116 C. K Lineharger — Apparatus for the Bapid 



Water at 40°-0 Water at 20°*0 



Tubes A and B. Tubes A and B. 



ni'lO 76-09 



11'IZ 76-13 



77-14 76-10 



77-09 76-16 



77-11 76-23 



77-16 76-14 



77-13 76-16 



77-12 76-18 



76-15 



Another series made 

 a fortnight later. 



76-12 

 76-18 

 76-21 

 76-13 

 76-19 

 76-14 

 76-18 



These series of numbers have been taken from my note- 

 books almost at random, and are but a small fraction of the 

 number of readings I have made. They show that for most 

 liquids differences in the readings amounting to more than 

 ■g^ of an incli seldom occur, and when a series of readings 

 are made and their average taken, the error almost vanishes. 

 Only in the case of water were greater differences of readings 

 observed ; yet, since water has such a large capillary constant, 

 the error committed is very slight. It was observed that tlie 

 nearer the extremities of the capillary tubes were, and, conse- 

 quently, the smaller the numerical value of the reading, the 

 more concordant were the readings. 



While the three sources of error just discussed seem to be 

 almost negligible, there is another inherent, not in the appa- 

 ratus, but in the method, which stands seriously in the way of 

 obtaining reliable absolute determinations of the surface ten- 

 sions of liquids. This source of error lies in the determina- 

 tion of the " apparatus constant," which may vary from one 

 liquid to another. A more detailed discussion of it will be 

 given in the next section, after the way in which the " appa- 

 ratus constant " is determined is described (p. 122). 



Y. Calculation of Eesults. 



Having now shown the degree of accuracy attainable in the 

 readings of the apparatus under discussion, I will pass to the 

 consideration of the manner in which the results obtained by 

 this indirect method may be converted into those arrived at by 



