CHEMICAL EXAMINATION O F SARSAPARILLA ROOT. 201XXV -Chemical Bxanaiization of Sarsaparilla Root.By FREDERICK BELDING POWER and ARTHUR HENRY SAUWAY.Sarsaparilla root is obtained from different species of Smdaa whichare indigenous to tropical America, from Mexico to Brazil. It hasbeen used medicinally for several centuries, and still maintains aplace in the various national Pharmacopoeias.The above-mentioned root has been the subject of numerouschemical investigations, which appear to have been conducted chieflywith the object of ascertaining the nature of its so-called activeprinciplee, or the constituents to which the assumed therapeuticvirtues of the root are due. The earliest of these investigations wa202 POWER AND SALIVAS:apparently that of Pallotta (J.Pharm. Chim., 1824, 10, 543), whoobtained a substance in the form of a white powder, possessing apeculiar odour and a bitter taste. This evidently impure substancewas considered to bo a new organic base, and was designatedpariglina or parillin. Berzelius, in 1826, suggested for it the namesmilacin. Products were subsequently obtained by Thubeuf ( J .Pharm. Chim., 1832, 18, 734; 1834, 20, 162, 679) and by Batka(Annden, 1834, 11, 313; J . Pharm. Chim., 1834, 20, 43), whichwere designated respectively salseparin and parillinic acid, but itwas shown by Poggiale (J. Phwm. Chim., 1834, 20, 553) that thesesubstances, although obtained by different methods, were practicallyidentical in composition and properties with the pariglina (parillin)of Pallotta.The investigations mentioned were followed by variousothers, which need not be here enumerated, until Fluckiger ( A rch.]'harm., 1877, 210, 535) isolated and described a substance, forwhich he retained the original name of parillin. This was definitelyshown to be a glucoside, and to belong to the class of saponins, theglucosidal nature of " smilacin " having previously been observedby 0. Gmelin (Annulen, 1859, 110, 174). Fluckiger proposed forparillin the alternative fo.I'TILulz, @40H70018 or C48H86018, and con-sidered that the saponins as a class might be represented by thegeneral formula CnH2~-ioO18.The most recent extended investigation of sarsaparilla root wasconducted by W. v. Schulz (Pharm. J., 1892, 52, 6; Arb. despharmkol.Inst. z u D o r p t , 1896, 14, 14), who has stated it tocontain three distinct saponin glucosides, to which the followingnames, formulae, and characters were assigned :I. Parillin, of Pallotta and of Fluckiger (" Sinilaciii " of Ber-zelius), C26H4,0,,,2~H2,0. M. p. 176-14O; [a]= - 42'33O. Crystalline,and almost insoluble in cold water.11. Smilasaponin (" Smilacin " of Merck), (C,,H,,0,0)5,12H20.[aID -26.25O. Amorphous, and soluble in water.111. Sarsasaponin, (C22~,0,,),,,24H,0. M. p. 220.26' ;[a]= - 16'25O. Crystalline, and readily soluble in water.It was noted by v. Schulz that although the analysis of the last-mentioned compound gave somewhat higher figures for the hydrogenthan the formula C,,H,,O,, requires, and the formula (&H&&would therefore, appear at first sight to be more correct, he hasadopted the former expression in order to bring the substance intothe series having the general formula CnH2n-8010.This generalformula has been considered by Kobert to represent the compositionof all the above-mentioned compounds, as well as that of a numberof other substances belonging to the class of saponins (compare vanRyn, Die GEykoside, p. 217). It has, furthermore, been stated bCHEMICAL EXAMINATION OF SARSAPARILLA ROOT. 203v. Schulz that the hydrolysis of the sarsapadla glucosida is noteasily effected by heating under ordinary conditions with diluteacids, and that it is most completely accomplished when the opra-tion is conducted in a sealed tube. The ultimate products of thehydrolysis, besides sugar, have been designated respectively a.aparigenin, smilasapogenin, and sarsasapogenin, and to all of thesethe formula C,,H,O, or C2,H,,04 has been assigned.It has been assumed by v.Schulz (Zoc. cit.) that the above-mentioned behaviour of the sarsaparilla glucosides on heating withdilute acids is due to the presence of several sugar complexes inthe molecule, which become successively eliminated, and that thefirst products of hydrolysis therefore still retain the character ofglucosides. This view has also been entertained by Rosenthalerand Strom (Arch. Phnrm., 1912, 250, 290) with respect to thesaponiu obtained from the white or Levant soaproot (from a speciesof G y p ~ ~ p h i l a ) . The last-mentioned authors propose to designatethe first product of hydrolysis as ‘‘ pro-sapogenin,” restricting theterm sapogenin to the final productl of hydrolysis, which is obtainedby heating the glucoside with acid under pressure.The so-calledpro-sapogenin, although crystalline, did not yield satisf actor1 resultson analysis, which it was thought might be due to the presence ofsome impurity, such as sapogenin.Although some glucosides are known which, by suitable methodsof hydrolysis, are capable of yielding intermediate glucosidalproducts, such, for example, as amygdalin and apiin (Ber., 1895,29, 1508; A?tnden, 1901, 318, 121), i t appears very doubtfulwhether the saponins in a pure state actually possess this character.The present investigation of sarsaparilla root has shown it to containbut one definite saponin glucoside, which agrees fairly well in itspercentago composition arid characters with the sarsasaponin ofv.Schulz (Zoc. cit.), and this name has therefore been retained,although a different formula has now been assigned to it. Thesarsasaponin of the present authors is, however, readily hydrolysedby