ORGANIC CHEMISTRY.PART ALIPHATIC DIVISION.Radimls and Ions.Electropositive and Electronegative Character .-Great interestattaches to the determination of the relative electro-positivity and-negativity of organic radicals owing to the important influencewhich small differences in polarity can exert on the course ofnumerous reactions. An attempt has been made in a recent papert o place a series of alkyl and aryl radicals in the order of relativenegativity by observing the behaviour of unsymmetrical mercurydialkyls with hydrogen chloride.1 It is assumed that the groupR’ which first dissociates from mercury and combines with thehydrogen ion in solution to form the hydrocarbon R’H accordingto the reaction : R’HgR + HC1+ R‘H + HgRC1, is the moreelectronegative of the two radicals ; that is to say, it has the greaterattraction for electrons and the greater tendency to split off as anegative ion by capturing the electron pair joining it to the metal.The order of increasing negativity amongst the radicals examinedis the following : CH2Ph*CH, < CH,Ph < n-C&,5 < n-C4H, <C,H, < C2H5 < CII, < n2-C6H4C1 < O-C,H&l< $I-c,H@ < C6H5 <m- CH3*C6H4 (p-CH,*C,H, < o-CH,*C,H, < a-nap ht hyl < O-MeO*C,H4<p-MeO*C,H,<CN, and it is seen that the order of the alkylgroups, relative t o one another, is that which has been hithertoaccepted (iso-groups are less negative than the correspondingn-groups, and tert.-groups less negative still), while amongst arylradicals examined the tolyl and methoxy-radicals are more negativethan phenyl, and the chlorophenyl radicals less so.Certain of the broader considerations relating to the classificationof groups are reviewed by R.Robinson in discussing the mechanismsof organic reactions.2 Ordinarily the electropositiveness or negative-ness of radicals is viewed in relation to hydrogen, which is arbitrarilytaken as the standard for comparison. Then R in R-X is consideredt o be electropositive or electronegative according as X receives orloses electrons (by induction or by electromeric change) whenH*X is transformed into R*X. I n this way alkyl groups are alwayselectropositive and NO,, CO, C02Et, and CN are always electro-A., 409.1 1%. S. Kharasch and A. L.Flenner, J . Amer. Chem. b‘oc., 1932, 54, 674;Rapports et Discussions, Inst. Internat. de Chim. Solvay, 1831, p. 423FARMER. 97negative, but it is evident that these designations cannot be appliedto those virtually amphoterk atoms or groups which produceinductive and electromeric effects in opposite directions, unlessthere is some knowledge as to the importance of these effects or oftheir relative magnitude. Unsaturated radicals like vinyland phenyl,although statically nearly always electropositive in character, areelectronegative in relation to the corresponding saturated system ;moreover, owing to structural flexibility and powers of co-operationin electromeric displacements which these groups display, no generalelectrical character can be assigned to them. Atoms possessingunshared electrons (N, 0, C1) and groups terminating in theseatoms exercise generally a negative inductive effect (GI +,0 +--, etc.) and are electronegative, but this can only be statedwith certainty in the absence of electromeric transformations :indeed, where the electromeric displacements of the atom leadnormally to reaction, e.g., when the atom or group occurs in systemssuch as O--Cx=C or @-CzO, the inductive effect acts byopposing or reversing the electromeric effect.The group NH2(not NR,) is electronegative in NB,*CH,*CH3 (dipole N-C), butin NH,Fh, and even more so in PU’H,*COR, the action of the phenylgroup is probably sufficient to reverse the polarity (dipole N-C).Chlorine on the other hand is always electronegative, but to a smallerdegree when it is linked to an unsaturated group, especially if this iskationoid.The high reactivity of organo-magnesium compounds is to beattributed to the tendency towards separation of negatively chargedalkyl groups (the tendency of the metal to assume kationoid functionsnot being thereby opposed) which, possessing only feeble affinityfor their charge, act as energetic anionoid reagents.The alkylgroups of alkyl halides, of alkyl sulphates and of all esters, on theother hand, tend to separate as positively charged ions of kationoidreactivity, thus assisting the halogen atom to form a stable halideRadical Transformations.-Many of the simpler transformationsand isomerisations of organic chemistry can be interpreted asactually occurring by the transfer of Hf, H-, Me-, or electronpairs from one position to another within an organic kation producedIt is doubtful whether the Wiirtz reaction, when appearing in competitionwith organo-magnesium halide formation in the Grignard reaction, can beformulated as simply as R - Mg-Br + R - Br -+R - R + MgBr, inview of the lack of interaction between these two types of organo-halide whenI 3 n6- 6+6+ 6-+ + - + if--heated together (see p. 101).REP.-VOL.ILXIX. 98 ORGANIC CHEMISTRY.-PART I.by the separation of an electronegative group, although the reasonfor these changes is not always plain to see. Thus reactions such asthe dehydration of isobut'yl alcohol to yield n-butylenes in additionto the expected isobutylene, or of tert.-butyl alcohol to give mainlyrearrangement products in any reaction in which hydroxyl is re-moved, can be summarised as due to the series of changes shown in(a)-a series which is analogous to that usually considered to beresponsible for the well-known anionotropic isomerisation of allylicor propenoid compounds, shown in ( b ) :+Ye * * * * * * :A:B:X: ---+ :A:B A A : B : ,+ :Y:A:B: ( a ) ......o f X e .... .... ............ separation tautomerisetion** Shift of electron pair together with the atom or group which it holds. ........ separation shift of 2e :A:B:D:X: ---+ :AiB:D ____f . . . . . OfX0 ..+ye . * * * * * . * .. ..A : B i D -+ :Y:A:BiD ( b )The relationship between various well-known " abnormal " re-actions of organic chemistry is discussed on these lines hy F.C.Whitm~re.~Organo-metallic Compounds.(Continued from Ann. Reports, 1928, 25, 92.)The papers which have appeared in this section since the lastReport are so numerous that reference can be made to only a smallproportion of them.Metal A1kyls.-Lithium alkyls, LiR, which were originally pre-pared by the action of lithium on mercury dialkyls, have beenshown to be obtainable by the action of the metal directly on alkyland aryl halides in ether or benzene.5 Ethyl-lithium reacts withdiethylthallium chloride to yield a true thallium trialkyl, T1Et3,6and by the action of gallium tribromide on methylmagnesiumbromide in ethereal solution the etherate of trimethylgallium,GaMe,,Et,O, is ~btainable.~ Tetraethylgermanium has beenfound to yield on bromination triethylgermanium bromide,GeEt,Br, and from this compound the tri- and di-ethyl oxides andimines of germanium, (GeEt,),O, GeEt,O, (GeEt,),NH, andG-eEt,:NH, and various other alkyl derivatives of germane anddigermane have been directly or indirectly ~btained.~ MagnesiumJ .Amer. Chem. SOC., 1932, 54, 3274; A., 1016.K. Ziegler and H. Colonius, Annaten, 1930, 479, 135; A., 1930, 590;H. Gilman, E. A. Zoellner, and W. M. Selby, J . Amer. Chem. Soc., 1932, 54,1957; A., 728... *.. .... ..H. P. A. Groll, ibid., 1930, 52, 2998; A., 1930, 1302.C. A. Kraus and E. A. Flood, ibid., 1932, 54, 1635; E. A. Flood, ibid.,p. 1663; A., 606FARMER. 99dialkyls and diazyls can be obtained smoothly by the action ofmagnesium on mercury dialkyls and diaryls in ether.Contrary toearlier observations, the latter compounds are soluble in ether.*Observations of the thermal decomposition of sodium andpotassium methyls indicate that t'he reaction proceeds ultimatelyaccording to the equation : 8M*CH, = 6CH4 + M,C, + 6M.Methylsodium is regarded as the salt of a weak acid, the anion ofwhich suffers the fundamental change : 4CH,- --+ 3CH4 + C----,owing to the capture of protons by some of the methide ions at theexpense of others. The stripped quadrivalent ions do not, however,persist, for what is found is the acetylide ion, which is formed pre-sumably by the combination of pairs of quadrivalent ions with lossof electrons : 2C-----+ CiC-- + 6e.9 In the spontaneous de-composition of ethylsodium at the ordinary temperature, the mainreaction can be represented as due to the transfer of a proton fromone ethide ion of every pair to the other, thus effecting the change2NaEt _I, Na,C,H, + C,H,; at 90-loo", however, the ethideions appear for the most part to lose it proton and an electron pair(hydride ion), the main decomposition being then expressed bythe scheme NaEt + NaH + C,Hp, although some ethaneaccompanies the ethylene.10 Study of the thermal decomposition ofdiethylmercury and tetraethyl-lead yields the interesting resultthat when depomposition is effected in ethylene at 250-300", theethylene polymerises, the initial reaction being most probablybetween the latter substance and the ethyl radicals, C,H, +2C,H, = C4H9 + C,H,, followed by C4H9 + 2C,H, = C,HI3 +C2H4, the radicals being ultimately eliminated by mutual interactionor owing to contact with exposed surfaces.llGrignard Reagents.-There now remains little doubt that inordinary ethereal Grignard solutions equilibria always occur inaccordance with the scheme :2MgRHal += MgRz + MgHal, 12, l3The part played by the ether in the formation of these solutions isdue to its solvent power, not only for the organo-metallic com-pounds, but also for the magnesium halides.12 The position ofCompare H.Clilman and R. E.Brown, J . Amer. Chern. SOC., 1930,52,5045; A,, 1931,206.W. Schlenk, jun., Ber., 1931, 64, [B], 736; A., 1931, 719.W. H. Carothers and D. D. Coffman, ibid., p.1254; A., 1930, 757.lo Idem, ibid., 1929,51, 588; A., 1929, 433.11 H. S. Taylor and W. H. Jones, ibid., 1930,52, 1111; A., 1930, 757.12 W. Schlenk and W. Schlenk, jun., Ber., 1929,62, [B], 920 ; A., 1929, 687.An alternative representation, R2Mg,MgHa12 r-" MgRS + MgHal,, is con-sidered less likely to be correct.1s H. Gilman and R. E. Fothergill, J. Amer. Ohern. Soc., 1929,51, 3149; A.,1929, 1432100 ORGANIC CHEMISTRY .-PART I.equilibrium depends on the nature of the halide : thus, for a numberof simple alkyl bromides and iodides the equilibrium is displacedtothe right withincrease in weight of the alkyl group, and substitutionof bromide for iodide has little effect ; with chlorides, the establish-ment of equilibrium occurs very slowly and is complicated by theseparation of magnesium chl0ride.1~9 15 Progress of the reactionin the backward direction has been demonstrated in the case ofphenylmagnesium halides 13, l4 and it has been found that a verysmall amount of magnesium halide is sufficient to initiate reactionbetween triphenylmethyl and magnesium, and to bring about,owing to the continuous regeneration of magnesium halide, completeconversion into the Grignard reagent thus : 2Ph,C + Mg +MgX, + 2CPh,-MgXBoth the alkylmagnesium halides and the magnesium alkylswhich are present in the Grignard solutions are active reagents, anddoubtless the two types of reaction product, zlix., R,C*O*MgHal andR,C*O*Mg*O*CR, which are produced by the action of Grignardsolutions on aldehydes and ketones, owe their respective origins tothe two types of compound.17 The ratio MgR,/(MgR, + MgRX),which can be determined experimentally with some precision,17is considerably modified (increased or decreased) when themagnesium employed is alloyed or mixed with copper, or whenmetallic zinc or salts such as mercuric chloride are,added to thereactants.But the addition of the latter substances has usuallyanother effect, vix., to suppress to a smaller or greater extent theformation of Grignard reagents and correspondingly to favour theWiirtz reaction (yielding hydrocarbons RnR) which commonlycompetes therewith.151 l8 Copper-magnesium alloy, like iodine,forms an excellent catalyst for starting the Grignard reaction, butin most of the instances examined the presence of either substancein significant proportion seriously reduces the yield of Gignardreagent and correspondingly enhances the yield of Wiirtz reactionproducts.19 The ratio R*R/(MgR, + MgRX) is thus increased by theadded substances, but in the case of the Grignard reagent from ally1bromide is notably decreased (from 72% to 6%) by the influenceof copper.There is no reason to think that the Wiirtz reaction as it(CPh,),Mg + MgX,.16l4 W. Schlenk, jun., Ber., 1931, 64, [B], 734; A., 1931, 718.l5 G. 0. Johnson and H. Adkins, J . Amer. Chem. SOC., 1932,54, 1943; A.,The proportion of active reagent present as MgR, is given as 6% for 728.EtI and 84% for BuCI.l6 W. E. Bachmann, ibid., 1930,52,4412; A , , 1931, 79.l7 W.Schlenk, jun.,Ber., 1931, 64, [B], 736; A., 1931, 719.G. 0. Johnson and H. Adkins, J . Amw. Chem. SOC., 1031, 53, 1520; A.,1931, 719.l9 H. Gilman and E. A. Zoellner, ibid., p. 1581; A., 1931, 719FARMER. 101occurs here results from the action of either type of active Grignardreagent on the original alkyl halide employed, since both types ofreagent have been found to be quite unreactive towards the corre-sponding alkyl halides (ally1 is an exception).15 Grignard reagentscan, however, be converted in smaller or greater measure intoproducts of the Wiirtz-Fittig type by the direct action of variousheavy-metal salts, and it is reported that when an arylmagnesiumhalide is added to a suspension of silver bromide a silver aryl isfirst obtained which decomposes into silver and diary1 when thereaction mixture is boiled.*')Organo-magnesium halides can in some cases be prepared freefrom ether by employing benzene as the medium in the Grignardreaction,21 or by avoiding the use of a solvent altogether.22 Theymay also be formed and undergo reaction in situ when a ketone orester replaces the solvent; in this case, however, exchange ofalkyl groups between the reactants may occur in presence of themetal.21 Certain of the lower alkylmagnesium halides can bedistilled (sublimed) in a stream of ether, but there is nothing toshow whether the organo-halides distil as such rather than as themixture in equilibrium therewith: in high vacuum distillationsmagnesium alkyls are obtained.%The reactive properties of Grignard reagents have been furtherstudied.An important property which has received a large amountof attention is the reductive action which manifests itself when thereagents react with certain aldehydes and ketones. Aldehydes andketones which themselves contain branched chains or are employedwith organo-magnesium compounds containing branched chains oftenbecome more or less completely reduced to the correspondingprimary or secondary alcohols-a circumstance which seriouslylimits the usefulness of the Grignard method for synthesisingsecondary or tertiary carbinols but affords in certain cases a con-venient means of reducing the aldehydes and ketones concerned.24Although this reducing action appears to be directly related to thecomplex character of the alkyl groups present in the reactants, nodefinite and exact correlation has been possible between the amountof reduction obtainable and the degree of complexity or the in-2o J.H. Gardner and P. Borgstrom, J . Amer. Chem. Soc., 1929,51, 3375 ; A . ,19301 76.2 1 W. Schlenk, jun., Ber., 1931, 64, [B], 739; A., 1931, 718.22 H. Gilman and R. E. Brown, J . Amer. Chem. SOC., 1930, 52, 5045; A.,28 Idem, ibid., p. 4480; A., 1931, 78; Rec. trav. chirn., 1929,48, 1133; A , ,24 J. B. Conant and A. H. Blatt, J . Amer. Chem. SOC., 1920, 51, 1227; A . ,1931, 206.1930, 76.1929, 676102 ORUANIC CHEMISTRY .-PART I.cidence of the complex groups in one or other (or both) reactants.25It appears to be the case, however, that for the synthesis of tertiaryaliphatic carbinols containing both st'raight and branched alkylgroups, it is best to employ (a-)branched-chain ketones and magnes-ium n-halides.Whether or no each of the organo-magnesiumcompounds present in the Grignard reagent can , without equilibration,effect reduct,ion has not yet been satisfactorily determined, but it hasbeen found that diisobutylmagnesium, which constitutes 75 yo ofthe reagent prepared from isobutyl bromide, reduces benzophenoneto benzyhydrol to the e'xtent of at least 64%, and the Grignardreagent before separation effects reduction to the extent of 84%.26Kohler and his collaborators, in examining the behaviour of keto-oxido-compounds towards Grignard reagents, showed that benzyl-ideneacetophenone oxide, PhCH-CH*COPh, with phenyl-magnesium bromide yields triphenylcarbinol, Ph,C*OH.27 Herethe first step is probably the scission of the benzoyl group as benzo-phenone, which is then transformed into the carbinol.With mostlieto-oxido-compounds a similar result is obtained and manyrelatively heavily phenylated esters suffer fission under the actionof the Grignard reagent. With benzylidene-p-phenylacetophenoneoxide, however, two reactions occur, in one of which both the ketonegroup and the oxide ring are attacked, yielding (I) or its isomeride,and in the other a preliminary scission of phenyl p-diphenylyl ketoneis apparently followed by the reduction of the latter to a pinacol (II).2810-Jp-Substituted ally1 bromides, the structure of which permits ofanionotropic change, can react with Grignard reagents t o giveisomeric olefins (I11 and IV), which are derived by combination ofthe alkyl group from the reagent with the tautomeric forms of theunsaturated kation ; 29 in other cases they react to give hydrocarbons(V and VI), which appear to owe their origin t o a funct,ional ex-25 A.H. Blatt and J. F. Stone, jun., J. Anter. Clzem. SOC., 1932, 54, 1495;A , , 598.26 C. R. Noller, {bid., 1931, 53, 635; A., 1931, 473.27 E. P. Kohler, N. K. Richtmyer, and W. F. Hester, ibid., p. 205; A., 1931,28 E. Bergmenn and H. A. Wolff, ibid., 1932,54,1644 ; A., 616.354.C. Pr6vost and J. Daujat, Bull. SOC. chim., 1930, [iv], 47, 688; A., 1930,1168FARMER. 103change between the reactants : R*CH:CH*CH,Br + MgEtBr -+R*CH:CH*CH,*MgBr + EtBr.30 The Wiirtz reaction is found toaccompany the latter type of reactivity, and both here,30 and wherethe substituted ally1 bromides react directly with magne~ium,~~the coupling of the hydrocarbon radicals may take place in one ormore of three ways yielding substituted hexadienes (VII, VIII, andIX).8 CBR*CH:CH*CH, =+= R*CH*CH:CH,~ -~R*CH:CH-CH,R' R*CH:CH*CH, [R*CH:CH*CH,*];(111.) (V.R*CH:CH-qH,R*CHR'*CH:CH, R;CH2*CH:CH2 CH,:CH*CHR(IV.1 (VI.) [CH,:CH*CHR*],(VII.)(VIII.)(=.IGrignard reagents react with arsenic trichloride and antimonytrichloride t o give trialkylarsines 32 and trialkylstibines respectively,and with a-tert.-amino-nitriles t o yield amino-ketones in additionto amino-hydrocarbons in which the nitrile group has been replacedby the alkyl radical of the Grignard reagent : 33Reference may also be made t o new studies of the action of Grignardreagents on dialkylamides, R1*CO*NR2R3,34 a-~yano-esters,~~carbonic esters 36 and sulphonyl chloride^.^'Compounds of Boron.Boron trifluoride reacts with simple organic compounds ofdifferent types to give addition compounds, and as early as 1891Patein obtained a solid product from acetonitrile and boron trifluoride.The manner of formation and the consbitution of these compounds30 C.PrBvost, Bull. SOC. chirn., 1931, [iv], 49, 1372; A., 1932,41.31 C. Prbvost and G. Richard, ibid., p. 1368; A., 1932,40.33 W. J. C. Dyke and W. J. Jones, J., 1930, 2426,463; A., 1931, 77; 1930,587.33 T.S . Stevens, J. M. Cowan, and J. MacKinnon, ibid., 1931, 2568; A.,1931, 1404.34 (Mlle.)M. Montagne, Ann. Chirn., 1930, [XI, 13,40; A., 1930,460; Compt.rend., 1931, 192, 1111; A., 1931, 831; S. P. Ti, ibid., 1930, 191, 943; A.,1931, 77.35 A. Mavrodin, ibid., 1930, 191, 1064; 1931, 192, 363; A., 1931, 205, 471.36 D. Ivanov, ibid., 1929,189,830; 1931,193,773; A., 1930,61; 1932,43.37 H. Gilman and R. E. Fothergill, J. Amer. Chem. SOC., 1929,51,3501; A.,1930, 462104 ORGANIC CHEMISTRY.--PART I.have recently been studied independently by different workersand it is now apparent that, when boron trifluoride is passed intofatty acids, the esters of fatty (and certain other) acids, ethers,amides, and nitriles, simple addition products are quite frequentlyproduced.Esters such as methyl and ethyl formate, methyl, ethyl and propylacetate, ethyl propionate and methyl benzoate, give distillableliquids or solids which have the general constitution R’*CO,R”,BF, ;acetic acid and propionic acid yield liquids which are distillable withsome decomposition under reduced pressure, and appear to havethe composition (CH,*CO,H),,BF, and (Et*CO,H),,BF, respectively ;acetonitrile yields a distillable solid product, MeCN,BF,, and aceticanhydride a solid Ac,O,BF,.Methyl and ethyl alcohols give withboron trifhoride solutions of high conductivity, due to the formationof addition compounds, R*OH,BF,, which ionise to yield hydrionand the corresponding complex organic ions ; the higher alcohols,however, suffer decomposition to hydrocarbons.Anisole andphenetole give undistillable liquid addition products with borontrifluoride.All the above compounds appear to be capable of formulation inthe manner \O-B-F, in which the link between boron andoxygen is a co-ordinat,e link or semipolar double bond. The additionproduct of boron trifluoride with alcohol has been used as a catalystin converting acetylene into acetals39 and in the esterification ofacetic and propionic acids wit,h various alcohols ; in the former casethe reaction is assumed to be due to the addition of the ions of thecomplex to the hydrocarbon :CHiCH + 2H@ + 2[RO+BF3]- ---+ CH,*CH(OR-+BP,), +x + -/FR/ \FCH,*CH(OR), + 2BP,Boron trichloride 4O reacts vigorously with methyl and ethylalcohols at low temperatures to give quantitative yields of low-melting, volatile alkoxy-compounds, BCI,*OR, BCl( OR),, andB(OR), (R = Me or, Et) ; these are readily hydrolysed by water oralcoholysed (if they contain chlorine) with excess of alcohol.Boron38 H. Bowlus and J. A, Nieuwland, J . Amer. Chem. SOC., 1931, 53, 3835;A., 1931, 1404; G. T. Morgan and R. Taylor, Chem. and Ind., 1931, 869; J.,1932, 1497 ; A., 1931, 1404; 1932, 728.39 J. A. Nieuwland, R. R. Vogt, and W. L. Foohey, J . Amer. Chem. SOC.,1930, 52, 1018; A., 1930, 745; H. D. Hinton and J. A. Nieuwland, ibid.,1930,52, 2892; 1932,54, 2017; A., 1930, 1160; 1932, 728.4 0 E. Wiberg and W. Sutterlin, 2. anorg. Chem., 1931, 202, 1, 22, 31, 37; A.,1932, 258FARMER.105trichloride and the monoalkoxy-derivative BCl,*OR react withmethyl or ethyl ether even at - 80" to form the additive compounds,BC13,R20 and (BCl,*OR),,R,O, but the di- and tri-alkoxy-derivatives,BCl(OR), and B(OR),, do not react therewith even at 100". Theaddition product BCl,,Et,O reacts vigorously with alcohol at- 40" t o yield the compound (BCI,*OEt),,Et,O, which readilydecomposes into the compounds BCl,,Et,O and BCQOEt),. Thethermal decomposition of these substances has been studied.Tri-tert.-butylboron, which has been obtained by the action ofboron trifluoride on tert.-butylmagnesium chloride in ether, is con-verted on oxidation into tert.-butylboric acid. Tri-sec.-propyl-boron resembles the tert.-butyl deri~ative.~~OleJinic, Di-oleJinic, and Acetylenic Compounds.Additive Reactions.-The various factors which enter into con-sideration in connexion with the problems of orientation in olefinicand polyolefinic additions were referred to in last year's Report.The new work on this subject refers mainly to the influence of thegroups present in unsaturated substances on the orientation oftheir addition products, but it is becoming increasingly evidentthat the influence of the reaction conditions (especially conditionsrelating to solvent and to catalytic influences) is of considerableimportance in determining the course of reaction in certain typesof addition.All the evidence available from the literature goes to showthat in the addition of an unsymmetrical molecule XY t o anethylenic hydrocarbon R1CH:CKR2, the ultimate distribution ofthe addenda1 components X and Y, is largely determined by theinfluence of the groups R1 and R2 on the direction of olefinic polar-isation.In general both of the compounds R1*CHX*CHYR2 andR1CHYCHXR2 are likely to be formed, but as yet there is littlein the way of accurate quantitative evidence to show how com-pletely the results of addition can be correlated with the characterof the groups R1 and R2, or how far the observed orientation ingiven examples h independent of the conditions of reaction andof the nature of the addendum XY. With respect to the mannerin which alkyl groups influence the orientation at a double bondR. Robinsone has pointed out that it is doubtful whether theinductive displacements which are usually attributed to alkylgroups (represented by CH+, C2H6+, etc.) can lead directlyto reactivity, the reason being that only unshared electrons can4 1 E.Krause and P. Nobbe, Ber., 1931,64, [B], 2112; A., 1931, 1280.42 " Outline of an Electrochemical (Electronic) Theory of the Come ofOrganic Reactions," p. 16.D 106 ORGANIC CHEMISTRY .-PART I.take part in co-valency formation, and inductive displacements donot appear t o modify the extent of sharing in a bond from thepoint of view of the numbers of quantised electrons associatedwith the atoms in question, although they admittedly disturb t,hesharing electrostatically. Rather must reaction be preceded bya degree of electronic polarisat,ion (symbolised by C==C) in whichelectxons actually dissociate themselves from one of the carbonatoms and become free to be shared with external atoms-thusproviding the means by which anionoid olefinic molecules react.The important effect of alkyl substitution would seem to beclearly shown by a series of additions t o Aa-, AB-, and Ay-unsatur-ated acids recently carried out.43 The extent to which the additionof hypochlorous acid t o each of three isomeric hexenoic acidsoccurs in opposite directions is shown in the scheme :Me*CH:CH*CH,*CH,*CO,HnlC1 ......OH 95%.OH. ..... C1 5%.Et*CH:CH*CH,*CO,HCI...... OH 20%.OH ...... C1 80%.Pra*CH:CH*CO,HOH ...... C1 100%.Here on the basis of the order of inductive effects usually acceptedfor the lower alkyl groups (viz., Pr>Et>Me>H), the orientationsobserved experimentally are completely in accord with expectation,provided it is assumed that the substitution of a carboxyl groupin an alhyl group, as instanced by *CH,*CH,*CO,H and *CH,*CO,H,does not diminish-or diminishes very slightly-the inductive effectof the alkyl group, and that the carboxyl group itself, when attacheddirectly t o an olefinic linkage, possesses little orienting influencecompared with alkyl, or none a t all..Under the conditions of the above reaction and with hypo-chlorous acid as the addendum it is doubtless true that the orientinginfluence of a carboxyl group in the system CH:CH*CO,H is muchsmaller than that of even a methyl group.44 But no general con-clusion can be drawn from the above-cited facts as to the efficacyof the carboxyl group as an orienting influence, either in the case(a) represented by the group *CH:CH*CO,H, in which a conjugated43 G.F. Bloomfield and E. H. Farmer, J., 1932, 2062; A., 930.44 Comparisons (unpublished) made by the Reporter, together with Dr.Bloomfield and Mr. C. Hose, between the orientations assumed by the compo-nents of hypochlorous acid in the systems Mc.CH:CH.CO,H, MeCH:CMe-CO,H,CH,:CMe-CO,H, and CH,:CMe*CO,Et show how important the influence ofthe methyl group can be, but also how materially the carboxyl group canaffect the orientationFARMER. 107system is present, or in the case (b) represented by the system*CH:CH*[CH,],CO,H, in which the carboxyl group is removed toa distance from the ethylenic centre.For in case (a) there is everyprobability that different addenda react in different ways : thushypochlorous acid, being a reagent for the ethylenic linkage p e r se,and having no tendency to add via the carboxyl group, is probablycomparatively little influenced by the latter in respect of the orient-ation assumed ; but addenda such as the hydrogen halides, althoughthese are also reagents for the ethylenic bond p e r se, appear to addpreferably at the ends of the conjugated system, the addition beinginitiated by the attachment of the proton to carbonyl oxygen.Similarly in case ( b ) it has recently been argued 2t44Q that, although inexamples such as the hydration of stearolic acid (X) 45 the observedorientation is improbably due to the influence of the carboxylgroup, transmitted from atom to atom along the carbon chain,CH,fCH,],*CiC*[CH,],*CO,H CH,:CH*[ CH,] ,*CO,HBr..... .H (in toluene) 20H. ..... 2H (57.6%)(X.) 2H ...... 20H (42.4%) H ...... Br (in ether) (XI.)yet it may be due t o the transmission.of this influence through theintervening space (“ field effect ”) : but here a further complicationenters, in that the medium may be such as to transmit the influencereadily or to dissipate it-a state of affairs which has been con-sidered t o receive illustration in hydrogen bromide addition toAc-undecylenic acid (XI),46 where the change from a toluene to anether medium completely reverses the mode of addition.Whetheror no the field effect has the importance thus attributed to it awaitsfuture verification, but there are not wanting indications that thecondition of the carboxyl group-ionised or un-ionised-an havea strong determining influence on orientation, so that it is quiteimprobable that an absolute mode of addition can usually beassigned to a mono-olefinic acid, irrespective of the experimentalconditions and of the character of the unsymmetrical addendum.Ammonia, methylamine, and diethylamine add readily to ethylcrotonate, giving the corresponding p-amino- or p-albylamino-butyric Beaction proceeds at room temperature andliquid ammonia or alcoholic ammonia may be employed. In thecase of ammonia ethyl Pp’-iminodibutyrate is also formed.A study of the lactonisation of A=- and’AB-unsaturated acids has445 R.Robinson, ‘‘ Outline of an Electrochemical (Electronic) Theory of the4 5 (Mrs.) G . M. Robinson and R. Robinson, J., 1926, 2204; A , , 1926, 1024.4 6 J. Walker and J. S . Lumsden, &id., 1901,79, 1191.4 7 K. Morsch, Monatsh., 1932,60, 50; A., 600.Course of Organic Reactions,” p. 32108 ORGANIC CHEMISTRY.-PART I.shown that all acid-y-lactone systems can be interpreted in onegeneral scheme : 48I I I -CH-CC-CO~OH &= ~=C~-&H-CO.OH 4 -+-C~H-#H0-coy-lactoneLactonisation represents an addition reaction in which the com-ponents of the carboxyl group, H and RCO.0, become attachedto the olefinic centre; consequently, analogies are to be expectedwith other additions in which the addendum is of the type HX,but here special limitations are imposed by the bound condition ofthe anionic component of the addendum.Originally R. Fittigand his collaborators had maintained that only AB-unsaturatedacids and allylacetic acid (the only Ay-acid then known) could beconverted by boiling 50% sulphuric acid into y-lactones. Thisgeneralisation, however, is incorrect, since simple Aa-acids giveappreciable quantities of lactone under Fittig’s experimental con-ditions. The variations between different systems can be attributedt o differences in the ratio of the velocity of tautomeric change(a and b ) to that of ring closure (c). Three, possibly four, types ofacid can be distinguished : (i) those in which both changes areslow, but lactonisation is much faster than tautomeric change (acidswith one y-alkyl group and no P-alkyl substituent); (ii) those inwhich lactonisation is fast and tautomeric change slow (acids withtwo 7-alkyl substituents) ; (iii) those in which tautomeric changeis faster than lactonisation (acids with one y- and one P-substituent) ;(iv) ( ? ) those in which tautomeric change is fast and irreversiblein the direction pr --+ ap, no lactonisation being possible (acidswithout y-substituents). No evidence of the re-formation of aAs-acid from a simple lactone by treatment with sulphuric acid orby prolonged heating has been discovered : consequently lacton-isation is formulated as irreversible, but the change, lactone --+ A=-acid, appears to be possible in more complicated systems such aslactonic acids of the paraconic type.The manner of addition of hydrogen chloride to conjugatedolefin-acetylenes resembles that of hydrogen bromide to the con-jugated b u t a d i e n ~ .~ ~ In the latter case the hydrogen of theaddendum always attaches itself (so far as existing evidence shows)its a proton to a terminal carbon atom of the chain, whilst thebromide ion becomes linked at either the second or fourth carbon48 R. P. Linstead, J., 1932, 115; A., 251.49 W. H. Csrothers, G. J. Berchet, and A. M. Collins, J . Arner. Chern. SOC.,1932,54, 4066 ; A., 1231 ; W. H. Carothers and D. D. Coffman, ibid., p. 4071 ;A., 1232FARMER. 