104 Proceedings of the Royal Irish Academy. 



This is just twice the value found by Kennedy, and would imply that only 

 half of the total collisions result in coalescence. If the ion be regarded as a 

 loose cluster, tlie value of -y will be greater, and coalescence must be assumed 

 to occur at a still smaller proportion of the collisions. We may note that the 



fact that the law — = - yn" holds over a considerable range shows that when 



two nuclei coalesce the body formed rapidly assumes the size peculiar to 

 the original nuclei. "We may easily imagine two spheres of water uniting to 

 form a sphere which rapidly evaporates to tlie stable size. But the conditions 

 of stability of an ion consisting of a loose cluster are not easily imagined, and 

 it would seem as if combination between such bodies must result in the 

 accumulation of still greater complexes, so that the rate of recombination 

 should progressively increase. 



Oombivation with Small Ions. 



The interaction between nuclei, large ions, and small ions is of considerable 

 importance in connexion with atmospheric electricity. We have seen earlier 

 in this paper that nuclei are produced by the action of ultra-violet light on 

 moist air, and that in the presence of small ions the nuclei pick up charges, 

 and are converted into large ions. There is hardly any doubt that this 

 process is going on very actively in the higher layers of the atmosphere. It 

 is possible that the bulk of the nuclei of condensation are formed in this way. 

 Where large ions, nuclei, and small ions are present, the equilibrium con- 

 centration of each will depend upon their rate of combination with one another, 

 as well as upon the rate of internal recombination peculiar to each. 



The fact that large ions with six electronic charges can be formed, and can 

 persist in sufficient numbers to be easily observed, shows quite clearly that 

 the sign and magnitude of the charge already present have not much influence 

 on the collisions and unions between large ions and small ions. This enables 

 us to simplify the treatment of tliis point very considerably, for while 

 recognizing that the cliance of union of, say, a negative small ion with a 

 positive large ion of five charges is greater than its chance of union with an 

 ion of five negative charges, we shall not be far wrong in working on the 

 basis of a uniform rate of combination between small ions and nuclei of all 

 sorts. The recognition of the fact that the charge in the nucleus has not very 

 much influence in the collisions helps us to another estimate of the size of 

 the ions. Townsend^ has shown that the number of ions coming in contact 

 with a spherical conductor is appreciably influenced by a charge on the 



' Townsend, Electricity in Gases, pp. 214-1(5. 