heating with dilute acids, without the formation of an inter-mediate product, but it has now been ascertained that it is accom-panied in sarsaparilla root by a phyhsterolin (phytosterol gluco-side), which, as has recently been shown (T., 1913, 103, 399, 1022),is not changed by the ordinary methods of hydrolysis. There canbe little doubt that the glucosidic products other than sarsasaponinwhich have hitherto been obtained from sarsaparilla root were notpure substances.The composition and characters assigned to the=called parillin would appear to indicate that it consisted essen-tially of a mixture of the substance now designated ils sarsasaponinwith varying proportions of a phytosterolin, whilst in the course o204 POWER AND SALWAY:the present investigation i t has been definitely ascertained thatsmilasaponin ((' smilacin ") is not a homogeneous compound.With consideration of the confusion which has existed respectingthe glumsides of sarsaparilla root, and the fact that nothing hasbeen known of its other constituents, apart from the recorded pres-ence of starch, traces of an essential oil, a little fatty oil, resin, andinorganic salts, it has seemed desirable to subject this root to a moresystematic and complete chemical examination.The results of thepresent research, together with the deductions from them, aresummarised at the end of this paper.EXPERIMENTAL.The material employed for this investigation consisted of a goodquality of commercial grey Jamaica sarsaparilla, such as is recog-nised by the British Pharmacopaeia, and which is there describedas the dried root of Smilax ornata, Hooker, fil.In order to ascertain whether an enzyme were present, a, quantity(500 grams) of the powdered material was mixed with water andkept for two days a t the ordinary temperature. The mixture waethen filtered under pressure, and alcohol added to the filtrate.Aflocculent precipitate was thus produced, which, when dried in avacuum over sulphuric acid, amounted to 2.6 grams, or 0.52 percent. of the weight of root employed. This substance slowly hydro-lysed amygdalin, thus indicating its enzymic activity.Another portion (10 grams) of the powdered root ww tested f o rthe presence of a.n alkaloid, but with a negative result.Twenty-five gram9 of the powdered root were next extractedsuccessively in a Soxhlet apparatus with various solvenb, when thefollowing amounts of extract, dried a t looo, were obtained :Petroleum (b. p, 35-50') extracted 0.11 gram = 0'44 per cent.Xther ,, 0.11 ,, = 0'44 ,,Chloroform ,, 0.07 ,, = 0.28 ,,Alcohol ,, 1-30 ,, = 5-20 ,, Ethyl acetate ,, 0.20 ,, = 0.80 ,,- -Total ...............1.79 grams== 7-16 per cent.For the purpose of a complete chemical examination 22.1 kilo-grams of the ground material were extracted by continuous perco-lation with hot alcohol. After the removal of the greater portionof the alcohol, a viscid, dark-coloured extract was obtained, amount-ing to 2-95 kilogramsCHEMICAL EXAMINATION OF SARSAPARILLA ROOT. 205Distillation of the Eztract with Steam. Separation of anEssential Oil.The whole of the above extract was mixed with water, and avigorous current of steam passed through the mixture for severalhours. The distillate, on extraction with ether, yielded 2.6 gramsof an essential oil, being thus equivalent to about 0.01 per cent.of the weight of root employed. This essential oil, when distilledunder diminished pressure, passed over between 70 and 200°/ 15 mm.as a pale yellow liquid, which possessed a pleasant, somewhataromatic odour, a density of 0.977 a t 15O/15O, and was not com-pletely soluble in 70 per cent.alcohol. It was found to containfurfuraldehyde, and also gave a bluish-black coloration with ferricchloride, thus indicating the presence. of a phenolic substance.a f t e r the above operation the steam distillation flask containeda considerable quantity of a brown resin, which formed with theaqueous liquid a viscid emulsion. Since the resin did not separateon keeping, the mixture was agitated with hot amyl alcohol. Bythis means a very dark-coloured aqueous liquid ( A ) was obtained,whilst the amyl-alcoholic extract contained the resin (B), partlyin solution and partly in suspension. This extract was filtered, thefiltrate well washed with water, the solvent then removed, and theresidual semi-solid resin, together with that portion of the resinwhich was insoluble in amyl alcohol, put aside for subsequentexamination.Examination of the Aqueous Liquid (A).Isolation of a new Dicarboxylic Acid, Sarsapic Acid, C,H,O,.The aqueous liquid from which the resin had been removed, asabove described, was repeakedly extracted with ether.The combinedethereal extracts were then concentrated to a convenient volumeand shaken with an aqueous solution of ammonium carbonate. Onacidifying the ammonium carbonate extract it yielded about 6 gramsof a semi-solid substance, which was dissolved in ether, the etherealsolution being washed, dried, and the solvent removed.The residue,which gradually solidified, was first freed from a little adhering oilby drying on a porous tile, and then crystallised from hot watercontaining a little alcohol. A substance was thus obtained whichseparated in slender, colourless needles, melting at 3 0 5 O :0.0850 gave 0.1307 CO, and 0.0200 H,O. C=41-9; H=2*6.0.1728 required for neutralisation 20.0 C.C. AT/ 10-KOH.C,H,O, requires 0=41*9; H=2.3 per cent.N.V. =64-9206 POWER AN]) SALWAY:A dicarboxylic acid, C,H,06, requires N.V. = 65.2.The molecular weight of the substance was determined by Barger’s0*0700 in 5.6 C.C. alcohol was between 0.065 and 0.075 mol.It is evident from