i8o 



NATURE 



[February 9, 1922 



ior the frictional resistance, T^acdfxv, must be divided 

 by 1+20!./^ where L is the mean free path of a 

 gas molecule and a is a constant depending on the 

 nature of the particle and its form. 



If, then, particles of duet enter the highly rarefied 

 upper hydrogen levels of the atmosphere, they will 

 experience very little retardation until they reach 

 that level (about 100 km.) at which viscosity 

 begins to increase rapidly to its normal or full 

 value, but then they will be very quickly brought 

 to rest even in spite of their high initial velocity. 



Hence none of this dust will penetrate below a 

 certain level in the atmosphere, probably from 60 

 to 80 km. It will be stopped and held by 

 air viscosity. The moment its velocity falls off the 

 forces tending to separate the oppositely electrified 

 particles will also decrease, and the oppositely 

 charged particles may then neutralise each other. The 

 result will be, as I think, to give the highly con- 

 ductive layer in the earth's atmosphere a tolerably 

 well-defined under-surface determined by the very 

 rapid rate at which air viscosity rises with increasing 

 air-pressure. Hence it is clear that when the dust 

 particles reach a certan level their earthward pro- 

 gress will be practically arrested. 



Meanwhile the region above this will be left im- 

 pregnated with the smaller and lighter negative ions 

 which are moving slowly or quickly in directions 

 oblique to the earth's magnetic meridians and wind- 

 ing their way spiral-fashion towards the regions of 

 the magnetic poles. The explanation of nvimerous 

 astronomical, meteorological, magnetic, and atmo- 

 spheric electric phenomena by the aid of this hypo- 

 thesis of electrified solar dust projected by light 

 pressure from the sun to the earth has been worked 

 out in great detail by S. Arrhenius, K. Birkeland, 

 W. J. Humphreys, and others. This solar dust 

 hypothesis seems to be supported by the observations 

 of Newcomb, Yntema, Abbot, and W. W. Campbell 

 on the fact that on clear moonless nights the sky 

 sends to us more light than can be accotmted for by 

 the sum total of starlight, and that this extra light 

 is notably greater near the horizon than at the 

 zenith, and also by the spectroscopic observations 

 which show the green auroral line in all parts of the 

 tropical sky on moonless nights. 



In addition to this hypothesis of a permanently 

 conducting upper region of the atmosphere we are 

 compelled to postulate that beneath this there must 

 be a region of variable ionisation due to solar light, 

 which is ionised during the day above the level of 

 clouds, dust, and water-vapour, but more or less 

 dis-ionised during the night. 



Dr. Eccles has worked out the consequences of 

 assuming an atmospheric region in which ions of 

 molecular mass are present, possibly formed by the 

 action of ultra-violet light on molecular groups 

 which are photo-electric. The presence of these 

 heavy ions acts so as to produce what is in effect a 

 reduction in the dielectric coefficient and therefore 

 an increase in the velocity of electric waves through 

 the ionised region. 



This action may be illustrated by a magnetic 

 parallel. If iron spheres were placed in a magnetic 

 NO. 2728, VOL. 109] 



field they would be magnetised, but owing to the 

 reverse action of the free poles the magnetic force 

 in the iron would be less than the force at that point 

 if the iron were not there. Hence the magnetisation 

 produced is not that which corresponds to the ex- 

 ternal impressed magnetic force, but to the reduced 

 magnetic force. 



In the same manner if heavy ions are present in 

 the air the orderly arrangement of them by the 

 impressed field reduces the effective electric force 

 in the space occupied by them, and this is equivalent 

 to a reduction in mean dielectric constant. But the 

 velocity of the wave is inversely as the square root 

 of the dielectric constant, and therefore the wave 

 speed is increased. From this it follows that if 

 there is a gradually increasing density of heavy ions 

 of both signs as we rise higher in the atmosphere, 

 there will throughout that region be a gradually 

 increasing electric wave velocity with height, and 

 therefore an effect which has been called ionic re- 

 fraction in virtue of which the higher levels of a 

 plane electromagnetic wave advancing over the earth 

 will advance more quickly than the lower parts. 

 Hence the wave track will follow round the earth's 

 curvature and an obliquely rising ray may even be 

 brought down again to eaiith by an action resembling 

 that of an inverted mirage. 



The very complicated phenomena connected with 

 freak signalling, the great effect on signal strength 

 of the sunset and sunrise periods, the curious ano- 

 malies in the difference between daylight and night- 

 time radio-transmission for various wave-length, and 

 the variation in range between north— south and east- 

 west transmission have all received certain plausible 

 explanations on the theory of a variable ionisation 

 by sunlight of the atmosphere, and its irregularities 

 at the bounding surface of the earth shadow cone as 

 it sweeps through the atmosphere. The atmo- 

 spheric ionisation at this surface will tend to become 

 " patchy," and will therefore bestow a certain in- 

 creased opacity and increased reflecting power on 

 that region for electric waves just as small air 

 bubbles in water give it a certain opacity for visible 

 light. 



The general increase in range of radio-communica- 

 tion by night is accounted for on this theory as due 

 to partial removal of the ionic refraction which in 

 the daytime brings the ray down again to earth at 

 ranges less than that due to the guiding properties of 

 the permanently ionised higher layer. 



There are, however, curious exceptions to this in 

 the case of certain long-wave transmission. Senatore 

 Marconi long ago pointed out that with certain wave- 

 lengths from 5000 to 6000 metres transatlantic radio 

 signals are often stronger by day than by night. 

 These anomalies and others recorded by Dr. Eccles 

 seem, however, to meet with reasonable explanations 

 on the ionic refraction theory- 



On the other hand, our difficulties are great in 

 bringing these hypotheses to critical test. The atmo- 

 spheric region in which the phenomena take place 

 is far beyond the reach of our meteorological sound- 

 ing balloons or possibilities of testing the actual 

 ionic distribution. We can only, therefore, patiently 



