414 Mr. E. H. M. Bosanquet on the Determination 



into the circuit S, after calibrating the galvanometer without 

 it. The value of the deflection is then altered in the ratio of 

 the change of resistance. 



Let E be the total increased resistance, 

 E the resistance without the addition. 



Then equations (2), (3), (4) become 



Q = -jp ........ (5) 



Qo= -g^ ? * * (' 



1 /3E o Gtan0 ^ 



The 2 disappears, as before, when reversal is employed. 



An example of case (a) is given later ; where, in the second 

 determination of my Gramme machine, a resistance of 1000 

 B.A. units was introduced into the secondary circuit, to mo- 

 derate the deflection. The determination of the resistances, 

 however, can hardly ever be accurate, as the copper earth 

 induction-coil is liable to great alterations through change of 

 temperature. 



An example of case (b) would be afforded by the determi- 

 nation of such an induction-coil as the one by Apps, with high 

 resistance S, mentioned in Prof. Foster's paper. In this case 

 the galvanometer would be calibrated by the earth induction- 

 coil, the high resistance-coil then introduced, and the known 

 values of E, E employed in formula (7). 



The determinations I have made are mostly of the mutual 

 coefficients of the coils of my earth-inductors, placed close 

 together, I have also made a couple of determinations of the 

 coefficient of field-magnets and armature in my A Gramme 

 machine. 



The determination of February 21 was rather rough; the 

 coils were placed together in their frames, but they could not 

 be brought very near. The numbers are useful, as illustrating 

 the increase of the discrepancy of Maxwell's formula, with 

 the increase of b. the distance of the central planes. In fact, 

 unless b is small compared with the diameters, the formula is 

 not properly applicable. Here it is about half either of the 

 radii, and calculated and observed values are nearly as 4 : 5. 



On February 22 the coils were dismounted and brought as 

 close together as possible, b being now between a quarter and 

 a third of either of the radii. The calculated and observed 

 values are nearly as 5 : 6. In order to see what part of the 