109atom. With vinylacetylene (XII) and the isopropenyl-, isobutenyl-and cyclohexenyl-acetylenes represented by f ormulz (XIII), (XIV),and (XV), the hydrogen component of the addendum attaches itselfto the terminal acetylenic carbon atom, whilst the chlorine com-ponent becomes linked at the fourth carbon atom in the case ofvinylacetylene, and at the second in the remaining instances.CH,:CMe*CiCH (XIII.)(XVIII.)The formation of the allene derivative (XVI) is particularly inter-esting.This substance can be isolated as the major product undercertain conditions of reaction, but in the presence of hydrogenchloride it undergoes isomerisation to p-chlorobutadiene so readilythat the latter substance always constitutes part of the reactionproduct. Certain salts reinforce the catalytic effect of hydrogenchloride, and when cuprous chloride is present none of the allenederivative survives.p-Chlorobutadiene has been designated" chloroprene " in analogy with isoprene, and has been made thestarting material for the production of synthetic rubber. Chloro-prene adds a further molecule of hydrogen chloride, yieldingay-dichloro- AP- butene.In the addition of hypochlorous acid to sorbic acid and to p-vinyl-acrylic acid the main influence of the carboxyl group is to decidewhich of the two double bonds of the conjugated system shall beattacked. The ap-unsaturated centre becomes de-activated, andthe anionic charge develops at the %carbon atom. Attachment ofthe chlorine component of the addendum then occurs at the &carbonatom in the manner : 50CH,*CH:CH*CH:CH*CO,H + CH,*CHCl*CH( OH)*CH:CH*C02HCH;CH*CH:CH*CO,H --+ CH2C1*CH( OH)*CH:CHC02HR1R2C:CR3*CR4:CR5*COzHreact with hydrogen in the presence of a platinum catalyst inSorbic acid and its homologues of the series5O G. F.Bloomfield and E. H. Farmer, J., 1932,2072; A,, 930110 ORGANIC CHEMISTRY .-PART I.different ways. 51 At room temperature and atmospheric pressurefour types of addition take place, wiz., ccp-, as-, y&, and orpyS-, thelast of these representing unpreventable attack at both ethyleniccentres simultaneously. The different acids of the series yieldmixtures of the different types of addition product, but examin-ation of the reaction mixtures after partial hydrogenation showsthat hydrogenation follows a different course for each acid; theconstitution of the conjugated compound ( i .e . , the character ofthe substitution in the chain) appears definitely, therefore, to bereflected in the additive mode. It is found, however, that thefigures representing the proportions of the various reduction pro-ducts a t the intermediate stages of reduction are by no means tobe regarded as solely determined by, or affording a true measureof, the structural influences at work in the individual acids; for,although they are reproducible if the catalyst is freshly preparedfrom platinum oxide, when a less active (" aged ") catalyst is usedthe course of hydrogenation (in the case of sorbic acid a t least) 52becomes considerably changed. Thus the nature and extent ofsubstitution in the butadiene chain is not, under the conditions oftemperature and pressure employed, the sole, or apparently themost important, influence in determining the course of reaction.The substitutional or constitutive influence, with its activating ordeactivating tendencies, appears t o be superimposed on a specificcatalytic influence which is capable of activating both unsaturatedcentres of the conjugated system simultaneously.Striking observations respecting selective hydrogenation in tlicnon-conjugated (diolefinic) linoleic acid series are due t o T.P.Hilditch and E. C. Jones.53 I n hydrogenating the unsaturatedglycerides contained in olive and cottonseed oils with a nickel-kieselguhr catalyst it is found that the linoleic groups are verylargely converted into oleic and isooleic groups before the latterare further hydrogenated ; after the linoleo-glycerides have dis-appeared, steady increase in the total saturated acid commences,but development of fully saturated glycerides is relatively slow untilthe final stages of hydrogenation.Further it is found that trioleins(mixtures of oleic and isooleic triglycerides) at first disappear muchmore rapidly than fully-saturated components are produced, indicat -ing that a molecule of triolein, adsorbed by nickel, is desorbed assoon as a single oleic group has undergone hydrogenation. Thusdirect transformation of molecules of trioleh into trktearin a t oneand the same contact with the catalyst does not occur and a singleunsaturated centre only is involved in each effective contact between51 E.H. Farmer and R. A. E. Galley, J., 1932, 430; A., 365.5 2 Idem, Nature, 1933, 181, 60. 53 J., 1932, 805; A., 498PARBEER. 111catalyst and 8r triolein molecule; in this way the hydrogenation ofMerent classes of unsaturated glycerides is definitely selective,the order of reduction being trioleins, di-oleo-compounds, andmono-oleo-compounds.Formudion of Dimerides, Trimerides, and Tetrarnerides.-Olejins.To the list of dimerides and trimerides (including XX-XXIIIbelow), the constitutions of which have been definitely establishedduring the past three years, must now be added dimeric indene(XXIV).541.2.3.4.5..AH2C=CMe2 + H+CH=CMe, +Hz&h2 + H+CH=CPh, -+H,eCHEt + H+CH=CH, jCH3*CMe2*CH:CMe2 (xx.)CH3*CPh2*CH:CPh2 (XXI.)CH,*CHEt*CH:CH, (XXII.)0H,C=CMe2 4- H+C(CMe3)=CMe2 4 CH3*CMe2-C(CMe3):CMe2(XXIII.)The self-addition of mono-olefins, whether of the same or of differentmolecular species, appears to be jointly dependent on (a) thecapacity of a hydrogen atom in one of the reactant molecules (theaddendum) to suffer incipient ionisation in the manner H+C,=C,,and (b) the occurrence of successful polarisation at t'he double bondof the other.55 The process (a) is promoted by electromeric changesin the opposite (usually less-favoured) direction to those utilisedin (b), and polymerisation can only occur to the extent that (a) isachieved; the addition product then retains the double bond ofthe addendum molecule, according to the generalised scheme :s+ s-s - n a+ S+ 8 -CHR1=CR2R3 + H+ CR4ZCR5R6 -+ CH2R1*CR2R3*CR*=CR5R6The formation of addition products by the interaction of styreneand different benzene hydrocarbons in presence of sulphuric54 E.Bergmann and H. Taubadel, Ber., 1932, 65, [B], 463; A., 507.Compare G. S . Whitby andM. Katz, J. Amer. Chem. Soc., 1928,50, 1160; A,,1928, 627; Canadian J . Res., 1931,4, 344; A., 1931, 833.6 5 Compare Ann. Reports, 1930,27,91, 92; 1931,28, 87; E. Bergmann andH. Weiss, AnwZen, 1930,480,49 ; A., 1930,901 ; E. Bergmann, H. Taubadel,and H. Weiss, Ber., 1931, 64, [B], 1493; A., 1931, 945112 ORGANIC CHEMISTRY.-PART I.acid 56 (e.g., CHPh:CH, + C,H, --+ CHPh,*CH,) appears to be ananalogous process.Acetylenes.Acetylene polymerises catalytically at low tem-peratures in the presence of ammonium and cuprous chlorides toproduce a dimeride (vinylacetylene), a trimeride (divinylacetylene),and a tetrameride (probably CH,:CH*CiC*CH:CH*CH:CH,). Thedimeride can definitely yield the tetrameride by self-addition underthe conditions of the reaction and appears to yield the trimerideby combination with a~etylene.~' Polymerisation is limited tocompounds containing the group GCH, consequently the trimerideand the tetrameride are the ultimate products of the reaction.The process is of the same general character as that occurring inthe polymerisation of the mono-olefins and it is noteworthy thatthe reaction ( A ) proceeds in preference to (B) :(A) CH,:CHGC-H + CHiCH + CH,:CH*CiC*CH:CH,(B) CH,:CH*CiCH + H-CiCH + CH,:CH*CH:CH*CiCHThe characteristic mode of dimerisation of conjugatedbutadienoid compounds is Closely related to the Diels-Alder reaction.In last year's Report reference was made t o the elucidation of themanner of dimerisation of cyclopentadiene : during the presentyear the manner of dimerisation in two other outstanding instanceshas been determined.The remarkably stable dimeride of cydo-hexadiene is constituted analogously to dicyclopentadiene, 58 andthe hydrocarbon derived by the combined decarboxylation andDioleJins.I;T. H /.\ /9cH-cH\Doebner 'shydrocarbonpolymerisation of sorbic acid (Doebner's supposed tricyclooctane)is now recognised to be o-propyltoluene.59 The latter mode ofreaction is reproduced in the case of cinnamenylacrylic acid andalso of p-vinylacrylic acid, which yield the aromatic hydrocarbons,o-Ph*C,H4*CH,*CH,Ph and ethylbenzene respectively.All dimerisations of butadienoid hydrocarbons in which the con-stitution of the product has been satisfactorily established yield,56 A.Spilker and W. Schade, Ber., 1932, 65, [B], 1686.5 7 J. A. Nieuwland, W. S. Calcott, F. B. Downing, and A. S. Carter, J . Amer.5 8 K. Alder and G. Stein, Annalen, 1932, 496, 197; A., 938.59 R. Kuhn and A. Deutsch, Ber., 1932,65, [ B ] , 43 ; A., 258.Chem. SOC., 1931, 53, 4197; A., 1932, 40FARMER. 113so far as is known, derivatives of cyclohexene. Hitherto, however,there has been little to indicate how the trimerisation and tetra-merisation of butadienoid compounds proceed.The work ofK. Alder and his collaborators on the trimerides of cyclopentadiene 6othrows some light on the subject. Tricyclopentadiene is formed bythe addition of cyclopentadiene to dicyclopentadiene at the doublebond in the bridged cyclohexene ring of the latter.+Dic yclopentadiene Tric yclopentadieneTwo forms of the trimeride are obtained and since the spatialconfiguration of that portion of the carbon skeleton representedby thick lines in the formula has been shown to be exactly thesame in each case, the isomerism is to be attributed to the alternativespatial arrangements that can be assumed by the cyclopentene ringwith respect to the plane of the adjoining ring. It is probable,indeed, that the two forms of the trimeride arise by addition tothe two known forms of the dimeride respectively.Polymerisationcan go still further to the tetrameride stage. Now since the'' mixed " addition products (XXVI) and (XXVII) can be respect-ively obtained by the addition of successive molecules of cyclo-pentadiene to the compound (XXV) (itself derived by the unionof cyclopentadiene with maleic anhydride) and by the addition ofcyclopentadiene and as-diphenylbutadiene successively to (XXV),Ph(XXVI.) (XXV.) (XXVII.)there is strong indication that the more reactive double bond inany butadiene dimeride or trimeride is normally the one in thebridged cyclohexene ring.A study of the polymerisation of eleven simple butadienes underthe influence of heat or of chemical agents shows that dimerides(one or more in each case) are invariably formed.In some caseshigher polymerides are also formed, but the presence of at leastthree unsubstituted hydrogen atoms on the terminal carbon atoms6 O K. Alder, G. Stein, and others, Annalen, 1932, 496, 204; A,, 938.61 G. S. Whitby and R. N. Crozier, Canadian J. Res., 1932, 6, 203; A., 361 ;G. S. Whitby and W. Gallay, ibid., p. 280; A., 496114 ORGAKIC CHEMISTRY.-PART I.appears to be necessary for the formation of a synthetic rubber.The dimerides are themselves polymerised in the cold by certaininorganic reagents (H,S04, SbCl,, SnCI,, etc.), but there is not,hingas yet to show that their formation constitutes an intermediatestep in the formation of normal rubbers.Refractivity, Condensations, and Structural Mobility.-The simplebutadienoid hydrocarbons, when prepared by the best methodsavailable, show considerable variations in boiling point , density,refractive index, and dielectric constant, although when submittedto the Diels-Alder reaction as a chemical test of homogeneity theyshow no signs of marked heterogeneity. These physical differ-ences, which do not occur in examples where geometrical isomerismis impossible, are attributed to the presence of cis- and trans-forms,or in the case of as-dialkylbutadienes, possibly to the presence ofcis-cis, &-trans, and trans-trans forms.The optical properties ofthe mono- and di-methylbutadienes are also discussed, but fordetails the reader is referred to the original paper.6, The opticalproperties of a number of aryl-polyacetylenes have also beeninvestigated and values for the molecular exaltation due to theCPhiC group obtained.63 These values vary considerably accordingt o the solvent employed, but an average value, EM, 3.29, is recorded.The t rammission of electrical influences through conjugatedcarbon systems is well seen from a number of new examples ofwell-known reactivities.The first of these concerns the formationof oxalyl derivatives by the action of oxalic ester on carboxylicesters or ketones in the presence of sodium or potassium ethoxide.AtH H1-3 p x I 0CHZ-CR'=CH-C<ORIT a R'*CH-C<oR(XXVIII.) (XXIX.)RQ,C*CO*CRR'*CO,R RO,C*CO*CH,*CR':CH*CO,R(XXX.)RO,C*CQ*CH,*CH:CH*CH:CH*CO,RThe detachment of a proton from the a-methylene group of a fattyacid ester (XXVIII) a t the instance of the reagent is rendered62 E. H.Farmer and F. L. Warren, J., 1931, 3221; A., 1932, 141.63 V. ICrestinslri and N. Perssianzeva, Ber., 1931, 64, [B], 2363; A., 1931,141FARMER. 115possible by the juxtaposition of the carbonyl group, so that anoxalyl group is enabled to enter the molecule at this point. Thesame effect obtains if an ethylenic grouping, C:C*, is interposedbetween the a-methylene group and the carbonyl group, as wasdiscovered by A. Lapworth G4 in working with p-alkylacrylic esters(XmX), and has now been shown to occur in the case of sorbicester (XXX).65The formation of aldols or their dehydration product>s by theinteraction of unsaturated aldehydes is an analogous reactivity,and from the recently synthesised P-methylcrotonaldehyde 66(2 mols.) the aldol CMe,:CH*CH(OH)*CH,*CMe:CH*CHO and thecorresponding t riene-aldehyde,CMe :CH*CH :CH*CMe :CH*CH 0,can be successively obtained.67aldehyde together give rise to phenylpentadienal,CHPh :CH*CH:CH*CHO, 68whilst crotonaldehyde (2 mols.) yields an unbranched aldol, doubt-less CHMe:CH*CH( OH)*CH,-CH:CH*CHO.69, Acetaldehyde andcrotonaldehyde, when condensed together, give hexadienal,CH,-[CH:CHI,*CHO, which yields with more acetaldehyde, octa-t rienal, CH,*[ CH:CH ],*CHO. 70 The same type of reactivity doesnot, however, extend to the condensation of aldehydes with un-saturated acids.Thus under the conditions of the Perkin reactionno condensation occurs between benzaldehyde, potassium crotonate,and crotonic anhydride, but in the presence of tert.-bases, benz-aldehyde and crotonic anhydride react to give, not 6-phenyl-AQr-pentadienoic acid, but a-vinylcinnamic acid,Probably here the mixed enolic anhydride,is formed intermediately.Likewise, benzaldehyde and croton-CHPh:C( C0,H) *CH:CH,. 71CHMe:CH*CO*O*C( OH) :CH*CH:CH,,A peculiar type of aldehyde condensation which appears to64 J., 1901,79,1276; see also L. Higginbotham and A. Lapworth, J . , 1923,6 5 W. Borsche and R. Manteuffel, Ber., 1932,65, [B], 868; A., 721.66 F. G. Fischer, L. Ertel, and K. Lowenberg, ibid., 1931, 64, [B], 30; A.,1931, 335.6 7 F.G. Fischer and K. Lowenberg, Annakn, 1932, 494, 263; A., 600;compare K. Bernhauer and E. Woldan, Biochem. Z., 1932, 249, 199; A,,834.R. Kuhn and A. Winterstein, Helv. Chim. Acta, 1929, 12, 493; A., 1929,699.123,1325.69 (Miss) I. Smedley, J., 1911, 99, 1627.70 R. Kuhn and M. Hoffer, Ber., 1930, 63, [B], 2163; 1931, 61, [B], 1977;A . , 1930, 1406; 1931, 1273.R. Kuhn and S. Ishikawa, ibid., p. 2347 ; A., 1931, 1413116 ORGANIC CHEMISTRY .-PART I.proceed in the manner of the Diels-Alder reaction is afforded bythe self-condensation of p-methylcrotonaldehyde in the presence ofsodamide. The aldehyde in this case gives rise to t,he cyclic alde-hyde (XXXI), and citral, with the same condensing agent, givesCMe, CMe,CH, C*CHO \\ CH, CHGHO Ch\,H*CHOOH --+ &Xe CMe C<H\Cg \Cd CH (XXXI.)I I I\ /--+ CMe CH I Ian exactly analogous product ; crotonaldehpde, on the other hand,gives a resin.*'Isomeric Change.-Anionotropic changes closely resembling thatwhich occurs during the isomerisation of the a@-dibromide of hexa-triene( CH,Br*?lH*CH:CH*CH:CH, + B: + CH,Br*CH:CH*CH:CH*CH2Br)have been observed to occur when furfuryl chloride (XXXII) andsorbyl chloride (XXXIII) react with potassium ~yanide.'~ InKCN1- CN*CHMe*CH:CH*CH:CH,CHMe:CH*CH:CHCH,*OBu CHMe:CH*CH:CH*CH,Cl(XXXIII.) CHMe( OBu)*CH:CH-CH:CH,these cases the transferen& of the cyanide radical to the remoteend of the conjugated system appears to reach completion, but inthe instance of the cyanide derived from 5-met hylfurfuryl chlorideand in that of the butoxy-compound derived from sorbyl chlorideby the action of silver butyrate a mixture of isomeric compoundsis obtained.Synthesis oj Ethylenic NitriZes.-Much new information is nowavailable concerning the methods of synthesis of cis- and trans-forms of A=- and AB-ethylenic nitriles, including compounds of the72 T.Reichstein, Ber., 1930, 63, [B], 749; A., 1930, 611 ; T. Reichstein andH. Zschokke, Helv. Chim. Acta, 1932, 15, 249; A., 519; T. Reichstein andG. Trivelli, ibid., p. 254; A., 498FARMER. 117~ e n t e n o - , ~ ~ hexeno-,74 i~ohexeno-,~~ h e ~ t e n o - , ~ ~ and i~ohepteno-~'series.Trienes and Tetraenes.The isolation by R. S. Cahn, A. R. Penfold, and J. L. Simonsen 78of an open-chain triene-carboxylic acid from the wood-oil ofCallitropsis araucurioicles has directed attention to the synthesisof conjugated triene acids.The acid of Simonsen and his collabora-tors, which contained one double bond more than geranic acid,absorbed three mols. of hydrogen on catalytic hydrogenation to givethe already-known dl-tetrahydrogeranic acid. It was therefore adehydrogeranic acid containing the carbon skeleton of geranicacid, and therefore one of seven triene isomerides possessing thenecessary skeleton and differing from one another only in the positionof the double bonds. The most probable formula for the acid wasthe fully-conjugated one, CMe,XH*CH:CH*CMe:CH*CO2H, in spiteof the fact that no addition product with maleic anhydride had beeno b t aim ble .Now the introduction of a methyl group into the p-position duringthe synthesis of a conjugated acid can usually be achieved by sub-mitting such a ketone as mesityl oxide to the Reformatsky reaction ;but here the ketone requisite for the production of a terminal iso-propylidene group in a conjugated chain was that homologue of thewell-known crotylideneacetone which should be capable of synthesisfrom F.G. Fischer, L. Ertel, and K. Lowenberg's P-methylcroton-aldehyde.CMe,:CH*CH:CH*CO*CH,,by Fischer and Lowenberg was followed by its submission to theReformatsky reaction with bromoacetic ester. Thus was obtaineda hydroxy-ester, CMe,:CH*CH:CH*CMe( OH)*CH,*CO,Me, from whichby dehydration and hydrolysis a solid dehydrogeranic acid identicalwith Cahn, Penfold, and Simonsen's acid was derived.79 Thisacid was accompanied by an oily acid which apparently representedThe successful synthesis of this ketone,7 3 P.Bruylants and G. Jmoudsky, Bull. Acad. roy. Belg., 1931, [v], 17,1161; A., 1932, 257.74 P. Bruylants and L. Ernould, ibid., p. 1174; A., 1932, 258; R. A. Letchand R. P. Linstead, J., 1932, 443; A., 371 ; P. Bruylants, Bull. SOC. chim.Belg., 1932, 41, 309; A., 1119; A. Dewael, ibid., p. 318; A., 1119.75 J. Baerts, ibid., p. 314; A., 1119; P. Bruylants, Bull. Acad. TOY. Belg.,1931, [v], 17, 1008; A., 1931, 1403; P. Bruylants and L. Ernould, ibid., p.1027; A., 1931, 1403.7 6 P. Bruylants, Bull. SOC. chi,m. Belg., 1932, 41, 333; A., 1119.7 7 G. Festraete, ibid., p.327 ; A., 1119.7 8 J., 1931, 3134; A., 1932, 144.79 Annalen, 1932,494,263 ; A., 600118 ORUANIC CHEMISTRY.-PART I.or contained a geometrical isomeride of the first, since it gave di-methylheptatriene on decarboxylation.The same synthesis was effected independently by R. Kuhn andM. Hoffer,so likewise employing the diene ketone derived fromP-methylcrot onaldehyde. Here, however, the Ref ormat sky productyielded, when dehydration was effected before hydrolysis, a mixtureof Cahn, Penfold; and Simonsen's acid (m. p. 186") and a solidisomeride (m. p. 137") thereof; but when hydrolysis and dehydra-tion were effected in one operation, only the lower-melting isomeridewas obtained. The two acids show almost identical absorptionspectra and doubtless differ in the geometrical configuration aboutthe ap-double bond.I h h n and his collaborators draw attentionto the approximately constant difference in melting point betweenall such pairs of trans- and cis-isomerides belonging to the mono-olefin, diene, and triene series of monocarboxylic acids respectively.Triene acids analogous to the above but containing the ring ofionone have also been derived by submitting p-ionone to the Re-formatsky reaction with bromoacetic ester and hydrolysing the trieneester (XXXIV) thereby immediately obtained.81 The acids (solidand liquid) in this case also were probably geometrical isomerides.A related synthesis has been effected by treating a-ionone withallylmagnesium bromide, the product being the hydroxy-hydro-carbon (XXXV), from which the corresponding trimethylcgclo-hexenylmethylhexatriene is obtained on dehydration.The alcohols corresponding to Kuhn and Hoffer's n-octatrienaland n-decatetraenal have been obtained by reducing the aldehydeswith aluminium isopropoxide.Both n-octatrienol,CH, *CH :CH*CH: H,HO*CH,*CH: 8 Hand n-decatetraenol, CH,*CH:CH*CH:QH, like sorbyl alcohol, arccrystalline solids unstable in air.s2HO*CH,*CH:CH*CH:CHso Ber., 1932, 65, [B], 651 ; A., 600; see also R. Kuhn and H. Roth, ibid.,81 P. Karrer, H. Salomon, R. Morf, and 0. Walker, Helv. Chim. Acta, 1932,82 T. Reichstein, C. Ammnnn, G. Trivelli, and others, ibid., p. 261;p. 1285; A., 1111.15, 878; A., 852.A., 496FARMER. 119Several interesting considerations arise out of the observed be-haviour of polyene-carboxylic acids on reduction.As was shownby J. T. Evans and E. H. Farmer,83 the reductive behaviour ofsorbic acid when treated by metals in aqueous media varies with theacidity or alkalinity of the medium, but both aP- and as-dihydro-genated (A?- and AP-dihydro-)products are always formed. Asimilar production of isomeric dihydro-acids by a@- and terminaladdition (possibly also by as-, etc., addition in the case of trienesand higher polyenes) is generally to be expected with all conjugatedpolyene-monocarboxylic acids, although theoretical considerationsindicate that the proportions must vary with the character of thesubstitution in the polyene chain and in some cases one of theforms may be absent, or present in but small amount, or may even beso unstable as to suffer double-bond migration immediately afterproduction.Although in the sorbic acid series dual modes of re-duction have been shown to apply generally,a4 experiments by R.Kuhn and A. Winterstein on the reduction of aw-diphenylpolyeneshave given only terminally additive products ; 85 also from the (mono-meric) reduction products of the triene- and tetraene-acids,CH,*[CH:CH],*CO,H and CH,*[CH:CH],*CO,H, only terminaladdition products have as yet been isolated.86 A noteworthypoint here, however, is that the Akl.ihydro-derivative of thetriene acid suffers transference of the conjugated system (as awhole) towards the carboxyl group when heated with alkali underthe usual conditions of the py,ap-change :CH,*CH,*[CH:CH],*CH,*CO,H --+ CH3*CH,-CH2* [CH:CH],*CO,HIn these reductions, as in those of the ao-diphenylpolyenes andindeed of all other reductions of polyene systems so far described,there has been no indication of the production of geometrically iso-meric polyene-reduction products.In all terminal additions thedistinction between cis- and trans-forms appears to be obliteratedduring the reaction, for reasons which are quite intelligible havingregard either to electronic or to Flassical representations of thevalency changes which occur in the processes; 87 moreover, so83 J . Soc. Chern. Ind., 1928,47,268; J., 1928, 1644; A., 1928,868.84 Ann. Reports, 1930, 27, 87.85 Helv. Chim. Acta, 1928,11, 123; A., 1928, 281.8 6 R.Kuhn and M. Hoffer, Ber., 1932, 65, [B], 170; A., 365.8 7 Compare E. H. Farmer and W. M. Duffin, J., 1927,405; A., 1927, 448;E. H. Farmer, B. D. Laroia, T. M. Switz, and J. F. Thorpe, ibid., p. 2937; A.,1928, 151; E. H. Farmer and F. L. Warren, J., 1929, 897, 901; A., 1929,812120 ORGANIC CHEMISTRY.-PART I.far as present evidence shows, the resulting‘ configurations aboutthe individual double bonds are all of tram-character.s8Pol yenes .Bixin and Crocetin.-The foregoing consideration respectingthe geometrical form of reduction products is of importance inconnexion with the inter-relationship of bixin and p-bixin, the latterof which can be converted into the former by the addition and re-moval of iodine. The probability that the two substances arecis-tram‘ isomerides seemed somewhat diminished by the fact thatQH:CH*8:CH*CH:CH*~:CH*CH:CH*C! :CH*CH:CH*F:CH*CH:?HC0,Me Me Me Me Me CO,HQH:CH*Q:CH*CH:CH*8:CH*CH:CH*CH:Q *CH:CH*CH:v*CH:rHCO,Me Me Me Me Me CO,Hthe formation of isocarotin from P-carotin, which likewise occursby the action of iodine, involves no cis-trans rearrangement : butin the latter case strictly stoicheiometrical proportions of iodine arefound to be necessary for the conversion, whereas in the former onlya trace of iodine is required.The two bixins give on reductiondihydro-compounds which are identical and of which the methylesters are identical: hence it appears that they are geometricalisomerides, but there is no indication as to whether only one, ormore than one, cis-unit is present in the molecule of bixinAnalysis of the absorption spectra of the various compoundsstrongly indicates that dihydrobixin arises by the terminal additionof hydrogen to the parent compounds, and in such a process geo-metrical considerations indicate that homogeneous trans, trans... . . -and cis, cis.....-systems ( A and C respectively) can alike pass intohomogeneous trans, trans.. . .-chains.Bixin (Kuhn and Winterstein, 1925) (XXXVI.)Bixin (symmetrical formulation) (XXXVII.)The case of j3B’-diphenylmuconic acid, ho-ever, appears a t present toafford an exception (Farmer and Duffin, Zoc. cit.). One of the two knowngeometrically isomeric forms of this acid yields the cis-dihydro-acid on reduc-tion with sodium amalgam, and the other a mixture of cis- and trane-dihydro-acids.This result may mean that the addition (owing t o the influence of thephenyl groups) does not take place terminally, or that certain types of sub-stitution can exert a determining influence on configuration during the passagefrom the intermediate (non-double bonded) stage of reduction to the final stagea$ which reducocl prodacts of definite reomotrical form appearFARMER. 121If trans-linkings usuahy occurred in the form (D) [obtainable from(A) by rotation about the single bonds], the appearance of cis-linkings in the building up of dihydro-polyenes might be geometricallyconceivable ; but such configurations in conjugated compounds-atleast in the crystalline condition-have not yet been observed andthe X-ray analysis of trans, tra ns....-p olyenes agrees better with theform (A).89 At present there is no reason to assume, as has formerlybeen done for glutaconic acid and muconic acid, that in long-chain polyenes all distinctions between cis- and trans-configurationsabout the individual double bonds disappear, and the revival byD.Riidulescu90 of electronic formulae (other than as reactionformulae) in which distinctions between cis- and trans-configurationsdisappear 91 has no experimental justification or demonstratedprobability-although the synthesis, for instance, of the 72 geo-metrical isomerides of 1 : 14-diphenyltetradecaheptaene seems asomewhat remote event.The reported non-equivalence of the carboxyl groups in bixinmay be due therefore to a single cis-linking placed unsymmetricallyin the chain, and the four methyl substituents, the positions of whichhave not yet been determined, may well occupy symmetrical positionsin the chain as in (XXXVII),s9 rather than as in (XXXVI).Ac-cording to (XXXVII) bixin would represent the middle fragmentof the lycopene molecule as formulated by P. Karrer and hiscollaboratorsg2 and it is possible that bixin is derived fromlycopene, or a polyene closely related thereto, by oxidativedegradation.A remarkable feature of dihydromethylbixin and of the di-methyl ester of crocetin is that these substances are smoothly de-hydrogenated to p-methylbixin and to the dimethyl ester of crocetinrespectively by secondary and tertiary amines in the presence ofair ; 93 both substances, moreover, yield vivid colour reactions whentreated with caustic alkali in pyridine, the coloured solutions(emerald-green and indigo-blue respectively) yielding the dehydro-genated products when shaken with oxygen.This behaviour isrelated to the capacity of glutaconic ester, As-dihydromuconicester, and other allied substances' to give deep yellow or red sodio-derivatives with alcoholic sodium ethoxide, or with caustic soda-89 R. Kuhn and A. Winterstein, Ber., 1932,65, [.€?], 646; A., 618.90 Ber., 1931, 64, [B], 2223; A., 1931, 1351.91 Compare G. Wittig and W. Wiemer, Annalen, 1930, 483, 144; A.,93 Helv. Chim. Acta, 1930,13, 1084; A., 1930, 1422.93 R.Kuhn andP.J.Drumm,Ber., 1O32,85, [B], 1468; A., 1138.1931, 92122 ORGANIC CHEMISTRY .-PART I.pyridine (aq.), and is doubtless due to the formation of enolic sodio-derivatives in the groups *CH:CH*CH,*CO,R.Sodio-glutaconic ester (yellow) Sodio-dihydromuconic ester (red)I-(blue)Dihydromethylbixin(orange)Dimethyl ester of dihydrocrocetin(yellow ) I1 CH:CH*CMe:CH*CH:CH*CMe:CH*CH:C( 0 M e ) d a(green)The auto-oxidation of dihydromethylbixin in the presence of causticsoda results rapidly in the removal of the two hydrogen atoms fromthe or-carbon atoms according to the equation :C26H,404 + 0, + C,,H& + H202.94Lycopene and p-Carotene.-Lycopene, C,H,,, yields on gentleoxidation with chromic acid a C8-ketone, CMe,:CH*CH,*CH,*COMe,the balance of the molecule being represented by a new aldehyde,lycopenal, C32H420, which contains an isopropylidene group and tendouble bonds in the carbon chain.g5 Spectroscopic examinationof the aldehyde and its oxime indicates that the aldehyde group isprobably conjugated with the double bonds of the chain as requiredby the formula (XXXVIII) suggested by P.Karrer and his col-laborator~,~~ but the evidence here is not decisive and does notexclude the possibility that lycopenal possesses a structure corre-sponding to the lycopene formula (XXXIX), which, like (XXXVIII),fits the facts so far known about the lycopene constitution.[ CMe2:CH*CH2*CH2.CMe:CH*CH:C-H*CMe:CH*CH:CH*CMe:CH*CH:],(XXXVIII.)CMe, :CH*CH, *CH,*CMe :CH*CH,*CH,*CMe:CH*CH:CH*CMe :CH*GHCMe, CH CH :CH*CMe :CH*CH :CH*CMe CH CM : CH*CMe: CH CH(XXXIX.)94 R.Kuhn, P. J. Drumm, M. Hoffer, and E. F. Moller, Ber., 1932, 85, [B],s5 R. Kuhn and C. Grundmann, ibid., p. 898; A., 749.D6 Helv. Chim. Acta, 1930,13, 1084; 1931,14, 435; A., 1930, 1422; 1931,1785.597FARMER. 123It has been shown that p-carotene suffers oxidation on carefultreatment with chromic acid to yield a p-oxycarotene which appearsto have the empirical formula C40H5803, containing without doubt onep-ionone ring undamaged-a feature which results in the preserva-tion of the growth-promoting effect characteristic of p-carotene.With a larger amount of chromic acid a second oxidation product,apparently C,H,,O,, is obtained which does not appear to bederived by way of P-oxycarotene.This substance, to which thename of p-carotenone has been given, appears to be derived by fhespecific oxidation of the double bonds of the p-ionone rings inp-carotene as shown in formula (XL).97Rubber.Natural Rubber.-A series of investigations on the structure ofnatural rubber has been carried out by T. Midgley, jun., and A. L.When cr@pe rubber is rapidly distilled, the distillateobtained, although consisting largely of isoprene, dipentene, andheveen (C,,-hydrocarbon), contains a number of other hydrocarbonsin smaller or greater quantity. All of the recognisable products?containing from five to ten carbon atoms in the molecule, werefound to possess structures which could have been derived from therubber unit, --C---?=C-C- or ---C-Q_b--C--, by simpleC Coperations (hydrogenation, dehydrogenation, double-bond migration,self-addition). From the proportions of the variws productsobtained, it appears that there is little tendency for the carbonskeleton of the rubber molecule? as represented in the Pickles formula,I--- C-~=C-~~-C-?=C-C-j-C-r=C-C .