CAPILLARY CONSTANTS 43 1 



in the case of many liquids that their contact-angles are zero 

 if they " wet glass." What is meant by the phrase " wetting 

 glass" is not usually defined, nor is it easy to frame a clear 

 definition. The mere spreading of the liquid is a very shaky 

 criterion, and there is nothing so markedly different between the 

 behaviour of benzene and of ether when poured on to horizontal 

 sheets of clean glass as to indicate that the one has a zero 

 contact-angle, the other a contact-angle of i6°. 



In fact, it is almost more satisfactory to reverse the definition, 

 and to define " wetting glass" as a property possessed by liquids 

 having zero contact-angles, rather than to define zero contact- 

 angle as the angle of contact of a liquid which wets glass. 

 (This assumes of course, which is quite practicable, that we have 

 measured the angle of contact of the liquid by some quite inde- 

 pendent method.) 



The capillary-rise method has been used for the determina- 

 tion of the surface-tensions of a very large number of liquids, 

 organic and inorganic, and almost all the results are of doubtful 

 value inasmuch as, in the great majority of cases, the question 

 of the contact-angle has been passed over in silence, or at best 

 the angle has been assumed to be zero on the strength of more 

 or less obscure "glass-wetting " phenomena. In some cases the 

 surface-tensions and molecular complexities of such liquids as 

 fused salts have been determined by the capillary-rise method 

 without the slightest apparent appreciation of the fact that such 

 liquids might have contact-angles large enough to render the 

 calculated results quite valueless. 



This is all the more unfortunate, since there are several 

 methods, as we shall see later, which are equally rapid and 

 easy to handle, and which are quite independent of any know- 

 ledge of contact-angles. 



But even if we assume that the contact-angle is known, 

 there exist other defects which are inherent in the method. 

 The surface-tension of the liquid under examination is, neglect- 

 ing small corrections, proportional to rh. Hence if one arranges 

 matters so that h is large, and can therefore be measured with 

 considerable percentage accuracy, r must be small, and the 

 difficulties avoided in the measurement of h are merely trans- 

 ferred to the measurement of r. As a matter of fact, it is usual, 

 where tubes of very small radius are employed, to determine the 

 radius of the tube indirectly by measuring the capillary elevation 



