A NEW CURRENT WKIQHKK, KTC. 



491 



measuring machine bed was much greater than this, but an analysis of the readings 

 leaves little doubt that the prolwble error is not more than 1/x. On each jwir of coils 

 not less than 112 observations were made, and the curves show that the error of a 

 le olwervation must l>e small ; hence the error of the mean diameter deduced from 



PLAMf. OF MCASURMNT 



s 7 a s 10 ii 



CLLIPT/C/TY. 



Fig. 14. Suspended cylinder N<>. _'. 



the 112 observations is not appreciably greater than the error of the gauge. We 

 conclude, therefore, that the relative diametral dimensions of the coils are correct to 

 5 in 1,000,000. The probable error of the axial lengths given in Tables I. and II. is 

 of the order 1 5 in 1 ,000,000, and the calculated value of the mutual induction should 



3 R 2 



