1892.] On certain Ternary Alloys. 389 



A. B. 



Bismuth 387 48 ! 2 



Zinc 35-3 31-8 



Tin.., 26-0 20'0 



100-0 100-0 



The point B evidently falls well outside the critical curve for 750, 

 although it is not distinctly outside that for 650. 



On applying Stokes' second system to the above values, the fol- 

 lowing figures result from a large-scale plotting: A + A' = 37, 

 B + B' = 120; whence the composition for the limiting point, L, 

 is : 



Bismuth 18'5 



Zinc 60-0 



Tin 21-5 



100-0 



These percentages of bismuth and zinc are not far from those re- 

 quired for the formula BiZn lo . 



Calculated. Found. 



Bismuth 24'4 18 5 = 23'5 



Zinc . , 75-6 60 = 76'5 



100-0 100-0 



Inasmuch, however, as an entirely different proportion is found 

 when silver is the "solvent" metal (approximately BiZn 2 ), this 

 cannot be regarded as much evidence of the existence of an atomic 

 compound of bismuth and zinc at the limiting point. 



On contrasting together the curves thus deduced for mixtures of 

 lead-zinc-tin and bismuth-zinc-tin, it is obvious that in each case the 

 curve for a higher temperature lies inside that for a lower tempera- 

 ture. When bismuth is the heavier immiscible metal, the curve lies 

 inside that obtained with lead instead of bismuth; apparently tbe 



same relationship holds in the case of bisi ^ u \ -zinc- silver alloys 



lead J 



and of I -aluminium-tin alloys, as will be hereafter discussed, 



lead / 



The limiting points above deduced in each case lie to the right (zinc 

 side) of the central line of the triangle, and the uppermost tie-lines 

 in each case slope to the right. With bismnth-zinc-tin alloys the 

 same disposition is also observed with the lower tie-lines, but with 

 lead-zinc-tin alloys the lower ties slope to the left, the reason for this 

 difference probably being that lead and tin exhibit a tendency to 

 combine together to form a definite compound, SnPb 3 , the formation 



