80 DETERMINATION OF THE MASS [CH. vjii. 



because it is of such insignificant weight compared 

 with its surface but falling always relatively to the 

 air. A cloud may readily be carried up by a current 

 of air, but that is only because the air is moving up 

 foster than the drops are trickling down through it. 

 No motion of the air disturbs the relative falling 

 motion : the absolute motion with reference to the 

 earth's surface is the resultant of the two. 



The fact that nuclei are required for mist precipi- 

 tation can be . proved by filtering them out with 

 cotton wool, and finding that as the nuclei get fewer 

 the mist condensation differs in character, becoming 

 ultimately what is called a Scotch mist, such as forms 

 in fairly clean air ; where since the dust particles are 

 comparatively few, the centres of condensation are 

 few also, and accordingly have each to condense a 

 considerable amount of vapour ; so that the drops are 

 not nearly so close together and are bigger ; wherefore 

 they fall quicker, like very fine rain. In perfectly 

 clean elaborately-filtered air the dew point may be far 

 passed without any vapour condensing, and the space 

 will remain quite transparent in spite of its being 

 supersaturated with vapour. 



The reason for this effect of, and necessity for, 

 nuclei, is. thrown into strong relief by Lord Kelvin's 

 theory concerning the effect of curvature on vapour 

 tension,* because the more a liquid surface is curved 

 the more it tends to evaporate, and an infinitely 

 convex surface would immediately flash off into 

 vapour. Consequently an infinitesimal globule of 

 liquid cannot exist; vapour can only, condense on a 

 surface of finite curvature, such as is afforded by a 

 dust particle or other body consisting of a large 

 aggregate of atoms. For it must be remembered 



* See, for instance, Maxwell's Theory of Heat, 1891 edition, p. 290. 





