1903.] The Properties of the Aluminium- Tin Alloys. 283 



illustrate the process in the case of the alloy containing 15 '42 per cent, 

 of aluminium. No. 4 is a polished area of the cast metal. Heating in 

 a current of dry hydrogen gas causes no marked rearrangement of the 

 constituents at 216 220 C., even when continued for 28 hours. 

 The polished surface, however, becomes roughened and presents a 

 warty appearance. Similar results are produced when the experiment 

 is repeated at 228J 232 C., that is at a temperature just over the 

 melting point of the eutectic. A more marked change is noticeable at 

 250 C., and a tendency to rearrangement is apparent, even after 

 9J hours at the temperature. Finally, when the annealing process is 

 repeated at 290 300 for an equal time, 9| hours, the repolished and 

 etched surface shows the structure of the alloy to have been profoundly 

 modified thereby, and the appearance under the microscope is now that 

 characteristic of the annealed metal (fig. 5). The round, oval, or some- 

 times irregular areas of aluminium-rich constituent, show up brilliantly 

 white in the midst of the darker coloured eutectic. The eutectic is 

 soft and on that account somewhat difficult to polish, but, if carefully 

 treated and etched with a 2 per cent, caustic potash solution so as to 

 remove the skin of smeared metal from the surface, it exhibits under a 

 magnification of less than 200 diameters the wavy or striated structure 

 characteristic of eutectic mixtures generally (fig. 6). 



The polished and etched plates of these alloys, both cast and 

 annealed, must be examined soon after they are prepared, since the 

 surface rapidly becomes pitted, owing to the action of atmospheric 

 moisture. In the annealed ingots this pitting is seen to be confined to 

 the eutectic areas, and is apparently due to the incomplete segregation 

 of small nuclei of the constituent of higher melting point from the 

 eutectic mixture. Until the percentage of tin in the mixture exceeds 

 that represented by the point C in the curve (Diagram 1), the amount 

 of eutectic visible in the polished plates is not proportional to each 

 addition of tin made. When this point is passed a general proportion 

 between the two may be stated to exist. 



Putting together the evidence from the curves, the hydrogen evolu- 

 tions (Diagram 2) and the microscopic structure, it may be concluded 

 that there is a substance or series of substances present in these alloys 

 marked by the common property of evolving hydrogen from water. 

 The only very marked interruption to the continuity of the curve is in 

 the region CD in the vicinity of an alloy corresponding to the formula 

 AlSn. In the fluid mixtures containing larger percentages of aluminium 

 than 18*5, there are more atoms of aluminium than of tin. From 

 these media there appears to separate at successively lower temperatures 

 solid matter which contains aluminium and tin in the proportions 

 AlgSnj,, where the value of x is at first greater than y, but is gradually 

 approximating to it. This separation is taking place from a fluid 

 which always contains more atoms of aluminium than of tin. The 



