118 SIR DAVID BREWSTER ON THE MOTION, ETC., OF LIQUID FILMS. 



themselves. If we now draw the bubbles gently apart, they will form two 

 truncated cones united by a small circular film at their apices, which may some- 

 times be reduced to a mere point before bursting, thus forming two perfect cones, 

 as shown in fig. 16. 



If, instead of bringing the two convex films into direct connection with one 

 another, we unite them by means of a small bubble placed between them, this 

 intermediate small bubble will assume the barrel shape shown in figs. 17 and 18. 

 If we now decrease the distance between the large convex films, the small bubble 

 will assume the flattened form of fig. 17. If we increase the distance between 

 the large films by drawing them apart, the small bubble will become elongated, 

 as in fig. 18. The larger end of the barrel will always be that in connection with 

 the film of least convexity; in other words, with the film whose radius of 

 curvature is the greater, and when the system is drawn apart so as to break the 

 chain of connection, it is to this film (that, namely, of greater convexity) that the 

 small bubble will always adhere. 



The form of the barrel is always that of a more or less perfect cylinder, all of 

 whose bounding lines are convexly curved. 



When the large convex films have their radii exactly equal, the barrel- 

 shaped cylinder will be perfectly symmetrical, i.e., the circles at its upper and 

 lower ends will be equal. When this is the case, or nearly so, the system can be 

 drawn apart without suffering disruption till the ends of the cylindrical bubble 

 become almost mere points. 



To make the two large films of exactly equal convexity is of course impossible 

 practically, but theoretically it would appear that in such a case the small 

 central bubble ought to part connection with each of the large bubbles simul- 

 taneously, and fall, resuming its original form of a perfect sphere. 



If a cylindrical or conical tube, such as that used in the previous experiments, 

 be slightly dipped into the solution, and then gradually raised vertically, the film 

 formed on its lower aperture will remain attached to the surface of the liquid. 

 The curvatures of the film as we gradually raise the tube are very curious. At 

 first, while the surface of attachment is large, the film is convexly curved ; its 

 sides then become straight lines ; then concave, then, just before parting with the 

 surface of the liquid, they become slightly convex above, concave below. In 

 fig. 19 these changes of curvature are shown. 



If the liquid into which we dip the tube be shallow and small in quantity 

 (that, for example, contained in a watch-glass), it will be elevated bodily when 

 we raise the tube, its whole surface becoming more and more convex as the film 

 becomes more concave. If we insert a film half way up the tube before making 

 this experiment, it will sink in the tube whilst the lower film is convex, and 

 again rise when the lower film begins to become concave. 



