ORGANIC CHEMISTRY.A. General.1. THE STRUCTURE AND STEREOCHEMISTRY OF SIMPLEORGANIC MOLECULES.THE purpose of this article is to take two quantities, bond lengthand the heat changes in hydrogenation, and discuss the informationobtained from accurately determined values of them as to thestructure and stereochemistry of simple molecules. Both quantitiescan be measured with a known degree of accuracy and a number ofconclusions can be based directly on them without the need forsubsidiary hypotheses. It will be seen that in many points theyconfirm what is already known from the ordinary methods oforganic chemistry and stereochemistry and show that the simplepictures of molecules deduced by those methods are a closeapproximation to the truth, much closer than appeared to be thecase when physical methods were first applied to organic molecules.In other respects, however, they give valuable information whichcould never have been obtained by older methods; some of thisinformation has an important bearing on the correlation ofstructure and chemical reactivity, and some raises points whichindicate that certain current theories of organic chemistry are inneed of revision.Bond Lengths and Angles.The quantitative data on bond lengths available at the presenttime are fairly extensive, and those for the hydrocarbons and theirhalogen derivatives are discussed in the following paragraphs intheir relation to the structural problems of organic chemistry.Hydrocarbons.-The carbon-carbon single bond distance wasfirst measured in the diamond crystal, in which each atom isbonded tetrahedrally (ie., with bond angles of 109" 28') to fourothers at distances of 1.541 & 0-001a.l This same value withinthe experimental error of & 0.02 or 0.03 A.is observed in ethane,propane, isobutane, neopentane , cycZopropane, cydopentane, andcyclohexane (Table I) according to recent electron diffractionmeasurements on the vapours.2 X-Ray measurements on com-W. H. Bragg and W. L. Bragg, Proc. Roy. SOC., 1913, A, 89, 277; W.Ehrenberg, 2. Krist., 1926, 63, 320.L. Pauling and L. 0. Brockway, J. Amer. Chern. SOC., 1937, 59, 1223.A. Miiller, Proc. Roy. SOC., 1928, A , 120, 437BROCKWAY AND TAYLOR : STRUCTURE AND STEREOCHEMISTRY. 197202 ORGANIC CHEMISTRY.since small differences in energies of activation due to smalldifferences in structure are often responsible for very largedifferences in reactivity.At the same time the resonating moleculeis consistent with all of the observed dimensions. The inadequacyof the single Kekul6 structure is further demonstrated bycalculations of resonance energy, which could not exist in a singlestructure.Anthracene= and chrysene25 also show average ring sizes of1.41 A., a value which is consistent with the intermediate positionwhich these molecules also occupy between benzene and graphite.In the substituted benzenes the ring almost certainly retains itssize 1 7 s 1**19 except in derivatives in which the ring has severalconjugated groups attached. Several methylbenzenes have beenmeasured particularly to determine the length of the bond attachingthe methyl group to the ring.The best X-ray investigation hasbeen made on durene 26 with a resulting Car.-Cal. distance of 1.47 A.The latest electron diffraction examinations o f 1 : 3 : 5-trimethyl-and hexamethyl-benzenes show the value 1.54 & 0.02 A. In these,two values of the ratio Cm.