METEORIC AND ARTIFICIAL NICKEL-IRON ALLOYS. 
103 
at about the air temperature, will begin to crystallise then as a mixture of nickel-rich 
and nickel-poor alloys. At any temperature above this there will he always some 
solid solution, which, after partial crystallisation in the condition of lability, leaves a 
stronger solution which is metastable until the temperature falls. 
During' cooling’ the solid solution will never become so rich in nickel that it contains 
more than 27 per cent., for any solid solution containing less than this percentage can 
begin to crystallise by becoming labile before the eutectic temperature (about 0° C.) is 
reached, and when the latter temperature is attained the solid solution remaining will 
crystallise as a mixture of the nickel-rich and nickel-poor alloys. 
In its final state the material will be roughly homogeneous, for practically the same 
process of crystallisation will have occurred in every small element of the mass. 
There will be mixed crystals containing every proportion of nickel from 2 or 3 per¬ 
cent. to about 6 - 5 per cent, (as well as the richer crystals of the second type formed 
at the eutectic temperature), for the cooling will have proceeded so rapidly that the 
continual redistribution of nickel necessary to produce uniformity in composition of 
the mixed crystals (of the nickel-poor type) will not have had time to occur. 
§ 15. On such a view of the process of crystallisation during cooling as is above 
indicated, ;t will be seen that a structure identical in form with that observed in the 
Widmanstatten figures and similar in composition to that shown by the magnetic 
properties is explicable. The fact that the breadth of the layers of kamacite some¬ 
times varies from one part of a meteorite to another would be explained by difference 
in the rate of cooling, or in the manner of deposition of nuclei in different parts of 
the material. The fact that the average thickness of the kamacite hands decreases 
as the total percentage of nickel in the material increases can be explained in the 
same way. Thus, for example, in a meteorite containing a relatively high percentage 
of nickel, e.g., 12 per cent., the whole material will exist “uncrystallised” until a 
temperature in the neighbourhood of 300° C. is reached. When this temperature— 
the temperature of lability of the metastable solid solution—is attained, crystallisa¬ 
tion will begin; but the rate of growth round the original nuclei will he so slow (on 
account of the lowness of temperature) that, unless the rate of cooling is excessively 
slow, there will be a succession of returns to the state of lability during cooling. If 
we assume that the successive growths of nuclei take place in parallel octahedral 
layers, the final result, after the eutectic temperature is passed, will he comparatively 
narrow bands of kamacite with intervening thin bands of taenite. 
When the percentage of nickel is higher than 15 or 20 per cent., the temperature 
at which the crystallisation begins will be so low that, as in the case of compara¬ 
tively rapid cooling considered above, the crystallised material may appear to he 
homogeneous (eg. Section V., §§ 8 and 9, p. 71). 
§ 16. It seems clear that if meteoric iron were kept for an indefinite period at a high 
temperature (below its melting-point), the structure which distinguishes it from 
artificial nickel iron, of the same percentage composition, at the air temperature 
