192 DR. S. W. J. SMITH AND MR. J. GUILD: A THERMOMAGNETIC STUDY OF 



fig. 5, the contribution of the eutectoid to the total magnetisation, at say 700 C., 

 indicates the amount of iron which it contains, (2) that the eutectoid mixture has 

 the same composition in all the steels. 



It is instructive to consider in detail the forms of the thermal hysteresis loops and 

 connecting curves shown in the figures, and to examine their significance with respect 

 to the sequence of changes within the material ; but considerations of space do not 

 permit further reference to this here. 







11. Evidence of the Relative Unimportance of the Effects of Diffusion in 



Loiv-carbon Steel. 



We proceed to describe further experiments of which the results are shown in 

 fig. 8. These were made upon a steel containing about O'l per cent, of cai'bon. Their 

 object was to test the inference, from a comparison of fig. 5 with figs. 6 and 7, that 

 the effects of the slowness of diffusion are relatively unimportant when the layers of 

 solid solution from which the eutectoid is about to separate are relatively thin. In 

 the present case the eutectoid patches form about one-ninth part only of the total 

 mass of the steel. 



The thicker-lined curves represent, as before, the results obtained during continuous 

 slow heating and cooling. The heating was stopped at about 815 C. The 

 equilibrium diagram indicates that, at this temperature, the solution in contact with 

 the still undissolved iron contains roughly 0'3 per cent, of carbon. The patches of 

 transformed eutectoid are therefore bordered by layers in which the concentration of 

 carbon drops to about 0'3 per cent. 



During cooling the iron in these border layers separates out again from the surface 

 inwards, the surface concentration rising continuously from 0'3 per cent, towards the 

 eutectoid percentage as the temperature falls. The re-crystallising iron closes in 

 upon the layers of solid solution ; but the re-precipitation of the excess iron is not 

 (juite complete at the temperature at which it begins to dissolve during heating. 

 For it will be observed that the cooling curve US rises slightly, but distinctly, more 

 rapidly than the heating curve PQ falls, and therefore that, if we assume no solution 

 of iron along PQ, we must suppose some precipitation along RS. 



It is easy to see, as before, how this effect may arise ; but it is also easy to show 

 how much less pronounced it is in the present case. To do this, we repeated the 

 experiment, but interrupted the cooling at various temperatures between R and S, 

 as in the experiments of figs. G and 7. 



The results are shown in the centre fig. 8. For the sake of clearness, we have 

 displaced the observed ordinates downwards. Their true positions are to be found by 

 raising all of them through the distance between the lower AB and the upper. It 

 will be observed that the effects of alternation during cooling, although perceptible, 

 are now very slight, compared with those of figs. 6 and 7, until the temperature 

 at which the eutectoid appeared during continuous cooling is approached. 



