MCRENZIE AND WREN : OPTICALLY ACTIVE GLYCOLS. 473LI1.-Optically Active Glycols Derived from 1-Benzoinand from Methyl 1-Mandelate.By ALEX. MCKENZIE and HENRY WREN.THE work described in the present communication was undertaken inconsequence of the racemisation phenomena which were observed byone of us (Wren, Trans., 1909, 95, 1583, 1593) in connexion with thestudy of I-benzoin and its derivatives. I-Benzoin, which is preparedfrom magnesium phenyl bromide and I-mandelamide (McKenzie andWren, Trans.,'1908, 93, 312), has [aID - 119O in acetone solution, andundergoes complete racemisation with great readiness in the presenceof alkali. The interpretation was suggested that the isomeric changein question was probably of a keto-enolic character, the hypotheticalo-dihydroxystilbene being formed as an intermediate phase :C,H,*QH=OH C,H,*S*OH C,H,*QH OHC,H,*C*OH C,H,*COLaevorotatory.Inactive. Inactive.C,H,*CO -+ -+The racemisation of 2-benzoin methyl ether also proceeds with greatreadiness in the presence of alkali.I n accordance with this view, the hydrogen attached to the carbonatom of the :CH*OH group migrates to the adjacent carbonyl-oxygenatom :C,H,*FEC*OHC&,*CO "3 fDesmotropic change of this nature is, however, impossible if thecarbonyl group in I-benzoin is displaced by the CRR-OH group, atransformation which can be effected by the application of Grignard'sreaction to kbenzoin. We find accordingly that optically activeglycols of the types :C,H,* CH( OH) CRR OH and C,H, CH( 0 H) *CRR'* OHdo not lose their activity in the presence of alkali.The tripheny474 McKENZIE AND WREN: OPTICALLY ACTIVE GLYCOLSglycol, obtained from methyl Z-mandelate and magnesium phenylbromide, has [aID +221.3' in acetone solution, and is perfectly stableeven when boiled with N/lO-alcoholic potash for thirty minutes.Similarly, the diphenylethyl glycol, obtained from I-benzoin andmagnesium ethyl iodide, retains its activity unchanged during twenty-five hours at the ordinary temperature when dissolved in N/10-alcoholicpotash, whereas I-benzoin methyl other, under the same conditions, isracemised completely within five minutes.The convereion of methyl I-mandelate, a highly active lsevo-rotatory compound with [a], - 236' in carbon disulphide solution(Wren, Zoo.cit.), into the highly active dextrorotatory triphenylglycol, referred to above, suggested the possibility of a Waldeninversion having occurred in this change. In order to obtain furtherevidence on this point, the action of magnesium phenyl bromide onI-benzoin was examined, when it was found that the resulting glycolwas dextrorotatory and identical with the product from methylI-mandelate. There is thus no evidence of a Walden inversion ineither of these actions, and i t is accordingly proposed to designate thedextrorotatory glycols, derived either from methyl I-mandelate or fromI-benzoin, as I-compounds.Desyl chloride has not yet been obtained in an optically active form,and we therefore treated Lbenzoin with thionyl chloride, in the hopeof obtaining an interesting compound in which the displacement ofthe chlorine by the hydroxy-group might be studied.Unfortunately,the product of the action was inactive desyl chloride. The displace-ment by the chlorine atom of the tertiary hydroxy-group i n the activeglycols, which are described in this paper, has not, so far, beenattempted, since the action both of thionyl chloride and of fuminghydrochloric acid on inactive triphenylethylene glycol caused theelimination of one molecular proportion of water with the formationof triphenylvinyl alcohol. Now Tiffeneau (Compt. rend., 1908, 146,29) has found that the latter alcohol is also formed from triphenyl-ethylene glycol by means of sulphuric acid; this behaviour is,however, abnormal, the researches of Tiff eneau showing that theelimination of water from the glycols of the type :Ar GI1 (OH) CRR'.OHresults in other cases in the formation of aldehydes of the type :ArRR'C CHO.A further point arose in connexion with the interaction of I-benzoinand magnesium methyl iodide, Whilst the former compound containsonly one asymmetric carbon atom, the carbon atom of the carbonylgroup becomes asymmetric during this action, so t h a t the formatioDERIVED FROM L-BENZOIN AND FROM METHYL L-MANDELATE. 