METEORIC AND ARTIFICIAL NICKEL-IRON ALLOYS. 
labile state at 0 2 . The immediate effect of the formation of the crystalline nuclei is, 
of course, to reduce the concentration of the salt left in solution. Hence, since the 
temperature of transition from the metastable to the labile condition decreases when 
the concentration is diminished, the remaining solution is metastable at the tempe¬ 
rature 0 2 , and further crystallisation can ensue only in the way already described as 
characteristic of the metastable state, viz., round the crystalline nuclei formed. It 
will be assumed for simplicity that any thermal effect accompanying the formation of 
crystals can he ignored. 
If the temperature is kept constant at 0 2 , there will be a very slow crystallisation 
of the salt in solution. The amount taking place will he very small if the time 
during which the temperature is maintained constant is short—even if, as before, the 
solution is kept stirred. If the temperature is maintained at 0 2 for a sufficient length 
of time, crystals will continue to grow round the original nuclei until the amount of 
salt left in solution is that required to produce a saturated solution at 0 2 . If the 
temperature is now assumed to fall slightly, the solution becomes supersaturated and 
metastable. Crystallisation will proceed slowly round the crystals already formed 
until the solution again becomes saturated, and will then cease. Hence extremely 
slow cooling from the temperature 0 2 at which the labile condition first appears will 
be such as to favour the growth of the largest and most symmetrical crystals it is 
possible to obtain. 
If, however, the rate of fall of temperature below 0 2 is rapid compared with that 
just contemplated, it is easy to see that the process of crystallisation will be different. 
The temperature will fall below 0 2 before the solution has deposited enough crystals 
to become saturated at that temperature. It is still supersaturated and, if the 
amount of crystallisation round the nuclei, since the advent of the labile condition at 
0 2 , has been small, it will pass almost immediately from the metastable to the labile 
state. A fresh deposition of crystalline nuclei will then ensue. This process will be 
repeated again and again as the temperature falls—the growth round each set of 
nuclei being a very slow process will not proceed far before, by reduction of tempe¬ 
rature, the solution has again reached the labile condition. Hence comparatively 
rapid cooling will cause the growth of a succession of sets of small crystals. 
An important deduction (of which the significance in the supposed analogous case 
of solid solution will be seen below) from the behaviour during comparatively rapid 
cooling, as set out above, is that at any temperature below 0 2 the amount of crystalli¬ 
sation will reach, practically at once, a value which will not increase appreciably until 
the temperature suffers a further fall. The rate of growth round the nuclei will be 
so small compared with the rate of formation of the nuclei themselves that the 
amount of crystallisation at a given temperature will appear to be approximately a 
single-valued function of the temperature. 
Again, for the purposes of the analogy which it is sought to develop between a 
crystallising solution and cooling nickel-steel, suppose that the comparatively rapid 
