1891.] On certain Ternary Alloys. 177 



its nature, so long as there was any of it left, differing from that of 

 the other alloy. If, however, the point c lies in the line CL, on 

 increasing the quantity of C the two alloys merge into one. 



On communicating to Dr. Alder Wright this mode of graphical 

 representation, he tried it on a large scale on the results of two pairs 

 of series from the former experiments. In one pair the temperature 

 was 650, and the proportion by weight of zinc to lead was 2 to 1 

 in the first case, and 1 to 2 in the second. In the other pair the 

 weights of zinc and lead were equal, and the temperature 650 in 

 one case and 800 in the other. In the first pair the agreement of 

 the critical curves was very good, but the agreement in the direction 

 of the ties was not by any means equally good. In the upper part 

 of the figure, corresponding to the case in which there was a con- 

 siderable quantity of tin, though not enough by any means to pre- 

 vent the formation of two layers in the entire mass, the difference of 

 inclination ranged to about 5", the ties in the first case being inclined 

 to those in the second as if they had been turned round in the direc- 

 tion of a line passing through the lead corner of the triangle, and 

 turning round in the direction from lead- zinc to lead-tin. In the 

 second pair of series in which the weights of lead and zinc were equal, 

 and the temperature was 650 in the first case and 800 in the second, 

 the critical curve for 800 was of the same general character as that 

 for 650, but lay a little inside it, which is just what was to be 

 expected, on account of the increase of solubility attending the higher 

 temperature. Moreover, the critical curve for 650 agreed very fairly 

 with those for the same temperature in the first pair, notwith- 

 standing the difference in the proportion of lead to zinc in the three 

 cases. 



I had not anticipated the greater accordance existing between the 

 critical curves in different cases for the same temperature than that 

 shown in the direction of the ties. But, when the plottings revealed 

 it, it seemed to me that the cause was not far to seek. When the 

 molten mass has as yet been but slightly stirred, the superposed 

 alloys, supposed to be severally homogeneous, will most likely be 

 represented on the diagram by points, one or both of which lie out- 

 side the critical curve. In this condition an alloy represented by an 

 external point, having the metal C to spare, will be capable of dis- 

 solving bodily a portion of the other. This process accordingly, 

 being something analogous to the solution of a salt till saturation is 

 obtained, will go on as the stirring proceeds, and be sensibly complete 

 in a moderate time. The two alloys will then be represented by two 

 points lying on the critical curve. Such alloys may be said to be 

 associated. But the passage from merely associated to truly conjugate 

 alloys, as the stirring proceeds, is likely to be decidedly slower. For 

 now neither alloy can bodily dissolve any portion, however small, of 



