44 
MESSRS. C. T. HEYCOCK AXD F. H. NEVILLE ON 
tli6rG IS still a consiclGralilG amoiint of tliG tin-ncli motliGr-siiLstancG, thus tliG ingot 
was not altogotliGi- solid at tliG moniGiit of chilling. TIig has a fino uniform grain, 
with no tracG of striation or otliGr changG, and tliG margins of tho combs arc quite 
sharp. There was evidently a considerable difference between the composition of the 
solid and the liquid at the moment of chilling. 
Sn 16. V.s.c. chill at 738° (fig. 39). 
The ingot was allowed an hour and a-half to cool from 752° to 738°, in addition to 
the time given to the slow cooling of the preceding alloy. The ingot was quite solid, 
showino- no trace of granulation. The hold regular gridiron of tlie more rapidly 
^ ^ • 1 
cooled alloys has disappeared from the outside. A polished and etched section makes 
it evident that the primary combs have now grown so as to touch each other nearly 
everywhere, but there are two small interspaces, covering together less than 1 per 
cent, of the area, and full of the white tin-rich mother-substance; here we can detect 
the rounded form of the lobes of primary. One also begins to see polygonal boundaries, 
meeting three together in angles of 120°. There are not many examples of this, and 
the boundaries are rather geometrical lines than ribands of perceptible width, but 
these lines run out of the two or three little pools of white, and it is clear that the 
lioundaries are where the primaries met. 
Here we observe a process of solidification that can only be explained by the 
theory of solid solution ; the copper-rich solid grows and at the same time assimilates 
the tin-rich liquid, while the unalisorhed liquid down to the last drop remains quite 
different from the solid phase; the process we observe is exactly predicted by 
Hoozeboom’s theory. It is clear that the alloy does not solidify by the same steps as 
those that a pure or nearly pure body would follow, for in the case of such a 
substance the solid first forming would differ very little from the remaining liquid, 
while we find that in the case of Sn 16, however little solid has formed, it is 
consideralily richer in copper than the mother-substance. If a pure body differing 
from the liquid in composition crystallised, it could never grow so as to absorb the 
whole of the liijuid, and the resulting solid would be a complex of two kinds of 
crystals, a condition which does not exist in Sn 16 immediately below the solidus. 
The alloy is very uniform, showing no iridescence or striation. 
These two chills would lead us to fix the temperature of complete solidification of 
Sn 16 at a point not liigher than 740°, but we had some reason for suspecting that 
the pyrometer had altered a little lietween its last standardisation and the moment 
when tliese two chills were taken. We therefore made two independent chills at 
750°. These were cooled very slowly before the chill, and on each occasion the 
pyrometer was carefully standardised before the experiment. Each of these ingots 
shows the alloy as completely solid, no mother-substance being visilfie, and the combs 
of /B having disappeared and l 3 een replaced by the chai'acteristic polygons bounded by 
