64 
MESSRS. C. T. HEYCOCK AND F. H. NEVILLE ON 
microscope amply confirms them. The evolution of heat is clue to the reaction 
7) + licjuid = H. With perfect equilibrium transformations the whole of the r) in all - 
alloys between the percentages given by the points H' and H ought to disappear at 
temperatures below 400° in consecjuence of this transformation, but in the alloys as 
usually prepared the H only forms a thin coating over the crystals of r], much of the 
7) remaining even in ingots that have not been chilled. We have however succeeded, 
by maintaining the ingots for a very long time at temperatures a little below 400°, 
in removing all 17 from alloys between the points H' and H. We shall return to this 
point. 
With more tin than that contained in Sn 87, that is, in the HI alloys, the primary 
crystallisation should consist of H, and, in fact, as an unchilled ingot of Sn 90 shows 
(fig. 89), the alloy consists of primary crystals of H in a eutectic of H and tin. The 
crystals of H are often hollow, as if they had grown round something that had 
afterwards disappeared; this again suggests the possibility that, through a surfusion 
of the H, the 77 may have formed first and been decomposed when the H began to 
crystallise. We are not cjuite confident that the above is the true explanation of the 
peculiar and well-known form of the H. It will be noticed that the exterior of the 
crystals of H is rectilinear and angular, and tliat in the neighbourhood of a large 
crystal of H the mother-liquid has been so much exhausted of that body that there 
are no particles of H in the eutectic, while in places remote from a large crystal of H 
the eutectic is full of minute specks of that body. The same feature can be detected 
in the long slender crystals of H in an unchilled ingot of Sn 95 (fig. 90). 
The IK Alloys. —-With more tin than corresponds to the point I, it is no longer a 
copper-rich body which crystallises first, but a tin-rich body which must he very 
nearly pure tin. A chill of Sn 99, when well polished, showed primaries of a new 
type in a eutectic, but etching reagents destroy the pattern of primaries and only 
leave the patchy appearance of a solid solution. This end of the series deserves more 
attention than we have as yet been aljle to devote to it. 
The Transformation 77 liquid = H. 
This reaction, which commences when any alloy of a composition between Sn 25 
and Sn 87 sinks below the temperature of 400°, is necessarily a slow one, for as soon 
as the plates of 77 have become covered with a coating of IT, the process can only 
proceed by a dilTusion through the thickening layer of H. The diffusion is fiir from 
rapid at temperatures near 400°, and at much lower temperatures it practically 
ceases. Consequently, with ordinary rates ot cooling, these alloys contain when cold 
tliree phases instead of two. This is recorded in tlie diagram by placing a bracket 
round the symbol of the third phase whose presence is due to an incomplete 
reaction. But, although this incomplete equilibiium is the usual condition of the • 
alloys, it can he avoided. We have succeeded in transforming all the alloys from 
