172 
MR. A. A. GRIFFITH ON 
4. Experimental Verification of the Theory. 
In order to test formula (13), it was necessary to select an isotropic material which 
obeyed Hooke’s law somewhat closely at all stresses, and whose surface tension at 
ordinary temperatures could be estimated. For these reasons glass was preferred to 
the metals in common use. A comparatively hard English glass,* having the following 
properties, was employed :•—- 
Composition—-Si0 2 , 69-2 per cent. ; K 2 0, 12-0 per cent. ; Na 2 0, 0-9 per cent. ; 
A1 2 0 3 , 11-8 per cent. ; CaO, 4-5 per cent. ; MnO, 0-9 per cent. 
Specific gravity—2-40. 
Young’s modulus —9-01 X 10 6 lbs. per sq. inch. 
Poisson’s ratio—0-251. 
Tensile strength—24,900 lbs. per sq. inch. 
The three last-named quantities were determined by the usual tension and torsion 
tests on round rods or fibres about 0-04-inch diameter and 3 inches long between the 
gauge points. The fibres had enlarged spherical ends which were fixed into holders 
with sealing wax. A slight load was applied while the wax was still soft, to ensure 
freedom from bending. The possible error of the extension measurements was about 
dz 0-3 per cent., and Hooke’s law was obeyed to this order of accuracy. No “ elastic 
after-working ” was observed with this glass, though more accurate measurements 
would doubtless have indicated its existence. 
The problem of estimating the surface tension of glass, in the solid state, evidently 
requires special consideration. Direct determinations appeared to be impracticable, 
and ultimately an indirect method was decided on, in which the surface tension was 
found at a number of high temperatures and the value at ordinary temperatures deduced 
by extrapolation. 
On the accepted theory of matter, intermolecular forces in solids and liquids consist 
mainly of two parts, namely, an attraction which increases rapidly as the distance 
between the molecules diminishes, balanced by a repulsion (the intrinsic pressure), 
which is due to the thermal vibrations of the molecules. It is reasonable to assume 
that the attraction, at constant volume, is sensibly independent of the temperature ; 
this amounts merely to supposing that the attraction exerted by a molecule does not 
depend on its state of motion. On this view, the temperature variation, at constant 
volume, of the intermolecular forces is determined entirely by the change in thermal 
energy. Hence, it may be inferred, on the accepted theory of surface tension,f that 
the surface tension of a material, at constant volume, is equal to a constant diminished 
by a quantity proportional to the thermal energy of the substance. In the case of 
solids, nearly the same result should hold at constant pressure, as the temperature- 
volume change is small. 
* Supplied in the form of test-tubes by Messrs. J. J. Griffin, Kingsway, London, 
f Poynting and Thomson, 1 Properties of Matter,’ ch. xv. 
