ON THE CHEMICAL COMPOUNDS CONTAINED IN ALLOYS. 



137 



see two well-marked summits, at o and o', the first being exactly at the 

 formula NaCdg. It is evident from these curves that the summits of 

 many of the branches lie below other branches, and therefore correspond 

 to unstable states. The position of the summit and the formula of the 

 compound must in such cases be a matter of speculation until experi- 

 ments of another kind have been made.^ 



M. Kurnakov has also studied the crystalline matter which separates 

 at the freezing-points, and has noted great variations in this as one goes 

 from one branch to another. He separated by filtration the hexagonal 

 plates which crystallise at points along the branch bc, and analysed them. 



R^Hg KUc, 



Fifi. 5. 



RHg 



RHg^ RHcj^R\ nncj^ Kffcj^ 



10 





40*' 50'" 60"" ./ 70> 



Atomic percentage. 



WJVa 



He found, however, that the composition of the crystals varied from 

 point to point along the branch. He attributes this to the presence of 

 mother liquor attached to the crystals, and he thinks that the crystals, if 

 free from mother liquor, would have had the formula NaaHg. This is a 

 very reasonable supposition, but it may be that he was examining mixed 

 crystals. The formation of mixed crystals of two bodies which crystallise 

 isomorphously is certainly common in alloys, and may be the cause of 

 singularities in the freezing-point curve. Indeed, it is doubtful if we 



' The amalgams of sodium and potassium have been examined by various other 

 methods ; for example, by the determination of their specific volume— a method 

 recently employed by E. Maey (Zeits. Phys. Chemie, xxix. p. 119). He finds a 

 number of angles in his plotted curve which he attributes to the existence of 

 compounds. 



