338 ON TWO NEW CRYSTALLINE COMPOUNDS 



especially in the ternary alloys of zinc, tin, and lead, which were shown by the Messrs. 

 Svanberg,* on the same grounds, to contain a definite compound, ZnSug, 2 PbSnj. 

 The fact that, in a melted mixture of bismuth and zinc, which do not alloy together, 

 there are two stationary points of temperature coinciding very closely with the melting 

 points of the two metals, seems to support Rudberg's opinion.f It is well known that 

 Newton's (or Arcet's) alloy BiaPb^^Sna, after it has been rapidly cooled from a melted 

 condition to 57° C, or even a few degrees lower, becomes suddenly heated by a sponta- 

 neous evolution of heat, accompanied with a very considerable expansion of the mass. 

 Since, after the expansion, the specific heat of the alloy is the mean of that of the 

 metals which compose it, and since the change is accompanied with a manifest altera- 

 tion of texture, Person J argues that the phenomenon indicates an actual chemical de- 

 composition in the whole mass of the alloy, and that the combination between the 

 metals is only momentary, and confined within certain temperatures. Similar facts he 

 shows to be true of Rose's alloy Bi2PbSn2, and also of BigPb^ and BigSn.!- 



Croockewit,§ by melting together copper, tin, lead, and zinc in different atomic pro- 

 portions, stirring the melted metals while cooling, and after partial solidification turn- 

 ing out the still fluid portion, obtained crystalline masses which, as he found by 

 analysis, approached, and sometimes very closely coincided with, the calculated com- 

 position of CuaSnj, CuSn, CugSn, CuaZn^, CuaZuj, CuaZn, CuoPbg, CuPb, SnZnj, SnZn, 

 Sn.jZn, SnPb^, SnPb, SngPbg. He obtained, moreover, similar results in regard to the 

 amalgams, though not quite so satisfactory, and draws the conclusion, that the binary 

 alloys of these metals are mixtures of the above compounds either with each other or 

 with one or the other of the metals. Rieffel || makes seven different compounds of tin 

 and copper, CuSn, CuSn24, CuSn^g, SnCug^, SnCu^g, SnCu^a, SnCugg, of which CuSn 

 " crystallizes in large and exceedingly characteristic plates," and CuSn24 and CuSn^g 

 " both in needles . radiating in all directions from numerous centres." These com- 

 pounds do not correspond to those of Croockewit, and the discrepancy is probably 

 owing to a variation in composition similar to that, hereafter to be explained, in the 

 compounds of zinc and antimony, which wUl be found to resemble those of copper and 

 tin, as described by Eieffel, at least in their crystalline characters. 



* Poggendorf, Annalen, Vol. XXVI. p. 280. 



t M. Fournet, Ann. de Chim. et de Phys., Vol. LIV. p. 247. 



I Ibid., Vol. XXIV. pp. 143, 148. 



§ Journal fur prakt. Chemie, Vol. XLV. p. 87. 



II Compt. Rend., Sept., 1853, p. 450. 



