RECENT PROGRESS IN THE STUDY OF ALLOYS. 93 



of the binary alloys that have been examined, the tem- 

 perature from N onwards does depend somewhat on the 

 original proportions of the two metals, and that this stage 

 does not appear to be altogether isothermal. An impurity 

 would produce this effect, but it may be that some other 

 cause is at work. 



It would probably require a very delicate thermoscope 

 to detect all these details during the cooling of one alloy, 

 but the whole series is well shown in the cooling - curves 

 of iron-aluminium alloys (fig. 5). This figure is taken from 



,.& 





Cooling Curve, for Iron- AUimmium Alloy* 

 ■»» ^O - 31 ■ .4o 4f SO J SS 



Professor Roberts Austen's "Third Report on Alloys". 1 The 

 curves were obtained by an automatic method of observa- 

 tion. A thermo-junction was placed in the centre of the 

 mass of molten alloy, and a galvanometer in the circuit 

 projected a spot of light on to a sensitive plate. The plate 

 possessed a uniform motion of translation at right angles to 

 that of the spot of light from the mirror of the galvanometer. 

 In this way, as the cooling progressed and the electro- 

 motive force of the thermo-couple, and consequently the 

 current passing through the galvanometer, decreased, curves 

 were traced out in which the rate of cooling is indicated by 

 the slope of the curve. Each curve in the figure belongs 

 to an alloy of different composition, the point where each 

 curve cuts the line of 1000 giving the percentage composi- 

 tion, which can be read off from the scale at the base of the 

 figure. 



It will be seen that all the curves approach the 

 horizontal at low temperatures. This is due to the fact that 



1 Engineering, 1895. 



