﻿Cooling exhibited by Glass and by Steel. 185 



Finally, after annealing in boiling sulphur (450°), the polari- 

 zation figure has wholly vanished. We have in hand a hollow 

 globule free from aeolotropic strain. We show below that 

 the substance of the quenched P. E. drop is under a strain of 

 dilatation. It is necessary to bear in mind that the isotropic 

 part of this strain is not demonstrable by optic means. The 

 drop annealed at 450° has, however, wholly lost its explosive 

 character. 



Here therefore we encounter the first important analogy be- 

 tween the optical behavior of quenched glass and the electrical 

 behavior of quenched steel. Change of strain in glass is in- 

 cipient at 200° ; in steel at 50°. Change of strain in glass is 

 complete at a temperature certainly greater than 350° and less 

 than 450°. The corresponding limit for steel cannot be so 

 well defined.* In both cases the essential dependence of the 

 result to be reached upon the temperature and the time of an- 

 nealing (asymptotic relations) is the marked feature of the 

 phenomenon. 



Solution. — In Bulletin U. S. G. S., No. 35, we communicate 

 a series of results on the resistance of consecutive conaxial 

 layers of steel rods. From these we select the following digest, 

 Table II. Having secured the rods (fine grained steel, 

 quenched hard, length 6 cm , diam. 0'6 cm ) in an especial clamp, 

 the resistances between points about 5 cm apart were measured 

 by a modification of Matthiessen and Hockin's method. A 

 process of galvanic solution enabled us to remove given thick- 

 nesses of shell consecutively. After each of these removals 

 resistance measurements were made. In the table W t denotes 

 the effective resistance in microhms, S the corresponding spe- 

 cific resistance referred to the cubic centimeter; i?, # the mean 

 radius and thickness, respectively, of the consecutive shells, s 

 their (calculated) specific resistance. 



It is difficult to arrive at accurate values for specific resist- 

 ance in these measurements ; for we encounter very small 

 values for total resistance (W t ) at the inception of the experi- 

 ments, and small and irregular values of sectional area at 

 the close. Inasmuch as sections are always measured too large, 

 the values for specific resistance (S ) are too large, and the error 

 increases rapidly as we pass from greater to smaller diameters. 

 The mean, error of S is certainly several per cent and hence 

 s is only an approximation. Nevertheless these results are 

 important; they show that in case of hard rods less than 0*6 cm 

 thick, the values of resistance taken from circumference to axis 



* • It is remarkable and significant that the electrical effects due to the anneal- 

 ing of a glass-hard steel rod vanish almost entirely at a temperature at 



which the density of a homogeneous hot rod is the same as the density of an 

 (otherwise) identical cold glass-hard rod. — Bull. 14, 97. 



