64 



THE STRUCTURE OK THE NUCLEUS. 



For the sake of coiiii>:iiis<.ii willi these computed results, the experiments 

 described in table 6 and figure 6 were specially investigated. Observations of the 

 fall, /■ of the |pl;inc uf demarcation in the lapse of time were made in full daylight. 

 The first 8 e.\hausti«>ns were successively examined in the lai'ge globe, and tho rates 

 of descent of the "id and 8th computed graphically. 



lABLE 6.— SUBSIDENCE OF FOG STR.ATA OF SUCCESSIVE ORDERS. 

 OBSERVATIONS. Cf. Fij (< 



The observed i-ate of subsidence is seen to be a remaikedly deci'easing quan- 

 tity, from which one at once suspects that it is largely apparent. Naturally, the 

 observations are vague from the difficulty of seeing a sharply defined upper fog 

 limit; but the curves as a whole nevertheless unmistakably indicate much raoie 

 lapid descent at the beginning than at the end of the curves, and a rate of descent 

 inrreasiutr with the number of exhaustions, i.e.^ with the mean size of the cloud 

 particles. The decreased rate of subsidence must be referred to evaporation at the 

 top, where the smallest particles congregate, and the following chaptei-s will show 

 how ini]>i)rtaiit this soui'ce of loss of cloud particles really is. To this must further 

 be added the tendency of cleared warmer air to rise on the outside of the Ijulk 

 contained, producing false subsidence or an apparent descent of the fog limit to 

 replace the removal by convection. The table shows Ikuv little confidence may be 



