Determination of the Surface Tensions of Liquids, 113 



the bath, the movable tube, preferably the one with the 

 smaller bore, is raised by turning the screw. This may be 

 done rapidly while bubbles are being forced out of the larger 

 tube every second or so. At a certain point, bubbles may 

 cease to issue from the larger tube and commence to come 

 from the smaller one, an indication that the correct adjustment 

 has been passed. The screw is now turned very cautiously in 

 the opposite direction until bubbles issue at about the same 

 time from both tubes. The final adjustment is generally best 

 made as follows : 



The air is compressed very slowly until it has arrived just at 

 the orifices of both tubes. The pressure is then not increased 

 further, and, as a rule, it will be observed that, if the adjust- 

 ment is perfect, bubbles will issue of their own accord from 

 both the tubes in a second or so ; sometimes one of the bubbles 

 lies a little behind the other. The success of the operation 

 depends upon the steady and gradual compression of the air ; 

 if this be subjected to abrupt or irregular changes of pressure, 

 not inconsiderable errors may be committed. It is advisable 

 to make at least two readings, one when the narrower tube is 

 being lowered, the other when it is being raised ; indeed, the 

 readings can be made so rapidly that it is a good plan to take 

 a half dozen or so, the tube being moved alternately up and 

 down. 



lY. Discussion of Soueces of Error in Apparatus. 



In the determination of the capillary constants by means of 

 the apparatus described in the preceding sections, measure- 

 ments of three kinds have to be made, viz., measurements of 

 specific gravity, of temperature, and of the distance between 

 the lower extremities of the capillary tubes. We pass to the 

 discussion of the errors inherent in each. 



I. The determination of the surface tension by any of the 

 direct methods such as measuring the height to which liquids 

 rise in capillary tubes, etc., cannot be said on an average to be 

 accurate to more than one part in two thousand and only in 

 exceptional cases to one part in ten thousand. On the other 

 hand, the accuracy of a specific gravity determination can 

 easily attain one part in ten thousand and in many cases to 

 even one part in a hundred thousand. In the formula to be 

 communicated in the following section, it will be seen that the 

 influence of the specific gravity upon the value of the capil- 

 lary constant is not very great, so that a determination of the 

 specific gravity to one part in two thousand is perhaps sufla- 

 ciently accurate for our purposes, although it is of course 

 better to be sure of the fourth decimal place. We may con- 



