THE CONSTITUTION OF THE COPPER-TIN SERIES OF ALLOYS. 63 
tin-rich material (black in the figure) crossing them; these suggest that the very first 
'Solid may have been y. In Sn 45 this feature is not so well marked. Phis peculiarity 
perhaps indicates that the curve FG might have been traced for some distance 
below GH, in a region of unstable equilibrium. 
If we compare chills, taken at, or a little below, 400°, of all the alloys from Sn 27 
to Sn 87, that is almost from E to H, we find the rj decreasing in amount from a 
quantity that almost fills the alloy at Sn 27 to a vanishingly small quantity, and 
with the larger percentages of tin the ingots are very liquid at the moment of 
chilling. Figs. 83 and 84 of ingots of Sn 38 and Sn 50 that had been kept in the 
vapour of boiling sulphur (at 445°) for 70 hours and then sharply chilled, show the 
decrease in the 'rj with increasing percentage of tin, and fig. 85 of Sn 80 chilled 
without a very slow cool at 433 ° illustrates this decrease in a more marked way. 
In the sulphur-boiled ingots in which the r) has been in contact with the liquid for 
a long time, no doubt at a slightly varying temperature, the bais are lounded, and in 
fig. 83 one sees a tendency for them to group themselves into families. This rounded 
appearance is not seen in the rj except in the sulphur-boiled ingots; the more usual 
appearance is of great angularity in the outline of the crystals, as in fig. 86 , of a slow- 
cooled chill of Sn 50 that was maintained at 420° for some hours and then chilled. 
In this ingot the plane of section coincided with the plane of the plates; an inspection 
of the outside of the ingot shows that a section at right angles would have cut the 
plates at right angles and given a pattern of bars. 
If one of the FGH alloys is cooled below 400° before chilling the amount of t] is not 
i^icreased but the bars are margined by a new substance, and the same new body is 
present as chill primary between the bars. This new substance H is the primary 
crystallisation of the branch HI of the liqiiidus. H and r) are both very little acted 
on by hydrochloric acid, especially if excess of tin is present, and etching consequently 
leaves them both white on a dark ground. But the 17 gradually oxidises in the air 
and darkens so that it is readily distinguished from the H. Fig. 87 ol Sn 73, chilled 
at 373 °, illustrates the effect. It will be noticed that the bars of 17 are true crystals 
bounded by plane faces. They remind one of the twinned bars of 77 seen in fig. 57. 
If we chill below 400° a little H can be detected even in Sn 26, ljut it is well seen in 
the ingot of Sn 29 that was chilled at 380°, after having been kept at this 
temperature for four hours (fig. 88 ). In this ingot, except a few holes which may 
have been full of tin-rich mother-substance before the etch, the 17 and H fill the 
ingot. The conversion of the 77 into H by reaction with liquid at the temperature of 
400° requires more detailed study, but we will reserve this point until the survey of 
the remaining groups is complete. 
The HI Group .—The singularity H occurs in the liquidus at Sn 87. The curve 
does not change its direction perceptibly here, but Boberts-Austen and Stansfield 
found a row of second halts in the cooling at this temperature in all alloys from 
about Sn 27 to Sn 87. We have not verified their results pyrometrically, but the 