these results that the above substance is adibasic acid possessing the empirical formula C6H406.The onlycompounds of this formula hitherto recorded are tetrahydroxybenzo-microscopic method, with the following result :Mean M.W. = 179. C,H,O, requires M.W. = 172.quinone, C O < ~ ~ ~ ~ ~ ~ $ ~ ~ { > C O , and 3 : 5-dihydroxy-4-pyrone-2-carboxylic acid, CO<C(oH)L--&a C( OH)‘C( CO H)> 0, neither of which, liowever,is identical with the above-described substance. Since the latteris therefore a new compound, it is proposed to designate i t snrscipicctcid, with reference to the source from which it has been obtained.Sarsapk acid, C4H,0,(C0,H),, is sparingly soluble in cold wateror ether, but very soluble in alcohol. It is moderately soluble inhot water, from which it separates, on cooling, in colourless needles.It gives no coloration with ferric chloride.Its metJtyZ ester,C4H20,(C0,*CH,),, prepared by passing dry hydrogen chloride intoa hot methyl-alcoholic solution of the acid, crystallises from alcoholin colourless leaflets, melting at 1 2 1 O . This substance is volatilein steam, and, when gently heated, possesses an odour suggestive ofsafrole. It was analysed with the following result :0.0995 gave 0.1746 CO, and 0.0363 H,O.C8H80, requires C = 48.0 ; H = 4.0 per cent.From the above-mentioned properties of sarsapic acid it wouldappear to contain two carboxylic groups, which would account forthe state of combination of four of the oxygen atoms. I n order toascertain the manner in which the two remaining oxygen atomsare combined, methyl sarsapate was heated for some time withacetic anhydride, but no change took place.The same ester washeated for several hours with sodium acetate and hydroxylaminehydrochloride in aqueous alcohol, but no reaction occurred. Itwould appear, therefore, that sarsapic acid contains neither ahydroxyl nor a carbonyl group. With consideration of these factsit is highly probable that the acid possesses the constitution :C=47.9; H=4*1.,-yC:H :C(CO,H>>,‘CH:C(CO,H) *The ethereal liquid , which had been shaken with aqueous alumon-iuni carbonate for the removal of the sarsapic acid, as describedabove, was subsequently extracted with aqueous solutions of sodiumcarbonate and sodium hydroxide.Both of these extracts, on acidiCHEMICAL EXAMINATION OF SARSAPARILLA ROOT. 207fication, yielded only a small amount of a brown, amorphous solid,whilst the ethereal liquid remaining after the treatment withalkalis also contained nothing definite in character.The aqueous liquid, after being extracted with ether as abovedescribed, gave on agitation a copious and persistent froth, andevidently contained some saponaceous substance. I n order to isolatethe latter, if possible, the aqueous liquid was extracted repeatedlywith a.myl alcohol. The combined extracts were first washed witha little water, and then concentrated under diminished pressure,when a considerable quantity (20 grams) of a pale brown, amor-phous solid was deposited.This substance was found to be gluco-sidic in character, and also to possess saponin-like properties, butno definite compound could be isolated directly from it. With theobject of ascertaining whether any definite hydrolytic productcould be obtained from the substance, the latter was heated forsome timz with dilute sulphuric acid, when a brown solid soon beganto separate. After the hydrolysis was complete the mixture wasextracted with ether, when some indefinite material remained undis-solved. The ethereal solution was then first washed with aqueoussodium hydroxide, which removed some colouring matter, andsubsequently with water, after which i t was dried and the etherremoved, when about 1 gram of a crystalline solid remained.Thissubstance was purified by recrystallisation from alcohol, whenit formed colourless, prismatic needles, melting a t 183O. It wasfound to be identical with sarsasapogenin, C,,H,,O,, the hydrolyticproduct of the glucoside (sarsasaponin) , which was subsequentlyisolated, as described below.The aqueous liquid obtained in the above-described hydrolysiswas treated with baryta water for the removal of the sulphuricacid, and concentrated to a small volume. It then readily reducedFehling’s solution, but no ositzone could be prepared from it.It is evident from the above results that the brown, amorphoussolid which had been obtained by extraction with amyl alcoholcontained a small amount of sarsasaponin, together with a consider-able proportion of indefinite material.The ampl-alcoholic filtrate remaining after the removal of theabove-described brown, amorphous solid also yielded a small amountof sarsasapogenin on hydrolysis with sulphuric acid.The original aqueous liquid, after having been extracted withainyl alcohol as above described, wits treated with an exceN of basiclead acetate, when an abundant, dark brown precipitate wasproduced.This was collected and washed with water, then sus-pended in water, and decomposed by hydrogen sulphide. Thefiltered liquid, which was very darkly coloured, was concentrate208 POWER AND $AL,WAY:to a small volume, but nothing crystalline separated on keeping.A portion of the liquid was therefore heated for a short time withaqueous potassium hydroxide, after which treatment the mixturewas acidified and extracted with ether.The ethereal liquid wasthen shaken with aqueous ammonium carbonate, which removed asmall quantity of a crystalline acid. This substance, after purifica-tion, melted a t 305O, and was identified as sarsapic acid, C6H406,which had previously been isolated as described above. Since thissmall amount of acid was only obtained