ON THE CRYSTALLINE STRUCTURE OF METALS. 
295 
tures. On the other hand, it has been very widely believed that annealing or 
re-crystallisation, particularly in iron and steel, are “ critical ” phenomena which can 
only occur at or above certain definite temperatures. Arnold has gone so far as to 
make a careful determination of such an “ annealing point.” Various of the “ arrest- 
points ” in the cooling of iron and steel have also been regarded as representing 
critical points in respect to annealing, but the connection between the two is by no 
means fully proved. Having found phenomena of annealing or re-crystallisation in 
other metals, it therefore became interesting to inquire whether any corresponding 
arrest-points could be found in the cooling of these metals. We investigated the 
matter by means of a pyrometric arrangement consisting of two thermo-electric 
junctions, a very sensitive D’Arson val galvanometer, and a potentiometer somewhat 
similar to that used by Sir W. Roberts-Austen ; the deflections of the galvanometer 
were, however, observed by means of a telescope and scale, instead of being photo¬ 
graphically recorded. It may be, therefore, that either from this cause, or from 
insufficient sensitiveness of the whole arrangement, some minute arrest-points were 
overlooked; hut between the melting-points and the ordinary temperature of the air 
no trace of an arrest-point was observed in the three metals tried, i.e., lead, tin and 
cadmium. Such arrest-points, if they exist at all, may be found at much lower 
temperatures than those to which our experiments were carried. 
Having thus failed to correlate the phenomena of re-crystallisation in lead, &c., 
with any definite point at which heat is evolved during the cooling of the metal, 
and having reason to believe that even in iron the arrest-points are not necessarily 
intimately connected with annealing, we look for a theoretical explanation of these 
actions in another direction. The theory of re-crystallisation which we shall now 
advance as a working hypothesis for the explanation of the phenomena describe^ 
in this paper ascribes an important part in the action to the impurities present even 
in “ pure ” metals. The impurities which we believe to be of importance are those 
which are capable of forming eutectic alloys, or fusible compounds, with the metal 
itself; they would therefore be mainly metals, particularly the more fusible metals, 
such as bismuth, tin, cadmium, mercury, sodium, or even rarer metals, such as 
gallium. It is well known that when a metal containing a small proportion of such 
impurities crystallises, the impurities are, for the most part, segregated in the inter¬ 
crystalline boundaries. The crystals themselves form at a temperature when the 
eutectic alloys are still quite fluid, and the growing crystals gradually push the 
remaining eutectic into the boundaries. Where the quantity of infpurities present is 
sufficiently great, this eutectic can be seen under the microscope forming an inter¬ 
crystalline cement. Mild steel, where the “ pearlite ” plays the part of a eutectic, 
is a good example of such a structure ; other examples can be found in the gold- 
aluminium alloys illustrated by Messrs. Heycock and Neville. # Where the 
quantity of eutectics present is very small, the meshes of inter-crystalline cement 
* “Gold-Aluminium Alloys,” Heycock and Neville, ‘Phil. Trans.,’ A, vol. 194, plates 4, 5. 
