ORGANIC CHEMISTRY.PART I ,-ALIPHATIC DIVISION.Pyrolysis of Hydrocarbons.STUDY of the action of heat on paraffin hydrocarbons has demon-strated that two types of reaction occur, degradative and additive.I n the thermal decomposition of the normal paraffins rupture of thechain appears to occur a t any position with production of an olefinand the complementary lower paraffin or, a t the limit, hydrogen.Methane represents a special case, possessing a higher decompositionpoint than its homologues. With this hydrocarbon little decom-position occurs below lOOO", a t which temperature both ethyleneand acetylene are formed; a t higher temperatures under atmos-pheric pressure decomposition into carbon and hydrogen is prac-tically complete. For the maximum production of ethylene andacetylene, reduced pressure is essential and a t 1200"/100 mm., withsuitable rates of flow, 4% of ethylene and 2.5% of acetylene areformed ; above this temperature the quantity of ethylene diminishesto zero while the amount of acetylene increases (14.5y0 a t 1500"/50mm.).With isobutane, fission a t CH linkings to yield hydrogenand isobutylene becomes a major reaction.The formation of ethylene and acetylene from methane affords asimple instance of the additive reactivity which is concerned in theaddition of olefin to olefin. To the interaction of the olefins whichare produced a t an early stage of the thermal decomposition of theparaffins, the formation of aromatic hydrocarbons is attributed.Consequently the behaviour of the olefins under the action of heatis a matter of considerable practical importance.Examination of the thermal decomposition of ethylene, propylene,and the two n-butylenes3 suggests that the principal primaryreactions in each instance are the formation of either two-carbonor four-carbon olefins (or of both) : 2C,H4 --+ C4H,; 2C,H, -+1 For a review of the extensive literature dealing with the decompositionof paraffin hydrocarbons by thermal means, with and without catalysts, byelectrical means, by photosensitisation, and by a-particles, see G.Egloff ,R. E. Schaad, and C. D. Lowry, jun., J. Physical Chem., 1930,34, 1617; A.,1268.2 E. N. Hague and R. V. Wheeler, J., 1929, 378; A., 1929, 536; F. deRudder and H. Biedermann, Bull. SOC.chim., 1930, [iv], 4'4, 704; A., 1268;C. D. Hurd and L. U. Spence, J. Amer. Chem. SOC., 1929,51, 3353, 3561 ; A.,1930, 58, 191.R. V. Wheeler and W. L. Wood, J , , 1930, 1810; A . , 1309ORGANIC CHEMISTRY .-PART I. 83C,H4 + C4H8 ; C,H, --+ 2C,H4. Thus polymerisation and degrad-ation occur side by side, but the primary processes are speedilyfollowed by dehydrogenation (a primary mode of decomposition inthe case of butylene). In the presence of the hydrogen so liberated,the olefins decompose through the scission of the carbon chain at thelast C - C linking, forming radicals which by hydrogenation producemethane and the next lower olefin : CH,*CH:CH, --+ CHI,* +*CH:CH, ?!$ CH, + CH2:CH, ; CH,*CH,*CH:CH, -+ CH3* +*CH,*CH:CH, ?$ CH, + CH,-CH:CH,.Butadiene, formed bydehydrogenation of butylene, is thought always to be present in theearly stages of decomposition and it is suggested that the combin-ation of ethylene and butylene, of propylene and butadiene, and ofbutylene and butadiene yields hydroaromatic hydrocarbons whichat higher temperatures lose hydrogen to yield benzene and itshomologues .The degradation of the paraffins can be viewed as proceeding bythe reversible elimination of a molecule HX, which may be either aparaffin of smaller molecular weight or hydrogen itself. The residueis then an olefin, and the process is comparable with the eliminationof acid and water from any olefin addition p r o d ~ c t . ~The additive process can probably proceed in two ways, whichare cha.racteristic of olefin polymerisation and butadiene-olefinaddition respectively.Both of these processes are discussed later(pp. 91,88) and need no further mention here, but it would seem fromrecent developments in these directions that of the two mechanismswhich have been suggested for the production of hydroaromaticcompounds, vix. (1) direct butadiene-olefin addition and (2) theintermediate formation of (open-chain) hexadienes or hexatrienesby addition of olefin to olefin, the former is the more probable.Carbon Unsaturation.Mono-ole$nic and Mono-acetylenic Compounds.-Inability toobtain homogeneous specimens or to determine the composition ofmixtures of isomerides is frequently a great disadvantage of thecommon methods of olefin preparation.For this reason the dis-covery of a general method for obtaining pure Aa-olefins by theaction of zinc on p-alkyloxy-n-alkyl bromides, and the elaborationof a method for estimating the proportion of admixed Aa- andA@-butenes by determining the rate of reaction of their dibromideswith potassium iodide, are of importance.T. M. Lowry, J . , 1929, 378; A., 1929, 536.H. B. Dykstra, J. F. Lewis, and C. E. Boord, J . Arner. Chem. SOC., 1930,R. T. Dillon, W. G. Young, and H. J. Lucas, ibid., p. 1953; A., 888 ;52,3396; A., 1269.W. G. Young and H. J. Lucas, ibid., p. 1964; A., 88884 FARMER :Announcements have been made with regard to the occurrenceof an unexpected variety of isomerism in AB-pentene.7 Thehydrocarbon as prepared by dehydrobromination of y-bromopentane,or dehydration of diethylcarbinol, is reported to pass under theinfluence of ultra-violet light into an isomeride.This differs fromthe fht compound only slightly in physical properties but verymarkedly with regard to the proportion of p- and y-bromopentaneproduced therefrom by the action of hydrogen bromide. Toexplain the relationship of the isomerides, the absorption spectraof which differ sufficiently markedly to exclude their representationas cis- and trans-isomerides, a species of electronic isomerismdependent on unequal sharing of electrons at the double linking ispostulated.The observation that cis-forms of ethylene compounds suffercatalytic reduction more readily and smoothly than the trans-varieties has been found to apply to isocrotonic acid, to isostilbene,and to cis-o-hydroxy- and cis-o-ethoxy-cinnamic acids, all of whichare more readily reduced to saturated acids than are the correspond-ing trans-compounds.* On applying the method to confirm theconfigurations of erucic and brassidic acid, it is found that thedifference in their rate of hydrogenation (to behenic acid) is lessmarked than in the case of oleic and elaidic acids, but is sufficient toestablish them as the cis- and trans-forms of the acidCH,*[CH,] ,*CH:CH*[ CH,], ,*CO,H.In spite of contradictory results obtained by other methods itappears from the study of eleven triple-bonded substances ofdifferent types that hydrogenation of acetylenes in the presence ofcolloidal palladium stabilised by starch uniformly gives the cis-ethylenic derivative .gThe very different courses followed by hydrogenation processeshave caused considerable speculation as to the state of hydrogen(H+, H(n), or H-) entering into reaction.By employment of anapparatus recently described, the action on organic compounds ofatomic hydrogen from a Geissler tube may be ascertained.10 Oleicacid and ethyl phthalate are reduced thereby, but considerablequantities of polymeric materials are produced in each instance.7 M. S. Kharasch and F. R. Darkis, Chem. Reviews, 1928, 5, 571 ; E. P. Carr,A., 1929, 1420; M. L. Sherrill, C. J . Amer. Chem. SOC., 1929, 51, 3041;Baldwin, and D. Haas, ibid., p. 3034; A., 1929, 1419.8 C. Pael and H. Schiedewitz, Ber., 1930, 63, [B], 766; A., 740.9 M.Bourguel, Bull. SOC. chim., 1929, [iv], 45, 1067; A., 1930, 317.Compare J. S. Salkind and V. K. Teterin, J . Russ. Phys. Chem. SOC., 1929,61, 1751 ; A., 1930, 574.10 W. Nagel and E. Tiedernann, Wiss. Ver68. Siemene-Konz., 1929, 8, [2],187; A,, 1930,577ORGANIC UHEMISTRY.-PbT I. 85The action of sulphur l1 and of amines l2 in bringing aboutcis-trans transformation of ethylenic acids has been studied. Theconversion of methyl maleate into methyl fumarate is catalysed byammonia and primary or secondary amines, but not by tertiaryamines. Mechanisms are suggested for this transformation and forthe Knoevenagel reaction : in the one case the disappearance ofthe configurational constraint associated with the double bond andin the other the production of appropriate kationic and anionicforms of the reactants are considered to be the consequence of theformation of co-ordinate links between the hydrogen atom of thebase and carbonyl oxygen >C=O-A HX.For the formerreactivity, however, the suggested mechanism is improbably ofgeneral applicability, since there is reason to believe that in someinstances the cis-trans isomerisation of Knoevenagel products(although not the formation) involves ap,py-isomerisation l3 andwith regard to the latter there is as yet no conclusive evidence toshow that the Knoevenagel reaction is uniquely a reaction involvingthe non-enolised carbonyl group of a ketone or aldehyde, althoughthere is much evidence to show that the reaction is not dependenton the enolisation of the carbonyl g r 0 ~ p .l ~The question of catalytic mechanism arises again in connexionwith the Michael reaction. As the result of the work of J. F.Thorpe lS on the addition of ethyl sodio-a-cyanopropionate toethyl P p-dimethylacrylate the conclusion was unavoidable that inthis instance eithr addition could occur to some extent by partitionniNaof the sodio-addendum in the manner or, ifaddition did indeed involve the partition Na-[-CMe(CN)*CO,EtMe-$!