Molecular Constants. 325 
of solid matter towards the end of its harmonic excursions 
is "visible with a vibrating rod whose excursions are many 
times longer than its thickness. Better with a long mono- 
chord wire. The shadow of such a wire shows most distinctly 
the increased path-density at its edges when it vibrates in a 
plane at right angles to the light. If such a cord be made to 
vibrate in a circular path, or if an upright rod be fastened ex- 
centricallv to the top of a humming-top, the shadow is also 
graduated, being densest at the edges. Here we have really 
of course the projection of a circular path of uniform density 
on a tangent to the circle, precisely as in § 12. The vibrating 
solids have skins. 
§ 14. The hypothesis which I submit, and which is quite 
independent of the facts of the preceding paragraphs, is 
briefly that, when a mass of liquid has a free surface, while 
the internal particles have paths more or less free, those that 
strike the bounding surface never have free paths unless they 
strike it at such angles as enable them to escape as vapour. They 
are momentarily at rest; and however small and elastic (in 
regard to the surface) they may be, they form a dynamic (ever 
renewed) skin, which in its turn acts as a check upon the pass- 
ing particles, delays them, and so thickens their crowd. 
There is accordingly at all such surfaces an increased den- 
sity due to diminished mean velocity; and it is this increased 
density which forms the so-called surface-tension or skin. 
Gases and vapours should have such skins at the bounding 
surface between themselves and liquids and solids; and perhaps 
it is for this reason that a solid has the power of what is vir- 
tually condensing a gas, even sometimes to liquefaction, upon 
its surface. 
§ 15. Liquid Slabs. — Whether or not the hypothesis in § 14 
as to the cause of surface-tension be correct, and whether or 
not the terms surface-tension or skin be either of them satis- 
factory, it is convenient to adopt some such expression for the 
apparent toughness of the surfaces of liquids. When a little 
liquid is poured upon a flat horizontal surface which is not 
attacked by the liquid, a circular disk of liquid is formed, the 
shape of the edge of which has been very fully examined by 
Quincke and others. In most such cases, one of the most 
important factors is the specific relationship in the sense of 
adhesion between the solid and the liquid. In fact the ques- 
tion, like all questions of capillarity, involves density (and 
gravitation), cohesion, adhesion, and surface-tension. Such 
experiments show the relationship between two bodies as well 
as the physical attributes of one. About twenty years ago I 
made an attempt to get rid of the factor adhesion, with partial 