475of two isomeric laevorotatory glycols, each one of which contains twoasymmetric carbon atoms, is theoretically possible :?GH5H-C-OH H-C-OH1 and OH-$-CH, CH,-F-OHCiP5 ‘6=5These glycols are not, of course, enantiomorphously related, andmight be expected to be produced in unequal amounts, Only oneisomeride was obtained, Again, when r-benzoin was acted on bymagnesium methyl iodide, only one of the two possible inactiveglycols was isolated, and it was obvious from the yield that the otherisomeride could have been present only in small amount, or notat all.The only glycol of the type R*CH(OH)*CRR*OH known up to1904 was the triphenylethylene glycol described by Gardeur (Bull.Acad.roy. BeEg., 1897, [iii], 34, 67). Acree (Ber., 1904, 37, 2753)has shown that similar glycols are obtained by the aid of Grignard’saction (compare also Tiffeneau and Dorlencourt, Ann. Chim. Phys.,1909, [viii], 16, 237, and other papers). These compounds are alloptically inactive.The Grignard action as applied to an ester of anoptically active hydroxy-acid was studied for the first time byP. F. Frankland and Twiss (Trans., 1904, 85, 1666), who preparedd-aa88-tetraphenylerythritol by the interaction of methyl d-tartrateand magnesium phenyl bromide. This ditertiary glycol is charac-terised by the high dextrorotation which it exhibits when contrastedwith that of the tartrate from which it is derived.I n the course of the large amount of work carried out by manychemists on the connexion between unsaturated or negative groupsand optical activity, the abnormal effects produced by the phenylgroup have been repeatedly observed. It is, therefore, not surprisingthat such effects should be encountered with the compounds which arenow described.Purdie has shown that the substitution of the hydrogen atom ofthe alcoholic hydroxy-groups in the optically active lactic, malic, andtartaric acids (or their esters) by an alkyl group causes a very pro-nounced rise of optical activity.Vhen a similar displacement iseffected in E-mandelic acid, no such effect is observed (McKenzie,Trans., 1899, 75, 753).C,H,-CH(OH)*CO,Me,has [a], -236’ in carbon disulphide solution (Wren, Zoc. cit.),whereas methyl I-phenylmethoxyacetate, C,H,*CR(OMe)*CO,Me (seeexperimental part) has [a], - 1 0 1 . 7 O in the same solvent. The com-Again, methyl I-mandelate4'76 McKENZIE AND WREN : OPTICALLY ACTIVE QLPCOLSparison between the effect on rotatory power which is brought aboutby the displacement of the hydroxy-group in kmandelic acid (or itsmethyl ester) by the methoxy-group, and a similar displacement inaliphatic hydroxy-acids (or esters), leads to the conclusion that thegreat difference between the two cases is due to t h e influence exertedby the phenyl group.When E-triphenylethylene glycol, OH*CHPh*CPb,*OH, is alkylatedby nieans of silver oxide and methyl iodide, only one of the twohydroxy-groups undergoes methylation.It is shown that the mono-methoxy-derivative, obtained in this manner, has the formulaOMe*CHPh*CPh,*OH,and not OH*CHPh-CPh,*OMe, and that the introduction of themethyl group into the molecule of the triphenyl glycol lowers thevalue for the specific rotation to + 185.3' in acetone solution.I n a recent exhaustive study of the influence of constitution on therotatory power of optically active compounds, Rupe (Annalen, 1909,369, 311 ; compare also ibid., 1903, 327, 157) points out that theI-menthyl esters of saturatttsd acids, derived from phenylcinnamic acids,possess a higher degree of optical activity than the correspondingesters of the unsaturated acids.The normal effect