(CN)*CO,Et’in the way usually supposed to be characteristic of the Michaelreaction, then the product could be further alkylated by methyliodide at the a-carbon atom of the resulting sodio-ppy-trimethyl- - Thorpe y- c yanoglutaric ester,held the view that the methyl group in the methyl-cyano-ester andthe hydrogen atom in the corresponding unmethylated cyano-esteractually separated, the sodium atom remaining in the ester residue.NaEtO,C*CMe( CN)*CMe,*CH*CO,Et*l* G.Rankoff, Ber., 1929,62, [B], 2712; 1930,63, [B], 2139; A., 1930, 65,la G. R. Clemo and S. B. Graham, J., 1930, 213; A., 452.l8 E. H. Farmer and J. Ross, J., 1926, 1570; A., 1926, 834.l4 J. C. Bardhan, ibid., 1929, 2225; 1930, 1509; A., 1929, 1462; 1930,Compme A. Michael, Ber., 1900, 38, 3731; A.,1406.1170; C. K. Ingold, a i d . , p. 184; A., 1170.1901, i, 123.Ibid., 1900, 77, 92386 FARMER :Convincing evidence has been furnished to show that in thepresence of a full molecular proportion of sodium ethoxide thepartition of ethyl methylmalonate involves entirely the separationof the methyl group as kationic component; with a fractionalmolecular proportion (conveniently 9) , however, the methyl groupfor the most part remains attached to the ester residue, but to aminor extent (about 10%) separates as the kationic fragment.16The action of sodium ethoxide in small proportion is considered tobe purely catalytic, contributing towards a reaction mechanismcomparable with that encountered in catalytic cis-trans-ethenoidisomerisation, but differing from that which governs addition whenthe addendum is a sodium enolate.17 Strong ground is afforded forpresuming that the characteristic partition of non-alkylated sodio-addenda (sodium enolates) involves the separation of hydrogen andnot of sodium.Investigations relating to the manner of addition to olefins ofhydrogen halides ,l8 carbonyl chloride ,l9 and halogenated amines 2ohave been carried out.The results are of importance as reflectingbroadly the polar influence of the ethylenic substituents on thecompetitive vformation of alternative additive products (>C,X*C,Yand >C,Y*C,X), but as yet the quantitative data available in thisimportant field are few.PolyoleJinic and Polyacet ylenic Compounds.-With the recognitionthat the substances reputed to be aa-dimethylbutadiene 21 andp-cyclopropylpropene 22 are actually ay- and aa-dimethyl- buta-diene respectively, the series of monomethyl- and dimethyl- buta-dienes becomes complete. It is now abundantly clear from themanner in which bromine becomes added to the as-, my-, and py-di-methylbutadienes 23 that the formation of both adjacent andterminal dibromides must be regarded as embodying primaryreactivities of the conjugated molecule : that is to say, the form-ation of as-, EC-, etc., additive compounds is not to be regarded astaking place via the formed molecules of corresponding ap-com-l6 A.Michael and J. Ross, J . Amer. Chem. SOC., 1930, 52, 4600.l7 For non-catalytic addition the addendum in the case of ethyl malonateis represented by H-C(C0,Et) : C(ONa)-OEt, analogous to J. F. Thorpe'srepresentation of ethyl sodiocyanoacetate.la C. C. Coffin, H. 8. Sutherland, and 0. Maass, Canad. J . Res., 1930, 2,267 ; A., 888 ; R . Lespieau, BUZZ.SOC. chim., 1930, [iv], 47, 847 ; A . , 1401lD E. Pace, Gazzetta, 1929, 59, 578; A . , 1929, 1419.2o Z. Foldi, Ber., 1930, 63, [ B ] , 2257; A., 1423.21 0. Diels and K. Alder, AnnaZen, 1929, 470, 98; A., 1929, 819; E. H.Farmer, C. D. Lawrence, and W. D. Scott, J., 1930, 510; A., 572.z2 P. Bruylants, Bull. Acad. roy. Belg., 1908, 1011; A., 1909, i, 226; N. vanKeersbilck, Bull. SOC. chirn. Belg., 1929, 38, 205; A , , 1929, 1163.43 E. H. Farmer, C. D. Lawrence, and W. D. Scott, Zoc. citORGANIC CHEMISTRY .-PBRT I. 87pounds. It is equally apparent that the bromination products,one or all in each case, are not necessarily mobile, The new factsare consistent with the additive mechanism of H. Burton andC. K. IngoldF4 but run counter (as in the writer’s view do all theobservations of the last few years concerning additive processes inrelation to anionotropic change) to a more recent formulation.25The latter is compatible with the possession by a butadiene of twopotential structures, di-ethylenic and conjugated, and obscures thestrong experimental indication that the orientation of the additiveproducts is related to the double degree of unsaturation only to theextent that the latter is able to contribute to the formation of amobile propene system when once simple ethylenic addition hasbeen initiated. In this connexion it is interesting to note thatexamination of the Raman spectra of butadiene, piperylene, iso-prene, and By-dimethylbutadiene has shown no reason for assumingthat the C:C linking in conjugated systems, whether open or closed,is in a state distinct from that of an ordinary double linking of theally1 type.26The variable addition to conjugated substances demonstrated forhalogen addition has now been shown to characterise sodio-esteraddition 27 and hydrogenation (non-catalytic) .28 Esters of themalonic type yield both aP- and as-addition products with the estersof the butadiene-a-carboxylic acids and likewise the reduction ofthe latter acids by metals which react with water gives both Ay- andAs-butenecarboxylic acids.In all these instances there appearsto exist a definite ratio between the amounts of the ap- and as-products formed which depends largely on the constitution of theconjugated compound but to some extent on other factors.Forexample, methyl substitution at the 6-, p-, and pa-carbon atoms inp-vinylacrylic acid causes a change in the proportion of the @- and as-reduction products in the direction required by considerations ofprototropy, with which hydrogen addition has been correlated.2q Itmust be remembered, however, that the experimental methods24 J., 1928, 912; A., 1928, 634. Some misconception has arisen (compareI. E. Muskat and H. E. Northrup, J . Amer. Chern. Soc., 1930, 52, 4048; A.,1553) from a statement of these authors to the effect that in the addition ofhalogens the initial product could only be a 1 : 2-compound. It seems clearfrom the context and from a subsequent re-statement (Ann. Report8, 1928,25, 132) that the authors have in mind the attack by the reagent at a singleethylenic centre and not the actual formation of the ajl-dibromide.26 Ann. Reports, 1929,26, 121.26 A.Dadieu and I(. W. F. Kohlrausch, Ber., 1930, 63, [ B ] , 1657; A,,2 7 E. H. Farmer and T. N. Mehta, J., 1930, 1610; A , , 1163.28 J. T. Evans and E. H. Faxmer, Chem. and Ind., 1928,47,268; J., 1928,1162.1644; A., 1928, 86888 FARMER :available for estimating the @,a6 ratio in all additive processesleave much to be desired and a t present the significance of quantita-tive results is not to be rated too highly.In many reductions of butadiene acids by metals simple reductionproducts are not solely or principally formed, but in their placehydrogenated double molecules appear.29 These are (so far as hasbeen ascertained) linked at the @-carbon atoms and are prone toundergo secondary intramolecular change as illustrated by theformation of a bicyclohexane derivative (A, below) from ethylcrotylidenemalonate. A similar type of reductive polymerisationis that whereby isoprene passes on treatment with potassium andalcohol to a mixture of Pc-, Pq-, and yc-dimethyl-ABf-octadienewhilst py-dimethylbutadiene passes with the same reagents toa mixture of pycq-tetramethyl-A~~-octadiene (B) and methylrubber 30 ; the interesting observation is also made that isoprenemay be simultaneously ethylated and reduced by the action ofpotassium and ethyl bromide to yield S-methyl- As-octene.CH,*QH*QH*y( CO,Et), yH2*CMe:CMe*CH,CH,*CH*CH-C( C02Et), CH,*CMe:CMe*CH,(A*) (B.1The simple addition products of the butadienes are for the mostpart oily substances and no satisfactory means of characterising thehydrocarbons has been available.Owing to the discovery thatmaleic anhydride, acetaldehyde, acraldehyde, acrylic ester, p-benzo-quinone, a-naphthaquinone, and other substances possessing thegroup CH:CH*CO unite quantitatively with the butadienes, fre-quently when the reactants are mixed at room temperature;l anelegant and trustworthy method of characterisation is provided.In addition, the underlying reaction is of great synthetic value andpresents a number of points of theoretical and practical interestwhich may be summarised thus : (1) No single instance has beendiscovered in which other than complete terminal attachment of thenon-dividing addendum occurs, so that all the known products maybe typified by those derived from the combinations isoprene-maleicanhydride, butadiene-acraldehyde, and butndiene-benzoquinone,(I), (11) , and (111) , respectively.29 C.M. Cawley, J. T. Evans, and E. H. Farmer, J., 1930, 622; A., 578.80 T. Midgley, in., and A. L. Heme, J . AmeT. Chem. SOC., 1929,51, 1293,1294; 1930,52,2075,2077; A., 1929, 674; 1930, 888.81 0. Diels and K. Alder, Annalen, 1928, 460, 98; A , , 1926, 1018; 0. Diels,K. Alder, and E. Naujoks, Ber., 1929, 62, [ B ] , 664; A . , 1929, 670; 0. Diels,K. Alder, W. Lubbert, E. Naujoks, F. Querberitz, K. Rohl, and H. Segeberg,Annalen, 1929,470, 62; A,, 1929, 819; 0. Diels, K.Alder, and P. Pries, Ber.,1929, 82, [B], 2081; A., 1929, 1297ORGANIC CHEMISTRY.-PART I. 89II IvCH2CH CH*CHO(111.)vCH2(1.1 (11-1(2) The butadiene acids (e.g., sorbic and muconic acids) yieldcarboxylated tetrahydrophthalic anhydrides with maleic anhy-dride,32 but in the formation of these, py,ap-double bond changemay occur.(3) Cyclic butadienoid compounds (cyclopentadiene, cyclohexadiene, furan, N-methylpyrrole) yield bridged anhydrocyclohexeneson treatment with maleic anhydride; 31 benzoquinone may unitewith one or with two molecular equivalents of open-chain or cyclicbutadienes to yield hydronaphthaquinone and hydroanthraquinonederivatives respectively ; ct-naphthaquinone combines with onemolecular equivalent of open-chain or cyclic butadiene to yieldhydroanthraquinone derivatives.The compounds arising from thecombinations butadiene-ct-naphthaquinone, cyclohexadiene (2 mols.)-benzoquinone ( 1 mol.) , N-methylpyrrole-maleic anhydride arerepresented by (IV), (V), and (VI).CO CH2 CH/ a $ ! H A C H / I \ IfI \scH@ I1 RH QH2 F!W.) (V.1 (VI.)Many of the hydroanthraquinone derivatives, including (V), areconvertible into the corresponding anthraquinones.(4) The hexatrienes and higher poly-ene hydrocarbons also yieldcyclohexene derivatives ; e.g., both the cis- and the trans-form ofhexatriene yield the same ethylidenecyclohexene 32 (VII) withmaleic anhydride, whilst a8-diphenyloctatetrane unites with thesame reagent (even one molecular equivalent only) straightway togive (VIII).34 The course of reaction of the higher hydrocarbons ofstructure Ph-[CH:CH],*Ph appears to be dominated by the strongreactivity of the terminal methine groups, so that with aK-diphenyl-32 E.H. Farmer and F. L. Warren, J., 1929, 897; A., 1929, 812.33 0. Diels, K. Alder, G. Stein, P. Pries, and H. Winckler, Ber., 1929, 62,34 R. KuhnandT. Wagner-Jauregg, Ber., 1930,&3, [B], 2662; A., 1680.