after heating with alkali,it must have been present in the aqueous liquid in some form ofcombination. The ethereal liquid, after the removal of the sarsapicacid, contained only indefinite colouring matter.The filtrate from the basic lead acetate precipitate was treatedwith hydrogen sulphide for the removal of the excess of lead, andthe filtrate concentrated to a syrup.On keeping the latter for sometime about 10 grams of a crystalline solid separated in long needles.This substance consisted of potassium nitrate, which had previouslybeen observed to occur in sarsaparilla root. The syrup also con-tained an abundance of sugar, since it readily yielded d-phenyl-glucosazone, melting and decomposing a t 2 1 2 O . I n order furtherto confirm the identity of the sugar a portion of the syrup washeated with acetic anhydride. An acetyl derivative wm thusobtained, which, when crystallised several times from alcohol, melteda t 130-131°, and consisted of 6-penta-acetyldextrose. The above-mentioned eyrup gave no precipitate with potassium-mercuric iodide,iodine, phosphomolybdic acid, o r mercuric nitrate, but, when heatedwith potassium hydroxide, ammoniacal vapours were evolved.Sinceit still possessed saponaceous properties, a portion of the syrup washeated with 5 per cent. aqueous sulphuric acid, and the mixturesubsequently extracted with ether, when a small amount of sarsa-sapogenin, C26H4203, was obtained, thus indicating the presence ofsarsasaponin in the original aqueous liquid.Examhation of the ResitL (B).The crude resin, which had been separated from the aqueousliquid ( A ) by means of amyl alcohol as above described, wasdissolved in alcohol, mixed with purified sawdust, and the driedmixture successively extracted in a large Soxhlet apparatus withlight petroleum, ether, chloroform, ethyl acetate, and alcohol.Petroleunt Bstrcici of *.the IZesi~z.Isolation of Sitosterol, C2,H460.The petroleum extract of the resin was a dark-coloured, viscidSince it consisted chiefly of fatty solid, amounting to 125 gramsCHEMICAL EXBMINATION OF SARSAPARILLA l-iOOT.209matter, it was heated for some time with 50 grame of potassiumhydroxide in the presence of alcohol. The greater portion of thealcohol was then removed, water added, and the mixture subse-quently extracted many times with ether. The combined etherealextracts yielded on evaporation 10 grams of an oily residue, whichrapidly solidified. This material was dissolved in a hot mixture ofalcohol and ethyl acetate, when, on cooling, a substance separatedin colourless needles, melting a t 135-13607 which gave the colourreaction of the phytosterols :0.0958, heated at llOo, lost 0*0040 H,O.H,O=4.2.0.0918 * gave 0.2818 GO, and 0.0995 H,O. C= 83.7 ; H = 12.0.C$,H,,O,&O requires H,O = 4.5 per cent.C2,H4,0 requires C= 83.9 ; H = 11.9 per cent.A determination of the optical rotatory power of the substance0*3048,* made up to 20 C.C. with chloroform, gave a, -Oo50/ in aThe acetyl derivative, when crystallised from a mixture of alcoholand ethyl acetate, separated in stellar clusters of colourless needles,melting a t 126-127O.It is evident from these results that the above-described substanceis sitosterol.gave the following result:2-dcm. tube, whence [aID -27.3O.Isolatiotb of Sitosterol-d-gl,ucoside (PliytosterolirL), C,,H,O,.The alkaline liquid, from which the abovedescribed sitoeterol hadbeen removed by ether, was acidified with dilute hydrochloric acid,and the precipitated material extracted with ether.A portion ofthe precipitate, amounting to about 3 grams, was very sparinglysoluble in ether. This w a ~ therefore separately collected, and thenpurified by crystallisation from amyl alcohol. It was thus obtainedin microscopic needles, melting and decomposing a t 280-285O :0.1062 gave 0.2802 CO, and 0.0984 H20. C = 72.0; H = 10.3.C,H,,06 requires Q= 72.3 ; H = 10.2 per cent.From the analysis and properties of the above-mentioned sub-stance it appeared to consist of a phytosterolin (phytosterol gluco-side). The correctness of this supposition was confirmed by hydro-lysing tho substance in amyl-alcoholic solution with hydrochloricacid (compare T., 1913, 103, 403), when it was resolved intodextrose, which was identified by means of its osazone, and asubstance melting a t 1 3 6 O .The latter gave the characteristiccolour reaction of the phytosterols, possmsed an optical rotation inchloroform of [aID - 35'2O, and yielded an acetyl derivative meltingVOL. cv.* Auhydrous substancc.1210 POWER AND SALWAY:at 124O. This hydrolytic product was thus identified as sitosterol,and the substance from which i t was obtained was consequentlysitasterold-glucoside.Zdenfiification of the Fatty Acids.The above-mentioned ethereal solution, from which the phyto-sterolin had been removed by filtration, yielded, on evaporation,about 70 grams of crude fatty acids.These were converted intotheir methyl esters, and then subjected to fractional distillationunder diminished pressure, when the greater portion passed overat 210--230°/15 mm., but a small fraction was collected above230°/15 mm. The latter fraction solidified in the receiver, and,when crystallised from ethyl acetate, separated in glistening leaflets,melting a t 58-59O:0.1174 gave 0.3354 CO, and 0.1381 H,O.The above substance appeared. to consist of methyl behenate, andthis view of its character was confirmed on hydrolysis, when afatty acid wm obtained melting a t 76-77O, and possessing anectralisation value of 157 (C~H4402 requires N.V. = 165). Thefatty acid was thus identified ils behenic acid.The fraction of methyl esters distilling a t 210-230°/15 rnm.washydrolysed, and the resulting fatty acids separated into liquid andsolid portions by conversion into the lead salts, and treating thelatter with ether.The Solid Acids.