Cremer Nar} Separation of Hydrocarbon Oils from Fat Oils. 161
for the most part of very limited applicability, and in some
1 I cases wholly untrustworthy.
T H E C H E M I C A L N E Ww S * When the hydrocarbon oil in admixture happens to be
of comparatively low boiling-point, it may often be driven
off by exposing the sample to a temperature of about
150° C., but the estimation thus effeGed is generally too
low, and often quite untrustworthy,
When it is merely desired to estimate approximately
the proportion of hydrocarbon oil present, and not to
isolate it or examine its exa& character, Keettstorfer’s
titration-process may be used, as suggested by Messrs,
Stoddart. But the best and most accurate method of
detecting hydrocarbon oils in, and quantitatively separating
‘hem from, fat oils, is to saponify the sample, and then
agitate the aqueous solution of the soap with ether.* On
separating the ethereal layer and evaporating it at or
below a steam heat the hydrocarbon oil is recovered in a
state of purity.
Either caustic potash or soda may be employed for the
saponification, but the former alkali is preferable, owing
to its greater solubility in alcohol andthe more fusible
character of the soaps formed. A convenient proportion
to work with consists of 5 grms. of the sample of oil, and
25 c.c. of a solution of caustic potash in methylated spirit,
containing about 8o grms. of KHO per litre, Complete
saponification may usually be effected by boiling down the
mixture in a porcelain dish, with frequent stirring, until
it froths strongly. In the case of butter, cod-liver oil,
and other fats which undergo saponification with difficulty,
it is preferable to precede this treatment by digestion of
the mixture for half an hour at 100° C. in a closed bottle,
After evaporating off the alcohol, the soap is dissolved in
water, brought to a volume of 70 to 8o c.c,, and agitated
with ether. The ethereal solution is separated, washed
with a little water and carefully evaporated, The agitation
 with ether must be repeated several times to effet a
somplete extraction of the hydrocarbon oil from the soap
solution.
The loregoing process has been proved to be accurate
on numerous mixtures of fat vils with hydrocarbon oils,
The results obtained are correc to within about 1 percent
in all ordinary cases. In cases where extreme accuracy
's desired, it is necessary to remember that most, if not
all, animal and vegetable oils contain traces of matter
wholly unaéed on by alkalies, In certain cases, as butter
and cod-liver oil, this consists largely of cholesterin,
Cz6Hy40.t The proportion of unsaponifiable matter
soluble in ether which is naturally present in fixed oils
and fats rarely exceeds 1} per cent, and is usually much
less. Sperm oil, however, constitutes an exception,
yielding by the process about 40 per cent of matter
soluble in ether.} This peculiarity has no practical effe@
on the applicability of the process, as sperm oil, being the
most valuable of commercial fixed oils, is never present
without due acknowledgment of the fa®. Spermaceti
and the other waxes yield, after saponification, large percentages
 of matter to ether, and hence the process is not
wailable for the determination of paraffin wax in admixture
 with these bodies, though it gives accurate results
with the mixtures of paraffin and stearic acid so largely
employed for making candles,
The following figures obtained in my laboratory by the
analysis of substances of known purity and of mixtures of
known composition, show the accuracy of which the pro-ON

 THE
SEPARATION OF HYDROCARBON OILS FROM
FAT OILS.»
By ALFRED H. ALLEN, F.C.s., F.LC.

THe extensive production of various hydrocarbon oils
suitable for lubricating purposes, together with their low
price, has resulted in their being largely employed for the
adulteration of animal and vegetable “oils. The hydro
carbons most commonly employed for such purposes are: —
1. Dils produced by the distillation of petroleum and
bituminous shale, having a density usually ranging
between 0'870 and o'g15.
2. Oils produced by the distillation of commor rosin,
having a density of 0'g65 and upwards,
3. Neutral coal oil: being the portion of the produés of
the distillation of coal-tar boiling about 200° C., and
freed from phenols by treatment with soda.
Solid paraffin, used for the adulteration of bees’-wax
and spermaceti, and employed in admixture with
stearic acid for making candles.
The methods for the dete®ion of hydrocarbon oils in
fat oils are based on the density of the sample; the lowered
flashing and boiling-points; the fluorescent characters of
the oils of the first two classes; and the incomplete
saponification of the oil by alkalies. The faste of the oil,
and its odour on heating are also useful indication..
If undoubtedly fluorescent, an oil certainly contains a
mixture of some hydrocarbon, but the converse is not
strictly true, as the fluorescence of some varieties - of
mineral oil can be destroyed by chemical treatment, and
in other cases fluorescence is wholly wanting. Still, by
far the greater number of hydrocarbon oils employed for
lubricating purposes are strongly fluorescent, and the
remainder usually become so on treatment with an equal
measure of strong sulphuric acid.
If strongly marked, the fluorescence of a hydrocarbon
oil may be observed in presence of a very large proportion
of fixed oil, but if any doubt exist the hydrocarbon oil
may be isolated. As a rule, the fluorescence may be seen
by Lolding a test-tube filled with the oil in a vertical posi.
tion in front of a window, when a bluish ¢ bloom" will be
perceived on looking at the sides of the test-tube from
above. A better method is to lay a glass rod, previously
dipped in the oil, down on a table in front of a window,
so that the oily end of the rod shall proje@ over the edge
and be seen against the dark back-ground of the floor,
Another excellent plan is to make a thick streak of the oil
on a piece of black-marble, or glass smoked at the back,
and to place the streaked surface in a horizontal point in
front of and at right angles to a well-lighted window. t
Examined in this manner, a very slight fluorescence is
readily perceptible. If at all turbid the oil should be
filiered before applying the test, as the refleGtion of light
from minute particles is apt to be mistaken for true
fluorescence. In some cases, it is desirable to dilute the
oil with ether, and examine the resultant liquid for
fluorescence, An exceedingly small amount of mineral
oil suffices to impart a strong blue fluorescence to ether.
The quantitative analysis of mixtures of fat oils with
hydrocarbon oils has till recently been very uncertain, the
published methods professing fo solve the problem being
* A Paper read before the Chemical Section of the British
Association, York Meeting, 1881.
t Either of these plans is infinitely superior to the polished tin-Plate
 usually recommended. In short. the back-ground should be
black, not white.

i,

* According to my experience, treatment of the dry soap with ether,
petroleum spirit, or other solvent, is liable to cause error from soly-:ion
 of the soap itself, if much hydrocarbon oil be present,
+ The process affords a very rapid and simple means of isolating
cholesterin. Thus, on dissolving the traces of unsaponifiable matter
left by butter in a little hot alcohol, and allowing the liquid to cool,
abundant crystals are deposited, which may be identified as cholesterin
 by their microscopic and chemical charaters, A sample of
Jutterine gave no cholesterin.
{ 1am investigating this interesting fa@, and have obtained ful]
confirmation of Chevreul’s observation that sperm oil when saponified
yields a peculiar solid aleohol instead of glycerin, Itis distin& from
cetyl alcohol, and distils, apparently without decomposition, at a very
high temperature.