684 On the Theory of Magnetic Storms. 



If, is suggested that the storms are due to the recombina- 

 tion o£ electrons and ions in the upper atmosphere which 

 arise from ionized clouds of gas ejected from the sun, such 

 as are observed in the protuberances. It is shown that no 

 appreciable electrostatic forces can exist in the chromosphere, 

 and it is shown that the observed velocities are therefore 

 probably due to light-pressure acting on the neutral atoms. It 

 is shown that the gas will become completely ionized and there- 

 fore invisible when the density has become sufficiently low. 



It is shown that recombination will not take place on the 

 way from sun to earth and that the normal expansion of 

 the gas will increase the size of the cloud sufficiently to 

 account for the observed duration of magnetic storms. 



It is shown that the usual assumption that only particles 

 of relatively high speed could penetrate to the levels at 

 which auroraa are observed is probably erroneous, since it 

 implies the existence of hydrogen which does not appear in 

 the spectrum and since it requires the isothermal region 

 to continue to an indefinite height. This is improbable since 

 the percentage concentration of the heavy absorbing gases 

 C0 2 , H 2 0, and possibly 3 , would diminish exponentially 

 with the height. 



It is shown that the total mass of the particles striking 

 the earth need not exceed 6.10' on this theory, the total 

 charge being of the order 2..10 21 E.S.U. The potential 

 difference which would be formed between the electrons 

 which would be stopped first and the positive ions which 

 would penetrate deeper would be of the order of 1000 volts. 

 It is shown that this is of the order to be expected since 

 ionization by collision would be important if it were 

 appreciably exceeded and the charges would recombine 

 until some such value were reached. 



It is shown that the total mass in one of these assumed 

 clouds need not exceed I'd . 10 12 grammes and that its 

 apparent diameter before it becomes completely ionized 

 would be of the order of 1'. 



It is shown that this theory does not lead to any change in 

 the surface atmospheric potential gradient, and it is pointed 

 out that the magnetic phenomena to which the recombina- 

 tion of these particles would give rise would coincide with 

 those explained by Dr. Chapman's theory. 



It appears therefore that the theory leads to results which 

 are of the right order of magnitude when the quantities are 

 examined and that it is free from the inconsistencies inherent 

 in the stream theories promulgated hitherto. 



Clarendon Laboratory. Oxford. 

 July 14. 1919. 



