us that the electron-current is 

 no longer in perfect quadrature with the electric vector. Dissipation 

 of energy occurs together with reflection, and if it is sufficiently great 

 — if, that is to say, the electrons collide often enough with molecules — 

 it takes the place of reflection. In the terms of my acoustic simile 

 on an earlier page, the signals are no longer echoed back from the hard 

 slopes of the mountain-range of Fig. ZA, but are swallowed up and 

 lost as if in something soft and woolly. 



One would expect absorption to occur in the lower reaches of the 

 ionosphere rather than in the upper, since the air is denser there and 

 the electrons suffer many more collisions. The absorbing layer, that 

 is to say, must be situated just where it is able to cut us off from the 

 reflecting layers by dissipating the signals which we send. Why 

 should it do so occasionally with such completeness, and most of the 

 time not do so at all, for any except the lowest frequencies? 



For an answer to this query, one looks again to the sun. Ordinarily, 

 we will suppose, the ionizing agent coming from the sun penetrates 

 deep enough into the air to form the reflecting layers high overhead, 

 but is nearly consumed in so doing. Occasionally, though, the sun 

 sends forth a quite abnormal transitory burst of radiation, so strangely 

 constituted that it passes the reflecting layers without contributing to 

 them or weakening itself, and continues so far down that at the level 

 where it at last engenders free electrons they constitute a layer ab- 

 sorptive and not reflective. This would be no more than an ad hoc 

 assumption, were it not that brilliant eruptions are frequently seen 

 on the face of the sun at the moments when fadeouts are commencing. 

 To some extent it is still an ad hoc assumption, for the light whereby 

 the eruption is seen is certainly not ionizing light, and we must assume 

 that the visible light is attended by rays of other wave-lengths having 

 just the properties desired. Coincidence of fadeout and eruption is, 

 however, so frequent, that it would now take a very sceptical mind 

 to reject the assumption. The trend of the curve "/m/.v" in Fig. 15 

 sustains it. 



And so I have now come back to the theory that it is radiation 

 from the sun which makes the upper atmosphere into an ionosphere, 

 by detaching electrons incessantly from the molecules thereof. (The 

 detaching must be incessant, for the electrons are always liable to 



