[ 453 ] 



the magnitude of the convection- 



XLVI. Diffusive Convections 

 By Albert Griffiths, M.Sc. (Vic), A.R.C.S. (Lon.). * 



Section I. 



rj^HE diffusion of a subtance through a liquid, e. g. water, 

 A may be regarded as merely the transmission of the sub- 

 stance through the stationary water. 



If the solution of the substance has a density different from 

 that of water, depending on its strength, the act of diffusion 

 tends to alter the density at any point ; in general this 

 tendency to alter the density prevents the possibility of 

 statical equilibrium and hence convection-currents are 

 produced. 



So far as the author is aware the only case in which 

 diffusion of the substance can take place without the pro- 

 duction of convection-currents is that in which diffusion takes 

 place along a perfect cylinder, the generating lines of which 

 are truly vertical. 



The determination of 

 currents is, in general, a difficult task, involving, as it does, 

 hydrodynamics as well as diffusion. 



The present contribution determines the magnitude of the 

 convection in a simple case, which arose in connexion with 

 an attempt to devise an apparatus for the accurate deter- 

 mination of the coefficient of diffusion of copper sulphate 

 through water. 



Omitting some details, it may 

 be said that the apparatus consists 

 of a vessel, V, divided into two 

 parts by a diaphragm D. 



Through the diaphragm pass 

 narrow tubes of which only two 

 are shown in the figure. Means 

 are taken for keeping the upper 

 compartment full of nearly pure 

 water, and the lower part full of 

 a solution of copper sulphate. 



Attention will, be confined to 

 two tubes AB, MN, only, and it 

 will be assumed that the upper 

 ends A and M are in contact with 

 perfectly pure water, and that 

 the lower ends B and N are in 

 contact with a solution of copper sulphate of constant strength. 

 Doubtless diffusion will tend to diminish the strength of the 

 * Communicated by the Physical Society : read June 10, 1898. 