--C C Cto suffer fission at points other than those single bonds which arefarthest removed from the double bonds.As the result of a study of the precipitation of rubber from analcohol-benzene medium it is claimed that individual long-chaincomponents of rubber can be characterised by the temperature a twhich precipitation occurs in a slowly cooled solution of standardcomposition. By means of this technique rubber is found to consist97 R.Kuhn and H. Brockmann, Ber., 1932,65, [B], 894; A., 749.98 J . Amer. Chem. SOC., 1929, 51, 1215; A., 1929, 702; ibid., 1931, 53, 203;A., 1931, 357; (with M. W. Renoll) ibid., p. 2733; B., 1931, 853; ibid., 193254, 3343, 3381 ; A., 1036124 ORGANIC CHEMISTRY .-PART I.to the extent of over SOY; of a single component; this is ac-companied by more soluble but less definitely characterised com-ponents. Milled rubber, on the other hand, is found to be made upof a continuous series of undefined components without a singlepredominating individual.An interesting contribution to rubber chemistry is afforded by anew investigation of the action of hydrogen peroxide on solutionsof rubber in chloroform-acetic acid.99 Extensive oxidative hydr-oxylation takes place at the double bonds of the rubber molecule,but some of the double bonds appear to survive, since the productis still unsaturated, and, after acetylation, further hydroxylationcan be brought about with hydrogen peroxide.During acetylation,however, some of the hydroxyl pairs appear to lose water to yieldoxide rings, and some of the methylene groups of the rubber unitto suffer replacement of hydrogen by oxygen. For details of thetransformations, which include the production of two dibasic acids,reference must be made to the original paper, but the main changesare shown in the following scheme, in which the primary hydroxyl-ation product is represented on a C,, basis, this being the smallestmolecular magnitude compatible with the analytical results obtained.Rubber H,O,i C50H,6(OH)16 --% C48H8,014(CH0)2&/' JHNO*c50H7 606(OAc)4 c48H86014(c02H)2 jH*@ %.%C~OH~~O~,(OH)~(OA~), C50H7606(0H)4 (CHo)Z48 74 1 (saturated)IC50H56016(0H)1Z(saturated)IJ.C48H7408(C02H)2Similar operations have been carried out with gutta-percha andbalat a.Artificial Rubber.-Owing to the fact that chloroprene ( p-chloro-butadiene), which can readily be prepared from acetylene in a stateof purity and in large quantity, polymerises about 700 times asrapidly as isoprene, it can be used as a starting material for theproduction of synthetic rubber.l Within ten days, under ordinaryconditions, chloroprene polymerises into a transparent resilient9g J.A. Mair and J.Todd, J . , 1932, 386.1 W. H. Carothers, I. Williams, -4. RI. Collins, and J. E. Kirby, J . Amer.Chem. SOC., 1931, 53, 4203FARMER. 125mass resembling vulcanised rubber; but if polymerisation is in-terrupted before it has proceeded to completion, a soft plasticprodvct (a-polymeride) resembling unvulcanised rubber is obtained.Under the action of heat the oc-polymeride changes rapidly into thep-polymeride, and by other means volatile (p)-, granular (a-), andbalata-like polymerides can be formed. The conversion of chloro-prene into p-polychloroprene is stated to proceed rapidly in aqueousemulsion, yielding a synthetic (vulcanised) latex of smaller particlesize than natural latex.With regard to the constitution of polymerised chloroprene itappears probable that the p-polymeride is the analogue of naturalrubber and should be represented by a formula strictly analogousto that of Pickles,* * * -CH~*~:CH*CH,*CH,*F:CH.CH,*CH,.Q:CH*CH,- * *In conformity with this formula it is extremely resistant to alkalis(as also to hydrochloric acid, hydrofluoric acid, ,and many otherreagents), is less reactive than natural rubber towards ozone, andyields succinic acid on oxidation.But these observations do notreveal whether a uniform attachment of chloroprene units in themanner 1 : 4-1 : 6, etc., obtains, or whether inverted unions ofthe type 1 : U : 1 or 4 : 1-1 : 4- occur in the chain. Suchinversions probably do not occur to any considerable extent insyntbetic isoprene rubber, since the behaviour of the latter toozone is normal,2 and here also the X-ray pattern indicates that thereis greater freedom from irregularities than in other synthetic rubbers.Of other known bay-chlorobutadienes, y-chloro- P-methylbutadieneis the only one likely to serve as a precursor of rubber.In this casepolymerisation occurs as rapidly as with chloroprene and theproduct is vulcanised by heat alone.3 When a second methyl groupis introduced into the chloroprene molecule, as in y-chloro-ccp-di-methylbutadiene, the ease of polymerisation is greatly reduced, butis still superior to that of isoprene.The nature of the product obtained by the pyrolysis of sodiumrubber (polymerised isoprene), in comparison with those obtainedfrom natural rubber, is held to indicate that the position of themethyl groups and the character of the unsaturation in methylrubber both differ from those in natural r ~ b b e r .~ The products fromsodium rubber are on the whole more saturated, showing a decreasea Compare R. Pummerer and A. Koch, M e d e r ’ s “ Handbuch der Kaut-W. H. Carothers and D. D. Coffman, J. Amer. Chem. SOC., 1932’54,4071;c1 c1 c1schukwissenschaft,” 1930, p. 270.A., 1232. ‘ T. Midgley, jun., A. L. Henne, and A. F. Shepard, ibirE., p. 381 ; A., 276126 ORGANIC CHEMISTRY .-PART I.i n the amount of isoprene and dipentene and an increase in theamount of pentenes and of saturated hydrocarbons present (saturatedhydrocarbons are absent from the natural rubber distillate).The pyrolysis results for sodium rubber agree with the arrange-ment of carbon atoms, but are incompatible with the regularrecurrence of double bonds shown in the carbon skeleton :Titration of sodium rubber with bromine indicates the presence ofone double bond per C,H, unit, and the ozonide of sodium rubbercorresponds in composition with [C5H8O3’Jn.The ozonide does not,however, break down into simple products on heating with waterand it is concluded that the sodium rubber molecule is a modification,possibly cyclised, of that shown above, the modification resultingin the presence of one chemically weak bond per C,H, unit.Polylsaccharides.Length of Chain and Molecular Weight .--Since the cryoscopicmethod of determining the molecular weight of polysaccharides hasin many cases proved to be unreliable, considerable interest attachesto the results afforded by purely chemical methods of estimation.A theoretically simple way of determining the length of chain of thosecomplex open-chain substances which are built up of recurrent unitsconsists in determining the percentage degree of occurrence of theterminal units (one or both), since these will usually permit ofdifferentiation from the intermediate units of the chain.Themethod will apply whether the complex substances in question ariseby the additive polymerisation of olehic units (polystyrenes, rubber,etc.) or by the condensation-polymerisation of hydroxylated mole-cules (polysaccharides), so long as the ends of the chain are notjoined. The practicability of the method, however, dependsentirely on the reactive capacities of the terminal units whilst stillforming part of the chain, on the facility with which one or other ofthe terminal units (or characteristic fragments thereof) can besegregated by quantitative scission of the chain, or on both properties.Thus if cellulose be represented, according to the conception whichhas gained almost general acceptance, as built up by the union ofglucosidically-linked p-glucose units (XLI), one of the terminal unitswill differ from the other and from the intermediate members of thechain in the number of non-glucosidyl hydroxyl groups capable ofmethylation (XLII).I n the same way, if starch and glycogen berepresented as built up of cc-glucose units (XLIII and XLIV), andinulin of fructofuranose units (XLV), the terminal C,-units will iFARMER.127every case be capable of differentiation from the intermediate units,and from one another.CH,*OHH OHCH,-OH -- H OHCH,-OHH OHc HxCH2-OH 1(XLV.)The method was found to be applicable equally to cell~lose,~starch (both amylose and amylopectin portions),6 glycogen,' andinulin.8 In the first three polysaccharides the non-reducing5 W. N. Haworth and H. Machemer, J., 1932,2270; A., 1022.6 E. L. Hirst, (Miss) M. M. T. PIant, and (Miss) M. D. Wilkinson, ibid., p.7 W. N. Haworth and E. G. V. Percival, ibid., p. 2277; A., 1022; see also8 W. N. Haworth, E. L. Hirst, and E. G. V. Percival, ibid., p. 2384 ; A., 11 17.2375; A., 1116.G . K.Hughes, A. K. Macbeth, and F. L. Winzor, ibid., p. 2026; A., 934128 OR,GANIC CHEMISTRY .-PART I.terminal unit in each case was estimated after cleavage of themethylated chain and glucosidation of the fragments as tetramethylmethylglucoside, and in the case of inulin as tetramethyl methyl-fructoside. The length of the chain and the corresponding molecularweights so determined are shown in the table.No. of C,-units. Mol. wt.Cellulose .................................... 100-200 20,000-40,000Starch (amylose) ........................ about 24 5,000Inulin ....................................... 30 5,000. , (am y lo pee tin ) .................. 9 ,Glycogen.. .................................. about 13 2,;ooThe value to be attached to these figures depends of course on theextent to which scission of the original polysaccharide chains hasbeen avoided in the preparation of the fully methylated product andthis consideration is of the utmost importance with a polysaccharideso susceptible of degradation as inulin.The production of fullymethylated polysaccharides, under conditions precluding degrada-tion, was studied prior to the above determinations and the poly-saccharide derivatives actually utilised were stated to be demon-strably free from scission products of small chain-length and to havebeen formed to all appearance without any degradation of theoriginal molecules. In this connexion it is to be remembered thatH. Staudinger and H. Freudenberger lo claim, on the basis ofviscosity measurements, that native cellulose is more complex thanthe most complex of the acetates.The values are, however, putforward as the average lower limits of the size of the macromolecules,and it is not regarded as impossiblle that the native celluloseshave molecular magnitudes of the order of the higher limit aboveindicated.A method analogous to the above, employing Willstatter andSchudel’s hypoiodite procedure for determining the free aldehydegroups of sugars, had previously been applied by different workersto determining the length of chain of the polysaccharides and theirderivatives. This method furnished evidence that cellulose can bebroken down by acetolysis to dextrins containing varying numbersof glucose units (e.g., 8-13, 3040, 20-60, etc.), but it proved tobe unreliable when applied to the more complex substances, includ-ing cellulose itself.ll H.Staudinger and H. Freudenberger place the9 For details, see the above cited papers. See also W. N. Haworth andH. R. L. Streight, HeZv. Chim. Acta, 1932,15,609 ; A., 724 ; W. E. Hagenbuch,ibid., p. 616; A., 725.lo Bey., 1930, 63, [B], 2331; A . , 1930, 1416.l1 Ann. Reports, 1930, 27, 112FARMER. 129limit of its reliability at chains of 60 glucose units,12 but its trust-worthiness up to even this degree of complexity is contested byK. Hess and his collaborators.13 For starch a chain-length of 25-30glucose units has been determined by its means-a value not differinggreatly from the foregoing estimate ; but for cellulose the correspond-ing value is one of only 50 units (average).llAnother purely chemical method which has been applied to thedetermination of the chain-length of celluloses from various sourcesrelies upon the estimation of the carboxyl content of the material : l4this is stated to be quite appreciable and to vary little with thesource of the cellulose.The length of the chain is in this wayassessed at 96 glucose units or a multiple thereof, whilst that of xylan,which is associated with the cellulose in wood, is similarly assesseda t 16 C,-units or a multiple thereof.Homogeneity of Structure.-In view of the structural relationshipwhich is held to exist between cellobiose and cellulose on the one handand between maltose and glycogen or starch on the other, the factthat very high yields of 2 : 3 : 6-trimethyl glucose are derivable fromthe methylated polysaccharides would seem to preclude the possi-bility that polysaccharide structures are other than homogeneouschains or large rings ; furthermore, the isolation of tetramethylglucopyranose suggests very strongly that each of the structures isan open chain of limited length. Nevertheless there are featuresconnected with the degradation products of the polysaccharideswhich are considered to throw grave doubt on the homogeneity ofthe molecular structures.Thus Sir J. C. Irvine maintains that whencellulose is methylated on a large scale and the product subjected tograded hydrolysis, only 2 : 3 : 6-trimethyl glucose is Liberated in thefirst instance, the 2 : 3 : 4 : 6-tetramethyl glucose being derived fromthe more resistant fractions of the methylated celldose.l5 Asimilar indication of non-homogeneity applies also to starch and toinulin, since it is stated in connexion with the former that one of thefractions of methylated amylose examined was convertible into amixture of sugars which contained as much as 23--26% of 2 ; 3 : 4 : 6-tetramethyl glucose together with 65--52% of 2 : 3 : 6-trimethylglucose, and 21% of 2 : 3- and 2 : 6-dimethyl glucose ; of the latterl2 Ber., 1930, 63, [B], 2331; A., 1930, 1416.13 K.Hess, K. Dziengel, and H. Maass, Ber., p. 1922; A,, 1930,1416; K. Dziengel, C. Trogus, and K. Hem, Annakn, 1931, 491, 52; A.,1932, 47.14 E. Schmidt, W. Simson, and R.Schnegg, Naturwiss., 1931,19, 1006; A.,1932, 149; E. Schmidt, K. Meinel, W. Jandebeur, W. Simon, and others,Celtulosechm., 1932, 18, 129; A., 934.15 Nature, 1932,129,470; Chem. and Ind., 1932,263; A., 502.REP.-VOL. XXIX. 130 OKGdNIC CKXEMISTRY .---PAl<T I.it is stated that a chain length of 25 or of 43 fructose units could bededuced according to whether the ~-methoxy-5-methylfurfural andtrimethyl anhydrofructose produced as by-products in the hydrolysiswere calculated as tetramethyl y-fructose (total 2.7%) or not (total1.7%).This question of the homogeneity of structure of wood and cottoncellulose as revealed by the composition of their hydrolysis productis considered a t some length by D. J. Be11.16 Two main considera-tions enter into the discussion : (1) the extent to which the hydro-lytic fission of directly methylated celluloses differs from that ofcelluloses which have been acetylated as a preliminary stage insecuring easy and complete methylation (the procedure employed byHaworth and Hirst and their collaborators) ; (2) the reason for theinertness towards hydrolytic reagents of certain resistant portionsof methylafed wood cellulose.Bell, like H. Staudinger and H.Freudenberger,ll considers that acetylation treatment-especiallyof the kind used in the production of acetone-soluble acetates-causes extensive depolymerisation of cellulose,17 so that the forma-tion of an appreciable amount of tetramethyl glucose from themethylation product of the so-called " depolymerised " celluloseand none from that of so-called " intact " cellulose is accounted for ;moreover, there is no necessity to assume that intact cellulosepossesses an open-chain structure, since ring fission of a closedsystem could occur by acetolysis where acetylation is employed as apreliminary to methylation.With regard to the resistant portionsof methylated wood cellulose, which are obtained both from " intact "cellulose and from celluIose which has been " depolymerised " by apreliminary acetylation process, hydrolysis is successful only afterthe material has been submitted to acetylation : then tri-, di-, andmono-methyl derivatives of glucose are formed and the conclusion isdrawn that the inertness of the methylated material towards hydro-lytic reagents is due to more fundamental reasons than degree ofpolymerisation, that is to say, presumably, to non-homogeneityof structure.I n connexion with the homogeneity of inulin it is to be noted thatthe very stable crystalline anhydrofructose derived by Sir J.C. Irvineand J. W. Stevenson l8 by the action of chloroformic nitric acid onl6 Biochem. J . , 1932, 26, 590, 598, 609; A., 934.17 E. Elod and A. Schrodt (2. angew. Chem., 1931,44,933; A., 1932,48) onthe contrary show by viscosity measurements that the conversion of primarycellulose acetates (insoluble in acetone) into secondary acetates (soluble inacetone) by means of acetic acid involves some deacetylation but no appreciablechange in molecular aggregation.18 J. Amer. Cliem.SOC., 1929, 51, 2197 ; A., 1929, 1046FARMER. 131inulin proves to be a clifructose anhydride doubtless possessing theformula : 89 l9HO*CI-I, KT/ \<"""oy/og O-CH, Yi: CH,.OHHO HH I i HO HThis substance, which was originally thought to represent in allprobability a heterogeneous component of the inulin chain, isobtainable not only by the action of nitric acid but (in the form of itsmethyl derivative) by the mild hydrolysis of methylated inulin ; itsresistance to hydrolysis, whereby it yields fructose, is probably to beattributed to the characteristic dioxan structure of the anhydridering. Three stereo-forms of the anhydride are theoretically capableof existence and three isomeric forms of difructose anhydride werereported last year, one of which proves t o be identical with Irvineand Stevenson's compound.The formation of the anhydrideappears to be brought about by rupture of the primary valencieswhich connect the fructose units in inulin and their reunion to formthe more stable compound; at any rate there is nothing in theevidence so far adduced to show that the anhydride is pre-formed ininulin or to conflict with the view that the inulin chain is homo-geneously composed of fructose units as represented in formulaThe most important evidence concerning the homogeneity orheterogeneity of structure of cellulose and starch must necessarilybe derived by detailed study of degradation. Here, however, thelines of investigation are manifold, embracing the kinetics ofhydrolysis, the range of possible cleavage products, and the correla-tion of rotational and X-ray data.Degradation of Gellalose and Starch.-Degradation of cellulose ormethylated cellulose to the di-, tri-, and tetra-ose stages has beendefinitely accomplished by acetolysis 2*) 21 and hydrolysis ; 22l9 E.W. Bodycote, W. N. Haworth, and C. S. Woolvin, J., 1932,2389; A.,1117; W. N. Haworth and H. R. L. Streight, Helv. Chim. Acta, 1932,15, 693;A., 724; H. H. Schlubach and H. Elsner, Ber., 1932,65, [B], 519; A,, 603. Thelast authors suggest a monofructose structure.2o K. Freudenberg, C. C. Andersen, Y. Go, K. Friedrich, and N. W. Richt-myer, Ber., 1930,63, [B], 1961; A., 1930, 1412; F. Klages, ibid., 1931,64, [B],1193; A., 1931, 827; W. N. Haworth, E. L.Hirst, and H. A. Thomas, J.,1931, 824; A., 1931, 941.a1 K. Freudenberg, K. Friedrich, and I. Bumann, Annabn, 1932, 494, 41 ;A., 501.22 L. Zechmeister and G. Toth, Ber., 1931, 84, [B], 854; A., 1931, 716.(XLV)132 ORGANIC CHEMISTRY.-PART I.degradation to the penta- and hexa-ose stages appears also to havebeen 21 Degradation to certain other of the simplerstages, as represented by celloisobiose, procellose, and the well-known“ biosan ” of K. Hess and H. Friese, has also been reported, but thereis now little doubt that the “ biosan ” is a cellodextrin or mixture ofcellodextrins of high molecular weight (at least 2000) and recentwork shows that celloisobiose and procellose represent one and thesame tri~accharide.~~ With regard, however, to the long series ofpolysaccharides which might be expected to range between hexaoseand undecomposed cellulose, the most elaborate fractionations ofdegradation products have failed to yield pure individuals.Thesize of the break-down products appears to be to some extentcontrollable by adjusting the conditions of acetolysis, but it isremarkable that, whereas W. N. Haworth and K. M a ~ h e m e r , ~ ~ likeH. Staudinger,26 K. FreudenbergY27 K. H. Meyer and H. Mark,28 andtheir respective collaborators, have found the cellodextrin acetatesderived by acetolysis to be separable into groups of different averagechain-length, K. Hess and his collaborators z9 have found that undercomparable conditions of acetolysis the product appears to be com-posed of a single crystalline variety of cellulose acetate (Hess’s celluloseacetate 11) and one or two of the early members of the cellodextrinseries.30 The impossibility of effecting the complete separation ofthe few components in the latter case is attributed to the formationof additive complexes between the latter.Haworth and Machemer’sresults for the methylated cellodextrins, which were achieved byestimating the amount of te t ramet h yl met hylglucoside obtainable oncomplete hydrolysis, differ little from those obtained directly for thecellodextrin acetates : the individual polysaccharides range in sizefrom chains of 11 to chains of 26 glucose units.From the acetolysis product of starch, maltotriose and malto-tetraose (in addition t o glucose and maltose) have been isolated inthe form of their fully-methylated derivatives. I n connexion with23 K.Hess and F. Klages, Annalen, 1932, 497, 234; A., 1022.24 W. N. Haworth, E. L. Hirst, and 0. Ant-Wuorinen, J., 1932, 2368; A . ,25 Ibid., p. 2372.2 6 Ber., 1930, 63, [B], 2313, 2331, 3132; A., 1930,’ 1415, 1416; 1931, 202;2 7 K. Freudenberg and E. Bruch, ibicl., 1930, 63, [B], 535; A., 1930, 457.1116.ibid., 1931, 64, [B], 1688, 1694; A., 1031, 1040.K. H. Meyer and H. Mark, ibid., 1928,61, [B], 2432; A., 1029, 51 ; K. H.Meyer, H. Hopff, and H. Mark, ibid., 1930, 63, [B], 1531; A., 1930, 1025.29 K. Dziengel, C. Trogus, and K. Hess, Annalen, 1931, 491, 5 2 ; A., 1932,47.30 These lower polysaccharides, including the bioses, trioses, tetraoses, etc.,derived from cellulose and starch, have been conveniently termed by K-Fre udenberg, ‘ ‘ oligosaccharides.’ FARMER. 133these substances and the corresponding oligosaccharides fromcellulose, K. Freudenberg and his collaborators 21 point out that ifthe molecular rotation, [MIn, of a polysaccharide of n units be repre-sented as the sum of the rotations due to the two (aldehydic and non-aldehydic) end units, [MI, and [MI,, respectively and to 72-2 inter-mediate units each furnishing a rotation-contribution of [MI, /a(this differing little from that due to the middle unit of a penta-saccharide), there is justification for writing the resulting expression,(n - 2)[M],/a, or, to show the average contribution of one unit ofthe chain, in the formW l n = “la + (m - 2)[M]m /a + [Mle, in the form [ J f b 8 = [MI, +The difficulty which attends the application of the expression to thecalculation of the rotation of the higher polysaccharides, owing tocomplication arising from the existence of a- and @-forms, is fullyrecognised ; the values calculated for the trioses and tetraoses byemploying the observed values for the relevant bioses and hesaosesare in good agreement with the experimental values obtained underequilibrium conditions in various solvents ; moreover, the value of0-3” calculated for cellulose in water is a credible one.The progress of the hydrolysis of cellulose with 50% sulphuricacid has been measured by titration with iodine, and from the resultsvalues for K , and K,, the hydrolysis constants concerned respectivelyin the initial rupture of the intact chain and in the final stage ofhydrolysis (cellobiose to glucose), have been cal~ulated.~l Now, ofall the possible assumptions that might be made in terms of theseconstants as to the course of hydrolysis of the cellulose molecule(e.g., that K , is the hydrolysis constant applicable to the fission of allglucosidic linkages present in the molecule of triose, tetraose and soon up to cellulose, and K, is the hydrolysis constant of cellobiose),it is shown, on the basis of certain calculations of W.Kuhn32 as tothe form of various alternative hydrolysis curves, that only two arereconcilable with the observed changes in the value of the velocityconstant as hydrolysis proceeds : of these two, only one is definitelysupported by independent values for the velocity constant obtainedby following the course of reaction polarimetrically.This assump-tion is that Kl represents the hydrolysis constant of the tetraose andall higher fragments up to cellulose, and K, the constant of cello-31 K. Freudenberg, W. Kuhn, W. Durr, F. Bolz, and G. Steinbrunn, Ber.,1930, 63, [B], 1510; .A., 1930, 1025; compare K. H. Meyer, H. Hopff, andH. Mark, {bid., 1929,62, [B], 1103; 1930,63, [B], 1531; A,, 1929, 799; 1930,1025.32 IbicE., p. 1503; A,, 1930, 1025134 ORGANIC CHEMISTRY.-PART T.biose and cellotriose. Obviously this represents only an approxima-tion to the truth, since if cellulose be conceived as a long chainmolecule a number of transition values must actually lie between thehydrolysis constant applicable to degradation of the long fragments(K,) and that concerned in the last step (K2).(Thus, although Kzis attributed to both the biose and the triose, the triose will possess alower mean constant than K,, and the tetraose and pentaose a higherconstant than Kl.) Nevertheless, provided Kuhn’s fundamentalexpressions are accepted,33 the experimental results afford strikingevidence as to the uniformity of the linkages throughout the cellulosechain.Starch behaves in the same manner as cellulose, but its constantsare higher and the difference between K , and K, is smaller than inthe case of cellulose.31 Owing to the latter fact distinction betweenthe long-chain formula and a biose anhydride formula for starchwould have been impossible without other evidence.In this con-nexion the fact that starch yields nearly 100% of maltose duringdiastatic hydrolysis appears a t first sight to be in opposition to theconception of the former as a chain of 1 : 4-linked a-glucose units;but this observation probably only means that the enzyme attacksone end of the chain in much the same way that the oxidation ofcellulose chains is held to occur.34Some interest attaches to the group of simple polysaccharideanhydrides derived from starch by the action of F. Schardinger’sbacilI~s,~5 although the formation of these has no clear bearing onthe constitution of starch. Two of the anhydrides originallyrecognised by H.Pringsheim 36 in Schardinger’s a-dextrin nowappear to be identical (the supposed diamylose is the same as thetetra-amylose) ;37 but more interesting is the fact that X-rayexamination of a-amylose (amylopectin) is stated to show theexistence of six distinct modifications which differ in X-ray patternand are capable of interconversion either in solution or during thetransformation of solvent -containing forms into solvent -free ones .3 8The nature of the molecular changes in starch and cellulose and their33 See the contention of F. Klages, Ber., 1932, 85, [B], 302; A., 370; see34 T. Nakashima, J . SOC. Chem. I n d . Japan, 1931, 34, 414; A., 1932, 149.35 Centr. Bakt. Par., 1911, ii, 29, 188; A , , 1911, i, 181.36 Ber., 1912, 45, 2533 and subsequent papers; A., 1912, i, 832; see alsoH.Pringsheim, A. Weidinger, and P. Ohlmeyer, ibid., 1931,64, [B], 2125; A.,1931, 1277.3 7 A. Miekeley, ibid., 1930, 64, [B], 1957; 1932,85, [B], 69; A., 1930, 1414;1932, 255; M. Ulmann, Biochem. Z., 1932, 251, 458; A., 1021; K. Hess andM. Ulmann, Naturwiss., 1932,20,296; A., 724. See, however, H. Pringsheim,9. Weidinger, andH. Sallentien, Ber., 1931, 84, [B], 2117; A., 1931, 1276.<38 M. Ulmann, C. Trogus, and I<. Hess, ibid., 1932, 65, [B], 682 ; A., 604.also the reply of K. Freudenberg and W. Kuhn, ibid., p. 484; A., 501FARMER. 135derivatives which are responsible for these modifications of theX-ray diagrams are at present obscure. The water-soluble oligo-saccharides (tetraose, pentaose, and hexaose) derived by acetolysisof cellulose are stated to give the same interferences as hydro-cellulose,39 whilst the native and mercerised forms of cellulose as wellas a number of modifications of starch are all reported to show quitedistinctive X-ray diagrams.40 It is suggested, indeed, that three orfour truly isomeric forms of cellulose and four or five such forms ofstarch exist, the isomerism being of a steric nature, due to thepliation of the pyranose rings.Cellulose, starch, inulin, mannan, xylan, silk fibroin, wool, horn,etc., are all reported to be broken down to mixtures of low-molecularanhydrides by the action of dry hydrogen chloride.41 It is suggestedthat addition compounds with the reagent are first formed, andthese, being more susceptible of degradation than the original com-pounds, yield small molecules (possibly unimolecular) whichsubsequently revert to a more or less highly polymerised condition.Degradation of cellulose acetate to the condition of a water-solublecarbohydrate has been effected by benzenesulphonic acid,42 and thatof starch to gentiobiose by heating with dilute hydrochloric acidunder pressure.43 In the latter case the disaccharide would appearto be a reversion product of glucose, but its formation by directscission of a polysaccharide in the cc-amylose portion of the starchis the view recorded.With respect to the amylose and amylopectin portions of starch,although the former is generally regarded as constituting the majorportion of the starch grain, it is now stated that the yields of amylo-pectin derived by the action of dilute methyl-alcoholic hydrogenchloride on different starches range from 63 to 83%.The amylo-pectin thus obtained, while probably not identical with the naturalsheath substance of the grains, is claimed to be (unlike formerpreparations) free from a m y l o ~ e . ~ ~Synthuis of Cellulose.-The action of bacteria in building upsubstances of polysacc haride character from sugar- cont ainingmaterials has long been observed. Now, however, it is shown thatthe membranous material produced by the action of Acetobacterxylinum on glucose behaves in all its observed physical and chemicsl30 K. Hess and F. Klages, Annalen, 1932,497,234; A., 1022.40 J. R. Katz and A. Weidinger, Rec.trav. chim., 1932, 51, 842; A,, 934.4 1 H. H. Schlubach, H. Elsner, and V. Prochovnick, Angew. Chem., 1932,45,42 H. Pringsheim and K. Ward, jun., Cellulosechem., 1932, 13, 6 5 ; A., 502.43 T. C, Taylor and D. Lifschitz, J. Arner. Chem. Soc., 1932, 54, 1054; A.,44 A, Eckert and A. Marzin, J. pr. Chem., 1932, [ii], 133, 110; A., 370,245 ; A., 502.500136 ORGANIC CHEMISTRY.-PART I T .properties (including rotation, X-ray structure, acetylation, methyl-at,ion, and hydrolysis) as a true cell~lose.~~E. H. FARMER.PART IT.-HOMOCYCLIC DIVISION.Ta.utomerism .(Continued from Ann. Reports, 1931, 28, 105.)Three-carbon Systems.-The work on the influence of structure ontautomerism in three-carbon systems terminated by one activatinggroup, of the general type(1.) >CH&CRX >C:d*CHRX (IT.)a,!?-form flyform(X = CO,H, CO,Et, CN, or COMe)is now reaching finality inasmuch as all the more accessible systemshave been examined, and it is of interest to review the results inthe light of the theoretical considerations advanced to explain them.The application of the electronic theory to this problem, origin-ally proposed by Ingold, Shoppee, and Thorpe and used in theseReports, was extended by R.P. Linstead,2 who pointed out thatthe influence of steric factors and the effect of conjugation (Lap-worth and Manske’s “ Thiele factor ”) 3 must be taken into accountin addition to the polar factor. The theory in that form gives asatisfactlory explanation of the interconversion of unsaturatedacids (I and 11; X = C02H) or, more strictly, their anions inalkali.*Thus, the introduction of an alkyl group in the a-position retardstautomeric change and causes a shift of the equilibrium towardsthe @-form; an allryl group introduced into the y-position causesa marked shift in the opposite direction without affecting themobility, that is, the rate a t which equilibrium is established; andan alkyl group in the P-position also favours the &-form. The45 H. Hibbert and J. Barsha, J . Amer. Chem. SOC., 1931,53, 3907 ; A., 1931,1401; Canadian J . Res., 1931, 5, 580; A., 1932, 256; E. Schmidt, 31. Atterer,and H. Schnegg, Cellulosechem., 1931, 12, 235 ; A., 1931, 1038.1 J., 1926, 1477; A., 1926, 939.J . , 1929, 2498; A., 1930, 64; compare Ann.Reports, 1931, loc. cit.J . , 1928, 2535; A , , 1928, 1245.