MESSRS. C. T. HEYCOCK AND F. H. NEVILLE ON 
richer in tin than the earlier formed y. The central parts of the y combs have also 
broken up into a complex, which reminds one a good deal of Sn 23 and Sn 24. This 
again is evidence that the y is not a pure chemical compound. A little of this 
separation could he detected in the surface when etched with ferric chloride before 
the allo}^ was repolished for the heat-oxidation. A slowly chilled ingot, very sharply 
chilled at 715°, was almost untouched by the strongest mixture of hydrochloric acid 
and ferric chloride, but heat-oxidation still reveals a little mother-substance. 
The lower chills wei-e etched in the usual manner with ferric chloride. Many of 
these sliow polygonal divisions, hut very little difference in the tint of the polygons, 
and no pattern in them. Even the strongly exothermic change at E' does not show 
itself by any })attern in the pure Sn 25, for the obvious reason that both before and 
after the change there are no two substances for the etching reagent to distinguish 
l^etween. In quite the lowest chills, and in the unchilled ingots, the patches 
sometimes assume a bar-like form, especially if the polished and etched surface is 
exposed to the air for some days before examination. This appears to be due to the 
fact that the plates of q, of which the ingot is composed when the temperature is 
below E', oxidise at different rates according to their position with reference to the 
plane of the section. This effect is seen in fig. 66, an unchilled ingot. A similar 
selective darkening of the rj by atmospheric action was observed in a chill of Sn 28 
at 725°. 
The EE Alloys. Sn 26 and 27.—This region is not marked by a singularity in the 
liquidus, and in harmony with this fact the first steps in the solidification are very 
similar to those of the preceding alloys. But there is a greater difference between 
tlie composition of the first solid and that of the liquid, and consequently the y combs 
of tlie upper chills are revealed by polisliing alone through the more brilliant network 
of mother-substance around them. This effect is common to the iqiper chills of all 
the alloys as far as the point G of the diagram. This network is rapidly eaten out 
by etching re-agents, leaving the unchanged y combs j^ale in a black network. 
Exanq^les of the appearance are given in the photographs of the succeeding alloys. 
Below our solidus this network of mother-substance vanishes, and the only pattern is 
that typical of a chemically uniform solid made up of grains. The chills of Sn 26 at 
677° and 656° are both (fi‘ this character. A chill at 640°, that is, below the 
transformation curve, is extremely uniform, and most of the surface defies all the 
efforts to etch it, but a high power enables one to discover some places where a new 
network has grown, due, we presume, to the laeaking-up of the solid solution into 
two phases, y and a little of a tin-rich phase. It is pi’etty certain that we may 
venture to he more precise, and say that the change is the crystallisation of the 
conq:)Ound CuySn out of a solid solution containing rather more tin. This change is 
not complete at 640°, and there is very little of the new tin-rich phase. The chill at 
612° shows the whole surface broken up into the characteristic bars of CiqSn, and the 
dark lines between, where the tin-rich phase has been eaten out by the etch. M' e 