-This portion of acid, amounting to about30 grams, was crystallised once from a mixture of alcohol and ethylacetate, when it melted a t 54-56O. The product was analysed withthe following result:C= 77.9 ; H = 13.1.C,,H,1,02 requires C = 78.0 ; H= 13.0 per cent.0.1148 gave 0.3168 CO, and 0.1289 H,O. C=75*3; H=12*5.C,,H,O, requires C?='75-0; H=12*5 per cent.It is thus evident that the solid acids consisted of a mixture of. palmitic and stearic acids.The Liquid A cids.-These acids, when distilled under diminishedpressure, passed over between 230° and 240°/15 mm., and amountedto 20 grams.An analysis and a determination of the iodine valuegave the following results :CIgHsO2 ,, C=76*1; H=12.7 ,, ,,0-1132 gave 0.3178 CO, and 0.1170 H,O. C=76*6; H=11-5.0' 1492 absorbed 0-2333 iodine.C,,H,O, requires C = 76.6 ; H'= 12-1 per cent. Iodine value =i 90.1.C,slI,20, ,, C=77.1; IT=11*4 ,, ,, Iodine ,, =181.4,Iodine value= 156CHEMICAL EXAMINATION OF SAR.SAPARILLA ROOT. 21 1From these results it would appear that the liquid acids consistedof a mixture of deic and linolic acids, the latter predominating.Ethereal Extract of the Resin.This extract was a dark-coloured solid, amounting to 32 grams.It was digested with a considerable volume of ether, when a portiorlwas found to be very sparingly soluble.The mixture was thereforefiltered, and the insoluble material purified by crystallisation froma mixture of amyl and ethyl alcohols, when about 1 gram of amicrocrystalline solid, melting at 295-300°, was obtained. Thissubstance was identified as sitosterol-d-glucoside, since it yielded, onhydrolysis, sitosterol and dextrose.The ethereal liquid from which the above-mentioned glumidehad been removed was next shaken with aqueous ammoniumcarbonate. The alkaline liquid, on acidification, yielded 1.5 gramsof an oily acid, which became partly crystalline on keeping. Thecrystals were freed from oily matter by pressing on a porous plate,and then recrystallised from dilute alcohol, when colourless needles,melting at 300-305°, were obtained.This acid yielded a methylester, melting at 121°, and was thus identified as sarsapic acid,c6H406, which had previously been isolated from the aqueousliquid.The ethereal liquid was subsequently shaken with aqueous sodiumcarbonate, but only a small amount of a brown, amorphous solidwas thus removed. It was then treated with a 10 per cent. solutionof sodium hydroxide, when an extract was obtained which, onacidification, yielded a quantity of a brown, amorphous solid. Thiswas collected and dissolved in hot dcohol, from which it separatedin an indistinctly crystalline form. Attempts were made to obtainthis substance in a purer condition, but after several separationsfrom alcohol it still retained a brown colour, and was not definitelycrystalline.It began to sinter a t 78O, and melted completely a t145O. It was glucosidic in character, for, on heating with hydro-chloric acid in the presence of alcohol, it yielded an aqueous liquidwhich readily reduced Fehling’s solution, and also a hydrolyticproduct melting a t 50-55O. The latter possessed the propertiesof a fatty alcohol, but the amount obtained was not sufficient forits complete characterisation. It would appear probable, however,from the result, of the hydrolysis, and also from the properties ofthe glucoside, that the latter was somewhat impure cetyl-d-glucoside,which is known t o sinter a t 78O, and become completely melted a t150° (Salway, T., 1913, 103, 1028).P 212 1’OWEH. AND GBLWAYldentificatioit of Stigmasterol, C,,H,,O.The ethereal liquid, which had been shaken with alkalis, as abovedescribed, was finally washed with water, dried, and the etherremoved. A crystalline solid, amounting t o 2 grams, was thusobtained, which, when recrystallised from a mixture of ethyl acetateand alcohol, separated in colourless leaflets, melting a t 1 4 0 O .Thesubstance possessed the properties of a phytosterol, and evidentlyconsisted for the most part of sitosterol, which had previously beenisolated from the petroleum extract of the resin. Its high meltingpoint indicated, however, the presence of some stigmasterol, andthe substance was theref ore successively acetylated and brominated,according to the method described by Windaus and Hauth (Ber.,1906, 39, 4378; 1907, 40, 3681).I n this manner a sparinglysoluble bromo-derivative was isolated, which crystallised from amixture of alcohol and chloroform in colourless leaflets, decom-posing a t 208O:0.0656 gave 0.1162 CO, and 0.0408 H20.C,Hb20,Br4 requires C = 48.7 ; H = 6.6 per cent.From the analysis and properties of the above compound i t isevidently identical with tetrabromoacetyl stigmasterol, thus provingthe presence of stigmasterol .in sarsaparilla root.C-48.3; H=6*9.Chloroform and Ethyl Acetate Extracts of the Resin.These extracts were dark-coloured, brittle solids, amounting to55 and 20 grams respectively. Both of these extracts were gluco-sidic in character, but no definite glucoside could be isolated fromthem. The ethyl acetate extract of the resin, however, on heatingwith dilute sulphuric acid, yielded a small amount of sarsapicacid, C,H,O,.Alcohol Extract of the Resin.Itwas dissolved in alcohol, and the solution kept for some time, whena quantity of a crystalline solid separated, which was collected.The alcoholic liquid was then concentrated to a convenient bulk,and heated for some time with aqueous hydrochloric acid.Afterremoving the alcohol