1917] on Soap Bubbles of Long Duration 205 



diameter, tied together bj thin akiminium wire ; but the drawn-out 

 glass rod C acts automatically when placed vertically just below 

 the drop, for, as this grows, it descends by its own weight, and, 

 enveloping the rounded end of the rod, drains down along the 

 drawn-out neck, and the lightened l)ubble rises off the rod. 



After the bubble has thus drained for several hours, the boiled 

 distilled water is run in. The collecting tube may be utilised for 

 this by turning down the nozzle and tilting the vessel, or a separate 

 inlet tube with stop-cock may be fitted. 



In order to isolate the bubble from the soap solution in A, and 

 prevent any distillation or intrusion of extraneous liquid, the reservoir 

 in some experiments was removable. For this purpose a conical 

 ground joint was made in the tube A above the india-rubber cork. 



The contraction of the bubble is somewhat irregular, because of 

 the constantly varying thickness of the film, shown by the beautiful 

 changes and gradings of colour. In the case of an air bubble the 

 contraction is also relatively slow. Thus (Fig. 12) one of 12 cm. 

 diameter only diminished in forty days to 11 cm. A second one, of 

 21*4 cm. diameter, contracted in five months to 16 '6 cm. A third 

 one, of 15 "8 cm. diameter, took a Uttle over six months to contract 

 to a plane film on the supporting nozzle. In wet hydrogen, however 

 (Fig. I'd), an 11 cm. bubble, — under conditions of somewhat lower 

 temperature, — contracted completely in about three months; almost 

 all this time, moreover, it was of sufficient thickness to show strong 

 colours. 



The form of the contraction curve of the hydrogen bubble is 

 fairly regular (considering the variations of thickness as shown by 

 the colour changes), and of similar form to that given by black 

 bubbles of constant composition. Its collection curve, of similar 

 type, in which the amounts collected by drainage are plotted as 

 ordinates to the time of collection as abscissae, is also shown in 

 Fig. l:]. 



Similar contraction and collection curves are shown in Fig. 12 

 and Fig. 13 for air and hydrogen respectively. The left-hand vertical 

 scales are for cms. in diameter for the contraction curves ; the right- 

 hand scales are for grams of liquid collected for the collection curves. 

 The collection curve appears roughly as an image of the contraction 

 curve, showing that, whereas the contraction is slow at first, increasing 

 parabolically to the final disappearance, the condensation and drain- 

 age begins rapidly when the bubble contains concentrated glycerine, 

 and falls off to zero as the dilution asymptotically approaches 

 completion. 



Careful measurement of such curves given by both air and 

 hydrogen bubbles revealed them to be parabolic during the initial 

 period before the surface became sensibly reduced by contraction. 

 From this it follows that during this period the rate of collection is 

 inversely as the total amount collected ; just as in the contraction 



