156 Mr. C. Barus on the Viscous Effect of 



of identity from an initial to a final configuration ; or it is 

 a break-up superinduced by the disintegration of one or more 

 molecules of each group. Whichever the change may be, 

 it must, from the nature of the problem in general, be dis- 

 tributed uniformly throughout the mass of the solid (§ 9). 

 Even without stress the said change may result from secular 

 subsidence. Moreover, it is conceivable that molecular disin- 

 tegration may occur in such a way as to elude detection. 



Now I have since been able to prove experimentally that 

 in glass-hard steel a change of the viscous quality may bo 

 obtained as the result of at least tioo distinct kinds of struc- 

 tural change, probably as the result of the two kinds of 

 break-up here in question. The first section of the present 

 paper purposes to show this by aid of the phenomena of 

 accommodation, or motional annealing, as they may be called 

 more uniformly in keeping with the following work. 



2. The term annealing in its most general sense refers to 

 a process by which strained solid structure, whether main- 

 tained by mechanical or chemical causes, is changed to 

 isotropic structure. In viscosity, inasmuch as strained 

 structure is ultimately accompanied by molecular instability, 

 annealing is a process by which viscosity is increased ; and 

 from this point of view annealing need have no direct refer- 

 ence to exposure to temperature. Hence I have designated 

 by the term motional annealing all such forced molecular 

 motion to and fro, in virtue of which the molecules of a 

 thoroughly soft solid are brought into new relations to each 

 other, to the effect that viscosity is increased at the expense 

 of the motionally less stable configurations of the soft solid. 

 There may appear to be some incongruity in the term, inas- 

 much as the solid motionally acted on always experiences 

 strain ; it is not the strain, however, but the increment of 

 viscosity of the solid, to which the term refers'^. 



Again, in order that a solid may be motionally annealed 

 the mechanical treatment (torsion, traction, &c.) must be 

 applied heloio a certain critical limit of intensity. Otherwise 

 this treatment introduces its own specific instability ; and in 

 proportion as stress is indefinitely increased, the viscosity of 

 the solid may now be reduced in any measure. Further 

 justification of the term is to be given in § § 6,9. At present 

 it is more expedient to indicate the points of crucial difference 

 between motional annealing and thermal annealing in their 

 effect on steel in the glass-hard state. 



3. When glass-hard steel is annealed at 100°, the effect is 



* Phil. Mag. [5] xxvi. pp. 199 to 203, 1888. 



