ON THE STRUCTURE OF METALS. 
423 
1. First a general structure, visible under small magnification (twenty diameters 
for instance), and sometimes even to the naked eye. This structure is that of ingots 
of steel, and, probably, also of most metals which have been poured in ingot moulds. 
The mass subdivides itself into jointed groups, the intimate nature of which we will 
consider later, but the groups may be distinguished from each other either by a 
different general orientation of elementary crystallization, or by being more or less 
energetically attacked (see Plate 9, photos. 1, 2, 3, 4, 5, 6). It follows from this that 
the same group may appear either brilliant or dark, golden, brick-red, or deep purple, 
according to whether it reflects more or less light into the objective under a definite 
mode of illumination. The photographs 1 and 6, for instance, represent the same 
micro-section lighted obliquely, but from opposite directions. The general orientation 
is, however, subject in a given group to local changes (photo. 12), but the limits of 
the subdivisions are then ill defined. 
If we eliminate the complications introduced into our observations by the obliquity 
of the sections, and by the deformation of the test pieces caused by traction, it will be 
seen (photos. 1, 2, 3, 4, 5) that the groups are of two kinds, and resemble those which 
Chernoff found in ingots of steel. 
Those at the periphery are prismatic and at right angles to the surface of cooling ; 
while those in the interior are, roughly speaking, equi-axial, so that a transverse 
section of a quarter of an ingot may be diagrammatically 
represented by fig. 2. The genesis of this double structure has 
already been explained by one of us in a paper on the “ Cellular 
Theory as applied to Metals.” # 
The absolute dimensions of the groups, and the relative 
importance of prisms and equi-axial polyhedra, vary in ingots 
of steel with the rapidity of cooling, with the temperature of 
the metal at the moment of pouring, with the amount of 
occluded gas and with chemical composition. We have detected 
in our alloys of gold analogous variations. From this point of 
view, the thirteen specimens may be divided into five classes, of which each is repre¬ 
sented by a photograph in which the micro-section is enlarged to seventeen diameters 
and illuminated obliquely by Sorby’s parabolic mirror. 
First, the prisms are of considerable size and occupy the whole section, even to the 
centre (Au, Bi, Zr, Rh, Zn, Pd, T1; see photos. 1 and 6). 
Second, the prisms do not occupy all the mass, their mean size is smaller than that 
of the metals in the first category, and a central core is occupied by polyhedra, which 
are roughly equi-axial (Li, Se ; see photo. 2 of the alloy with lithium). 
Third, the prisms stop about midway between the periphery and the longitudinal 
axis of the section, while the central core becomes of considerable importance (In, Sb, 
see photo. 3 of alloy with indium). 
* ‘ Ann. des Mines,’ 8th Series, vol. 8, pp. 5-84. 
