Discussion of Atmospheric-Electric Observations 415 



Perhaps one of the chief difficulties resulting from a comparison of the ionization over 

 the land and ocean arises from the fact that if we assume, as we must do, a sufficiently large 

 value for the ionization caused by the penetrating radiation to account for the ionization 

 over the ocean, i. e., a rate of formation of 1.6 pairs of ions per c. c. per second, it would 

 seem that we must consider this cause to be active over land also. Thus, the rate of pro- 

 duction of pairs of ions which can be accounted for on land is at least 1.6 plus the rate of 

 production which can be accounted for by the radioactive materials in the soil and atmos- 

 phere on land. The latter contribution amounts to 4.5 according to the estimates of Eve,' 

 hence the value of q which can be accounted for on land is at least about 6, and this is much 

 more than is necessary to account for the observed number of small ions, the latter being no 

 greater than the value found at sea. It would seem that the explanation of this difficulty 

 must be sought in the slowly moving ions, which, though they contribute very little to the 

 conductivity, nevertheless have to be maintained, since they are continually suffering 

 recombinations. The total number of ions, small and large, over the ocean must be sup- 

 posed to be smaller than that over land, but the greater purity of the air over the ocean will 

 result in the fraction of the ions which exist there as small ions being greater than the cor- 

 responding fraction over land. The practical equality of the measured numbers of small 

 ions over land and sea drives us to the conclusion that, as far as this fact is concerned, the 

 decrease in the total ionization over the sea is just compensated by the greater fraction of 

 the ions which there function as small ions. Thus, for example, if we were to assume that 

 there are no slowly moving ions over the sea, we should have q' = 1.6 as the total ionization 

 over the sea. Hence, total ionization over land = 1.6+4.5 = 6.1. Thus, if n, and nj 

 refer to the total numbers of small ions per c. c. over sea and land, respectively, N to the 

 number of large ions per c. c. over the land, and if a is the same for both classes of ions, 



N+ni_ 



V1.6/ 



= about 2 



Hence, since n, = ns approximately, we should on this basis find iV = n,, i e., the number of 

 small ions per c. c. on the land would be about equal to the number of large ions. If the 

 ions over the sea are not entirely of the quickly moving class, an argument of this type 

 would lead to the conclusion that the number of large ions over the land is greater than the 

 number of small ions. 



It will thus be gathered from the foregoing remarks that the two outstanding problems 

 which are of primary importance for the satisfactory clarification of our ideas on the 

 ionization over the land and sea are, (1) the problem of determining how much of the 

 ionization produced in a closed vessel is to be attributed to causes other than the vessel itself, 

 and (2) the problem of determining, over land and sea, the average number of slowly moving 

 ions per c. c. 



Although the radium emanation over the great oceans is insufficient to contribute 

 markedly to the ionization there, its effect over a small ocean like the Atlantic is not 

 insignificant. In this connection a point of some interest shows itself when we compare the 

 average values of the conductivities over the Atlantic and Pacific Oceans. Taking the 

 average of the values in Table 91 for the Atlantic Ocean, we find X+ = 1.57X10~* e.s.tj., 

 and X_ = 1.31X10~* E.s.u. The mean of these values is 1.44X10~* e.s.u. Taking the 

 average values for the Pacific Ocean as obtained by utilizing the values in Table 85 and 

 the corresponding values in Table 91 , the very low values 0.08 X 10 "^ e. s. u. being, however, 

 omitted, we obtain X+ = 1.46X10"* e.s.u., and X_ = 1.22XlO~^ e.s.u. The mean of 

 these values is 1.34X10"'' e.s.u. Writing X^ and X^for the mean unipolar conductivities 

 over the Atlantic and Pacific Oceans respectively, we thus have X„/Xp = 1.44/1.34 = 1.07. 



^Phil. Mag., S. 6., vol. 21, p. 34, 1911. Eve gives the value 4..3.5 instead of 4.5. The slight difference is to be accounted 

 for by the fact that Eve took 80X10~" curie per cubic meter as the normal radium-emanation content of the atmosphere, 

 whereas the value 88X10"'- curie per cubic meter has been adopted in this report. 



