00 



Professor Sir James Dewar 



[Jan. 21, 



in widtli ; the larger bubbles are not often over 0*008 mm. diam. 

 'J1ie appearance of the colonies are roughly represented in Fig. ">. 



•035 mm. >i 



Between the fairly equally spaced colonies there was a regular 

 distribution of smaller bubbles of from 0*005 to 0*001 mm. Among 

 these a small proportion were colonies, but of a low order of aggre- 

 gation. 



In one straight line of length, 1 * 58 mm. across a film, there 

 were : — 



Total 



10 colonies of from 0-01 to 0*02 mm. 



15 „ ,, 0-005 „ 0-01 „ 



about 50 „ about 0-001 „ 0-003 ,, 



0-15 

 0-11 

 0-10 



0-36mm. 



The proportionate length of bubbles =^ ■ 36 mm. in 1 * 53 mm. 

 of film, or 2oJ per cent, so that the total relative area of bubbles 

 to continuous film = (0'235)- = 0*055, or 5| per cent. 



The appearance of double and triple films under the microscope 

 was simply that of a more thickly sown appearance of the bubbles. 

 (Jne perfectly black film Ih cm. diam. showed much smaller bubbles, 

 and much more thinly spread ; the most numerous were 0*0(4 mm. 

 diam. In a slightly thicker film the bubbles are somewhat larger. 



From the dimensions of these bubbles and colonies it is evident 

 that they must be very flattened in shape in the film, because its 

 thickness is at most one-hundredth of the length of the bubbles. 

 The upper and lower bounding surfaces of these bubbles must, there- 

 fore, be extremely thin relatively to the mean film thickness. This 

 suggests that the main diffusion most likely occurs through these 

 extremely thin portions of the film. When two films are superposed, 

 the bubbles will not be likely to coincide to any great extent, as they 

 only occupy about 5| per cent of each film. The exaggerated diffu- 

 sion through the thin places will thus be largely checked. The 

 addition of a third film would further reduce this by a similar factor, 



