1882.] Unit of Resistance in Absolute Measure. 399 



could not be perfectly liarmonized on the basis of an d priori calcula- 

 tion of the self-induction. 



Table VII.— Third Series. 





60. 



45. 



35. 



30. 



Mean. 



Resistance of standard at \ 



23 '619 



23 621 



23 -630 



23-638 



23 -627 



Correction proportional to T 



0-006 



Oil 



0-018 



0-025 





Resistance of standard at"! 



23 -613 



23 -610 



23 612 



23 613 



23 612 



Table VII gives the results of this series. The " number of teeth ' r 

 in the first row is inversely as the speed of rotation. The second 

 row gives the resistance of a certain platinum-silver standard at 13° 

 in absolute measure, as calculated with a value of the self-induction 

 derived from evidence independent of the spinnings. The simple 

 mean of these numbers is 23"627 ( X 10 9 C.Gr.S.), but they exhibit a 

 well-marked tendency to rise with the speed. In the third row are 

 numbers proportional to the squares of the speeds, by subtraction of 

 which a practically perfect agreement is arrived at. The mean 

 23'612 thus represents the result of the investigation, if the effect of 

 self-induction be determined from the spinnings themselves, and is to 

 be preferred for reasons explained in the paper to the simple mean 

 23*627. The difference is, however, less than one part in a thousand. 



The resistance at 13° of the same coil in terms of B.A. units is 

 23*935, from which we find 



1 B.A. unit = -98651 eartIx q na[W , 



second 



This number is somewhat lower than that which we obtained 

 (*9893) with the original apparatus,* but it agrees with that required 

 to reconcile Dr. Joule's thermal determinations. Rowland's value is 

 distinctly higher (*9911), while Kohlrausch obtained 1*02. No satis- 

 factory reconciliation of these results is arrived at, but some remarks 

 are made upon the relative merits of the various methods. 



* " Proc. Roy. Soc," vol. 32, p. 141. 



