178 
MR. A. A. GRIFFITH ON 
it should be, approximately at least, a constant of the material. Traube* gives the 
following as the intrinsic pressures of a number of metals, at ordinary temperatures :■— 
Table IV. — Intrinsic Pressures of Metals (Traube). 
Metal. 
Intrinsic Pressure. 
Nickel. 
lbs. per sq. inch. 
4-71 x 10 6 
Iron. 
4-70 X 10 6 
Copper . 
3-42 x 10 6 
Silver . 
2-34 X 10 6 
Antimony. 
1-74 x 10 6 
Zinc. 
1*58 X 10 6 
Tin. 
1-06 X 10 6 
Lead. 
0-75 X 10 fl 
These are of the same order as the direct estimate obtained above for glass, but they 
are from 20 to 100 times the strengths found in ordinary tensile and other mechanical 
tests. 
In the case of liquids, the discrepancy between intrinsic pressure and observed tensile 
strength is much greater. According to Van der Waal’s equation, water has an 
intrinsic pressure of about 160,000 lbs. per sq. inch, whereas its tensile strength is found 
to be about 70 lbs. per sq. inch. It has been suggested that this divergence may be 
due to impurities, such as dissolved air, but Dixon and Jolv|* have shown that dissolved 
air has no measurable effect on the tensile strength of water. 
Thus the matters under discussion appear to be of general incidence, in that the 
strengths usually observed are but a small fraction of the strengths indicated by the 
molecular theory. 
Some further discrepancies between theory and experiment may now be noticed. 
In the theory it is assumed that rupture occurs in a tensile test at the stress corresponding 
with the maximum resultant pull which can be exerted between the molecules of the 
material. On this basis the applied stress must have a maximum value at rupture, 
and hence, if intermolecular force is a continuous function of molecular spacing, the 
stress-strain diagram must have zero slope at that point. This, of course, is never 
observed in tensile tests of brittle materials ; in no case has any evidence been obtained 
of the existence of such a maximum anywhere near the breaking stress. 
Again, the observed differences in strength as between static and alternating stress 
tests are at first sight inexplicable from the standpoint of the molecular theory, if the 
* * Zeitschr. fur Anorganische Chemie,’ 1903, vol. xxxiv., p. 413. 
f ‘ Phil. Trans.,’ B, 1895, p. 568. 
