82 Art 9.— T. Terada : 



rôle in this respect, since it acts in differentiating the gases with 

 different absorbing powers and molecnlar sizes. '^ 



Again, if there be present any sensible fluctnation of the 

 solar radiation^^ with a short duration, the thermal effect, and 

 also the radiation pressure in some measure, may contribute 

 to the initiation, if not the direct excitation of the vertical 

 motion of the upper atmosphere. The remarkable irregularities 

 of the day-waves may partly be accounted for in this way.^^ 



As to the origin of the magnetic waves with the periods 

 longer than 10'" we may probably suggest the slow atmospheric 

 waves possible in the case Avhen two layers with different 

 temperatures are superposed. ^^ 



The fact described in § 17, that the hourly distribution of 

 the senses of rotation of the disturbing magnetic vectors are quite 

 different for the short waves and the long undulations, requires 

 an explanation. It may only be suggested for the present that 

 the difference is in any case due to the difference in the modes 

 of atmospheric free vibration. Any further discussion must be 

 postponed till the nature of the vibrations of the atmosphere 

 has been more fully investigated from the theoretical side, 

 especially for a spherical atmosphere. 



1) Adopting Debye's approximate formula for an ahaorhinfj sphere (Ann. d. Pliys., [4] 30, 

 1909, p. 117), tlie pressure of tlie light wave with the wave length ), on a sphere of radius 

 (( may be put 



„ 27Trt ,S' 



1-83 . w;2 , 



> c 



where S is the energy of the radiation per sec. per cm.- Assuming « = 10-^, ). = 5xlO-5 and 



3 

 Ä = — ; — X4"10'', we obtain a pressui-e 0'48XlO-22 dynes. Taking the mass cf the sphere as 



l"GXlO~'-*gr. for a hydrogen atom, we obtain an acceleration of 30 cm/sec. The mean 

 free i^ath at the pressure of 0007 mm. being about l-7tî cm. the mean velocity may be 

 estimated to be of the order of 5'1 cm/sec The path traversed is therefore only 184 m. per 

 hour. This is probably too small to answer our purpose. But if we take a " Dipole " 

 instead of the absorbing sphere, the value of the pressure may in some cases be estimated 

 to be decidedly greater. Near 200 km. from the earth's surface, the pressure will vary 

 rapidly with the height ; in the higher layer, the path traversed by molecules or atoms 

 may be enoi-mously greater than the above estimation. 



2) Together with the back-radiation from the earth. 



3) The corjmsculnr radiation being deviated by the terrestrial field, will not be confined 

 to the day hemisphere, or rather prefers the night side. The fluctuation of this radiation 

 may be considered to be of the type considered in p. 79. 



4) H. Lamb, Proc. R.S., A 84, 1911, p. 551. 