-Cal. to C,,-C,, were used (equal to1.47/1-39 and 1.54/1.39). Qualitative differences between the twotheoretical curves when compared with the photographs led to achoice of the second ratio. Quantitative measurements then gavethe result and estimated error quoted above. It is hard to believethat either result is in error by as much as the difference betweenthem, and a discrepancy exists which can be resolved only withfurther investigation.The high value is also observed for methylethylenes (see above),so the occurrence of a double bond in the benzene ring would notbe expected to shorten the bond to the methyl group.It has beensuggested2' that the resonance in the ring may shorten theexternal bond; the same effect would then be expected in theparallel case of the bond attached to a carboxyl group. In oxalicacid dihydrate this bond is shortened by the conjugation of thetwo carboxyl groups ; but in ammonium oxalate monohydrate,28where the conjugation is inhibited by the twisting of the carboxylgroups into different planes, the carbon-carbon bond is longer than1-54 A. in spite of the resonance between the two C-0 bonds in eachcarboxyl group.The question of the bond length by which a non-conjugatinggroup such as methyl is attached is of some importance as a24 .J.M. Robertson, R o c . Roy. SOC., 1933, A, 140, 79.26 J. Iball, ibid., 1934, A, 146, 140.26 J. M. Robertson, ;bid., 1933, A , 142, 659.27 Ann. Reports, 1936, 33, 76.28 S. B. Hendricks and M. E. Jefferson, J . Chem. Physics, 1936, 4, 102BROCKWAY AND TAYLOR : STRUC!LWRE AND STEREOCHE.MISTRY. 203starting point from which to measure the effect of bond lengthwhen the benzene is part of a conjugated system. An interestingcomparison showing the effect of conjugation can be made amongthe three molecules, d i b e n ~ y l , ~ ~ stilbene,17 and tolan,ls whosestructures have been determined by Robertson (see Fig.6, p. 187).The three distances of interest are the bond length fixing the size ofthe benzene nucleus, that connecting the phenyl to the first atomin the side chain, and that connecting the two side-chain atoms(i.e., in the centre of the molecule). For dibenzyl Robertsonreported the three distances 1.41, 1.47, and 1-58 A., respectively.The accuracy of the location of the atomic positions is less in thismolecule than in the other two because of its more complicatedshape; and recently the investigator suggested 30 that the size ofthe benzene ring in dibenzyl should be reduced t o 1.39a., whichwould automatically raise the second distance to 1.49 A. Althoughthis still leaves a large difference between the side-chain bondlengths, the uncertainty of 0.03 A.in the positions of the methylenecarbon atoms would allow the difference to be materially reduced.In stilbene the three distances are 1.39,1.45, and 1.33 A. The thirddistance corresponds to the double bond. In tolan the distances are1.39, 1.40, and 1.19 A., the last representing the triple bond. As atpresent reported, the bonds holding the phenyl groups in the threemolecules have the distances 1.49, 1.45, and 1.40 A., respectively.The decrease in the second and third shows the effect on a singlebond which is involved in conjugation between a benzene ring anda double or triple bond, respectively. The extra effect in thelatter case may be due in part to the interaction of a triple on asingle bond even without conjugation as observed in methyl-acetylene.X-Ray investigations have been reported for diphenyl,31p-dipheny1benzeneF2 and 4 : 4’-diphen~ldiphenyl,~~ where now thebenzene rings are bonded together without intermediate atoms.In all these the ring size is reported to be 1.41 or 1 .4 2 ~ . and theconnecting bond lengths 1 . 4 8 ~ . If the bond from phenyl to asaturated carbon atom is 1.47-1.49~. long, as suggested by thedurene and dibenzyl results, then the distance 1 . 4 8 ~ . betweenphenyl groups seems much too large, as the following comparisonshows. Accepting the distance 1 . 5 4 ~ . for the bond between twosaturated carbon atoms (Le., each bonded to four atoms) and thedistance 1-48 A. for that between a phenyl group and one saturated29 J.M. Robertson, PTOC. Roy. Soc., 1935, A, 150, 348.30 Private communication.31 J. Dhar, Indian J. Physics, 1932, 7, 43.32 L. W. Pickett, Proc. Roy. SOC., 1933, A, 142, 333.33 Idem, J. Amer. Chm. SOC., 1936, 58, 2299204 ORGANIC CHEMISTRY.atom, we should expect the bond between two phenyl groups to benearer 1 . 