Barus and Strouhal— Viscosity of Steel. 35 
stress values it ultimately approaches very closely to or even 
exceeds the maximum viscosity of steel; that it is much 
greater than the viscosity of hard steel. 
4, With these accordances before us, we may venture a final 
suggestion. The viscosity of iron increases with great rapidity 
with the amount of deformation. For a certain interval of 
(low) values of apphed torsional stress it even exceeds the 
maximum viscosity of steel. Hence if there be an inherent 
bearing of viscosity on retentiveness, there is given us in iron 
a condition of magnetic instability: if relatively to the (decreas- 
ing) amount of deformation or of magnetization viscosity de- 
creases rapidly, then a mere shock may at once deprive the 
magnet of the whole of its induction. The retarded curvature 
of the steel curves (tables 1 to 26) does not invalidate this 
inference, since retardation is here an expression of the fact 
that the absolute amount of change of form for each rod de- 
creases as time increases. 
We do not wish to push these inquiries into further detail at 
present because we have experiments with iron now in pro- 
gress. Nevertheless we invite the reader to peruse for compar- 
ison Mr. Hwing’s magnetic results as summarized by him, I. «., 
pp- 962, 563. 
Conclusion— Among the chief results of this paper is the 
light thrown on the crucial importance of the physical changes 
which steel undergoes during annealing at high temperatures, 
1. e. when subjected to the action of tem peratures between 500° 
and 1000°. It is within this interval that a variety of nearly 
coincident phenomena occur: Gore’s* sudden volume expan- 
sion; Tait’st sinuously broken thermoelectric diagram - line; 
we infer an irregular behavior of electrical resistance; Gore- 
Baur’st sudden disappearance of magnetic quality; the passage 
of carbon from uncombined to combined; Jean’s§ critical ce- 
mentation temperature. Furthermore, Chernoff-Barus’|| sudden 
appearance of hardness in quenched steel, Fromme’s{ maxi- 
mum density, Chwolson’s** minimum resistance, our own unique 
maximum of magnetizationtt and probable maximum of vis- 
cosity, are all referred to this interval, approximately to the 
* Gore: Proc. Roy. Soc., xvii, p. 260, 1869. 
+ Tait: Trans. Roy. Soc., Edinburgh, xxvii, 1872-73, p. 125. 
¢t Gore: Phil. Mag., IV, xxxviii, p. 59, 1869; Baur. Wied. Ann., xi, p. 408, 
1880; Ibid., xl, p. 170, 1870. 
§ Jeans: ‘Steel, its history, ete.’”’ London, Spon, 1880. 
|| Chernoff: Vortrag. geb. in der Russ. techn. Ges., April and Mai, 1878. 
Barus: Wied. Ann., vii, p. 405, 1879. 
4] Fromme: Wied. Ann., viii, p. 354, 1879. We were unable to find Fromme’s 
maximum: but we shall search again. All these results are crucially dependent 
on the method of annealing, as stated in the text. 
** Chwolson: Carl’s Rep., xiv, p. 26, 1878. 
++ Barus and Strouhal: U. 8. Geol. Survey, Bull. 14, pp. 145, 198. 
