ON GOLD-ALUMINIUM ALLOYS. 227 



running through them. This soft grey easily-scratched body increases in amount as 

 we add more aluminium up to the end of the curve. The purple body seems to retain 

 its character unchanged, although decreasing in area. With 99 '3 atoms of aluminium 

 one still sees small occasional spots of the unchanged purple, on the dirty grey ground, 

 which is no doubt aluminium. Hence there is probably no new compound between 

 H and the aluminium end of the curve. The alloys with more than 66' 6 atoms of 

 aluminium that were extracted for analysis polished so badly that we have not 

 attempted to take photographs of them. But a specially made ingot of alloy, con- 

 taining about 20 per cent, by weight of gold, that is 96 '7 atoms of aluminium, was 

 successfully polished (fig. 29). On the curve this alloy would be a little to the left 

 of the eutectic angle I, and the pattern of the polished section naturally resembles 

 that at a corresponding position on the other branches. It consists of rather scanty 

 rows of small spots of the purple body arranged in straight lines at right angles to 

 one another. 



We give also a Rontgen ray photograph of a thin section of this alloy cut 

 parallel to the surface of the previous section. This has been enlarged 5 diameters. 

 In this shadow-photograph the opaque AuAl 2 is dark, while the eutectic ground, 

 which is almost wholly composed of the transparent metal aluminium, is light. The 

 resemblance between the pattern of the two is complete, but the Rontgen photograph 

 naturally shows the numerous crystals in the body of the alloy as well as those on 

 the surface. 



Slowly -Cooled Alloys. 



It might be thought that the rapidly-cooled samples of alloy extracted by the Jena 

 pipettes would not give a fair picture of the character of the crystallisation of a more 

 slowly-cooled mass. For this reason, and also to obtain larger detail, we prepared a 

 number of slowly -cooled ingots of alloy by the method described in our paper on 

 X-ray photography (' Journ. Chem. Soc.,' 1898, p. 721). Each ingot weighed about 

 100 grammes. We found, however, that these ingots when cut and polished gave 

 results identical with, though on a larger scale than, the quickly-cooled alloys. The 

 slowly-cooled alloy containing 26'1 atoms of aluminium (fig. 12) shows by its marked 

 rotation effect that the slow-cooling facilitates the grouping of the crystals into 

 aggregates having the same orientations, and other slowly-cooled alloys confirm this. 

 But as we have not found the slowly-cooled alloys contradict the evidence from the 

 quickly-cooled ones, we do not give a detailed description of the former. There is, 

 however, one exception to this due to the great difference in specific gravity of 

 aluminium and the body AuAl s . A slowly-cooled alloy of these two substances, 

 that is, one on the branch HI, will show a settlement of the heavy crystals of 

 AuAlj to the lower part of the crucible. The Rontgen ray photographs (figs. 30 

 and 31) of the slowly- and quickly-cooled alloy, with about 96'6 atoms of aluminium, 

 show this difference. We see in the photograph of the slowly-cooled alloy 



2 G 2 