4 2 ~ . than 1 . 4 8 ~ . I n addition the conjugation betweenthe rings should have a further shortening effect. On the otherhand, if the bond from the phenyl group to the saturated atom isalso near to 1.54 A., as suggested by the electron diffractionmeasurements on the methylbenzenes, then the shortening to1 . 4 8 ~ . in the diphenyl type molecules is easily understood as theconjugation effect between the benzene nuclei.The existence ofconjugation in these molecules is indicated by their planarstructures, a condition due to a degree of double bond character inthe connecting links, which is to be contrasted with the non-planararrangement of the dibenzyl molecule.Thus the carbon-carbon distances in nearly all of the hydro-carbons which have been investigated support a scheme of bondlengths in which the bonds between saturated atoms are 1.54 A.,double bonds are 1.34 A., and triple bonds are 1.20 A. long. Thesingle bond is affected little, if a t all, when it is adjacent to a doublebond or a benzene ring, but is shortened about 0 . 0 8 ~ . by anadjacent triple bond. I n molecules whose structures as ordinarilywritten contain alternate multiple bonds (conjugated systems), all ofthe bonds between carbon atoms linked to fewer than four atomsare resonance hybrids with distances intermediate betweenthe appropriate pair of the above standard lengths.(Refer-ences 2* lo* 12* 34 have been given to papers discussing the predictionof the lengths of hybrid bonds.) The two important exceptions tothis scheme, durene and dibenzyl, may be noted with the hope thatfurther investigations on molecules of this type may be undertaken.The lengths of carbon-hydrogen bonds have been measured inthe three typical cases with single, double, and triple bonds onthe carbon atom to which the hydrogen atom is bonded. Veryaccurate results have been obtained for the first and the third fromspectroscopic data. I n methane 35* l 3 the C-H distance is 1.093 A.and it is probable that practically this same value would beobserved whenever the carbon atom involved forms four singlebonds. When the C-H bond is adjacent to a triple bond, itslength is 1.057 A., this value being reported for acetylene andhydrogen cyanide.36 I n ethylene the spectroscopic analysis hasbeen performed only for the molecule containing the lighter isotopeof hydrogen, and hence only two of the three parameters requiredto fix the structure of the molecule have been supplied.Althoughthe C--H distance cannot be fixed with certainty, it very probably84 W. G. Penney, Proc. Roy. SOC., 1938 (in the press).35 N. Ginsburg and E. I?. Barker, J. Chem. Physics, 1935, 3, 668.36 P. Bartunek and E. F. Barker, Physical Rev., 1935, 48, 616HAWORTH : TRITERPENES.331This general scheme of origin of the products is also supported bythe identification of 2-hydroxy-1 : 8-dimethylpicene 27 from amyrin,and the assumption regarding the migration of the methyl group isstrongly supported by observations on some amyrin derivatives.34By Wolff reduction of the semicarbazone of amyrene (CO for CH*OHin 11), a hydrocarbon (111; R’ = R” = H) was prepared. Inaccordance with the hypothesis, 1 : 2 : 3 : 4-tetramethylbenzeneY1 : 2 : 5 : 6-tetramethylnaphthaleneY and phenolic compounds werenot produced by dehydrogenation of this compound, but 2 : 7-di-,1 : 2 : 5-, and 1. : 2 : 7-trimethylnaphthalenes and 1 : %dimethyl-picene were isolated. The dehydro-genation products of the compound(I11 ; R’ = Me ; R” = OH) were alsoA in agreement with anticipations ;HO*HC</)1 : 2 : 5 : 6-tetramethylnaphthalene (inthis case from rings A and B), 1 : 2 : 7- “’>n trimethylnaphthalene, and a new x picene homologue, probably 1 : 2 : 8-PV.) trimethylpicene, were obtained.OtherformulaeY35 e.g. (IV), consistent with the isoprene rule also supplya rational explanation of the dehydrogenation results.