10 
MESSRS. C. T. HEYCOCK AXD F. H. NEVILLE ON 
from the liquid, but appears as a recrystallisation out of homogeneous solid ^ or y. 
All these substances, except /S and y, are stable under certain circumstances at 
ordinary temperatures and are therefore found in slowly cooled unchilled ingots ; on 
the other hand, the two latter are unstable at low temperatures, and are only found 
in certain chilled alloys. 
The angle C of the licpiidus indicates that the composition of the solid phase here 
changes abruptly,"^ for while tlie branch ABC corresponds to the solidus A5, the branch 
CD corresponds to the solidus Ic. The angle at C was a great difficulty to us so lono’ 
as we only examined alloys that had not been chilled, but Roozeboom’s theory 
explains in the most perfect manner all the pheiiomena connected with this angle. 
It tells us that just above the temperature C the cooling saturated liquid deposits, 
and is in equilibrium with the a solid solution whose composition is given by the 
point 6, while just below the temperature C the liquid forms /3 solid solution, much 
richer in tin and given in composition by the point 1. Thus as the saturated liquid 
cools through the temperature C an isothermal transformation + liq^ = takes 
place. The heat evolved by this reaction is well marked in the cooling curves. No 
uniform solid solution of percentages between h and 1 can exist. 
Tlie angle D indicates another break in the series of solid solutions, and we have 
indicated a corresponding break between c and m in the solidus. The study of the 
chills in this region affords, as will be seen later, some evidence for the break cm, but 
this evidence is not altogether conclusive. However, we propose to speak of the solid 
solutions of the branch mdef as y crystals, to distinguish them from the yS crystals of 
the branch Ic. Both the cooling curves and the microscope indicate that in alloys 
between Sn 17 and Sn 20 an isothermal exothermic reaction occurs when the 
temperature falls to that of the point D. This reaction must be either /3 + liq = y 
or /3 = y. The first equation corresponds to our figure, in which c and m are 
separated, the second to the case of m and c being really coincident. 
Thus the branch ABLC of the liquidus deposits a solid solutions, the branch CD 
deposits 73 solid solutions, and the branch DEFG deposits y solid solutions. 
The branch GH of the liquidus deposits crystalline plates of the substance p, which 
in this region consists of nearly or quite pure CusSn. 
The branch HI deposits crystals of the substance H, which is certainlv very near 
in composition to CuSn, although it has a slight excess of copper, probably due to a 
small amount of p in solid solution. 
The liquid of the branch IK deposits crystals that must be very nearly j^ure tin. 
* During the experiments for the determination of the liquidus, described in a jjrevious paper (‘ Phil. 
Trans.,’ A, vol. 189, p. 50), the cooling alloy was sometimes stirred by means of a hand stirrer during the 
freezing-point experiments. On such occasions we noticed a marked difference between the precipitates 
formed a little above and a little below C; above C the solid forming was soft, and we compared it to 
mud, but as soon as the temperature C was reached the precipitate became hard and gritt}', and we 
compared it to a sharp sand. 