which Rupeobserved in this particular group appears to be that negative groupslower the rotatory power iu a pronounced manner. Thus, in I-menthyl/3-phenylcinnamate, CPh2:CH*COf*CloH19, with [a],, - 37-92' inbenzene solution, the optical effect caused by the asymmetricmenthyl group is influenced by the electronegative character of twophenyl groups and of one double linking.When one of these negativefactors is eliminttted, for instance, the double linking, the rotatorypower is enhanced, thus I-menthyl PP-diphenylpropionate,CHPh,*CH,*CO,*C,,H,,,has [a], - 61.72' in benzene solution, Now this behaviour isopposed to the deduction from work of Tschugaeff, Haller, Walden,and others, and, indeed, also from Rupe's own work on this subject,namely, t h a t unsaturated groups (phenyl group, double linking) tendto enhance optical rotation. For example, Rupe finds tbat I-menthylcrotonate has [.ID - 91-06O in benzene solution, whereas I-menthyln-butyrate has [a], -70.56", the elimination of the double linkinglowering the rotatory power in this case. Again, Frankland andSlator (Trans., 1903, 83, 1349) show that d-tartranilide has a higherdextrorotation than has d-tartramide, and that aromatic groups raisethe rotation of the latter compound very considerably.Finallyanother example of the same effect, and a very striking one, isthe comparison between p-hydroxyphenyliminocamphorDERIVED FROM L-BENZOIN AND FROM METHYL L-AT ANDELATE. 47 7with [alD +1363O in chloroform solution, and the product of itsCH*NH*C,H,*OHreduction, p-hy droxy phenylaminocamphor, C,H,,< I co 9wit,h [a]= + 83' in chloroform solution (Forster and Thornley, Trans.,1909, 95, 942).I n the course of his important work in this field, Rupe emphasisesthe depression in rotation exerted by the displacement of a methyl bya phenyl group, thus :[alD i nCH,:CMe'CO,H ........................... - 91.76"CH,:CPhCO,H ..........................63.03CMe,:CH*CO,H .......................... 88.60Z-Menthyl esters of the acids. Leiizene soliition.CMePh:CH-CO,H ........................ 65-89CPh,:CH*CO,EI ........................... 37'92These figures are quoted here for comparison with the valuesobtained for tho glycols :[alu inOH*CHPh'CMe,'OH .................. - 21 '6'OH'CHPh'CMePh'OH ................. + 34-0OH*CHPh*CEtPh*OH ................. +27'4OMe'CHPh'CPh,'OH .................. + 185.3OH'CHPh*CPh,*OH .................. 4- 221 '3The latter compounds are, of course, of a very different type fromthose of Rupe, and it is of interest that the effect of the phenyl groupis so pronounced.EXPERIMENTAL.Z:Glycols. acetone solution.Action of Magnesium Methyl Iodide on 1-Benzoin.The Grignard reagent, prepared from 1.4 grams of magnesium(4 mols.), 7-8 grams of methyl iodide (4 mols.), and 50 C.C.of ether, wascooled in ice-cold water, and 3 grams of finely-powdered Lbeneoin(1 mol.) were added in small quantities at a time. The action wasvigorous, After the mixture had been boiled gently for three hours,it was decomposed in the usual manner by ice and dilute sulphuricacid, and the liberated glycol extracted with ether. After drying theethereal solution and removing the ether, the resulting viscid, brownproduct solidified when stirred with a small quantity of carbondisulphidb. It was purified by crystallising from carbon disulphideseveral times, 10 C.C. of solvent being used on each occasion, until itsrotation was con stant.l-ap-Dihydroxy-a/3-&phenyZpopune, OH CHPh- CMePh OH, separ-ates from carbon disulphide as a colourless, amorphous solid, andmelts at 81-82O.It is very readily soluble in boiling carbondisulphide, and sparingly so in the cold solvent. It is veryreadily soluble in cold benzene, methyl alcohol, ethyl alcohol, ether, o478 McKENZIE AND WREN: OPTICALLY ACTIVE GLYCOLSacetone, and less so in cold carbon tetrachloride.in boiling light petroleum (b. p. 60-80°), separating on cooling.is very sparingly soluble in water.cold concentrated sulpburic acid.It dissolves readilyItIt gives a violet coloration withFor analysis, the glycol was dried a t 78O until constant in weight :0.11'75 gave 0.3383 CO, and 0.0743 H,O.C,,H,,O, requires C = 78.9 ; €3.