(b) Oxidative Degradation of Oleanolic Acid and Hederagenin.-This section will be confined to the behaviour of oleanolic acid andhederagenin towards oxidising agents, and the constitution of othertriterpenes will be deferred until later in the Report. Oleanolicacid, C30H4803,5* and hederagenin, C30H4,04,6* are mono- anddi-hydroxy-acids respectively, and esterification and hydrolysisexperiments indicate the tertiary character of the carboxyl groups.The acids contain a double bond, which resists catalytic reduction,but its presence is deduced from the formation of lactones36 andbromo-la~tones.~* 36 Although rigid proof is lacking, the acids areusually regarded as y8-unsaturated acids.Rings A and B.Important evidence concerning the structure ofthese rings has been obtained by oxidising derivatives in which thecarboxyl group is protected by ester or lactone formation.Oxidation of the methyl ester of hederagenin (V) with perman-ganate 37 yields the methyl esters of a hydroxy-keto-acid (VI) anda dibasic hydroxy-acid (VII) by oxidation of the secondary and theprimary alcoholic group respectively. When the methyl ester ofI vlu<I34 L.Ruzicka, H. Schellenberg, and M. W. Goldberg, Helu. Chim. Acta,35 Z. Kitasato, Acta Phytochirn., 1937, 10, (l), 199.313 Z. Kitasato and C. Sone, ibkl., 1932, 6, (2), 179.37 W. A. Jacobs, J. Biol. Chem., 1926, 88, 631.1937, 20, 791332 ORGANIC CHEMISTRY.hederagenin is oxidised with chromic one carbon atom iseliminated and a mixture of the methyl esters of a keto-acid,C2gH440, (VIII), and a dibasic keto-acid, C2gH,& (IX), is obtained.By oxidising the acids (VIII) and (IX) with hypobr~rnite,~~~ 39* 40the tribasic acids, C2gH4406 (X) and C28H2206 (XI) respectively,are obtained, the formation of (XI) involving the loss of oneadditional carbon atom. These and later results indicate a1 : 3-arrangement of the diol group, which must be present in aterminal ring (A), and necessitate the partial formulz given below.The synthetical product was identical with rubrene; it gave+-rubrene with acids, and yielded an oxide which evolved oxygenon heating.24 The intermediate compound (VIII), stereoisomericwith dihydroxydihydrorubrene,6* was converted into colourlessdehydrorubrene (IX) 9e 25* 26 by dehydration.An independent syn-thesis of rubrene has been effected by c. F. H. Allen and L.The reversible oxidation is, in itself, an argument in favour of (11) forrubrene. The formation of such colourless oxides is chaxacteristic ofbenzenoid compounds, and anthracene and diphenylanthracene exhibitreversible oxidation (C.Dufraisse a.nd M. Girard, Compt. rend., 1935, 201,428; C. Dufraisse and A. Etienne, ibid., p. 280; E. Willemart, ibid., p. 1201 ;1936, 202, 140).18 G. Schroeter, Ber., 1921, 54, 2242; L. I?. Fieser, J. Amer. Chem. SOC.,1931, 53, 2329; L. F. Fieser and E. L. Martin, ibid., 1935, 57, 1844.10 C. Dufraisse and M. Loury, Compt. rend., 1935, 200, 1673.2O C. Dufraisse and L. Velluz, Compt. rend., 1935, 201, 1394; Bull. SOC.21 A. Guyot and A. Haller, Bull. SOC. chim., 1904, 81, 795.chim., 1936, 3, 1905392 ORGANIC CHEMISTRY.Gilman,22 and the intermediate (VIII) obtained by these authors isidentical with dihydroxydihydrorabrene.CO CHPhCO PhThe formation of polynuclear hydrocarbons occurs frequentlywith meso-phenylnaphthacenes. The conversion of (VIII) into(IX) mentioned above provides one example, and 5 : 12-diphenyl-naphthacene is converted into the violet hydrocarbon (X) or theblue hydrocarbon (XI) by mild oxidation; (XI) is probablyidentical with a compound obta,ined previously from r~brene.