1892.] on Metals at High Temperatures. 615 



haps the most remarkable point about the purple alloy is its melting- 

 point, which I have shown to be some degrees higher than that of 

 gold itself.* See diagram, Fig. 7, in which curves of several con- 

 stants of these alloys are given. This fact affords strong evidence 

 that the alloy AuAl^ is a true compound, having analogies to the 

 sulphides, for in every other series of alloys the melting-points of all 

 the members of the series are lower than that of the least feasible 

 constituent. There is one other fact of much interest connected 

 with this alloy. When it is treated with dilute hydrochloric acid, 

 chloride of aluminium is formed, and gold is released in a singularly 

 voluminous form. The heat of formati(m of the gold-nlumijiiura 

 alloy lias not been determined, but hydrochloric acid, which will not 

 attack gold, will readily split up this compound, of which more than 

 three-fourths is gold : the compound, in fact, behaves like a distinct 

 metal, having special heats of oxidation and chlorination of its own. 



(3) Lastly, we come to the question of solutions of metals in each 

 other. One very remarkable instance of the behaviour of metals at 

 high temperatures reveals the fact that the presence of a small amount 

 of metal in a mass of another lowers the freezing-point of the mass. 

 In the industrial world this has long been known. Cellini tells us, 

 for instance, that when the bronze for his great figure of Perseus, at 

 Florence, was running out of the furnace, it suddenly showed signs of 

 setting, and he therefore threw pewter plates and dishes into the ducts 

 through which the metal had to pass — "a thing," he says, "never 

 before done." The fluidity of the metal was immediately increased, 

 and he found every part of the casting " to turn out to admiration." 



The excellent work of Heycock and Neville,f on the lowering of 

 the freezing-points of metals by the addition of other metals should, 

 I vi^ould suggest, form the subject of a lecture in this Institution at 

 an early day. 1 cannot attempt to deal with the matter here. In 

 leading up to these questions of solution, as applied to metals, I 

 would remind you that Lord Eayleigh told us a few evenings since 

 that it was by no means certain that a gas rushing into a vacuous 

 globe ever completely fills it, as there may still be tiny spaces into 

 which " odd molecules " fail to find room to vibrate in. If it is 

 difficult for a gas to entirely fill a vacuous space, you would think it 

 impossible for a small quantity of a metal to rapidly permeate a fluid 

 mass Qf another metal; nevertheless, so far as analysis can detect, 

 this does happen. 



It may be incidentally observed that the relations of the ordinary 

 gases to metals are far more intimate than they were formerly sup- 

 posed to be, and this was proved by Graham's work on the absorption 

 of gases by metals, which has often been dealt with in this Institu- 

 tion. To take only the case of iron, more than twenty years ago 



* Proc. Eoy. Soc. vol. i. 1891, p. 367. 



t Chein. Soc. Journ. vol. Iv. 1889, p. 666; vol. Ivii. 1890, pp. 376, 656; 

 vol. lix. 1891, p. 936 ; vol. Ixi. 1892, p. 88S. 