= 7.1 per cent.The specific rotation was determined in acetone solution, the glycolC= 78.5 ; H= 7.1.having been dried at 78" :l = 2 , ~ = 1 . 6 4 8 , a:'' +1.12O, [u]? +34*Oo.No racemisation was observed with the solution of the glycol inalcoholic potash.Action of Magnesium Methyl Iodide on r-Benzoin.Since a second asymmetric carbon atom is generated by the actionof magnesium methyl iodide on benzoin, the behaviour of r-benzoinwas studied in order to find out if only one glycol is formed. Theexperiment indicated that, if an isomeric glycol is produced, i t can bepresent only in small amount.Twelve grams of r-benzoin were gradually added to the Grignardreagent, prepared from ,5.5 grams of magnesium, 30.3 grams ofmethyl iodide, and 100 C.C.of ether. The crude product, obtained asin the previous experiment, amounted t o 12.5 grams. After onecrystallisation from carbon disulphide, the glycol melted at103.5-105.5°, and the yield was 11.5 grams. After a secondcrystallisation from the same solvent, the compound was pure.Inactive up-dihydroxy-up-diphenylpropane separates from carbondisulphide in colourless needles, and melts at 103*5-104*5°. It is lesssoluble in carbon disulphide or light petroleum than is its 2-isomeride.When dried in a vacuum over sulphuric acid at the ordinary tempera-ture, it retains persistently small quantities of solvent. For analysis,it was accordingly dried at 78O until constant in weight :0-1223 gave 0.3520 CO, and 0.0764 H,O.C15H1602 'requires C = 78.9 ; H = 7.1 per cent.The preparation of this glycol has also been described by Tiffeneauand Dorlencourt (Ann.Chim. Phys., 1909, [viii], 16, 237), who givethe melting point as 104O. There is also no evidence in the workrecorded by these authors of the formation of a second isomeride.C = 78.5 ; H = 7.0.Action of Magnesium Ethyl Iodide on 1-Benzoin.Four grams of I-benzoin (1 mol.) were treated with magnesium ethyliodide, obtained from 1.8 grams of magnesium (4 mols.), 11.8 grams oDERIVED FROM L-BENZOIN AND FROM METHYL L-MANDELATE. 479ethyl iodide (4 mols.), and 50 C.C. of ether. The crude diphenylethylglycol (4.3 grams) was crystallised from successive small quantities ofcarbon disulphide.The first crop obtained melted sharply, and, whendried at 78' until constant in weight, gave the following value for itsspecific rotation in acetone solution :l = 2, C = 1.288, aD + 0*64O, [=ID + 24.8'.After a second crystallisation, the melting point was the sameas before, but the value for the specific rotation determined asbefore was somewhat higher. The concentration, however, wasgreater :I = 2, c = 4,674, a:" + 2.56', [a]:' + 27.4'.The value obtained after another crystallisation was practicallyidentical with this.I-ap-Dihydroxy-ap-d~phe~ ylbutane, OH*CHPh* CE tPh* OH, separatesfrom carbon disulphide in colourless prisms, capped by pyramids,and melts at 96-5-975'. It is very readily soluble in boiling carbondisulphide, and sparingly soluble in the cold solvent. It is solublewith difficulty in light petroleum or water.It is easily soluble in coldacetone, ethyl alcohol, benzene, chloroform, or ether, and less soin carbon tetrachloride. I t s solution in cold concentrated sulphuricacid is magenta-coloured, and becomes green on heating.For analysis, it was dried a t 78':0.1509 gave 0,4366 CO, and 0-1018 H,O.The value for the specific rotation in acetone solution is + 27.4", asC = 78.9 ; H = '7.5.C,,H,,O, requires C = '79.3 ; H = 7.5 per cent.given above. I n ethyl-alcoholic solution :I = 2, C = 2.883, ctF5 + 1-13', + 19.6'.I n chloroform solution :The corresponding inactive compound has been prepared by Acree(Amer. Chem. J., 1905, 33, 193).When the 1-glycol was dissolved in ethyl-alcoholic