32 Barus and Strouhal— Viscosity of Steel. 
the substance of greater viscosity, must have retained a similarly 
high intensity of strain. This reasoning, however, is incom- 
plete and must be approached with caution. We pointed out* 
that the low degrees of thermal conductivity possessed by glass 
were favorable to the retention of strain. We also remarkedt 
that the occurrence of Gore’s phenomenon of sudden expansion 
at red heat distinguished iron and steel from all other metals, 
and that in spite of relatively good thermal conductivity, iron 
and steel possessed virtually all the favorable conditions of 
glass for retaining strain. Whether differences of viscosity are 
sufficient to account for the unlikeness of behavior during 
quenching we are unable to say. It will be necessary to com- 
pare iron and steel at higher than ordinary temperature and 
also under greater values of applied stress than was easily fea- 
sible in the above experiments. In other words, iron and soft 
steel are nearly equally viscous for small values of stress. 
After stress exceeds a certain intensity, the permanent set sud- 
denly imparted to iron is enormously great as compared with 
steel. Similar relations are true for soft steel when compared 
with hard steel. It is this class of ‘‘sudden” phenomena 
which come into play during quenching. The primary effect 
of quenching steel is chemical hardness. Strains are retained 
in the steel so modified, just as they are in glass: whereas in 
soft steel or soft iron the result would be permanent set. The 
occurrence of sudden and gradual deformation in a single sub- 
stance suggests that ordinary séatic friction is probably a viscous 
phenomenon. 
We add in passing that the observed inefficiency of tempera- 
tures less than 200° in producing marked viscous deformation 
in a Rupert’s drop proves that mere interference of thermal 
expansion with the conditions of strain cannot be the primary 
cause of its variations; that strain variations essentially depend 
on diminished viscosity. 
We also add in passing that the importance of strain as asso- 
ciated with glass-hardness is emphasized by the mass-constants 
of the cast-irons. The densities of these metalst range between 
the maximum of ca. 7°6 for white cast iron and the minimum of 
ca. 6°9 for gray cast iron. Hence density increases in marked 
degree in proportion as total carbon is more and more nearly 
combined. Quite the reverse of this is true for steel where 
density decidedly decreases as total carbon is more and more 
nearly combined. ‘This discrepancy we interpreted as a strain 
of dilitation and carefully compared it with the analogous be- 
havior of glass in our earlier papers 
* This Journal, xxxi, p. 451, 1886. + Bull. 14, p. 99. 
+ Cf. Bloxham’s Chemistry, p. 342, Lea, Philad., 1873. Bull. 14, pp. 76, 77. 
