KOSACEJE 411
in the dark for twelve hours. 200 c.c. of water are now added, and
the liquid filtered through a Gooch crucible, and the precipitate washed
with cold water, ,then with 10 c.c. of 10 per cent, alcohol, dried for three
hours in vacuo over sulphuric acid at 70° to 80°, and weighed. The
weight multiplied by 5'408 gives the number of grams of benzaldehyde
in 100 c.c. of the solution.
Kleber's method, based on the use of phenylhydrazine is also ac-
curate for the determination of the free benzaldehyde. Slightly modified
by Schimmel & Co. this method is as follows:—
To 2 c.c. of a solution of the oil (about 5 per cent, strength), 10 c.c~
of a fresh alcoholic solution of phenylhydrazine are added and allowed to
stand for an hour in a stoppered flask holding about 50 c.c. Twenty c.c
of deci-normal hydrochloric acid are then added and the flask gently
shaken. Ten c.c. of benzene are then added and the mixture is shaken
well and poured into a separator. The acid layer separates clear,
and is measured and filtered, and 20 c.c. of the filtrate is titrated with,
deci-normal potash, using di-ethyl-orange as indicator. From this the
titration figure for the full volume of the acid liquid is calculated. A
blank experiment is carried out without the almond oil, and each c.c. of
deci-normal potash in excess of that used in the estimation = 0'0106 gram
of benzaldehyde.
Dodge has proposed another method for the quantitative estimation,
of benzaldehyde. This is based upon Cannizzaro's reaction:—
2C 6 H 5 CHO + KOH = C 6 H 5 COOK + C 6 H 5 CH 2 OH.
He made tests with alcoholic potash solution, which showed that the
reaction does not proceed quantitatively when the solution is of less than
twice normal strength. Good results were obtained with a 25 normal
solution. Ten c.c. is allowed to stand for twenty-four hours at ordinary
temperature with from 1 to 2 grams of benzaldehyde, after which the
unattacked alkali is titrated back with semi-normal hydrochloric acid. A
blank experiment is carried out without the almond oil. The difference
shows how much of the alkali has entered into reaction, expressed in
c.c. of semi-normal potash. By multiplying the number of c.c. by the
molecular weight of benzaldehyde (106) and dividing by the quantity of
oil employed in the test, the aldehyde-content is obtained.
In the case of pure artificial products, Dodge found this figure to be
from 9903 to 100*3 per cent. He lays stress upon the fact that in
measuring the 2'5 normal potash solution it is necessary to proceed with
great care, because so small a difference as 0'05 c.c. represents an error
of not less than 1 to 2
- 5 per cent, benzaldehyde.
In the case of natural bitter almond oil the method gave no useful
results, presumably because the saponification of the benzaldehyde-
cyanohydrin into mandelic acid was incomplete.
It is apparent that in natural bitter almond oil all these methods only
estimate the free benzaldehyde, and not that which is present in com-
bination with hydrocyanic acid. This is shown by the following results
(see table on p. 412).
The Determination of Hydrocyanic Acid.—Hydrocyanic acid may be
approximately estimated by dissolving 1 gram of oil in 5 c.c. of alcohol,
and adding 50 c.c. of water. Then add ammonio-silver nitrate solution
and shake well. Acidify slightly with nitric acid, and collect, wash, and
dry the silver cyanide precipitated. Ignite and weigh the silver, four
parts of which correspond to practically one of hydrocyanic acid.