ON DYNAMIC ISOMERISM. 139 
stood, as may be shown by referring to some of the examples that have 
been most fully investigated. It is, for instance, generally recognised 
that reversible isomeric change may occur in such a way as to involve 
the interconversion of a coloured and a colourless isomeride.' Usually 
both will be present in solution but, on crystallising out, one form only 
will separate in an approximately pure condition; in some instances, 
however, the two forms are capable of crystallising with such equal 
readiness that it is possible to separate out the coloured or the colourless 
form at will by varying the solvent? or the temperature of crystallisa- 
tion * ; even when this cannot be done it is often found that rapid crystal- 
lisation causes the separation of a mixture of the two isomerides. 
It is then only necessary to assume that the two forms of the substance 
are endowed with the property of forming isomorphous mixtures or 
solid solutions, in order to realise the conditions for the production of 
stained crystals of absolutely constant colour-intensity, since it will 
follow (in accordance with a general law) that the constancy of the 
equilibrium-proportions in the liquid state will be reproduced in the 
crystals, although the actual ratio of the two components need not be 
the same in the two phases. Reversible polymeric changes involving 
the inter-conversion of coloured and colourless compounds are familiar 
in the cases of nitrogen peroxide and the colourless (bimolecular) and 
blue (unimolecular) forms of ter-nitrosobutane*; here again it would 
probably only be necessary for the two forms to be isomorphous, in 
order to give rise to stained crystals of uniform composition and constant 
intensity of colour.® 
The importance of these examples consists in the fact that they pro- 
vide forthe production of stained crystals from which the colour could 
only be removed by the discovery of methods even more refined than 
those which are required on the one hand to arrest isomeric change in 
solution and on the other to effect the separation of isomorphous mixture. 
C. Crystallisation in Relation to Lwminous Phenomena. 
Since chemical change is usually-checked, if not actually arrested, 
on passing from the gaseous or liquid to the solid state, it is clear that 
this should exert a most important influence on any luminous properties 
which depend for their development on the chemical rather than on the 
physical activity of the molecule. This is clearly seen in the case of 
triboluminescence, phosphorescence, and fluorescence, whilst colour may 
be quoted as an illustration of an optical phenomenon which is not 
affected in any marked degree by change of state. 
1 This point is in itself sufficient to dispose of the idea that the development of 
colour is due to oscillatory isomeric change, since if this were the case the separate 
isomers must necessarily be colourless and only the mixture coloured. 
2 #4g., isonitrosomalonanilide, Whiteley, Zrans., 1903, 83, 34. 
% Hg., p-methoxyphenylphthalimide, Piutti and Abati, Ber., 1903, 36, 1000. 
* Bamberger and Seligman, Ber., 1903, 36, 689. 
5 Equilibrium between colourless and coloured forms has also been postulated as 
attending certain cases of ionisation (e.g., violuric acid has been supposed to acquire 
its characteristic colour only on passing into the ionised state), but the development 
_of colour is usually due to a change of structure of the kind that has already been 
considered and in any case the extension of the theory of ionisation to the staining 
_ of a colourless crystalline compound by its own ions would be too bold a conception 
to merit serious consideration. 
