RELATION OF VAPOUR AND AIR. 570 



itself into a fine watery powder, the diameter of 

 the particles of which this powder consists being 

 very small : they are estimated by various writers, 

 from 1 -100,000th to l-20,000th of an inch* 1 . Such 

 particles, even if solid, would descend very slowly ; 

 and very slight causes would suffice for their sus- 

 pension, without recurring to the hypothesis of 

 vesicles, of which we have already spoken. Indeed 

 that hypothesis will not explain the fact, except we 

 suppose these vesicles filled with a rarer air than 

 that of the atmosphere ; and, accordingly, though 

 this hypothesis is still maintained by some* 2 , it is 

 asserted as a fact of observation, proved by optical 

 or other phenomena, and not deduced from the sus- 

 pension of clouds. Yet the latter result is still vari- 

 ously explained by different philosophers : thus, M. 

 Gay-Lussac 23 accounts for it by upward currents of 

 air, and Fresnel explains it by the heat and rarefac- 

 tion of air in the interior of the cloud. 



Classification of Clouds. A classification of 

 clouds can then only be consistent and intelligible 

 when it rests upon their atmological conditions. 

 Such a system was proposed by Mr. Luke Howard, 

 1802-3. His primary modifications are, Cirrus, 

 Cumulus, and Stratus, which the Germans have 

 translated by terms equivalent in English to feather- 

 cloud, heap-cloud, and layer-cloud. The cumulus 

 increases by accumulations on its top, and floats in 



91 Kjpmtz. Met. i. 393. M Ib. i. 393. Robison, ii. 13. 



- :I Ann. Chin;, xxv. 1822. 



I' I' --' 



