208 Mr. C. Barus on Maxwell's Theory of the 



The phenomenon considered as a whole must be continuous, 

 both as regards temperature and time. In proportion as 

 temperature is higher, however, Maxwell's theory predicts 

 that the effect of the same increment of the temperature 

 of annealing will produce increments of viscosity successively 

 diminishing at a very rapid rate. Supposing molecular con- 

 figurations originally present in all states of instability, it 

 follows at once that the groups which retain this quality after 

 annealing must very soon vanish when the temperature of 

 annealing is increased. The data prove this in a convincing 

 way. Eods annealed at 300°, 400°, 500°, . . . 1000°, show 

 about the same viscous behaviour (relatively speaking), not- 

 withstanding the fact that chemical analysis proves that the 

 decomposition incident to the successive application of these 

 temperatures on glass-hard steel continues steadily to in- 

 crease*. Indeed the chemical decomposition above 300° is 

 more marked than below 300°; yet its bearing on Maxwell's 

 theory is now without interest because in none of the high 

 annealed rods do configurations unstable at mean atmospheric 

 temperatures survive after annealing. 



11. Having analyzed the phenomena at mean atmospheric 

 temperature, I come next to consider the conditions of mean 

 relative viscosity of steel at 100°. The glass-hard state must 

 here be withdrawn, for consideration in § 13 ; because such a 

 rod would undergo annealing during the viscous measurements 

 at 100°. 



Hard steel annealed at 100° bears the same relation to 100° 

 that glass-hard steel does to mean atmospheric temperature. 

 Hence the reasoning of the preceding paragraph, mut. mut., 

 applies at once. It is merely necessary to bear in mind that 

 100° is now the temperature of incipient annealing, and that 

 therefore the temperatures which produce corresponding 

 viscous effects are proportionately higher. Rods An. 200° 

 now occupy about the same relative position that rods An. 

 100° did in § 9 ; An. 300° the same relative position as An. 

 200° in § 9 ; &c. Moreover, for equal increments of the 

 temperature of annealing, the increment of viscosity shown 

 by the rod at 100° diminishes rapidly as temperature increases; 

 &c. 



In one respect the present results differ from the above. The 

 phenomena are here spread out over a scale (roughly estimated) 

 about ten times larger. This means, following Maxwell's 

 theory, that at 100° the number of unstable molecular confi- 

 gurations is relatively much larger than at mean atmospheric 



* Am. Journal, xxxii. pp. 277, 282, 1886. 