* The statement that the interconversion of acids is studied “in sodiumethoxide a t 25” ” (Ann. Reports, 1931, 28, 108) should read “with 10 equiv-alents of 25% potassium hydroxide solution a t lOO”.” The footnote on thesame page refers to the cyuilibration of the ester, not the acid.-G. A. R. IconKON. 137principal discrepancy is the position of equilibrium in the cyclicacids, in which the &form is favoured in every case :85% By at equil.(111.)this will be discussed below.In nitriles (I and 11; X = CN) the equilibrium is largely in-fluenced by the great tendency of the cyano-group to become partof a conjugated system: this was again observed in a recentinvestigation.4 This conjugative effect accounts for the fact that,of all the unsaturated nitriles examined, only two are preferentiallystable in the py-form.These are isohexenonitrile (VI)4 andphenylcrotononitrile (VII) 5 : the former owes its stability to tlheWI.1 CMe,:CH*CH,*CN CHPh :CH CH,*CN (VII* 1polar effect of the y-alkyl group, and in the latter the phenyl groupexercises a strong conjugative influence opposed to that of thenitrile group.In nitriles, therefore, the effect of a p-alkyl group is negligible;but a y-group produces its normal effect, doubtless because it actsdirectly on one of the carbon atoms concerned in tautomerism.Unsaturated nitriles have also been the subject of repeated studyby Bruylants and his schoo1,G but mainly from the stereochemicalpoint of view.Bruylants criticises some ofLetch and Linstead's conclusions, more especially as regardsthe purity of the nitriles prepared by different processes and thevalue of the analytical methods employed by the two schools;these differences in no way affect the main conclusions discussedabove.*In unsaturated ketones (I and 11; X = COMe), although theeffect of 8- and y-substituents on tautomerism accords well withIn his latest paperR. A. Letch and R. P. Linstead, J., 1932, 443; A., 371.A. Kandiah and R. P. Linstead, J., 1929, 2139; A., 1929, 1294.P. Bruylants, Bull. SOC. chim. Belg., 1930, 39, 572; R. Breckpot, ibid.,p. 462; P. Colmant, ibid., p. 568; G. Heim, ibid., p. 458; A., 1931,472, 194,472, 205; P.Bruylants and H. Minetti, Bull. Acad. TOY. Belg., 1930, 16,1116; P. Bruylants, ibid., 1931, 17, 1008; P. Bruylants and L. Emould,ibid., p. 1027; A., 1931, 205, 1403; A. Dewael, Bull. SOC. chim. Belg., 1932,41, 318, 324; J. Baerts, ibid., p. 314; G. Festraete, ibid., p. 327; P. Bruy-lants, ibicl., p. 333; A., 1119.P. Bruylants, ibid., p. 309; A., 1119.* Dr. Linstead informs the Reporter that a re-examination of the nitrilesused in Let& and Linstead's work leaves no doubt as to their purity;Bruylants's criticism on this point appears to be based on a misunderstanding.E 138 ORGANIC CHEMISTRY .-PART 11.theory, that of an a-alkyl group, founded on numerous observationsJgcauses not only a great diminution of mobility but also a pronouncedshift in equilibrium towards the @form, that is, in the oppositesense to that predicted by theory.Moreover, rema,rkable differ-ences, not predictable from considerations of polarity, are observedbetween the different cyclic derivatives : 9TH2*CH2)C:CHX CH~(CH:.CH~ CH 'CH'2)C:CHX ~H2*CH2*CH2)C:CHX(X = COMe)CH2*CH2 CH2*CH2*CH277% up at equil. 23% aP 40% UPFinally, the equilibria in a representative series of unsaturatedesters have been examined; lo these follow in the main the ruleslaid down for acids, but ethyl CycEopentylideneacetate constitutesan exception to this, the equilibrium favouring the ap-compoundas in the corresponding ketone ; the compound is also exceptionallymobile, apparently a characteristic of all cyclopentane derivatives.The effect of the a-alkyl group is irregular, for it favours the&form in the cyclohexane derivative but in no other case.Ester (up-form).Equilibrium (yo up). Mobility.CHMe,.CH:CH*CO,Et * ............... 10 (?) HighCH,Et*CH:CH*CO,Et .................. 92 (?) 153CH,Et.CH:CMe.CO,Et .................. 95 151CMeEt:CH*CO,Et ........................ 75 26CMeEt:CMe*CO,Et ..................... 05 2CH,*CH,1 \C :CH *C 0 ,E t 60 835CH2*CH2/CH2\CH,CH,/CH2*CH2\CH2*CH2CH2\CH,*CH,/...............,CH2*CH2\C:CH*CO,Et -1 ...... 38 8.11 /C:CMe*CO,Et ............... 88 84,CH,*CH,\C:CMc.CO,E t ...... 5 0.15* Linstead, J . , 1929, 2498. 7 Kon and Linstead, J., 1929, 1269.It is difficult to account satisfactorily for these anomalies.An attempt was made to explain l1 the preferential existence ofcyclohexane derivatives in the py-form by assuming that the wander-8 G.A. R. Kon and E. Leton, J . , 1931, 3496; A., 1931, 1274; compares Ann. Reports, 1929, 26, 117. The figures given above are the latest10 G. A. R. Kon, R. P. Linstead, and G. W. G. Maclennan, J., 1932, 2454;11 S. F. Birch, G. A. R. Kon, and W. S. G. P. Norris, J., 1923, 123, 1361;Ann. Reports, 1931, 28, 109.available; Kon, J., 1930, 1616; A., 1930, 1184.A., 1111.G. A. R. Kon and E. A. Speight, J . , 1926, 2727; A., 1926, 1246KON. 139ing of the double bond into the ring relieves the strain in the six-membered ring which is postulated by Baeyer's strain theory.Conversely, it has been suggested by Bennett (Ann.Reports, 1929,26, 118) that the greater stability of the ap-form in cyclopentanederivatives is due to the fact that the migration of the double bondinto the five-membered ring introduces a strain : a particularlygood example of this difference is afforded by the substitutedmalonic acids (VIII) and (IX) ; the former exists exclusively in theap- and the latter in the @-form and their respective isomeridesdo not appear to be capable of isolation.12 *The equilibria in the monocarboxylic acids (111), (IV), and (V),however, cannot be reconciled with these views. In addition, thefoundation for them has been considerably weakened by theobservations of R. S. Thakur 13 on a number of dicyclic acids,ketones, and esters of the type (X) and (XI).He finds thatCH, CH,/\ /\\/\/QH, QH QXRXCH, CH CH2CH, CH,CH, CH,/ \ /\\/\/QH, QH GCHRXCH, CH CHCH, CH,p-decalin derivatives behave in all respects like the correspondingcyclohexane compounds, the figures in the two series agreeing withina few units yo :X = C0,H : R = H, 88% By at equil.; R = Me, -yo By at equil.X = COMe : R = H, 63% By at equil. ; R = Me, 100 yo By at equil.X = COzEt : R = H, 60% By at equil. ; R = Me, 90% By at equil.The anomalous effect of the or-alkyl group is again observed.In the corresponding tram-hexahydrohydrindene derivatives ofthe general type (XII) and (XIII), Thakur l4 k d s that the equili-brium is in every case, including the monocarboxylic acids, on thel2 W. E. Hugh and G. A. R. Kon, J., 1930, 775; A., 1930, 1162.l3 J., 1932, 2129, 2139; A., 1032.l4 Xbid., pp.2147, 2157; A., 1032.* The statement (Ann. Reports, 1931, 28, 110) that the ester of the acid(VIII), when regenerated from its sodio-derivative, contains 30-50y0 of theBy-isomeride is erroneous. The original authors did not give an estimate ofits By-content, but the ester appeared from its physical properties to bepractically pure ethyl cyclopentenylmalonate; this view is c o h e d byunpublished experiments140 ORGANIC CHEMISTRY.-PART II.side of the ap-form; the compounds display even greater mobilitythan cyclopentane derivatives.With regard to the decalin derivatives, the results may be summedup by saying that the effect of the cyclohexane ring remains the samceven when the ring becomes part of the dicyclic system.The strainless nature of the decalin system is now generallyaccepted, although there is not the same measure of agreementregarding the simple cyclohexane ring.If the latter is strainless,no difficulty arises in the interpretation of the striking similaritybetween cyclohexane and decalin derivatives, but this means that the“ strain factor ” originally postulated has no existence. If, on t’heother hand, the single ring is regarded as strained, it must be inferredthat this strain has no influence on tautomerism, unless it be assumedthat the decalin system is strained to the same extent. It appearspreferable, therefore, t o abandon for the present all attempts tocorrelate ring strain with tautomerism.A similar argument applies to the cycEopentane and hexahydro-hydrindene compounds.The ring system of the former is strainlessand should not, therefore, differ in a radical manner in its influenceon tautomerism from the decalin system. The resemblance betweencyclopentane and hexahydrohydrindene compounds is not un-expected; the ring system of the latter appears from models t o beslightly strained (in the trans-form), but this strain is in the oppositcsense to that in the planar six-membered ring. The great tendencyof these compounds to exist in the ap-form might therefore be heldto support Bennett’s view (see above), since the entrance of the doublebond into the ring would tend to produce an even more strainedcondition than that present in cyczopentenyl derivatives, just as thefacts require.The Glutaconic Acids.-Some progress has been made in thcinvestigation of three-carbon systems terminated by activatinggroups at both ends of the general typeXCHR*CR‘:CR”X and X*CR:CR‘*CHR”X(X may be CO,H, CO,Et, or CN; R, R’ and R” may be dkyl or aryl groups,CO,H, or C0,Et)A brief reference to the subject has already been made in theReport for 1931, p.111; further investigations published in thKON. 141course of the year serve to establish several generalisations regardingthe simpler glutaconic acid~.l~-~O An unsymmetrically substitutedglutaconic acid, such as a-benzylglutaconic acid, can theoreticallyexist in two isomeric forms (I) and (11), each of which can giverise to two stereoisomerides :CO,H*C(CH,Ph):CH*CH,*CO,H CO,H*CH(CH,Ph)*CH:CH*CO,HIn practice two forms are usually encountered, namely, the cis-formof the ap-acid and the trans-form of the py-acid, the remaining twoisomerides being comparatively unstable ; Kon and Watson l8have, however, converted the cis-ap-form of a-benzyl- p-methyl-glutaconic acid into its trans-stereoisomeride by ultravioletirradiation.The relative stabilities of the two more stable modifications varywidely according t o the nature of the substituents present; e.g.,a-benzylglutaconic acid is stable in its trans-py-form, and the cis-ap-form is converted into this even on treatment with warm waterand therefore represents a true " labile " rnod&ation.l* The sameis probably true of other simple a-substituted acids; the cis-formof glutaconic acid itself has only lately been obtained 21 and is alsoextremely unstable.The a-benzyl- p-methyl acids (111) and (IV)are both stable, but the cis-acid passes integrally into its isomerideon treatment with alkali, and the reverse change is brought aboutby acids ; we therefore have one alkali-stable and one acid-stableform : 1,CO,H*C( CH,Ph) :CMe*CH,*CO,H e(I.) up-Acid (cis and trans) (11.) Py-Acid (cis and trans)alkali(111.) cis (ap) acidCO,H*CH( CH,Ph)*CMe:CH*CO,HThese acids thus differ from the majority of tautomeric compoundsin which the equilibrium is independent of the reagent employedand incidentally provide an interesting example of the conversionof a trans- into a cis-acid by the agency of hydrochloric acid, areagent commonly used t o bring about the reverse change.Similar relationships obtain in the p-phenyl- a-met hylglut aconicacids; l9 the presence of a phenyl group exercises the expected(IV.) trans (py)l5 G.A. R. Kon and E. M. Watson, J., 1932, 1 ; A., 252.l6 B. S. Gidvani, G. A. R. Kon, and C. R. Wright, ibid., p. 1027; A . , 601.1 7 G. A. R. Kon and H. R. Nanji, ibid., p. 2426; A., 1127.18 G. A. R. Kon and E. M. Watson, ibid., p. 2434; A., 1127.10 B. S. Gidvani and G. A. R. KOR, ibid., p. 2443; A., 1127.Z o G. A. R. Icon and H. R. Nmji, ibid., p. 2557; A., 1247.21 I. R. Malachowski, Ber., 1929, 62, [B], 1323; A., 1929, 794142 ORUANIC CHEMISTRY.-PART 11.stabilising influence (owing to its & T effect) on both possible forms(V) and (VI) : this is exemplified by the isolation, though in smallCO,HGMe:CPh*CH,GO,H CO,H*CHMeCPh:CH*CO,Hamount, of a third acid which appears to be the &-modificationThe stability of the different forms of unsymmetrically sub-stituted acids appears t o depend largely on steric factors in additiont o the expected polar effects and it has been suggested l8 that thesymmetrical distribution of groups about the doubly-bound carbonatoms is the principal one, thus :CH2Ph$*C02H CO,H*CH(CH,Ph)*EMe CO,H*CH(CH,Ph)*GMeCO,H*CH,*CMe HC*CO,H CO,H*CHStable StabIe Unstable(V.)of (VI).An investigation of some cyclic glutaconic derivatives showsthat ability to exist in stereoisomeric forms is not essential t o“glutaconic character.” The acid (VII) is just lilre an ordinary&substituted glutaconic acid ; it forms an enolic anhydride (VIII),its ester yields a stable potassio-derivative, and the latter can beconverted into the a-methyl ester (IX).These properties are there-fore solely connected with the existence of a mobile hydrogen atom.The acid (VII) and its ester and also the analogous cyclopentanederivatives only exist in one form, no sign of the expected isomerideswith the double bond outside the ring, such as (X), having beenencountered.CH2 CO/\/\\/\//QH2 $ ? (VIII.) Y?Y?(VII.) QH2 G*CO2HCH, C*CH2*C0,H CH, C C*OHCH, CH \/CH, P? 7H2 (IIHCO2H (X*) (IX.) YH, fi*CO,EtCH, C*CHMe*C02Et CH, C:CH*CO,H\ /CH2\/CH,A similar behaviour might have been ant,icipated in the acid (XI)and its ester, in which all three carbon atoms of the three-carbonchain are included in the ring :(XI.) CMe<CH*co2RC*CO,RSome of the reactions of these substances were indeed originallKON.143held to support such a view, although the " normal '' formulation(XII) was adopted for the acid and its solid ester.22A reinvestigation of the whole subject has led to some unexpectedconclusions.23 The ester, which is correctly represented by (XI),does not possess a mobile hydrogen atom in the ordinary senseand the supposed formation of a sodio-derivative cannot be sub-stantiated. The action of sodium ethoxide on this ester causesthe rapid addition of ethyl alcohol to the double bond to give theethoxy-ester (XIII) and no evidence has been obtained of theexistence of the Al-isomeride (XIV) of the ester (XI).The mostremarkable fact ascertained relates to the action of heat on theester (XI), which was stated to give rise to the " labile " modific-ation, formulated as (XI) by the earlier workers. This has now beenshown to consist of the straight-chain acetylenic ester (XV), thechange involving not only the opening of the cyclopropane ringbut the absorption of the methyl group into the chain.The behaviour of the ester (XI) with sodium ethoxide recalls thereaction of the esters of itaconic, citraconic, and mesaconic acids,which all undergo conversion into the same ethoxy-ester, directlyderived from itaconic ester : 24GH2 $?H,*OEt YH3$*CO,Et =+ Q*CO,Et -+ yH*CO,EtCH*C02Et CH,* C 0,E t CH,*CO,Et(cis and trans) Itaconic ester w-EthoxymethylsuccinicesterIt has, however, been shown that an equilibrium mixture of allthree unsaturated esters is produced and that the two stereoisomerica@-esters (citraconic and mesaconic) predominate in this; it isclearly unsafe t o conclude that the equilibrium in this case favoursthe &form (ethyl itaconate) because derivatives of the latter, suchas the ethoxy-ester and' the addition product with ethyl sodio-malonate, are preferentially produced,25 for additions are known tobe influenced by such factors as the solubility of the sodio-derivativesformed.The carbethoxyglutaconic esters l6 are generally similar to the22 F.R. Goss, C. K. Ingold, and J. F. Thorpe, J., 1923, 123, 327, 3342;25 G. A. R. Kon and H. R. Nanji, J . , 1932, 2557; A., 1247.24 E. H. Coulson and G. A. R. Kon, J., 1932, 2568; A., 1234.25 C. W. Shoppee, J., 1930, 968; A., 1930, 912.1924,125, 1927; 1925, 127, 460144 ORGANIC CHEMISTRY.-PART 11.corresponding cyanoglutaconic esters 26 in occurring as mixtures ofap- and py-isomerides,( C0,13t),C:CR*CHR'*C02Et (CO,Et),CH*CR:CR'*CO,Et(4) ( B Y )from which the py-forms can usuaIly be obtained in the pure stateby conversion into the potassium derivatives (which always havethe &structure ; no abnormal cases have been encountered) andacidification with a weak acid in the absence of water.* Ethyltx-carbethoxy-p-phenylglutaconate gives rise t o two different,metallic derivatives : 16? l9 the yellow, sparingly soluble sodio-derivative obtained in the condensation of ethyl sodiomalonatewith ethyl phenylpropiolate appears t o have the structure (XVI) ,whereas that formed from the ester itself on treatment with sodiuniethoxide has the sodium attached to the other end of the chain(XVII).A compound similar to (XVI) is also formed from ethyl(CO,Et),C*CPh*CH*CO,Et)Na Na((CO,Et),C*CPh*CH*CO,Et(XVI.) (XVII.)sodiomethylmalonate and ethyl phenylpropiolate, a fact which shouldhave a considerable bearing on the mechanism of the Michaelreaction, since it can be proved that in this case it is the sodiuniand not the methyl group of the addendum which separates andbecomes attached to the negatively polarised end of the unsaturatedmolecule.The cyanoaconitic esters26n arc difficult to purify imd do not lenrlthemselves to detailed study ; in general, however, they behave lilrothe cyanoglutaconic esters, particularly in forming abnormaly-alkyl derivatives : 26(XVIII.C0,Et *CH( CN)*C( CO,E~):CH~CO,E t -+(XIX.)The sodio-derivative of ths ester (XVIII) must also be derived fromthe isomeric ap-ester, and a product consisting mainly of the latteris obtained on acidification.26 G. A. R. Kon and H. It. Nanji, J., 1931, 560; A , , 1931, 608; compare2 b R. D. Desai, J . , 1932, 1088; A . , 602.* The repeated observation of such " false equilibria " lends strong supportt o the view that the anionic charge originally localised on one or other endof the three-carbon system (in the sodio-derivative) need not necessarily beredistributed and therefore lead to an equilibrium mixture when the mobilehydrogen atom is reintroduced by acidification.This redistribution appearsto take place under the influence of the acid (or water) and is analogous tothe phenomenon discussed on p. 146, namely, the selective reaction of benzyl-magnesium chloride to give a- or y-reaction products according to the reactantemployed.CO,Et*C( CN) :C( C0,Et )*CHR*CO,EtAnn. Reports, 1931, 28, 111RON. 145Methyleneamethine Xystems.-The effect of different meta-substituents on the mobility 27 in the systemx.2 R*C,H,*CR:N*CH,Ph s R*C,H,*CH2*N:CHPhx.2follows the order of the dipole moment of the compound R-Ph moreclosely than was the case with the corresponding para-compoundsY28in agreement with theoretical considerations. Similarly, the agree-ment between the effect on mobility and side-chain reactivity iscloser in the meta- than in the para-series.The effect of differentgroups on equilibrium should follow their effect in facilitatingmeta-substitution, but as all the groups employed (except NO,) areop-directing, a comparison is impossible and their effect on side-chainreactivity is theref ore employed.Annionotropy.-Comparatively little work has been carried out onthis subject. A. Kirrmann and R. Rambaud29 have found thatacetylation of the ester (I) with acetic anhydride leads to the corre-sponding acetate (11), but if the hydroxy-compound is treated withphosphorus tribromide, the bromine enters the y-, not the a-position ;the bromide (111) on treatment with sodium acetate gives an acetate(IV), isomeric with (11).(1.) CH,:CH*CH( OH)*CO,Et CH,:CH*CH(OAc)*CO,Et (11.)(In.) CH,Br*CH:CH*CO,Et CH2( OAc)*CK:CH*CO,Et (IV.)R.Rambaud 30 subsequently found that the ct-bromide isomericwith (111) was also produced but passed into (111) on distillation.The corresponding methyl ester undergoes the change even morereadily, but the a-chloro-ester can be prepared in the pure state andis stable ; on treatment with calcium bromide, however, the y-bromidealone is produced.Some abnormal cases of the Grignard reaction have been inter-preted on the assumption of an anionotropic mechanism.Cinnamylchloride (V) gives a magnesium derivative which is converted bycarbon dioxide into the acid (VI),31 and this is explained by themigration of the MgCl residue from the y- to the a-carbon atom :(V.) CHPh:CH*CH,Cl+ [CHPh:CH*CH,- -CHPh*CH:CH,](m.) CHPh(CO,H)*CH:CH, ---+ CPh(C0,H):CHMe (VII.)+ MgC1-2 7 C. W. Shoppee, J., 1932, 696; A., 384.28 Idem, J., 1930, 968; 1931, 1225; A., 1930, 912; 1931, 834; Ann.Reports, 1931, 28, 106.2s Compt. rend., 1932, 194, 1168; A., 600.30 Ibid., 195, 389; A., 930.31 H. Gilman and S. A. Harris, J . Amer. Chem. SOC., 1931, 53, 3641; A . ,1931, 1290146 ORGANIC CHEMISTRY .-PA4RT 11.The acid (VI) readily undergoes further change with acids or alkalis,or merely on heating, into methylatropic acid (VII), this time by aprototropic mechanism.The formation of the intermediate ionsor free radicals accords well with the isolation of as-diphenyl-AQc-hexadiene32 as a by-product in the above reaction, since thiscan result from the union of the two different radicals.The reaction of benzylmagnesium chloride with formaldehyde,giving o-tolylcarbinol, is interpreted on similar lines,33 the three-carbon system here involving the phenyl group :The two forms me probably in equilibrium, because the action ofcarbon dioxide on the Grignard reagent gives the normal product.phenylacetic acid.A somewhat similar explanation is advanced by P. R. Austin andJ. R. Johnson,34 who find that the course of the reaction is dependenton the nature of the compound reacting with the Grignard reagent ;for instance, o-tolyl derivatives tend to be produced from acidchlorides, anhydrides, formaldehyde, and its derivatives ; it issuggested that these influence the production of one or other electro-meric form of the benzyl radical, which then reacts.The tautomericchange can involve the p-position of the benzyl group if both o-posi-tions are substituted, leading to a&-migration of the MgCl group;thus the Grignard reagent from 2 : 6-dichlorobenzyl chloride giveswith acetyl chloride 3 : 5-dichloro-4-methylacetophenone (VIII),although with carbon dioxide the normal product, 2 : 6-dichloro-phenylacetic acid (IX), is obtained :&*llgC1 ()BF*co2H(VIII.)The migration of a PhSO, ion is held by D. T. Gibson 35 to explainthe production of the compound (XI) from benzenesulphonylacetone(X) and methyl p-toluenethiolsulphonate : 36(X.) Ph*SO,*CH,*COMe + C6H4Me*S0,*SMe ---+C6H4Me'S0,'CH(SMe)*COMe (XI.)and the compound (XI) can be converted into the correspondingbenzenesulphonyl derivative by means of methyl benzenethiol-32 H.Gilman and S. A. Harris, J . Amer. Chern. SOC., 1932,54, 2072; A,, 730.33 H. Gilman and J. E. Kirby, ibid., p. 348; A., 410.34 Ibid., p. 647; A., 385. 35 J., 1932, 1819; A . , 837.36 I d e m , J., 1931, 2641; A . , 1931, 1394KON. 147sulphonate in excess; the excess of PhSO, or C,H,Me*SO, ionsdetermines the course of the reaction, but the presence of the SMegroup is shown to be essential for the exchange to occur. The thiolgroups are also readily exchanged, such as SMe for SEt, in compoundsof the type (XI).The exchange of groups here is essentially inter-molecular and differs in this respect from the majority of tautomericchanges in which the separation of the mobile group is generallyinferred rather than proved.Further cases of tautomerism are discussed on pp. 173, 174.Other Rearrangements.-Owing to limitation of space the dis-cussion is deferred of a number of interesting papers on thepinacol-pinacolin change,3’ on pinacolinic dearninati~n,~~ and onthe rearrangement of quaternary ammonium salts 39 and ofo- hydroxy- and o-amino-arylsulphones .40A remarkable example of group migration is provided by triphenyl-methyl o-tolyl ether,41 the displaced group entering the side chain :The constitution of the product has been confirmed by G.S. Parsonsand C. W. Porter.42Terpenes.Progress in this group of natural products continues, especially inthe elucidation of the structure of the more complex compounds ofthe sesqui- and di-terpene group; the ,new investigations bearremarkable testimony to the usefulness of the so-called “ isoprenehypothesis,” which has frequently led t o the prediction of thecorrect structure when the experimental evidence was indecisive.In the monoterpene group few major problems remain to besolved, and of these the ever-green question of camphene has latelyundergone some notable developments.S. Nametkin and L. Briissoff 43 have found that a-methylcamphene(11), obtained by the dehydration of tert.-methylfenchyl alcohol (I)37 M.Tiffeneau, J. L6vy, and collaborators, Bull. SOC. chim., 1931, [iv], 49,1595 et seq.; W. E. Bachmann and F. H. Moser, J . Amer. Chem. SOC., 1932,54,1124; A., 515; W. E. Bachmann, ibid., p p . 1969,2112; A., 745,737; C.H.Beale and H. H. Hatt, ibid., p . 2405; A., 854.38 A. McKenzie and J. R. Myles, Ber., 1932, 65, [B], 209; A., 382; A.McKenzie and (Miss) E. M. Luis, ibid., p. 794; A., 746.39 T. Thomson and T. S. Stevens, J., 1932, 55; A., 262; J . L. Dunn andT. S. Stevens, ibid., p. 1926; A., 816.4O L. A. Warren and S. Smiles, ibid., pp. 1040, 2774; A., 735; A. A. Leviand S. Smiles, ibid., p. 1488; A., 735.4 1 P. Schorigin, Ber., 1926, 59, [B], 2506; A., 1927, 54.42 J. Amer. Chem. SOC., 1932, 54, 363; A., 267.43 Annalen, 1927, 459, 144; A ., 2928, 182148 ORGANIC CHEMISTRY .--PART TT.or of tert.-methylborneol,44 is hydrated to a methylisoborneol whichis not the 6-methyl compound (111) that should be formed by theusual Wagner mechanism, but the isomeric 4-methyl compound(IV) ; it gives on dehydration, in addition to a-methylcamphene, theisomeric p-methylcamphene (V), and this is readily reconverted into4-methylisoborneol :(1.1 (11.) (111.)H2C( l4 ‘CMe, EaOH&!Y,AH/”:”. -Ha0 H2C\ I ,CH*OHThe structure of the new products is definitely established andhas been independently confirmed by later work; 45 and a similarseries of reactions has also been carried out starting with tert.-p henylbornyl alc o hoL4These facts cannot be explained by the usual (Wagner) mechanismof the camphene change and the explanation advanced is that, justas dehydration of 2 : 2-dimet hylcyclohexanol can take place withoutchange of ring structure but is then accompanied by the wanderingof a methyl group (Meerwein),(V.) CHz2 ! 7 H2(f/7Me\CH2 (IV.)CMe\CMe, I -ACMeIsomer-withoutpinacolic I 4---H,C- Me2C//’CH,I13ina,coliccl1nngc%GZ+isomer-isat ionCH,rileso the reverse process also can take place; in this way a-methyl-camphene is hydrated as follows,CH*OH44 L.Ruzicka, Helw. Chim. Acta, 1018, 1, 110; A., 1915, i, 398; S. Namct-kin and M. Schlesinger, J. R ~ s s . Phys. Chem. SOC., 1919, 51, 144.45 M. Bredt-Savelsberg and J. Buchkremor, Ber., 1931, 64, 600; A., 1931,625; S.Nametkin and L. Brussoff, J. pr. Chem., 1932, 135, 165.46 S. Nametkin, A. Kitschkin, and D. Knrssanoff, J. p r . Chem., 1930, 124,144; A., 1930, 216RON. 149the camphene hydrate then passing into the more stable methyl-isoborneol.Now in the hydration of P-methylcamphene this pinacolic changedoes not take place ; owing to the fact that in this compound a CHgroup is present next to the CXH, group, the intermediate camphenehydrate (VI) can be formed by the simple addition of water tothe double bond :W e ) (VI.) (IV.) CH:OHThis simplified mechanism cannot apply to a-methylcamphene,because it would lead to an unstable tertiary alcohol; the primaryaddition product first undergoes a pinacolic change without iso-merisation of the ring, followed by the conversion of the p-methyl-camphene hydrate into 4-methylisoborneol.The dehydration of the latter yields, as already stated, mainly@-methylcamphene formed by the simplified process ; some cc-methyl-camphene is, however, produced at the same time, evidently by areversal of the sequence of changes just described, thus confirmingthe reality of both reaction mechanisms.When these considerations are applied to camphene itself, it isclear that both processes must lead to the same final product;thus, the dehydration of isoborneol must give rise to camphene :I Or i CHH 2 d c H b M e 2 H,C/ I \CMe2H,C-- I b e H,G,dH,C:CH2I CH2 I/\ CH'-OH -,isoBorneol CampheneThere are, however, implications in these changes which appear tohave escaped the original authors, and these have since been pointedFor instance, the two formulae of camphene given above are47 J.Houben end E. Pfankuch, Annulen, 1931, 489, 193; A., 1931, 1300150 ORGANIC CHEMISTRY.-PART II.not identical but represent mirror images (owing to the two rings notbeing in the same plane). Moreover, different products will beformed according as the methyl and the hydroxyl group which areinterchanged are cis or trans to one another. In the former casethere will be a reversal of the sign of rotation; in the latter, thetrans-isomeride will be formed, having the same sign of rotation asthe initial material :c CrM" Mec----Me /OH I /OHC---Mec----OH I /Meboth the products belong to the opposite optical system from thatof the initial material.Owing to the symmetrical nature of thebridge heads in camphene these changes involve a change not ofstructure but merely of optical properties; thus, in the reversibleconversion of camphene through the hydrochloride into isobornylchloride, the two forms corresponding to (a) and (b) would be inequilibrium and equal quantities of d- and Z-product should beformed, leading to complete racemisation, which is actually observed.The same process accounts for the racemisation of isobornyl chloridein boiling cresol. A similar mechanism explains the production 48of active camphene from bornyl chloride and aniline under mildconditions, whereas the inactive hydrocarbon is formed under moredrastic conditions.When one of the bridge heads in camphene or its derivativescarries a substituent, the Nametkin change leads to isomerisationas in the case of methylcamphene, as well as to optical inversion :this has been well illustrated by Houben and Pfankuch4' by stnumber of examples.For instance, Z-camphorcarboxylic acid wasprepared from d-camphor, its amide (VII) converted into theamine, and the latter into 4-hyiiroxycamphor (VIII) :Y 70-NH, NH2d f k H 2 H 2 d I CH,CMe2IH2+CH2 H,H.c\dna,/co "2QMi 0 \&jyfOI CMe21 + I CMe2h --+(VII.) (VIII.)48 P. Lipp and G. Stutzinger, Ber., 1932, 65, [B], 241; A., 398EON. 151From the latter the amide of 4-hydroxycamphenecarboxylic acid(IX) was obtained, which on hydrolysis with hydrochloric acidunderwent addition of water, the product suffering the Nametkinchange at the same time, and Snally forming camphorcarboxylicacid (XI) itself. The complete cycle thus involves the conversion ofthe Z- into the d-acid :CO*NH2 CO,H(IX.) (X.) (XI.)In a later paper 49 the conversion of d-camphor into its antipodeis described (via camphor dichloride, 4-chloroisoborneol, 4-chloro-camphor, and the reduction of the semicarbazone of the latter) :this supplies a proof of the reaction mechanism proposed byNametkin, not depending on racemisation phenomena.In connexion with camphor chemistry, the preparation ofd-B-homocamphor by the following series of reactions should benoted :C02H C0,Me C02Mec8H14<82 -+ <CHO <CHO -+ <C(OH)CH2-C02Et +-cGH:CH*C02H <CH2*CH2*C02H Gg>cH2 2C0,H C02HThe strongly dextrorotatory B-homocamphor was then convertedinto p-camphor 51 (epicamphor), which was la3vorotatory :An interesting experiment from the point of view of the " isoprenehypothesis " is that of T.Wagner-Ja~regg,~, who treated isoprenewith acetic acid containing a little sulphuric acid. Amongst thecondensation products were geraniol, cycbgeraniol, halo01 anda-terpineol, 1 : 4- and 1 : 8-cineole, and a monocyclic sesquiterpenehydrocarbon with three double bonds, convertible by formic acidinto a dicyclic one of the caryophyllene group.49 J. Houben and E. Pfankuch, Ber., 1931, 64, [B], 2719; A., 1932, 62.so F. Salmon-Legagneur, Compt. rend., 1931, 192, 748; A., 1931, 626.5 1 Idem, ibid., 1932, 194, 467; A., 399.6 2 Anncxkn, 1932, 496, 6 2 ; A,, 866152 ORGANIC CHEMISTRY.'-PART 11.The original hypothesis was that 1 : 4-dimerisation of isopreneresidues would take place, followed by hydration, and the formationof geraniol seems t o support such a view.In acid solution, however,linalool is known to give geraniol, and this process may accountfor its formation thus :8esquiterpenes.