400 



Mr. C. Barus on the Secular 



With the exception of No. 9, the variations are, as a rule, 

 well within one per cent, of the electrical value of the interval 

 hard-soft. Where the resistance is larger or smaller than the 

 mean value, it is possible that the parts were accidentally 

 splashed by the water immediately before quenching, or 

 quenched at a different degree of red heat. But inasmuch as 

 the series of measurements for each rod takes some time, and 

 as the effective resistances are necessarily small, I believe that 

 the rods are even more homogeneous than Table I. indicates ; 

 for the errors introduced by variations of temperature, by the 

 difficulty of clamping the thin, very brittle rods (soldering 

 would change the temper enormously), and by other incon- 

 veniences of manipulation, are by no means negligible. 

 Hence the degree of homogeneity found may be considered 

 quite satisfactory. 



In Table II. I give the mass m, length I, and density at 

 0° C, A , of some of the rods. The latter datum is essential 

 for the determination of the sections. I shall also use it in 

 the future in determining the secular volume-changes of these 

 wires. Rods Nos. 11 to 15 are too thin for direct measure- 

 ment. Hence a mean value of A was here assumed. I may 

 add that in the soft state the density of these rods is about 

 A =7-80. 



Table II. — Mass Constants of the Glass-hard Steel Rods, 

 June, 1885. 



No. 



m. 



/. 



Ao- 



No. 



m. 



I 



V 



9- 



centira. 



<7/cm 3 . 





9- 



centim. 



g/cm. 3 



1 



•9910 



24-95 



7-710 



11 



•3257 



27-94 





2 



•9544 



24-02 



7-689 



12 



•3322 



2830 





3 



•9970 



25-25 



7-703 



13 



•3170 



2720 





4 



1-0483 



26-30 



7705 



14 



•2888 



24-35 





5 



10079 



25-46 



7-703 



15 



•3004 



2607 





6 



1-0243 



25-80 



7-687 



16 



2-4713 



25-63 



7-676 



7 



1-0089 



25-47 



7-670 



17 



2-4857 



25-70 



7671 



8 



1-0074 



25-50 



7-682 



18 



2-1043 



22-50 



7-544 



9 



1-0666 



26-68 



7-726 



19 



2-5327 



26 06 



7-685 



10 



1-0293 



25-58 



7-719 



20 



2-3557 



24-40 



7-653 



Table III. contains the electrical constants of these rods ; 

 r t being the observed resistance per centimetre, at £°, a the 

 temperature coefficient, and s the corresponding specific re- 

 sistance at 0° C. The wires were quenched on June 1st and 

 2nd, 1885, and the measurements made at the time given, 

 only a few days after hardening. Sections and radius are 



