T H E J O C R Y A L O F I i V D C S T R I A L A N D ENGINEERIiVG C H E M I S T R Y
M a y , 1916
t o pass a twenty-mesh screen. The hides t a n n e d quickly, t h e grain being completely struck through in a short time. I t was found t h a t this material did not plump t h e hides properly. The hides were in all three cases completely t a n n e d without harmful discoloration. A side of hide t h a t had been prepared f o r t h e one-bath chrome, one-bath vegetable t a n , was submitted t o these leaves for t h e vegetable t a n . Here, t o d , t h e hide was well tanned b u t not properly plumped. Another objection raised was t h e bulkiness of t h e leaf residue in t h e tanning vats. T o do away with this nuisance it would be necessary t o make this material u p into a n extract. T o this extract suitable plumping organic acids could be added. The preliminary d a t a above described would indicate t h a t these leaves are a n extremely satisfactory source of tannin for t a n n e r y purposes. Work along this line is progressing in t h e tannery. SUMMARY
Ceanothur oelutinus, a widely distributed plant, has been found t o contain 7.3 per cent a n d 17.3 per cent tannins. T h e wax was composed of free hydrocarbons, free creotic acid a n d in a great p a r t of palmitic a n d stearic acids in combination with ceryl a n d myricyl alcohols. A trace of glycerides appeared t o be present. The tannin was found t o be of t h e catecholic variety. Tests in t h e t a n n e r y have led t h e authors t o believe t h a t a suitable extract for tanning purposes could be made from these leaves. From t h e quantity of this shrub available, its objectionable presence in t h e forests a n d t h e value of t h e products obtainable from it. this material should become of considerable economic importance. I n concluding, we would wish t o t h a n k Mr. Carl A. Kupper a n d Mr. C. S. Smith, of t h e Vnited States Forest Service, for obtaining t h e material for t h e investigation a n d for t h e survey of its occurrence; Professors W. C. Blasdale a n d H. C. Biddle, of t h e Chemistry Department of t h e University of California, for valuable suggestions given during t h e progress of t h e work; and Mr. David Bloom, for permitting t h e experimental work in t h e tannery. UKIVERSITY OF
CALIFORNIA, BERKELEY
SOME QUALITATlVE TESTS FOR GUM ARABIC AND ITS QUANTlTATIVE DETERMINATION’ By C.
E. WATERSAXD J. B TUTTLE
Received December 16, 1915
INTRODUCTIOS
T h e group of polysaccharides includes such diverse substances as t h e starches, cellulose, t h e dextrins, t h e t r u e gums a n d t h e plant mucilages. T h e y possess i n common t h e property of being decomposable hydrolytically into one or more sugars, usually pentoses or hexoses. By oxidat;on t h e acids corresponding t o these sugars are formed, b u t t h e first step toward this reaction appears t o be hydrolysis. The true gums, of which gum arabic is typical, dissolve in cold water, yielding clear solutions which, 1 Published by permission of the Director of the Bureau of Standards. Copies of the complete paper, from which this is abridged, can be obtained by application t o t h e Director, Bureau of Standards, Washington, D. C.