-The notable progress in this group is principallydue to the researches of L. Ruzicka, who has attacked the problemin a broad and fundament,al manner. In order to place beyonddoubt the identity of some of the naphthalene derivatives obtainedby the dehydrogenation of polyterpenes, he has synthesised all thepossible trimethylnaphthalenes ; 53 t,he synt,hesis of some importantphenanthrene derivatives, including pimanthrene and retene,was, however, carried out by R.D. Haworth, B. M. Letsky,and C. R. Mavin,54 and L. Ruzicka and H. Waldmann 55 haveindependently synthesised pimanthrenequinone. An interestingsynthesis of phenanthrenes is also due to 5. C . Bardhan and S. C.S e n g ~ p t a , ~ ~ who have obtained pimanthrene, retene, and 1 : 4-di-methylphenanthrene.For comparison with naturally occurring hydronaphthalenederivatives, decalin and all the possible methyldecalins have beenprepared 57 by a method which is an extension of that originallyused 58 for the preparation of decalin-1 : 3-dione :co co ' co5 3 L. Ruzicka and H. Ehmann, HeEv. Chirn. Acta, 1932, 15, 140; A., 277.5 4 J., 1932, 1784, 2720; R.D. Haworth, ibid., p. 1125; R. 1). Haworthand F. M. Bolam, ih;tl., p. 3248; A., 839, 608, 1024.55 Helv. Chiiii. Acta, 1932, 15, 907; A., 948.5 6 J., 1932, 2520, 2798; A., 1241.5 7 L. Ruzicka, D. R. Koolhaas, and A. H. Wind, H e h . Clii?n. Acta, 1931,58 G. A. R. Kon and M. Qudrat-i-Khuda, J . , 1926, 3071 ; A . , 1927, 150.14, 1151, 1171; A., 1931, 1302RON. 153The diketone undergoes reduction by Clemmensen’s method t o thesaturated hydrocarbon, which is found t o belong t o the truns-series-another illust,ration of the remarkable ease with which thetruns-locking of the second ring occurs.‘A comparison of the physical properties of the dicyclic sesqui-terpenes and their fully saturated reduction products with thesynthetic decalins shows that the former all belong to the cis-series.Bisabolol and bisabolene trihydrochloride have been synthesisedas follows : b9 B-terpineol was ozonised, giving the ketone (I), whichwas condensed with the Grignard reagent from z-bromo- p-methyl-Ap-pentene, giving hisabolol (11), converted into a trihydrochlorideidentical with that prepared from nerolidol :\C:CMe2(1.) (11.)This synthesis removes all doubt as to the position of the doublebonds, which formerly rested on analogy.The structure of eudesmol may now be considered as settled.60The formula (111) was regarded as probable,61 although the alt’erna-t,ive (IV) had not been finally disposed of :Me MeHO II(IV. 1CH,IICH2MeH(111.) Mea-form /3-formA hydroxyl group attached to a side chain is much more readilybenzoylated than one directly connected t o the ring; 62 it is nowfound that eudesmol readily gives a benzoyl derivative. Ifeudesmol is correctly represented by formula (111), the dihydro-chlorides of eudesmenc and of selinene should be identical or, atmost, stereoisomeric, and this has now been confirmed; the di-hydrochloride of selinene exists in two stereoisomeric forms, m.p.52” and 74”, respectively. The final proof of the position of thehydroxyl in eudesmol was obtained by dehydrating dihydroeudesmol59 L. Ruzicka and M. Liguori, Helu. Chim. Acta, 1932, 15, 3 ; A., 277.6o L. Ruzicka, A. H. Wind, and D. R. Koolhaas, ibid., 1931, 14, 1132;A., 1931, 1302.L. Ruzicka and E.Capato, Annalen, 1927, 453, 62 ; A . , 1987, 570.E2 L. Ruzicka and A. G. van Veen, ibid., 1929, 476, 109; A., 1929, 1305154 ORGANIC CHEMISTRY .-PART TT.under very mild conditions; the hydrocarbon (V) was then almostexclusively produced, and gave acetyldimethyldecalin (VI) onozonisation :CH20(v.) Me (vI.) MeThe formation of the isopropylidene analogue of (V), previouslyobserved,61 was inconclusive, since this could have been formedfrom an eudesmol of the structure (111) or (IV).Eudesmol, however, is a mixture of a- and @-forms, for oxidationproducts of both have been obtained from it :a-form1Me *n - *(/,) A/A >A/\/ OH I/ OH co/!-form CH2The a-form, which gives rise t o acidic products on ozonisation, isprincipally found in eudesmol prepared from selinene dihydrochloride(Semmler and Risse's selinenol),63 since elimination of hydrogenchloride usually tends to produce the form with the double bondin the ring; the @-form predominates in the natural product (fromthe oil of Eucalyptus Macarthuri). The melting point and rotationof eudesmol do not appear to be notably affected by its composition.Several alcohols of the sesquiterpene series, such as machil01,~~cryptomerad01,~~ and the alcohol from the bark of MugnoZia Ovata,66are identical with eudesm01.~~Of more than passing interest is the discovery of three crystallineketones closely related t o eudesmol, which have been isolated fromthe wood oil of Eremophila Mitchelli, since they are amongst63 Ber., 1912, 45, 3305; A., 1913, i, 66.64 S.Takagi, J . Pharm. SOC. Japan, 1921, 41, 473; A., 1921, i, 721.65 H. Wienhaus and H. Scholtz, Ber. Schimnzel and Co., 1929, 267; A . ,66 Y. Sugii and H. Shindo, J . Pharm. SOC. Japan, 1930, 50, 103; A., 1931," L. Ruzicka, D. R. Koolhaas, and A. H. Wind, H e h . Chim. Acta, 1931,1929, 1308.267.14, 1151; A., 1931, 1302RON. 155the first cyclic ketones of the sesquiterpene series so far discovered ;moreover, their structure has been completely elucidated.68 Theyare eremophilone (VII), 2-hydroxyeremophilone (VIII), and2-hydroxy- I :2-dihydroeremophilone (IX) :Me(VII.) (VIII.) (IX.)Eremophilone is dehydrogenated by selenium t o eudalene andcontains two double bonds, one of which is ap- to the keto-group,since it forms an unstable addition product with hydrogen sulphideand an oxide with hydrogen peroxide ; the formation of a hydroxy-methylene compound shows that there is a CH, group next to theketo-group, i.e., the system :CH*COCH2-.In accordance with this,reduction with sodium and alcohol leads to dihydroeremophilol(X), which on ozonisation yields formaldehyde and a hydroxy-ketone, C14H2402 (XI), further oxidised by hypobromite to anacid, C1,H2,O, (XII) :(X.)The double bond which(XI.) (XII.)escapes reduction and is here shown tobe in the side chain, must be independent of the conjugated system:CH*CO*CH,* and the latter must therefore be situated in the ringnot carrying the isopropenyl group. Only two such positions areavailable and a decision between that adopted and the alternativewith t’he keto-group in position 2 can readily be made.Thus, theoxide of eremophilone on treatment with acetic acid and sodiumacetate passes into 2-hydroxyeremophilone, identical with the naturalproduct (VIII). The benzoate of the latter gives acetone and onlya trace of formaldehyde on ozonisation and must therefore consistlargely of the isopropylidene compound; in eremophilone and itsderivatives we are clearly dealing with another case of isomerismsuch as that of eudesmol discussed above. Apart from acetone, theprincipal oxidation product is a mixed anhydride (XIII), formed6s A. E. Bradfield, A. R. Penfold, and J. L. Simonsen, J., 1932, 2744.A. Pfau (Helv. Chim. Acta, 1932,15,1481) claims to have isolated two isomericketones, C15E200, from oil of cedar and another closely related compound fromoil of turmeric, but no experimental details are available156 ORGANIC CHEMISTRY.-PART IT.by cyclisation of the original oxidation product with loss ofwater :The formation of this neutral compound finally establishes theposition of the ketonic group in these compounds.The compound(VIII) is evidently analogous t o diosphenol in the monoterpeneseries.2-Hydroxy- 1 : 2-dihydroeremophilone is oxidised by ozone t oformaldehyde, a ketone (XIV), and only a trace of acetone; thiscompound, like the parent ketone, consists mainly of the isopropenylEorm :(IX.) (XIV.)The position of the keto-group must be the same as in the parentcompound, because 2-hydroxytetrahydroeremophilone, which isobtained from (IX) on catalytic reduction, can be further reducedwith sodium amalgam to tetrahydroeremophilone, in the same waythat hydroxycamphor is reduced to camphor; 69 and the tetra-hydro-compound is oxidised by hydrogen peroxide to a dibasic acidC15H2404, showing that the hydroxyl group is next to the csrbonyl.Artemisin.-Artemisin, which accompanies santonin in ArtemisiaMarina, has been proved to be 7-hydroxysantonin (I).70 Energeticreduction in the presence of platinum oxide gives hexahydro-artemisin (11), which is dehydrogenated to l-methyl-7-ethyl-riapht halene :Me Me(1.) (11.16Q J.Bredt and M. Bredt-Savelsberg, Ber., 1929, 62, [B], 2214; A., 1929,7 O K. Tettweiler, 0. Engel, and E.Wedekind, Annulen, 1932, 492, 105; A , ,1308.371EON. 157The formation of desmotropoartemisin, which is a true phenol,71shows that both the double bonds must be in the same ring ; and theymust occupy the positions shown, because artemisin does not reactwith Caro's acid and therefore has no CH, group next t o the carbonyl.The hydroxyl group in artemisin is tertiary and its position isshown by hydrolysis with alkali, artemionic acid (111) being formed :(I) -+ MeCHl l\i MeCH -\/\ Pl'+MeCH I(111.)OH A/C0,H ' k H (!JO,H dHThis acid can be hydrogenated to a tetrahydro-compound, which canalso be obtained by alkaline hydrolysis of x-tetrahydroartemisin(there are four of these, as required by theory), and is further reducedby sodium amalgam to hexahydrosantonin (IV) :Me(IV.1The hydroxyl group of artemisin can be eliminated by heating withformic acid, artemisene (V) being formed; and this on opening thelactone ring passes into artemionic acid (VI), thus confirming themode of reaction given above :HeZenin.-Helenin, the bitter principle of elecampane root , isrelated to santonin and contains a mixture of lactones, in whichalantolactone and isoalantolactone were identified many yearsago ; 72 K. W. F. H a n ~ e n , ~ ~ and L. Ruzicka and P. Pieth 74 have inaddition isolated dihydroisoalantolactone, identical with the productobtained by reduction by S p r i n ~ . ~ ~ The lactones are accompaniedin the original oil by sesquiterpenes which belong to the eudesmolgroup, being dehydrogenated to eudalene by heating with selenium.7671 P. Bertolo, #azzetta, 1923, 53, 867; A., 1924, i, 304.l2 J. Kallen, Ber., 1876, 9, 154; A., 1876, i, 917.73 Ber., 1931, 64, [B], 67, 943; A., 1931, 360, 734.7 4 Helu. Chim. Acta, 1931, 14, 1090; A., 1931, 1301.7 5 Ber., 1901, 34, 775; A., 1901,76 L. Ruaicka and J. A. van Melsen, Helv. Chim. Acta, 1931,14, 397; A.,325.1931, 734158 ORGANIC CHEMISTRY.-PART II.The lac tones are dehydrogenated to 1 -methyl- 7 - e t hylnap ht haleneand evidently differ only in the position of the double bonds, becauscthey are reduced to the same tetrahydro-derivative, very similar to,but not identical with, deoxytetrahydrosantonin, with which it isprobably stereoisomeric. 77The naturally occurring dihydro-lactone gives on ozonisation aketo-lactone, also obtained by K.W. F. Han~en,'~ which on reduc-tion by Clemmensen's method gives an acid, C,,H,O,, and a hydro-carbon : the latter, also obtained by decarboxylation of the acid, is9-met hyl- 3-ethyl- cis- decalin 78 and yields p - e t hylnaphthalene whenheated with selenium. These changes are formulated as follows :p-Ethylnaphthalene has also been obtained from the dihydro-lactone by Hansen, but by a more direct and therefore less instructiverout'e. These reactions establish the carbon skeleton present in thetwo lactones.The position of the carboxyl and carbinol groups can be inferredfrom the fact that ozonisation, followed by mild oxidation withpermanganate, of a mixture of alantolactone and isoalantolactonegives the acid (XII).76 py) (XI.)H0,C pn +co HOA /\/H0,C COUnstable intermediatc compound.Both lactones are readily reduced to dihydro-derivatives, whiclisuggests that one of the double bonds is in the up-position to thccarbonyl group. This would give rise to formulze (XIII) and (XIV)for the iso-lactone :CO---b CH,(XIII.)As the dihydro-derivatives and &hydrochlorides of the two lactones7 7 Ber., 1931, 64, [B], 1904; A., 1065.' 8 L.Ruzicka, D. R. Koolhaas, a,nd A. H. Wind, Helv. Chim. Acta, 1931,14,1171; A., 1931, 1302KON. 159are different, the double bond not adjacent to the carbonyl groupmust also be differently situated in the two isomerides. Ozonkationof dihydroalantolactone gives a ketonic acid (XV), showing that thedouble bond in question is situated in the ring.The addition ofhydrogen chloride to this double bond must inevitably produce anarrangement identical with that resulting from the semicycliciso-lactone ; and as the two dihydrochlorides are different, theymust be represented by the formula? (XVI) and (XVII). The twolactones therefore most probably have the structures (XIII) and(XVIII) :(XVI.) (XVII. ) (XVIII.) AlantolactoneHansen 77 favours the formula (XIV) for the iso-lactone; shouldthis prove to be correct, the formula (XVIII) for alantolactone wouldalso require revision, since both (XIV) and (XVIII) would give riseto the same dihydrochloride.The Resin Acids (continued from Ann. Reports, 1927, 24, 124).-Considerable progress has been made in the elucidation of thestructure of abietic and dextropimaric acids and the formulaassigned to them have attained some degree of certainty.The new facts are as follows : oxidation of either acid with a largeexcess of permanganate leads to two acids, CllH1606 and C1,H1,06.79The former is dehydrogenated by selenium to m-xylene, and thelatter to 1 : 2 : 3-trirnethylben~ene.8~ In these acids ring I of theparent acids is intact and it can be concluded that the methyl groupwhich is lost in the dehydrogenation of abietic acid to reteneoccupies position 12, not 11 as hitherto supposed :Me Me M U M079 L.Ruzicka, J. Meyer, and M. Pfeiffer, Helw. Chim. Acta, 1925, 8, 637 ; d.,*O L. Ruzicka, M. W. Goldberg, H. W. Euyser, and C.F. Seidel, HeZv.1925, i, 1419; P. Levy, Ber., 1929, 62, 2497; A,, 1929, 1448.Chin&. Acta, 1931, 14, 545; A., 1931, 736160 ORGANIC CHEMISTRY .-PART 11.Later work showed that these formulz required revision nridincidentally supplied the final proof of the position of the carboxylin abietic and dextropimaric acids. F. Vocke 81 found that the C,,acid has two carboxyls attached to tertiary carbon atoms, and thesame conclusion was independently reached by L. Ruzicka, G. B. R.de Graaff, and H. J. Miiller.S2 It had previously been found thatmethylabietin, in which a methyl group replaces the carboxyl, isdehydrogenated to a methylretene, and methylpimarin, similarlyderived from dextropimaric acid, yields a meth~lpimanthrene.~~It has now been found 82 that these compounds are oxidised to thesame phenanthrene-1 : 7-dicarboxylic acid and that they are there-fore homoretene (111) and homopimanthrene (IV) respectively :(111.) W.) (V.) WI.)This means either that the carboxyl in the resin acids is attached toa methyl group as in (V) or that both groups are attached to thenucleus in position 1 as in (VI). The former alternative is contraryto the " isoprene rule " and does not account for the difficult esteri-fication of these acids, which is adequately expressed by formula (VI).The formation of an ethyl group in the dehydration of abietinol isdoubtless due to a pinacolic change :CH,-CH, - EtMe CH,.OH- [ 6 ] + I\ fi I 1The same conclusion was independently reached by R. D.HaworthYs5 who, in addition, synthesised homoretene and homo-pimanthrene and placed their structure beyond all doubt.The acid andits ester combine with maleic anhydride 86 and therefore contain asystem of conjugated double bonds, the possible positions of whichare represented in formula (VII) by thick lines; of these, theThe skeleton of abietic acid is thus finally settled.81 Annalen, 1932, 497, 247 ; A., 1036.82 Helv.Chim. Acta, 1932, 15, 1300; A . , 1255-1.83 Ann. Reports, 1827, loc. cit.85 J., 1932, 2717.a 6 L. Ruzicka, P. J. Ankersmit, and B. Frank, Helw. G'him. Acta, 1032, 15,1289; rl., 1254KON . 1619 : 14-position is considered to be excluded for steric reasons and thearrangement (VIII) is the most likely :Dextropimaric acid does not form an addition product withmaleic anhydride and therefore does not contain a system of con-jugated double bonds.On oxidation with excess of permanganateit gives the same two acids (I) and (11), and therefore has the samesubstituents in ring I, as abietic acid; 87 on ozonisation it givesformaldehyde, proving the presence of a terminal CH, group.The following formule are therefore considered likely :Of these, (IX) is much the more probable because tetrahydro-dextropimaric acid is dehydrated by selenium to pimanthrene, aswould be expected; from a compound of the alternative structure,some 1-methyl-7-ethylphenanthrene should be formed.The suggested formuh for abietic and dextropimaric acids arederived from irregular isoprene chains :~ .~ . ~ . ~ . ~ ~ . ~ . ~ . ~ . c! Q ' F ! $-J.c.c.c. Q -Abietic acidDextropimaric acid ~ ~ ~ ~ ~ ~ ~ * ~ ~ F! ' Q 9 '1 :iQi 1 ICAgathic (more strictly, agathicdicarboxylic) acid, C,H,,O, (XI),p<esent in manila or kauri copal, contains two rings and two doublebonds 88 and gives on dehydrogenation with selenium 1 : 2 : 5-tri-methylnaphthalene and a hydrocarbon C,,H,, (XII). On treatmentwith formic acid a further ring is closed, forming isoagathic acid8 7 L. Ruzicka, G. B. R. de Graaff, M. W. Goldberg, and B. Frank, HeZw.Chi,m. Acta, 1932, 15, 915; A., 948; R. D. Haworth, loc. cit.8 8 L. Ruzicka and J. R. Hosking, Annakn, 1929, 469, 572; A , , 1929, 572.REPe-VOL. XXIX. 162 ORGANIC CHEMISTRY .-I'ART 11.(XIII),89 which contains one double bond and is dehydrogenated LOpimanthrene; a small amount of the latter is also produced fromagat hic acid it self.Me\/Me/\A Me,/Me I A/\ '\(,"" JY I I1 I- H0,C c1\@2H AAk H ,Me H0,C(XI.) (XII.) (XIII.)One carboxyl group is lost by agathic and isoagathic acid onmelting, noragathic and norisoagathic acid respectively beingformed.One ester group in methyl agathate is readily hydrolysedand the free carboxyl also is easily split off, leading to methylnoragathate ; the latter is only hydrolysed with extreme difficulty,and the same applies also to the second ester group in methylagathatc and to both ester groups in methyl isoagathatc. The firstcarboxyl is therefore probably situated next to a double bond in theside chain, on a carbon atom which is involved in the conversion ofagathic acid into the iso-acid, since the character of this carboxyI .changes completely on cyclisation of the acid.The same doublebond is also t'hc only one t o bc reduced by sodium and alcohol andif it is o$ to the carboxyl, it is probable that there arc no othersubstituents in the vicinity. The results of ozonisation, which givesformaldehyde, formic and oxalic acids, also support this formulationand confirm the presence of the terminal CH, group in agathic acid.The second carboxyl must occupy a sterically protected position,being quite different from the carboxyl in abietic or dextropimaricacid ; and on the assumption that the acid is derived from a regulararrangement of isoprene units, the position 12 for this group is to bepreferred to the alternative one, 11.Moreover, one carbomethoxyl group of methyl isoagathate can bereduced, though not so easily as that in methyl agathate ; the carb-inol group is then converted into methyl, and the product ondehydrogenation with selenium passes into a new trimethylphen-anthrene. This shows that the carboxyl which is not reducedoccupies a different position from that in abietic acid :Me\/Me Me\/Mc:AA (SIIl'r _3 llile _3 1 l,Xe G."/\&HO,C Li\(jcF* OH H02C j / M c8 9 L.RiuiCka aiid J. t:. tlouliiiig, Helu. C'hiiu. dcta, 1930, 13, 1402; 1031,14, 203; A . , 1931, 231, 359KON.The diterpenic alcohol sclareol, from the oilappears to be closely connected with agathic acid.alcohol with a terminal CH, group oxidisableconverted into 1 : 2 : 5-trimethylnaphthalene on - -The probable formula (XIV) is put forward for it.90Me\ /Me163of Salvia sclarea,It is a ditertiaryto formaldehyde,dehydrogenation.(XIV.)A considerable amount of work has lately been carried out onelemic acid, a triterpene derivative, and the related amyrins andsapogenins but the results are not yet ripe for summarising; mostauthorities, however, now agree that the sapogenins are triterpenederivatives with a C3,, formula.Polycyclic Hydrocarbons.Apart from the technical importance of a number of the morecomplex hydrocarbons of this group, increasing attention has latelybeen devoted to the possibility of determining the fine structure ofthese compounds ; the results indicate that speculation has pro-gressed beyond mere guesswork and give a logical explanation ofotherwise obscure differences between closely related compounds,such as the great difference in meso-reactivity between phen-anthrene and anthracene or, better still, 1in.-benzanthracene, whichdoes not immediately follow from their formuke.A sensitive testof such reactivity is the reaction of these compounds with maleicanhydride 91 or benzoquinone ; 92 thus, phenanthrene does not addthe former reagent, whereas anthracene does so readily.93 Theabsorption spectra provide another means of investigation andnumerous measurements of extinction coefficients have been made ;the results of these physical measurements are in good agreementwith those of purely chemical investigations.E. Clar’s basic assumption is that these hydrocarbons exhibitM.M. Janot, Corn@. rend., 1930,191,847; 1931, 192, 845; A., 1931, 94,9 1 0. Diels and K. Alder, Annalen, 1931, 486, 191 ; A., 1931, 848; E. Clar,92 E. Clar and F. John, ib,id., 1930, 63, [BJ, 2967; E. Clar, ibid., 1931, 64,93 E. Cler, ibid., 1932, 65, [B], 1485; A., 1131.737.Ber., 1931, 64, [ B ] , 2194; A., 1931, 1292.[BJ, 1676; A., 1931, 209, 1044164 ORGANIC CHEMISTRY.-PART II.valency tautomerismg4 leading to two or more forms, which inanthracene, for example, are (I) and (11) :o-quinonoid formThe second form, referred toelectrons on the meso-carbonH”diradical formas the “R” state, has unsharedatoms and is responsible for thereactivity of these positions, which varies according to the perman-ence of this phase in different compounds.95 It is also responsiblefor the characteristic anthracene absorption bands in the long-waveregion of the spectrum.The catalytic hydrogenation of anthracene, which has lately beenre-e~amined,~~ is readily interpreted in the light of this conception :e.g., the formation of the 9 : 10-dihydride is not an intermediate stepin, but takes place concurrently with, the formation of the 1 : 2 : 3 : 4-tetra- and the 1 : 2 : 3 : 4 : 5 : 6 : 7 : 8-octa-hydride.The dihydride,being a true benzene derivative, is much more slowly hydrogenatedthan anthracene itself, and similarly the octahydride is only slowlyattacked. The dihydride is presumably derived from the “ R ”state, whereas the quinonoid form is directly hydrogenated t o thehigher hydrides, without passing through the dihydride as postulatedby G.S~hroeter.~’ *The degree of mobility of the unshared electrons in the “ R ” phasevaries with the valency demand of the arylene residues fused to themeso-ring and with it the depth of colour of the substance. Forinstance, 2 : 3-benzanthracene (111), which is more reactive thananthracene, is orange and is photochemically oxidised to the quinonewhen its xylene solution is shaken with air ; and 2 : 3 : 6 : 7-dibenz-anthracene (IV) is deep blue,g* forms a bimolecular peroxide onphotochemical oxidation, and instantaneously reacts with maleicanhydride. This indicates that it exists solely in the “ R ” state andit is actually termed 2 : 3 : 6 : 7-dibenzanthracene-9 : 10-diyl; inagreement with this, its blue solutions do not darken further on94 E.Clar and F. John, Zoc. c i t .95 E. Clar, Ber., 1932, 65, [B], 503; A., 608.O 6 K. Fries and K. Schilling, ibid., p. 1494; A., 1123.07 Ibid., 1924, 57, 2003 ; A., 1925, i, 127.98 E. Clar and F. John, ibid., 1930, 63, [B], 2967; A., 1931, 209.* The formation of the tetrahydride from the dihydride, which has beenestablished by Schroeter, may proceed through the intermediate dehydrogen-ation to anthracone, which has been isolated from the hydrogenation productof the dihydrideRON. 165heating, an effect observed in some anthracene derivative^,^^ anddissociation is evidently complete.\/\/+A/ //H" H"(" R " phase alone shown.)On the other hand, 1 : 2 : 5 : 6-dibenzanthracene (V) reacts slowlywith maleic anhydride and is colourless ; it is a curious fact that itscarcinogenic activity is apparently not directly connected with itschemical reactivity : 1(VI.) (VII.)Now these hydrocarbons can be considered to be derived fromquinones; thus, anthracene and 2 : 3-benzanthracene, in its sym-metrical form (VI), both embody the skeleton of o-benzoquinone.Clar suggests that there is a simple relationship between the reduc-tion potentials of the quinones and the extinction curves of the relatedhydrocarbons.A number of these potentials have been determinedby L. F. Fieser and his collaborators ; when plotted against log cmax.of the corresponding hydrocarbons, they give a straight line, andthus the reduction potential of an unknown quinone can be predictedif the absorption spectrum of the relevant hydrocarbon is known.Hydrocarbons related to a quinone with a low reduction potentialhave a short " R " phase and exhibit diminished meso-reactivityand vice versa; and the corresponding effect is observed in thedisplacement of the " A " absorption bands of the spectrum.2 : 3-Benzanthracene in its unsymmetrical form (VII) is derivedfrom the (presumably highly reactive) unknown 2 : S-naphtha-quinone.Similarly, 2 : 3 : 6 : 7-dibenzanthracene can be derivedfrom the latter or from the likewise unstable 2 : 3-anthraquinone;owing to the instability of these quinonoid forms the compound tendsto exist more and more in the free radical or " R " state, until in thelimiting case of the 2 : 3 : 6 : 7-compound this becomes the onlystable condition.n9 C.K. Ingold and P. G. Marshall, J., 1926, 3080; A., 1927, 141.1 J. W. Cook, J., 1931, 3273; A., 153.J. Amer. Chern. Soc., 1924, 46, 1864; 1929, 61, 3101; 1930, 52, 5204;A:, 1924, ii, 839; 1929, 1452; 1931, 173166 ORGANIC CHEMISTRY.-PART II.All the extinction curves show, in addition to the ‘( A ” bands, anintense band of clearly benzenoid character, though somewhatdisplaced. The paleic anhydride addition product of anthraceneshows a typical benzene spectrum, although it is displaced in thesame way, but the corresponding compound from 1 : 2-benzanthra-cene (VIII) is a typical naphthalene derivative.The free hydro-carbon also shows some characteristics of the naphthalene spectrum,but no trace of this is seen in 2 : 3-benz- and 2 : 3 : 6 : 7-dibenz-anthracene, although it reappears in their addition products withmaleic anhydride. It is clear that in the latter the effect of theunshared electrons due to the “ R ” state is no longer observed andthe spectrum is entirely due to the benzene (or naphthalene) rings a tthe side :H*C 0 I / H T I CH*CO I /H I 1 I t /yI CHdO H H IIt can be concluded that the naphthylene residues must have adifferent structure in ang.- and Zin.-hydrocarbons derived fromanthracene; in the former it is symmetrical and in the latterunsymmetrical :It also follows that the Zin.attachment of rings t o anthracenestabilises the diradical state, whereas the aromatic state is favouredby their ang. attachment, which causes a general diminution ofreactivity and increase in stability. In accord with this is theobservation that the stability of dihydro-derivatives of these com-pounds goes hand in hand with their increasing meso-reactivity. Asimilar parallelism has already been recorded in the benzonaphth-azines .4Phenanthrene has clearly a less pronounced “ R ” phase thananthracene and is therefore less reactive in the meso-positions ;at the same time the 1 : 4-positions display a certain reactivity, e.g.,L. F. Fieser, J . Amer. Chem. SOC., 1931, 53, 2329; A., 1931, 1064.0. Hinsberg, Annalen, 1901, 319, 264; A ., 1902, i, 238.E. Clar and L. Lombartii, Uer., 1932, 65, [R], 1411 ; A., 1123RON. 167towards lithium,s and a second diradical phase (XI) may have to betaken into consideration :The structure (XI), like (X), is that of a naphthalene derivative, andthis is distinctly shown in the extinction curve of the hydrocarbon,in addition to the intense benzene band already noted in the anthra-cene spectrum. Chrysene (XII) evidently has a similar structurebut with a longer 9 : 10-diyl phase. Triphenylene shows a naphth-alene-like spectrum; its dirndical phase is evidently the 1 : 4-diyl(XIII) :(XII.) (XIII.)On the other hand, the curve of 2' : 3'-naphtha-2 : 3-phenanthrene(XIV), a deep orange, highly reactive substance, shows no vestigeof phenanthrene character ; it is just like that of 2 : 3-benzanthra-cene, but reverts t o the phenanthrene type when (XIV) combineswith maleic anhydride.CO-CH- IOn the assumption fhaf naphthalene has an unsyminetricalstructure derivable from that of o-benzoquinone, chryscne, phen-anthrene, and triphenylene could be respectively linked up with1 : 2-phenanthraquinone, 1 : 2-naphthaquinone, and 9 : 10-phen-anthraquinone ; the reduction potentials of these quinones are notsimply related to the log E ~ ~ ~ .as in the anthracenes and it is concludedthat in all hydrocarbons formed from naphthalene by ang. attach-ment of phenylene residues there is a double bond between any twoW. Schlenk and E. Bergmann, Annalen, 1928, 463, 84; A., 1928, 1031.7 L.F. Fieser and M. A. Peters, J. Amer. Chent. Roc., 1931, 53, 703; A . ,1033, 480168 ORGANIC CHEMISTRY .-PART 11.rings ; this explains the inability of these hydrocarbons to react withmaleic anhydride.Perylene presents special interest, being the basis of an importantgroup of colouring matters, and has been the subject of repeatedinvestigation. It is reduced by sodium in amyl alcohol,s givingsimultaneously an octahydro- and a tetradecahydro-compound,(XVI) and (XVIII) ; the first is formed in the same way as tetralinfrom naphthalene, and the formation of the second is like thehydrogenation of anthracene. It is suggested that the isomericforms of the hydrocarbon (XV) and (XVII) give rise to thesereduction products, just as two quinones, the 3 : 9- and the 3 : lo-,are formed on oxidation :/LkI3 -(XV.)7 81 11 111 \A/ 9 10(XVI.)I II I I/\,AII 1 I(XVII.) (XTX.)E. Clar regards formula (XV) as improbable, because the middlering is triquinonoid and, although the positions 6 : 7 and 1 : 12 areexactly equivalent, perylene reacts with only one molecule of maleicanhydride : this and the preferential formation of the 3 : 9-quinoneare only explained by the formula (XIX), which is also supported byspectrographic evidence. Thus (XV) contains two naphthaleneresidues which should find expression in the extinction curve andthis is not observed ; also, the spectrum of the 2 : 3 : 10 : ll-dibenzo-compound (XX) is very similar to that of the parent hydrocarbon,whereas the addition of two rings to the structure (XV) should causea marked alteration, similar to that observed on passing fromnaphthalene to phenanthrene. On the other hand, the curveof the 1 : 12-benzo-compound (XXI), though still of the perylene* A.Zinke and 0. Benndorf, Monatsh., 1932, 59, 241 ; A., 507.9 Ber., 1932, 65, [B], 846; A., 731RON. 169type, is less intense and evidently corresponds to a more saturatedstructure ; the bands are displaced towards the ultra-violet, doubtlessowing to the inclusion of the double bonds of the quinonoid systemin the new benzene ring.\/The anthraquinonoid formula (XVII), however, accounts for someof the reactions of perylene and is retained ; the distinctive characterof perylene is attributed to the conjugated system 3 : 10 (markedwith asterisks in formula XIX).Benzanthrone is easily reduced, taking up four atoms ofhydrogen,1° and behaves towards the Grignard reagent as an unsatur-ated ketoneg This behaviour is held to indicate that an equilibriumexists between the forms (XXII) and (XXIII) :9-(XXIII.)(XXIV.)Benzanthrene (XXIV), formed by replacing the carbonyl groupof benzanthrone by a methylene group, gives a typical naphthalenespectrum as would be expected from the above, but not from thealternative, anthracene-like, formulation ; its reactivity, e.g., withmaleic anhydride, is due t o the conjugated system marked byasterisks (compare perylene).This arrangement is not present inbenzanthrone, which, like phenanthrene and chrysene, does not forman addition product.These differences in reactivity are inexplicablelo J. v. Braun and 0. Bayer, Ber., 1925, 58, 2667; A., 1926, 172.F 170 ORGANIC CHEMISTRY .