400 Professor Sir James Dewar [Jan. 18, 



Much higher values were given by a black column of five equal 

 segments, initially 161 c.c, very completely drained by two cellulose 

 strands on opposite sides of the bubble contours. No glass ring 

 support was employed, so that the column contracted freely without 

 the segments being strained out of their normal curvature. In four 

 days the volume diminished 30 per cent. The mean rate of air 

 transference being nearly 11 c.c. per day when each segment had a 

 volume of approximately 125 c.c. The corresponding value of the 

 curved surface was about <S(> (cm.)-, so that the loss through unit 

 area was ' 14 c.c. per day. The solution used for this contained 

 1 per cent ammonium oleate ta 10 per cent, of glycerin. 



GrAS Diffusion Through Black Films. 



The rate at which gases diffuse through liquid films was found 

 to be very much greater than the slow escape caused by the small 

 excess-pressure by which bubbles are distended. The latter only 

 becomes appreciable some days after the black stage is reached ; but 

 if a thick coloured air bubble has a small percentage of hydrogen 

 blown into it, a contraction takes place in a few minutes ; or, con- 

 versely, an expansion is quickly produced by circulating some 

 hydrogen in the closed vessel in which the bubble is hanging. In 

 the same way when a plane film Avas formed across a cylindrical 

 vessel containing only air and some soap solution, and hydrogen was 

 then circulated in the space between the film and the neck, a steady 

 movement of the film towards the neck at once began ; the hydrogen 

 passed through to the enclosed space Ijeyond the film more quickly 

 than the contained air diffused out. The reverse process was easily 

 carried out by first filling the bottle with pure hydrogen, and 

 expanding the film as before. A current of clean air was then cir- 

 culated on the same side of the film as before ; the resulting move- 

 ment of the film was now away from the neck, because the volume 

 enclosed beyond the film lost hydrogen more quickly than air could 

 diffuse in. When oxygen was used instead of air the motion was 

 more rapid. 



When these experiments were made after the black stage was 

 reached, the diffusion was sufficiently rapid to cause a large distortion 

 of the plane film, due to partial adhesion round the periphery of the 

 film, although the glass walls had previously been well moistened. 

 For the laboratory measures of the relative diffusion rates of various 

 gases, an 8 litre bottle, 19 cm. in diameter, was calibrated, and the 

 movements of the film (which would thus have an area of some 

 300 cm.-) were noted for successive small time intervals. The stopper 

 was fitted with inlet and outlet tubes, of which the former was long 

 enough to reach the bottom, and could also slide easily in an airtight 

 fitting. The film was thereby started by pushing the inlet tube down 



