CAPILLARY CONSTANTS 429 



in quality, and it is the main object of this paper to discuss these 

 methods critically, and to attempt to obtain some criterion which 

 will serve to discriminate between them. 



In attempting to describe the various methods that have 

 been used from time to time for the determination of surface- 

 tensions it is convenient to adopt some scheme of classification 

 such as is given in the table on p. 430. 



The table explains itself, and, whilst not assuming to be 

 complete, it contains, I believe, all the methods which have 

 been used at all extensively. 1 



In seeking for a criterion to discriminate between these 

 various methods we first note that the primary conditions to be 

 fulfilled by any method which can be used successfully and 

 widely must be those of a high order of accuracy, reasonable 

 rapidity in performance, perfect and easy control of temperature 

 conditions, and unimpeachable rigour in mathematical details. 



And it therefore follows that all the methods tabulated under 

 the heading " Dependent on contact-angle " stand at once con- 

 demned. 



All these methods give only T cos (where T is the surface 

 tension, and 6 the angle of contact between the liquid and — 

 usually — glass), or some other function of T and the angle of 

 contact, and therefore a separate knowledge of 6 is required 

 before T can be obtained. It will be seen that included in this 

 list are two methods, the " capillary-rise " method and Wil- 

 helmy's method, which may almost be considered classic. 



The capillary-rise method, in particular, is very widely used 

 in physico-chemical research work, and is almost the only method 

 described at any length in the great majority of physico-chemical 

 text-books. In the practice of the method, surface-tensions are 

 determined by measurement of the height h to which the liquid 

 rises in a tube of small radius r. As a simple calculation shows, 

 neglecting minor corrections, T is given by the equation 



T = rh Pg 

 2 cos 6' 



where p is the density of the liquid, and g the acceleration due 

 to gravity. If we know the value of the contact-angle for the 

 liquid under examination the method can be considered a fairly 

 reliable one. But as a matter of fact it is too readily assumed 



1 With one exception, which will be referred to afterwards. 



