* 
Floating Bodies and the Theory of Capillarity. 309 
it must be borne in mind, that in problems of this character, no 
deduction from analysis is worthy of confidence which does not 
admit of a rational physical interpretation, capable of being 
tested by observation or experiment. 
Restricting myself to the consideration of the physical cause 
of the horizontal motions of small floating bodies when brought 
near to one another, it is evident that if the surface tension is 
Precisely the same in all parts of the liquid surface and is not 
at all modified by the formation of the adjacent meniscuses,— 
and further, if the angle of contact of the liquid with the solid 
remains constant—the horizontal as well as the vertical com- 
Eapent of the elastic reaction of the intervening tense film must 
e independent of its radius of curvature. Now, the question 
1S, are these two conditions or postulates realized in the class 
of phenomena under consideration? Some payee have 
thought it possible that the angle of contact might vary with 
the curvature of the film adjacent to the wall of the solid; but 
the invariability of this angle, at least in the case of liquids 
Which wet the surfaces of the solids,—seems to be generally 
nen . 
_ 4n fact, the assumption of the constancy of 
liquid films at given temperatures, un 
