Barus and Strowhal— Viscosity of Steel. 33 
We add finally the following data from the elastics of glass 
and of steel. The eolotropic expansion produced by quench- 
ing glass or steel we showed elsewhere* to amount to 0-000. 
The volume resilience of glass and of steel, according to Pro- 
fessor Hverett’s measurements,t is 4x10” and 2x10” respec- 
tively. Hence if per square centimeter p be the stress-value 
for the given expansion, we find aproximately p=10X10° 
‘dynes for steel and 2Xx10° dynes for glass. Now per square 
centimeter the tenacity of steel is 810° dynes; the tenacity 
of glass 0°6x10° dynes. The approximate ratio of stress to 
tenacity is therefore estimated at 1:3 for steel and 38°3 for glass. 
This shows that in both cases stress and tenacity are of the 
same order; that stress is in excess; that but for the peculiarly 
favorable, symmetrically arched structure of the quenched 
globule, rupture would ensue in glass certainly ; probably also 
in steel. This accords with the explosive tendency of a P. R. 
drop and with the less pronounced Hability of steel to crack 
on quenching.t If therefore quenched glass and quenched steel 
are under mean stress intensities of several thousand atmos- 
pheres, then in discussing the corresponding viscous properties 
they must be brought into relation with these high values of 
pecuhar stress. 
Magnetic relations.—1. If again we abstract from the extreme 
states of hardness, we find’ that both the viscosity and the © 
moment of linear magnetization per unit of mass, of a perma- 
nently saturated steel rod, increase in marked degree from 
hard to soft. This isa singularly striking result, inasmuch as 
the conditions of magnetic stability (following Hopkinson we © 
shall call them coercive force), conditions which at first sight 
we would be inclined to associate with viscosity, decrease as 
viscosity increases. Hence permanently saturated linear mag- 
netic intensity and viscosity on the one hand, magnetic stability 
or coercive force and hardness on the other, seem to belong 
together. 
2. The minimum of permanent linear intensity of saturated 
steel rods has no viscous equivalent; but we have not yet 
studied the viscosity of extremes of hard steel minutely, nor 
have we as yet sufficiently sharp data for the magnetization of 
very long rods (length/diameter>100) in its relation to temper. 
In the extreme soft region, on the other hand, the occurrence 
* This Journal, xxxii, pp. 190, 191, 1866. By eolotropic expansion we refer 
to the part of the total expansion which produces explosiveness and the polari- 
scopic phenomenon. 
+ Everett: Phil. Trans, p. 369, 1866. The above numbers are rounded from 
Everett's values. 
¢ Mr. J. M. Batchelder has just communicated to us his interesting results 
(Journal Franklin Inst., (3) viii, p. 133, 1844), in which of twelve massive pieces 
of quenched steel, eight subsequently cracked and one actually exploded. 
Am. Jour. Sci.—THIRD SERIES, Vou. XX XIII, No. 193.—Janvary, 1887, 
3 