[ B ] , 2337; A., 1929, 130390 FARMER :decapentaene, and apparently with still higher members of the series,cyclohexene formation is initiated at both ends of the unsaturatedchain as would be anticipated.C/(VII.)A A[A = *CO*O*CO*](VIII.)When many of the pairs of reactants in the Diels-Alder reactiona.re mixed, colorations rapidly appear, which disappear whenreaction is complete.These probably mark the first stage ofadditi0n,~2,34 presumably corresponding to attachment of theaddendum at one end of the conjugated system. Another interestingobservation concerns the colour of conjugated hydrocarbons them-selves.35 It has been concluded from the study of synthetic andnatural dyes that an aliphatic hydrocarbon must possess 5-6double bonds in unbroken conjugation for colour to develop;moreover, the colour-conferring value of a free carboxyl groupconjugated with double linkages would seem to correspond to about14 double linkages. On this basis the first member of the sorbic acidseries which would be expected to show absorption bands extendinginto the visible part of the spectrum is decatetraenoic acid,CH3-[CH:CH],*C0,H.By the isolation of both octatetraenoicacid and decatetraenoic acid this expectation has been fulfilled,since the former acid is colourless whilst the latter is intenselyyellow.The constitution of the well-known crystalline compound ofisoprene and sulphur dioxide has now been el~cidated.~~ Thesulphur dioxide acts as a non-dividing addendum and in this instancebecomes attached at the terminal carbon atoms of the isoprenechain, yielding P-methyl-As-butene-ct8-sulphone.By the action of iodine (2 atoms) on acetylenic Grignard reagentsof the type CRiCMgX (2 mols.) a series of conjugated diacetylenes(R*CiC*CiCR) has been ~btained.~' These include A*C-decadi-inene,Aq-dodecadi-inene, Ace-tetradecadi-inene and a6-diphenyldi-acetylene. The behaviour of the last of these when catalyticallyhydrogenated is remarkable : the product contains both dihydro-35 R.Kuhn and M. Hoffer, Ber., 1930, 63, [B], 2164; A., 1406.36 E. Eigenberger, J . pr. Chem., 1930, [ii], 127, 307; A., 1405.37 V. Grignard and Tchdoufaki, Compt. Tend., 1929, 188, 357, 1531; A . ,1929, 290, 907ORGANIC CHEMISTRY .-PART I. 91and tetrahydro-derivatives which in form at least are representableas products of as-, apy6-, and aa66-addition.CHPh:C:C:CHPhCH,Ph*CiC*CH,PhCPhiCCiCPh -+ CHPh:CH*CH:CHPh (c~s-c~s)Polymer isation.Hydrocarbons.-The more immediate problems connected withthe formulation of the additive and degradative processes associatedwith the different stages of polymerisation and depolymerisation ofolefinic hydrocarbons concern (1) the structure of dimerides and theuniformity (or otherwise) of the additive processes underlying theirformation, and (2) the genetic relationship of monomerides andhigher polymerides.aayy-tetra-phenyl- A.8-butene, by the act,ion of aluminium chloride, iodine, orsulphuric acid.The last reagent, however, converts the unsaturateddimeride initially produced into a saturated isomeride which (likethe dimeride obtained by the action of aluminium chloride onstilbene) has been deemed to be a derivative of cy~lobutane.~~ Itis now shown that the saturated dimeride is 1 : 3 : 3-triphenyl-hydrindene, formed by a simple cyclisation of the unsaturatedi~omeride.~~ Likewise a specimen of diisobutylene has been shownto contain p8B-trimethyl-AP- and -Aa-pentene~.~OThe formation of cyclobutane derivatives cannot be considered asa characteristic of the dimerisation of mono-olefins : indeed suchpolymerisation is increasingly found to conform to an additivetype which resembles the Michael reaction in that hydrogen separatesfrom the addendum molecule, the components becoming added atthe double bond of the second molecule :as-Diphenylethylene gives an unsaturated dimeride,CPhz:CH[H] + CPh2:CHZ -+ CPh,:CH*CPh,*CH3Addition of precisely this kind appears to account for the syn-theses of open-chain olefins during the pyrolysis of hydrocarbons(p.82) and for the production of Aa-butene and Aa-hexene fromethylene under the action of an electric discharge.41 The latterS.V. Lebedev, I. A. Andreevski, and A. A. Matinschkina, J . Rust?. Phy8.Chem. SOC., 1922,54,223; A., 1923, i, 770; H. Wieland and E. Dorrer, Ber.,1930,63, [ B ] , 404; A., 464.E. Bergmann and H. Weiss, Annalen, 1930, 480, 49, 59; A., 901, 902;C. S. Schoepfle and J. D. Ryan, J . Amer. Chem. SOC., 1930, 52, 4021; A.,1568.40 R. J. McCubbin and H. Adkins, ibid., p. 2547; A., 1017.‘1 G . Mignonac and R. V. de Saint-Aunay, Compt. rend., 1929, 189, 106;A., 1929, 103792 FARMER :result is particularIy noteworthy, since not only is 80-90% of theethylene converted into the butene-hexene mixture, but there isdefinite experimental indication that the polymerisation proceedsin the stagesCH2:CH2.z+ CH,:CH*CH,*CH, % CH2:CH*CH,*CH,*CH2*CH,.The mechanism underlying the polymerisation brought aboutwhen acetylene is submitted to an electric discharge is not quite soclear, but, since a mixture of dipropargyl, y-methyl- Ass-pentadi-inene, and diethenylacetylene is addition probablyproceeds in an analogous way--first giving ethinylethylene, whichunites with acetylene in more than one way.The process wouldseem to become somewhat modified when the polymerisation iseffected by heat.& Acetylene begins to polymerise at about 300°,polymerisation continuing as the only reaction of importance up to600". To a large extent benzene and other aromatic hydrocarbonsare produced. Analogously, methylacetylene might be expectedto yield mesitylene and other aromatic hydrocarbons : instead, itfirst suffers three-carbon isomerisation into allene, ultimatelyyielding (as does allene when taken initially) a mixture of dimeride,trimeride, and higher polymerides, but apparently no aromatichydrocarbons.For the polymerisations brought about by the action of electricdischarge and ultra-violet light 45 on saturated paraffins, however,quite different mechanisms are suggested.These involve notdegradation to olefins, succeeded by polymerisation, but polymeris-ation with hydrogen fission as indicated for the respective processesin (1) and (2) :The products, however, are not wholly saturated.The synthetic relationship of the monomeride, dimeride, trimeride,etc., of olefins such as isobutylene has long been a subject of interestbut has not been satisfactorily determined.It has now been shown,however, that, whereas isobutylene when heated in a glass tube a t43 G. Mignonac and R. V. de Saint-Aunay, Compt. rend., 1929,188, 959; A , ,1929, 637.49 R. N. Pease, J . Amer. Chem. Soc., 1929, 61, 3470; A., 1930, 58; R. N.Meinert and C. D. Hurd, ibid., 1930, 52, 4540; A., 1931, 61; P. SchlapferandM. Brunner, Helv. Chim. Acta, 1930,13, 1125; A., 1400.44 S. C. Lind and G. Glockler, J . Amer. Chem. Xoc., 1929, 51, 3655; A.,1930, 190.4 5 W. Kernula, Row. Chem., 1930,10, 273; A., 887ORGANIC CHEMISTRY .-PART I. 93200" yields only triisobutylene, and when treated with sulphuricacid yields the trimeride with a very small amount of dimeride, yetif polymerised by contact with Florida earth it yields polymeridesrecognisably ranging from di- to hepta-merides along with higherpolymerides which suffer depolymerisation on attempted distill-a t i ~ n .~ ~ Although prolongation of the period of contact of hydro-carbon and earth leads to an increase in the proportion of the morehighly polymerised forms, not all the individual polymerides seemto suffer further polymerisation. Indeed, the changes which takeplace appear to be the following :The polymerising action may be reversed by heating the productsa t 200" in contact with the ~atalyst.~' By studying the decom-position of the individual polymerides it is found that each one ofthe foregoing polymerisations may be reversed but in all cases somemonomeric isobutylene is produced, attributed in the cases of thetetra- and penta-merides to progressive dissociation of the initialdi- and tri-meric forms.Practical dficulties have prevented astudy of the decomposition of the higher polymerides, but it hasbeen noted that prolonged contact of the polymeric forms withfloridin at a high temperature causes the production of stable formsof high molecular weight. Diisobutylene is the most stable poly-meric variety and the stability rapidly decreases with increasingmolecular weight.HaZogeno-ok$ns.-Allyl chloride is converted in the absence ofair, slowly on keeping, but more rapidly under the influence ofultra-violet light, into polyallyl chloride, (C3.H5Cl)n, which is not asingle substance but a mixture of polymerides48 (mean value ofn = lo), formed by normal valency unions and probably possessingthe structure :CHz*CH(CH2Cl)*[CH,*CH(CHzCl)],*CH,~CH(CH2Cl) *The polymeride may be separated by fractional precipitation andextraction methods into a series of polymerides (n = 9,12,5,25,11,and 7) which decrease in solubility with increase in molecular weight.Vinyl bromide and as-dichloroethylene also polymerise in thelight to yield non-homogeneous substances, the components of whichrepresent different stages of polymerisation.49 Owing to inferior4 6 S.V. Lebedev and C. G. Kobliansky, Ber., 1930,63, [ B ] , 103; A., 316,4 7 Idem, ibid., p. 1432; A., 1017.48 H. Staudinger and T. Fleitmann, Anmlen, 1930, 480, 92; A,, 889.dU H. Staudinger, M.Brunner, and W. Feist, Hdv. Chirn. Actu, 1930, 13,805; A , , 1402; H. Staudinger and W. Feiet, ibid., p. 832; A., 140294 FARMER :solubility in cold and instability (leading to dehydrobromination)in hot solvents the molecular weight cannot be determined. Thepolyvinyl bromide obtained from the former suffers replacement ofthe halogen atoms by methyl groups when treated with zinc methyl(ethyl groups with zinc ethyl) to yield a hydrocarbon (C,H,), whichresembles cyclocaoutchouc and is less unsaturated than caoutchouc ;from indications afforded by reduction experiments it appears toconsist of a mixture of long straight-chain compounds. Theas-polydichloroethylene from the latter may be fractionated byextraction from benzene and reduced to hydrocarbons of widelydiffering molecular weight.It is completely saturated and to it isassigned the formula :* CH2*CC12*[CH2*CCI,],CH,*CC12 * * [X = 50-1001.Po@-esters.