1902.] Experimental Researches on Drawn Steel. 



29 



point the magnetic intensity is the same hot or cold. The explanation 

 of the fact that a self-demagnetising force of given amount produces a 

 zero temperature coefficient in drawn steel now becomes clear. 



The influence of magnetic intensity on the magnitude of a is next 

 investigated in a long series of observations. The wire was mag- 

 netised step by step, and at each step the temperature coefficient 

 and the percentage permanent change of intensity were determined, 

 both for induced and residual intensity. The temperature coefficient 

 in general follows the susceptibility and becomes least at the highest 

 intensity but always incremental; on demagnetising it increases, 

 slightly and then falls towards zero, which occurs at some very low 

 residual intensity. 



The behaviour of the residual magnetism as demagnetisation pro- 

 ceeds is very interesting; when a small part of the magnetism is 

 removed, the effect of alternate heatings and coolings is to leave the 

 intensity higher than before the application of heat and cold, and the* 

 more the magnetism is removed the greater the recovery by heating 

 and cooling ; and, at last, on applying so large a demagnetising force- 

 as to remove all the magnetism and to leave a small residual intensity 

 inverse to the direction of the original intensity, then, heatings and 

 coolings clear this out and restore some of the original magnetisation. 

 The behaviour of the coefficient during these changes is also discussed.. 



For comparison, the same series of experiments was performed upon 

 an annealed iron wire about 400 diameters long. The hot curve of 

 magnetisation crosses the cold curve a little beyond maximum 

 susceptibility and the maximum residual intensity is less hot than 

 cold; it would therefore appear that at some small initial intensity 

 the coefficient should be positive, at a later stage zero, and finally 

 negative. The zero coefficient was not obtained in these experiments,, 

 but the subsequent negative coefficient was traced, and it was found 

 that when the susceptibility is large the negative coefficient is a 

 minimum; the coefficient is throughout negative during demagneti- 

 sation. 



The intensity under heating and cooling is similar in behaviour to 

 the intensity of drawn steel; on magnetising, heatings and coolings 

 always cause a loss of residual magnetism, but on demagnetising 

 heatings and coolings restore some of the magnetism removed. It 

 would no doubt be advantageous for the sake of producing a magnet 

 of constant intensity to apply a small reversed force after magnetising 

 to saturation, any fluctuations of temperature afterwards would then 

 tend to increase rather than diminish the intensity. 



It is pointed out that the gradual decay of magnetism in observa- 

 tory magnets will generally be accompanied by some change in the 

 coefficient, and that remagnetisation of magnets will likewise alter the 

 magnitude of the coefficient ; the general effect of this change will be 



