46 SECTIONAL ADDRESSES. 
the temperature falls, the iron changes into a modification which is stable 
at lower temperatures and loses its power of holding the carbon or carbide 
molecules (for the X-rays have so far failed to determine how the carbon 
atoms are grouped in the space lattice) in solution, so that separation 
occurs, and a-iron and cementite, Fe,C, crystallise from the mass, two 
solid phases now being present in place of one. The scale of the separation 
may vary greatly according to the time occupied by the process. No 
separation can occur without diffusion, and the transport of atoms or 
molecules through the solid mass takes an appreciable time, which is 
greater the lower the temperature, so that it is much less perfect when 
the steel is cooled rapidly than when ample time for diffusion is permitted. 
Consequently, the size of the molecular aggregates of cementite may vary 
from that of ultramicroscopic particles, so small and offering so large a 
surface to the action of chemical reagents that the mass is stained black 
or brown by acids, in which case the mixture is known as troostite, to 
the comparatively coarse, although still microscopic scale of the well- 
known laminated pearlite, in which the thin alternate sheets of ferrite and 
cementite, like the fine sheets in mother-of-pearl, can produce colours by 
the diffraction of light, whence the pearly appearance noticed by Sorby 
in the first exact scientific study of the microscopic structure of a metal. 
Now let the cooling be so rapid that a distinct separation into two 
phases, even on an ultramicroscopic scale, does not occur. The rearrange- 
ment of the iron atoms in their space lattice, in this instance from the face- 
centred cubic arrangement of the y-iron into the body-centred cubic 
arrangement of «-iron, still takes place, but the crystallisation of cementite 
as a separate phase is prevented. The result is that a new structure is 
obtained, known as martensite, in which the iron is, at least for the greater 
part, in the a-form, as is proved by its X-ray examination and by its 
magnetic properties, but in which the carbide is held, either in unstable 
solid solution in «-iron, in which it is normally insoluble, or as sheets of 
molecules parallel with the octahedral planes of the iron. Both views 
have their supporters, but I must profess a leaning towards the second. 
Whichever be correct, it is certain that this unstable condition is associated 
with great hardness and lack of plasticity, and it is necessarily present 
in fully hardened steels. Still more rapid cooling may suppress both the 
change in the lattice and the separation into phases, the solid solution 
which is stable at high temperatures being preserved during cooling, so 
that a part of the iron is still in the y-condition, and holds carbon atoms 
in a homogeneous fashion within its structure. As such a cooled solid 
solution is not hard, the steel is actually rendered less hard and brittle 
when the quenching is so severe than if it had been cooled somewhat less 
rapidly. The transformation of the iron, however, occurs with such ease 
that it is only when the proportion of carbon is rather large, or when 
some other metal is present, that this condition can be observed. 
It is the addition of foreign metals which has brought about the most 
remarkable changes in the properties of steels, out of which there has 
grown a new and important industry—that of the alloy steels. The 
presence of foreign elements in the original solid solution has a powerful 
influence on the rate of change in the system. As a general rule, the change 
from one lattice to another and the passage of a constituent, such as 
carbide, out of or into solution are greatly retarded by the presence of 
