ON COLLOID CHEMISTRY AND I'TS INDUSTRIAL APPLICATIONS. 23 
cannot be produced directly. By a slightly different procedure 
Hatschek obtains strata of lead chromate in gelatin, also as pseudo- 
morphoses after silver chromate, although the reaction between lead 
nitrate and potassium chromate in gelatin leads only to a continuous 
band of lead chromate in the ordinary way. 
Microscopic observation of the stratifications, both while in 
course of formation and after completion of the reactions, have been 
made by Liesegang himself and by later investigators. Only ina 
few reactions—of which the original silver chromate one is the best 
example—are the ‘clear’ spaces between the rings practically free 
from the insoluble compound: in most cases the rings contain a 
- large number of small, and the clear spaces a small number of large, 
crystals or crystalline aggregates. A striking macroscopic illustration 
is afforded by cadmium sulphide in silicic acid gel, which exhibits 
no clear spaces at all, but a continuous succession of alternately 
yellow and pink bands. As is well known, the two shades are due 
to the difference in the size of the particlesand both may be obtained 
by precipitating aqueous solutions of different concentrations. It is 
an open question whether, in many cases, the actual amount of 
reaction product in equal volumes of ring and clear space is not 
approximately the same. 
As regards the theory of the phenomenon, the first explanation 
of the origin of periodic deposits was given by Wilhelm Ostwald. 
It is based on the assumption of ‘ metastable’ supersaturation, and is, 
in fact, the principal experimental evidence adduced by him for the 
existence of solutions in that condition. According to him, silver 
chromate is formed as the silver nitrate diffuses into the bichromate- 
gelatin, but at first remains in supersaturated solution until the limit 
of metastability is reacned. When this happens, the silver chromate 
is precipitated and ‘on the precipitate thus formed the silver 
chromate, in respect of which the vicinity of the ring is supersaturated 
(the italics are mine, E. H.) is deposited and reinforces it; this ° 
continues until the soluble chromate removed from the vicinity has 
gone into the precipitate.’ 
H. Bechhold, while generally accepting this theory, has shown 
that it is at least incomplete by demonstrating conclusively that the 
phenomenon is profoundly affected by factors which it fails to take 
into account, such as the solubility of the precipitate in the second 
reaction product, ¢.g., the solubility of silver chromate in ammonium 
nitrate, ammonium bichromate having been employed in his 
ex periments. 
R. BE. Liesegang also realised that the experimentum crucis would 
be the failure to produce a second ring system in a layer of gelatin 
already containing one, since the crystals forming the latter should 
obviously prevent supersaturation anywhere in their vicinity. He 
nevertheless succeeded in obtaining such a secondary system, but 
found that it was formed at a level below that of the primary 
stratifications. This result satisfied him that the experiment did not 
refute the supersaturation theory, a conclusion which it is difficult to 
accept. There is no obvious reason why a deposit of silver chromate 
which ex hypothesi (see the italicised passage in the quotation from 
Ostwald) prevents supersaturation radially over some distance 