potassiumhydroxide, no racemisation was detected.Thus 0.2977 gram, whenmade up t o 10 C.C. in cold alcoholic potash (0-104N), had a,, + 1.20' ina 2-dcm. tube. This value did not alter during twenty-five hours atthe temperature of the laboratory.1 = 4, c = 2.044, + 0*26O, [a]$'5 + 3.2'.It melts at 115-116'.Action of Mclgnesizcm Phenyl Bromide on Methyl 1-Madelate and on1- Benzoin.A solution of methyl I-mandelate (8 grams) in ether (50 c,c.) wassiphoned within an interval of eight minutes into a solution of theGrignard reagent, prepared from magnesium (4.6 grams), bromobenzen480 McKENZIE AND WREN: OPTICALLY ACTIVE GLYCOLS(30-3 grams) and ether (50 c.c.). After the vigorous action hadsubsided, the mixture was boiled gently for two hours and thendecomposed in the usual manner.The resulting brown solid wascrystallised several times from methyl alcohol until i t s rotation wasconstant.For analysis and determinations of its specific rotation, tihe glycolwas dried a t 100' until constant in weight. It retains methylalcohol with considerable obstinacy, determinations indicating,however, that the alcohol is not present in the air-dried productin definite molecular proportions.l-up-I)i~ydroxy-uPP-~ri~?~~~~yZet?~ane (tripheny lethy lene glycol),OH*CHPh*CPh,*OH,separates from methyl alcohol in colourless needles, and, after beingdried at loo', melts a t 128-129', the meltiug point of the inactiveisomeride being 167',according to Acree (Ber., 1904, 37, 2762).Theactive glycol is readily soluble in boiling methyl alcohol. It may alsobe crystallised from light petroieum, in which it is sparingly soluble.It is practically insoluble in water, but dissolves with ease in coldacetone, benzene, ether, or chloroform. I t gives an emerald-greencoloration when beated with concentrated sulphuric acid ; the solutionbecomes colourless in presence of excess of water, and a yellow tintappears when an excess of alkali is added. These colour reactions arealso exhibited by triphenylvinyl alcohol :0.1719 gave 0.5230 CO, and 0.0960 H20.. C=83-0; H=6*2.C,,,H,,O, requires C = 82.7 ; H = 6.3 per cent.A determinatiou of its specific rotation in acetone solution gave theresult :I = 4, c = 1.0156, + 8*99*, [ u ] E ' ~ + 221.3'.Its rotation was also determined in chloroform solution :I = 4 , C = 1.3196, u: + 12-33', [u]: + 233.6'.It was of interest to find out if a dextrorotatory glycol would alsobe obtained by the interaction of I-benzoin and magnesium phenylbromide.Finely-powdered I-benzoin (3 grams) was accordingly addedgradually to the solution obtained from magnesium (1 gram), bromo-benzene (8.1 grams), and ether (60 c.c.), the Grignard reagent havingbeen cooled previously in ice-cold water. After the addition of thebenzoin, the mixture was boiled gently for ninety minutes. The glycolwas isolated as in the preceding experiment.0.1343 gave 0.4085 CO, and 0.0758 H20.C,,H,,O, requires C = 82.7 ; H = 6.3 per cent.The compound was identical with that obtained from methyI-mandelate, as shown by determinations of melting point and specificrotation made with it.On analysis :C = 82.95 ; H = 6.3DERIVED FROM L-BENZOIN AND FROM METHYL L-MANDELATE.481The action of ethyl-alcoholic potassium hydroxide m a s examined.I-Triphenylethylene glycol (0.8 gram) was boiled for thirty minuteswith 25 C.C. of 0.1121-alkali. A violet coloration appeared a t first,and, when the solution was shaken, this became reddish-brown.The mixture was poured into 200 C.C. of water, and the precipitatedglycol filtered and dried. Polarimetric examination in acetonesolution showed that the glycol had remained unracemised evenafter this drastic treatment with alkali.Action of Magnesium Methyl Iodide on Methyl l- Mandelate.A solution of methyl I-mandelate (7 grams) in &her (40 c.c.) wassiphoned within an interval of ten minutes into an ice-cold solution ofthe Grignard reagent, prepared from magnesium (4.4 grams), methyliodide (26.1 grams), and ether (50 c.c.).The crude product obtainedin the usual manner was an oil, which was dried and obtained crystal-line by being stirred with light petroleum. After two further crys-tallisations from light petroleum containing a little ether, the glycolwas obtained pure.1-up-Dihydroxy-aphertylisobutane, OH*CHPh*CMe,*OH, melts a t33-5-35', It gives an orange coloration with cold concentratedsulphuric acid, and the solution becomes strongly fluorescent whenwarmed :0.2657 gave 0.7000 CO, and 0.2029 H,O.I t s specific rotation was determined in acetone solution :C = 71.9 ; H = 5.5.C1,H1,O, requires C = 72.2 ; H = 8.5 per cent.I = 2 , ~=2*7352, - 1*18', [a]: - 21.6'.Action of Thionyl Chloride on 1-Benzoin and on Inactive Triphen$-ethylene Glycol.The action of thionyl chloride on I-benzoin was studied in the hopeof obtaining optically active desyl chloride, but racemisation occurredin the displacement of the hydroxy -group by chlorine.LBenzoin(3.2 grams) was covered with thionyl chloride (3.5 grams). Theevolution of hydrogen chloride began at the ordinary temperature,whilst the product gradually became liquid. After three hours, thetemperature was raised gradually from 15' to 55' during the course oftwo hours, when very little action appeared to take place.A t 55',hydrogen chloride was again evolved somewhat briskly, Thetemperature was then raised to 90'. The dark reddish-brown liquidbecame semi-solid when placed over soda-lime in a vacuum. Aftersome days, the specific rotation of the product in acetone solution was-9*7", but this activity may have been due t o a little unchange482 MCKENZIE AND WREN: OPTICALLY ACTIVE GLYCOLSbenzoin. After three crystallisations from ethyl alcohol, desylchloride, melting a t 66-68O, was obtained. This was quite inactivewhen examined polarimetrically.Thionyl chloride (28 grams) was added to inactive triphenylethyleneglycol ( 5 grams) prepared from r-benzoin according to Acree (Zoc. cit.).The glycol dissolved within fifteen minutes at the ordinary tempera-ture, the red solution slowly evolving hydrogen chloride.Thetemperature was raised slowly to the boiling point of thionyl chlorideduring one hour, and maintained at this point for two hours longer.After drying over soda-lime in a vacuum, the resulting solid wascrystallised twice from light petroleum (b. p. 60-80O). The compoundobtained in this manner was quite free from chlorine, and its meltingpoint (135-136') and analysis showed that it was triphenylvinylalcohol :0.1654 gave 0.5352 CO, and 0.0889 H,O.C,,H,,O requires C = 88.2 ; H = 5.9 per cent.Triphenylvinyl alcohol has been prepared by Delacre (Bull. Soc.chim., 1895, [iii], 13, 857 ; compare also Saint-Pierre, Bull. SOC. chim.,1891, [iii], 5, 292 ; Gardeur, Bull.Acad. roy. Bely., 1897, [iii], 34, 67),who employed the Friedel-Crafts' reaction with trichloroacetyl chlorideand benzene. The proof, however, that the compound in question wastriphenylvinyl alcohol, CPh,:CPh*OH, and not triphenylethanone,CHPh,-COPh, was supplied by Biltz (Ber., 1899, 32, 660). Anschutzand Forster (Annalen, 1909, 368, 89) have observed recently thattriphenylvinyl alcohol is formed by the interaction of acetylmandelylchloride, benzene, and aluminium chloride, whereas the formation ofbenzoin acetate might have been expected. The same authors alsoprepared the vinyl alcohol from desyl chloride, benzene, and aluminiumchloride.The action of fuming hydrochloric acid' on inactive triphenylethyleneglycol is similar to that of thionyl chloride.Three grams of theglycol were added to 40 C.C. of aqueous hydrochloric acid, saturated a t0". After several weeks in a stoppered bottle a t the ordinarytemperature, the product was diluted with water, and the solidcrystallised twice from ethyl alcohol. It melted at 135*5--136*5O, andwhen it was mixed with an equal amount of the glycol, obtained bythe aid of thionyl chloride, the melting point did not change.C = 88.2 ; H = 6.0.Inactive p-Hydrox?