in a current of steam, the remaining aqueousliquid was separated by filtration from a dark-coloured resin, andthe filtrate extracted with ether. The resin also was dige8ted withether, which, however, dissolved but a small proportion of it. Thetwo ethereal liquids were united, washed first with aqueous sodiumhydroxide, subsequently with water, then dried, and the solventThis extract was a dark brown solid, amounting to 45 gramsCHEMICAL EXAMINATZON OF SARSAPARILIA ROOT.213removed. A crystalline residue (0.2 gram) was thus obtained, whichmelted a t 183O, and was found to be identical with the saraa-sapogenin described below,Isolation of Sarsasaponirb, C4$3760,,,7H20.The above-mentioned crystalline solid, which separated from thealcoholic solution of the resin, was purified by several crystallisationsfrom alcohol, when colourless, elongated needles were obtained,which began to sinter a t about 200°, and melted completely a t248O. The substance was glucosidic in character, and its aqueoussolution, when agitated, yielded a copious and persistent froth. Itthus possessed the properties of a saponin.The amount of puresubstance isolated from the above extract of the resin was 1.5 grams,It was analysed, with the following results:0.1089, on heating a t 125O, lost 0.0130 H,O.0.0934 * gave 0.1967 CO, and 0.0663 H,O.H,O=11.9.C = 57.4 ; H = 7.9.0.1048 * ,, 0.2213 CO, ,, 0.0758 HZO. C=57*6; H=8.0.C44H760zo,7H,0 requires HL20 = 12'0 per cent.C44H76020 requires C = 57.1 ; H = 8.2 ,,c44H7,02, ,, c=57'4; H=7*8 ,,The above analytical figures will be seen to be in somewhatbetter agreement with the formula C44H7zOzo than with C44H,6020,but the data subsequently obtained by the hydrolysis of theglucoside are more in accordance with the latter formula, and thishas therefore been adopted as having the greater probability ofcorrectness.W. v.Schulz (Zoc. cit.) had previously isolated from sarsaparillaroot a saponin glucoside, which possessed nearly the same per-centage composition as that above described (Found, C = 57.1 ;H = 8.1), and was designated sarsasaponin. This compound, towhich the formula (C22H,0,,),,,24Hz0 was assigned, was stated tomelt a t 220*26O, to have [a], -16*25O, and to yield an indefinitehydrolytic product of variable composition. Its hydrolysis was,however, considered t o result in the formation of sarsasapogenin,C28H4604, dextrose, and an undeterxined acid or mixture of acids,C4H606, in accordance wi€h the following equation :2(C,,H360,0) + 2H,O = C I , J & 3 0 4 + 2C~H,206 + C4H606.The saponin glucoside, which has now been isolated fromsarsaparilla root, yields a well-defined, crystalline, hydrolyticproduct, which will subsequently be described.Although thisglucoside differs in some of its other characters, such as meltingpoint and optical rotation, from the sarsasaponin of v. Schulz, and* Dried a t 125"214 POWER AND 8ALWA4Y:a somewhat different formula has now been assigned to it, there canbe no doubt of the fundamental identity of the compounds. It istherefore deemed desirable that the name sarsasaponin should beretained for the glucoside which is here described.Sarsasaponin is readily soluble in water or hot alcohol, but onlyvery sparingly soluble in ether. It cani?ot be crystallised fromwater, and is best purified by crystallisation from 95 per cent.alcohol.It shows, however, a great tendency to separate fromconcentrated alcoholic solutions in a gelatinous form.Sarsasaponin can be removed for the most part from its aqueoussolution by shaking the latter with finely divided substances whichare insoluble in water, and it had been observed by v. Schulz(Zoc. cit.) that when its lead compound was decomposed by hydrogensulphide, the glucoside was contaiiied to a large extent in theprecipitated lead sulpliide. I n the course of the present investi-gation sarsasaponin couici only he isolated in a pure state fromthe resinqus material, notwithstanding the fact that i t is readilysoluble in water, and was evidently contained, in part, in theaqueous liquid obtained by treating the alcoholic extract of theroot with water.From the facility with which i t is mechanicallyprecipitated, as well as from- the results of cryoscopic observations,it seems highly probable that sdrsasaponin forms with water onlycolloidal solutions. The property of forming such solutions wouldalso account for the persistent froth which is produced by thesubstances designated as saponins when shaken with water.When sarsasaponin 3s dissolved in acetic anhydride, and sub-sequently a few drops of concentrated sulphuric acid added, a yellowcolour is produced, and the liquid rapidly develops a greenfluorescence. On keeping for some time, o r on the addition of alarger amount of sulphuric acid, the liquid acquires a reddish-brown colour.The specific rotatory power of sarvasaponin was determined withthe following result :0-2130,* made cp to 20 C.C.with water, gave a, - l02/ in a 2-dcm.v. Schulz (Zoc. cif..) has recorded that the substance designatedtube, whence [a],, - 48*fi0.by him as sarsasaponin had 3 rotation of [a], -16.25O.H y d ~ o l y s i ~ o{ Sarscisnponin.Formation of Snrsccsnpoyc~tl.n,, CiGH4203, atid Dextrose.One gram of sarsasaponin, in aqueous solution, was heated with5 per cent. sulphuric acid, when, after a short time, a gelatinous,* Anhydrous substanceCHEMlCAL EXAMINATION OF SARSAPARILLA ROOT. 21 5hydrolytic product separated, which gradually became crystalline.The heating was continued for several hours, and the mixture thendistilled in a current of steam, but no volatile product of hydrolysiswas found in the distillate.After this