1901.] on Some Recent Work on Diffusion. 549 



of the solutions in the ordinary sense of the word, for the gelatine is 

 virtually a solid. The effect has been produced by the molecules of 

 the coloured copper salt, by reason of their rapid movement in all 

 directions, gradually penetrating into the spaces between the mole- 

 cules of the gelatine layer. Given a sufficient length of time, and 

 there would be an equal partition of the coloured substance between 

 the two layers. 



Diffusion takes place, as is well known, much more rapidly with 

 gases than with liquids. Had our cylinder contained, for instance, 

 carbonic acid in the lower half and air in the upper, a complete 

 mixing would have taken place in a comparatively short time, even if 

 all convection currents had been prevented. 



The classical researches of Graham on the diffusion of gases 

 through thin porous septa established the general law that the rate 

 of diffusion of the different gases, under identical conditions, varies 

 inversely as the square roots of their respective densities. Graham's 

 results, however, only acquaint us with the relative velocities of dif- 

 fusion, whereas for the particular problem which we have before us, 

 we must know the absolute velocities of diffusion under strictly defined 

 conditions. 



It is mainly to the Viennese school of physicists, and especially 

 to Professor Loschmidt, that we owe our present knowledge of the 

 actual rate of penetration of one gas by another in free diffusion. 



By observing the speed with which different pairs of gases 

 spontaneously mix in a tube, Loschmidt was able to deduce certain 

 absolute values expressing the velocity of their interpenetration. 



Some of these results for different pairs of gases are given in the 

 diagram, the last column representing the " constant of diffusivity," 

 expressed in centimetre-gram-second units. 



Let us consider the constant for carbonic acid and air, which at 

 0° C. is * 142. This means that when air and carbonic acid gas are 

 freely diffusing into each other at this temperature, an amount of either 

 gas corresponding to • 142 cubic centimetre will pass in one second of 

 time, across an area of one square centimetre, when the partial pressure 

 of the gas varies by one atmosphere in one centimetre of length. 



Now when we come to apply these absolute values of diffusivity 

 to the passage of the extremely dilute C0 2 of the air into the leaf 

 stomates (whose dimensions can of course be determined), we find 

 that free diffusion through these openings is apparently able to 

 account for only a portion of the gas which we know must enter the 

 leaf, unless we make some extremely improbable assumptions as to 

 the very low point at which the partial pressure of the carbonic acid 

 is maintained immediately under the apertures. 



I shall not, however, trouble you with the calculations on which 

 this statement is based, since I prefer to put the matter in a more 

 concrete form, which has also the advantage of emphasising the 

 extraordinary power which an assimilating leaf possesses of extracting 

 carbonic acid from its surrounding air. 



