C. Barns — Viscosity of Solids. 195 



effect of —Aji—A x n — A"n, and the succeeding ^-quantities, 

 respectively. 



The scheme admits of simplification; but inasmuch as the 

 period of oscillation is arbitrary, the phenomenon remains in- 

 definitely complex. 



8h. The second part of table 4c shows that the viscosity gained 

 in virtue of consecutive alternate twisting of glass-hard steel is 

 permanently gained. The results indicate some recuperation ; 

 but the amount is small in comparison with the havoc of con- 

 figurations made by twisting. Mere molecular motion has 

 therefore permanently broken the more unstable configura- 

 tions. I will note that the viscous effect of prolonged twisting 

 to and fro in case of glass-hard steel, is of the same order of 

 magnitude as the effect of prolonged annealing at 100°. This 

 indicates the importance of the motional effect in question. 



Tf, following the analogy of steel, I consider annealing a pro- 

 cess by which unstable configurations are broken up, I may 

 with due caution designate the phenomenon here in question 

 as motional annealing. Experiments for which there is no 

 room here showed that motional annealing is relatively without 

 electrical effect. For increasing rates of twist thick wires show 

 viscous deformation sooner than thin wires. Hence motional 

 break up commences at the external surface where stress is 

 most intense, and proceeds thence toward the axis where stress 

 is least Thus there appears an essential dependence on the 

 dimensions of the twisted rod. Elsewhere * I pointed out' that 

 the limits of torsional resilience of soft iron are reached when 

 the obliquity of the external fiber (shear) somewhat exceeds 

 •003 radians. Regarding the laws of* motional annealing, cf. 

 § 16. It follows from the absence of electrical effect, that mo- 

 tional annealing probably presents a pure case of Maxwell's 

 " break up " of configurations of molecules. 



Sc. Streintz and Wiedemann's phenomenon "accommoda- 

 tion " admits of representation from a different point of view. 

 Returning to the data of table 4, suppose the experiment so 

 conducted that the twelve twists are immediately consecutive. 

 Suppose furthermore that time, m, instead of being reckoned 

 positively onward from the beginning of each of the said twists, 

 were reckoned alternately positive and negative, conformably 

 with the sign of the deformations (<p — <p f )/T. In this oscilla- 

 tory march (time as abscissa), since each deformation (ordinate) 

 now begins where the preceding deformation ceased, a contin- 

 uous series of open cycles is necessarily generated. The posi- 

 tions of these cycles shift at a gradually retarded rate, toward a 

 final very flat cycle, which for the constant values of time and 

 stress is fixed in position and closed. 



* This Journal (III), xxxiv, p. 183, 1887. 



