1891.] Properties of Metals in relation to the Periodic Law. 355 



portant part, and the introduction of a small quantity of it renders it 

 possible to cast very mild steel or even " wrought " iron into forms 

 which are remarkable for delicacy and soundness. The mode of action 

 of the aluminium on iron in the " mitis castings " has given rise to 

 wide divergence of opinion, but the view that it acts by the removal 

 of oxide, or of occluded oxygen, has gained much favour. In the case 

 of gold, which has neither occluded oxygen nor oxide to lose, the 

 castings of the metal with 0'2 per cent, of aluminium are also remark- 

 ably sound, and, as experiments prove, very tenacious ; the action cf 

 the aluminium is, therefore, probably a molecular one of much com- 

 plexity. 



It may be pointed out that the presence in gold of quantities of 

 silver which vary from O'l to 4'0 per cent, does not lower the freezing 

 point of the mass. Messrs. Heycock and Neville, who witnessed 

 certain of the experiments above described, inform me of the hitherto 

 unpublished fact, observed by them, that the presence of thallium 

 does not lower the freezing point of lead. 



The close concordance in both these cases between the atomic 

 volumes of the mass of metal and the added impurity is of special 

 interest in connexion with the generalisation given in my earlier 

 paper and re-stated on the first page of this. Silver has the same 

 atomic volume as gold, and if present in small quantity, produces no 



I effect on either its tenacity or its freezing point. 

 Throughout these experiments gold has simply been employed for 

 the sake of its freedom from liability to oxidation, but other metals 

 must be studied, and it is worthy of record that Hadfield has recently 

 shown that the parts played by aluminium and by silicon in steel are 

 almost identical. Most of the physical properties of aluminium and 

 silicon, in a free state, are totally different, but they possess the same 

 atomic volume, and when they are alloyed with iron they affect it in 

 precisely the same way. 

 I have to express my thanks to my assistant, Mr. H. C. Jenkins, 



his aid in conducting these experiments. 



[April 20, 1891. In the course of the investigation, it became 

 ident that, as is the case when aluminium is alloyed with copper or 

 >n, the addition of aluminium to gold is attended with evolution of 

 it. The following experiment was therefore arranged, with a 

 lew to obtain evidence on this point : 



A mass of 30 grams of gold, contained in an unglazed porcelain 

 icible, was placed in the centre of a block of firebrick and strongly 

 teated up to well above the melting point of the metal. The thermo- 

 junction was inserted directly in the gold, and the spot of light from 

 te galvanometer allowed to fall in the usual way on to the sensitised 

 late (fig. 1). A piece of cold aluminium; equal in weight to 1 per 

 mt. of the mass of gold, was then added and rapidly stirred. The 



