Trouton — On the Creeping of Liquids ^ etc. 5 



1^0 simple relation, however, could be found connecting the 

 depression with the properties of the pure substances. In some 

 few cases, it should be remarked, it was found that the effect of 

 one substance on another was roughly in inverse proportion to their 

 molecular weights. Thus, the depression produced on the surface- 

 tension of chloroform below the calculated value by the addition of 

 small quantities of alcohol is at the rate of about 1"7 for each 

 percentage added, while the corresponding depression produced by 

 chloroform on alcohol is only about '7. These are in the ratio of 

 about 2'4, which is also nearly the ratio of their molecular weights, 

 59*5/23 = 2-5. The effects produced are, in this case at all events, 

 simply proportional to the number of molecules added, and the 

 failure in general to find similar relationships holding with other 

 liquids may perhaps be due to the masking of the effect through 

 molecular association. 



A number of experiments have been made to ascertain if there is 

 any specific effect produced by the kind of material over which the 

 creeping takes place : this both in the initial stages and for con- 

 tinuous creeping. !N"o consistent quantitative results have been so 

 far reached. This is probably to be attributed to the difficulty of 

 presenting clean and unaltered surfaces for the liquids to creep over. 

 The experiments, however, undoubtedly point to a decided difference 

 between different metals, both in the rate the creeping goes on at 

 in the continuous stage, as well as in the initial stage. 



That the latter should be the case is not surprising, but it is not 

 easy to see how the kind of material the surface is made of can have 

 effect once the layer of liquid has become established, for its thickness 

 is found to be great compared to molecular distances. 



It is possible, however, that the effect may be wholly due to 

 specific roughness or corrugosity incident or natural to surfaces pre- 

 pared from different materials, for the state of roughness of a given 

 metal is found to have a great influence on the effect. 



Attraction between the solid and liquid is a necessary condition 

 for creeping to occur at all; thus, that the rate the liquid in the 

 first instance establishes the layer should be dependant upon the 

 material over which the creeping occurs, is not surprising. 