413
though viscous a n d adhesive, can be filtered. The closely related a n d quite similar mucilages, such as gum tragacanth a n d cherry-tree gum, simply swell up a n d form more or less homogeneous suspensions t h a t cannot be filtered. Perhaps most of t h e SOcalled gums are mixtures of one or more representatives of t h e above classes. As might be expected from their chemical nature, t h e gums do not readily lend themselves t o reactions of a definite qualitative or quantitative value. One of us learned how unsatisfactory some of t h e qualitative tests are when samples of mucilage first came t o t h e Bureau of Standards for examination. This led t o an extended s t u d y of t h e literature and of t h e various qualitative reactions t h a t have been published, a n d finally t o a comparatively accurate quantitative method. I n t h e last part of this work t h e two of us joined forces, hoping t o carry on a much more extended investigation t h a n was found later t o be possible. The object of this paper is t o discuss briefly some of t h e more important qualitative a n d quantitative methods a n d give references showing, as far as possible, what other methods have been published. Q U A L I T A T I V E TESTS
I n nearly all cases a 2 per cent solution of gum arabic was used. Similar solutions of dextrin and of gum ghatti, a substitute for arabic, were subjected t o the same tests. ( I ) FERRIC CHLORIDE A N D ALCOHOL-A mixture of 2 . 5 volumes of jo per cent alcohol a n d I volume of neutral ferric chloride solution containing 2 j g. of t h e salt i n I O O cc., precipitates gum arabic, though often only on long standing.’ Gum ghatti gives no precipitate, a n d dextrin a very slight one. ( 2 ) P O T A S S I U M HYDROXIDE-According t o Liebermann,2 solutions of gum arabic a n d of dextrin become amber-yellow when warmed with potassium hydroxide, while t h e closely related gum senega1 gives a t most a faint yellow color. Sollman3 s t a t e d t h a t dextrin when so treated t u r n s more or less brown, while some sugars a n d gums other t h a n arabic, behave similarly. Rideal a n d Youled came t o t h e conclusion t h a t this test is of no value. Two samples of gum arabic, among a number tested b y them, gave a green color; a solution of ghatti turned pink, and dextrin a very d a r k red or almost black. Nevertheless this tesf is among those recommended in t h e latest edition of “Lunge.”‘ Our own experiments amply confirm t h e statements t h a t this test is of no value. ( 3 ) COLOR R E A C T I O N S W I T H PHENoLs-Reiche6 found t h a t g u m arabic gives a flocculent blue precipitate when boiled with orcinol a n d concentrated hydrochloric acid. Other carbohydrates gave yellow or brown colors. Other investigators o b t a h e d reactions Roussin, J. Pharm. Chim., [41 7 (1868). 251. See also Allen’s “Com. Org. Anal.,” 4th E d . , Vol. 111,p 443. Chem.-Ztg., 14 (1890). 665. 3 A m . J. Pharm., 85, 176; Chem. Zenlr., 83, I (1911), 1560. 4 J . SOC.Chem. Ind., 10 (1891), 610. “Chem.-Tech. Untersuchungsmeth.,” 6th E d . , Vol. 111 (191 1 ) . pp. 167-8. a Ber. Ges. Flirdet. Chem. I n d . , 1819, 74; Chem.-Ztg.. 4 (1880), 191.
414
T H E J O 1-RAT,4 L 0F I N D C S T RI A L A N D E N G I N E E R I N G C H E M I S T R Y
with other phenols.’ I n our n o r k , nine different phenolic compounds were used in alcoholic hydrochloric acid solution. Quite a v-ide range of colors was obtained, b u t t h e test was found t o be unreliable for t h e carbohydrates which were studied. (4) B A S I C L E A D A C E T A T E - L h aqUeOUS Solution Of neutral lead acetate precipitates neither dextrin nor gums arabic and ghatti. 4 n alcoholic solution of t h e salt does, however, slowly precipitate gum arabic.% Xs might be expected from its employment in clarifying sugar solutions before polarization, basic lead acetate is a n excellent precipitant for many gums. T h e solution may be made basic by adding ammonia3 or by boiling with litharge. T h e precipitate formed when gum arabic solution is added, is of a peculiar curdy consistency, and of such characteristic appearance t h a t it is hardly necessary t o make a n y confirmatory tests. Allen4 speaks of this precipitate as a “white jelly.“ Dextrin solutions are, at most, made slightly cloudy by t h e reagent, while gum ghatti yields a very small amount of precipitate. most of which remains suspended in the liquid. T h e basic acetate solution used in this work was made b y long continued boiling of 40 g. of crystallized T.4BLE
