of the Molecular Forces of Liquids. 437 



then turned very gently around its sharp edge so as to put the 

 globule, placed very near this edge, into contact with the water. 

 This contact established, the knife was carefully withdrawn, and 

 the globule of mercury left floating on the surface of a liquid 

 thirteen and a half times less dense than itself. 



This phenomenon suggests several remarks. In the first place, 

 why is the globule not made wet by the liquid ? The fact is 

 due, I believe, to the layer of air condensed at the surface of the 

 globule: this explanation appears to be the more probable, 

 since I was able to make the globule float immediately after it 

 had remained immersed in the water for more than a quarter of 

 an hour, though not after it had been submerged for an hour or 

 upwards ; the layer of air in the latter case had been displaced, 

 or at least partially so. In the second place, it may be asked, 

 will it suffice, in explaining the present phenomenon, to say that 

 the weight of the mercurial globule is equal to that of the dis- 

 placed water, the depression formed around the mercury being, 

 of course, included ? Must we not take into account the fact 

 that, the water being concave immediately beneath the globule, 

 the capillary pressure there must be less than at surrounding 

 points? or does there come into play in this case a special 

 effect of cohesion, for instance a resistance opposed by the sur- 

 rounding liquid to the deformation of its surface ? I have made 

 a great number of experiments and calculations in order to elu- 

 cidate these questions decisively, but up to the present time I 

 have not been successful. 



As already remarked, the experiment above described furnishes 

 a very convenient method of showing, clearly, capillary attractions 

 and repulsions. In fact at the moment of withdrawing the 

 blade of the penknife, the globule of mercury is seen to suffer a 

 quick repulsion. This is evidently a capillary effect due to the 

 elevation of the water along the blade, and to the depression of 

 this liquid around the mercury. The sides of the capsule also 

 exert an energetic repulsion. To prove attraction, I endeavoured 

 to make two globules float in such a manner that, when still, 

 their distance asunder might be about 20 millimetres. In a few 

 moments they moved towards each other with a velocity which, 

 from being very small at first, increased rapidly as the distance 

 between them diminished. Immediately after contact the two 

 globules united and formed one, the layer of air adhering to each 

 of them having been so far displaced by the shock as to allow 

 cohesion to produce its effect. This union of the globules does 

 not readily take place unless the mercury is sufficiently pure^ 

 and the surface of the water free from small filaments and cor- 

 puscles ; for the latter interfere considerably with the capillary 

 actions, and render the distance between the mercurial surfaces 