-PABT 11.on the basis of such formulz as J. J. Thornson’s (Pt5dulescu’sformula, XXVI) and give strong support for a formulation withactual double bonds disposed in a definite rnanner.llRecent experiments suggest that the electromeric phases in poly-cyclic compounds can have considerable stability ; l2 e.g., 2-retenol(XXV) does not couple with diazotised amines. The addition of thelatter would norma.lly occur either at a. double bond or at the end ofa conjugated system (Le., ortho or para to the hydroxyl), but neitherof these alternatives (marked by asterisks) is available and nocoupling occurs :MeD.Riidulescu and his collaborators l3 have measured the extinc-tion curves of a number of organic compounds, including ant,hracene,phenanthrene, pyrene, perylene, and 2 : 3-benzanthracene, andinterpreted them in the light of his theoretical views, which cannot beconveniently summarised . He also assumes valency tautomerismin polycyclic hydrocarbons and his conclusions, based on spectro-graphic work, are similar to Clar’s. His formulae are, however,different ; those adopted for anthracene, for example, are (XXVI)and (XXVII) ; the latter thus corresponds t o a para-bridged phase,and the former is J. J. Thornson’s original electronic formula :The spectral bands due to the bridged phase are the “ A ” bands inthe long-wave region attributed by Clar to the “ R ” phase.Theessential difference, therefore, between these formuh is that in thebridged one the odd electrons are controlled by both meso-carbons,whereas in the diradical formula t’hey are unshared. E. Clar andmany others are strongly opposed to the bridged formulation, whichin any case cannot account for the behaviour of compounds such as2 : 3 : 6 : 7-dibenzanthracene-9 : 10-diyl, the diradical nature ofwhich appears to be beyond dispute, and it seems to the Reporterl1 E. Clar, Ber., 1932, [B], 65, 1425; A., 1131.l2 L. F. Fieser and M. N. Young, J . Amer. Chern. SOC., 1931, 53, 4120; 44.,163. The experiments of A. A. Levine and A. G. Cole (ibid., 1932, 54, 338 ;A., 259) on the ozonisation of o-xylene prove the separate existence of the twoKe kul6 individuals.13 Ber., 1031, 64, [B], 2223 et seg.; A., 1931, 1361EON.171that, even if the existence of a bridged phase is admitt'ed, this cannotsupersede the diradical phase. The limiting phases required toaccount for the facts are the quinonoid and the diradical.The synthesis of coronene (hexabenzobenzene) by R. Scholl andK. Meyer l4 was made possible by the discovery l5 that the chloridesof anthraquinone-l-carboxylic acids react in two tautomeric forms(I) and (11), in both of which the chlorine can be replaced by anaromatic residue by the Friedel-Crafts reaction. Thus the chlorideof anthraquinone-1 : 5-dicarboxylic acid condenses with m-xylene,giving mainly the compound (111) ; this is oxidised by permanganateto the lactonic acid (IV), which is reduced with hydriodic acid andphosphorus to the hexacarboxylic acid (V) :(111.)l4 Ber., 1932, 65, [B], 902; A., 731.l6 R.Scholl, H. Dehnert, andL. Wanka, Annakn, 1932,493,56; R. Scholi,H. K. Meyer, and W. Winkler, ibid., 494, 201 ; A., 274, 617172 ORGANIC CREMISTRY .-PART II.The acid (V) can undergo ring closure in several different ways :20% oleum produces the compound (VI), which can be reduced withfurther ring closure to the acid (VII), from which, by heating withsoda-lime and copper powder, dibenzocoronene (VIII) is obtained.This, treated with dilute nitric passes through a red diquinone (IX),giving a blue vat, into the acid (X), which is decarboxylated withsoda-lime to coronene itself (XI).HI+P ___,4C0,H(VII.)H0,C C0,H @H0,C C0,H(X.1(VIII. )The hydrocarbon is pale yellow with a blue fluorescence, high-melting,and extremely stable. These properties are best expressed by theformula (XI), in which the central ring is a true benzene ring and thedouble bonds in the peripheral ring system form a continuousconjugated system.The dibenzo-compound (VIII) forms a complete contrast tocoronene; it is highly coloured, and its red solutions have a greenfluorescence and are readily oxidised in the air ; it is readily convertedinto the quinone (IX) and gives a dark blue dibromide (probably acarbonium salt). Such reactivity suggests a diradical structure(XII) ; the authors, however, prefer the alternative structure(VIII), which in their opinion suffices t,o explain the behaviour of thesubstance.They also suggest electronic formulae (XIII) and (XIV)for dibenzocoronene and coronene, in which the thick lines representthree-electron linkages and the fine lines two-electron linkagesxox. 173These are open to the mme objections aa the non-committal formulsdiscussed on p. 170.(XIII.) (XIV.)The Reporter wishes to thank Dr. J. W. Cook for many helpfulsuggestions.Rubrenes.No account has yet been given in these Reports of a remarkablegroup of substances discovered in 1926.16Rubrene, the type of the series, is a red hydrocarbon, C42H28,dissolving in benzene with a yellow fluorescence, and is prepared byheating the chloride (I) :On oxidation with chromic acid 17 rubrene gives o-dibenzoylbenzeneand carbon dioxide, in accordance with the structure (11; Ar = Ph)assigned t o it.18(1.) 2Ph2CC1*CiCPh = C4,H2, + 2HC1A r / '.Ph(11.) / (, \p" /&Rp + I /\ ~~-cOPh+CO, COPh\\ y$ ,,x& // \/-The dibenzofulvene formula (11) accounts satisfactorily for thecolour and fluorescence of rubrene and its mode of formation. Themechanism suggested for the latter as the most likely 18 involves,fist of all, the migration of the chlorine from the CI- to the y-carbonatom of the acetylenic compound, a process analogous to themigration of the hydroxyl group in the corresponding carbinol19(111), which readily passes into the ketone (IV) :1@ C. Moureu, C. Dufraisse, and P. M. Dean, Compt. rend., 1926, 182, 1440;A ., 1926, 945.1' C. Moureu, C. Dufraisse, and L. Enderlin, ibid., 1928,187,406; A., 1928,1127; C. Dufraisse and L. Enderlin, Bull. SOC. chim., 1932, [iv], 51, 132; A.,507.l* A. Willemart, Compt. Tend., 1928, 187, 385; A., 996; Ann. Chim., 1929,[XI, 12, 345; A., 1930, 334.19 K. H. Meyer and K. Schuster, BeT., 1922, 55, 819; A,, 1922, i, 556174 ORGANIC CHEMISTRY.-PART II.(111,) Ph,C(OH)PhCiCPh++ Ph2C:C:F*Ph =OHCPhPh,C:CH*COPh (1v.1CPh -1J. Robin 2O has made a further study of the formation of rubrene.He finds that the chloride (I) passes on keeping or on treatment withammonia or an aliphatic amine into a compound still containinghalogen, evidently formed by loss of one molecule of hydrogenchloride from two of the acetylenic compound.The new compoundpasses easily and quantitatively into rubrene and is formulated as ahydrochloride such as (V) or (VI) :PhC CPh CPh CPhC1the corresponding hydroxy- and et hoxy-compound have also beenobtained .With aniline (and other primary nrylamines) the chloride (I) formsa colourless (VII) and a yellow (VIII) nitrogenous derivative ; theformer is converted into the latter when heated with aniline hydro-chloride, and its hydrochloride undergoes the same change, especiallyin aniline solution.( ~ 1 1 . 1 Ph2(i*CICPh -+ Ph,C:CH$Ph (VIII.)NHPh NPh(VIII) can be hydrolysed to the ketone (IV) and its constitution isthus proved. The change therefore provides an interesting exampleof tautomeric change similar to that observed by Meyer andSchuster.lg Robin prefers to formulate it as involving the additionof hydrogen chloride, which is then eliminated as aniline hydro-chloride, on the ground that hydrogen chloride is indispensable forthis transformation :20 Ann.Chim., 1931, [XI, 16, 421; A., 1932, 260; compare C. Moureu, C.Dufraisse, and J. Robin, Compt. rend., 1929, 188, 1582; A., 1929, 922RON. 176Similarly, he assumes the intermediate addition of hydrogen chlorideto the chloride (I) in the conversion of the latter into rubrene :A/\ CPh reh=] --+ I II c -+\/ // \ \/ClCPh Cl*CPh 9The cyclisation of the compound may take place either before or afterthe elimination of hydrogen chloride .This view of the process is, like all similar theories, difficult todisprove, but nevertheless appears t o be superfluous when accountis taken of the proved occurrence of tautomeric changes in similarlyconstituted ethylenic systems ; in addition, Meyer and Schuster 19have shown that the analogous transformation of the correspondingcarbinol cannot be due t o the addition and elimination of water,since it proceeds most readily in the presence of strong dehydratingagents (acetic anhydride) and bears a marked resemblance to thechanges subsequently examined by Burton and Ingold.21The early experiments indicated that only chlorides of the typePh,CCl*CiCAr gave rise to rubrenes, but unsymmetrical compoundsof the type PhArCClCiCAr have since been successfully employed : 22the former can furnish only one rubrene ; a compound such as (IX)can give rise to three different rubrenes, one of which should beidentical with that obtained from the isomeric chloride (X). Thisprediction is verified by experiment and provides good support fort,he scheme of formation of rubrenes given above.\/ //CPh Ph Phc1 .*C 'IPhCPh21 J ., 1928, 904; A . , 1928, 634; compare Ann. Repo~ts, 1928, 25, 127.32 C. Dufraisse and &I. Loury, Cornpt. Tend., 1932,194, 1664, 1832; A . , 732176 ORQANIC CHEMISTRY.-PART If.C. Dufraisse and R. Buret 23 have prepared a rubrene (XI), inwhich two of the phenyl residues are replaced by chlorine, from theketone (XII) by the action of phosphorus pentachloride.Ph PhThe most remarkable property of the rubrenes is their ability toabsorb oxygen on irradiation; no reaction t'akes place in the dark.15A solvent such as benzene is also necessary ; no reaction takes placein water or acetic acid, in which rubrene is insoluble,24 but even arapidly oxidisable solvent such as benzaldehyde can be employed." Antioxygens '' such as quinol or resorcinol retard the reaction, buttheir effect is purely chemical and not due to screening.The completion of the change is marked by the disappearance ofcolour and especially fluorescence.The product is an oxide, RO,,crystallising with half a molecule of benzene or other solvent ofcrystallisation, from which it cannot be freed without decomposition ;it has the remarkable property of giving up its oxygen when heated,regenerating the parent hydrocarbon, the change being accom-panied by an emission of light.Its formation and dissociation cantherefore be written : R + 0, + lightThis dissociable dioxide, oxyrubrene, is reduced by zinc and aceticacid to a monoxide, metrubrene, which can also be obtained by themild oxidation of rubrene with permanganate or chromic acid; 25this can be reduced to rubrene but cannot be oxidised to oxyrubrene.Oxyrubrene is isomerised by the action of a Grignard reagent toisooxyrubrene, which no longer dissociates when heated, but canbe reduced to rubrene26 and therefore contains the same carbonskeleton. The formulae (XIII) and (XIV) are suggested for theseRO,.two oxides.Ph Ph Ph Ph// *A/\ II II C=C II I 4 4 A/\I II C-C II I \/yqV \\/Y.\r.Ph h h Ph(XIII.) Metrubrene (XIV.) isooxyrubrene23 Compt.rend., 1932, 195, 962.24 C. Moureu, C. Dufraisse, and C. L. Butler, ibid., 1926, 183, 101 ; A,, 1926,25 C. Moureu, C. Dufraisse, and L. Enderlin, ibid., 1929, 188, 1528; A.,$ 6 C. Dufraisse and &I. Bucloche, ibid., 1930, 191, 104; A., 1930, 1173.945.1929, 922EON. 177The relationship between rubrene, metrubrene, and oxytwbrene isremarkably like that between hamoglobin and its oxidation productsand the names of the compounds are intended to emphasise thissimilarity : 26Rubrene Hemoglobinred.Oxyhaemoglobin -3 Methaemoglobin Oxyrubrene -3 MetrubreneSuch reversible oxidations are unknown except in the respiratorypigments, in which this property has generally been connected withthe presence of iron in the molecule.Oxidations are usually accom-panied by a considerable liberation of energy and would not beexpected to be reversible for that reason.The heats of formation of rubrene and its oxides are as follows : 27Rubrene. Oxyrubrene. isooxyrubrene. Metrubrene.- 131 - 108.4 - 50.4 - 92-4The difference between the heats of formation of rubrene andoxyrubrene (23 cals.) is less than the heat required for t'he formationof an oxide ring (50 cals.) or a carbonyl group (100 cals.). Neverthe-less, the photochemical formation of oxyrubrene is exothermic andthe activating action of light can presumably be replaced by someother catalytic influence.An interesting transformation has been observed as the result ofthe prolonged action of a Grignard reagent on isooxyrubrene.28The product is a magnesium derivative which on treatment withwater gives phenol and a rubrene (I1 ; Ar = H) containing one phenylgroup less than the initial material; similarly, with iodine, an iodo-rubrene (11; Ar = I) k produced, and with carbon dioxide thecorresponding carboxylic acid :C,,H2,O2 + 2Mg = C36HB*Mg*OPh + MgO -+ PhOH + Ca6H2*.The compound (I1 ; Ar = H) is a typical rubrene, forming a dissoci-able oxide and having the other characteristic properties of the class.The iodo-compound (11; Ar = I) on treatment with sodiumet hoxide loses hydrogen iodide and forms a violet hydrocarbon (XV),the change probably taking place as follows : 29/-\I \=(27 c.Dufraisse andL. Enderlin, Compt.rend., 1930,191,1321 ; A., 1931, 171.2 8 C. Dufraisse and M. Badoche, ibid., 1931, 193, 242; A., 1931, 1151.*Q Idem, ibid., p. 529; A., 1931, 1407178 ORGANIC CHEMISTRY .--PART 111.The intensification of colour is to be expected, since a naphtha-quinone group is present in the new compound in addition to therubrene skeleton, which remains intact.Sterols and Rile Acids.The year under review has seen important developments of thissubject, discussion of which must be deferred until next year owingt o lack of space. A summary of the present position with regardto the bile acids and cholesterol has been published by A.Windau~.~OG. A. R. KON.PART III.-HETEROCYCLIC DIvrsroN.HETEROCYCLIC chemistry of to-day is almost synonymous with thechemistry of heterocyclic natural products. The past 50 yearshave been spent in perfecting the implements with which the organicchemist of to-day works.Discoveries of new methods of wideapplicability, such as the remarkable " diene '' synthetic andanalytic methodof 0. Diels and K, Alder,l must of necessity be fewand far betwecn. A new chemistry is, however, arising which has aboundless future before it-chemistry under natural or biologicalconditions. Already we have some striking examples of its applic-ability, especially in the field of alkaloidal chemistry,I n 1905 A. Pictet suggested that alkaloids arofie from proteinsand amino-acids and the view was developed by E. Winterstein andG. Trier,2 who appear to have been the first to suggest a biogeneticmechanism for the isoquinoline group. The thesis was however,put in an unassailable position by R.Robinuony3 who illustrated itby syntheses of tropinone 4 and +pelletierine by methods whichmight occur in the plant, and extended it by the synthesis of N -methylhomogranatoline 6 and some analogues of $-pelletierine.7From Angostura bark, E. Spath and his co-workers isolated andidentified a series of quinoline alkaloids which an account of thcirsimilarity of structure were thought to arise in the plant from somecommon parent benzene derivative, probably a derivative ofanthranilic acid. This line of thought has been followed up ex-perimentally by C. Schopf and G. Lehmmqs who find that when30 Z. physiol. Chem., 1932, 213, 147.1 Annalen, 1932, 498, 1, 16; A., 1144.J., 1917, 111, 876; A., 1917, i, 664.R.C. Menzios and R. Robinson, J . , 1921, 125, 2263 ; A ., 1924, i, 13%;.B. K. Blount and R. Robinson, J . , 1932, 1429; A . , 759." Die Alkaloido," 1910.Ibid., p. 782; A., 1917, i, 581,7 Idem, ibid., p. 2485; A . , 1147. Annulen,, 1932, 497, 7 ; A . , 104GKING. 179o-aminobenzaldehyde and acetone are brought together in diluteaqueous solution of p H 3, 5, 7 or 9 no quinaldine is formed but atpH 12-13 a smooth Friedlander synthesis occurs with formation ofquinaldine. If, however, acetone is replaced by acetoacetic acid,a 90% yield of quinaldine may be obtained afterpH 6.8.16 days at 25" and+ co, + 2H,OIf the reaction is carried out at pH 13, no loss of carbon dioxideoccurs and quinaldine-3-carboxylic acid is formed almost quant'i-tatively. In a similar manner 2-n-amylquinoline may be obtainedfrom hexoylacetic acid and o-aminobenzaldehyde.The isoquinoliiie group of alkaloids is exceedingly rich in itsvariety of examples and no one who surveys the different types canfail to be convinced that they owe their biogenetic origin to theunits suggested by Winterstein and Trier.The four main groupsare the papaverine group (I), the berberine group (11), the phenan-thripyridine group (111), and the morphine group (IV).NMeRO (111.)The units in papaverine and the phenanthripyridine groupare 3 : 4-dihydroxyphenylethylamine and 3 : 4-dihydroxyphenyl-acetaldehyde, the latter giving rise to the benzyl portion of themolecule, and in the berberine group the intervention of formalde-hyde is also necessary.The morphine group, which includessinomenine, also contains these two units, although in the case ofmorphine there is some doubt as to how the ethylamino side-chainarrives on the tertiary carbon atom. One plausible explanation ha180 ORGANIC CHEMISTRY.-PART III.been given by R. Robinson and S. Suga~awa.~ Morphine, however,is accompanied in opium by over 22 other alkaloids, the majority ofwhich are clearly framed from t’he same two units, so there can beno doubt whatever that morphine has a similar biogenetic origin.It is therefore surprising to find that H. Emde lo regards morphineas having arisen from three molecules of a hexose and one of methyl-amine with loss of carbon dioxide and water, and to visualise thishe postulates a pro-morphine (V) in the plant, which loses carbondioxide.The two intact hexose chains seen in (V) are shown in (VI)in bolder print and it will be at once clear that the scheme, whilstingenious, fails to carry conviction.A synthesis of an isoquinoline alkaloid under biological conditionshas not yet been effected, although a synthesis of tetrahydropa-paverine in 8 yo yield from homoveratraldehyde and homoveratryl-amine in the presence of 19% hydrochloric acid and at a water-bathtemperature, based on the Winterstein-Trier mechanism, has beeneffected by E. Spath and F. Berger.llCH, CH,Stereochemistry .An interesting example of spiro-asymmetry has been recorded by(Sir) W.J. Pope and J. B. Whitworth.12 spiro-5 : 5-Dihydantoin(I) has been resolved by crystallisation in alcoholic solution withtwo molecular proportions of brucine. The behaviour is somewhatnovel, for a hot solution of the components deposits a salt of theform 1B,ZA in an almost pure condition and on allowing the filtratelo Naturwiss., 1930, 18, 539; A., 1930, 1072. 9 J . , 1931, 3163; A., 174.11 Ber., 1930, 83, 2098; A., 1930, 1454.12 Proc. Roy. XOC., 1931, [ A ] , 134, 357; A., 171KING. 181to stand an almost pure salt of the form 2B,dA separates. Evidencehas also been obtained pointing to the existence of 3 tautomericforms (I), (11), and (111), for when the hydantoin (I) with similarNHCO NH.70 NH*CO>c(NH*$OH fJ*CO NH*R*OH(50-NH%O*NH do*m CO-N HO*C*NH%O*N(1.) (11.) (111.)rotations in alcohol, water, and pyridine is dissolved in one molecularproportion of sodium hydroxide solution the rotation falls to aboutone-half its value through formation of the keto-enolic form.When(I) is dissolved in two or more molecular proportions of alkali, thesolution is temporarily stable and shows a change of sign of rotationthrough formation of a di-enolic form. Incidentally the usefulnessof brucine for forming salts with hydantoins is thus well establishedand should lead to the direct resolution of hypnotics of the nirvanol,luminal type, examples of which are hitherto unrecorded.12QT. Nishikawa 13 obtained crystalline brucine salts from his a- and @-forms of C-methylbarbituric acid, but failed to resolve either,whilst C.M. Hsueh and C. S. Marvel l4 were unable to isolate stablesalts of alkaloids with ethyl-sec.-butylbarbituric acid.Another contribution l5 from the Cambridge laboratories fulfilsthe prediction made in an earlier paper l6 that quaternary salts of8-substituted derivatives of quinoline should exhibit a moleculardissymmetry which is dependent on restricted rotation about asingle bond. Previously it had been shown that benzenesulphonyl-8-nitronaphthylglycine (IV) could be resolved into dextro- andhvo-modifications and now it has been demonstrated that8-benzenesulphonylethylamino- 1 -ethylquinolinium iodide (V, R =Et) can also exist in optically active forms. In either case the groupon the adjacent peri- or l-position restricts the free rotation of thetrebly substituted nitrogen atom about the bond linking it to thenucleus.See, however, H.Sobotka, M. F. Holzman, and J. Iiahn, J . Amer.Chem. SOC., 1932, 54, 4697.lS Mern. Ryojun Coll. Eng., 1931, 3, 277.l4 J . Amer. Chern. SOC., 1928, 50, 855; A., 1928, 529.l5 W. H. Mills end J. G. Breckenridge, J . , 1932, 2209.W. H. Mills and K. A. C. Elliott, J., 1928, 1291; A . , 1928, 748; Ann.Iieports, 1928, 25, 117182 ORGANIC CHEMISTRY.-PART III.The determination of the configuration of optically activesubstances is an aid to classification, but it often taxes the ingenuityof chemists. There are two methods employed, both of limitedapplication. I n the first and more direct method the change ofrotation is observed in transformations which do not involve re-placements of groups directly attached to the asymmetric carbonatom and where optical inversion is presumably excluded. A classicexample is the demonstration by K.Freudenberg l7 that Z-glycericacid, Z-lactic acid, d-malic acid, and &tartaric acid all possess thesame relative spatial configurations of hydrogen atoms, hydroxyland carboxyl groups attached t o the asymmetric centre. The secondmethod, introduced by G. W. Clough,ls is a less direct one. It isassumed that “ the optical rotatory powers of similarly constitutedcompounds possessing the same configuration are in general in-fluenced similarly by t,he same changes in the external conditionsand also by the introduction of the same substituent into a givenradicle attached to the asymmetric carbon atom.’,Both methods have now been used by W.Leithe19 for a deter-mination of the configuration of coniine and a-pipecoline in terms ofamino-acids which are standards of reference. Willstgtter on theone hand showed that (+)*-conhydrine (VI) by oxidation gavea laevorotatory (-)-pipecolinic acid (VII); on the other hand,K. Loffler and G . Friedrich 2O showed that (+)-conhydrine could beconverted through P-coniceine (VIII) into (-)-coniine (IX), so that(+)-coniine corresponded in configuration t o (+)-pipecolinic acid.This acid, however, by an application of Clough’s principles has been17 Ber., 1914, 47, 2027; A., 1914, i, 924.18 J . , 1918, 113, 526; A., 1918, ii, 255.20 Ibid., 1909, 42, 107; A., 1909, i, 180.* (+) and (-) indicate the observed signs of rotation of the materiaIsCompare A. Wohl and K. Freudenberg, Ber., 1923, 50,lD Ber., 1933, 05, 927; A., 865.under discussion.309; A., 1923, i, 182KING. 183shown to belong to the d-series of amino-acids. (+)-Pipecoline(Zmethylpiperidine) is also spatially related to (+)-coniine, as isshown by the analogous optical behaviour of the bases in solventsand as salts, so that (+)-coniine, (+)-pipecoline, and (+)-pipe-colinic acid belong stereochemically to the d-series of amino-acids.In a somewhat similar manner W. Leithe 21 has been able todetermine the configuration of a-phenylethylamine and of bases ofthe type of laudanosine and tetrahydroberberine.When (-)-a-phenylethylamine (X) is benzoylated and the benzoyl product madesusceptible to oxidation by introduction of a hydroxyl group in theplienyl nucleus, it can be oxidised to E( +)-benzoylalanine identicalwith that prepared from natural l( +)-alanine. (-)-a-Phenylethyl-amine has therefore the Z-configuration. Now bases of the laudan-osine and tetrahydroberberine type may be regarded as substituteda-phenylethylamines, as the following formulze (X-XIII) show.I CH2Ph(XIII.)On this basis Leithe, using Clough’s principles, has shown thatE( -)-phenylethylamine (X), its N-ethyl derivative (XI), (-)-1-methyltetrahydroisoquinolhe (XII), (- )-protolaudanosine (XIII),and ( - )-tetrahydroprotoberberine are all configurationally similarand thus belong to the Z-alanine series.During the past year two attempts have been made to classifythe complex stereochemical system of cinchona alkaloids withtheir four asymmetric centres.22 Consideration of cases of resolutionof such substances as l-methylcyclohexylideneacetic acid of Pope,Perkin, and Wallach, where there is no asymmetric carbon atom,raises doubts as to the legitimacy of attributing particular signs ofrotation to asymmetric centres in such substances as the cinchonaalkaloids.The results obtained, however, when accepted withreserve do seem to justify the means adopted. P. Rabe 23 with awealth of experimental material accumulated over a span of yearshas followed the methods used by H. King and A. D. Palmer 24 indeducing the sign of the contribution of the asymmetric centres in21 Ber., 1931, 64, 2827; A,., 177.22 Cornpare J.Keriner, Ann. Reports, 1022, 19, 157.23 Annalen, 1932, 492, 242 ; A., 289. 84 J., 1922, 121, 2677184 ORGANIC CHEMISTRY.-PART 111.these alkaloids, arrives at the same conclusions, and has beenable to classify sixteen closely related alkaloids. The principlesinvolved are briefly these. The cinchona alkaloids of generalformula (XIV) all give rise to weakly rotatory “ toxines ” (XV),N--CH( OH)---(4)(XIV.)the same “ toxine ” being obtainable from four stereoisomericalcohols (XIV). Thus hydrocinchonine, hydrocinchonidine,epihydrocinchonine, and epihydrocinchonidine, where R = H,R’ = Et, a 1 give rise t o the same hydrocinchotoxine, so that t,he(XV.) AH’differences between the alkaloids must lie in the spatial arrangementaround carbon atoms 3 and 4.Hydrocinchonine and epihydro-cinchonine give rise by two different experimental methods t o thesame highly dextrorotatory deoxyhydrocinchonine, where -CH(OH)-has been replaced by -CH,- just as hydrocinchonidine and epihydro-cinchonidine give rise to an isomeric deoxyhydrocinchonidine oflaworotation. The assumption is therefore made that the asym-metric carbon atom 3 is dextrorotatory in its contribution in theformer pair of alcohols and lzevorotatory in the latter pair. Hydro-cinchonine and epihydrocinchonine can thus only differ in thearrangement around carbon atom 4, and the same applies tohydrocinchonidine and epihydrocinchonidine. Another assumptionis now necessary.It i s assumed that, since hydrocinchonine hasa high dextrorotation and epihydrocinchonine a low dextrorotation,carbon atom 4 is dextrorotatory in its contribution in hydro-cinchonine and hvorotatory in its contribution in epihydro-cinchonine ; similarly, since epihydrocinchonidine has a dextro-rotation and hydrocinchonidine a laevorotation, carbon atom 4 isdextrorotatory in the former and lzevorotatory in the latter. Thesame reasoning has been applied by Rabe and his pupils to sixteen oKING. 185these closely related bases, all of the general formula (XIV), with thefollowing results.Optical sign of Optical sign ofc3. c4. c3. c4.R = OMe; R’ = CHXH,. R = OMe; R’ = Et.- - Hydroquinine ............ - -...............epiHydroquinine ...... - + + epiQuinine -epiQuinidine ............... + - ......Quinidine Hydroquinidine ......... + + .................. + -t- - Hydrocinchonidine ...... _ -epicinchonidine ......... - + epiHydrocinchonidine - -tepicinchonine ............ -t- - epiHydrocinchonine ... + -Cinchonine Hydrocinchonine ...... + +Quinine ..................epiHydroquinidine + -R = H; R‘ = CHZCH,. R = H; R’ = Et.Cinchonidine ........................... + +It will be noted that in the eight common naturally-occurringalkaloids the sign of rotation attributed to carbon atom 3 is alwaysthe same as that of carbon atom 4. In further just’ification of thissystem of classification it may be pointed out that the same resultfor the eight common alkaloids may be arrived at from a consider-ation of the magnitude of the specific rotations alone of the fourisomeric bases in any one group ; quinine and quinidine, for example,have the highest and the lowest lzevo- and dextro-rotations respec-tively of the four isomerides.P.Rabe and S. Riza25 have extended the results to the fourstereoisomeric rubanols (XIV ; R = R’ = H), obtained by syntheticmethods, in which carbon atom 1 has lost its asymmetry, andhave again obtained concordant results.A dserent view is adopted by H. Emde26 but on a less satis-factory basis. He invokes the principle of optical superposition ina form used by Hudson and then abandons its legitimate deductions.Cinchona alkaloids only occur in nature in two out of the possiblesixteen optically active forms and since, according to Emde, allexamples of epimerism among naturally occurring substances are dueto secondary carbinol groups, this must be the case in the cinchonaalkaloids.The asymmetric centre 3 must therefore have the sameconfiguration in all the naturally occurring alkaloids and the observedisomerism must be entirely dependent on the spatial arrangementaround carbon atom 4. On such a basis it is difficult to see howEmde would approach the problem of the configuration of carbonatoms 3 and 4 in the epimeric bases.Dipole measurements have been invoked2’ in determining thespatial- arrangements around the sulphur atoms in thianthren25 Anmlen, 1932, 496, 151 ; A., 865.26 Helv.Chim. Acta, 1932, 15, 557; A., 759.27 E. Bergmann end M. Tschudnowsky, Ber., 1932, 65, 458; A., 507186 ORGANIC CHEMISTRY .-PART III.(XVI) and the two isomeric disulphoxides (XVII) and (XVIII) withdipole moments respectively of 1.68, 1-7, and 4.2.(2x1.) (XVII.) (XVIII.)Thianthren cannot be planar and it is supposed that the moleculeis slightly folded at the sulphur atoms. The disulphoxide of struc-ture (XVII) has practically the same dipole moment as thianthren,since the moments of the SO groups compensate one another. I n(XVIII), however, the SO groups reinforce one another. It naturallyfollows that the oxygen atoms cannot be in the same plane as the CSbonds. A similar conclusion has of course already been reachedthrough the resolution of sulphoxides by Phillips and Kenyon. TheGerman workers are, however, loth to accept the postulate of semi-polar double bonds as an explanation of the phenomena, as in theiropinion the formation of a decet of electrons around the sulphuratom is not excluded.Oxide Rings in NaturaE Products.Fish Poisons.-This group of natural poisons has attractedconsiderable attention during recent years through the need forefficient insecticides.An arrow poison used by the Malays underthe name Ipoh is said t o be obtained from Derris elliptica (Fam.Leguminosm), known to the Javanese as tuba root and employed bythem as a fish poison. The active principle, tubatoxin, was firstisolated by T. Ishikawa 28 and later shown by T. Kariyone, K.Atsumi, and 11.Shimada 29 to be identical with rotenone, firstlisolated by K. Nagai 3O from Millettia taiwaniana Hayafn, obtainedfrom Formosa.Through the efforts of the chemists of four different nations theconstitution of rotenone is now agreed to be (I). On gentle oxidationit readily loses two hydrogen atoms with formation of dehydro-rotenone 31 (11). The latter substance can be saponified by alcoholicpotassium hydroxide with formation of derrisic acid (III), withaddition of two molecules of water, from which dehydrorotenone (TI)28 Jap. M e d . Lit., 1917, 1, 7 ; A., 1918, i, 94.29 J . Pharm. SOC. Japan, 1923, 500, 739; A . , 1924, i, 950.30 J . Tokyo Chem. Soc., 1902, 23, 744.31 A. Bntenantlt, Aw)zcrlc>i, 1928, 464, 2 5 3 ; A . , 1928, 1017KING.187is re-formed by the action of acetic anhydride,32 possibly throughformation of an intermediate lactone.Me0(11.1CH, 0CH,-CH*CMe:CH,Me0(111.)The molecule of derrisic acid is seen to be built up of two benzenenuclei with various addenda. When oxidised with hydrogenperoxide, derrisic acid yields derric acid33 (IV), and on furtheroxidation with permanganate this gives the lower homologue,risk wid (V), which can be decarboxylated to decarboxyrisic acidMe0()0*CH2*C02W g 8 3 0 * & 2 * 8 0 2 H CH COH Me0 MeO()O*CH,*C02H CO,H Me0(IV.) (V.) (VI.)The constitutions deduced for these acids by LaForge have sincebeen confirmed by syntheses of derric acid and risic acid by A.Robertson,35 of risic acid by S. Takei, S. Miyajima, and M.O ~ O , ~ ~and of decarboxyrisic acid by LaF~rge.~'The determination of the constitution of the second half of themolecule of rotenone proved to be a more difficult problem. By32 F. B. LaForge and H. L. Haller, J . Ainer. Chem. SOC., 1932, 54, 813;A., 401.33 F. B. LaForge, ibid., 1931, 53, 3896; A., 1931, 1415.s4 F. B. LaForge and L. E. Smith, J . Amer. Chem. SOC., 1930, 52, 2878(VI) *A., 1930, 1187; S. Takei, S. Miyajima, and M. Ono, Ber., 1931, 64, 248;A., 1931, 490.3 5 J., 1932, 1380; A., 751.37 J . Amer. Chem. SOC., 1931, 53, 3896; A., 1931, 1415.36 Ber., 1932, 65, 1041 ; A., 860188 ORGANIC CHEMISTRY.