REAGENT Ferric chloride and alcohol Basic lead acetate Copper sulfate and sodium hydroxide, cold Copper sulfate and sodium hydroxide, boiled Iodine solution
I--CHIEF
GUM A R A B I C
QL-ALITATIVE
TESTSFOR
Precipitate Kearly clear Dense precipitate Cloudy Blue precipitate, colorless Dark blue solution, often solution cloudy Precipitate darker, pale blue Slight reduction solution S o characteristic color reactions
Ihl, Chem.-Zlg., 9 (18851, 2 3 2 ; Clermont and Chautard, Compt. vend., 94, 1254; Jahvesber., 1882, 684. 2 Chauvin, M o n . Sci., [SI 1 ( l Y l I ) , 317-8; Chem. Z e n l r . , 82, I (1911), 1656. 3 T . Lippmann, “Chemie der Zuckerarten,” 3rd Ed., p . 1616. 4 0 ) . c i t . , p. 441 3 Rideal and Youle, J . SOC.Chem. I n d . . 10 (1891), 610; Fromm, Z . anal. Chem., 40 (1901), 143; Papasogli, L’Orosi, 2 1 (1.8981, 263-5; Scheibler, Z . i - w . Zuckerind., as, 288; B a t t u t , S M C UI n. d i g . et Colo?t., 33, 285; v . Lippmann, “Chemie der Zuckerarten,” p . 1616. 6 “Ziindwaren” in Lunge’s “Chem.-Tech. Untersuchungsmeth.,” 6th Ed,, To]. 111, pp. 167-8. 7 Chem.-Ztg., 14 (18901, 635. 1
(6) M I S C E L L A K E O U S Q U A L I T A T I V E TESTS-x?*Iany other reagents for gum arabic have been suggested, b u t t h e y need only be referred t o as follows: ( a ) Oxidase reaction for distinguishing between gums arabic and tragacanth.’ ( b ) Warming with hydrofluoric acid as a test for gum arabic. glue and dextrin.2 (c) Color reaction with alkali and diazobcnzenesulfonic acid.3 ( d ) Reaction with acidified egg albumen as a test for gum arabic and d e ~ t r i n . ~ ( e ) Kessler’s reagent t o detect gelatin in gum arabic solutions.’ (f) Reduction of mercuric acetate in the presence of sodium chloride.6 (g) Reduction of molybdic acid.’ ( h ) Color reaction with cobalt nitrate and alkali.s (i) Color reaction with colloidal gold s o l ~ t i o n . ~ Other reagents for g u m arabic, such as sodium or potassium silicate, borax or ammonium oxalate,1° would seem t o depend largely upon the natural calcium content of t h e gum, though some of t h e reagents are said t o cause gelatinization. I n our work numerous other reagents were tried, b u t even such promising G U M S AND
GUM GHATTI
lead acetate, dissolved in z;o cc. of water, with a n excess of litharge. I t was filtered: after which it remained clear for a long time. Basic lead acetate has been recommended probably more frequently t h a n any other reagent as a qualitative test for gum arabic, as well as for removing this and other gums from wines. liqueurs, etc., in which other constituents are t o be determined.j ( 5 ) C O P P E R S U L F A T E A N D S O D I U M HYDROXIDE-Jette16 does not mention t h e basic lead acetate test, b u t relies chiefly upon t h e behavior of dextrin: gum arabic, etc., towards copper sulfate and sodium hydroxide. T h e test was devised b y Liebermann’ for detecting dextrin and gum senega1 in t h e presence of gum arabic. The details are given quite fully b y Jettel. We need only say t h a t gum arabic gives a blue precipitate, t h e supernatant liquid being colorless. Ghatti gives a dark blue solution which is sometimes turbid, b u t which clears u p on warming gently; there is a little reduction on boiling. Dextrin gives a precipitate which dissolves on warming; a t t h e same time there is considerable reduction t o cuprous oxide.
