DUBOIS A N D LOTT ON COCOA S H E L L S I N COCOA POWDER. the strength of the hydrochloric acid solution in which the titration is made is of much importance. The directions, therefore, as given above should be closely followed in regard t o the strength and amounts of hydrochloric acid used. The following results were obtained with a number of alloys: No.
Weight taken.
1A
0.2000 0,2000 0.2000 0.2005 1 .oooo 1 .oooo 0.5000 0,5000 1 .0000 0.5000 0.5000 1 .oooo
2-4
1B 2B
1c 2c 1D 2D
1E 1F 2F 1G
CC.
Per cent. Per cent. of Sb by of Sb. thiosulphate meth
KIOa used. 6.70 6.70 6 .OO 6.03 2.70 2.70 5.90 5.90 2.40 2.85 2.85 2.80
13.40 13.40 12.00 12.03 1 .08 1 .08 4.72 4.72 0.96 2.28 2.28 1.12
13.43
...
12.07
... ...
1.04 5.11'
... 0.96 2.32
...
1.16
The first two alloys are antimonial leads and the others are commercial solders, some of which were of poor quality on account of the antimony present. The time required for a n analysis was about a n hour. Arsenic is rarely present in appreciable quantities in the alloys under consideration, but it is to be noticed that, if present, it would be titrated with the antimony. The following results were obtained b y dissolving metallic arsenic in concentrated sulphuric acid and proceeding exactly according to the method that has been given. As taken. Gram.
KI08 used. cc.
As found.
0.0100 0.0100 0,1019
4.10 4.10 40.60
0.0102 0.0102 0.1015
I n a case where an alloy contains an appreciable amount of arsenic it is best t o carry out the process exactly as directed as far as filtering off the lead sulphate and washing it with I : I hydrochloric acid. Then pass in hydrogen sulphide t o precipitate the arsenic, pass air through the liquid for half an hour or so t o remove the hydrogen sulphide and t o oxidize any iron that may be present, filter, wash with I : I hydrochloric acid, and titrate as usual. This process was tested b y the use of alloys mixed with known quantities of pure metallic arsenic with the following results: Alloy taken. Gram. 0.5000 0,9560 0.2000 0.2000 1 .oooo
As taken. Gram.
Per cent. Sb found.
0.0050 0,0118 0.0460 0.0520 0.0097
4.72 4.77 12 .oo 12.10 1.04
Per cent. Sb in alloy. 4.72 4.72 12.00 12 .oo 1 .08
SHEFPIELD LABORATORY, NEW HAVEN. CONN.
-----
25'
depends upon the fact that the heavier constituents of cocoa shells sink in a solution of calcium chloride of specific gravity 1.535,whereas the lighter constituents of the cocoa float in the same. The parts sinking are dried and weighed. Goske determined the amount of heavy material in a number of varieties of cocoa shells of known purity and found this value to vary from 15.4t o 38.76 per cent. on the fatfree shell. The highest figure is taken as the basis of calculation when applying the method to cocoa. The weight of the heavy shell constituents, calculated t o fat-free cocoa, is divided by the factor 38.7 and multiplied by 100, t o give the percentage of shells in the cocoa. I n studying this method, the writers determined factors on a number of varieties of cocoa shells, after extracting the fat. I n the table below is given the results obtained: No. 1 2 4 5 6
7 8 9 10 11 12 13
Percentage heavy constituents, sp. gr. over 1.535.
Variety of cocoa shells. Maracaibo Guayaquil Caracas Bahia African Sancher Surinan Java Amba Venezuela Trinidad Pto Cabello
22.72 21.26 32.32 39.68 31.44 36,66 43.52 42.7 39.46 33.62 52.2 48.12
The average of the figures above is 36.96. The factor selected, therefore, for the determination on cocoa powders was 37. I t will be noted that shell No. 2 showed the lowest percentage of heavy constituents, No. 1 2 the highest, and that No. 7 was the nearest to the average. These three shells were added in known quantities t o a cocoa powder on which the fat and shell contentthe latter by Goske's method-were first determined. With these mixtures Goske's method was followed identically as laid down, with the results shown in the table below, using factor 37 : Sample. 2x 2u
zc 7-4 7B
7c 12A 12B 12c
Percentage shells added. 4.55 10.
12. 4. 9. 14. 2. 6. 10.
Total shells found.
Percentage added shells found.
10.56 15.6 15.7 13.0 12.5
19.1 9.9 16.3 21.5
33.12 8.2 8.3 5.6 5.1 11.7 2.5 8.9 14.1
Total shells found on the cocoa used for this experiment was 7.44 per cent. This value was deducted from total shdls found in the above mixtures to give the value for added shells expressed in the last column.
DETERMINATION OF COCOA SHELLS IN COCOA POWDER.
DISCUSSION O F RESULTS.
B Y W. L. DUBOISAND C. I . L o n .
I t will be seen a t a glance that the results are neither uniform nor accurate. This could well be expected, however, when the great variation in heavy constituents in the various grades of shells is considered. Goske examined seven samples and obtained the range of results cited above. The writers secured values which were uniformly higher than those shown
Received January 28, 1911.
The work herein described was undertaken t o test a method proposed by Goske.9 This procedure Solder D contained a little copper or iron which caused the thiosulphate method to give high results. Also alloys A and B were found t o contain traces of copper. 2. .Vahr. Genussm, 19, 154;C. A , . 4, 1328 (1910).
