210 Mr. C. Bams on Maxwell's Theory of the 



fluid. I have shown * that if a glass -hard and a soft rod 

 (cwt. par.) be identically twisted and heated to 350°, the 

 former will have lost the whole of its strain, whereas in the 

 soft rod only about one third will have vanished. 



14. Viscosity in the above pages has been considered apart 

 from the stress (intensity) under which the viscous deformation 

 takes place. This is apt to lead to confusion, unless the stress- 

 intensity, relative to which the constants of viscosity are de- 

 fined, be clearly kept in mind ; or unless the term viscosity 

 be applied to solids in the restricted sense of " elastische 

 Nachwirkung." Thus, if a glass-hard and a soft steel rod be 

 subjected alike and at ordinary temperatures to torsional 

 stress of continually increasing magnitude, a stress-value 

 will be reached for which the viscosity of the hard rod 

 will be equal to, and eventually overtake the viscosity of the 

 soft rod. I was able to exhibit this phenomenon in even a 

 more striking way at 190° ; finding that for rates of twist less 

 than t= 3°, the steel rod (radius = 0*041 cm.) is much less 

 viscous and, as regards viscosity, much more susceptible to 

 the influence of temperature in proportion as it is permanently 

 harder f; whereas for rates of twist greater than t = 6°, steel, 

 ccet. par., is less viscous and more susceptible to temperature 

 in proportion as it is softer. 



Here I may profitably insert certain considerations postu- 

 lated in an earlier paper J : — Suppose stress to be so distri- 

 buted in a solid that its application at any interface is nowhere 

 sufficient to produce rupture. Then that property of a solid 

 in virtue of which it resists very small forces (zero-forces) 

 acting through very great intervals of time (go -times) may be 

 termed the viscosity of the solid. That property in virtue of 

 which it resists the action of very large forces (qo -forces rela- 

 tively) acting through zero-time may be termed the hardness 

 of the solid. Since the application of forces in such a way 

 as accurately to meet either of these cases is rare, we have, in 

 most practical instances, mixtures of viscous resistances and of 

 hardness to encounter. We may reasonably conceive that in 

 the case of viscous motion the molecules slide into each other 

 or even partially through each other per interchange of atoms, 

 so that the molecular configuration is being continually re- 

 constructed ; that in the other case (hardness) the molecules 

 are urged over and across each other, and that therefore the 

 intensity of cohesion is in this case more or less thoroughly 



* Amer. Journ.[3]xxxiv. pp. 4, 5, 1887. Experiments made by annealing 

 twisted systems. 

 f An. 190° being, of course, the maximum hardness admissible, 

 t Amer. Journ. [3] xxxiii. p. 28, 1887. 





