226 MESSES. C. T. HEYCOCK AND F. H. NEVILLE 



mixture, not an almost pure substance like the alloys at B and D, somewhat differ- 

 entiate the transition at G from those at B and D. The most probable explanation 

 of what happens at G is that the curve FG, along which the X body is in equilibrium 

 with the liquid, would, but for the formation of the purple alloy, be continued to a 

 summit, X, on the right of G, and then no doubt downwards towards the aluminium 

 end of the system. In the same way we must suppose that the equilibrium curve 

 HG between the purple crystals and the liquid would, but for the existence of the 

 X body, be continued below GF. Our experiments give, as is usually the case, only 

 those parts of the two equilibrium curves which have no other branch above them. 

 Our method of stirring the liquid vigorously until solid begins to form prevents very 

 great surfusion and destroys any chance of finding freezing points on the hypothetical 

 branch GX. But a cooling curve, conducted without stirring, might show such 

 points. We have no conclusive evidence as to the formula of the X body, but as at 

 G it forms considerably more than half of the alloy, the summit cannot be very much 

 to the right of G ; the formula AuAl seems to us an extremely probable one. The 

 microscopical study of alloys near the point G shows how rash it would be to assume 

 that an angle in a freezing point-curve, together with a horizontal line of second 

 freezing points, necessarily gives the formula of a compound. 



Alloys with more than 44 atoms of aluminium contain increasing quantities of 

 purple the nearer we get to the summit H, at which point the whole area of the 

 section is full of purple separated into large patches by slender lines of white that do 

 not form a complete network. The alloy with 6 5 '3 atoms of aluminium (fig. 28) 

 shows the structure well. It is necessary to examine these alloys without etching 

 them, as all etching reagents destroy the purple and cause an electrolytic deposit of 

 gold on the other bodies present. Most of the alloys between G and H had been 

 polished by the dry method, and hence the polished surface shows three things : the 

 beautifully uniform crystals of purple AuAl 2 surrounded by the white X body, and, 

 generally in the middle of the patches of white, some golden spots of eutectic. This 

 eutectic is the matter solidifying under the conditions of the point F. As we have 

 explained, the gold can be removed by wet polishing, in which case the eutectic is 

 white, but distinguishable from the pure X. It appears to be the harder Au 2 Al 

 that retains the gold. Even at H the thin bands of X contain thinner lines and 

 spots of eutectic. The purple is so generally surrounded by a border of pure white 

 X, that the purple and the F eutectic are rarely in contact (fig. 27). 



The alloys at 66'6 and 67"4 atoms have about the same amount of thread-like 

 mother-substance, not more than 1 per cent, of the area, and at the latter percentage 

 one does not see any marked change in the material of the threads. At 69 '2 atoms 

 the area of the bands of mother-substance has trebled at least, and it is greyer than at 

 the same distance from the summit on the other side. At 72 '6 atoms the mother- 

 substance has of course increased, and it is now a soft greyish body full of scratches and 

 quite unlike the substance X. Moreover these bands of grey have no threads of gold 