-PART IT”.the action of alcoholic potassium hydroxide on rotenone S. Takei 38had obtained an acid called tuhaic acid, now known to be C,,H1,04.39The constitution of tubaic acid (VII) was finally proved by H.L.Haller and F. B. LaForge40 by a close study of its properties,although much supplementary experimental evidence was suppliedby S. Takei and his co-~orkers.~l Tubaic acid was optically activeand contained a hydroxyl group, an indifferent oxygen atom, anda double bond which could readily be reduced, yielding dihydro-tubaic acid. On further hydrogenation it gave tetrahydrotubaicacid (VIII) , which could be decarboxylatecl to 24soamylresorcinol(IW‘CH2dH*CMe:CH2 hH,*CH,-CHMe, hH,*CH,*CHMe,(VII.) (VIII.) (IX.)The orientation of groups in (IX) is established, since it is knownthat alkyl groups meta to hydroxyl groups inhibit the fluoresceintest and one of the hydroxyl groups in (VIII) is indifferent tomethylating agencies.The loss of optical activity observed when(VII) is converted into (VIII) shows that the isopropenyl side-chain is attached as shown and not t o the neighbouring carbonatom.The constitutions of the two halves of the rotenone moleculehaving been determined with a considerable degree of certainty,a satisfactory formula (I) was suggested almost simultaneously byF. B. LaPorge and H. L. Haller,42 by A. Robertson,43 by S. Takei,X. Miyajima, and M. O ~ O , ~ ~ and by A. Rutenandt and W.M~Cartney.~~The numerous degradation products have almost all been assignedconstitutions which in some cases have been confirmed by elegantpartial syntheses. Thus, when rotenone is treated with zinc dustand alkali, it yields derritol46 (X) with a loss of two carbon atoms.The properties of this substance are consistent with the structureshown, for when treated with ethyl chloroacetate it yields derrisic38 J .Chem. SOC. Japan, 1923, 44, 841; Biochem. Z., 1925, 157, 1; A . ,39 T. Kariyone and S. Kondo, J . Pharm. SOC. Japan, 1925, 518, 376.40 J. Amer. Chem. SOC., 1931, 53, 4461; 1932, 54, 1988; A . , 165, 739.41 Ber., 1928, 61, 1103; 1929, 62, 3030; A., 1928, 765; 1930, 216.42 J . Arner. Chem. SOC., 1932, 54, 810; A., 401.43 J., 1932, 1380; A., 751.45 Annalen, 1932, 494, 17 ; A., 619.4 6 A. Butenandt, ibid., 1928, ‘464, 259; A . , 1928, 1017.1925, i, 761.44 Ber., 1933, 65, 1044; A., 860KING. 189acid (111) mixed with dehydrorotenone 47 (11).The constitutionof rotenonone, a product of strong oxidation of dehydrorotenoneMe0and differing from it by the substitution of an oxygen atom fortwo hydrogen atoms, has been finally solved by the observationthat rotenonone is hydrolysed almost quantitatively by alcoholicpotassium hydroxide into derritol (X) and oxalic acid.47 Con-versely, rotenonone has been synthesised from derritol (X) andethyl oxalate or chloro-oxalyl ethyl ester, so its constitution mustbe that represented by 48In the practical application of 'derris extracts as insecticides itwas observed by American workers that extracts poor in rotenonecould be very active insecticides. E. P. Clark, following up theclue, showed that derris roots also contained toxicarol, deguelin,and tephr0sin,4~ the first two of which a t a dilution of 1 in 5 x lo6will kill goldfish in 3 or 4 hours.Clark was also able to show thatcube', a Peruvian fish poison, contained rotenone, tephrosin, anddeguelin and that Tephrosia toxicaria from British Guiana containedtoxicarol and d e g ~ e l i n , ~ ~ whilst T . vogelii from Africa and Sumatracontained tephrosin and deguelin,51 an observation made indepen-dently by A. Butenandt and G. Hilgetag.52There is a close similarity in the molecular formulze of thesesubstances, suggesting a phytochemical relationship. Such is thecase, for E. P. Clark has shown that all three substances give thesame derric and risic acids as were obtained from rotenone.53 Theconstitution of deguelin (XII) has been determined by Clark.54474 849505152sa54S.Takei, S. Miyajima, and M. Ono, Ber., 1932, 65, 1043; A., 860.F. B. LaForge, J . Amer. Chem. SOC., 1932, 54, 3377; A., 1039.Science, 1930, 71, 396; A., 1930, 967.J . Amer. Chm. Soc., 1930, 52, 2461; A., 1930, 1223.Ibid., 1931, 53, 729; A., 1931, 491.Annalen, 1932, 495, 172; A., 751.J . Amer. Chern. SOC., 1932, 54, 1600; A., 619.Ibid., p. 3002; A., 950190 ORGANIC CHEMISTRY.-PART III.It passes on gentle oxidation with loss of two hydrogen atoms intodehydrodeguelin, which is also obtained by dehydration with aceticMe0 Me0(XII.) (XIII.)anhydride of tephrosin (XIII) and is~tephrosin.~~ In these twosubstances the relative positions of the added elements of the watermolecule are undecided.Dehydrodeguelin, which is optically in-active, on oxidation with permanganate yields a new tricarboxylicacid, nicouic acid, C,,H,,O,, which at its melting point losesa-hydroxyisobutyric acid. As it also gives the fluorescein reactionfor resorcinol and yields the latter on boiling with aniline, theconstitution (XIV) is assigned to it.Furthermore, when rotenone is reduced catalytically, it givestogether with other products a so-called rotenonic acid 56 (XV),which was shown by H. L. Haller 57 to isomerise under the influenceof acetic and sulphuric acids into P-dihydrorotenone (XVI) isomericwith dih ydrorotenone. p -Dihydrorot enone on gentle oxidationgave dehydro-P-dihydrorotenone, which proved to be identical withdihydrodehydrodeguelin obtained by Clark by catalytic reductionof dehydrodeguelin .Me0 Me0These observations prove conclusively that the oxide ring inthe second half of the molecule is a six-membered ring as shown.The same constitutions for deguelin and tephrosin on the basis ofClark's experimental results have also been advanced by A.Robert-55 E. P. Clerk and H. V. Claborn, J . Amer. Chem. Soc., 1932, 54, 4455.G6 F. B. LaE'orgc aiicl L. E. Smith, ibid., 1929, 51, 2574; A., 1181.o 7 Idem, ibid., 1931, 53, 733; A., 1931, 491KING. 191son 58 and by A. Butenandt and H. Hilgetag.59 The constitutionof toxicarol is still unsettled, the formulm proposed by Clark andby Butenandt and Hilgetag 59 showing considerable difference.From the solution of the constitution of these fish poisons andinsecticides arises the interesting and practical question as to thesimplest structure which retains these properties.Peucedanin, abitter principle from Peucehnurn oficinale, is known to be it fishpoison and recently 60b its constitution has been elucidated as(XVII). Still simpler structures but of the same type are pos-sessed by bergapten (XVIII) and xanthotoxin (XIX), which arealso known to be fish poisons.6lCH Me0 CHCHCunnabinol.-A resinous secretion of Indian hemp (CannabisIdicu) is known as hashish or bhang and has been used as anintoxicating drug for centuries in the East. By fractional dis-tillation of the resin, T. B. Wood, W. T. N. Spivey, and T. H.Easterfield 62 were able to isolate a series of substances, one ofwhich, a high-boiling oil, they concluded to be the active principle.Cannabinol is a constituent of this oil and is isolated as its crystallineacetyl derivative.On oxidation of the active fraction with nitricacid, nitrocannabiiiolactone, CI,HllO,N, was obtained, from whichcannabinolactone, C11H,202, was isolated by removal of the nitro-group. Cannabinolactone gave m-toluic acid on fusion withpotassium hydroxide. The constitution of cannabinolactone hasnow been definitely determined by R. S. Cahn63 and confirmedby F. Bergel and K. Vogele 64 by synthesis. Cahn found that5* J., 1932, 1384; A., 751.6o J. Amer. Chem. SOC., 1932, 54, 2537; A . , 855.60a Trier, “ Chem. d. Pflanzenstoffe,” 1924, 268.608 E. Spiith, K.Klager, and C . Schlosser, Ber., 1931, 64, 2203; A., 1931,61 C. Pomeranz, Monatsh., 1893, 14, 29; H. Thoms and E. Baetcke, Ber.,62 J . , 1899, 75, 20.63 J . , 1930, 986; 1931, 630; 1932, 1342; A . , 1930, 913; 1931, 625; 1932,64 Annalen, 1932, 493, 250; A . , 382.59 Annalen, 1932, 495, 172; A , , 761.1298.1912, 45, 3705; A., 1913, i, 192.747192 ORGANIC CHEMISTRY .--PART 111.nitrocannabinolactone could be converted into the correspondinghydroxycannabinolactone, which on fusion with potassium hydr-oxide gave 6-hydroxy-m-toluic acid (I), identical with the syntheticmaterial, and acetone. Wood, Spivey, and Easterfield were alsoable to oxidise cannabinolactone to cannabinolactonic acid (11),which Cahn has shown to be further oxidised to trimellitic acid,Me CO,H Me Mebenzene-1 : 2 : 4-tricarboxylic acid, whence it follows that canna-binolactone must be represented by (111).The synthesis by F.Bergel and K. Vogele G4 starts from p-cymene-3-carboxylic acid(IV), which is oxidised by chromic acid to cannabinolactone (111)and other products. On further oxidation of the synthetic materialwith alkaline permanganate cannabinolactonic acid (11) was pro-duced identical with a product obtained by G. Bargellini andG. Forli-Forti 65 from 4-aminodimethylphthalide 63 and also identicalwith an acid obtained from santonin by Cannizzaro and Gucci.The constitution of cannabinol has not been determined withcertainty, but sufficient is known to make the constitution (V)possible.63 Cannabinol can be acetylated, forms a methyl ether,and can be oxidised to n-hexoic acid, a product previously reportedby earlier observers as being obtainable from crude high-boilingcannabis resins.The positions assigned to the hydroxyl and then-amyl group are in accord with the products of oxidation andnitration and with the observation that cannabinol does not reactwith diazomethane in ethereal solution and does not dissolve insodium hydroxide solution.Simple Furan Derivatives.-Y. Asahina and collaborators 66showed that elsholtxione, a ketone obtained from the essential oil6 5 Gazzettu, 1910, 40, ii, 74; A . , 1910, i, 744.66 Arch. Pharm., 1914, 252, 435; A., 1915, i, 429; J . Pimrm. SOC. Japan,1922, 485, 565; A., 1922, i, 1047 ; Acta Phytochim., 1924, 2, 1 ; &4., 19334, i,976KING.193of Elsholtzia cristata, Willdenow, was almost certainly 3-methyl-2-fury1 isobutyl ketone (I).Me -Me -Me -I' IICO*CH2*CHMe, II I1CO2H I1 llCN Y (11.) Y (111.) v 0 (1.)This has been confirmed by its synthesis 67 from p-methylfuran,which by Gattermann's method gave 3-methylfurfuraldehyde andthis on oxidation gave elsholtzic acid (11) identical with an acidobtained from elsholtzione by Asahina. 3-Methylfurfuraldehydewas converted through its oxime into the nitrile (111), and asynthesis of elsholtzione (I) effected by allowing the nitrile toreact with isobutylmagnesium bromide.6'T. Reichstein and H. Zschokke 68 have also synthesised furan-(3-carboxylic acid in two ways and have shown that it is identicalwith the naturally occurring furan- p-carboxylic acid isolated byH.Rogerson from Euonymus atropu~pureus.~~ The synthesis ineach case depends on the partial decarboxylation of furandicarb-oxylic acids (IV) and (V) with loss of the carboxyl group in thea-position. The former dicarboxylic acid was obtained by F.Feist 70 by the action of potassium hydroxide on methyl bromo-coumalate (VI). The second synthesis depends on the intermediate,OH 0, q C o r -3 C02€€*C\~C*C0,H GH PH -+ CO,H.C!.JICH CH/\C-CO,HC02Me*C1!,&B(VL) CHsynthesis of the diethyl ester of 3-carboxyfuryl-%acetic acid (VII)by condensation of chloroacetaldehyde and ethyl acetonedicarb-oxylate. This ester on saponification and decarboxylation gaveQHO yH,*CO,R C0,R -CO,HCH, CO*CH,*C02R + It IIc~,.co~R + II ltMe\Cl Y (VII.) \/ (VIII.)2-methylfuran-3-carboxylic acid (VIII), which could be oxidisedto furan-2 : 3-dicarboxylic acid.67 T.Reichstein, H. Zschokke, and A. Goerg, Helv. Chim. A N , 1931, 14,1277; A., 166.68 Ibid., 1932, 15, 268; A., 519.70 Ber., 1901, 54, 1992; A., 1901, 657.69 J., 1912, 101, 1044.REP.-VOL. X-. 194 ORGANIC CHEMISTRY.-PART III.Anthocyanins.-The past year has witnessed the culminatingpoint in the chemistry of the anthocyanins. R. Robinson andA. R. Todd ‘1 have succeeded in synthesising the five diglucosidesknown as hirsutin, malvin, pelargonin, peonin, and cyanin chlor-ides, identical with the products from natural sources. All proveto be p-diglucosides substituted in the 3- and the 5-position of theanthocyanidin nucleus.Great experimental difficulties had to beovercome in the preparation of the intermediates, which whenonce obtained were condensed in the usual manner, of which thesynthesis of pelargonin may be taken as a typical example. 2-0-Monoacetyl- p-glucosidylphloroglucinaldehyde (I) and a-O-tetra-acetyl-~-glucosidoxy-4-acetoxyacetophenone (11) were condensedin dry ethyl acetate by hydrogen chloride. The resulting flavyliumsalt (111) was kept in alkali in a hydrogen atmosphere t o removeacetyl groups and acidified. Pelargonin chloride (IV) then separatedin a state of purity.c1 c1P+ +-IH O e o H H o d e o A c0 C6H1105 $) C6Hi’0(0Ac)4b6H1105 (Iv.) C6H,,O,*OAc (111.)c1H o d - 6 OH OH (V.)5)C6H 1 lo 5In a similar manner, by use of the appropriate initial materialscyanenin chloride (V) and malvenin chloride, the partial hydrolyticproducts of cyanin and malvin chlorides have been synthesised.72If confirmation were needed of the view that malvin is a digluco-side containing the glucose residues attached to different hydroxyl7 1 J., 1932, 2293, 2299, 2488; A., 1140.72 A. Le6n and R. Robinson, J . , 1932, 2221; A., 1038"NQ. 195groups, such has been furnished by P. Karrer and G. de Me~ron.'~Karrer and his collaborators had previously 74 shown that antho-cyanins could be oxidised by hydrogen peroxide and that in twocases it was possible to isolate crystalline intermediate products,malvone and hirsutone, from malvin and hirsutin respectively.A re-investigation of malvone and hirsutone by Karrer and deMeuron has demonstrated that these ketones readily give a quanti-tative yield of syringic acid and glucose by solution in 2N-sodiumhydroxide a t room temperature.If treated with phenylhydrazine,they readily give up one molecule of glucose as phenylhydrazoneand on acid hydrolysis of the residue the remaining glucose groupmay be found. For this reason malvone (VI ; R = H) and hirsutone(VI ; R = Me) are now regarded as esters of glucose and of syringicacid, with glucose attached to different parts of t,he molecule.n OMeBy application of the phenylhydrazine test to the solutions con-taining the hydrogen peroxide oxidation products of peonin, cyanin,and monardin it has been demonstrated that these three antho-cyanins also contain two glucose groups in different positions inthe molecule and one of them must be attached to position 3, aconclusion in agreement with the syntheses of Robinson and Todd.Nitrogenous .Anthocyanins.-In 1918 R. Willstatter and G.Schudelshowed how many pigments, rosaniline, methylene-blue, and others,could be removed from aqueous solution by extraction with anorganic solvent containing picric or dichloropicric acid.75 Thistechnique was applied by Schudel 76 to the colouring matter ofbeetroot (Beta vulgaris) and resulted in the isolation of an unstablenitrogenous diglucosidic anthocyanin, named betanin chloride,which gave betanidin chloride on hydrolysis. Similar pigmentswere found in Celosia cristata and in winter-spinach (Atriplexhortensis atrosanguineus) .On the synthetic side L.R. Ridgway and R. Robinson 77 hadprepared 3-carbethoxyamino-4'-methoxy-8-ethoxy-2-phenylbenzo-pyrylium chloride (VII) from 2-hydroxy-3-ethoxybenzaldehyde anda-carbethoxyamino-p-methoxyacetophenone, but attempts to obtain73 Helv. Chim. Act&, 1932, 15, 507, 1212; A., 520.74 Ibid., 1927,10, 729; A., 1927, 1197.75 Ber., 1918, 51, 782; A., 1918, i, 399.7 7 J., 1924, 125, 2240; A., 1925, i, 54.'~3 Dissert., Ziiricli196 ORGANIC CHEMISTRY .-PART III.the 3-aminoflavylium salt were unsuccessful, the amino-groupbeing replaced by hydroxyl. R. Robinson and (Mrs.) A. M. Robin-c1FO*C,H,*OMe - HO CH,*NH*CO,Etson 78 have now succeeded in preparing pure 4'-aminoflavyliumperchlorates (VIII) and (IX) by use of 4-aminoacetophenones.c10,.-%c10,+-l(VIII.) (IX.)It is of interest that in these substances the amino-group is abetter auxochrome than hydrosyl, and with this may be coupledthe observation that betanidin is the bluest-red of all the antho-cyanidins.The authors regard it as possible that the dihydroxy-aminoflavylium salt (IX; R = H) may be identical with betanidin,since the colour reactions and the absorption spectra, in the visibleregion, of extracts of beet and atriplex in 0.1% hydrochloric acidand of 4'-amino-3 : 7-dihydroxyflavylium chloride showed closecorrespondence. Later unpublished work would, however, seemto suggest a possibility of betanidin being identical with a trihydroxy-flavylium salt.79A1 kuloids.Derivatives of I%doZe.-The developments in alkaloidal chemistrywithin the last few years have shown that there are a number ofalkaloids containing the indole nucleus, derived presumably fromtryptophan. The constitution of very few of these alkaloids isknown with certainty ; exceptions are physostigmine and theharmine group. The following account is confined to the moreimportant developments in this field.Ergot AZEaZoids.-It may be recalled that for a number of yearsergotoxine and ergotinine were the only two definitely recognisedalkaloids of ergot. In 1922 A. Stoll 8o isolated a second pair ofisomeric alkaloids, ergotamine and ergotaminine, which chemicallyand pharmacologically show very close similarities to the formerpair of alkaloids.S. Smith and G. M. Timmis 81 were able to78 J., 1932, 1439; A., 750.61 J . , 1930, 1393; 1031, 1888; .4., 1930, 1050; 1931, 1171.78 Naturwiss., 1932, 33, 613.Schtoei;.. Apoth.-Ztg., 1922, 60, 341; A., 1923, i, 137RING. 197confirm the isolation of ergotamine and ergotaminine but onlyfrom unofficial ergots, such as that growing on a New ZealandPesturn. In 1931 the same observers showed that ergotoxine wasaccompanied by two alkaloids, ergotinine and +-ergotinine, theproportion varying in different ergots. Both ergotinine and $-ergo-tinine are converted into ergotoxine by boiling with alcohol" andphosphoric acid and +-ergotinine is partly converted into ergotinineby boiling with methyl alcohol. The composition of these alkaloidsis still doubtful.A. Soltys 82 gives good reasons for concludingthat ergotamine and ergotaminine may be represented by theformula C,3H350,N5 and ergotinine by C35H3905NS. He also findsall four alkaloids to be phenolic or weakly carboxylic, to yieldammonia on hydrolysis, and to give benzoic acid on oxidationwith permanganate and p-nitrobenzoic acid on oxidation withnitric acid. W. A. Jacobs 83 confirms the latter observation onergotinine and has isolated a new base, C,,H908N, containing threecarboxyl groups and one N-methyl group by the action of nitricacid. On the other hand, Smith and Timmis 84 have shown thatall four alkaloids on hydrolysis with alcoholic potassium hydroxidegive ammonia and a base ergine, Cl,Rz1ON3, which likewise con-tains one N-methyl group and gives many colour reactions associ-ated with indole derivatives.The only other important observationbearing on the constitution of these alkaloids is the well-knownone of G. Barger and A. J. E w i n ~ , ~ ~ that ergotoxine and ergotininegive isobutyrylformamide, CHMe,*CO-CO*NH,, on heating. Thisis probably the amide group present in the molecule of these alkaloidswhich appears as ammonia on hydrolysis.Physoatigmine.-The pioneering experimental work of G. Bargerand E. StedmanS6 and of M. and M. Polonowski*' led to theelucidation of the structure (I) now accepted for physostigmine(eserine). The greatest difficulty in the synthesis of such a structuremight be anticipated in the closure of p-substituted dihydroindolesMewith formation of the third ring.MeThis has, however, been accom-plished by three different methods which have led to the synthesis82 Ber., 1932, 65, 553; A., 629.83 J .Biol. Chem., 1932, 97, 739; A., 1147.84 J., 1932, 763, 1543; A., 526, 759.8 6 J . , 1925,127, 247; A., 1925, i, 392.87 Compt. rend., 1924,178, 2078; 179, 334; A . , 1924, i, 1094.J., 1910, 97, 290198 ORGANIC CHEMISTRY .-PART III.of structures related to or identical with those of physostigminederivatives.T. Hoshino and K. Tamura 88 allowed p-indolylethylamine t oreact with excess of methylmagnesium iodide (4 mols.) and treatedthe product with methyl iodide. A 30% yield of dinordeoxy-eseroline (11) was thus obtained.R. Robinson and H. S u g i n ~ m e , ~ ~after preliminary syntheses of indolenine derivatives, have beenable t o prepare dl-noreserethole (111) in an instructive manner.Me-l--FH2 EtO --~MeCH2*CH,*N(CO)C,H, co\/C*CO,E tN (IV.)(111.1 EtOO\/?/cB, NMeNH 0Ethyl 8-phthalimido-a-acetyl-8-methylvalerate was coupled withp-etlioxydiazonium chloride in alkaline solution to yield ethyl8 -phthalimido - a - keto - 8-methylvalerate-p-ethoxyphenylhydrazone,C,H4: (CO),:N-CH,*CH,.CHMe*C (C0,Et) :N,H*C ,H,*OE t , with loss ofan acetyl group. When this hydrazone was submitted to the actionof ethyl-alcoholic hydrogen chloride it gave ethyl 5-ethoxy-3-methyl-3- p -pht halimidoet hylindolenine-2 -carboxylat e (IV) . Saponiikationof this by ethyl-alcoholic potassium hydroxide gave a dicarboxylicacid (V) which on decarboxylation in boiling xylene gave theEt O~~~~;Y*CH~-~H yo (V.)N C,H,*CO,Hindolenine (VI).The methosulphate (VII) of (VI) was deprived ofthe phthalic acid by short boiling with hydrazine hydrate in alcoholicsolution and on acidification cyclisation took place by simpleaddition of the amino-group in the side chain to the unsaturatedindoleninium system with production of dl-noreserethole (VIII) .(VII.) (VIII.)As two asymmetric centres are produced, a mixture of two racematesmight be expected; but only one has been observed.8 3 Proc. Imp. Acad. Tokyo, 1932, 8, 171; A., 952.89 J., 1932, 304; A., 287Irma. 199Yet another method for closing the third ring has been workedout by F.E. King and R. Robinson.90 In a preliminary researchH. S. Boyd-Barrett and R. Robinson 91 were able t o prepare deseth-oxydehydroeseretholemethine (IX) by synthesising the indole (X)from y-phenoxypropyl methyl ketone phenylhydrazone. This**CH,*CH,*OPh(IX-) b e NR (X.)~ ~ e e * C H 2 * C H 2 * N M e 2indole was methylated under pressure with methyl iodide to yield1 : 2 : 3-trimethyl-3-p-phenoxyethylindoleninium iodide (XI), thebase corresponding to which on oxidation with permanganate gaveO--xgFCH2*OPh o G g e * C H 2 * C H ,* OP hNMe (XI-) NMe (XII.)an indolinone (XII). By the action of fuming hydrobromic acidthis indolinone was converted into a reactive bromoethyl derivative(XIII ; R = H), the bromine of which could be replaced by methyl-amino- or dimethylamino-groups with production in the latter caseof desethoxydehydroeseretholemethine (XIV ; R = H).QG~~*CH,*CH,B~ + R(&+wCH,*CH2*NMe2By extending the synthesis to the corresponding methoxy-derivative (XIV; R = OMe), King and Robinson were able tosynt hesise and eventually resolve dehydroesermetholemethine,(XIV ; R = OMe) into its optical isomerides, as the quaternary salt.One of these proved to be identical with the natural product, andincidentally the interpretation given to the various stages in thesyntheses received confirmation. Cyclisation of ethylaminoin-dolinones was fmally effected in a simple manner. The methoxy-NMe (XIII.) NMe (XIV.)MeO@$lMe-CH,*CH2*NH2 MeO(J-?-P;g2 2NMe N vo "0(XV.1 (XVI.) Meoo- \/CH\/CH2 (XVII .p5e--p32NMe NHbromide (XIII ; R=OMe) was converted through its reactionproduct with phthalimide into the corresponding ethylamine (XV)90 J ., 1932, 1433; A., 759. O1 Ibid., p. 317; A., 287200 ORGANIC CHEMISTRY .-PART III.which was dehydrated by phosphoric acid in boiling xylene to forman amidine (XVI), and this on catalytic reduction gave noreser-methole (XVII). The latter was characterised as a crystallinequaternary salt, dl-esermet hole methopicrate, which showed a greatsimilarity to the natural salt.There is little doubt that physostigmine is derived in the plantfrom hydroxytryptophan. The possibility, however, of indolenuclei arising by oxidation of amino-acids is shown by the workof H. S. Raper 92 on the conversion of tyrosine and of 3 : 4-dihydroxy-phenylethylmethylamine into indole derivatives under the influenceof tyrosinase.A striking example of the somewhat analogousoxidation of a tertiary base into a quaternary salt by indole ringformation has been discovered by R. Robinson and S. S ~ g a s a w a . ~ ~Laudanosoline (XVIII) is oxidised by chloranil in alcoholic solutionand in the presence of potassium acetate to a dehydrolaudanosolineMe0OH --+OH(XVIII.)hydrochloride, a quaternary dihydroindole (XIX) which containsthe same carbon skeleton as laudanosoline, for on exhaustivemethylation and Emde degradation it gives the same product(XX) as laudanosine itself. Almost the same ground was coveredby C. Schopf and K. Thierfelder,94 who were able to effect thesame dehydrogenation by tetrabromo-o-quinone, by oxygen inpresence of platinum, and by potassium ferricyanide in presenceof a phosphate buffer at pH 6-9-7.1.Strychnos Alkaloids.-During the past year great advances havebeen made in the elucidation of the structural formuh of brucineand strychnine.The advances are such that, although furtherwork is necessary to confirm the internal structure of the molecule,the main features of the outside skeleton may be regarded as settled.The two formulae for brucine (R = OMe) and strychnine (R = H)which find most favour are the following :92 Biochem. J., 1927, 21, 89; A . , 1927, 278. W. L. Dulihre and H. S.Compare also H. Burton, J., Raper, &id., 1930, 24, 239; A., 1930, 814.1932, 546; A., 402.Q3 J., 1932, 789; A., 527.g4 Annalen, 1932,497, 22 ; A., 2046KING.201(I) is that put forward by H. Leuchs 95 and it contains most ofthe features characteristic of the formula advanced by K. N. Menonand R. Robinsong6 except that the bridge from the basic N-atomin the Menon-Robinson formula is attached to the 8-carbon atomof the dihydroindole structure. B. K. Blount and R. Robinsong7prefer the structure (11) in which the bridge C,H, may be inter-preted as GHMe, in which case strychnine would contain theskeleton of tryptophan and also that of harmine. The argumentsin favour of the positions assigned to the bridge are not conclusiveand its final position must await further experimental work.The measure of agreement expressed by these two formula ismainly a result of the interpretationof the experimental observationson the oxidation of these alkaloids with nitric acid, chromic acid,and permanganate.The important product of the oxidation of strychnine with nitricacid, named dinitrostrycholcarboxylic acid by Tafel, has beenshown 96 to be 5 : ,7-dinitroindole-2 : 3-dicarboxylic acid (111).Thisfinds expression in the suggested formulae and is in agreement withQHz-QHz/ \/\CH NQ" Q0\/\ Q0ZHCH CH-0(111.)60 CH (IV.1NH CH, O-CH,the results of E. Spath and H. Brets~hneider,~~ who obtainedN-oxalylanthranilic acid (IV) as a product of the oxidation ofstrychnine with permanganate in alkaline solution, and N-oxalyl-4 : 5-dimethoxyanthranilic acid from brucine. The permanganateoxidation of brucine in acetone solution was the subject of Leuchs'Ber., 1930, 63, 2997; A., 1931, 242.S 5 Ber., 1932, 65, 1230; A., 953.9 7 Ibid., p.2305; A., 1147.O 6 J., 1932, 780; A., 527.6 202 ORGANIC CHEMISTRY .-PART 111.earliest contribution to the chemistry of these alkaloids, but theinterpretation of the results was first correctly advanced by R. C.Fawcett, (the late) W. H. Perkin (jun.), and R. Robinsong9 andhas now been accepted by H. Leuchs and F. Kr0hnke.l Leuchsfound that when brucine, C2,H2,0,N2, was oxidised by permangan-ate, a keto-acid, brucinonic acid (V), was formed which could bereduced by sodium amalgam to brucinolic acid (VI), and this bythe action of alkali gave brucinolone (VII) and glycollic acid.Using formula (I) for simplicity of representation, the changes areas follows :( p 3 2 - p 2CH N/ /\ / \ / \(v’) -7H \;I€ 70 -YH cH co(VI-1 )(\CH/ (VII.) 4 CH CH----CH*OHCO CH 60 CH\/CH2\/\CH, O-CH,The characteristic feature of brucinonic acid (V) is its a-ketonicacid amide group and as such it should be oxidisable by hydrogenperoxide.H. Leuchs and F. Krohnke now k d that such is thecase, for an amino-acid, C,0H2206N2, possibly (VIII), is formed.p 2 - p 2CH NHA further striking degradation has now been re~orded.~ Whenbrucinonic acid (V) or brucinolic acid (VI) is oxidised with chromicacid in sulphuric acid solution, a red crystalline o-quinone (IX) isformed, but the resinous by-products on further oxidation with thesame reagent give a very small yield of an amino-acid, Cl,H,,05N2.The suggestion is made that this is formed by the destruction ofthe truly aromatic portion of the molecule, as happens in thepreparation of Hanssen’s acid, C,,H,,O,N,, from brucine, C,,H,,O,N,,together with loss of glycollic acid and oxidation of the a-ketonic99 J., 1928, 3087; A., 1929, 82.2 Ibid., p.980; A., 866.Ber., 1932, 65, 218; A., 407.Ibid., p. 1230; A., 953KING. 203acid amide group of brucinonic acid as already described.basis the natural result of the degradation is as shown below.On this7H2-(iH2CH NH/ \/\ / \/ R~\-CH YH 70 CO,H.FH YH dH/ \/\ /CH2/cH*co2H N H R N \CH' -> TH QH 60 CH CO CH Or 60 CHCHR ' d v C H CH-CO\/ \/\CH, O-CH,(V.1 (X.) Wa.1As is pointed out by Leuchs, a substance of formula (X) onoxidation with permanganate should yield 5-oxalylaminohexa-hydroindoline-4 : 6 : 7-tricarboxylic acid, which would give oxalicacid on hydrolysis.( X a ) on the other hand should give a @-ketonicacid, namely, 5-malonylamino-6- ketohexahydroindoline-4 : 7 -dicarb-oxylic acid, which should lose carbon dioxide and malonic acid onhydrolysis. On the basis of the Blount-Robinson formula, (X)and ( X a ) should be replaced by (XI) and (XIa).CH/ \2VH (7H\ CO,H*Cf€ ,CH60 CH YH (XIa.)CO CHWhen strychnine is reduced electrolytically to tetrahydro-strychnine, the amide group >N*CO- becomes >NH,CH,*OH-and it was shown4 that on oxidation of this with chromic acid anamino-acid, C21H2204N2, was formed in 14% yield.When hexa-hydrostrychnine was similarly oxidi~ed,~ an amino-acid, C21H2204N2;resulted which was also obtained by catalytic reduction of Leuchsacid. On this view Leuchs' acid should be C,,H,04N2. Thechanges involved, however, appeared to be complex, since theacid C21H,,0,N, of Briggs and Robinson formed a benzylidenederivative, and therefore presumably contained the reconstituted>N*CO*CH,- group of strychnine, but did not give the usualH. Leuchs and W. Wegener, Bw., 1930, 63, 2220; A., 1930, 1455; H.Leuchs, ibid., p. 3187; A,, 1931, 242.5 L. H. Briggs and R. Robinson, J., 1931, 3160; A., 178204 ORGANIC CHEMISTRY .-PART 111.strychnine colour reactions. It has now been demonstrated thatthese acids, almost certainly, contain an extra carbon atom, presentas a carboxyl group in the p-position to the indole N-atom andthat it is derived from a double molecule analogous to the redamorphous dyes obtained by H.Wieland and collaborators fromstrychnine and characterised as meriquinonoid diphenyl derivatives.This gives a satisfactory explanation of the origin of the new carb-oxyl group and of the failure t o give the normal colour reactions,since these are dependent on a free p-position.The new formulae for brucine and strychnine also give a satis-factory explanation of the neo-bases. When strychnidine metho-salts (amide group reduced to -NH*CH,-) are digested with methyl-alcoholic potassium hydroxide, they yield a methoxymethyl-dihydrostrychnidine which can be interpreted by the schemeSC*#Me S0,Me + 3C*OMe, YMe.When the quaternary salt is reconstituted by boiling dilute acid,an isomeric methylneostrychnidinium salt is formed which onheating loses methyl chloride, for instance, and gives neostrychnid-ine. Both strychnidine and neostrychnidine give the same dihydro-strychnidine on catalytic reduction, so the difference in the basesmust reside in the position of the double bond.s Strychnidine(XII) and neostrychnidine (XIII) are therefore assigned the struc-tures shown, the double bond occupying adjacent positions.TheI 1oxidation of neostrychnidine to the diketone strychnidone by per-manganate,g on this view of the constitution, would consist in thedisruption of the double bond by addition of two oxygen atomswith formation of a 10-membered ring.Subsidiary Stryc7mos Alkaloids.-The study of the constitution' H.Leuchs and H. Beyer, Ber., 1932, 65, 201; A., 407.* 0. Achmatowicz, (the late) W. H. Perkin (jun.), and R. Robinson, J . ,' G. R. Clemo, (the late) W. H. Perkin (jun.), and R. Robinson, J., 1927,Annakn, 1931, 491, 107; A., 179.1932, 486; A., 406.1589; A., 1927, 888KIN#. 