Hf, Mil. F. F, SMITH OX T1IK mNSTIMVTloN OF 



In addition, a potential lead P r (fig. 9) is thrown into series with 11. [The value 



w 



Fig. y. 



of P, is of the order O'OIG ohm, and is conveniently measured to 1 per cent, of its 

 value.] The complete expression for the value of P is therefore 



(T 



_Q 



+ Q/W 



R(l 



" 



_ _ 



a + ft + d +~Pa/ I " S/(l + S/W'j 



</(<* + P tt ) 



a + /8 + d ~+ P 



1 



' 



Owing to the magnitude of P r , P a , and d, and in order to reduce the corrections to 

 small quantities, 1000-ohm coils were employed for R and S, while for a and ft 

 100-ohm coils were used. Q was a standard-unit coil. All the coils were of manganin 

 and had small temperature coefficients. 



The ratio R/S was equal to 1 '00002 ; that of a to ft was 0'99998. Also, in no 

 observation did (l + S/W')/(l + Q/W) differ by as much as 0'2 per cent, from 

 unity. 



With this information, the above equation may be simplified to the form 



P, , j/lOP. P, J04 



1000 " * c V 1000 



p _ QR(l + S/W') 



-S(1+Q/W)(1+R/W") 



1000 . 1000\ 



w w*/' 



Such simplification depends, of course, on the values of the coils employed. It will 

 presently be seen that P r had a resistance never so great as 0'03 ohm; consequently, 

 a value accurate to 1 per cent, was ample. The same remark respecting knowledge of 

 value applies to P a . The value of (P./100 P r /1000 0*00004) was of the order 

 0'00015 ohm ; that of d averaged 0'15 ohm, so that a value of d accurate to 1 per- 

 cent, was also sufficient to secure a final accuracy in the measurement of P of O'OOOl 

 per cent. The method adopted for determining P r , P , and d was, however, sufficiently 

 L;'" >(1 to ensure tlirir estimated values being correct to O'l per cent. 



VYitli respect to the arrangements. The coils (,), 11, and S were manganin coils of 

 the British Association anil made by WOLFF. They were immersed in oil contained 



