182 
MR. A. A. GRIFFITH ON 
In 1858, Karmarsch* found that the tensile strength of metal wires could be repre¬ 
sented within a few per cent, by an expression of the type 
p = A +! 
( 22 ) 
where d is the diameter and A and B are constants. In this connection it is of interest 
to notice that the figures in Table V. are given within the limits of experimental error 
by the formula 
F = 22,400 
4.4 -f- d 
0.06 + d 
(23) 
where F is in lbs. per sq. inch and d is in thousandths of an inch. Within the range 
of diameters available to Karmarsch, this expression differs little from 
F = 22,400 + 
98,600 
d 
5 
(24) 
which is of the form given by Karmarsch. Moreover, the values of B found by him 
for the weaker metals, e.g., silver and gold, in the annealed state, are of the same order 
as that given by formula (24) for glass. 
To a certain extent this correspondence suggests that the mechanism of rupture, as 
distinct from plastic flow, in metals, is essentially similar to that in brittle amorphous 
solids such as glass. 
The remarkable properties of the unstable strong state referred to on p. 180 above 
are exhibited most readily in the case of clear fused silica. If a portion of a rod about 
5 mm. diameter be made as hot as possible in an oxyhydrogen flame, and then drawn 
down to, say, 1 mm. or less and allowed to cool, the drawn-down portion may be bent 
to a radius of 4 or 5 mm. without breaking, and if then released will spring back almost 
exactly to its initial form. If instead of being released it is held in the bent form it 
will break spontaneously after a time which usually varies from a few seconds to a 
few minutes, according to the degree of flexure. To secure the best results the drawing 
should be performed somewhat slowly. 
When fracture occurs it is accompanied by phenomena altogether different from those 
associated with the fracture of the normal substance. The report is much louder and 
deeper than the sharp crack which accompanies the rupture of an ordinary silica or 
glass rod, and the specimen is invariably shattered into a number of pieces, parts being 
frequently reduced to powder. This shattering is not confined to the highly stressed 
drawn-out fibre ; it usually occurs also in the unchanged parts of the original thick 
rod and sometimes in the free ends, which are not subjected to the bending moment. 
The experiment is most striking, for it appears at first sight that the 5 mm. rod has 
been broken by a couple applied through the fibre, which may be only 0-5 mm. in 
* 1 Mittlieilungen des gew. Yer. fur Hannover,’ 1858, pp. 138-155. 
