132 REPORT— 1900. 



AuCd, even when the proportions of gold and cadmium in the original 

 mixture varied within wide limits. Again (-), by alloying gold with excess 

 of cadmium and distilling off the excess of cadmium they obtained a 

 residue having the. composition AuCd. Now that many metals have been 

 distilled w vacuo this method may meet with success in other cases. 



M. Lebeau (^) dissolves metals in excess of sodium and distils off the 

 excess of sodium by the prolonged passage of ammonia gas followed by 

 that of nitrogen. He thus obtains the bodies SbNa^, BiNa^, Sn]Sra4 in a 

 pure state. He is also succeeding by the same method with the other 

 alkali metals. M. Joannis (■*) some years ago applied a similar method 

 successfully. 



The method of fractional solution enabled Debray (") to isolate the 

 bodies PtSn^, RhSng, RuSn;,, by the action of dilute hydrochloric acid on 

 alloys containing excess of tin. M. Le Chatelier {^) has in the same way 

 isolated the compound Cu^Sn. He emphasises the opinion that by 

 choosing a suitable solvent, suggested by electro-chemical considerations, 

 the method will be found generally applicable. For example, by sub- 

 jecting alloys of copper and zinc to the prolonged action of a paste of 

 lead chloride he has obtained crystals of pure ZnqCu. Mr. Heycock, in 

 a research not yet published, obtained large crystalline grains of PtAl3 

 by the action of hydrochloric acid on a slowly cooled alloy of aluminium 

 and platinum. Mr. Stead (^) has isolated in this way crystals of SnSb, 

 Au4Pb, AusPbg, SnaAs,, and probably of some other alloys. The work 

 is in some cases not yet published, but he has been kind enough to com- 

 municate the results for the purpose of this report. 



Other cases could no doubt be quoted in which fractional solution 

 leaves a residue ha^^ing a formula, but there is a great risk of the solvent 

 attacking the crystals ; and, as Mr. Stead has found, the existence of mixed 

 crystals, or, at all events, of crystals having a core different from the out- 

 side, is a sei-ious drawback to this method i-egarded as an independent 

 method of discovery. It would seem that the proper moment for the 

 application of these methods comes when by the microscope, by the 

 freezing-point curve, or by potential determinations the existence of a 

 compound has been already indicated. 



In a systematic study of intermetallic compounds I should therefore 

 put first that of the chemical equilibrium of the binary system : this is 

 generally expressed by the freezing-point curve. Next, and perhaps of 

 equal importance, comes the microscopic examination of the solid alloys. 

 Thirdly, as more limited in scope, but sometimes more emphatic in its 

 indications, comes the determination of the difference of electrical potential 

 existing between a metal and its alloys. 



The method of studying chemical equilibrium which we owe so largely 

 to Professors Bakhuis Roozeboom and Le Chatelier is now familiar to 

 most chemists, and in the case of a binary system it can be sufficiently 

 described in a few words. A mixture of two substances, A and b, in 

 certain proportions is melted. It is allowed to cool slowly, and the 

 temperature is noted at which solid matter begins to separate from 

 the liquid. This is the 'freezing-point.' It tells us the temperature 

 at which this particular mixture becomes saturated — that is to say, 

 comes into equilibrium with a particular solid. By repeating the 

 experiment with a series of mixtures of A and b we get as many 

 points as we need for plotting the freezing-point curve. In this curve, 

 one ordinate is the percentage composition of the mixture expressed 

 either in weight per cent, of A and b, or, better, in atomic or molecular 



