ON DILUTE SOLUTIONS OF ALUMINIUM IN GOLD. 
269 
Microscopic Structure. 
« 
All alloys whose percentage composition places them to the left of the line Aaa' 
show the same patterns of uniform polygons or counties; these are closely packed, 
there being no interstitial matter, and they often meet in straight lines (Plate 4, fig. l). 
In other words, the alloys contain only one phase—the solid solution that we call alpha. 
The alpha of the point a contains 2 per cent, by weight of aluminium, but at lower 
temperatures alpha can apparently contain a little more aluminium, as indicated by 
the slope of the line aa !; we are, however, not quite satisfied as to the exact position 
of the point a!. 
In the region aa'Ybb the ingots consist of a complex of alpha and beta, the alpha 
decreasing in amount as we go to the right and vanishing as we cross the line bh 
(figs. 2 and 3). The alpha vanishes at or about A1 19, and for about 2 atomic per 
cents, we have in the solids at a high temperature beta only (figs. 4 and 18). 
Polished and etched surfaces of beta that have been chilled at a high temperature 
always present the appearance of angular rocks seen in relief, a high power showing 
groups of fine parallel lines, the direction of the lines changing from grain to grain so 
as to give the effect of shading. This is due to a laminated structure in the beta, 
perhaps to an incipient decomposition. In some cases the ruled structure is extra¬ 
ordinarily regular and resists etching, remaining bright when the rest of the beta is 
darkened (fig. 5). 
The triangle bhc contains all the alloys that can exist in equilibrium as pure beta ; 
it will be seen that this substance becomes unstable at 424° G. To the right of the 
point C, which appears to be an eutectic point, we get primaries of the substance D 
embedded in beta, the D generally etching bright; indeed everywhere to the right of 
the line Lc the alloys are a complex of D and beta (figs. 6 and 16). 
We are hardly entitled, as far as experimental evidence goes, to separate the points 
C and c, for we have no certain evidence of eutectic structure in any alloys of the BC 
branch, but we think it probable that the solidus of these alloys ends a little to the 
left of C as we have drawn it. It must- be remembered that the G eutectic can only 
contain a very small proportion of D, and would be hard to detect; alloys such as 
A1 22, A1 23, and A1 25, when appropriately treated, do show a minute network of D 
in the matrix, this is perhaps the eutectic (fig. 7). 
The relations of the three bodies alpha, beta, and D are, so far as this diagram 
extends, very similar to other cases already studied, for example, in the copper-tins. 
The lines &L and cL are the lines of incipient crystallisation of alpha and of D out of 
the solid beta ; they correspond to the liquidus in crystallisation out of a liquid. The 
point L is an eutectoid point; when any alloy containing beta cools slowly through 
the L temperature the residual beta in it breaks up into a complex of alpha and of D. 
We have verified this in alloys ranging from A1 18 to A1 23 (figs. 8, 9, and 14). If 
the composition of the alloy as a whole lies to the left of L, the primary, or pseudo- 
