1865.] Prof. Guthrie on Bubbles. 29 



The chemical nature of the gas, therefore, has also very little influence 

 upon bubble-size. The two purely physical influences active in determin- 

 ing the bubble-size are the density of the gas and its solubility in water. 

 These act to produce opposite effects. Increase of density in the gas 

 delays the departure of the bubble, and thereby increases its size ; increase 

 in the solubility of the gas in water impairs the stubborn cohesion of the 

 water, and thereby diminishes the bubble-size. If p and q be the specific 

 gravities of two gases P and Q referred to water, the buoyancy of two equal 

 bubbles of them will be respectively 



where W is the weight of an equal volume of water ; that is, W (q p) is 

 the difference in buoyancy of the two bubbles. The gases arranged in 

 their order of density are 



C0 2 , O, Air, N, H. 

 Arranged in order of solubility (at 20 C.), 



CO 2 , O, H, Air, N. 



The properties density and solubility are of course incommensurable, 

 so that we cannot predict the extent to which they may counteract one 

 another in the same gas to determine its bubble-size. But the order of the 

 gases in Table y is quite consistent with our previous knowledge. Thus 

 the bubble- size of air is intermediate between the bubble- sizes of nitrogen 

 and oxygen. It would, however, at present be premature to attempt to 

 make use of bubble-size to furnish an additional equation in gas-analysis. 



Effect of temperature and of tension. The first of these has also a 

 twofold action, by changing the density of the gas, and by changing the 

 cohesion of the liquid. Within a natural range of 10 C. change of tem- 

 perature takes no appreciable effect upon bubble-size. Also a variation of 

 three-quarters of an inch in the natural barometric mercurial column is 

 without sensible influence. These two influences were not made matters of 

 special study, but were only examined with the view of ensuring absence of 

 error from other experiments. 



Effect of change in the geometrical distribution of solid : size of orifice. 

 The change examined in this sense was the alteration hi the size of the 

 orifice through which the gas bubbled. For this purpose the ends of six 

 tubes of various internal diameter were ground flat, and until they had 

 exactly the same length. One end of each tube was stopped by a little 

 glass disk covered with a film of wax. The tube was then filled to over- 

 flowing with distilled water, and another little disk was pressed on the 

 top, the superfluous water being wiped off. The tube was then weighed, 

 emptied, and dried and reweighed. The same being done for each tube, the 

 volumes of the tubes are known to be in the same proportion as the weights 

 of their liquid contents, the diameters or radii of the tubes being in the 

 ratio of the square roots of the same weights. To calibrate tubes in this 

 manner, water is to be preferred to mercury, because the latter leaves a film 

 of air between itself and the glass, and thereby introduces a considerable 

 error in the deduced calibre of very narrow tubes. The tubes were inserted 