-The polymerism of the glycol esters of dibasic acidsis a polymerism of intermolecular condensation and not of theaddition of monomeric species. When in the interaction ofR(C02H), with R’(OH), the number of atoms in the system[*CO*R*CO*O*R’-O*] is too great to permit of five- or six-memberedring formation, the reaction becomes polymolecular and the poly-merides are usually linear in type. The ethylene, trimethylene,hexamethylene, and decamethylene esters of malonic, adipic, andsebacic acids have been obtained 50 as crystalline compounds and allare highly polymerised : the structural unit ranges from 7 to 22atoms and the molecular weight from 2300 (ethylene malonate) to5000 (trimethylene sebacate).Ethylene and trimethylene carbon-ates are obtainable, as would be expected, in the monomeric form,but the latter can be converted on heating with potassium carbonateinto a true polymeride, from which the monomeric ester is regener-ated on distillation in a vacuum.Neutral and acidic ethylene succinates showing different degreesof polymerisation have been prepared,51 and mixed polymerideshave been obtained by the interaction of ethylene glycol withequivalent quantities of succinic and sebacic acids, but not byfusing together (polymerised) ethylene succinate and ethylenesebacate. Both chemical and osmotic methods give concordantvalues for the molecular weights of these substances.Monomeric ethylene oxalate is obtainable but gradually poly-merises a t room temperature and rapidly on heating.By extractionof a partly polymerised ester with acetonitrile, soluble and insoluble6O W. H. Carothers and J. A. Arvin, J . Amer. Chem. SOC., 1929, 51, 2560;A., 1929, 1165; W. H. Carothers and F. J. van Natta, ibid., 1930, 52, 314;A , , 319.51 W. H. Carothers and G. L. Dorough, ibid., p. 711; A . , 452; W. H.Carothers, J. A. Arvin, and G. L. Dorough, ibid., p. 3292; A,, 1272ORGANIC CHEMISTRY.-PART I. 95polymerides have been extracted, either of which can arise spontane-ously from the other on keeping at the ordinary temperature.A similar process of pol yintermolecular reaction occurs whenc-hydroxydecoic acid is heated alone or with various inert solvents.52Mixtures of acids of the typeHO*[CH,],~CO,~([CH,]~*CO,.),.[CH,l,.CO,Hare produced, the components of which are partly separable byfractional crystallisation, and have molecular weights ranging fromabout 1000 to 9000.These complex acids yield the original acidon hydrolysis.Dirneric Acetoins.-An important example which diff ers from allthe polymeric types discussed above is that of acetoin.63 Observ-ations of the ultra-violet absorption spectrum of the monomericform of this substance indicate the presence of the carbonyl group andhence of the preponderating, if not exclusive, existence of theketo-form, COMe*CHMe*OH. There are two crystalline dimeridesof acetoin which give almost the same extinction curve in alcoholas the monomeride.It is clear that dissolution in alcohol does notcause fission of the dimerides, since they can be recovered crystallineafter some hours, and molecular-weight determinations of the iso-meride of higher melting point in the boiling solvent give normalvalues which decrease only when ebullition is prolonged. In the twocrystalline forms it would seem, therefore, that the two acetoinmolecules are united in such a manner that the keto-group, andconsequently the whole acetoin molecule, remains unchanged in itsmain valency arrangement.The possibility that groups other than carbonyl can give a similarextinction curve is negatived by observations with epichlorohydrinand triacetyl glucose anhydride.A dioxan structure is excluded,since dioxan itself is non-absorbent. Two ethereal oxygen atomsattached to the same carbon atom do not cause absorption, as isshown by the behaviour of methoxyacetaldehyde dimethylacetal.The dimeric methylacetal of acetoin does not exhibit the character-istic ketonic absorption, and hence is regarded as a dioxan derivative,CHMe*CMe( OMe)-O<CI~(~M~),CHM~>O, a view that is in harmony with itsreldtively di6cult polymerisation. From the fact that the dimericacetoins have considerably higher melting points when very rapidlyheated than when very slowly heated it appears probable that thetransformation temperature of the dimerides into the monomeridelies below the melting point.62 W.H. Lycan and R. Adam, J. Amer. Chem. SOC., 1929, 51, 3450; A.,63 W. Dirscherl and E. Braun, Ber., 1930, 63, [B], 416; A., 454.1930, 6596 FARMER :To account for the properties of dimeric acetoin a structure,[Me*hO*CHMe*OHI,, has been proposed in which the association ofmolecules is dependent on the exercise of definitely located auxiliaryvalency forces. On account of the divergence of views as to thenature of the associative links holding in such highly polymerisedsubstances as the polysaccharides, the further investigation of suchrelatively simple examples of residual valency polymerism as appear )from the evidence discussed) to be presented by the acetoins will beawaited with interest.Alcohols and Ketones.It was clearly desirable, for reasons discussed in last year’s Report,that all supposed esterifications in the p-position of glycerol shouldbe subjected to careful scrutiny.Further study of the constitutionof a number of supposed P-monoglycerides 64 supports the generalconclusion that previous announcements of the isolation of suchp-forms are incorrect. The instance of the @-benzoate of B. Helferichand H. Sieber 55 is, however, an exception and the claim of theseauthors to have isolated the first true p-glyceride is upheld. Like-wise, the assumption of a@- and ay-isomerism in diglycerides isshown by further cases to be based frequently on insufficient evidence.A list of apparently trustworthy methods for preparing a-, p-, a@-,ay-, and say-esters and ethers is given.It is pointed out that in some of the instances the cause of theproduction of unexpected isomerides is doubtless inherent in thesyntheses themselves : for instance, the production of ap- or ay-ethersfrom both ap- and cry-dihalogenohydrins is probably to be attributedto intermediate ap-oxide formation. This explanation may extendto similar preparations of the glycerol esters. Nevertheless, inter-changes of acid radicals with one another, and also with differentradicals, take place so easily that an explanation of these changesnot involving a large or difEcult movement of such heavy radicalsas the stearyl and palmityl groups is needed.If a tendency towards ortho-ester formation, as originally sug-gested by E.Fischer, be taken into consideration, the migration ofacid radicals can be represented merely by a rearrangement ofvalencies in which little or no relative movement of the atomsthemselves has to be presumed.ICH2*O*CO*C17H,5ay-Distearin ,%palmitate.CH2*O*C0 GI,€€, 5/3y-Dis tearin a-palmitate.64 A.Fairbourne, J., 1930, 369; A., 574.2. physiol. Chem., 1927, 170, 31; A., 1928, 44ORGANIC CHESI1STRY.-PART I. 97Moreover, if conditions can exist in which the Hantzsch type ofcarboxyl group actually occurs, the above two formula may becomeidentical ,C15H31*c[0 o * * * * ? H 2 * . * * . . . . YH . . . . E}c*c17H35CH,*O*CO*C, ,H3,and in any case, a tendency towards the transitory formation ofsuch @-oxide rings is held to be probable. On this basis themigrations which so frequently appear to take place while otherreactions are in progress (e.g., the migration of an acyl radicalsimultaneously with the elimination of a halogen atom, even if theglycerol molecule does not contain a hydroxyl group) can beexplained.The acid complexes formed between boric acid and glycols areshown by methods not involving their separation from solution tohave the composition HBD,, D representing a diol residue.56 Onesuch method is based on hydrogen-ion determinations in aqueoussolutions of boric acid and diol in various proportions, considerationof the equilibria involved giving a relation - ApR = (n/2) log a,connecting the difference of pa of two solutions and the ratio a of theamounts of diol present in each, n being the number of moleculesof diol combining with 1 mol.of boric acid. A second methodinvolves the consideration of the partition of hydroxyl ions betweenthe acid complex and free boric acid during neutralisation, and athird depends on cryoscopic measurements. The structure andmode of ionisation of these complexes are considered to be repre-sented by the general formulaA phenomenon which appears to involve the formation of ringcomplexes of somewhat similar character to the foregoing, and hasa direct bearing on the spatial distribution of the hydroxyl groupsof glycols, is revealed by the study of the solubility of arseniccompounds, particularly of arsinoacetic acid, in 99% acetic acid inthe presence of various It is found that two adjacenthydroxyl groups in the glycol generally increase the solubility ofarsinoacetic acid, arsenic trioxide, and resorcinolarsinic acid, andthis effect is intensified when three or four hydroxyl groups are ina position favourable to the formation of a five-membered ring, asin ethylene glycol, glycerol, pyrocatechol, etc.p-Glycols andm- and p-dihydric phenols have a smaller effect. Stereoisomerides56 J. Boeseken, N. Verrnaas, and A. T. Kuchlin, Rec. tmv. chim., 1930, 49,5 7 B. Englund, Svenek Kern. T'idskr., 1928, 40, 278; A., 1929, 52; J . pr.711; A . , 1018.Chem., 1929, [ii], 122, 121; 1930, [ii], 124, 191; A., 1929, 946; 1930, 330.REP.-VOL. XXVII. 