200 Professor Sir James Dewar [Jan. 19, 



of constant composition, (2) bubble of glycerine solution subjected 

 to absorption of water vapour. In this second case (see also Table 9) 

 the rate of gas transference d D/2 (plotted as abscissas) is increasing 

 more quickly than the inverse of the corresponding diameters 1/D 

 (plotted as ordinates) — a fact which could be explained (1) hj an 

 increase in the surface tension, with a resulting proportional incre- 

 ment in the internal pressure of the contracting bubble : and (2) by 

 a more rapid percolation of hydrogen through the diluted film. 



It is evident that further comparative experiments are necessary 

 with more dilute films of constant composition before these effects 

 can be differentiated. Preliminary attempts showed that a solution 

 containing only 0*15 per cent, of ammonium oleate with 1 • 5 per cent, 

 glycerine can give a fairly long-lived bulible (one such in hydrogen 

 contracted completely in 2^ days from <s cm. diameter : see p. 29). 

 But the maintenance of constant thickness is not easy, because with 

 bubbles of such dilution even slight variations of temperature cause 

 recurrent colour changes : it is specially difficult to maintain the 

 black stage. 



No great alterations of temperature occurred in any of the above 

 experiments, but distinct evidences of variations from this cause were 

 noted on several occasions. The laws governing this factor have yet 

 to be elucidated. 



Alteration of Thicks' ess by Contraction. Silvery 

 ])iscs. Constancy of Black Zone. 



A contracting bulible that remains constant in weight will, of 

 course, become thicker as the surface diminishes. The thickness will 

 vary inversely as the square of the diameter, provided the capillary, 

 viscous, and gravitational forces did not come into play. For 

 small changes, on the simple theory, the rate. of logarithmic increase 

 of thickness will be twice the rate of logarithmic diminution of 

 diameter. This is a property of the sphere that may ])e expressed 

 as follows : — When S = surface, T = thickness, p = specific gravity 

 of the liquid composing the l)ubble, then, since S = ttD-, 



Mass = S T p = TT D- T p 

 and as the mass and composition are assumed unaltered, Uiis gives 

 J)'- T = constant, or T varies as yv, • ■ • (1) 



i.e. the thickness varies inversely as the square of the diameter, and 

 is true for all magnitudes. Differentiating, 



2 T 1).^/D + \)\dT= 0, or '\^t = - '2^^ . (2) 



which can be put in the form 



rf(logT) = -2. /(log))) .... (3) 



