338 Mr. W. Grant on the Curves of 



same as that in the diagram, omitting the coil B. The 

 coil was brought as near as possible to the point P, 

 and clamped in that position, sufficient room being left 

 for the coil D to pass without touching it. D was then 

 moved to a balancing position on one side of P, and then to 

 the corresponding position on the other side of it, and the 

 half-difference of the readings on the scale M N was taken as 

 the value of the ordinate, the reading on the scale P for 

 the position of the coil C during the experiment being taken 

 as the value of the corresponding abscissa. 



The least distance between the centres of the coils C and 

 D at which readings could be obtained when their axes were 

 coincident was 1*6 centimetre, as their thickness prevented a 

 nearer approach; the least distance between their axes at which 

 readings could be obtained, when the mean planes of the coils 

 were coincident or at a less distance than 1*6 centim. from 

 each other, was about 9*2 centim. There was, therefore, in 

 each quadrant a rectangular area of which these dimensions 

 are the sides within which no results could be obtained. 



In order to extend the experimental curves within this area, 

 two coils, whose centres could be brought within 2 millim. of 

 one another, were constructed from the remainder of the piece 

 of wire from which the other four were made. Their diameters 

 were approximately the same as those of the other coils ; but, 

 owing to their thinness and the consequent small number of 

 convolutions of wire, their coefficient of mutual induction was 

 much less ; but this defect was compensated to some extent by 

 their proximity. The lines of force due to them were slightly 

 different in form from those due to the thick coils ; but as in 

 most positions of the thin secondary coil the lines of force 

 passed through it nearly at right angles, there was less objection 

 to the use of the thin coils than if they had been at a greater 

 distance apart. They were fitted one to each of the coils 

 and D, so that the same mountings served for both, and the 

 readings were obtained from the same scales. That which was 

 fitted to the former we shall call C , that which was fitted to 

 the latter we shall call D . 



In continuing the measurements within the area already 

 mentioned, the coils C and D were placed in the same 

 balancing pair of positions which C and D had occupied when 

 the last point of a given positive curve was determined. The 

 coil B was then moved towards F along the guide E F to a 

 position in which the mutual induction between A and B 

 balanced that between C and D i this gave the desired value 

 to the induction-coefficient of the curve ; and the remaining 