98 FaRMER :with different spatial configuration of the hydroxyl groups havedifferent effects, e.g., active and meso-forms of tartaric acid, cis- andtruns-forms of cyclohexanediol.The magnitude of the effect isaltered by substitution; it is increased by the introduction of alkylgroups, but the extent of increase is dependent on the number andnature of the substituents. The influence of carboxyl or carbonylgroups, e.g., in succinic acid or benzil, on the solubility is small.Investigation of the rate of dehydration of the unsaturated glycolCMe,( OH)*CH:CH*CMe,*OH in water in the presence of hydrogenions 58 shows the strong effect produced by the spatial configuration.The dehydration of the cis-form of the diol in solutions of mineraland organic acids is catalysed by hydrogen ions. The reaction isunimolecular and complete, the oxide being the only product.Measurements of the velocity at different temperatures, employinghydrochloric acid, show the temperature coefficient to be high, whilstcomparative measurements made in different acids show that thevelocity of catalysis is' only approximately proportional to thehydrogen-ion concentration, increasing more rapidly than thelatter. So far as the experiments show, the velocity in weakly acidmedia is constant, the acidity being unaffected by addition of theditertiary glycol ; in more strongly acid solution, the velocitycoefficient decreases slightly but regularly as the concentration of theglycol increases.Dehydration of the trans-form, on the other hand,is slow, yielding not a cyclic oxide but open-chain polyolefins.The result of replacing the hydroxyl groups in the acetylenicglycol CMe,( OH)*CiC*CMe,*OH by bromine is of interest.Thechange has been effected 59 with phosphorus tribromide and, inaddition to a product which doubtless has the constitutionCMe,Br*CiC*CMe2Br, two other isomeric dibromides have beenobtained.CMe,:CBr*CBr:CMe,obtainable from the first on heating, or directly from the glycol bythe action of hydrobromic acid ; to the other dibromide, no formulahas been definitely assigned, although on the basis of its oxidationproducts it cyclic formula has been thought to apply. Neverthelessit seems to the writer that the oxidation products, while not lendingsupport to this suggestion, are precisely those to be expected froman allene derivative of the formula CMe,:C:CBr*CMe,Br, both thisand the isomeric diolefin being produced during, or subsequentlyto, replacement of hydroxyl, by anionotropic change.One of these appears to be the compound6 8 M.Bourguel and R. Rambaud, Bull. SOC. chim., 1930, [iv], 47, 173;69 V. N. Krestinski and L. J. Bashenova-Koslovskaia, J . Ru88. Phys. Chem.A., 574.SOC., 1929, 01, 1691; A., 1930, 574ORGAN10 aHEMISTRY.-PART I. 99Various useful methods have been devised for the production ofacetals and ketals from glycols, hydroxy-acids, and hydroxy-ketones. Derivatives of the enolic form of acetol have beenobtained by reacting on isopropylideneglycerol a-monochlorohydrinwith potassium hydroxide, and a dimeric derivative of the enolicform of methylglyoxal has been analogously obtained from thep - toluenesulp hon y 1 derivative of g 1 y ceraldeh y de met h ylc y clo -acetal; 6O the diethylacetal of hydroxyacetone is shown to beobtainable 61 if the acetol acetate is allowed to react with ethylorthoformate and the product hydrolysed with lime, although theinteraction of free acetol or acetol formate with the same reagentyields only acetol ethylcydoacetal ; ethylidene derivatives of glycolsand hydroxy-acids can be obtained by interaction of the latter withacetylene in the presence of a catalyst.G2 In the last reaction thecatalysts employed to effect condensation are solutions of boron orsilicon fluorides in alcohols with mercuric oxide.The activeconstituents of these solutions, which possess high electrical con-ductivity, are probably hydrofluoboric and hydrofluosilicic acids.The compound BF,,Et,O formed by dissolving boron trifluoride inether is also stated to be efficacious as a catalyst.An examination of the thermal decomposition of acetone in thegaseous state carried out several years ago showed that between506" and 632" the process constitutes a homogeneous unimolecularreaction, the products of decomposition being about one-halfsaturated hydrocarbons and hydrogen, one-third carbon monoxide,and the remainder carbon dioxide and ethylene.63 A new examin-ation 64 has shown that for every 100 mols.of acetone vapourpassed (with nitrogen) through a quartz tube heated in an electricfurnace, approximately 60 mols. of keten can be recovered. Hencethe primary reaction in the unimolecular decomposition is probablythe separation of a molecule 'of methane and the formation of keten,which undergoes a bimolecular decomposition into ethylene andcarbon monoxide at the high temperature.If this is the correctexplanation, the proportions of methane, ethylene, and carbonmonoxide which would then be found agree fairly closely with theprevious experimental result.[ B ] , 1732; A., 1164.ao H. 0. L. Fischer, E. Baer, L. Feldmann, and L. Ahlstrom, Ber., 1930,63,V.V.Evlampiev, J. Ru88.Phys. Chem. Soc., 1929,61,2017; A., 1930,580.62 J. A. Nieuwlend, R. R. Vogt, and W. L. Fookey, J . Amer. Chem. SOC.,C . N. Hinshelwood and W. K. Hutchiscn, Proc. Roy. SOC., 1926, [A],13* F. 0. Rice and R.E. Vollrath, Proc. Nat. Acad. Sci., 1929, 15, 702;1930,52, 1018; A., 745.111, 245; A., 1926, 691.A., 1029, 1425100 FARMER :The extent of enolisation produced in a number of ketones on theaddition of various organo-magnesium halides seems to bear norelation to the structure of the given ket0ne,~5 but, on the whole,increases with the atomic weight of the halide present in the Grignardreagent, and is greater where tertiary organic radicals are presentthan primary or secondary. Enolisation, using magnesium tert . -butyl chloride, amounts for di-n-butyl ketone and carvone to about20%, for acetophenone 31%, for cyclopentanone 326y0, for thujone41 yo, for 4-methylcyclohexanone 46% ; for cyclohexanone 50.5%, formenthone 51y0, and for mesityl oxide 60%.I n relation to thesevalues cyclohexanone contains originally 8.2 yo of enol, 4-methylcyclo-hexanone 6-3%, and mesityl oxide 6.3%, whilst the remainder arenormally exclusively in the keto-form. No connexion appears thusto exist between the tendency towards allelotropism of a given sub-stance and its degree of enolisation under the influence of theGrignard reagent.As a result of a determination of the extent of enolisation ofethyl acetoacetate and acetylacetone in alkaline solution, P. Gross-mann came to the conclusion that, with a considerable excess ofalkali, the high results obtained corresponded with the presence of adienol. In determining the amount of enol the usual practica ofacidifying the solution before treatment with bromine was replacedby direct addition of an acid bromine solution and removal of theexcess of bromine by aniline hydrochloride.The question has beenre-examined by A. Hantzsch and W. Krober,67 who state thattitration of alkaline solutions of ethyl acetoacetate and of acetyl-acetone with bromine solutions shows that the excess of brominetaken up by the enolic compound varies with the concentration ofthe bromine solution and is due to secondary reactions. Measure-ments of the amount of hydrogen liberated by the action of acetyl-acetone and of benzoylacetone on sodium or potassium suspendedin benzene or xylene show that only one atom of the alkali metal istaken up by the enol. The formula COMe-CH:C(ONa)*OEt forethyl sodioacetoacetate and the existence of dienol salts with thegroup *C( ONa):C:C( ONa)* proposed by Grossmann are consideredto be incorrect.Distillation of free ethyl acetoacetate a t the ordinary temperatureand mm.pressure in quartz or Pyrex, but not in soft, glassvessels, has yielded fractions containing 30---40~0 of the enol, thehalf-life period of which is about 500 hours.68 These are approxim-65 V. Grignard and H. Blanchon, Rocz. Chem., 1929,9, 547; A., 1930, 67.8 6 2. physikal. Chem., 1924, 109, 305; A., 1924, i, 834.6 7 lbid., 1930, 147, 293; A., 1021.6 8 F. 0. Rice and J. J. Sullivan, J . Arner. Chem. SOC., 1928,50, 3048ORGANIC CHEMISTRY.-PART I. 101ately ten times as stable as any previously obtained. No substancehas, however, been found capable of stabilising the product andattempts to increase the stability by removing traces of water withsilica gel or acetyl chloride proved unsuccessful.In this connexionit is of interest that dimethylpyruvic acid as obtained by distillationat atmospheric pressure is the pure keto-form; when this is dis-solved in water, an equilibrium between enolic and ketonic formsis obtained, a 1-OM-solution containing o.4770, a 0-1M-solution0.28y0, and a 0-O1M-solution only a trace of the enolic form.69After addition of a small amount of alkali a third form is reportedto be detectable spectrographically.Bromomalondialdehyde behaves in aqueous solution as a trueacid.70 In alcohol the keto-enol equilibrium, as determined byK. Meyer’s bromination method, is established only after 48 hoursand corresponds with 24% of the enol a t the ordinary temperature.A rise in temperature raises the proportion of the enol.The sodium salts of acetone, methyl propyl ketone, camphor, andfenchone have been prepared by the action on the ketones of sodiumin liquid ammonia; benzophenone yields with sodamide andpotassamide the compounds C13Hlo0,NaNH, and C,,Hl0O,NK3,respec tively.71The possibility of synthesising cis- and trans-forms of n-alkylidene-acetones has been demonstrated in the instance of butylidene-acetone.72 The two forms of this ap-unsaturated ketone resistconfigurational change and interconversion has been effected onlyin the direction of cis-+ trans, through the hydrobromide.Theself-additive tendency of these substances in the presence of alkalinereagents has not permitted examination of the relationship existingbetween geometrical configuration and @,py-isornerisation, but ithas been possible by the action of dilute acid to convert the corre-sponding py-ketone into a mixture of cis and trans @-forms.Optimum conditions for the condensation of ketones by hydrogenchloride are obtained when the molecular ratio of acid to ketone is2 : 3.73 It is a notable fact that with the exception of methyl ethylketone reaction in the presence of this condensing agent is confhedto the ketone group of one molecule and the methyl group of thesecond molecule.Methyl ethyl ketone suffers condensation (inconformity with former results for an acid condensing agent) at the69 C.Fromageot and S. Perraud, Biochem. Z., 1930, 223, 213; A., 1272.70 J. Grard, Compt. rend., 1930, 190, 187; A., 324.7 1 H. H. Strain, J . Amer. Chem. Soc., 1930,52, 3383; A., 1273.72 E. N. Eccott and R. P. Linstead, J., 1930,905.73 V. Grignard and J. Cologne, Compt. rend., 1930,190, 1349; A., 1022.Compare H. Stobbe andF. J. Wilson, Annalen, 1910,374, 237; J., 1910, 97, 1722102 FARMER :ethyl group of the second molecule, yielding y-chloro-ys-dimethyl-hexan- @-one. This gives with alkali the corresponding ketone,CMeEt:CMe*COMe. If hydrogen chloride be replaced by hydrogenbromide or hydrogen iodide in the same molecular ratio of acid toketone, higher yields of the condensation products are obtained ;moreover, the condensation of methyl ketones with secondarycarbon groups, but not of ketones with tertiary groups, is facilitated.Carbohydrates.The synthesis of 2 : 3 : 4 : 5 : 6-penta-acetyl