~'.28Ph(IX.) (X.) (XI-)The new rubrene formula (11) gives a satisfactory explanation ofthe chemistry, and C. Dufraisse 3 has published a list of revisedformula for rubrene derivatives. One point, illustrated bydiphenyldichloronaphthacene, requires special comment. This sub-32 C. I?. H. Allon and L. Gilman, J . Arner. Chem. Soc., 1938, 58, 937.23 AUen and Gilman state that the diphenylnaphthacenequinone does notreact with phenylmagnesium bromide and therefore cannot be an intermediatein the Dufraisse synthesis.In a reply to this criticism (Bull. SOC. chim., 1936,3, 2175) it is pointed out that a large excess of the Grignard reagent isnecessary for the reaction.24 The oxides of naphthaceno and 5 : 12-diphenylnaphthacene do notevolve oxygen on heating; naphthacene oxide yields some 5-keto-5 : 12-dihydronaphthacene. The elimination of oxygen is facilitated by accumul-ation of meso-phenyl groups as in rubrene.?5 C. Dufraisse and L. Enderlin, Compt. rend., 1932, 194, 183.56 L. Enderlin, ibid., 1938, 202, 1188.27 C. Dufraisse and R. Girard, Bull. Soc. chim., 1934, 1, 1359.M. Radoche, Ann. China., 1933, 20, 200IIAWORTII : RUBRENES AND AZULENES. 393stance, readily obta,ined by the action of heat on (XII) 1i29 andpreviously represented on the basis of (I) by the planar symmetricformula (XIII), is now given the centro-symmetric formula (XIV).The latter formula is more in accordance with its conversion intothe violet (XV) 30 or blue (XI) compounds, by loss of one or twomolecides of hydrogen chloride respectively.c1 CPh CPh Ph C1(XII.) (XIII.) (XIV.)The new formula, however, still experiences a difficulty inexplaining the conversion of rubrene into 4-rubrene. The presenceof four meso-phenyl groups is essential for this reaction and it issuggested that these facilitate the addition of acid, HX, to the9 : 10-positions to give a compound of type (XVI), which isconverted into +rubrene (XVII) by loss of HX.3lPh C1 H Ph H Ph\-/(XV.) (XVI.) (XVII.)Several mechanisms for the formation of rubrene have beensuggested.39 22Axulenes.The blue colour of camomile oil was first observed in the fifteenthcentury, and since then it has been shown that, in about 20% ofthe cases investigated, essential oils contain or give rise to blue orviolet substances known as azulenes.1 I n general, the developmentof a blue colour in essential oils may be induced by dehydrogen-ation with selenium , sulphur, nickel or palladiurn-charcoal.2 Thepresence of azulenes or azulene-yielding compounds is detected byaQ C.Dufraisse and R. Buret, Compt. rend., 1932, 195, 962.30 C. Dufraisse, R. Buret, and R. Girard, Bull. SOC. chim., 1933, 55, 702.31 The oxidation of naphthacenequinone to naphthoic and benzoic acidsIt is therefore suggested Chat proves the 5 : 12-structure of the quinone.HX adds to the 5 : 12-positions.1 A.S. Pfau and P, Plattner, HeZv. Chim. Acta, 1936, 19, 858.2 L. Ruzicka, S. Pontalti, and F. Balas, ibid., 1923, 6, 855394 ORGANIC CHEMISTRY.the formation of' a blue-violet colour on the slow addition ofbromine to a chloroform or acetic acid solution of the essential 0i1.3In 1915, A. E. Shernda14 observed that the blue colour wasremoved from ethereal or petrol solutions of the oils by shakingwith concentrated phosphoric acid solution, from which theazulenes were regenerated by dilution with water. The azulenesobtained from cubebs, camphor or gurjun oils, yielded dark-colouredpicrates, and by titration of the picrate, the formula C,,H,, wasdeduced for the hydrocarbon from oil of cubebs.4~A large number of supposedly different azulenes have beenreported and named from their respective natural oils, but it nowappears that in reality only four or five isomeric and distinctsubstances are involved.S-Guaiacazulene, blue needles, m. p.30°,14 occurs naturally in geranium oil1** and in some coal tars.12It has been obtained by the action