operation there remainedin the distillation flask a crystalline solid, which was collected byfiltration, the filtrate being set aside for the subsequent examinationof the sugar. The solid substance was washed with water, andrecrystallised from alcohol, when i t separated in slender, colourlessneedles, melting at 183O, and containing water of crystallisation :0.1936, heated a t 120°, lost 0.0113 H,O.0.0961 * gave 0'2724 CO, and 0.0906 H20. C=77.3; H=10.5.The molecular weight of the subdance was determined by boththe cryoscopic and the microscopic method, with the followingresults :0,1681,* in 19.8 C.C. benzene, gave A t - O'lOOo, whence M.W. = 424.0.1820,* in 10 C.C.alcohol, was between 0.05 and 0.04 mol.H,O=5'8.Mean M.W. = 404.C26H4203,1iH20 requires H20 = 6.3 per cent.C20H4203 requires C = 77.6 ; H = 10.4 per cent,. M.W. = 402.From the above results it is evident that the hydrolytic productof sarsasaponin, for which the name sarsasapogenin may be retained,possessea the empirical formula C2,H,,O3. This is further con-firmed by the analysis of its acetyl derivative, described below.It has been stated by W. v. Schulz (Zoc. tit.) that the completehydrolysis of sarsasaponin is only effected with great difficulty,and that the resulting hydrolytic product, sarsasapogenin, possessesthe formula C2,H4,04. Inasmuch as the sarsasaponin obtained inthe present investigation was easily and completely hydrolysed bythe above-described treatment, it is probable that the glucoside ofv.Schulz was contaminated with some phytosterolin, such as is nowknown t o be present in sarsaparilla root, and which is only veryslowly liydrolysed by aqueous acids.Sarsasnpogeiziia, C,,H4,02*OH, is readily soluble in chloroform orbenzene, but only moderately so in ether or cold alcohol. Whenthe substance is dissolved in acetic anhydride, and a few drops ofconcentrated sulphuric acid subsequently added, a yellow colorationis produced, and the liquid soon develops a green fluorescence. Onkeeping for some time, or on the addition of a larger amount ofsulphuric acid, the liquid acquires a reddish-brown colour. Sarsa-sapogenin is optically active :0.1133, made up to 20 C.C.with methyl alcohol, gave a, - 0 O 4 1 'in a 2-dcm. tube, whence [aJD -60-3O.* Anhydrous substance21 6 POWER AND SALWAY:Monoacetylsarsasapogenim, Cz6R4,O3*CO*CH,. - This compoundwas prepared by heating sarsasapogenin for an hour with aceticanhydride, the solution being then concentrated to a small bulk,and a little alcohol subsequently added. A-fter a short thee anacetyl derivative separated, which, when collected and recrystallisedfrom alcohol, was obtained in colourless needles, melting a t 137O.It was analysed and its specific rotation determined, with thefollowing results :0.0968 gave 0.2684 CO, and 0.0872 H20.0.1160, made up to 20 C.C. with chloroform, gave aD -0O40‘ inC=75*6; H=10.0.CZ8H4,O4 requires C = 75.7 ; H = 9.9 per cent.a 2-dcm. tube, whence [aID -57’5O.The acid, aqueous liquid resulting from the hydrolysis of sarsa-saponin was treated with just sufficient barium hydroxide to pre-cipitate the sulphuric acid completely, and the mixture thenfiltered.The filtrate, on evaporation, yielded a syrup, from whichd-phenylglucosazone (m. p. 2 1 4 O ) was prepared, thus proving thepresence of dextrose. No evidence could be obtained of theformation of any other sugar as a product of hydrolysis of theglucoside.The above results have shown that sarsasaponin is resolved onhydrolysis into sarsasapogenin, C26€1420s, and dextrose. I n orderto ascertain the relative proportion of sarsasapogenin which isyielded by the glucoside, a definite quantity of the latter, dried at1 20°, was hydrolysed, and the respective hydrolytic productcollected, dried a t l l O o , and weighed.The amount of sarsa-sapogenin thus obtained was 38 per cent. of the weight of sarsa-saponin employed (C44H76020 requires C26H42O3 = 43.5 per cent.).The hydrolysis of the glucoside may thus be represented by thefollowing equation :C,,H7G0,, + H20 = C5-&34203 + 3C,H1,06.The assumption of v. Schulz (loc. cit.) that by the hydrolysis ofsarsasaponin an acid or a mixture of acids is formed, together withthe other products mentioned, cannot be confirmed, and is obviouslyincorrect. The acid observed by him was doubtless only such as isproduced by the action of the hydrolysing agent on the sugar.Examination of “ Smilclcin ’’ ((( Smilasapomh ” of ‘N.Schulz).The substance described by v. Schulz (loc. cit.) as ‘( smilasaponin,”and recorded in the literature undei that name or as “smilacin,”was evidently regarded by him as a dstinct glucoside of sarsaparillaroot. Although an amorphsus product, he assigned to i t thCHEMICAL EXAMINATION OF SARSAPARILLA ROOT. 21’7formula (C20H320,,),,1 2H,O, and observed its optical rotation t o be[uJD - 26.25O. Inasmuch as the “smilasaponin ” examined byv. Schulz was stated to have been the preparation known in corn-inerce as ‘( smilacin,” i t was deemed desirable in connexion with thepresent investigation to determine its character. A small quantityof the preparation was therefore obtained from the same soiirceof supply as had been indicated by v.Schulz.“Smilacin,” as procured from the source indicated, was a palebrownish-yellow, amorphous powder. It dissolved readily in water,yielding a yellow solution, which frothed strongly on agitation, anddid not reduce Fehling’s solution. The substance, when heated ina capillary tube, began to sinter a t 140°, and decomposed withevolution of gas a t 1 6 0 O . When heated a t l l O o , it lost 17-7 percent. of its weight, and on ignit:on it left a small amount ofinorganic residue. I n view of the character of the substance, itwas not deemed suitable for analysis, and on account of the colourof its aqueous solution the optical rotatory power could only bedetermined with approximate accuracy.