124 SCIENCE PROGRESS 
liquid. At 26° the cooling liquid passes from the unsaturated 
condition to a supersaturated one, in which crystals can 
grow, but require to be introduced from without; at 16° it 
passes into a condition in which crystals can be made to appear 
in the liquid by mechanical means such as shaking. Between 
these temperatures the solution is in what Ostwald has termed 
the ‘‘metastable” condition; below 16° it is in what he terms 
the “ labile” condition: and accordingly we distinguish the two 
showers, the first by the name “ metastable shower” and the 
second by the name “ labile shower.” 
In any laboratory the air must contain innumerable particles 
of every material that is being employed in the laboratory, and, 
so far as we know, immeasurably small germs are sufficient to 
inoculate a supersaturated solution and make it crystallise; in 
fact, the crystallisation of an exposed solution must be an 
extraordinarily sensitive test for the presence of such crystalline 
germs. 
If now the exact temperature at which the dense labile 
shower makes its appearance be determined for a series of 
solutions of different strengths, the observations will trace on 
the temperature-concentration diagram a new line, which we 
call the “supersolubility curve,” giving the relation between 
temperature of spontaneous crystallisation and strength of 
solution; this is the curve TT in fig. 1. 
We have determined this curve for various solutions such 
as sodium nitrate and sodium chlorate. We have also, by 
observations upon the refractive index, traced the changes of 
concentration in a cooling solution, and we find that it is not 
possible to obtain a clear representation of what actually occurs 
without taking account of the supersolubility curve. It is 
customary to think of a cooling solution as beginning to 
crystallise so soon as it becomes supersaturated, and as 
crystallising at such a rate that it remains just supersaturated 
and no more. If this were so, the changes in the temperature 
and constitution of the liquid would be represented by the 
line xaf. In reality, unless the cooling be very slow, the order 
of events is more faithfully represented by the line xabf; a few 
crystals appearing by inoculation at @ and continuing to grow 
very slowly, whereas a very rapid growth, which is generally 
accompanied by a cloud of crystals if the liquid be shaken, 
sets in at 6. Messrs. Hartley and Garrod Thomas have traced 