of sulphur on the sesquiterpeneor sesquiterpene alcohol fraction of the oils from guaiacum WOO^,^* 9callistris,1° patchouli,l gurjun balsam 1s ' 8 l1 and EucalyptusgZobuZus,l. and also by nickel dehydrogenation of gurjun balsamoi1s.l. l2 Camazulene, m. p. 132",l6 isolated from camomile andyarrow oils,'^ lactarazulene, a liquid recently l6 obtained from thefungus, Lactarizbs deZiciosus L., and vetivazulene, violet needles,m.p. 32",15 obtained by the action of sulphur on vetiver oil,l*15resemble S-guaiacazulene in properties. Elemazulene is obtainedin the form of a liquid in 1% yield by the action of selenium onelemol, a crystalline sesquiterpene alcohol of unknown structure.The product obtained from an essent.ia1 oil depends to some extenton the dehydrogenating agent and no azulene is obtained bytreating elemol with sulphur. A violet azulene previously knownas Se-g~aiacazulene,~ prepared by the action of selenium onguaiacum wood oil, has recently1*14 been proved a mixture ofazulenes from which S-guaiacazulene may be isolated.1920, p. 417.3 R. T. Baker and H. G. Smith, " k Research on the Eucalypts," Sydney,4 J .Amer. Chem. SOC., 1915, 37, 167, 1537.5 L. Augspurger, Science, 1915, 42, 100.6 R. E. Kremers, J. Amer. Chern. SOC., 1923, 45, 717.7 L. Ruzicka and E. A. Rudolph, HeZv. Chim. Acta, 1926,9, 118.8 P. Barbier and L. Bouveault, Compt. rend., 1894, 119, 281.9 L. Ruzicka and A. J. Haagen-Smit, HeZv. Chim. Acta, 1931,14, 1104, 1122.10 Y. Asahina and S. Nakanishi, J. P h m . Soc. Japan, 1932, 52, 1, 2, 5, 12.11 W. Treibs, Ber., 1935, $8, 1751.l2 J. Herzenberg and S. Ruhemann, Ber., 1925, 58, 2249; 1927, 60,13 J. Melville, J. Amer. Chem. Soc., 1933, 55, 3288.l4 K. S. Birrell, ibid., 1934, 56, 1248; 1935, 57, 893.15 A. St. Pfau and P. A. Plattner, Helv. Chim. Acta, 1937, 20, 224.2459 ; I?. Schlapfer and 0. Stadler, HeZv.Chim. Acta, 1926, 9, 185HAWORTH : RUBRENES AND AZULENES. 395The Sherndal method of isolation with phosphoric acid has beenextensively used, and the azulenes are characterised by picratesand styphnates, or better by the more stable and more highlycrystalline compounds with trinitrobenzene l6 or trinitrotoluene.Decomposition of the picrate and styphnate is effected withammonia, and the azulene is recovered from the trinitrobenzenecompound by reduction with ammonium sulphide, followed bydistillation with steam. Recently l5 the decomposition of thesepolynitro-derivatives has been improved by chromatographicadsorption on alumina; when cydohexane is used as solvent, theazulene derivative dissociates, the polynitro-compound is adsorbed,and the azulene is recovered from the filtrate in a very pure state.Catalytic reduction of the azulenes gives an octahydro-compound,C15H26,4 but refractivity measurements 7p suggest that theazulenes contain it dicyclic system with five double bonds, one ofwhich resists reduction.Oxidation of the azulenes with perman-ganate 6m gives acetic, isobutyric and ovalic acids and acetone, andozonolysis of the partly reduced azulenes (mainly tetrahydro-derivatives) 13* l4 yields formaldehyde, acetone, formic, isobutyricand a-rnethylglutaric acids. When the azulene-containing fractionsof vetiver or guaiacum wood oils are heated with hydriodic acidand red phosphorus, substances are produced which give smallyields of the naphthalene homologues (III) and (IV) whendehydrogenated with sulphur.The simplest explanation of theformation of (111) and (IV), which have been identified bycomparison with synthetic specimens,l7 would be to postulate aeudalene skeleton for the azulenes. This, however, is inconsistentwith the observations that fully reduced compounds of the eudesrnoltype, unlike decahydro-S-guaiacazulene, give naphthalenes but noazulenes on dehydrogenation.1 Oxidation of decahydroguaiac-azulene results in ring