362 Professor Sir James Dewar [Jan. 18, 1918 



long before such complete drainage as this is reached any movement 

 of the film occurring over a relatively dry surface will quickly 

 exhaust the liquid in the channel, and the film will be ruptured. 

 However, when the proper precautions were taken, the displacements 

 due to atmospheric variations did not amount to more than 1 or 

 2 mm. for weeks together. 



Until they become completely black very striking colour effects 

 can be seen in these large horizontal films. They have of course 

 only a small gradient of thickness, so that the bands of colour are 

 drawn out into large areas. During this period convection in the 

 films is most easily seen, because of the brilliant contrasts afforded 

 by the moving portions. Should there be much fluctuation of tem- 

 perature, the complete development to the black state may be 

 arrested, the last part of the coloured area being kept in intermittent 

 circulation for several days. Superfluous liquid draining down out- 

 side the central tube will produce a similar result by becoming spread 

 out on the black film surface as a coloured clot round the tube. 

 When this clot grows sufficiently it will draw off slowly, and pass 

 down the radius of greatest inclination of the film. It formed on 

 one occasion an ellipse (1 cm. by 1^- cm.) which took 35 seconds to 

 traverse the last one-fourth of its path. On approaching the 

 periphery where the film curves up slightly into contact with the 

 glass, the clot flattened out and curled into the most intricate con- 

 volutions, finally sweeping round close to the edge, a writhing 

 mass of coloured streaks, many of which became roving stars on the 

 black film. The whole coloured area was finally drawn into the 

 Gibbs ring, from which frequent escapes of small drops of liquid 

 took place. 



A favourable condition for longevity is a quiet temperature rather 

 below 10° C. Sudden alterations of 3° or 4° C. are most likely to ])urst 

 the films by provoking too rapid convection therein, as is evidenced 

 by very rapidly moving streams of silvery stars drawn from the 

 Gibbs ring. 



The vibrations which occurred during the operation of the air 

 compressors in the Laboratory did not affect the final stability of the 

 black films. A stationary circular wave was produced having an 

 amplitude of over 1 mm. vertically all round a circle midway 

 between the central tube and the walls of the globe. This motion 

 continued the whole time the machines were working, but the films 

 always survived. 



Instead of having one large plane black film, a mesh-work of 

 black films can be obtained completely filling the globe. To pro- 

 duce this, the lower end of the blowing tube A (Fig. 1) is inserted 

 slightly below the surface of the soap solution. The air current 

 being regulated suita])ly to the diameter of the tube A, a steady 

 succession of uniform bubbles is produced. The surface of the glass 

 having previously been well moistened, the issuing bubbles conse- 



