﻿Temperature Coefficients of Ferro-Mag netic Elements, 37 



magnetic intensity of the pure annealed ferromagnetic elements 

 in corresponding states are inversely proportional to the absolute 

 critical temperatures. 



2. The experimental work leading to this result was made 

 in two series. In the first wires of nickel and iron of 

 large dimensions were employed. They were annealed, 

 magnetized to saturation, heated to a little below the critical 

 temperature, and then inserted in a glass tube containing 

 olive oil, and in some of the later experiments in a brass 

 trough containing sand. The heating was effected by a 

 double row of many small Bunsen flames, and the magnetism 

 of the wire was tested by placing it " end on v to a mirror 

 magnetometer, the reflected beam of light being received on 

 a scale two metres away. An electrical resistance thermo- 

 meter which was laid side by side with the magnetized wire 

 was used to measure the temperature. With this apparatus 

 experiments were made up to 435° 0. on a wire of nickel, 

 two of soft iron, and a bundle of seven fine iron wires. 



Difficulties, however, were experienced in procuring cobalt 

 of the requisite dimensions, and the samples obtained con- 

 tained too much impurity to yield trustworthy results. It 

 was decided, therefore, to prepare cobalt electrolytically, 

 and this led to the second series of experiments in which 

 much smaller magnets were employed, These had the 

 advantage that they could be heated uniformly with less 

 probable error, but on account of their small mass, their 

 magnetic moment was not large enough to give a deflexion 

 of the magnetometer needle which could be read with 

 accuracy. 



This consideration necessitated the construction of a special 

 magnetometer in which the single needle w T as replaced by a 

 vertical astatic pair, 0*7 mm. apart and 13 cm. long, sur- 

 rounded by a stout copper tube which served to damp the 

 oscillations of the system. The orientation of the needles 

 was read as before by a beam of reflected light on a scale 

 two metres from the magnetometer. 



This instrument was highly sensitive to any force acting 

 on one pair of poles in the same horizontal plane, and was 

 but little affected by extraneous disturbances so that it 

 was very suitable for the purpose in hand, but it was 

 not so convenient for the determination of absolute values. 



Cobalt was now deposited electrolytically from a solution of 

 the double sulphate of cobalt and ammonia on wires of copper 

 and platinum 0*55 and 0*35 mm. in diameter respectively, 

 and from 10 to 15 cm. in length. Each wire thus coated was 



