230 MESSES. C. T. HEYCOCK AND F. H. NEVILLE 



Summary. 



The results obtained as to the equilibrium between gold and aluminium may be 

 stated briefly. 



Each point on the diagram of Curve 1 corresponds to a mixture of a certain 

 composition at a given temperature. 



All points above the curve correspond to homogeneous liquids. A point P on any 

 branch of the curve corresponds to a state of possible equilibrium between a solid 

 whose composition is given by the summit, real or imaginary, of the branch and a 

 liquid whose composition is given by P itself. 



A point Q below the branch of the curve, but above its lower end, corresponds to a 

 mixture of a solid and a liquid. The composition of the solid is given 

 by the summit of the branch, and that of the liquid by the point R, 

 in which a horizontal through Q cuts the branch. An intersection of 

 two branches, such as C, D, F, or G, corresponds to a state in which 

 the two solids given by the summits of the two branches can both 

 exist in equilibrium in the presence of the liquid given by the 

 intersection point. 



The phase rule for the case of two components in a system where the pressure is 

 constant and the vapour pressure nil forbids the existence of more than three phases 

 in true equilibrium. But the microscope shows that if an alloy containing 45 atoms 

 of aluminium be cooled to the temperature of F, it must have contained the three 

 solids AuAl 2 , X, and Au 2 Al in contact with a liquid. This is true even in the case of 

 slowly-cooled alloys ; in other words, the purple AuAl 2 appears able to exist in the 

 presence of a liquid that is not saturated with it. But the sections which present 

 this paradox also give its explanation. In the slowly-cooled sections at 45 atoms the 

 crystals of purple are always surrounded by a coat of the white X, which forms on 

 them as soon as, by partial solidification, the state G is reached. Hence the crystals 

 of purple take no part in the later equilibrium corresponding to points on GF. 



The binary metallic system treated of in this paper has many points of resemblance 

 with the system iodine-chlorine that has been already worked out. It illustrates 

 general principles already accepted and adds nothing to them. But we hope it may 

 have some value as a contribution towards the slowly accumulating proof that metals 

 combine with each other according to the same laws that hold good for compounds not 

 wholly metallic. 



The metallurgist, moreover, who applies, as has already been partly done, the double 

 method of this paper to pairs of metals likely to be of use in the arts would probably 

 arrive at results of value to him. 



We have to thank Dr. J. C. PHILIP for his untiring and most valuable assistance in 



