Viscosity of Solids and its Physical Verification. 207 



hard steel is always in a state of incipient change. A part, 

 though not the whole, of this change must be of a permanent 

 kind ; and I wish to keep in mind that at the temperature of 

 incipient annealing the heat-motion is such that broken con- 

 figurations are sometimes reconstructed. 



Inasmuch, therefore, as glass-hard steel contains a greater 

 number of unstable configurations than any other state of 

 temper at the same temperature, it follows from Maxwell's 

 theory that glass-hard steel, despite extreme hardness, must 

 be the least viscous member of the whole group of tempered 

 and soft steels. This strikingly anomalous prediction of the 

 theory is verified both by the results in Table II. and in 

 figure 1, as well as in earlier work* in a way so pronounced 

 as be irrefragable. 



If glass-hard steel is annealed at 100°, the greater number 

 of the unstable configurations are broken up in virtue of the 

 increased molecular vibration at the higher temperature. The 

 cold rod, after annealing, will show increased viscosity in 

 proportion as the number of unstable configurations has been 

 decreased. Experiment proves this in a strikingly con- 

 clusive way : the increase of viscosity thus produced is 

 marked, being nearly half the difference between the soft and 

 hard states of steel. This, too, is an observation favourable 

 to Maxwell's theory; for if there be configurations with an 

 inherent tendency to collapse at ordinary temperatures, it is 

 clear that but a small fraction of them will survive the ap- 

 plication of 100°. Moreover, the configurations broken up can- 

 not be reconstructed without the expenditure of fresh energy 

 (quenching). Since no such energy is ordinarily available, 

 the viscous properties of the annealed rod are of a permanent 

 kind. 



Again, if glass-hard steel (or steel annealed at 100°) be 

 softened by annealing at 200°, a greater number of unstable 

 groups will be broken up than in the foregoing case. The 

 viscosity of the cold rod must be considerably greater than that 

 of the harder rod. Experiment proves the viscous increase to 

 be about two thirds of the whole viscous difference between 

 hard and soft steel. Analysis gives definite evidence of the 

 occurrence of decomposition; and inasmuch as the unstable 

 groups are permanently broken, the annealed rod shows 

 determinate viscous properties. 



If glass-hard steel be annealed at 300°, 400°, 500°, &c, 

 effects of the same nature as those just discussed, but differing 

 from them in the degree of thorough removal of the unstable 

 molecular configurations, will result. 



* Am. Journ. xxxiii. pp. 2o, 26, 1887. 