Vol. 8, No. 5
DEXTRIN GUM GEDDA
Slight precipitate Less precipitate t h a n arabic Light blue precipitate, dark blue solution Precipitate darker
DEXTRIK Slightly cloudy Slightly cloudy Dark blue solution Complete reduction Reddish purple solution
ones as basic zinc salts and sodium zincate were found t o be useless. I n Table I is given a comparison of the more reliable reactions for t h e identification of dextrin, gums ghatti and gedda, and for gum arabic from different sources. Some of these samples !\-ere obtained after most of t h e work described above had been completed, and only the tests tabulated were applied. A11 samples of gum arabic behaved alike. QL-AKTITATIVE
UETHODS
Since t h e chemical nature of the gums is such as t o render most of the qualitative tests of a t least doubtful value, it is not surprising t h a t there are difficulties involved in their quantitative separation a n d determination. Some of t h e methods t h a t have been proposed will be mentioned below, after which o u r own procedure will be described ( I ) FERRIC
CHLORIDE
AND
CALCIUM
CARBONATE-
The details of this method, which was proposed by Roussin,ll are given b y Allen.’l Payet, A p ~ t h . - Z t g . . 26, 116; Z . anal. Chem., 44 (1905). 493. Oesterv. Chem.-Ztg., 3, 1888; Z . anal. Chem., 40 (1901). 131. 3 Petri, 2.physiol. Chem., 8 291 ; Jahrerber., 1884, 1328. 4 Giinsberg, J . pvakt. Cirtrn., 88, 2 3 9 ; Z . anal. Chem., 2 (1863). 218. 6 Vamvakas, A n n . chim. iinal. apgl., 12 (190i), 12, 139; Anaiysl. 32 (1907), 193, 226. 6 Hager, Pharm. Centv., 18, 313; Z . anal. Chem., 17 (18i8), 380. 7 Hager, “Commentar zur ersten deutschen Pharmakopoe,” Vol. 11. p. 116; ci. Stohmann in “Muspratt,” 4th Ed., Vol. I11 (!89l), p . 1917. 8 Papasogli, L’Ovosi, 2 1 (1898), 263-5. 9 Zsigmondy, Z . anal. Chem., 40 (1901), 697. 10 Allen, O p . cit., p . 441. ” O p tit,; also, Auguet. A n n . P a l s . , 2, 136-8; Chem. Rbs., 6 (191!), 538. 1
: Borntrager,
N a y , 1916
T H E JOURNAL OF I N D U S T R I A L A N D ENGINEERING CHEMISTRY
41 5
with comparative ease. I t was dried, weighed, ig( 2 ) FERRIC C H L O R I D E A N D S O D I C M CARBONATEThis method was employed b y Landwehr‘ for t h e de- nited and t h e ash weighed. T h e results b y this method termination of glycogen and, incidentally, of .gum varied a few per cent below a n d above 100. Copper acetate was chosen for making u p the rearabic. I n connection with this he determined t h e percentages of water held b y ferric hydroxide when agent because no sodium sulfate or other salt indried t o constant weight a t different temperatures. soluble in alcohol of t h e strength used could be formed. Even this reagent is not ideal, and the final step taken ( 3 ) L E A D ~ C E T A T E - T ~ ~use Of this salt, dissolved in alcohol, was proposed b y Chauvin2 and recom- is described below. mended b y Rocques and Sellier.3 As far as we know, ( 7 ) N E T H O D E I K A L L Y ADOPTED-Copper acetate dist h e basic salt has not been used quantitatively. Yari- solved in ammonia possesses certain advantages over ous formulas have been ascribed t o t h e precipitate, all the other solutions employed. I n order t o asand it is not unlikely t h a t it varies in c o m p ~ s i t i o n . ~certain t h e best conditions for precipitation, a few (4) ALCOHOL-chauvin employed alcohol acidified preliminary tests were made. One-quarter gram of with hydrochloric acid t o precipitate gum. T h e acid gum arabic was dissolved in j o cc. of water, 2 j cc. of is unnecessary, for strong alcohol has often been used ammoniacal copper acetate solution (see below) added, t o precipitate’ gum from plant extracts, wines, etc.6 and then enough alcobol t o give t h e required percentage of t h e total volume of the mixture. Thirty ( j ) MIs CE LL A N E 0 us XI E T II ODs- 51e t h o ds de pen ding upon hydrolysis,’ followed by estimation of hexoses per cent of alcohol gave no precipitate. With 4 0 or pentoses, and those by which