glucose, a true open-chain aldehyde form of a sugar,74 has been followed by the synthesisof several analogous compounds.Penta-acetyl galactose 75 andtetra-acetyl Z-arabinose,76 derived from the diethylmercaptals ofthe appropriate sugars, have rotation values in chloroform of - 25"and - 65" respectively, showing that the carbonyl group can producea relatively high rotation in spite of the absence of rings. Aldehydo-pentabenzoyl glucose has been obtained in an analogous manner.77Other notable syntheses in the monosaccharide group are those ofd-glucoheptulose,78 obtained by the Lobry de Bruyn rearrangementof d-a-glucoheptose and clearly related to the glucoheptulose ofBertrand and Nitzberg 79 as the optical antipode, Z-threose 8O (insolution and in the form of its osazone), and epifucose (Z-talo-methylose) .81The elucidation of the constitution of the sugar acetones hasresulted in the extensive employment of the isopropylidene deriv-atives of sugars, sugar mercaptals, and sugar acids for syntheticpurposes.This is exemplified in the preparation of a, number ofgem-dialkyl derivatives of d-galactose 82 and of fructofuranoseand in the synthesis of the 4-methyl derivatives of d-mannose 84 andd-galactose.85 The dicarbonates of glucose, fructose, mannose, andarabinose, which have now been obtained as well-defined and easilycharacterised products,86 will doubtless prove of equal utility, and74 Ann. Reporte, 1929, 26, 92.7 5 M.L. Wolfrom, J. Amer. Chem. Soc., 1930, 62,2464; A., 1023.76 M. L. Wolfrom and M. R. Newlin, ibid., p. 3619; A., 1411.7 7 P. Brig1 and H. Muhlschlegel, Ber., 1930, 63, [B], 1551; A., 1022.78 W. C. Austin, J . Amer. Chem. SOC,, 1930, 52, 2106; A., 894.7O Compt. rend., 1928, 186, 925, 1172, 1773; A., 1928, 510, 620, 867.80 V. Deulofeu, J., 1929, 2458; A., 1930, 68.81 E. VotoEek and V. KuEerenko, J . Czech. Chem. Comm., 1930, 2, 47; A.,82 H. Ohle and C. Dambergis, Annalen, 1930, 481, 255; A,, 1274.83 H. Ohle and 0. Hecht, ibid., p. 233; A., 1274.84 E. Pacsu and C. von Kary, Ber., 1929, 62, [B], 2811; A., 1930, 70.85 E. Pacsu and A.Lob, a i d . , p. 3104; A., 1930, 197.86 W. N. Haworth and C. R. Porter, J., 1930, 151; A., 326.325ORGANIC CHEMISTRY .-PART I. 103for certain purposes, owing to their less ready hydrolysis by diluteacids to the parent sugars and their immediate attack by even coldalkali, will prove superior to the sugar-acetones. The constitutionalformulse suggested for these substances are based on the similarityin properties to the diacetones.Formulation of Sugars.-A considerable amount of new workrelating to the representation of the sugars as pyranose and furanosering-structures has been carried out. This includes the isolation ofmissing forms in one or other series, and the investigation of thecomparative stability and optical rotatory powers of representativesof both series.New lactones of Z-rhamnonic acid *' and 2 : 3 : 4-tri-methylrhamnonic acid 88 have been isolated, the former of whichprobably, and the latter more certainly, has the pyranose structureof a true a-lactone ; a-methylmannofuranoside, tetramethylmanno-f uranose ,s9 trimethyl -1yxof uranoside , and trime t hy I-ly xof uranosehave also been obtained and found to resemble closely other y-sugarderivatives.Although representatives of the two types of ring-structure showa large difference in the ease of oxidation by permanganate in thepresence of an acid phosphate b ~ f f e r , ~ l yet an attempt to estimatethe comparative stability of pyranose and furanose sugar derivativeshas met with only slight success; 92 observations relating to thedifferential absorption of ultra-violet light by the two types of ringstructure have disclosed only very slight spectrographic difference^.^^Various examples are known of the wandering of acyl groupslinked with polyhydric alcohol residues. One of the fist to observethe change was E.Fischer,9* who suggested that the migrationmechanism might be explained by the intermediate formation of anortho-carbonic ester group in the manner :$!H,*OAc TH,*OAc $!H,*OAcCH,*OH CH,*O>C<OH CH2*OAcThe adoption of Fischer's explanation has been shown to provide aE. L. Jackson and C. S. Hudson, J . Amer. Chem. SOC., 1930, 52, 1270;A., 744.a0 J. Avery and E. L. Hirst, J., 1929, 2466; A., 1930, 68; F. E. Wright,J . Amer. Chem. Soc., 1930,52, 1276; A., 744.8s W.N. Haworth, E. L. Hirst, and J. I. Webb, J., 1930, 651 ; A., 748.90 H. G. Bott, E. L. Hirst, and J. A. B. Smith, &id., p. 668; A., 747.QH-OAC -+ QH-0 CH, + QH*OHC. H. Whitnah and J. E. Milbery, J . Amer. Chem. SOC., 1930, 52, 1627 ;A , , 748.92 H. Ohle and V. Marecek, Ber., 1930, 63, [B], 612; A., 581.93 F. Goos, H. H. Schlubach, and G. A. Schroter, 2. phyaiol. Chem., 1930,94 Ber., 1920, 53, 1624; A., 1920, i, S08.186, 148; A., 455104 FdRMER :wayg5 not only of representing the simple migration of acetyl asillustrated in the conversion of 3- into 6-monoacetyl glucose deriv-atives 96 but also of relating 1 : 2 : 3 : 4-tetra-acetyl @-glucose (I) 97and 2 : 3 : 4 : 6-tetra-acetyl @-glucose (11) 98 t o the supposed1 : 2 : 3 : 6-tetra-acetyl glucose of B.Helferich and W. K l e i r ~ . ~ ~There is definite evidence for the view that the supposed 1 : 2 : 3 : 6-tetra-acetyl glucose which is formed from, and is known to exist inequilibrium with, (I) in alkaline solution represents the cyclic inter-mediate form (111) produced during the migration of an acetylgroup from the position C, to C, in the carbon chain of (I) when thelatter is alkylated with methyl iodide and silver oxide. The finalproduct so obtained is the methyl glucoside (IV) corresponding to,and derivable from, (11).OAc OAc (h ?HCH,*OH CH2-O-V-CH, bH,H 1-0 OAc H 1-0 H 1-0 OMe H 1-0 OH@L7>l -+ @=>I --+ i<OLFr>I f- i@->lAcO 1 I H AcO I 1 H AcO 1 1 H AcO 1 i HH OAc H OAc H OAc H OAc(1.1 (111.) (IV.) (11.)The migration of the 4- rather than the l-acetyl group, as wasoriginally suggested, is not finally excluded,l but would seem to beunlikely in view of the fact that a further transposition of acetylresidues must then occur when the resultant compound undergoesconversion into 2 : 3 : 4 : 6-tetra-acetyl p-methylglucoside (IV).A similar explanation has been applied to the formation of theanomalous third (" y ") forms of triacetyl methylrhamnoside, tetra-acetyl methylmannoside , and hepta-acetyl chlorornaltose.2 Theexistence of the first of these had been tentatively explained byassuming a novel form of stereoisomerism; now, however, theresistance to hydrolysis of one of the four acetyl groups in thiscompound-the feature distinguishing it from the two already9 5 W.N. Haworth, E. L. Hirst, and E. G. Teece, J., 1930, 1405; 8., 1022.96 K. Josephson, Ber., 1929, 62, [B], 317, 1913; A . , 1929, 428, 1278;Annalen, 1929, 472, 217; A., 1929, 1044; Svensk Kern. Tidskr., 1929, 41,99; A., 1929, 912.9 7 J. W. H. Oldham, J . , 1925,127, 2840; A., 1926, 151.98 E. Fischer and K. Delbruck, Ber., 1909, 42, 2778; A., 1909, i, 633.O9 Annalen, 1927, 455, 173; A , , 1927, 858.1 Compare B. Helferich, Ber., 1930, 63, [B], 2142; A . , 1411.2 H. G. Bott, W. N. Haworth, and E. L. Hirst, J . , 1930, 1395; A., 1024;K. Freudenberg, Natumuise., 1930, 18, 393; A., 894; K. Freudenberg andH. Scholz, Ber., 1930,63, [ B ] , 1969; A . , 1412.8 W. N. Haworth, E. L. Hirst, and E.J. Miller, J., 1929, 2469; A., 1930,68ORGANIC CHEMISTRY.-PART I. 105known triacetyl methylrhamnosides (ct- and p-forms)-is at*tributedto its participation in the complex Ig>C<E%e . According to thenew explanation, which receives some support from absorptionmeasurements,4 the third or obstructed varieties of the three corn-pounds are to be represented thus :QH3 MeO-C-0H OAc OAc H H,C*OACThird form of triacetyl Third form of tetra-acetyl Third form of hepta-aceby1me thylrhamnoside. methylmannoside. chlo romaltose.It is contended by C. S. Hudson that change of ring form may anddoes occur during the methylation of sugars, a circumstance held tovitiate certain of the conclusions of W. N. Haworth with respect tothe structural representation of the sugars.New formulz deducedby applying the principle of optical superposition to available opticaldata have accordingly been put forward for various mono- andpoly-saccliarides.5 Quite apart, however, from the coherence of thechemical evidence which supports the existing formulation, there isimportant new evidence that the principle of optical superpositiondoes not apply uniformly throughout the sugar group.6 I n thefructose series, for example, difficulties are encountered in applyingHudson’s metliods,7 since the differences in the activities of thecc-series in different solvents are unusually great, whereas smallerdifferences are observed in the p-series : hence the magnitude of theincrement depends greatly on the choice of solvent and it is con-cluded from the data obtained for a variety of solvents that thc:increments calculated from the values of the aldose series cannot beapplied in the ketose series, in which the values are markedly higher.The limitations to the usefulness of the optical method in diagnosingconstitution also appear from a study of derivatives of 5-p-toluene-sulphonyl-3 : 6-anhydroglucose 8 and from other observation^.