of Coefficients of Mutual Induction. 415 



error was due to the dimensions of the section of the coils, 

 the coefficient was then calculated by the more accurate for- 

 mula referred to below ; but the improvement is very slight. 

 It appears that this formula does not in any way cure that 

 defect, which arises from the distance of the central planes. 

 So that in coils similar to mine we cannot expect to get nearer 

 by calculation than the number last referred to, which is to 

 the observed number nearly as 6 : 7. 



On February 23 experiments were arranged with a view 

 to test the consistency of the method, the coils examined 

 being made primary and secondary alternately. Though the 

 mode of observation does not admit of great accuracy, it 

 appears that there is a systematic difference between the 

 results of the two arrangements, amounting to about 1^ per 

 cent. This I have not been able to explain. The coefficients 

 of self-induction of the coils should have no influence ; but 

 that of the 250 will be very much greater than that of the 42, 

 and this is the only source of error that I can suggest. The 

 number calculated by Maxwell's original formula is to the 

 mean of observation nearly as 6 : 7. 



In all the experiments difficulty was experienced in con- 

 sequence of the continual fall of the battery-current. With- 

 out care in charging, and freshly charged cells, the experi- 

 ments could hardly be made, as the fall of the current affects 

 the galvanometer in the intervals. 



Rowland's method of control, with a magnet and coil, was 

 employed. 



The determinations of the coefficient of the Gramme machine 

 were very rough. In fact the fundamental equation (2) is 

 not really applicable to the dynamo at all, except perhaps in 

 the upper part of its range, where the current is, say, 10 

 amperes or more. For small currents do not do more than 

 shake the subpermanent magnetism, which is considerable 

 compared with the magnetism due to small currents. And 

 in motion the machine makes use of this subpermanent mag- 

 netism, although it does not enter into the electrical coefficient 

 of mutual induction. 



The observations of Feb. 24 were very irregular ; it ap- 

 peared as if the current sometimes shook the subpermanent 

 magnetism and sometimes not. Still a mean was fairly 

 deducible. 



On Feb. 25 a stronger current was employed, and the ob- 

 servations were fairly regular ; but this current, though it 

 moves the subpermanent magnetism, is quite insufficient to 

 reverse it as the reversal of a large current does ; so that 

 even here we do not get a representation of the whole effect. 