/-a-methoxy-app-tripl~enylethane.Methylation of Inactive Triphenylethylene Glycol,-Four grams of theinactive glycol (1 mol.) were added to 9.6 grams of silver oxide(3 mols.), 24 grams of methyl iodide (12 mols.), and 35 C.C. of acetone.The mixture was boiled gentlyfor six hours and filtered.The solvenDERIVED FROM L-BENZOIN AND FROM METHYL L-MANDELATE. 483was removed from the filtrate, and the product again alkylated withhalf the above quantities of oxide and iodide. The product obtainedin this manner was crystallised once from much light petroleum (b. p.60-80°), and then twice from methyl alcohol :0.1166 gave 0.3557 CO, and 0-0675 H20. C = 83.2 ; H = 6.5.0.3284 ,, 0.2392 AgI ; OMe = 9.6.C,,H,,O, requires C = 82.9 j H = 6.6 ; OMe = 10.2 per cent.This compound is obviously a monomethyl derivative of triphenyl-ethylene glycol. The following experiments showed that it was/3 - hydroxy - a - methoxy - aP/3 - triphenylethane, and not a - hydroxy-P-methoxy-a/3P-triphenylethane.Action of Nagnesiurn Phenyl Bromide on Inactive Benzoin -&ethylEther.-The Grignard reagent, prepared from 0.4 gram of magnesium(1.3 mols.), 2.5 grams of bromobenzene (1.3 mols.), and 14 C.C.ofether, was boiled gently for two hours with 2.7 grams of inactivebenzoin methyl ether (1 mol.). The product, obtained in the usualmanner, was a crystalline solid, and amounted to 3.2 grams. Aftertwo crystallisations from much ethyl alcohol, P-hydroxy-a-methoxy-aPP-triphenylethane, melting at 139O, was obtained. Its identity withthe compound prepared from triphenylethylene glycol was shown bythe melting point of a mixture of the two, and by the analysis :0.1818 gave 0.5510 GO, and 0*1080 H,O.Action of Magnesium Pheny2 Bromide on Inactive Methyl Phenyl-C = 82.65 ; H = 6.6.C21H2002 requires C = 82.9 ; H = 6.6 per cent.methoayacetate.-Inactive phenylmethoxpacetic acid,C,H,*CH(OMe)*CO,H,was prepared by the interaction of sodium methoxide and methylphenylchloroacetate.It was isolated by the aid of its sparinglysoluble sodium salt (compare McKenzie, Trans., 1899, 75, 753), andthen esterified by means of methyl alcohol and sulphuric acid.Eighteen grams of methyl ester, boiling at 118-1 19O/8 mm., wereobtained from 20 grams of acid.Five grams of the methyl ester (1 mol.), dissolved in 40 C.C. ofether, were added to a solution of 2 grams of magnesium (3 mols.) in13.2 grams of bromobenzene (3 mols.) and 40 C.C. of ether within aninterval of six minutes. The crude product obtained by decomposi-tion of the additive compound was crystallised twice from ethylalcohol.The yield of pure methoxy-glycol amounted to 5.4 grams.It was identical with the products obtained from triphenylethyleneglycol and benzoin methyl ether respectively.Inactive P-h y drox y-a- me t hox y-app- triphen, y Zet hane,OMe*CHPh CPh,. OH,melts at 138*5-139*5°. It is fairly soluble in boiling ethyl alcohol,VOL. XCVII. K 484 McKENZIE AND WREN : OPTICALLY ACTIVE GLYCOLSand sparingly so in the cold. It crystallises from ethyl alcohol incolourless prisms. It may also be crystallised from light petroleum(b. p. 60-80"), in which it is fairly soluble on heating. It is fairlysoluble in cold ether, easily so in cold acetone or chloroform, somewhatless readily so in cold carbon tetrachloride, and sparingly soluble inwater.It gives a colour reaction with concentrated sulphuric acidsimilar to that of I-triphenylethylene glycol.1-P- Hydroxy-a- methox y- a/3/3-triphenyZethane.Action o j Magnesium Phenyl Bromide on Methyl I- Phenylmethoxy-ctcetate.