638 Mr. C. T. R. Wilson on 



will never be attained below the first-formed drops, and the 

 rising air will remain supersaturated to nearly the critical 

 extent till it reaches these drops. They will thus grow 

 extremely rapidly until large enough to fall through the 

 supersaturated layer and lower cloud. 



The supersaturated layer above the top of a cumulus cloud 

 which is rapidly developing into cumulo-nimbus would pro- 

 bably in most cases be of quite small vertical thickness. 

 Under the conditions just considered we might expect a thin 

 cloud-cap to be formed suddenly over the head of the cumulus 

 and sink rapidly into it. (I have on several occasions ob- 

 served something very like this taking place ; but it is of 

 course very easy to be deceived in such matters.) It is 

 possible that a process of this kind would determine the 

 transition from cumulus to cumulo-nimbus. The drops 

 formed on the ions may themselves fall through the lower 

 cloud and reach the ground as rain. At the same time the 

 electric field between the free positive ions above and the nega- 

 tively charged drops will by the fall of the latter be extended 

 through the upper part of the lower cloud. This field may 

 help coalescence both between droplets of the lower cloud 

 and between them and the larger drops which have come from 

 above it. When this stage has been reached the effect of 

 splashing referred to in the last paragraph quoted above 

 from 'Nature' may be expected to become effective. It 

 would be interesting to know from direct experiment what 

 is the nature of the splash occurring when a large falling- 

 drop overtakes a smaller one. It seems not unlikely, however, 

 in the light of Worthington and Cole's photographs of the 

 " splash of a drop " falling into a vessel of water (Phil. Trans, 

 vol. clxxxix. A. p. 137, 1897), that droplets may be thrown 

 out from the underside of the combined drop if coalescence 

 takes place. If we have a vertical electrical field already in 

 existence the splashing process will in that case continually 

 add to the intensity of the field. 



It is in such ways that I think condensation upon negative 

 ions may have to be taken into account in relation to thunder- 

 storms ; the first onset of the rain may be due to this cause, 

 and it may also determine the direction of the initial electric 

 field. The intense electrical fields of thunderstorms and the 

 large charges carried down by the drops to the ground are 

 possibly due to electrostatic induction effects accompanying 

 collisions of the drops ; the prevailing sign of the field and of 

 the charge carried by the drops may be determined by the 

 initial condensation on negative ions. 



Elster and Geitel long ago suggested ( Wied. Ann. vol. xxv. 



