Surron—Study of Evaporation from Water-Surfuces. 149 
is probably only exact in the case of a permanent gas. For con- 
sider an extended volume of air containing aqueous vapour, whose 
temperature at the one end A is say 60°, and at the other end 
B 100°. let the vapour-tension at A be ‘52 inch, and at B 
1:00 inch. At A the air is saturated at its temperature; at B it is 
much drier. According to Clerk Maxwell’s statement, the aqueous 
vapour would diffuse from B to A, because there is a less quantity 
of it at_ A. Now one of the purposes of this paper is to inquire 
whether the unbalanced flow of vapour may be not along the 
absolute humidity gradient, but along what might be called the 
potential humidity gradient, which is probably a sort of compro- 
mise between the absolute and relative humidities. We do indeed 
often meet with the assertion, made in a loose, descriptive sort of 
way, that the rate of evaporation depends upon the dryness of the 
air. But most often the author is not quite clear as to what he 
means by “dryness.” If, however, for “ dryness ”’ we substitute 
“yelative humidity,’ then it seems highly probable that the 
statement contains a large element of truth. 
Tf this be so, then the air must be ina certain sense hygroscopic, 
and in part analogous to a hair hygroscope. Abbe has remarked 
that the rate at which vapour diffuses through a porous diaphragm 
depends on the difference of vapour-tension between one side and 
the other of the diaphragm, and claims that, ‘“‘ undoubtedly, a 
similar diffusion into animal and vegetable cells through their 
thin walls isthe basis of the hygrometric action that is utilized in 
the hair hygrometer of Saussure.” This, however, is hardly likely 
to be quite the case. For it is a matter of ordinary observation 
that if a wisp of human hair, which has been thoroughly cleansed 
in a soda solution and dried, be placed in a delicate balance in 
unsaturated air containing a given constant quantity of aqueous 
vapour, any decrease of temperature (which of course increases 
the relative humidity) will cause the wisp of hair to become 
heavier, and any increase of temperature (decrease of relative 
humidity) will cause it to become lighter—.e., to absorb or give 
out moisture without any regard to the actual vapour-tensions 
inside or outside the hair.! Is it not a more likely explanation 
1 My own observations, however, only show this in a rough way: I have so far 
failed in obtaining a definite relation between the increase of relative humidity and 
the increase of weight of the hair. 