-Methyl I-mandelate was alkylated with silver oxide andmethyl iodide, the mixture of methoxy-ester and unchanged methylmandelate saponified, and the resulting acid converted into sodiumsalt. Since sodium Lphenylmethoxyacetate is sparingly soluble inwater (McKenzie, Zoc. cit.), it can be separated readily from the sodiummandelate present.I-Phenylmethoxyacetic acid, obtained from the sodium salt, had[u]" - 150.1O for c = 3.597 in ethyl-alcoholic solution, the value quotedpreviously being [a]:" - 150*0° f o r c = 6.7656.I-Phenylmethoxyacetic acid (10 grams) was converted into its methylester by the Fischer-Speier method, using methyl alcohol and sulphuricacid. The yield was 8.5 grams.Methyl l-phenylmethoxyacetate, C,H,-CH(OMe)*CO,Me, is a colourlessoil, which boils at 1175-1 1 S0/8 mm.:0.2885 gave 0.6988 CO, and 0.1775 H,O.C,,H,,O, requires C= 66.6 ; H = 6.7 per cent.The following determinations of specific rotation of this ester invarious solvents were made for comparison with the activity of methylI-mandelateC = 66.1 ; H = 6.9.In carbon disulphide solution :I = 1, c = 2.93, uF5 - 2-98', - 101.7".In acetone solution :I = 4, C = 2,6948, CZ: - 10*38", [a,]: - 96.3'.In benzene solution :I = 4, c = 2.0968, a: - 8.324 [a]: - 99.2'.A solution of methyl I-methoxymandelate (5 grams, 1 mol.) in ether(40 c.c.) was siphoned with constant shaking within an interval ofsix minutea into a solution of the Grignard reagent (3 mols.) preparedfrom magnesium (2 grams), bromobenzene (13.2 grams), and ether(40 c.c.).Towards the end of the addition a bulky precipitateseparated, which became grey on warming. The mixture was heatedfor two hours. The crude product resulting from the action amounteDERIVED FROM L-BENZOIN AND FROM METHYL L-MANDELATE. 485t o 9 grains. After two crystallisations from ethyl alcohol, thecompound is pure.I - P - N y c l r o x ? / - a - n t e t / i o x ? / - a ~ ~ - t ~ ~ ~ r ~ ~ ~ ~ ~ ~ e t ~ i a ~ e , OMe*CHPh*CPh,*O€l[,separlttes from ethyl alcohol in colourless needles, grouped in rosettes,and melts at 143-144'.Its colour reaction with concentratedsulphuric acid is similar to that of the inactive isomeride :0.1459 p v e 0.4423 CO, and 0.0852 H20.0.239'7 ,, 0.1742 AgI. OMe = 9.6.Its specific rotation was determined in a number of solvents.C = 82.7 ; H = 6.5.C,,H,,O, requires C = 82.9 ; H = 6.6 ; OMe = 10.2 per cent.I nacetone solution :I = 1, ~ = 5 * 4 2 8 , a: + 10*06O, [u]F + 185.3'.I n chloroform solution :I = I, C = 4.579, ag + 10.76', [a];' + 235.0'.In benzene solution :I = 1, C = 3.667, uk + 10*8', [a]: + 294.5'.I n ethyl-alcoholic solution :I = 4, c = 1.0176, a: + 6-77', [a]: + 166.3'.The glycol is sparingly soluble in cold ethyl alcohol. It may also becrystallised from methyl alcohol or light petroleum (b. p. 60-80'). Itis easily soluble in cold acetone, chloroform, carbon tetrachloride,ether, or benzene, and sparingly so in water.Methylat ion of 1- Triphen ybtlby Zene Glycol. -Seven grams of I- triphenyl-ethylene glycol (1 mol.) were heated with 11.2 grams of silver oxide(2 mols.), 42 grams of methyl iodide (12 mols.), and 10 C.C. of acetoneduring seven hours. The alkylation was found, however, to beincomplete even after a second alkylation under the same conditionsas before. After a third alkylation, the product mas crystallised untilits rotation was constant, four crystallisations from light petroleumbeing necessary. The glycol obtained in this manner was identicawith I-P-hy droxy -a-me thoxy-aPP- triphenyle thane described above. Itmelted at 143-144', and its identity was confirmed by the mixedmelting-point method. A determination of its specific rotation inacetone solution gave the result :I=3, ~=2*885, a: + 10.68', [u];' + 185.1'.No racemisation was detected with the solution of this glycolin alcoholic potash. 0.254 Gram was made up to 25 C.C. with0*168N-alkali. This solution gave aD + 6 * 8 4 O , a value which had notchanged after twenty hours at the ordinary temperature.K K 486 HARDING AND HAWORTH : SYNTHESIS OF A~-CYCLOPENTENE-The bulk of the expense of this investigation has been defrayed bygrants from the Government Grant Committee of the Royal Societyand from the Research Fund Committee of the Chemical Society, forwhich we desire to make this grateful acknowledgment.BIRKBECK COLLEGE,LONDON, E.C