205of the subsidiary alkaloids which accompany the chief alkaloid ina plant is of considerable importance, for it throws light on thegeneral scheme of phytochemical synthesis adopted by nature ina given species. Within the last few years four new alkaloids havebeen isolated from the residues accumulated by manufacturers inthe isolation of strychnine and brucine.Whether these newalkaloids are specific to Xtrycltnos Nux-vomica or Str. Ignatii orboth is unrecorded.In 1931 K. Warnat lo described the isolation of three newstrychnos alkaloids which he named a- and p-colubrines and+-strychnine. The first two dif€er in composition from strychnineby a methoxy-group and on oxidation by Spath and Bretschneider’smethod give two isomeric monomethoxyoxalylanthranilic acids.These were characterised as their dimethyl esters, that from a-colu-brine being identical with synthetic dimethyl N-oxalyl-4-methoxy-anthranilate and that from p-colubrine being identical with theisomeric 5-methoxy-derivative. This suggests that strychnine,brucine, and the two colubrines stand in the following relationship :There is no proof of the identity of the remainder of the moleculeswith that contained in strychnine or brucine, but the similarity ofproperties of all four bases suggests identity.The third alkaloid, $-strychnine, has been examined by B.K.Blount and R. Robinson 11 and the original preliminary observationsof Warnat confirmed. Its composition is that of a hydroxy-strychnine and on reduction in acid solution it yields strychnine.It is relatively stable to ferricyanide, forms an N-nitroso-derivative,and on crystallisation from methyl or ethyl alcohol yields methylor ethyl derivatives which are readily hydrolysed in cold acidsolution with formation of +strychnine. These properties andothers suggest that +-strychnine carries a tertiary alcohol group ona carbon atom adjacent to the basic N-atom in strychnine.Onthis view the nitrosoamine would be >CO,NO*N< and the ethers>C(OR)-N<.The fourth subsidiary alkaloid isolated from manufacturers’residual liquors is vomicine, l2 C,,H,,O,N,, differing from strychnine(C,,H,,O,N,) by CH202. It contains an aromatic ring, since it iseasily brominated, and has one nitrogen in combination as a lactam,10 Helv. Ghim. Actu, 1931,14, 997; A., 1931, 1312.11 J., 1932, 2305; A., 1147.12 €1. Wieland and G. Oertel, Annulen, 1929, 469, 193; A., 1929, 708206 ORGANIC CHEMISTRY .-PART 111.since boiling alcoholic potassium hydroxide gives vomicinic acid,C,2H2G03NZ, a substance which is very readily autoxidisable and isconverted into the original base by the action of acids.On catalyticreduction vomicine adds on two hydrogen atoms, indicating thepresence of a double bond. It readily forms a benzoyl derivativeand on reduction with hydriodic acid gives deoxyvomicine,C,2H,,0,N,, these properties suggesting the presence of a tertiaryalcoholic group. Unlike strychnine, vomicine does not add onmethyl iodide at the basic N-atom, but vomicinic acid on methyl-ation yields N-methylvomicinic acid and its methyl ester, togetherwith two parallel products each containing a CH, group more,vix., an acid C,,H,,O,N, and an ester C,,H,,0,N,.13 The acidC,,H,,O,N, can be converted by alcoholic potassium hydroxidewith difficulty into N-methylvomicinic acid. This suggests thepresence of a phenolic group, which would also account for theready autoxidation of vomicinic acid.This finds expression inthe partial formula (I) for vomicine and (11) for vomicinic acid.Vomicine, like brucine, is seiisitive to chromic acid and givesan important series of degradation product8s.14 The chief acidformed has the composition C,,H,,O,N, and very readily losescarbon dioxide. The carboxyl group is therefore probably adjacentt o a hydroxyl group. By analogy with the results obtained by theaction of chromic acid on brucine and strychnine the degradationof vomicine may be represented thus :OH OHI CH,-The base C,,H,,O,N, obtained on decarboxylation of this acidIt can also contains a hydroxyl group which can be benzoylated.13 H.Wieland and F. Calvet, AnnaZen, 1931, 491, 117; A., 179.1 4 H. Wieland and G. Oertel, ibid., 1929, 469, 193; A., 1929, 708; H.Wieland, F. Holscher, and F. Cortese, ibid., 1931, 491, 133; A., 179KING. 207be reduced catalytically to C16H2602N2, probably through reductionof a double bond and of the ether linkage :sC-O-C< + 5CH CHE + H20.The two extra oxygen atoms in vomicine which are not foundin strychnine are thus accounted for as a tertiary alcohol groupand a potential phenolic group. If the remainder of the ringsystems of strychnine and vomicine is the same, and there is no evi-dence so far inconsistent with this view, the only additional differencewill be in the possession by vomicine of an extra methyl group.Yohimbine.-A group of alkaloids of which yohimbine is thechief representative has been obtained from Corynanthe Johimbe,of the natural order Rubiacece to which the cinchona plants belong.Six alkaloids, yohimbine, yohimbene,15 aZZoyohimbine,16 iso-yohimbine,16 a-l' and yl8-yohimbines, have been definitely identifiedand they all appear to have the formula C2,H2,0,N2.In addition,a seventh isomeric alkaloid, corynanthine, has been obtainedby E. Fourneau and Piore l9 from Pseudocinchom africuna, of thesame natural order as the above. All these alkaloids are mono-methyl esters which on hydrolysis yield monocarboxylic acids,and four of them, yohimbine, yohimbene, y-yohimbine, and iso-yohimbine, give isomeric acids which on decarboxylation yieldone and the same alcohol, yohimbol.20 aZZoYohimbine gives anisomeric alcohol, alloyohimbol.Corynanthine and a-yohimbinegive neither yohimbol nor alloyohimbol. The four alkaloids whichgive yohimbol, when treated with selenium and soda-lime in a vacuumsublimation apparatus, give one and the same substance, to whichthe formula ClgH12N,*O~C,gHlzN2 is assigned by G. Hahn and W.Schuch,2O but which according to F. Mendlik and J. P. Wibaut 21is a base, C,,H1,N,, t o which the name yobyrine is given. Thedifference between the four alkaloids named must accordingly bedue t o a different situation for the carbomethoxy-group in themolecule.A number of degradation products have been obtained fromyohimbine, some of which have been identified. G. Barger andl5 G. Hahn and W. Brandenberg, Ber., 1926, 59, 2189; 1927, 60, 707;A., 1926, 1263; 1927, 471.l6 Idem, Ber., 1927, 60, 669; A., 1927, 471; K.Warnat, ibid., 1926, 59,2388; 1927, 60, 1118; A., 1926, 1263; 1927, 681.17 R. Lillig and H. Kreitmair, Merck's Jahresber., 1928, 42, 20; B., 1930,485; G. Hahn, and W. Schuch, Ber., 1930, 63, 1638; A., 1930, 1194.1s G. Hahn and W. Schuch, Eoc. cit.19 Bull. SOC. chim., 1911, [iv], 9, 1037; A , 1912, i, 49.20 Ber., 1930,63, 1638, 2961; A., 1930, 1194; 1931, 243.21 Rec. trav. chim., 1931, 50, 91; A,, 1931, 369208 ORGANIC CHEMISTRY .-PART III.(Miss) E. Field 22 showed that yohimboaic acid when distilled withlime gives a dimethylindole with an odour of scatole, and whichgives a, crystalline picrate. The same dimethylindole, m.p. 55", wasobtained by E. Winterstein and M. Walterz3 by distillation of theacid and by K. Warnat24 by heating the acid with soda-lime orzinc dust. This indole is not identical with any known dimethyl-indole, and synthetic experiments by F. Mendlik and J. P. Wibaut 25show that it is not identical with 3 : 5-, 3 : 6-, or 3 : 7-dimethyl-indole. The colour reaction of yohimbine with sulphuric acid andpotassium dichromate is similar to that given by strychnine andsuggests a relationship. In fact E. Spath and 1%. Bretschneider 26find that both alkaloids on oxidation with alkaline permanganategive N-oxalylanthranilic acid (I). In the case of strychnine thisacid is known to arise from an indole nucleus, but in the case ofyohimbine a quinoline structure is not excluded.(1.1 (11.) (111.)When yohimbine is boiled with acetic anhydride and sodiumacetate, it gives a crystalline ON-diacetylyohimbine and an amor-phous monoacetyl derivative.27 The former on oxidation withdilute nitric acid gives succinic acid and a 6-nitroindazole-3-car-boxylic acid (11), the constitution of which follows from its de-carboxylation to 6-nitroindazole identical with the synthetic pro-duct.28 The indazole structure is probably not present in the originalmolecule, but arises by the action of nitrous acid on an o-amino-phenylacetic acid group as is shown in (111).The formation of theindazole is exactly analogous to the formation of 4-nitro-5-(3-pyridyl)-pyrazole from nicotine by the action of nitric acid.29Two other degradation products of yohimbine which have beendefinitely identified are o-oxycarbanil, Ph<-o>CO, which is NH22 J ., 1915, 107, 1025.23 Helv. Chim. Acta, 1927, 10, 5 7 7 ; A., 1927, 1205.24 Ber., 1927, 60, 1118; A., 1927, GSl.25 Rec. Irav. chim., 1931, 50, 91; A., 1931, 369.2o Ber., 1930, 63, 2997; A., 1931, 242.27 A. Schomer, Arch. PharnL., 1927, 265, 500; A., 1927, 1097.28 G. Hahn and F. Just, Ber., 1932, 65, 717; A., 760.G. A. C. Gough and H. King, J., 1931, 2968; A . , 68; Ann, Reports,Compare also I<. Warnat, Ber.,1926, 59, 2388; A., 1926, 1263.1931, 28, 166KING. 209obtained by the action of permanganate on yohimboaic acid indilute alkali at room temperature,30 and isoquinoline, obtained insmall yield by distillation of yohimboaic acid with zincWhen yohimboaic acid is distilled or fused with potassiumhydroxide or heated with zinc dust 23 or with lime,22 it yields abase, C13Hp$2, which, like the parent alkaloid, must contain atertiary N-atom, since it forms a methiodide and according toWinterstein and Walter is accompanied by a second base, C12Hl&..The disposition of the ring systems in these bases is unknown,but it is probably the same as occurs in the products of the actionof selenium on yohimbine.When this alkaloid is heated withselenium,25 it yields three substances : yobyrine, C,,H,,N, ;dihydroyobyrine, C1,H2,N2 ; and ketoyobyrine, C ~ O H ~ ~ O N ~ .Yobyrine differs from yohimbine by a carbomethoxy-group, amolecule of water and four hydrogen atoms, and ketoyobyrine mustbe formed by condensation of the carbomethoxy-group with elimin-ation of methyl alcohol and formation of a bridge.On fusion withpotassium hydroxide ketoyobyrine yields a basic substance,C,,H,,O,N,, and an acid identified as 2 : 3-dimethylbenzoic acid.It has been suggested25 that this is the benzene nucleus whichappears as dimethylindole by other methods of degradation, inwhich case o-oxycarbanil, oxalylanthranilic acid, and indazole-carboxylic acid must arise from a second benzene nucleus with anitrogen atom adjacent to it. Furthermore the base C13H12N2(IV), differing from yobyrine, C1,H,,N2 (V), by C,H6, is consideredby the same authors t o have lost four carbon atoms of the benzenenucleus with the two o-methyl groups attached, as is shown below :MeThis benzene nucleus is supposed t o be present in yohimbine as ahydroaromatic structure which is dehydrogenated by selenium butdegraded by other reagents.Pyrrole Pigments .31The remarkable progress which has been made in the last 30years in the elucidation of the structure of blood and leaf pigmentsis mainly due t o the efforts of Nencki, Kuster, Piloty, Willstiitter,and Hans Fischer and their pupils.The brilliant researches of the30 K. Warnat, Rer., 1926, 59, 2388; A., 1926, 1263.31 Valuable summaries by H. Fischer can be found in ibid., 1927, 60, 2611 ;Naturwias., 1930, 18, 1026; Nobel Vortrag, Dec. Ilth, 1930210 ORUANIC CHEMISTRY.-PART III.last-named were fittingly recognised by the award of the Nobelprize in 1931.The following account has been written to enable the reader t oappreciate the position which has been reached in the chemistryof the blood and bile pigments and also as an introduction to thechlorophyll problem.Hamin, obtained from blood by heating with acetic acid andsodium chloride (Teichmanii's test), has the constitution (I), as isshown by its properties and by synthesis.It consists of fourpyrrole-like nuclei joined in the cc-position by methine groups and apeculiar conjugated system of double bonds responsible for thecolour. The iron is held in complex combination, since it does notshow the ordinary ionic reactions of iron. Inspection of the formulashows how it can give rise to four different hzmopyrrole bases (11-V) by drastic reductive fission with hydriodic acid and also t o thefour corresponding hzmopyrrolecarboxylic acids (3'1-IX).Me-EtMe1] IIH \/(11.) NH\/(VI.) NHHaemopyrrolc.Me-XMe" IIHHzmopyrrole-carboxylic acid.Me - EtHI1 (/Me \/(111.1 NHMe-XCryp topyrrolc.HI1 IIMe v (VII.) NHCryptopyrrole-carboxylic acid.Me--Et Me-EtMe1' IlMe HI1 IIH\/ (v.) NH\/(IV.) NHPhyllopyrrole.Opsopyrrole.Me-X Me - XMe1/ '/Me HI1 I/H\/(IX.) NH\/(VIII.) NHPhyllop yrrole - Opsop yrrole-carboxylic acid. carboxylic acid.On oxidation, the nuclei bearing vinyl groups are completelydegraded but the acidic nuclei appear as hzmatic acid (X). If thevinyl groups are converted into ethyl groups by reduction of hzminwith hydriodic acid or catalytically to mesoporphyrin, the hematicacid obtained by oxidation is accompanied by methylethylmalein-imide (XI).(x.) Me---X o\)oThese reductive and oxidative processes can nowadays be sKING.21 1oontrolled that they give a clue t o the number of nuclei of agiven type within the molecule of a pigment. When the iron isremoved from haemin by various reagents, a series of porphyrinsis obtained characterised by their fluorescence, dichroism, and thepossession of absorption spectra in the visible and the ultra-violetregion.32The more important porphyrins derived from haemin are shown inthe following table. In column 3 is shown the number of isomeridesobtainable by altering the sequence of groups around the pyrrolerings in the P-position with the condition that each pyrrole nucleusmust bear one methyl group, whilst in column 4 is recorded thenumber of isomerides synthesised to date by H.Fischer and hiscollaborators.Isomerides. Side Chains.Possible. Synthesised. ~ A \Haemin C3,H3,0,N4FeCl 15 P3 4Me 2X 2CH:CH,Protoporphyrin C3,H3,04N4 15 233 4Me 2X 2CH:CH,Haematoporphyrin C3,H3 ,O ,N4 15 233 4Me 2X 2CH(OH)MeMesoporphyrin C34H3 @,N, 15 1234 4Me 2X 2EtXtioporphyrin C32H3 ,N4 4 4s5 4Me- 4EtDeuteroporphyrin C3,H3,04N, 15 336 4Me2X -Deuterohaemin C30H,,0,N4FeC1 15 233 4Me2X -A number of porphyrins have been found to occur naturally inyeast, pearl-oysters, mussels, feathers, egg-shells, and in cases ofhzematoporphyrinuria, and representatives of some of these havebeen synthesised.Coproporphyrin C3,H380 ,N4 4 437 4Me - 4 XConchoporphyrin C3,H3,0 4Me 3X 1 succinic acidOoporphyrin C34H3404N4 15 233 4Me 2Et BCH:CH,Uroporphyrin CIoH3 ,O ,N, 4Me - methylmalonic &Isornerides.Side chains.Possible. Synthesised. ~ - \succinic acid~~~32 A. Treibs, Z. physioE. Chem., 1932, 212, 33; W. Hausmann and 0.Krumpel, Biochem. Z., 1927, 186, 203; A., 1927, 893.33 H. Fischer and A. Kirstahler, Annalen, 1928, 466, 178; A., 1928, 1385;H . Fischer and L. Niissler, ibid., 1931, 491, 162; A., 173.34 H. Fischer and co-workers, ibid., 1927, 452, 289; A., 1927, 469; ibid.,459, 74; A., 1928, 76; ibid., 1928, 468, 166; A., 1928, 1384; ibid., 1929,475, 274; A., 1929, 1466; ibid., 1930, 480, 260; A., 1930, 932; ibid., 482,211; A., 1931, 101; ibid., 484, 85; A., 1931, 240.35 Idem, ibid., 1926, 448, 186, 201; A., 1927, 962, 963; ibid., 450, 190;A., 1926, 1261; ibid., 1927, 452, 285; A., 1927, 469; ibid., 459, 71; A.,1928, 76; ibid., 1928, 466, 211; A., 1928, 1382; ibid., 1932, 495, 26; A., 756.36 Idem, ibid., 1928, 466, 183; A., 1928, 1385; ibid., 1929, 473, 245; A.,1929, 1184; ibid., 1931, 491, 173; A., 173.37 Idem, ibid., 1926, 450, 214; A., 1926, 1261; ibid., 1927, 457, 97; A.,1927, 1088; ibid., 458, 132; A., 1927, 1206; ibid., 1928, 461, 276; A., 1928,776; ibid., 462, 249; A., 1928, 902; jbid., 466, 156; A., 1928, 1384; 2.physiol. Chem., 1929, 182, 265; A., 1929, 940; ibid., 1931, 196, 163, 236;A., 1931, 747, 853212 ORGANIC CHEMISTRY .-PART 111.Incidentally a large number of importanthave been prepared.Isomerides.Possible.Synthesised.Porphinmonopropionic acid 8 838Porphintripropionic acid 8 139Tetramethyltetrapropylporphyrin 4 440P yrroporphyrin 24 841PyrroEtioporphyrin 8 8 4 2Deuteroaetioporphyrin 15 9 4 3Rhodoporphyrin 24 .)44isouroporphyrin 4 345related porphyrinsSide chains.4Me 3Et 1X4M0 1Et 3X4Me 4Pr -4Me 2Et 1X4Me 3Et -4Me 2Et -4Me 2Et 1C02H 1X4Me 4CH(C02H),Ztioporphyrin, C32H38N4, which is obtained indirectly by a,pyrogenetic reaction from hematoporphyrin, is a fully substitutedporphyrin and is a convenient reference substance for all otherporphyrins. Four atioporphyrins are possible which, followingthe recognised procedure, are shown in abbreviated form as follows,each bracket representing a pyrrole nucleus, since the remainderMe Et Me E t Me Et Me E tI I -Me Me-l--l Et Me-] I-& Et 1 I-Et.I 1 L II 1 1 111: 1 [ Iv jlMrI -Me -Et Et Me Et- -Et Me I-I I-! Et Me Et Me Me Etof the molecules are identical in most porphyrins.All the porphyrinstabulated can be referred t o one of these four structures and it isknown, for instance, that natural hEmin has its substituents basedon atioporphyrin (111) and so has chlorophyll. If the ethyl groupsare replaced by propionic acid groups (X), there will be four possiblecoproporphyrins, all of which have been synfhesised and two ofthem, coproporphyrins (I) and (111), are identical with naturallyoccurring coproporphyrins.If, however, only two of the four38 H. Fischer and co-workers, Annalen, 1928, 461, 237; A., 1930, 651;ibid., 1929, 471, 293; A., 1929, 941; ibid., 475, 254; A., 1929, 1465; ibid.,1931, 492, 27, 50; A., 173.Idem, ibid., 1930, 479, 32; A., 1930, 621.40 Idem, ibid., 1931, 486, 20; A., 1931, 746.4 1 Idem, ibid., 1929, 473, 229; A., 1929, 1184; ibid., 1930, 480, 136; A.,1930, 931; ibid., 482, 199; A., 1931, 101.42 Idem, ibid., 1928, 466, 211; A., 1928, 1382; ibid., 1929, 473, 243; A.,1929, 1184; ibid., 1930, 480, 126; A., 1930, 931; ibid., 482, 199; A., 1931,101.43 Idem, ibid., 1928, 466, 217; A., 1928, 1382; ibid., 1930, 482, 209; A.,1931, 101; 2. phy5iOZ. Chem., 1931, 198, 56; A., 1931, 967.44 Idem, Annalen, 1929, 473, 237; A., 1929, 1184; ibid., 1930, 480, 109;A., 1930, 931.45 Idem, ibid., 1927, 457, 91; A., i927, 1088; ibid., 1930, 483, 1 ; A., 1930.1599; 2.physiol. Chem., 1932, 204, 68; A., 285KING. 213ethyl groups of aetioporphyrin are replaced by propionic acidgroups as in mesoporphyrin, then fifteen mesoporphyrins arepossible, 1 and 2 derived from Ztioporphyrin (I), 3-5 fromatioporphyrin (11), 6-11 from aetioporphyrin (111), and 12-15 fromaetioporphyrin (IV), and of these twelve have been synthesised.Me X Me Et Me XI 1Me XX MeI I MeEt MeMe X Me X Me X Me Et -- n -Et Me-Me[ ' AM, E t LEt u X MeMe Et Me Et Me X Me EtMe Me Me Me 1 Me X 179 ] [ 10 ] [ 11 ] [ 12 1; Et Me 1-1Me X I x xEt Me I X Et XP M e Et MeMe X Me X Me X n X X '1 13 ] [ 14e I I Me 'At MeMe Me EtTheir dimethyl esters have characteristic melting points and areespecially valuable for characterisation, since the decarboxylationproducts, the aetioporphyrins, have no defhite or very high meltingpoints. There are as many hamins possible as there are meso-porphyrins, since the ethyl groups are replaced by vinyl groupsand introduction of iron does not so far as is known increase thenumber of possible isomerides.Natural hamin corresponds tomesoporphyrin 9, as will be seen by comparison with the formulafor hamin given a t the beginning of the section. The first proofthat hamin had this particular orientation of groups in the p-positionswas afforded by the synthesis of mesoporphyrin 9, identical withthe product obtained from natural h a m i r ~ .~ ~Space does not allow an account of the variety of syntheticmethods used by H. Fischer and his collaborators in the preparation46 H. Fischer and G. Stangler, Annalen, 1927, 450, 53; A., 1928, 76214 ORGANIC CHEMISTRY .-PART 111.of the many porphyrins so far synthesised. Reference can only bemade t o the representative synthesis of natural hemin (haemin IX)and hsmin I11 corresponding to mesoporphyrins 9 and 3 respec-tively.2 : 3-Dimethylpyrrole and 2 : 4-dimethylpyrrole-&aldehyde werecondensed by alcoholic hydrobromic acid t o 4 : 5 : 3’ : 5’-tetra-methylpyrromethene hydrobromide (A), whilst cryptopyrrole-carboxylic acid (VII, p. 210) on bromination gave 5 : 5’-dibromo-3 : 3’-di-p-carboxyethyl-4 : 4’-dimethylpyrro-2 : 2’-methene hydro-bromide (B) 47 with loss of a methylene group.When (A) and (B)were heated in succinic acid a t 180-190”, deuteroporphyrin (C;R = H) was obtained.48Me-H Me=====€€ R CH Me(A*) Melt II-cH-1 1 Me/\/ \/\R\/ NH - \ 2 I e NHBr yJyH Ni=/ --+ c H Y H N , d M e \CH (C.1NH NHBrB r A CH=/NBr Me\ \\ A//Me![ II? XI-lMe X CH XThe latter was converted into deuterohaemin by the action offerrous acetate, acetic acid, sodium chloride, and ‘hydrochloricacid, since the pyrrole nuclei in the iron complexes are more reactivethan in the iron-free porphyrins. On treatment of deuterohaeminwith acetic anhydride in presence of stannic chloride, diacetyl-deuterohzmin was obtained and it was characterised by conversioninto diacetyldeuteroporphyrin (C; R = Ac).On reduction ofthis porphyrin with alcoholic potassium hydroxide the &-secondaryalcohol, hzmatoporphyrin [C ; R = CH(OH)Me], was obtainedin 24% yield, calculated on the deuterohzmin used. This por-phyrin proved to be identical with the natural product obtainedfrom haemin. It was quantitatively converted by heating ina high vacuum at 105” into protoporphyrin (C; R = CHZCH,),which with ferrous iron gave hemin (p. 210) identical with thatobtained from haemoglobin.It is a remarkable fact that the structure of hzmin as now deter-mined by synthesis was suggested as early as 1912 by W. ICii~ter,~~with the same orientation of groups in the P-positions except that amethyl and a vinyl group on one nucleus had to be interchanged.Moreover, it was not until hemin was synthesised that its constitu-tion was known with certainty, since immediately preceding its47 H.Fischer and H. Andersag, Annalen, 1927, 458, 135; A., 1927, 1206.48 H. Fischer and A. Kirstahler, ibid., 1928, 466, 178; A., 1928, 1385.49 2. physiol. Chem., 1913, 82,463; A., 1913, i, 210m a . 216synthesis it was still believed to contain one acetylene and one vinylgroup.50More recently H. Fischer and L. Niissler 51 have synthesisedhaemin I11 with an orientation of groups identical with that suggestedby Kuster for natural hzemin. The corresponding deuteropor-phyrin I11 (D) was obtained by condensation of 5 : 5’-dibromo-4 : 4’-dimethyl-3 : 3’-di-p-carboxyethylpyrro-2 : 2’-methene hydro-bromide (E) and 4 : 5 : 4’ : 5’-tetramethylpyrro-2 : 2’-methene hydro-The subsequent steps were then on the same lines as for thesynthesis of natural hamin IX.All authors are agreed that in haemin the iron is in the tervalentstate and on its reduction to the bivalent state a haemochromogenis obtained accompanied by a change in the spectrum and a decreasein stability.Hitherto haemochromogen was always stabilised byformation of complexes with nitrogenous bases such as nicotine,pyridine, and hydrazine, but it has now proved possible to isolate thehsms or complex salts with bivalent iron but without nitrogenousaddenda.52 Thus, if suitable porphyrins are treated in nitrogenwith a ferrous salt, crystalline haems are obtained. This has beenfound possible for proto-, aetio- and meso-porphyrins and theiresters and the products on addition of pyridine give intense hzemo-chromogen spectra. Hzemoglobin is a molecular compound ofhsm and globin, but it is exceptional in that it has no hzemo-chromogen spectrum. I n this connexion it is of interest thathamatin (hzem-oxide) can form additive products with glyoxaline,methylglyoxaline, and pilocarpine.53 The additive product withtwo molecular proportions of glyoxaline separates unchanged frompyridine solution and does not part with its glyoxaline even a t 100”.Glyoxaline has thus the greatest affinity for hzematin and sincehsmatin and globin combine t o give methzemoglobin,M the sugges-H. Fischer and G. Stangler, Annalen, 1927, 459, 53 ; A., 1928, 76.51 Ibid., 1931, 491, 162; A . , 173.52 H. Fischer, A. Treibs, and K. Zeile, 2. physiol. Chem., 1931, 195, 1;53 W. Langenbeck, Ber., 1932, 65, 842; A , , 757.A., 1931, 633.R. Hill and H. F. Holden, Biochem. J., 1926, 20, 1326; A., 1927, 67216 ORGANIC CHEMISTRY .-PART 111.tion is made 53 that the combination is through the glyoxalinenuclei of histidine, the characteristic amino-acid of globin.Bile Pigments.-Bilirubin, C,,H,,O,N,, is a bile pigment which issupposed to arise from the break-down of haemoglobin in cells ofthe reticulo-endothelial system, especially those of the liver. Itis a rare substance difficult to obtain, the best source being oxgall-stones. It differs from hzmin by having one carbon less, twooxygen atoms more, and no iron. It has no characteristic absorptionspectrum, so the porphyrin ring-system is probably absent. It ischaracterised by striking colour reactions, of which Gmelin’s is themost important. This may be carried out by carefully adding nitricacid containing nitrite to a solution of the pigment in chloroform.The latter passes through the colour changes green, blue, violet,red, and finally yellow.55 On energetic oxidation with chromicacid bilirubin gives haematic acid56 (X, p. 210) and on energeticreduction with hydriodic acid, cry-ptopyrrole (111, p. 210) andcryptopyrrolecarboxylic acid 57 (VII, p. 210).The most important derivatives are those obtained by gentlereduction. Sodium amalgam or catalytic reduction yields meso-bilirubin, C3,H,,06N4 ; this on further reduction yields meso-bilirubinogen (urobilinogen), which gives an intense colour reactionwith Ehrlich’s p-dimethylaminobenzaldehyde reagent. Meso-bilirubin on oxidation gives haematic acid (X, p. 210) and methyl-ethylmaleinimide (XI, p. 210) just as mesoporphyrin does. Onreduction with hydriodic acid and acetic acid mesobilirubin yieldsa mixture of two acids which differ from each other in compositionby a methylene group : bilirubic acid, C1,H2,03N2, and neobilirubicacid, Cl,H2,0,N,, both of which on further reduction give rise tocryptopyrrole (111, p. 210). The acidic reduction products are, how-ever, different, for bilirubic acid gives cryptopyrrolecarboxylicacid (VII, p. 210) whilst neobilirubic acid gives haemopyrrolecarboxylicacid (VI, p. 210). These observations receive a ready explanationon the formulm (I) for mesobilirubin, (11) for bilirubic acid, and (111)for neobilirubic acid.55Me-Et Me-X X-Me Et-Me CH= I-IoH (I). HO~-I=CH--II ll-c~~-Il I(_ v \.YN v NH NH\/NMe-Et Me-X X-Me Et-MeHI1 II-CH -11 llOH \/ \/HOll Il-cH2-II IlMeNH (111.1 NH\/ \/NH (11.) NH55 H. Fischer and R. Hess, 2. phpiol. Chern., 1931,194, 201 ; A., 1931, 497.56 W. Kiister, ibid., 1898, 26, 314; A., 1899, 314.5 7 H. Fischer and E. Adler, ibid., 1931, 197, 237; A., 1931, 967KINU. 217These two acids (11) and (111) are the leuco-compounds corres-ponding to xanthobilirubic acid (IV) and neoxanthobilirubicacid (V), into which they are respectively converted by the actionof sodium methoxide 58 or potassium meth0xide.5~Conversely, neoxanthobilirubic acid (V) is converted by condens-ation with formaldehyde into mesobilirubin (I), which gives theGmelin reaction and is identical with the product obtained frombilirubin. This incidentally supports the symmetrical structureassigned to mesobilirubin, and mesobilirubinogen will be the corres-ponding trimethane derivative. The orientation of the groupsattached to the pyrrole nuclei follows from the synthesis of xantho-bilirubic and neoxant hobilirubic acids. 5-Aldehydo-3-methyl-4-e t hylpyrrole-2 - carboxylic acid (VI) and crypt op yrrolecarboxylicacid (VII) condense in presence of a large excess of hydrobromicacid to give 5-carboxy-4 : 3’ : 5’-trimethyl-3-ethylpyrromethene-4‘-propionic acid hydrobromide (VIII), which on bromination gives5- bromo -4 : 3’ : 5’ - trimet hyl-3-et hylpyrromet hene-4’-propionic acidhydrobromide 59 (IX). The latter on treatment with potassium orsilver acetate yields xanthobilirubic acid (X) identical with theproduct from natural sources.Me-Et Me-X Me-Et Me-XC02HII IIcHo HI/ - C02Hl-I=CH--/J v NH\/NHBr v NH v NH(VI.1 (VII.) (VIII.) 1 Me-Et Me-X Me-Et Me-X I1 llMe t- BJ---I=CH--II l l ~ , v \/ v NHBr (IX.) NHThree isomerides of xanthobilirubic acid are possible and all weresynthesised by suitable methods. When xanthobilirubic acid wasbrominated, it gave an unstable substance which on treatment withpyridine passed into mesobilirubin identical with that from naturalbilirubin.60 During this synthesis one carbon atom is lost (theethanes are unstable), for which there are many analogies in theH O ~ ) = CH-N (x.) NH58 H. Fischer and H. Rose, Ber., 1913, 46, 439; A., 1913, i, 382.59 H. Fischer and H. Berg, Annalen, 1930. 482. 189; A., 1931, 101.6o H. Fischer and E. Adler, 2. physioE. Chem., 1931, 200, 209; A., 1931,1420218 ORQANIC CHEMISTRY .--PART 111.syntheses of dipyrromethanes from methylbrominated pyrroles.61Since mesobilirubin on reduction with hydriodic acid gives neobilirubic acid or by fusion with resorcinol gives neoxanthobilirubicacid, the synthesis of xanthobilirubic acid is also a synthesis ofneoxanthobilirubic acid. In a similar manner bromination of theisomeric mesobilirubic acids gave isomeric mesobilirubins, all ofwhich are important, since there is some evidence that more thanone bilirubin occurs in nature.On the basis of the results recorded above, the most probableformula for bilirubin is (XI).Me---CH:CH, Me-X X-Me CH,:CH,===,Me HoI-cH-~I 11-\//OHCH,-lI 11- CH=N v v NH (XI.) NHSuch a structure has an orientation of groups in the P-positionscorresponding to a mesoporphyrin 13 (p. 213) or an aetioporphyrinIV (p. 212), which has no other natural representative. It followsfrom this that the conversion of haemoglobin into bilirubin mustinvolve intermediate degradation to pyrrole units. Against theabove structure for bilirubin is the observation that on oxidationwith nitrous acid a substance is formed which has a composition,C7H7O2NY agreeing with that of methylvinylmaleinimide, but oncatalytic reduction takes up two hydrogen atoms to give a substancewhich is not methylethylmaleinimide. Formulae have been suggestedfor these substances containing fused pyrrofuran rings producedby ring-formation between the vinyl group and an a-hydroxyl group,but this is difficult to reconcile with the relative orientation of thevinyl and hydroxyl groups necessitated by (XI). The final decisionwill no doubt depend on a synthesis of the structure (XI).The nomenclature of the porphyrin group is applicable to thebilirubin group. If the vinyl groups in (XI) are replaced by prop-ionic acid (X) groups, a coprobilirubin which may occur naturallyin certain pathological conditions is produced corresponding toaetioporphyrin IV and such a coprobilirubin (XII) has beensynthesised. 62Me-X Me-X X-Me X---Me\/--NHO(\~=CH-~(J~-CH,-,,- II I1 CH=I-lOH ,,/N NH (XII.) NH NIf the vinyl and propionic acid groups are replaced by ethylgroups, an aetiomesobiliru bin or decarboxylated mesobilirubin isproduced, of which one example has also been ~ynthesised.~~61 H. Fischer and P. Halbig, Annalen, 1926, 447, 123; A., 1926, 621.E2 H. Fischer and E. Adler, 2. physiol. Chem., 1932, 210, 139; A., 1045.63 H. Fischer and E. Adler, ibid., 1932, 206, 187; A., 627KmQ. 219When mesobilirubin (I) is oxidised with ferric chloride, a ferri-chloride of a new pigment, ferrobilin, is obtained,64 which by theaction of alkali gives glaucobilin (XIII). The particular orientationof pyrrole linkages ascribed to it follows from its synthesis by theaction of formic acid on neoxanthobilirubic acid (V).The vinyl analogue of this substance in which the ethyl groupsare replaced by vinyl groups has now been found in nature,65 but atrue pyrrole structure is preferred for a terminal nucleus ratherthan for one of the central nuclei as shown in (XIII). This sub-stance occurs as uteroverdin in the dog’s placenta, as a constituentof the pigments of birds eggs (sea-gulls eggs gave 250 mg. from5 kg. of shells), and as a by-product in the extraction of bilirubinfrom ox gall-stones. It has also been obtained artificially byoxidation of bilirubin with ferric chloride and is hence called de-hydrobilirubin. The constitution assigned is supported by the re-sults of reductive fission-formation of bilirubic acid-and oxid-ation-formation of methylethylmaleinimide after preliminaryreduction but not before.HAROLD KING.c4 H. Fischer, H. Baumgartner, and R. Hess, 2. physiol. Chem., 1932, 206,201 ; A., 627.65 R. Lemberg, Annalen, 1931, 488, 74; A., 1931, 1066; R. Lemberg andJ. Barcroft, Proc. Roy. Soc., 1932, [B], 110, 362; A., 627; R. Lemberg,Annalen, 1932, 499, 25