142 



NATURE 



[February 2, 1922 



minus ions are so far separated that they do not 

 readily recombine. True gaseous photo-ionisation 

 always produces ions of both signs in equal number 

 mixed up together, and the conductivity quickly dis- 

 appears when the ionising agency is withdrawn. 



(2) The resulting electric conductivity must be 

 sufficiently high, say, as good as that of ordinary 

 fresh water, to act as a true wave guide. This 

 implies that the ions must be very numerous per 

 c.c. and very mobile or have high ionic velocities 

 under unit electric force. 



Bearing in mind that the upper regions, of the 

 earth's atmosphere above the 100 km. level prob- 

 ably consist chiefly of hydrogen, and that the velo- 

 city of ions in hydrogen under unit electric force is, 

 according to measurements, from two to three times 

 that in oxygen or nitrogen at the same pressure, it 

 is easily seen that in the upper hydrogen levels of 

 the atmosphere a very moderate amount of ionisa- 

 tion, say, lo^ ions per c.c, might give a conduc- 

 tivity of the order of that of fresh water, or about 

 700,000 ohms per c.c. Another quality this 

 conducting layer must possess if it is to act as 

 a true reflector of long waves is a somewhat sharply 

 defined lower surface. 



It has already been remarked that observations 

 on signal strength over long distances show an enor- 

 mous difference between the actual measured values 

 and those predicted by a simple diff'raction formula. 

 Attempts have been made to find an empirical 

 formula for the received current in terms of the other 

 quantities involved. At first these efforts started 

 W'ith the erroneous assumption that the attenuation 

 might be regarded as due to an " absorption " 

 caused by the atmosphere, and therefore mathematic- 

 ally represented by an exponential factor appended 

 to the simple Hertzian expression for the magnetic 

 or electric force at a known distance on the equa- 

 torial plane of a small oscillator. 



Prof. G. N. Watson finds that if in place of a 

 perfectly conducting spherical earth in free space 

 we assume an earth having a conductivity about the 

 same as sea water, enclosed in a spherical sheath 

 or shell of material having a conductivity of about 

 1-44x10-^5 E.M.U., equal to a specific resistance 

 of 700,000 ohms per c.c. or not far from that of 

 ordinary fresh water, the interspace being about 

 100 km., then the diff^raction- formula for the 

 receiving aerial current would have to be modified 

 and the exponential factor becomes e-9■ee/^/A_ 

 Watson therefore considers that if we are able to 

 assume an upper conducting layer in the atmosphere 

 at a height of about 100 km. having a fairly sharp 

 under-surface and a specific resistance of about 

 700,000 ohms or, say, 0-75 megohm per c.c, then 

 guided wave propagation through the included 

 spherical shell of insulating air would account for 

 the observed attenuation in actual terrestrial long- 

 distance radio-telegraphy. 



We have then to consider what are the probabili- 

 ties and possibilities for the existence at a height of 

 100 km. or so of such a conducting layer and 

 how it may be supposed to become ionised. Gaseous 

 conductivity is always and onlv due to the presence 

 NO. 2727, VOL. 109] 



of ions, and in the above case these are created by 

 the strong electromotive forces brought into play. 

 In gases contained in glass vessels there are always 

 some few free ions or electrons present for some 

 reason. If a high frequency magnetic field is made 

 to act on the gas these ions are driven with great 

 force against the gas molecules and ionise them, 

 thus producing very quickly a copious supply of 

 ions and giving the gas high conductivity. We 

 cannot, however, say that a rarefied gas is a good 

 conductor per se for very feeble impressed electro- 

 motive forces as we can say that a metal is a good 

 conductor. Hence mere rarefaction due to height 

 will not bestow the required electric conductivity on 

 the atmosphere. Neither can the required ionisation 

 be produced by solar light, because then it would 

 vanish in the night-time by recombination of the 

 ions. 



The suggestion I wish to make as to the cause of 

 this ionisation is based upon a modification of hypo- 

 theses already advanced by S. Arrhenius, K. Birke- 

 land, and W. J. Humphreys concerning the pro- 

 jection of dust by light pressure from the sun. 



We know that the sun's photosphere is in a con- 

 tinual state of disturbance due no doubt to violent 

 explosions in regions beneath this light-giving 

 locality. Above this photosphere lies the so-called 

 reversing layer composed of metallic vapours which 

 produce the Fraunhofer lines in the spectrum. 

 These eruptions carry up not only metallic vapours, 

 but also vast masses of the superlying chromosphere 

 composed chiefly of hydrogen and helium gases in 

 the form of solar prominences or red flames which 

 are often seen rising to a height of several hundred 

 thousand kilometres in a few minutes, thus indi- 

 cating velocities of several hundred kilometres per 

 second. When these solar metallic vapours are thus 

 carried up into colder regions they must be con- 

 densed into a metallic mist or rain composed of 

 particles of various sizes. We know also from ex- 

 periment as well as theory that light exercises a 

 pressure on solid objects and that this pressure 

 per square centimetre for totally absorbing or 

 black bodies is numerically equal to the light 

 energy in the cubic centimetre. Measurements 

 made of the so-called solar constant at the 

 earth's surface when corrected for atmospheric 

 absorption give a value of 2-5 gram calories 

 per sq. cm. per min. Hence the energy 

 of light per c.c. is nearly 6/10^ ergs and the light 

 pressure therefore 6/10^ dynes per sq. cm. on a 

 black surface. But at the sun's surface this pressure 

 is 46,000 times greater, or 2-75 dynes per sq. cm. 

 As this pressure varies as the square of the linear 

 dimensions of the particle w^hilst gravitation varies 

 as the cube, it is clear that as the dimensions of a 

 particle decrease a limit will be reached at which 

 the light pressure will overbalance the gravitation 

 attraction. 



It is easy to prove from known data that at or 

 near the sun's surfaces black particles of the density 

 of water would be just repelled if they had dia- 

 meters of 15,000 A. U. = 150/10^ cm. 



If their density is 5-5, then the critical diameter 



