430 
MR. F. OSMOND AND PROFESSOR ROBERTS-AUSTEN 
4. The strength of the joint, which is nil towards the end of the solidification, and 
remains insignificant for a longer time if the solidification is prolonged by the presence 
of impurity in the way indicated in fig. 3. The presence of a crystalline envelop 
covering the joints, which appears to he an unfavourable condition, as it is specially 
marked in the alloys of low tenacity, is probably also connected wfith the prolonga¬ 
tion or want of sharpness in solidification occasioned by the presence of certain 
foreign bodies. These envelopes without doubt represent the parts which have 
solidified last. 
But all the factors we have just reviewed co-operate directly in determining the 
condition of the joints, and indirectly the mechanical properties of the cooled metal. 
The co-efficient of expansion, the elastic limit and plasticity of the bodies of the 
grains and the delayed solidification) are directly dependent on the molecules and 
atoms. It is therefore not surprising that in a case which is relatively so simple as 
that of gold the influence of the atomic volume of the added elements can be directly 
observed. Furthermore, the relation between the mechanical properties of the metals 
and the atomic volume of the dissolved impurities is an experimental fact. We can 
now explain the greater part of the apparent exceptions. The exceptional behaviour 
of indium is explained by the presence of a cement, and that of lithium and of 
aluminium by the development of secondary crystallization.* But as the conversion 
of heat into work is effected in the steam engine by a series of intermediate stages, 
so here also a more or less complex mechanism intervenes between the extreme terms 
which are in relation to each other. The foreo-oiim research is a small contribution 
n o 
to the study of one of the elements of this mechanism. 
Part II. — On some Phenomena of Annealing. 
Etching by sulphuric acid at a temperature of between 200° to 250° revealed a 
network of joints on the gold alloyed with bismuth and thallium, which the previous 
attack with aqua regia had not rendered apparent. All the specimens (except the 
alloy with indium which had not been subjected to the action of sulphuric acid) were 
then repolished and again etched with aqua regia under the same conditions as before. 
The pure gold and the alloys with K, Pd, Zn, Rh, Li, Se, and Zr did not undergo 
any change, but the appearance of the four alloys with Bi, Tl, Sb, and A1 were more 
or less altered by the treatment. 
The bismuth alloy, which was initially adequately represented by the zirconium- 
gold, Photograph No. 6, for a magnification of 17 diameters, and by the thallium- 
gold, Photograph No. 15, which shows details (magnified 150 diameters), is now 
represented by Photographs No. 7 (17 diameters) and No. 20 (150 diameters). The 
large grains of the melted metal become sub-divided, after heating for five minutes 
to between 200° to 250°, into a number of small polyhedral grains. The effect is just 
* Zinconium still presents a very striking exception. 
