346 
Journal of Agricultural Research 
Vol. XXVI, No. ft 
acid was employed, no color was obtained. Evidently, the acid serves 
first of all to hydrolyze the acetal and liberate free epihydrin aldehyde, 
which then condenses with phloroglucin even in weak acid solution, in 
the same manner as does the product of the reaction between acrolein 
and hydrogen peroxid. 
The colored phloroglucids obtained from epihydrin aldehyde diethyl- 
acetal on one hand, and from mixtures of acrolein and hydrogen peroxid 
on the other, were compared spectroscopically in concentrated hydro¬ 
chloric acid solution. The spectra of the two solutions appeared to be 
identical, each spectrum exhibiting a well-defined band centered at 6.0 
on the arbitrary scale of the instrument. 
While the foregoing experiments point strongly toward the conclusion 
that substance K is identical with epihydrin aldehyde, the evidence is 
not yet final. There is still the possibility that the spectroscope em¬ 
ployed may not have been delicate enough to show minor differences 
between two nearly identical spectra. Again, it would be conceivable 
that epihydrin aldehyde might undergo some transformation in presence 
of hydrochloric acid, and that the colored phloroglucid is derived from 
some secondary decomposition product thus formed rather than from 
the epihydrin aldehyde itself. 
As to the first of these possibilities it may be said that the results of 
the spectroscopic examination were completely confirmed by use of a 
very delicate spectrophotometer, the results obtained thereby being 
described in another section of this paper. The identity of the colored 
phloroglucids obtained from substance K and from epihydrin aldehyde 
is therefore beyond question. 
The second of the possibilities suggested above, viz, that the epihydrin 
aldehyde, under the conditions of the Kreis test, might experience some 
indeterminate change prior to its condensation with phloroglucin, seemed 
to be rather remote, since such carbonyl compounds as might conceivably 
result from the decomposition or rearrangement of epihydrin aldehyde 
have already been eliminated as factors in the production of the red 
phloroglucid which we have described. The desirability of obtaining 
analytical evidence regarding the composition of this compound is 
nevertheless apparent; and a small amount of the phloroglucid was 
prepared from acrolein and hydrogen peroxid, and submitted to analysis 
by combustion. 
Twenty-five cc. of a 5 per cent aqueous solution of freshly pre¬ 
pared acrolein were treated with an excess of pure 30 per cent hydrogen 
peroxid and about 300 cc. of concentrated hydrochloric acid. Three 
hundred cubic centimeters of a 1 per cent ethereal solution of phlo¬ 
roglucin were added immediately, and the mixture was shaken in a 
separatory funnel. An intense red color was immediately imparted to 
the hydrochloric acid layer, followed rapidly by the separation of a deep 
purple precipitate. The hydrochloric acid layer containing the preci¬ 
pitate was separated from the ethereal phase, diluted with about 500 cc. 
of water, separated from the ether that was thrown out of solution, and 
filtered. The precipitate, which had been collected on a Buchner funnel, 
was washed thoroughly with water, air dried, and tested for solubility 
with a view to its recrystallization. It proved to be insoluble in alcohol, 
ether, benzene, chloroform, acetone, carbon tetrachlorid, carbon bisul- 
phid, isoamyl alcohol, butyl alcohol, methyl alcohol* petroleum ether, and 
pyridin; very slightly soluble in glacial acetic acid and in concentrated 
