1865.] 



Prof. Guthrie on Bubbles. 



TABLE ft. 



From glass, air-bubbles through water. 

 T=23C. 

 B=767millims. 



27 



Hence, under these conditions, rate has little or no influence upon bubble- 

 size. In order to see whether a tube of different calibre would give rise to 

 bubbles more sensitive in regard to their rate, a narrower orifice at h 

 was employed. The flask M had a capacity of 50 cub. centims. The 

 following mean results were obtained, each mean being derived from two 

 experiments : 



. Number of bubbles in Absolute volume of 



50 cub. centime. single bubble. 



/( cub. centim. 



0-33 1927 0-02595 



1-00 1945 0-02571 



This result, taken together with Tables a and ft, shows how small is the 

 effect of rate upon bubble-size. If anything, there is, on the whole, a very 

 slight tendency to diminution in bubble-size asfft diminishes that is, as the 

 rate increases. This is just the reverse of what was found to be the case 

 with SLG. Most probably, however, this effect is not due specifically to 

 the rate, but to the alteration in the diameter of the orifice at different 

 rates. When a rapidly succeeding series of bubbles passes through the 

 orifice h, the sides of the delivery-tube are swept more completely dry than 

 when the bubbles pass more slowly ; so that in the former case the 

 opening is, in fact, a little larger than in the latter. We shall see in the 

 sequel how sensitive bubble-size is to variation in the width of the delivery- 

 tube. 



It may be here noticed that, unless the tube H remains strictly in the 

 same position, it is hopeless to attempt to get uniformity in results. This 

 is especially the case when the opening h is turned half up in the shape of 

 a siphon ; for then the least displacement out of the vertical causes virtu- 

 ally an alteration in the available size of the opening, and a consequent 

 variation in bubble-size. A great and otherwise unaccountable variation in 

 the bubble-number, under circumstances apparently identical, directed 

 attention to this source of error. By taking a wider tube, and allowing 

 the end to contract in the blowpipe flame, a rounded opening is produced, 

 the horizontal projection of which is much less variable with alteration in 

 the verticalness of the tube H. 



