ON DILUTE SOLUTIONS OF ALUMINIUM IN GOLD. 
273 
recalescence produces a fine diaper of alpha, but the alpha primaries previously 
existing remain uniform, as apparently the alpha cannot change at this temperature. 
An ingot of A1 19 was, as the etched surface shows, forced to recalesce at 507° C. ; 
although no thermal effect was noticed it shows a network of alpha with no primaries. 
Another ingot of A1 19 was recalesced at 504° C., the temperature rising to 515° C., 
and was chilled a few minutes later at 510° C. It shows a magnificent pattern of 
bars, many of them large, and of feathery masses of Y, this substance being every¬ 
where outlined by bright alpha. It is very evident in this case that a complete 
recrystallisation has taken place throughout the ingot, the new pattern being 
determined by the shape of the crystals of Y. The control ingot was fine-grained, 
rocky, very uniform beta. 
We have forced the recalescence of A120 at 397° C. and 445° C., these two 
experiments only confirming our previous results. A forced recalescence at 498° C., 
rising to 520° C., is a remarkable example of the recrystallisation of a solid being 
brought about by a stimulus. The picture is one of broad bands, bars, and sprays 
of Y, margined as usual by alpha; at the highest temperature reached the Y must 
have been very near its melting point. Another ingot, recalesced at 514° C., rose to 
524° C. and remained at that temperature for some minutes. The pattern shows that 
a good deal of the Y melted in consequence of the high temperature attained. The 
pools of matrix are not pure alpha but a complex resembling the feathery pattern 
of a high chill of beta. Here we have a case of a solid that, when touched by a cold 
wire, rises ten degrees, completely recrystallises and partly melts. Another ingot 
was recalesced at 508° C. and rose to 523° C. The figures 18 to 21, inclusive, illustrate 
forced recalescence. 
These results confirm the view that Y cannot exist much above 525° C. Another 
experiment supports this conclusion. A portion of the ingot recalesced at 498° C., 
and described above, was reheated to 525° C., kept at this temperature for 12 hours, 
and then chilled. The pattern shows that quite half the Y melted, or disappeared, 
the solid part appearing to have remained in equilibrium with the liquid ; but we do 
not know whether the matrix continued' liquid until the chill or solidified previously 
as beta. The solidified matrix has a coarse striation and might be a complex of alpha 
and Y. The equilibrium between Y and liquid, that we perhaps see here, could only 
be arrived at in an indirect way ; if we had attempted to reach it by cooling molten 
A1 20 the alloy would, at 525° C., have become a mass of solid beta. 
A1 22 was recalesced at 478° C., the temperature rising to 520° C., the ingot cooling 
in a few minutes to 486° C., when it was chilled. A control was treated in the same 
way but not recalesced. Both ingots have large crystals of D approximating in 
shape to those that separate along the cL line, but the recalesced ingot, alone, has 
a golden network of alpha in the matrix identical with that of the recalesced 
A1 20. We have also forced a recalescence in A1 23, but the pattern shows very 
little change. 
VOL. CCXIV.-A. 
2 N 
