McClelland and Kicnnedy — 'Hie Large Ions in the Atmosphere. 89 



•considerable doubt lias been thrown on their accuracy as giving the correct 

 number of small ions by Pollock, who asserts that in the Ebert apparatus a 

 very considerable fraction of the large ions are collected by the insulated 

 terminal, the result being to give far too high a value for the number of small 

 ions. -In Pollock's own measurements of the small ions, using an apparatus 

 and voltage such as to collect all the small ions and only an insignificant 

 fraction of the large ones, he obtained at Sydney values, the maximum of 

 which was 157, the minimum zero, the mean of 128 observations giving 39 for 

 the munber of positive and 38 for the number of negative ions per c.e. 

 Langevin measured the small ions in the same way ; and his statement that 

 the number of large ions was about fifty times the number of small ones would 

 give the number 200 for the small ions per c.c. This also is far below tlie 

 values ordinarily quoted. The discrepancy between the values of n quoted by 

 observers who take no account of the large ions, and the values given by 

 those who have considered the large ions, is further emphasized by our 

 measurements made on twenty-three days at different periods of the year. 

 The table shows the maximum observed value of n to be 197 and the minimum 

 17, the mean value being 78. 



Direct observations on the number {q) of small ions produced per 

 second in a cubic centimetre of air by natural ionisation give 5 or 6 as a mean 

 value for this quantity. In the case of natural ionisation in ordinary air, in 

 finding the relation of n and 5 we must take into account the recombination 

 •of the small ions with the large ions of the opposite sign,'and also the diffusion 

 of small ions towards uncharged nuclei as affecting the steady state. In the 

 steady state we will therefore have 



c[ = an- + (5N^n + jNn, (2) 



■where a is the ordinary coefficient of recombination . of small ions of 

 opposite signs, and the other letters have the same meaning as above. In 

 this equation the terms iuvoh'ing the number of large ions and of neutral 

 nuclei will be of far greater importance than the term aJi' in the case of ordi- 

 nary air, where n is so small in comparison with iV^or JV-^. We have at present 

 but very meagre data to compare the value of q obtained from this equation with 

 that found by experiment. Suppose, however, we take /3 and y of the same 

 order as a, say 10"^* and suppose io\:A\ we take the value 50,000, whicli would 

 be an average value for the number of nuclei per c.c. obtained by Barus by his 

 condensatioiTapparatus ; then assuming for n and iVthe average values 80 and 



*The value of a obtained by McClung in filtered air is 1-5 x 10-« (Phil. Mag., 1902, 

 pp. 283-305). 



[14^=] 



