ON THE STRUCTURE OF METALS. 
431 
the same as is caused by annealing steel castings at a temperature of about 800°. 
Nothing remains of the original structures. In the new grains the paste, the 
secondary crystallization, and the joints present the same characteristics as in the 
old grains. 
The alloy with thallium undergoes a strictly analogous transformation, but the 
dominant lines of the initial structure have been preserved in several places, notwith¬ 
standing the internal rearrangement (photo. 19, 150 diameters). The alloy with 
antimony behaved in a very different fashion ; this alloy, which was adequately 
represented before annealing at 200° by the gold-indium, photograph 3(17 diameters) 
and by 14 for the details (150 diameters), is now represented by the photographs 8 
(17 diameters) and 22 (150 diameters). The original structure has disappeared, but 
the new organization shows neither the polyhedral grains nor the continuous network 
of joints ; the incipient grains, indicated by the crystalline orientations, have not 
clean faces, and the sulphuric acid only traces fragments of broken joints. This 
structure recalls that of hardened steel of medium hardness. 
The alloy with aluminium undergoes the same transformation, but only locally and 
partially. Generally speaking, it seems that bismuth, thallium, antimony, and 
aluminium, when present in the proportion of about 0’2 per cent., behave in respect 
to gold in the same way as carbon does with regard to steel, but at a much lower 
temperature. 
It also appears to be evident that the bodies in question must have been present 
in the solid metal in a state closely resembling the fluid. None of these bodies 
possess a very high melting point, and, as is natural, this circumstance favours the 
maintenance of fluid molecules at a low temperature. The temperature, however, 
of annealing remains much below that of melting aluminium or antimony, and even 
below the melting points of the eutectic alloy of aluminium and gold (about G00°) or 
of antimony and gold (440°), and it at most reaches that of the fusion of bismuth. On 
the other hand, lithium, with a point of fusion below that of bismuth, and zinc with 
a fusing point below that of antimony, have not exerted a similar effect in lowering 
the temperature of annealing.'" The melting point of the impurities, although it is not 
without influence, is not the sole factor to be considered. It should be observed that 
lead, judging from what is known of its action on gold, probably behaves like bismuth 
and thallium. Amalgamation also appears to be a phenomenon of the same kind, 
possibly occurring, owing to the liquidity of mercury, at a still lower temperature. 
But gold, mercury, thallium, lead, and bismuth follow each other in the classification 
of elements based on increasing atomic weight, and are grouped on the same horizontal 
line of Mendeleef’s table. This coincidence is curious. Whatever it may signify, 
this transformation of the structure of a metal, at a temperature so far below its 
melting point, in the presence of only two-tenths per cent, of a foreign body, is 
probably not an isolated fact, and appears to open a new field for research. 
* We are not speaking of potassium, -which appears to be concentrated in a cement. 