T H E J O U R N A L OF I X D Y S T R I A L Ah’D EXGIAEERIh’G C H E N I S T R Y .
by Goske, the average of the same being practically what he obtained as the highest figure. Goske selects his highest value as the factor for calculation. I n using this factor, results wrould be too low on cocoas t o which shells having a lower percentage of heavy constituents were added, and would be too high for those samples to which cocoa shells having a higher percentage were added. The method, it appears to us, wbuld be only approximately accurate where shells having practically that factor had been used. . There is no way, however, of determining this point in the examination of any cocoa submitted to the analyst. I n determining the accuracy of the method it 9ppears more logical t o select the average factor rather than the highest factor, although in the examination of commercial samples it is probably better to take the highest value obtained, because the benefit of the doubt is thereby given fo the sample in each case. The table above shows lack of uniformity of results. Samples 12A, I Z B and I Z C contained shells, the factor of which was 5 2 . 2 0 , much higher than the average selected for the calculation. I n this case it would naturally be expected that the results would be too high. This, it will be seen, is the case. Samples 2A, zB and zC contained shells, the factor of which was 2 1 . 2 6 , and in this case results would be expected t o be too low. This expectation is realized. Samples 7A, 7B and 7C, however, contained shells, the factor of which was 36.66, practically the figure used in the calculations. Results are, in the three samples respectively, 1.6 per cent. high, 4.9 per cent. low, 2 . 3 per cent. low. If the highest shell value j 2 . 2 0 had been used in the formula instead of 37, i t will be seen that in the case of mixture 2A the amount of added shells indicated would have been 0 , when in fact 4.5 per cent. of shells had actually been added t o the cocoa. I t would seem, therefore, that this method does not reliably indicate the addition of smaller quantities than 5 per cent. of cocoa shells and cannot be said to afford any accurate idea of the amount of husks actually placed in the commercial product. Filsinger and Botticher found that the method of A. Goske gives low results.1 L-. s. FOOD& D R U G I N S P E C T I O N L A B O R A T O R Y , BUFFALO,
N. Y .
T H E ESTIMATION OF ESSENTIAL OILS. BY CHARLESD. HOWARD
Received February 10. 1911.
During 1908 the writer published a methods for the determination of essential oils in extracts s n d pharmaceutical preparations, involving precipitation and extraction in a Babcock milk bottle, the small quantities of chloroform and ether used being volatilized b y rapid evaporation in a water bath. This method was applied t o a variety of extracts, including, incidentally, benzaldehyde, although there was no intent to claim t h a t it afforded results of any value in the latter in-
’Z
bffentl C h m , 16, 3 1 1 2Jour. A m Chem S o c , 30, 608 (1908)
April, 1911
stance.’ The method was subsequently criticized b y Hortvet and West,z the authors claiming t h a t the procedure did not serve t o eliminate all of the chloroform and that extraction was not complete. As a result of further experience, it is admitted t h a t this method sometimes affords erratic results in the case of certain oils and t h a t the average worker is liable to encounter some difficulty in securing concordant and accurate figures. I t h?s proved of value, however, in the case of such oils as lemon, orange, wintergreen and peppermint. A method of this character has the advantages of being simple and quickly carried out, requires no large quantity of solvent, and avoids the difficulties in connection with drying and weighing the oil-the latter, in our experience, proving not inconsiderable. Based upon the conviction of the writer that a much smaller quantity of solvent than that prescribed by Hortvet and West can be made to serve for complete extraction, the following modification was devised and has been used in this laboratory during the past year with good results, having been applied to most varieties of essential oil preparations. The procedure involves application of the principle, reierred to b y the writer in his original paper and since confirmed by Hortvet and West,s that when an ethereal solution of an essential oil is rncidly evaporated, no appreciable loss of oil occurs. Procedure.-Transfer 20 cc. of the extract to a fourounce separatory funnel; in the case of preparations containing more than five per cent of oil, take but I O cc. Add 50 cc. of water and (except in the case of oils of the type of cinnamon and clove) two drops of strong hydrochloric acid. Shake out with three portions of ether, using 1 5 cc., I O cc. and 5 cc. Aftereach extraction except the last, the ether solution may be run out into a small flask, which is kept stoppered if a series of determinations is being run simultaneously. The combined ether extracts are washed once with I O cc. of ether-saturated water for removal of the bulk of the alcohol, then cautiously transferred t o a I O per cent. milk bottle, rinsing the flask and tip of the funnel with an additional two or three cc. of ether. Attach a bulb tube t o the stem of the bottle and connect with a filter-pump, immerse the bottle in nearly boiling water, start the pump and shake with a gentle rotary motion a t first. When all danger of spirting has passed, shake violently and toward the last immerse in boiling water for a few seconds, or until the application of a match flame demonstrates the complete elimination of the ether. The removal of most of the latter should require not more than two or three minutes. Finally add cold water and centrifuge. I n the case of oils heavier than water, salt solution must be used as the floating agent, except with wintergreen, for which cold sulphuric acid ( I : 2 ) may be safely and most conveniently used. STATE
LABORATORY OF H Y G I E N E , CONCORD, F. H.
Loc. c d . . pp. 608, 610.
* T H I SJOURNAL, 1, 84 (1909) 3 LOC.
c i t . . p. 88.