‘0.2250 of anhydrous substance, made up to 20 C.C.with water,gave a, about - loo’ in a 2-dcm. tube, whence [a],, -44’4O.When a little o+ the substance was dissolved in acetic anhydride,and a few drops of concentrated sulphuric acid subsequently added,a reddish-brown coloration was produced, the liquid also showinga faint green fluorescence.Hydrolysis of ‘( &‘milacin.”-Three grams of the substance weredissolved in 30 C.C, of amyl alcohol, and 10 C.C. of 15 per cent.aqueous hydrogen chloride added, together with sufficient alcohol(about 1 c.c.) to render the mixture homogeneous. This solutionwas heated for three hours, and the amyl alcohol then removedin a current of steam, when a dark-coloured solid remained in thedistillation flask. The solid was extracted with ether, and theethereal solution washed with aqueous sodium carbonate, whichremoved a small amount of brown, amorphous material.Theethereal liquid was then dried, and the solvent removed, when0.3 gram of a substance was obtained, which, when crystallised fromalcohol, separated in colourless, feathery needles, melting a t184-185O. When the substance was mixed with a specimen ofpure sarsasapogenin (m. p. 183O) no depression of the melting pointensued. It also showed tho same behaviour as the last-mentionedcompound when dissolved in acetic anhydride and a drop of con-centrated sulphuric acid subsequently added. The identity ofthe substance with sarsasapogenin was further confirmed byanalysis 218 CHEMICAL EXAMINATION OF SARSAPARILLA ROOT.0.0712 gave 0.2035 CO, and 0.0665 H,O.C2,H4,03 requires C = 77.6 ; H = 10*4 per cent.The aqueous liquid resulting from the above-described hydrolysiswas exactly neutralised with sodium carbonate, and then evaporatedto dryness under diminished pressure.This residue was digestedwith hot alcohol, the mix$ure filtered, and the filtrate evaporated.A syrupy liquid was thus obtained, which readily reduced Fehling’ssolution, and yielded an osazone melting a t 212O.It is evident from the above results that the so-called “smilacin,”which has now been examined, contained a relatively small pro-portion of the glucoside sarsasaponin, but that i t consisted for themost part of indefinite amorphous products.C=77*9; H=10*4.2Yummary and Con c1iisiot)s.The material used for the present iiivestigatiori c.onsistetl of n.good quality of grey Jaiuaica sarsaparilla root.The root was found to contain a small amount, of azi enzyme,which slowly hydrolysed ainygdalin.An alcoholic extract of the root, when distilled in a current ofsteam, yielded an amount of essential oil equivalent to about 0.01per cent.of the weight of root employed. This essential oil wasa pale yellow liquid, which distilled between 70° and 200°/15 mm.,and had a density of 0.977 a t 1 5 O / 1 5 O .The alcoholic extract was found to contain the following definitecompounds : (i) a crystalline glucoside, sarsasaponin, C,,H7,0,,,7H,O(m. p. 248O; [aID -48*5O), which, on hydrolysis, is resolved intosarsasapogenin, C,,H,,O, (m. p. 183O; [ a ] , - 60’3O), and dextrose.Sarsasapogenin yields a mo~~oacetyE derivative, C,,H,,O,-CO*CH,(m. p. 137O ; [alp - 57.5O) ; (ii) sitosterol-d-glucoside (phytosterolin),C33H5606 (m. p. 280-285O); (iii) sitosterol, C,,H,,O (m. p.135-136O; [a],, - 27‘3O) ; (iv) stigmasterol, C,,H,,O, identified byits tetrabromoacetyl derivative, CmH490Br4*CO*CH, (In. p. 208O) ;(v) a nsw, crystalline, dicarboxylic acid, sarsapic acid,C,H,O,(CO,H),, melting a t 305O, and yielding a dimethgl ester,C,H80, (m. p. 121O) ; (vi) dextrose, from which B-penta-acetyldextrose (m. p. 131-132O) and d-phenylglucosazone (m. p. 212O)were prepared; (vii) a mixture of Fatty acids, consisting of palmitic,stearic, behenic, oleic, and linolic acids. The alcoholic extract con-tained, furthermore, a small quantity of a substance which possessedthe characters of cetyl-d-glucoside, and a considerable quantity ofpotassium nitrate was also present. The total amount of resinousmaterial was equivalent to about 1’25 per cent. of the weight ofthe rootIDENTITY OF THE SUPPOSED B-2 : 5-DIMETHYLPIPERAZINE. 219The results of the present investigation have shown that Jamaicasarsaparilla root contains but one definite saponin glucoside, namely,sarsasaponin. This is accompanied, however, in the root by aphytosterolin (sitosterol-d-glucoside), which represents a class ofcompounds that has only quite recently been known t o occur inplants (compare T., 1913, 103, 399, 1022). It is probable that thesame or similar conditions exist with respect to the constituents ofother commercial varieties of sarsaparilla root, and the compositimand properties of the compound designated by v. Schulz and earlierinvestigators as parillin would, in fact, indicate that it consistedof a mixture of sarsasaponin and a phytosterolin. The-so-called‘I smilacin ” ( ’ I smilasaponin ” of v. Schulz), as examined by tEepresent authors, has been ascertained not to be a homogeneoussubstance, but t o contain some sarsasaponin, together with indefiniteamorphous products.It may finally be noted that v. Schulz (Zoc. cit.) had foundsarsasaponin to be, physiologically, the most active of the glucosidicproducts described by him, and this observation ia quite inaccordance with the above conclusions respecting the character of(( parillin ” and (‘ smilasaponin.”THE WRLLCONUE CHEMICAL RESEARCH hBORATORIES,LONDON, E.C