becomes 



Theory of Magnetic Storms. 681 



It is clear that the constant adiabatic temperature gradient 



— ~ ~ obtains as long as the energy of convection gV 



<y _K 



is large compared with the energy interchanged by radiation 



F(#, T). When T becomes much smaller than T' however, 



then, provided the gas absorbs, i. e. if x is appreciable, 



F(.r T 

 ¥(x, T) must become large and the term — ' 



important. ^ 



The absorption of radiation in the atmosphere is mainly 

 due to C0 2 , H 2 and possibly 3 , (>0 2 especially being at 

 any rate partly responsible for the isothermal layer. But 

 the percentage of C0 2 in the atmosphere decreases extremely 

 rapidly as the height increases and must diminish, for 

 example, to 1/100 of its original value in 5*6 . 10° cm. 

 It cannot therefore reasonably be supposed that the tempera- 

 ture of the stratosphere, which is partly conditioned by the 

 C0 2 , remains constant up to heights of this order. The same 

 holds good, though at a greater height, for H 2 0, and it may 

 be doubted whether the feeble absorption of the other gases 

 is capable of overcoming the up and down currents, which 

 must tend to form in an atmosphere in which pressure 

 differences and winds of considerable magnitude are known 

 to exist. 



These considerations would seem to show that it is 

 impossible a priori to predict anything about the tempera- 

 ture and constitution of the outer atmosphere. The fact 

 that aurora? occur at heights of the order of 10 7 cm. cannot 

 therefore be brought in evidence that the particles producing 

 electrification must have a normal range of 3*0 cm. in air 

 and a velocity of the order of 1*7. 10° cm./sec. 



The last point to be considered is the total number oE 

 particles involved, and the conditions they would produce in 

 the upper atmosphere. Since the range is approximately 

 proportional to the mass the electrons, whose energy corre- 

 sponds to a potential drop of some 1*8 volts, will be stopped 

 at their first encounter. The positive particles will continue 

 until their kinetic energy is used up in encounters or electro- 

 static potential relative to the outermost layer, which is 

 predominantly negatively charged. They will then spread 

 out under their mutual repulsion and a radial current will be 

 formed by the inflow of negative and outflow of positive 

 particles on their way to recombine. 



Dr. Chapman gives the total energy of an average 

 magnetic storm as 10 23 ergs. [{' this is equated to the 

 kinetic energy of the incident stream, and if one assumes the 



