Strains Mechanically Aijplied. 165 



effects of the stress t' (temi)erature t'), and is not influenced 

 by subsequent application of stress t' (temperature t'), pro- 

 vided r>T' {t>t'). In case of incomplete deformation 

 (partial annealing) induced by stress r (temperature t), tliis 

 law applies more fully as the ultimate effect of t [t) is more 

 nearly reached. Again, the effect of t' succeeding r {l! suc- 

 ceeding t) is more nearly ?i// as the effect of t>t' (t>t') 

 api iroaches completeness. 



To discern the cause of this detailed analogy it is sufficient 

 to call to mind, (1) that in thermal annealing viscous defor- 

 mation is produced by thermal diminution of viscosity under 

 conditions of the initially given stress stored up in the 

 tempered solid ; in motional annealing by increase of the 

 mechanical stress applied from without under conditions of 

 initially given viscosity. Finally, (2) the configurations, 

 molecular or atomic, which just break up under the action of 

 stress r or temperature i, respectively, are necessarily limited 

 by a higher order of stability, and are greater in number 

 than those just surviving under less intense conditions of 

 stress or temperature. This is the crucial feature of the 

 analogy. However unlike the instabilities may be in the 

 two cases of motional and of thermal annealing, however un- 

 like the treatment to which they are here respectively sub- 

 jected, its effect in modifying the occurrence of instability is 

 similar, and hence the similarity of viscous results. 



So far as we have observed, however, residual phenomena 

 are absent in thermal annealing of glass-hard steel, and this 

 is a point of difference between the break-up of atomic groups 

 and molecular groups. The former are not reconstructed. 

 Again, in motional annealing for increasing rates of twist, 

 thick wires show viscous deformation at earlier dates than 

 thin wii'es. The break-up commences at the external surface, 

 where stress is most intense, and proceeds thence to the 

 axis, where stress is least. The history of motional annealing 

 is therefore essentially dependent on the dimensions of the 

 deformed wire and varies for different values of radius. I 

 pointed out * that in soft iron the limits of torsional resilience 

 were reached when the obliquity of the external fibre (shear) 

 somewhat exceeds '003 radians. Finally, the ultimate anneal- 

 ing effect (time = CO ) of any temperature t acting on glass- 

 hard steel increases at a retarded rate with temperature, and 

 practically reaches the limit of variation below 350°. In 

 case of motional annealing stress may be applied in any 

 degree from without, and increasing effects obtained limited 



* Am. Juurnal, xxxiv. p. 183, lti87. 



