s 
BULLETIN 399, U. S. DEPARTMENT OE AGRICULTURE. 
on the contrary, more completely emulsified it. Treatment with 
kieselguhr (infusorial earth) and subsequent filtering under differ- 
ence in pressure yielded a clear oil. but the process was entirely too 
slow in operation. After several experiments the following method 
was found to work very satisfactorily. 
All the water possible was removed with a separatory funnel, and 
the emulsion warmed on a water bath to not over 90° C. To this was 
added one-tenth its volume of a hot solution containing 2 per cent of 
gelatin and the whole thoroughly mixed. To the warm mixture was 
added one-fourth its volume of a hot solution containing approxi- 
mately 10 per cent of tannin. The emulsifying agent was thus brought 
into a state of coagulation, and the oil was released. The mass was 
Fig. 5.— Experimental still used in extracting orange oil: a, Still: b, condenser : c. receiver; d,separatory 
funnel. 
now thrown into a sack made of heavy Canton flannel and again 
pressed. The oil and water thus pressed out were separated by means 
of the separatory funnel, and the oil dried with quicklime and filtered. 
The filtered oil constitutes the marketable product. 
When the method was thus proved to be practicable, a cheap and 
abundant supply of tannin was sought. This was found in the 
rhizome of the common saw palmetto, which contains approx- 
imately 7.58 per cent of tannin. 1 A portion of this rhizome was 
boiled with water in a copper vessel for several hours and the liquor 
then strained off. Ten pounds of the fresh rhizome was found to 
yield a supply of tannin solution sufficient to treat the emulsion 
from at least a ton of orange 
1 Trimble. H. The tannins of the palmetto-. In Gard. and Forest, v. •. no. 42>. p. tsa- 
