102 
MR. S. W. J. SMITH ON THE THERMOMAGNETIC ANALYSIS OF 
Suppose that these original nuclei are arranged in the form of an octahedral 
network (cf, e.g., the crystals of steel described by Tschernoff, ‘ Metallographist,’ 
vol. 2, 1899, p. 74) and that the material is of comparatively uniform composition, so 
that the octahedral lamellae of crystalline nuclei are at nearly uniform distances apart. 
These nuclei will contain a relatively small percentage of nickel (see Section V., § 2, 
p. 67, also fig. 23), and the remaining solid solution will contain a greater percentage 
of Ni than the meteorite as a whole. If the cooling continues extremely slowly, 
crystallisation will proceed around the nuclei from the metastable solid solution. The 
crystals (both the nuclei and the deposits around them), and also the solid solution, 
will gradually get richer in nickel (cf Section VI., § 3, p. 85). 
If the rate of cooling is sufficiently slow, the strength of the solid solution (in the 
spaces between the octahedrally arranged bands forming round the first growth of 
nuclei) may continue to increase, so that its temperature remains always above that 
at which it passes into the labile condition. Under these circumstances there will be 
no further spontaneous generation of nuclei (above the eutectic temperature) after 
that which occurs at 600° C. 
The “crystallisation” (i.e., regular structural re-arrangement of the material) will 
proceed outwards from each of the octahedral planes along which the spontaneous 
re-arrangement began. The layers of solid solution (material not yet structurally 
re-arranged) will become gradually narrower. At the same time the layers of 
“ crystallised ” material will become wider. Throughout the whole process there will 
be a continual re-adjustment of the percentages of iron and nickel in the crystals and 
in the solid solution. 
T1 ie final state of the material below the air temperature will be such that it 
contains layers of homogeneous nickel-poor alloy and thin intervening layers of 
eutectic. The actual width of each layer of nickel-poor alloy will depend upon the 
distance apart of the first lamellae formed when the material reached the temperature 
of lability ; but the ratio of the widths of the nickel-rich and nickel-poor layers 
(taenite and kamacite) will be nearly constant if the original distribution of lamellae 
is symmetrical. 
§ 14. A perceptible change in the process of crystallisation will occur if the rate of 
fall of temperature is more rapid than that above contemplated, so that, after the first 
deposition of nuclei during cooling, the solid solution does not always remain above 
its temperature of lability. Then the growth of crystals may be insensible between 
the time of deposition of one set of nuclei and the next. At each deposition crystals 
slightly richer in nickel than those last deposited will be formed, and a solution richer 
in nickel will be left. The latter will remain metastable while the temperature falls a, 
little ; then it will become labile and there will be a fresh deposition of nuclei. The 
crystallisation will thus proceed by spontaneous generation of successive sets of nuclei 
from a solid solution which gets continuously richer in nickel as the temperature falls. 
The strongest solution will .contain about 27 per cent, of nickel, and, becoming labile 