t h e gum is oxidized per cent there was a faint precipitate, so finely divided t o mucic acid,8 are of no value because samples of t h a t it was practically impossible t o filter. With jo, 60, and 7 0 per cent of alcohol, respectively, dense g u m from different sources yield different percentages of t h e products. Wide variations in the iodine ab- and easily filtered precipitates were formed. The sorption, acidity, and t h e amount of alkali taken u p addition of more alcohol t o the filtrates from these three caused no further precipitation. These prefrom alcoholic potash, are also found.g (6) C O P P E R S U L F A T E 4 S D ALKALI--Madsen’ esti- cipitates gave on ignition t h e following amounts of mated t h e gum in licorice juice (Succus Ziquiritiae) ash, practically all cupric oxide: Percentage of Alcohol Employed: 50 60 io by throwing it out with strong alcohol, dissolving Ash, g r a m . . . . . . . . . . . . . . , . , , , . . . . , . , . . . . . 0.0412 0.0436 0.0544 0.0402 0.0446 0,0564 in water, a n d precipitating with copper sulfate a n d 0.0398 0.0452 0,0548 _ _ _ _ _ _ _ sodium carbonate. The gum is not identical with Averages. . . . . , , , . . . . . . . . . . . . . , . . . , 0,0404 0.0445 0.0552 gum arabic, b u t for botanical reasons may well be From this it will be seen t h a t in t h e presence of j o quite similar t o i t . During the course of our work, while still unaware per cent of alcohol, t h e gum is precipitated quantitaof t h e work of Madsen, i t was attempted t o use t h e tively and carries down with it less occluded matter same reagents, together with alcohol, t o precipitate t h a n when t h e mixture contains more alcohol. T h e following procedure was finally adopted for t h e g u m arabic quantitatively. The filtrates were always cloudy. Slightly better, b u t still far too low, results determination of gum arabic: Fifty grams of copper were obtained when sodium hydroxide instead of car- acetate were dissolved in water, a n excess of ammonia bonate was used, and having the combined solutions added, and the solution diluted t o 1000 cc., using water contain 60 per cent of alcohol. Since a n y excess of and alcohol in such proportions t h a t the final solucopper over t h e gum present would be thrown down tion contained j o per cent of alcohol. For each deas hydroxide, t h u s increasing t h e difficulty of washing termination a jo-cc. portion of a gum arabic solution, a n d drying t h e precipitate, some modifications of representing 0 . 2 5 g. of gum, was pipetted into a 2 jo-cc. Fehling’s solution were next tried. The best results beaker, a n equal volume of alcohol added. and then were obtained with a reagent made u p with copper z j cc. of t h e copper reagent, with constant stirring. acetate, sodium potassium t a r t r a t e and sodium hy- The precipitate was allowed t o settle, was filtered on droxide. When this was added t o a solution con- a tared paper, washed with j o per cent alcohol containing about 0.25 g. of g u m arabic in 40 cc. of water, taining ammonia, then with 7 j per cent, and finally there was no precipitation, or a t most a slight turbidity. with 95 per cent alcohol. I t was dried t o constant An amount of 9 j per cent alcohol equal t o t h e volume weight a t IO j O, ignited in a porcelain crucible and the of t h e mixed solutions was then added, with vigorous ash weighed. T h e amount of ash was deducted from stirring. This resulted in t h e formation of a fine- the weight of t h e original precipitate and the differgrained precipitate t h a t could be filtered o f f and washed ence called “net gum arabic.” The amount of moisture in t h e gum originally taken for analysis must be ’ Z . physiol Chem., 9 (1884), 164. See footnote 9 ; also Ann. Fals., 6, 27-30; Chem. Zanlv., 83, I (1912). allowed for. This is determined b y drying in a current 756. of hydrogen at 1 0 5 ” . No allowance is made for t h e A n n . chim. anal. appl.. 16 (1911), 218-20; Chem. Zcnlr., 82, I1 (1911), potassium and calcium which form a n integral part 394. Riegel, Arch. Phavm., [21 54, 155; Jahresber., 1 (1847-8), 795; Battut, of t h e gum. These may be t o some extent retained SuCr. I n d i g . el Colon., S2, 2 8 5 ; Scheibler, Z. Vev. d. Zuckerind., 23, 288. in t h e precipitate a n d , therefore, be included in t h e 6 LOC. Cil. ash. Any error t h a t may be introduced b y neglecting Diehl, Phaum. Rundschau, 1, 31;. Z . anab. Chem., 22 (1883), 622; Ifadsen, Pharm. Centr., 20, 144; Z . anal. Chem.. 22 (1883). 