202 



Prof. J. A. Ewing and Mr. W. Low. [Mar. 24, 



throughout their whole length. Very few experiments were made 

 with test-pieces of this form, for it was found that they gave by no 

 means an exceptionally high value for the magnetic induction. This 

 is to be ascribed to the fact that the ends of the cylinder, which were 

 in contact with the pole-pieces, necessarily shared that value of the 

 induction which existed in the part of the pole-piece faces which they 

 touched, and this comparatively low induction in and near the ends of 

 the cylinder neutralised the much higher value in the middle portion. 

 The induction coil, being wound from end to end of the bar, gave a 

 mean value for the whole length. To obtain higher values, it was 

 obviously necessary to restrict the measurement of the induction to 

 the middle portion, where the induction was greatest ; and, further, 

 it was desirable to furnish the bar with conical or some form of 

 spreading ends, which would present an easy path for the lines 

 of induction to converge towards the central neck. Accordingly, 

 test-pieces were turned of the form and dimensions of Sample A, 

 shown in Plate 2, fig. 1, where the bobbin is sketched in place between 

 the pole-pieces. These were wound along the whole length of the 

 narrow central neck with an induction coil consisting of a single 

 layer of No. 36 S.W.G. silk-covered wire. In Sample A the 

 diameter of the iron neck was 0'923 mm., and the diameter measured 

 to the middle of the thickness of the wire forming the induction coil 

 was - 9495. Hence there was but little space, outside the section of 

 the iron, enclosed by the coil ; and the small amount of magnetic 

 induction in this non-ferrous space was allowed for by a method to be 

 explained below. 



In test-pieces of the form of Sample A the loss of magnetism 

 observed on suddenly withdrawing the piece from its place between 

 the pole-pieces of the field magnet, is less than the whole magnetism 

 by the small but somewhat uncertain quantity of residual magnetism 

 which the piece retains. To avoid this source of uncertainty another 

 form of teskpiece was used, which is shown in fig, 2, Sample B. 

 Here the bobbin has its conical ends rounded at the base to form 

 portions of a circular cylinder, and the pole-pieces are hollowed to 

 correspond. The bobbin can now be turned completely round about 

 a central axis at right angles to the paper, so that the direction of its 

 magnetism is reversed, and half the ballistic effect of the reversal 

 measures the magnetic induction. This method was used in the 

 greater number of the observations. Again, by merely withdrawing 

 the bobbin from the field, and comparing the effect of this withdrawal 

 with half the effect of reversal, an estimate was arrived at of the 

 amount of error to which the former experiments were subject on 

 account of residual magnetism. 



To determine the intensity of the magnetic field in the space 

 immediately surrounding the narrow neck in which the greatest 



