[shaw] THE KELVIN-VARLEY SLIDE POTENTIOMETER 97 



, . , , . . . , ., c . , HAS . 100000 



multiply the right-hand side ol equation (14) by innnnn , — • Hence 



the true reading for the slider at n is given by . 



101 



There may also be small errors in the vernier coils, but these may 

 generally be neglected because they form such a small fraction of 

 the whole reading. Such errors only become important when the 

 slide reading is at or near zero in connection with the measurement 

 of the ratio of a very small resistance or electromotive force, to a large 

 one. Except for these special cases nothing can be gained by a calibra- 

 tion of the vernier coils. The correction to be applied to the vernier 

 if the reading on it is m, would be 



101 



m f_ , _ 2 ? ÛM , „ \ (20) 



+00 



(a H+ i + a n +i ~l 1" e ) 



where 2 r m is the sum of the errors of the vernier coils up to m expressed 



101 



2 2 a, 



in mean ohms of the vernier and 



{(a n+ i + a„+o YqI *~ e ) u 



the difference from 1000 ohms of the vernier coils shunted by R n+ i 

 and R n +2 of the main slide, as derived in equation (17). It will be seen 

 that if n is very small the second term in the above expression (20) 

 may become very important and the first term quite appreciable if m 

 is large. 



Practical Details of the Calibration. 



The actual procedure and the subsequent calculations may be 

 performed quite simply. In order to illustrate the method fully, 

 and to show the magnitude of the measurements and corrections 

 an account is given below of the calibration of the Kelvin-Varley 

 Slide at McGill University. 



The best kind of galvanometer at our disposal for this work was 

 found to be a four coil Thomson reflecting galvanometer of the usual 

 type with the coils in parallel giving a resistance of 1500 ohms. The 



value of — ^ — was determined first and repeated subsequently 



during and after the main observations. The circuit is connected as 

 described above in preparation for obtaining the equations (8). By 

 altering one of the two thousand ohm coils by one ohm and also by 



Sec. Ill, 1914—7. 