134. this is small and very much less t h a n the error inherent ’ v. Lippmann, Lac. c i t . in t h e method. Kiliani, Ber., 16 (1882), 34-i. Williams, Chem. Y e w s , 58, 224: 2. anal. Chem., 28 (1889), 732. I n some of the preliminary work the gum-copper
416
T H E JOI;'R,VAL O F I - V D C S T R I A L AilrD E - V G I S E E R I X G C H E i M I S T R Y
precipitate was dried a t 95' and then at IO^'. The additional loss in weight a t the higher temperature was usually about 3 or 4 mg.. or 1.j per cent. The following results were obtained with 0 . 2 j - g . portions of gum: R E S C L T OBTAIKED S BY T H E A m r o m ~COPPERSULFATE METHOD SERIES A B C D Percentage of gum f o u n d . . 96.6 100. I 100.4 101.6 98.0 99.0 101.2 102.8 96.9 98.3 99.1 103.3 98.4 99.3 99.7 ... 96.7 100,i 98.8 ...
...........
Averages. ...................... 97.3 General average.. . . . . . . . . . . . . . . .9 9 . 5
99,s
99.9
102.6
I t is evident from a n inspection of the figures t h a t the method as finally modified is capable of giving results as accurate as could be expected. It is realized, ho\Tever. t h a t much work could yet be done upon mixtures of gum arabic with ghatti. dextrin. etc.. as well as upon mucilages of known composition. A few preliminary determinations, made while Fehling's solution was still being experimented with. indicate t h a t dextrin and ghatti tend t o be carried down with g u m arabic. At t h e same time a sample of mucilage mas prepared according t o the formula in the U. S. Pharmacopoeia; from this the gum vias precipitated with strong alcohol. dried and analyzed. The net gum found in four determinations was about 91 per cent. This low result may be due t o a fault of the method or t o partial hydrolysis of the gum. The publication of the results so far obtained has been so long delayed, and the opportunity t o complete the v o r k seems so remote, t h a t i t has been decided t o present this paper without furiher postponement. S T U 11A R Y
T h e most characteristic qualitative test for gum arabic is the precipitate formed with basic lead acetate. AIixtures of copper sulfate and sodium hydroxide, and of neutral ferric chloride and alcohol are of value as confirmatory tests. A summary of t h e more important methods t h a t have been proposed for the estimation of gum arabic is followed b y a description of the steps t h a t led the authors t o the use of alcoholic copper acetate-ammonia solution for this determination. BUREACOF STANDARDS, WASHINGTON
THE OCCURRENCE OF AZELAIC ACID AS A PRODUCT OF T H E SPONTANEOUS OXIDATION OF FATS B y BEN H. P\-ICOLET
.ISD
LEONARD M. LIDDLE
Received October 13, 1915
I t is very well known t h a t azelaic acid, COOH(CH2)?COOH, is a normal product of the artificial oxidation of various unsaturated f a t t y acids. All of t h e common eighteen-carbon unsaturated f a t t y acids (oleic, linolic. linolenic) very probably have a double bond located between the ninth and tenth carbons, that is, in the center of the carbon chain. I t is further pretty well established that neither linolic nor linolenic acid has a double bond between this position and the carboxyl group. Consequently it was t o be expected, and it has been found, t h a t azelaic acid results from t h e oxidation of these acids with alkaline permanganate, from t h e decomposition of their ozonides, and from the breaking down of their nitrogen peroxide addition products.
