48 
.MESSES. C. T. HEYCOCK AND F. H. NEVILLE ON 
traces of separation are as yet so slight that they have to be very carefully searched 
for. We do not reproduce this, because there is so little to see, and because what 
pattern there is closely resembles that of Sn 18 chilled at 550° (fig. 48), with this 
difference only, that the white is mucli more abundant in the Sn 18. 
The chill of Sn 17 chilled at 495°, when etched as usual with dilute FeClg, is still 
largely blank (fig. 46), but the white hands bordering the polygons are now much 
broader, and near some of the borders and angles there is a delicate but perfect 
growth of the white S in sprays that we are in the habit of calling the fein-leaf 
crystallisation.” The central and much the larger part of each polygon is, however, 
blank, so that we have in this ingot a transition from the uniform solid solution to the 
two-phase state jiroper to the region below the curve XD'. According to the diagram, 
this ingot was chilled a little below the line. It is important to notice that the daik 
ground is uniform even to a high power, and that when examined by oblicjue light it 
shows no iridescence; the ground has not yet broken up into the C eutectic. But 
the chill at 450° (fig. 47) shows the ground broken up into a minute complex,* and 
also the fe^ii-leaf scattered umformly over the whole surface, ihis ingot etches with 
very weak FeClg and has the patchy iridescent efiect of the low chill of bn 15 5 
due to the formation of the C' eutectic. These lower chills of Sn 1/ were not slovly 
cooled before the chill, and hence the detail is small, hut two ingots of Sn 18 (fig. -±8) 
and Sn 19 (figs. 49 and 49a), both chilled at 550° after a very slow cooling, show the 
nature of the change of the Xh)^ curve even lietter, and fix a point on it accuiately. 
In the chill of Sn 18 at 550° (fig. 48), prolonged etching with FeClg brings out very 
little pattern. There are, as the photograph shows, only a few small specks and 
threads of S. This ingot was evidently chilled when in the triangle /XE' of uniform 
solid solutions. On the other hand, the chill of Sn 19 at the same tempeiatuie, e^en 
when lightly etched with FeClg, develops a splendid pattern of white tin-rich S in the 
form of hands, fern-leaf, and liars. Hence the ingot must have been chilled belov 
the transformation curve. We also give a chill of Sn 18 at 501 (fig. 50), in vhich 
the H' crystallisation has taken place. It must be remembered that the curve 
XD'E' has lieen determined from the halts in the cooling curves and not from the 
chilled alloys, but we see that the latter confirm its position. 
The copper-rich gi'ound in the chill of Sn 19 at 550 is uniform, hut iinchilled 
ingots, or those chilled below X', show the ground to be the C' complex, fig. 53. It 
must also be remembered that m the chill of Sn 19 at 550 the crystallisation of the 
S was arrested by the chilling liefore it was complete, hence this ingot does not show 
the 8 fern-leaf so closely packed together as would he the case in an unchilled ingot. 
Sn 20. 31‘8 y>er cent, hy iveight of tin. 
This alloy, the “ speculum metal ” of Lord Bosse, has the formula CkpSii, and the 
* This feature requires a higher maguificatiou thau that of fig. 4 1 . 
