BANCROFT. — TERNARY MIXTURES. 329 



the mixtures which are saturated in respect to water, and not in respect 

 to chloroform. Either water or chloroform, when added to these solu- 

 tions, produces a precipitate of water. These two sets of solutions :ire 

 easily distinguishable qualitatively, because in the first case the new 

 phase, containing a large percentage of chloroform, is denser than the 

 mixture from which it separates, while in the second case the new 

 phase, containing chiefly water, is lighter than the original solution. 

 The point where the new phase changes from being denser to beino- 

 lighter than the first phase is the point of intersection of the two curves. 

 At this point only is the nature of the precipitate determined by the 

 nature of the infinitely small excess added. The intersecting point 

 represents the concentration at which, were chloroform and water 

 solids at that temperature, both could be in equilibrium with the solu- 

 tion and its saturated vapor. It corresponds to the concentration of a 

 solution containing two salts with a common ion which is in equilibrium 

 with the two solid salts, formation of a double salt being excluded. In 

 one respect the analogy between a system having three liquid compo- 

 nents and one composed of two solids and a liquid does not hold. If 

 to a saturated solution of silver bromate silver acetate is added, the 

 precipitate is silver bromate, and, conversely, the precipitate is silver 

 acetate if silver bromate be added to a saturated solution of silver 

 acetate. The salt with the less concentration precipitates the one with 

 the greater, up to a certain point. In a chloroform-water-alcohol mix- 

 ture in which chloroform is present in large quantities, the precipitate 

 is water, or the substance with the greater precipitates the one with the 

 lesser concentration. This difference of behavior is due to the new 

 phase being a solid in the one case and a liquid in the other. By a 

 suitable choice of the three components, and by varying the tempera- 

 ture, the substance in respect to which the solution was saturated 

 could be made to separate either as a liquid or a solid phase, and this 

 difference could be made zero. The transition point would come 

 when the equilibrium was between four phases, one solid, two liquid, 

 and one gaseous. 



There is no apparent theoretical reason why the two curves should 

 not be prolonged beyond their intersection ; but there is a very good 

 practical one. Beyond the point of intersection the curves denote 

 saturated but labile solutions, and a supersaturated system composed of 

 liquids is almost impossible to realize. When I come to the study of 

 ternary mixtures having one or more solid components, I hope to be 

 able to follow one of the curves at least beyond the intersecting point; 

 but in the present work I have made no such attempt. 