scission to a dibasic Cl5-acid, which onpyrolysis gives a CI4 ketone. Catalytic dehydrogenation of thisketone has furnished a, phenolic compound, thus indicating the16 H. Willstaedt, Ber., 1935, 68, 333; 1936, 69, 997.17 L. Rueicka, P. Pieth, T. Reichstein, and L. Ehmann, Ber., 1933, 66,268396 ORGANIC CHEMISTRY,presence of a seven-membered ring in the azuleiies.15 Thesuggestion is made that the naphthalene derivatives (111) and (IV)are formed from guaiacazulene (I) and vetivazulene (11) respectivelyby a retropinacolinic change, and structure (11) for vetivazulenehas been established by the synthesis of the phenol obtained bythe degrada,tion outlined above.lsThese suggestions are supported by the synthesis of a number ofcompounds of type (VIII).By ozonolysis of octalin, W. Hiickel,A. Gercke, and A. Gross l9 obtained the diketone (V), which wasreadily converted by aqueous sodium carbonate into the cyclo-pentenocycloheptanone (VI).20 A. S. Pfau and P. Plattriercondensed this ketone with Grignard reagents and dehydrogenatedthe products (VII) with nickel or sulphur to give the substitutedazulenes (VIII).The products (VIII; R f= Me, Et or Ph), isolatedR N n RW e ) (VI. 1 (VII.) (VIII * )by the phosphate method, were purified as picrates or thetrinitrobenzene compounds. Azulene, the parent hydrocarbon(VIII; R = H), was prepared l5 from (VI) by reduction, first tothe dihydro-compound and then to the carbinol, and the latter wasdehydrogenated with palladium-charcoal to ( V I I I ; R = H). Theyields of azulenes are about 5% and as they, particularly (J7111;R = H), are unstable to acids and air, the chromatographicprocedure is frequently employed in the decomposition of thepolynitro-derivatives. The synthetic compounds have an intenseblue colour, indistinguishable from that of S-guaiacazulene in dilutealcoholic solution and a comparison 1 5 3 21 of the absorption spectrabetween 2300-7500 A.suggests that azulene, S-guaiacazulene, andvetivazulene belong to the same class. Consequently vetivazuleneand S-guaiacazulene are regarded as 4 : 8-dimethyl-2-isopropylazulene(11) and 1 : 4-dimethyl-7 -isopropylazulene (I) respectively.The occurrence of this cyclopentane-cycloheptane system in theterpene family is of great interest.22 The famesol chain can beThe experimental details of the degradation of khe azulenes to, and thesynthesis of, these phenols have not yet been published.Is Ber., 1933, 66, 563.80 W. Huckel and L. Schnitzspahn, AnnaZen, 1933, 505, 2742 1 B. Susz, A. S. Pfau, and P. A. Plattner, Helv. Chim. Acta, 1937, 20, 469.33 The occurrence of azulenes, as such, in fungi or essential oils has notbeen definitely established ; they may be produced during extraction 16 (seealso Chem. Zen&., 1934, I, 81)IIAWORTH : RUBRENES AND AZULENES. 397arranged so as to give nine structures of the cyclopentane-cyclo-heptane type and vetivazulene (11) and guaiacazulene (I) representtwo of these forms.Azulenes may be obtained from sources other than the terpenes.The blue distillate reported by W. Hentzschel and J. Wislicenus 23as a by-product during the preparation of cyclopentanone bydistillation of calcium adipate, contains azulene, which has beenidentified spectro~copically.1~ The production of azulene in thisreaction is probably due to traces of impurity which facilitatedehydrogenation and it has been shown that the addition of nickelassists the azulene formation. The ketone (VI) is probably formedby the route outlined below, and converted into (VIII; R = H),[CH214*CO&Adipjc acid --+ CO --t\[CH2],*C02Hby the action of traces of nickel or impurities ; it has been shownthat (VI) is partly converted into (VIII; R = H) with selenium orpalladium-carbon . R. D. H.L. 0. BROCKWAY.R. D. HAWORTH.R. P. LINSTEAD.T. G. PEARSON.S. PEAT.T. W. J. TAYLOR.2:i Annalen, 1893, 275, 312