^E.Braun, Naturwiss., 1930, 18, 393; A., 896; Bey., 1930, 63, [ B ] , 1072;J. Amer. Chem. SOC., 1930, 52, 1680, 1707; A., 747.W. N. Haworth, Nature, 1930, 126, 238; A., 1273; J . Amer. Chem. SOC.,H. H. Schlubach and G. A. Schroter, Ber., 1930, 63, [B], 364; A., 456.H. Ohle and E. Euler, ibid., p. 1796; A., 1165.F. Micheel, ibid., p.347; A., 455; H. H. Schlubach and R. Gilbert,A., 1411.1930, 52, 4169.ibid., p. 2292; A.. 1412.D 106 FARMER :From measurements of the rotation of Q- and p-glucose andct-methylglucoside in borax solutions 10 it appears that the glucosidicand 2-carbon hydroxyl groups are essential for the reaction withborate, and occupy in a-glucose cis-positions. A means of determin-ing directly the configuration of ct- and p-forms in the sugar serieshas been sought, and consistent results have been obtained in con-necting the reactivity of a number of glucosyl halides towardstrimethylamine with the cis- or trans-configuration of the substituentgroups at the 1- and 2-carbon atoms of the chain.ll Only a cis-relationship of these groups would seem to permit of quaternarysalt formation with the amine, but the method is not generallyapplicable in this simple form, since the interaction of acetobromo-I-rhamnose with trimethylamine is found to lead to the productionof a diacetyl anhydrorhamnose. On the assumption that Waldeninversion does not occur at the C,-atom during ring closure theanhydride ring is regarded as produced by the removal of brominealong with the 4-acetyl group, a procedure which may be renderedpossible by the approach in space of the two groups in a folded phaseof the p p a n ring (formula A).It is suggested that the correspond-ing stability of acetobromomannose towards trimethylamine maybe due to the greater distancegroup as indicated in formula(A) cH3of the bromine atom from the 4-acetylB.Further progress has been made in the investigation of twonaturally occurring sugars, digitoxose, and hamamelose (hama-melihexose).To the former, which had been shown by H. Kiliani 12to be a 2 : 6-deoxyhexosej a formula has been assigned : l3 for thelatter, obtained as methylhamameloside from the hamameli-tannindescribed by Freudenberg and Bliimmel,14 a branched-chain formuladifferent from that originally suggested by the latter authors hasbeen advanced. l510 M. Levy, J . Biol. Chem., 1929,84, 749, 763; A., 1930, 69.11 F. Micheel and H. Michoel, Ber., 1930, 63, [B], 386; A., 455.12 Ber., 1922, 55, [B], 75; A., 1922, i, 224.13 F. Micheel, Ber., 1930, 63, [B], 347; A., 455.14 Annalen, 1924, 440, 45; A., 1925, i, 51.l5 0.T. Schmidt, Annalen, 1929,476, 250; A., 1930, 197ORGANIC CHEMISTRY.-PART I. 107$?H,*OHQH*OH 1HO*Q--$?H*OMeQHOTH2 H*$?*OHH&H CH,*OHDigitoxose. Methylhamameloside.H* *OH YH- 0MeThe probability that a genetic relationship explains the existencetogether in natural products of sugars and of pyran or pyrone nuclei16is increased by a demonstration of the convertibility of acetobromo-glucose and acet obromogalac tose into tetra-acetyl 2 - hydroxyglucaland tetra-acetyl 2-hydroxygalactal respectively, and thencesmoothly into diacetylkojic acid.17Synthetic GZucosides.-The nature of the reagents employed toeffect the conversion of glucosyl halides into alkyl glucosides wouldappear to determine, to some extent at least, the a- or the @-form ofthe product.It is possible to obtain either an a- or a p-methyl-glucoside from both 2-trichloroacetyl-3 : 4 : 6-triacetyl-(3-glucosylchloride and 3 : 4 : 6-triacetyl-~-glucosyl chloride by selectingappropriate conditions. This may be due to each reagent havingits own isomerising effect, but it is considered to be more probablethat the first stage in these processes consists in the addition of thereagent to the glucosyl halide and that the subsequent behaviourof the additive compound so obtained is determined by the nature ofthe reacting systems.ls Direct conversion of p- into a-glucosideshas been found to result from the action of titanium tetrachloridein chloroform,19 and a-alkylglucosides admixed only with minorproportions of the corresponding 13-forms have been obtained byheating the dibenzylmercaptals of the aldomonoses with mercuricchloride dissolved in the requisite alcohol.20It has now been established that dihydroxyanthraquinonediglucosides, including such as contain the sugar residues attachedto the same benzene nucleus, can exist.21 The preparation of analizarin diglucoside should therefore be possible.a-Hydroxylgroups, when protected by (3-hydroxyl groups, cannot be caused toreact with halogeno-sugars by using an excess of the latter, by16 W. N. Haworth, " Constitution of Sugars," p. 38.1 7 K. Maurer, Ber., 1930, 63, [BJ, 25; A., 326; K. Maurer and A. Muller,18 W. J. Hickinbottom, J., 1930, 1338; A., 1023.19 E. Pacsu, J . Amer. Chem.SOC., 1930, 52, 2563; A., 1023.20 E. Pacsu and N. Ticharich, Bey., 1929, 62, [B], 3008; A., 1930, 197.21 A. Mi.iller, ibid., p. 2793; A., 1930, 71; A. Robertson, J., 1930, 1136;ibid., p. 2069; A., 1412.A., 895108 FARMER :prolonging the reaction, or by raising the temperature. It appearspossible, therefore, that ruberythric acid is a monobioside, butowing to the demonstrated difference of its octa-acetyl derivativefrom that of an octa-acetyl maltoside derived synthetically fromalizarin it is apparently not a maltoside. Rubiadin glucoside hasnow been shown definitely by synthesis to be 3-p-glucosidoxy-1 - hy drox y -2-methylanthraquinone .22Di- and Tri-saccharides.In supplementing the evidence furnished by methylation methodsas to the correct formulation of disaccharides, use has been made ofthe 1 : 2- and 5 : 6-unsaturated derivatives of sugars (glucals andglucoseens). Thus new evidence in favou? of the C,-linking of thecomponents in melibiose is provided by the behaviour of melibialtaken in conjunction with the capacity of the sugar to form both1 : 4- and 1 : 5-glucosides and lac tone^.^^ New evidence as to themanner of coupling and the pyranose ring structure of trehalose islikewise afforded by the formation of a bis-triphenylmethane deriv-ative of this sugar and of a non-reducing trehalosediene which ishydrolysable to isorhamn~nose.~~Gentiobial, a new compound of the glucal class, has been shownto yield a mixture of hydroglucal and glucose when submitted to theaction of emulsin after reduction.25 It is considered, therefore, tobe a 6-gl~cosidoglucal, the glucal structure containing a pyran ring.Analogously, the already known cellobial and lactal are consideredto be 4-glucosidoglucose and 4-galactosidoglucal respectively.Glucal itself passes rapidly on oxidation with perbenzoic acid tomannose,26 which fact corresponds with an earlier observation ofM.Bergmann and his collaborators that lactal yields 4-galactosido-mannose with the same reagent.27 It has been found, however, that3-methyl glucal yields 3-methyl glucose with perbenzoic acid 28and that the 4-galactosidomannose from lactal is a mixture of atleast two sugars.29A new disaccharide ketose has been derived from a-lactose by the22 E. T. Jones and A.Robertson, J . , 1930, 1699; A., 1167.23 P. A. Levene and E. Jorpes, J . Biol. Chem., 1930, 86, 403; A , . 749.24 H. Bredereck, Ber., 1930, 63, [ B ] , 959; A., 748.2 5 M. Bergmann and W. Freudenberg, ibid., 1929, 62, [ B ] , 2783; A.,28 C. Tanaka, Bull. Chem. SOC. Japan, 1930, 5, 214; A., 1273.2 7 Annalen, 1923, 434, 79; A., 1924, i, 265.2 8 P. A. Levene and A. L. Raymond, J . Biol. Chem., 1930, 88, 513; A.,Z g A. J. Watters and C. S. Hudson, J . Amer. Chem. SOC., 1930, 52, 3472;1930, 70.1411.A., 1275ORGANIC CHEMISTRY.-PART I. 109Lobry de Bruyn method. This substance, d-lactulose,3° whichyields d-fructose and d-galactose on hydrolysis, is considered to be4- p-d-galactosido-a-d-fructose.In connexion with the synthesis of sucrose, recent condensationsof tetra-acetyl y-fructose with tetra-acetyl glucose have yieldedocta-acetyl isosucrose as the only crystallisable product (occasionallyocta-acetyl isotrehalose is isolated) .31 As isosucrose is less stablethan sucrose, it is improbable that octa-acetyl sucrose is first formed.The isolation of two varieties of octa-acetyl sucrose which have thesame optical rotation in chloroform but different melting points hasbeen These do not correspond with the " A " and " B "varieties of sucrose but are considered to be stable and labile forms.The investigation of several natural products has proceeded astage further.From the calcium aldobionate obtained by hydrolysisof gum arabic, a crystalline aldobionic acid has been isolated whichyields a dicarboxybionic acid on oxidation.The fission productsof both acids confirm the view that the aldobionic acid is aglycuronogalactose and the slow rate of hydrolysis of the two methyl-glucosides derived from the aldobionic acid shows that both of thesepossess the pyranose structure.% The aldobionic acid is thereforeregarded as glycurono-3(or 6)-~-galactose.The third sugar constituent of scammonin has been recognised asrhodeose, which remains together with rhamnose when the hydrolysisproduct of scammonin is fermented to remove glucose.34 Thesugar of ct-crocin when acetylated yields a product which is found tobe identical with octa-acetyl genti~biose.~~Acid hydrolysis of the glucomannan obtainable from " konjakpowder " (powdered tubers of Amorphophallus Eonjack, C.Koch)confirms the view that this polysaccharide contains mannose andglucose in the ratio of 2 : l.36 Glucomannan gives a triacetate andyields by acetolysis a mixture from which glucomannotriose, gluco-mannobiose, and mannobiose may be obtained by deacetylation.On hydrolysis the triose yields two molecules of mannose and onemolecule of glucose ; similarly, glucomannobiose yields equal30 E. M. Montgomery and C. S. Hudson, J . Amer. Chem. SOC., 1930, 5.2,31 ( S i r ) J. C. Irvine and J. W. H. Oldham, ibid., 1929, 51,3609; A., 1930,3% A. Pictet, Helv. Chim. Acta, 1930, 13, 698; A., 1166.33 M. Heidelberger and F. $1. Kendall, J . Biol. Chem., 1929, 84, 639; A.,34 E. VotoEek and F. Valentin, J . Czech. Chem.Comm., 1929, 1, 606; A.,35 P. Karrer and K. Miki, Helv. Chirn. Acta, 1929,12, 985; A., 1929, 1427.36 K. Nishida and H. Hashima, J . Dept. Agric. Kyushu, 1930, 2, 277; A . ,2101 ; A., 894.197.1930, 66.1930, 71.1413110 FBRMER :quantities of mannose and glucose, but mannobiose yields mannosePolysaccharides.Uncertainty as to the trustworthiness of the cryoscopic methodfor the determination of the true molecular weights of the poly-saccharides and their derivatives enhances the interest attached tothe investigation of the relationship between viscosity and molecularcondition. This relationship is the subject of a series of papers byH. Staudinger and his ~ollaborators,~7 the theoretical considerationsdeveloped being based on the following hypotheses.(1) Variationin the specific viscosity, qm., of a colloidally dissolved substance withpressure is an indication of structure in solution which may becaused by the presence of macromolecules in such concentrationthat they are mutually impeditive. (2) The specific viscosity of themacromolecular hydrocarbons increases proportionately with theconcentration in dilute solution but more rapidly than the concen-tration in concentrated sol solution and in the region of gel solution.