1016] on Problems in Capillarity 801 



course, the whole film can be obtained in a condition too thick to 

 show colouring. Such a film, however, when once more left quiet, 

 thins out again at a much greater rate than a newly thrown film, 

 which of course has an excess of undrained liquid on the glass surfaces 

 in the neighbourhood of the Clibb layer. In fact it has been found 

 that a film some 3 or 4 cm. diam. thins with extraordinary slowness if 

 formed, e.g., on a cylindrical ring of fine gauze wrapped outside with 

 bibulous paper or worsted wetted with the soap solution. In such 

 a case only feeble colouring is seen even after some hours, showing 

 how thick the film has remained. 



To illustrate the stabihty of the film the portable box form was 

 placed in an ordinary laboratory shaking machine oscillating at 

 about three or four times a second, and the film remained intact for 

 hours. If the movement was in the plane of the film, it l)ecame 

 thickened as described above ; but if in the perpendicular plane, 

 very little alteration was produced in the distribution of thickness of 

 the bands. 



The results produced by difference of temperature in the Gibb 

 ring are well seen in the double bulb. Here again an interchange 

 of liquid between film and Gibb ring takes place. A black or partly 

 black film shows the effects in the most striking manner. If a pad of 

 wool soaked in water about 30° C. to 40° 0. is held on the top of a black 

 film (see 1, Fig. 9), in a few seconds a reflecting streak appears at 

 the bottom, and steadily extends up to a broad segment which would 

 cover the whole film surface in a short time. Now place the warm 

 pad at the bottom of the half black film. At once there begins a 

 movement in the surrounding Gibb ring : increasing streams moving- 

 up show coloured streaks mounting the arch of the film, breaking 

 out as brilliantly coloured starlike discs, often leaving a streak 

 curving after them through the black, like a stream of meteors 

 across a dark sky. Meanwhile the coloured lower segment of the 

 film thins out and passes through the appropriate stages of colotn-, 

 while being steadily absorbed downwards, as if to feed the upward 

 moving streams in the Gibb ring. 



In an unexhausted bulb the convection currents caused by this 

 treatment give rise to quite different appearances. For example, 

 starting with a film about two-thirds black, while the lower third is 

 banded with colour, and applying the warm pad below, a current of 

 warm air then streams up across the film displacing the liquid 

 against the gravitational and surface tensional forces which tend to 

 maintain it in its initial condition. A coloured peak is steadily 

 elevated into the dark upper region, and a thin stream of coloured 

 particles begins to rise from the part so elevated, and fall again 

 round the periphery on to the disturbed coloured part below. The 

 original parallel bands have now the appearance of stiata bent into 

 finely curved patterns ; a very brilliant appearance occurring from 

 the contrast of the silvery and steel-blue portions of the film, 

 YOL. XXI. (Xo. 110) 3 G 



