964 
is 54 degrees above the transition temperature!) there does not 
occur any change of colour. 
3. If they are heated in a perfectly dry condition to 130°, the 
transformation into the yellow modification occurs in a few seconds. 
+. If traces of any solvent are added (water, ethyl alcohol, methyl 
alcohol, ether) the transformation velocity increases enormously. 
For instance: If a trace of water is added the transformation occurs 
immediately at 100°. 
5. If the yellow erystals are dried thoroughly they remain unchanged 
below 46° (e.g. at room temperature) even if they are in contact 
with the red ones. Having put into a dilatometer 150 grams of each 
modification, not the slightest transformation could be observed after 
some months at 15°. (The bere of the capillary tube was 0.8 m.m. 
while the transformation would be accompanied by a change of 
volume amounting to as much as 6 per cent). 
12. These facts prove that, both above and below the transition 
temperature, we find in this case the same retardations as have 
been found with the metals: consequently the compound, as we 
know it in a pure state, is a metastable system. 
With the metals this is the general rule; stabilisation occurs only 
on using special means. On the other hand, in the case of chemical 
compounds stabilisation takes places as it were spontaneously, because 
under those conditions which are met with in the laboratory and 
in daily life, the substance which accelerates enormously the rate of 
stabilisation, a solvent, is always present. Special precautions must 
be taken here to exclude its presence (even in traces). If these 
substances are not dried very thoroughly, traces of the solvent are 
included in the erystals and even these are sufficient to bring about 
the stabilisation with great velocity. (Compare § 5). 
The pure compound is thus to be regarded as a metastable system; 
that we often get the stable modification is to be attributed to the 
presence of impurities (traces of any solvent). 
13. We shall deal later with a large number of such cases, but 
one other example may be mentioned here. If KNO, is heated above 
its transition temperature (129°) it becomes thorougly dried in con- 
sequence of the high temperature. If the dry salt is now cooled 
quickly in dry surroundings, a range of temperatures is reached, 
where the rate of ‘stabilisation (transformation of 8-KNO, into the 
a-modification) is so small, that the two forms remain coexisting. 
As there does not exist here a difference in colour between the 
