THE PHENOMENA OF RUPTURE AND FLOW IN SOLIDS. 
183 
diameter. As a matter of fact, however, the shattering is probably merely one of the 
means of dissipating the strain energy of the strong fibre, which at fracture is perhaps 
10,000 times that of silica in the ordinary weak state. An elastic wave is doubtless 
propagated from the original fracture, and the stresses due to this wave shatter the 
rod. 
Confirmation of this view is obtainable if the fibre is broken by twisting instead of 
by bending. The thick part of the rod is in this case found to contain a number of 
spiral cracks, at an angle of about 45° to the axis, showing that the material has broken 
in tension, but the cracks run in both right- and left-handed spirals, so that the surface 
of the rod is divided up into little squares. This shows that the cracking must be due 
to an alternating stress, such as would result from the propagation of a torsional wave 
along the rod. 
Another phenomenon which has been observed in these fibres is that fracture at any 
point appears to cause a sudden large reduction in the strength of the remaining pieces. 
Thus, in one case a glass fibre was found to break in bending at an estimated stress 
of 220,000 lbs. per sq. inch. One of the pieces, on being tested immediately afterwards, 
broke at about 67,000 lbs. per sq. inch. 
7. Molecular Theory of Strength Phenomena. 
From the engineering standpoint the chief interest of the foregoing work centres 
round the suggestion that enormous improvement is possible in the properties of 
structural materials. Of secondary, but still considerable, importance is the demon¬ 
stration that the methods of strength estimation in common use may lead in some cases 
to serious error. 
Questions relating to methods of securing the indicated increase in tenacity, or of 
eliminating the uncertainty in strength calculations, can scarcely be answered without 
some more or less definite knowledge of the way in which the properties of molecules 
enter into the phenomena under consideration. In this connection it is of interest to 
enquire whether any indication can be obtained of the nature of the properties which 
are requisite for an explanation of the observed facts. 
For this purpose it is convenient to start with molecules of the classical type, whose 
properties may be defined as {a) a central attraction between each pair of molecules 
which decreases rapidly as their central distance increases, and which depends only on 
that distance and the nature of the molecules ; (6) translational and possibly rotational 
vibrations whose energy is the thermal energy of the substance. In the unstrained 
state, the kinetic reactions due to ( b) balance the central attractions (a). 
In a‘body composed of such molecules, the flaws which have been shown to exist 
in real substances might consist of actual cracks. But experiment shows that under 
certain conditions the strength of glass diminishes with lapse of time. On the 
present hypothesis this would require the potential energy of the system to increase 
