342 
Journal of Agricultural Research 
Vol. XXVI, No. 8 
The peroxid tests were all made without addition of acid, under which 
conditions the hydrogen peroxid employed in this experiment reacted 
negatively. The results of these tests were as follows: 
Table II .—Results of tests of solutions of acrolein and hydrogen peroxid 
No. 
Kreis test. 
Peroxid test. 
1. 
Negative. 
Negative. 
Do. 
Strongly positive. 
Do. 
Do. 
Do. 
All negative. 
2 ... 
.do. 
2. 
Strongly positive. 
4.. .... 
.do. 
c. 
.do. 
6. 
.do. 
ia to 6a. 
All negative. 
The hydrochloric acid series was not tested further,* but at the expira¬ 
tion of 22 hours the water series was tested again, with the same results 
as before. Addition of hydrogen peroxid to solution 2, until the solution 
contained a total of three and one-half molecules of hydrogen peroxid 
per molecule of acrolein, failed to elicit a positive response to the Kreis 
test. 
The results of this experiment were quite surprising. It was to have 
been expected that substance K would still be formed, though in di¬ 
minished quantity, in solutions containing less than a stoichiometric 
quantity of hydrogen peroxid. But as a repetition of these tests (with 
the same material, to be sure) led to similar results, it was concluded 
(erroneously, as will appear later), that substance K is formed in an ir¬ 
reversible reaction from two molecules of hydrogen peroxid and one of 
acrolein, that it contains peroxid oxygen, and that it must therefore be 
acrolein peroxid. It appeared, furthermore, that the concentrated 
hydrochloric acid favored the destruction of substance K, and probably 
functioned, not in the formation of substance K, but merely in the con¬ 
densation reaction with phloroglucin. 
On the basis of this experiment many unsuccessful attempts were 
made to isolate substance K from aqueous solutions of acrolein and hydro¬ 
gen peroxid. Evaporation of such solutions in a vacuum desiccator led 
to small, colorless, sirupy residues, possessing no characteristic odor, and 
reacting but weakly with the Kreis reagent. Distillation in vacuo led 
to similar residues, though positive Kreis tests could be obtained on the 
aqueous distillates. The tests on the distillates, however, were always 
weaker than those obtained on the original solutions, and could also be 
variously interpreted as indicating the volatility of substance K as 
such, or as pointing to a gradual decomposition of substance K into 
acrolein and hydrogen peroxid and the separate distillation of these com¬ 
pounds. Finally, an attempt was made to prepare acrolein peroxid by 
the method employed by Pastureau and Launay (29) in the preparation 
of the peroxid of mesityl oxid, that is, by subjecting acrolein, instead of 
mesityl oxid, to the action of 3 per cent hydrogen peroxid in dilute 
sulphuric acid solution. An abundant, white, flocculent precipitate was 
obtained, which failed to dissolve in the usual solvents, and was ap¬ 
parently a polymerization product of acrolein. 
In the course of this work, a number of observations were made that 
could not be reconciled with the previous finding that substance K is not 
