RECENT PROGRESS IN THE STUDY OF ALLOYS. 87 



tinuous, as will be the case when some of the solid is already 

 present, the thermometer will indicate a slower rate of cool- 

 ing, or, if the system is protected from too rapid loss of 

 heat, the temperature may become constant for a time. 

 But molten metals, like other liquids, show the phenomenon 

 ot sin-fusion, that is to say, the liquid, if no solid is present, 

 can be cooled below its proper melting point. It is then in 

 an unstable condition, and sooner or later this state ends 

 by the rapid formation of a considerable amount of solid 

 with the evolution of a corresponding amount of heat. If 

 the system is so arranged that losses of heat by radiation 

 and conduction are slow, the heat evolved at this stage may 

 raise the temperature of the whole mass considerably, the 

 highest temperature attained being at, or very nearly, that 

 at which solid and liquid can exist in permanent equilibrium. 

 This temperature, the freezing point as it is convenient to 

 term it, is thus often indicated by the thermometer with 

 great delicacy. 



If we arc dealing with a pure metal the freezing point is 

 remarkably constant, the temperature after the period of 

 surfusion is over, remaining absolutely the same until the 

 whole mass is solid. Conversely, if heat be slowly added 

 to a mass of the pure solid metal, the temperature will 

 become constant as soon as fusion commences, the melting 

 and freezing points being absolutely identical. 



When the metal is no longer pure, but is alloyed with 

 a small quantity of some other metal soluble in it, the 

 process of solidification obeys the law of Raoult. The 

 metal present in excess, which we may term the solvent, x no 

 longer solidifies at its true freezing point ; but it does not 

 begin to do so until a lower temperature is reached, the 

 pure solvent separating at continually lower and lower 

 temperatures as the liquid becomes a stronger solution of 

 the dissolved metal. 



MVhen two substances are mixed and form a homogeneous liquid, it 

 is difficult to give a rule for deciding which of the two is the solvent. 

 There is a tendency to regard that substance as the solvent which separates 

 out in cooling. But this rule breaks down when we consider the cooling 

 of a saturated solution. 