1-01. 8, S o . 5
I t seemed probable t o us t h a t the difference between artificial and natural oxidation would be one of d e gree. or perhaps only one of time, rather t h a n a coniplete difference in kind. In this case azelaic acid should be a product, and perhaps an important p r o d uct, in t h e development of rancidity, and particularly in its more advanced stages. Griigerl exposed some oils t o air on filter paper and examined them after four years. One of the products isolated was an impure azelaic acid, supposed t o he mixed with some suberic acid. Scala2 isolated from rancid material a number of f a t t y acids ranging from formic t o pelargonic, and also most of the corresponding aldehydes. He obtained one fraction of nonvolatile, water-soluble acids, which yielded crystals t h a t he suspected of being azelaic acid. Here apparently the matter rests to-day. I t is the purpose of the present note t o call attention to the occurrence in very highly rancid cottonseed oils of amounts as high as I O per cent of azelaic acid or a glyceride of this acid. The glyceride, whose presence is indicated, would belong t o a type heretofore unknown. Certain grades of fuller's earth are largely used in the bleaching of refined cottonseed oil. The oil remaining absorbed in t h e earth after filtering and pressing! is in a n ideal condition for oxidation by air. I n fact the oil-soaked clay, in large scale work, frequently becomes heated t o the ignition point b y simply standing in contact with air. The oil from a specimen of earth t h a t had been used t o bleach cottonseed oil. and had since been kept for a year and a half without any effort t o protect it from t h e air, was extracted for examination. From IOO g. of clay, low-boiling petroleum ether extracted only 4 g; of oil. H o t alcohol extracted z j g. The characteristics of the oils from these two extracts are given below. EXTRACT Ether Amount of Extract (grams). . . . . . . . . . . . . . . . . . . . . . . . . 4.0 Acid No. (mg. KOH per g . oil), . . . . . . . . . . . . . . . . . . . . . . 145.4 Saponification N o . (mg. KOH per g . oil saponified). . . . . 282 . O Volatile Acids ( m g . KOW t o titrate steam-volatile acids from 1 n.) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.0 Azelaic Acid (after saponification). . . . . . . . . . . . . . . . . . . 3%
Alcohol 25 . 0 144 8 260.0 35.0 10%
I n spite of its high acid number, the fraction extracted b y alcohol gave very little azelaic acid when extracted with hot water. I t was accordingly saponified, and the liberated f a t t y acids extracted with hot water. The water solution thus obtained was concentrated and g a r e in various fractions crystals of crude azelaic acid amounting t o about I O per cent of t h e fat saponified. This crude acid,' after two or three recrystallizations from water, showed the constants of pure azelaic acid. Found Melting p o i n t . . . . . . . . . . . . . . . . . . . . . . . . . . . 105' Equivalent weight (by titration). . . . . . . . . . . 9 3 , 9
Pure azelaic acid 106O 94.0
Another sample of clay was extracted wholly with alcohol, and azelaic acid determined in the hot water extract both from the original material and from t h e acids obtained on saponification. Z . angew. Chem., 1889, 62; J . SOC.Chem. I n d . , 1889, 202. Slue. s f i e v . U E Y . i t a l . , 30 (1897). 613; Goza. chim. ital., 38, I (1908). 307; abstract Chem. Zenlr.. 1908, I. 2085. 1 2
