of Cold Hard Steel. 491 



same principles that a method for constructing invariable 

 electric standards of resistance will have to be based, though 

 the desideratum of hardness and resilience (whatever metal 

 may be chosen) is not of special moment here. 



5. Returning from this brief digression to the special 

 results of the present paper, I will refer to Table II., in which 

 the softening effect of atmospheric temperature acting through 

 prolonged time is exhibited. Twenty rods (No. 18 broken) 

 of different diameters ('0806 centim., '044 centim., "126 centim. 

 in the different batches) were selected. Columns 3 to 8 show 

 the resistances, r t , in microhms per linear centimetre, at the 

 temperature t° and on the dates stated at the heads of the 

 columns. Columns 9-13 give the corresponding specific 

 resistances, % in microhms between opposed faces of a 

 centim 3 ., at zero centigrade, for the same dates. The 

 reduction to zero was made by aid of the relation investigated 

 for steel * 



(s + -78) (a - -0001435) = '0682, 



wherein an essential dependence of the temperature-coefficient 

 a and corresponding specific resistance s for any state of 

 temper is evidenced. Finally, the columns Ss denote the 

 resistance decrements since the first measurement (1885). 

 This was made only a few days after hardening in the electrical 

 apparatus f. It will be seen that there is a fair order of 

 agreement between the corresponding values of Ss 0j quite as 

 much as one may expect from the small resistances and the 

 long interval between measurements. 



The table also contains mean values for each group, showing 

 that the thickest rods seem at first to have softened more 

 rapidly than the thinner rods. The mean values of s are 

 inscribed in the chart forming the curve marked " secular,'"' 

 the ordinates being specific resistances as above, and the 

 abscissas the time elapsed since hardening, in months instead 

 of minutes. It will be seen that the present rods (from a 

 perfected mode of tempering) are initially much harder 

 (greater s ) than the rods used in studying temper. I do not 

 as yet know what effect the initial hardness J has upon the 

 limiting softness of a rod tempered at any temperature. I 

 can merely note therefore that atmospheric temperature acting 



* Bulletin U.S.G.S. No. 14, p. 15, et seq. ; Wied. Ann. xx. p. 525 

 (1883); furthermore, Bull. U.S.G.S. No. 54, p. 1G0 (1889); American 

 Journal, xxxvi. p. 427 (1888). 



t Bulletin U.S.G.S. No. 14, p. 29 ; Bull. U.S.G.S. No. 73, p. 7 

 (1891). 



% For the same steel one would expect this effect to he nil ; but the 

 experiments have not been conclusive. 



