316 Mr. A. K. Chapman on the 



this same disk last year, nothing of unusual interest was 

 found. Upon carrying the fields below 1390, which was 

 the lowest used in the previous experiments, a remarkable 

 effect came to light. The value of m/R is zero up to a 

 field of about 760, rises to a sharp maximum at 950, falls to 

 half the maximum value at 1500, and then decreases slowly 

 as the field increases up to 10,000, which was the largest 

 field used. 



In order to investigate this region accurately, it was 

 necessary to read small differences in field with a certainty 

 greater than that afforded by the fluxmeter. Accordingly, 

 a standard ohm, immersed in an oil-bath, was put into the 

 magnet circuit and the current flowing determined by taking 

 the potential drop across it ; a Wolff potentiometer was 

 employed, together with a standard cell prepared by Professor 

 Hulett of the Department of Chemistry. By taking corre- 

 sponding current and fluxmeter readings for several fields of 

 very different magnitudes, it was found that, for any given 

 region, the field is proportional to the current. Having 

 determined the factor of proportionality, it was then possible 

 to fix accurately small differences in the field by the 

 corresponding current readings. 



During the course of preliminary experiments on this 

 metal, some difficulty was experienced in duplicating the 

 numbers obtained at various times. In view of the fact 

 that the magnetic properties of iron vary rapidly with the 

 temperature, as does the Hall effect, the disk together with 

 the inductance-coils was put into a brass box, so arranged 

 between the poles of the magnet that a constant stream of 

 water could be kept flowing through it. In this way the 

 temperature was, at all times, kept stationary to within one 

 degree, which proved to be sufficiently constant for the work 

 on hand. 



Residual magnetism also interfered seriously with the 

 measurements at low fields. This was eliminated by applying 

 an alternating current to the magnet before the desired 

 direct current was passed through it. In every case the 

 initial alternating current employed was much larger than 

 the direct current to be used and was gradually reduced to 

 zero, thus effectively demagnetizing the magnet and disk. 



Having taken these two precautions, it was possible to 

 obtain fairly consistent results, as shown by the table. 



