214 Mr. 0. Barus on Maxwell's Theory of the 



to the influence of change of temperature, offers sufficient 

 proof of its occurrence. The laws of annealing* hard steel 

 are as follows : — 



(1) The annealing- effect of any temperature acting on 

 glass-hard steel increases gradually at a rate diminishing 

 through infinite time ; diminishing very slowly in case of low 

 temperatures ( < 100°) ; diminishing very rapidly at first, and 

 then again slowly at high temperatures ( > 200°) ; so that the 

 highest and hardest of the states of temper possible at any 

 given temperature is approached asymptotically. 



(2) The ultimate annealing-effect of any temperature, t°, is 

 independent of the possibly pre-existing effects of the tempe- 

 rature t /0 , and is not influenced by subsequent applications of t fo , 

 provided t > t' . In the case of partial annealing at t° (time 

 finite), this law applies more fully as the ultimate effect of 

 t° is more nearly reached. 



Postulating the strain discussed in § 13, these laws follow 

 at once from Maxwell's theory ; and the explanation (rtiut, 

 mut.) is identical in character with that given in §§9 to 18, 

 with reference to the applied torsion-strain. Inasmuch as 

 annealing is accompanied by chemical decomposition, the 

 conditions under which the temper-strain is reduced are those 

 of § 13. 



The second law of annealing asserts that the heat-effect is 

 analytic, but not in the same degree synthetic. The carbon 

 configuration definitely broken up by annealing does not re- 

 combine on cooling. In a pure metal, and up to a certain 

 limiting (small) stress, configurations broken up by stress 

 may recombine when stress is released or reversed. 



17. In certain comparisons between the strain-effect exhi- 

 bited by glass and by steel t, we were led, both by gravimetric 

 and by polariscopic observations, to this distinction : the strain 

 in hard steel is very perceptibly affected by annealing-tempe- 

 ratures as low as 50°, whereas in the case of quenched glass 

 (Rupert's drops) perceptible annealing is incipient only at 200°. 

 The bearing of this result on the present discussion is manifest; 

 the difference of behaviour is due to the absence in glass of 

 anything equivalent to the unstable carbon configuration in 

 hard steel. The case of glass is nearly that of soft steel, and 

 the behaviour as regards viscosity in these two instances is 

 similar. 



Schroeder's J important result has relevancy here ; in the 



* Phil. Mag. [5] viii. p. 341, 1879; Wied. Ann. xi. pp. 962, 965, 967, 

 1880 ; Bull. U. S. G. S. no. 14, p. 195, 1885. 

 f B. and S., Amer. Journ. [3] xxxii. p. 185, 1886, 

 X Wied. Ann. xxviii. p. 369, 1886, 



