24 REPORTS ON THE STATE OF SCIENCE.—1918. 
amounting to millimetres, should not do so equally at right angles to 
that direction, viz., into the depth of the layer of gelatin. 
E. Hatschek carried out experiments of a more direct character in 
the test tube. Crystalline lead iodide was suspended in agar contain- 
ing also potassium iodide. When lead nitrate solution was allowed, 
to diffuse into this gel, the usual, very perfect stratifications of lead 
iodide were formed, although crystalline nuclei were disseminated 
through the gel and should have made supersaturation impossible. 
R. E. Liesegang does not consider that this result is incompatible 
with the supersaturation theory and suggests that the ‘radius of 
action’ of the nuclei is too small to prevent the formation of strata 
which, with the particular reaction, occur at very small distances 
from each other. Without going-into any speculations regarding the 
mechanism of this action, it can be said a priori that {Liesegang’s 
objection, ifjjvalid, would be equally fatal to Ostwald’s theory 
itself. 
The reason suggested by Liesegang to prove his own experiment 
inconclusive is, finally, eliminated in a somewhat similar experiment 
by L..J. de Whalley (published by E. Hatschek). He obtained a 
second system of stratifications of lead chromate in agar, which 
already contained a very fine system of strata of the same compound, 
in a test tube. As the reacting solutions in this arrangement must 
inevitably pass through, and meet in the vicinity of, existing strata, 
the case is different from that of the flat layer used by Liesegang, and 
appears to leave no escape from the conclusion that supersaturation 
was prevented throughout the formation of the second system of 
strata. 
An alternative theory of periodic precipitation has been advanced 
_by S. C. Bradford and supported by some experimental evidence, the 
most striking of which consists of photographs of preparations in 
strongly coloured solutions, such as the alkali chromates. According 
to him, one of the reacting solutes is adsorbed by the layer of 
precipitate, the result being a zone practically free from it, so that 
the clear space between the strata is at once accounted for. Consider- 
ing the uncertain and conflicting results of adsorption experiments 
with solutions of electrolytes, it would be desirable to support what 
is undoubtedly an attractive suggestion by direct evidence—which 
Bradford so far has not obtained—that the solutes in question are 
actually adsorbed by the appropriate precipitates, e.g., that potassium 
chromate is really adsorbed by lead chromate, or potassium sulphides 
by lead sulphide. 
' Apart from the difficulties already set forth, any theory musi be 
pronounced inadequate which leaves—as do both Ostwald’s and 
Bradford’s—the gel out of account altogether. There is an abun- 
dance of material to show that the same reaction, if carried out in 
different gels, leads to entirely different results, in other words, that 
the gel has a specific effect and does not merely act as an indifferent 
medium which prevents mixing or currents. In this connection a 
suggestion made by H. Freundlich incidentally in a paper on 
another subject is of interest : that the formation of periodic strata 
may be an instance of the coagulation by electrolytes of a suspensoid 
sol, While at first sight the distinction between a sol of, say, silver 
