224 Messrs. W. B. Hardy and H. W. Harvey. [June 21, 



With this value, and on the assumption stated above, I find that the 

 pressure at the interface would vary as follows : — 



Depth of film, in ftp. 



Per cent, of maximal 

 pressure. 



50 



95 



40 



90 



30 



80 



20 





10 



45 



5 



25 



The pressure would, however, not increase so rapidly as this with increase in 

 the thickness of the film, owing to the variation of density in the film itself 

 which is not taken account of in the above calculations. This is obvious 

 when we remember that the film as it gains in thickness also gains in mean 

 density owing to compression by the increase in the mean Laplacian pressure. 

 The compressibility of the film will be relatively great since it is a transition 

 layer between gas and fluid. 



Turning to the observations themselves, the film was always so thin as to 

 produce very slight effect upon the movements of shreds of camphor. This 

 is what might be expected, since the distilled water was drawn from the 

 bottom of a large glass reservoir, and all the chambers were thoroughly 

 rinsed. From Eayleigh's measurements of such films* the thickness may 

 be put with tolerable certainty as less than 2 /a/a — probably 1"5 /a/a. 



At 2 /a/a the interfacial pressure will be considerably less than 10 per 

 cent, of its maximal value, and the coefficient of sliding friction y should 

 have diminished proportionately. 



Therefore, for the same potential gradient, the velocity of the water past 

 the film should be much greater than it would be past a film 100 /a/a thick, 

 or past the glass if we assume that the layer of electric density at the 

 glass-water interface does not differ widely from that at the film-water 

 interface. 



So much for theory. Observation shows that the velocity of the water 

 past the surface film differs very slightly from that past the glass at the 

 bottom of the trough. Thus a surface film of a thickness far below the 

 accepted estimate of the range of molecular action acts like a mass of solid 

 of, relatively, infinite thickness. 



In considering this surprising result the three variables on which the 

 relative velocity at an interface depends have to be remembered. The 



* ' Eoy. Soc. Proc.,' 1890, vol. 47, p. 364. 



