432 



C. Barm — Certain generic Electrical 



similarly favorable to the reduction to linear forms is obtained 

 by adding cast-iron to the steel series, I have assumed the 

 equation 



x(y-\-m)=n. 



Using this equation as a basis for the application of the method 

 of least squares, the observations of Dr. Strouhal and myself 

 lead to the results contained in the following table. These 

 results are easily understood, and I need hardly add that 

 according to § 1, the conductivity and temperature-coefficient of 

 steel are respectively 



m and 





Table II. — Showing Strouhal and Barns' results for the electrics of iron-carburets, 

 /(<>)(/' (0)//(0) + »)=*. 



Metal. 



Temper. 



/(0). 



/W/(o). 



/W/(0). 



Diff. • 









Observed. 



Computed. 





Steel 



Soft 



15-9 



00423 



0-00417 



+ 6 





Annealed light blue 



18-4 



360 



366 



- 6 





full blue 



20-5 



330 



333 



- 3 





'• yellow 



26-3 



280 



269 



+ 11 





" light yellow 



28-9 



244 



249 



— 5 





G-lasshard 



45 •'7 



161 



173 



-12 



Cast iron 







78-5 



129 



119 



+ 10 



m= — 0-000438^0-000097 

 n = + 0-05930 ±0-00151 



Applied to steel alone, the constants computed by the method 

 of least squares show even better agreement, viz : 



m=— 0-000303rfc0*000079 



^ = + 0-0620 ±0-0017 



I omit the details, with the mere remark that the difficulties 

 of measurement with a molecularly unstable brittle body like 

 hard steel are great. * 



5. I desire now to add to these remarkable results the new 

 data which I found for platinum alloys. I shall endeavor to 

 make my series more nearly complete ; to investigate points 

 more nearly contiguous and nearer in position to the pure 

 metal than was the case in the foregoing series, as well as to 

 introduce a great variety of platinum alloys. For both in 

 Matthiessen and Yogt's results, and in Dr. Strouhal's and my re- 

 sults, the position of the individual points is not always near 



* Prof. J. H. Langley kindly called my attention to recent data for manganese 

 steel. Looking up Fleming's work (Lum. elect r., xxvii, p. 589, 1888), I found his 

 results to be 5 = 68 and a = 0-0012. In the above diagram s = 68 corresponds 

 to a=0'0013, showing satisfactory accordance with experiment even in this region 

 of abnormally high specific resistance. 