(3) Approximate constancy in the specific viscosity of a solution overa considerable range of temperatures is evidence of the presence ofmacromolecules in solution and against that of strongly solvatedmicelles. (4) Measurements of viscosity in different solvents affordevidence of the influence of the medium on the macromolecules.Several polysaccharides and their derivatives have been examinedexperimentally and the deductions from the viscosity of theirsolutions are referred to in the sections below.Inulin and 1,ichenin.-It has been reported that the molecularweight of inulin in liquid ammonia and in molten acetamide is inagreement with the formula (C6Hl,O5), and that the isolation of thepolymerised product from the latter mixture yields an inulan ofcomposition ( C6H1005)2 which is freely soluble in water but becomesas insoluble in cold water as inulin on keeping.Observations of asimilar kind have been recorded with respect to the depolymerisingaction on inulin of glycer01,~8 acetamide 39 and f~rmamide,~O but thereality of the alleged depolymerisation is denied by E.Berner.*lThis author suggests that specimens of inulin and other polysacch-37 H. Staudinger and R. Nodzu, Ber., 1930, 63, [ B ] , 721; A., 571; H.Staudinger and W. Heuer, ibid., p. 222; A., 333; H. Staudinger, K. Frey,R. Signer, W. Starek, and G. Widmer, ibid., p. 2308; A . , 1415; H. Staudingerand 0. Schweitzer, ibid., p. 2317; A , , 1414.38 H. Vogel, Ber., 1929, 62, [ B ] , 2980; A., 1930, 198; H. H. Schlubachand H. Elsner, ibid., 1930, 63, [B], 362; A., 456.39 J. Reilly and P. P. Donovan, Sci. Proc. Roy. Dublin SOC., 1930, 19, 409;a,, 896.4O H. Pringsheim and W. G. Hensel, Ber., 1930, 63, [B], 1096; A., 896.4 1 E. Berner, ibid., pp. 1356, 2760; A., 1025.onlyORGANIC CHEMISTRY.-PART I.111arides of apparently low molecular weight (determined cryoscopic-ally) may be obtained by treatment with glycerol or ethylene glycolat the ordinary temperature, but attributes the observed effect to aphysico-chemical process caused by the adsorption by the poly-saccharide of appreciable quantities of the " depolymerising " andprecipitating media, whereby water solubility is conferred. Theadsorbed materials may only be removed by protracted heating in avacuum.The correctness of this conclusion is contested by H. H. Schlu-bach 42 and H. Pringsheim and their collaborators and it is pointedout by the former that inulin, when treated wth benzamide, yieldsa product which is separable into two portions of different specificrotation; also the values recorded by different workers for thespecific rotation of depolymerised products are such as to permitthe products to be divided into two groups, (1) those which have thesame specific rotations as the initial inulin, into which they can betransformed, and (2) those which differ in specific rotation from theoriginal material and are not capable of spontaneous reconversion.Viscosity measurements 44 of inulin dissolved in formamide showthat its behaviour differs greittly from that of polystyrene or balatasolutions.Since qSp. is much greater at lower than at higher tem-peratures, a change in the structure of the colloidal particles is heldto occur, which is attributed to alteration of the co-ordinate linkingsof the molecules among themselves and with the solvent; thesechanges are reversible, since particles of the original size result whenthe solutions are cooled.Examination of solutions of lichenin in formamide shows thesolute to be present in the molecular form a t great dilution.In themore concentrated solutions co-ordinative union is considered toexist between solvent and solute and between the molecules ofsolute; the latter linkings are resolved when the temperature israised and for this reason a more pronounced fall in viscosity isobserved than in more dilute solutions, in which only the co-ordin-ative linkings between solute and solvent are ruptured. Themolecular weight of lichenin is markedly higher than that ofinulin; the chain appears to contain about 300 unit molecules.Starch.-The experimental conditions necessary for effecting asharp separation of the two main constituents of potato starch havebeen investigated.45 The a-amylose from the outer envelope of the42 H.H. Schlubach and H. Elmer, Ber., p. 2302; A., 1415.43 H. Pringsheim, J. Reilly, W. G. Hensel, W. Burmeister, P. P. Donovan,44 H. Staudinger and 0. Schweitzer, ibid., p. 2317 ; A., 1414.4 6 M. E. Baldwin, J . Amer. Chem. SOC., 1930, 52, 2907; A . , 1167.and N. Hayes, ibid., p. 2636112 FARMER :grains forms 84 & 1% of the original starch; the p-amylose fromthe interior of the grains constitutes within 1% the remainder of thestarch. The properties of these preparations are described and it issuggested that the great difference in the temperature coefficientsof the rotations of the two components probably accounts for thedivergent values previously recorded.Products having the characteristics of Pringsheim's amylobioseand amylotriose have been obtained from amylose and amylopectin,respectively, by the action of cold concentrated hydrochloric acid 46as described by this worker and his collaborators.Evidence isfurnished to show that they are not definite compounds and theirproperties are quite inconsistent with the formulae assigned to them.47Products of similar character are stated to be obtainable by theaction of cold hydrochloric acid on glycogen and on glucose.Viscosity measurements carried out on solutions of potato starchor soluble starch in formamide reveal a behaviour differing com-pletely from that of polystyrenes and caoutchouc and markedlyfrom that of lichenin.Since even dilute solutions do not obey theHagen-Poiseuille law, deductions concerning molecular weight areinadmissible. Freely mobile colloid molecules are absent. Thespecific viscosity diminishes with rise in temperature, the starchsuffering change. Since the qSp./c values in quite dilute solutions ofdiffering concentration are not constant, irregular changes in thestructure appear to occur ; it cannot therefore be determined whethermacromolecules are present in solution under definite conditions.It is thought probable that the colloidal particles of starch are com-posed of associations of extremely unstable, readily cracked mole-cules, and that the solution contains molecular aggregates (micelles)and not macromolecules.Cellulose .-All the isolable products of the degradation of celluloseacetates reduce alkaline iodine solution and the iodine consumptionis stated to be an accurate and reproducible measure of the freealdehydic groups and the molecular weight of the compounds.48Technical cellulose acetate has been further degraded with hydrogenbromide in ghcial acetic acid, and the product separated by hotmethyl alcohol into two portions, acetylsaccharides A and B.Theiodine numbers of the respective materials indicate the presence inthem of mixtures of polysaccharides containing as an average 8 and9-11 hexose units. The product obtained by the acetolysis of46 R.Weidenhagen and A. Wolf, 8. Ver. deut. Zucker-Ind., 1930, 80, 265;'' Compare Ber., 1926, 59, [B], 991, 096, 1001; A . , 1926, 715.48 M. Bergmann and H. Machemer, ibid., 1930, 63, [B], 316; A., 457.Compare R. Willstatter and G. Schudel, ibid., 1918, 51, 7 8 0 ; A., 1918, ii, 461.A., 1168ORGANIC CHEMISTRY .-PART I. 113cellulose according to the method of K. Hess and H. Friese49 haebeen separated into a number of fractions which appear to contain8-13 hexose residues.To the latter substance, the hexa-acetyl biosan considered byHess and Friese to be a true dimeride of the fundamental cellulosemolecule, a molecular weight of 556587 has recently beenassigned 50 : this value, determined in acetic acid, is considered to bea true one, although the behaviour of different samples of acetone-soluble cellulose acetate in acetic acid is irregular and the relationshipof cellulose acetate to glacial acetic acid apparently complex.Thevalue is rejected by other workers 51 on the grounds of the unsuit-ability of the cryoscopic method for determining the (very large)molecular weight of this siibstance and of the demonstrable non-homogeneity of the biosan acetate.With regard to the more complex acetolysis products of cellulose,the terminal groups form so small a proportion of the molecule thattheir accurate determination by the iodine method is held to beimpossible; 52 nevertheless, on the basis of iodine consumption, thechain in cellulose itself has been computed to contain 50 glucoseresidues on an average as against 25-30 such residues in the starchmolecule.53The course of degradation of the cellulose molecule on heatingwith acetic acid and zinc chloride has been followed by viscositymeasurement^.^^ Degradation of the initial cellulose moleculeoccurs very rapidly, but the simpler products are much moreresistant. The initial, highly viscous solutions are computed tocontain cellulose triacetates of average degree of polymerisation60-100; after 20 hours’ action, the degree is reduced to 10. Atthe ordinary temperature, reaction proceeds extremely slowly,giving products of a degree of polymerisation 150. After 7 days at30°, the degree is 130 and after 10 days it is 100. Since nativecellulose is more highly polymerised than the most complex of itsacetates, its molecular weight is considered to be above 24,000, andthe cellulose molecule probably contains a chain of 500--1000glucose residues, thus having dimensions similar to those of thecaoutchouc molecule.The simpler cellulose acetates dissolve without swelling to give** Annalen, 1926, 450, 40; A., 1927, 44.6o K.Hem Ber., 1930, 63, [ R ] , 518; A., 456.51 K. Freudenberg, E. Bruch, and H. Rau, ibid., 1929, 62, [ B ] , 3078; A.,1930, 198; K. Freudenberg and E. Bruch, ibid., 1930, 63, [ B ] , 535; As, 457:K. H. Meyer and H. Hopff, ibid., 1930, 63, [ B ] , 790; A . , 750.62 H. Staudinger and H. Freudenberger, ibid., p. 2331 ; A . , 1416.s8 K. Freudenberg, W. Kuhn, W. Diirr, F. Bolz, and G. Steinbrunn, ibid.,p.1510; A., 1025114 BENNETT AND CHAPMBN:mobile solutions, the viscosity of which increases proportionally tothe concentration ; slightly degraded cellulose acetates swell verymarkedly, their viscosity in solution increasing far more rapidy thanthe concentration, so that very viscous solutions can be obtainedat relatively high dilution. I n them, the total region of the macro-molecules is greater than the available volume in relatively dilutesolution. I n place of sols in which the macromolecules have freemovement, these viscous solutions correspond to gels in which themacromolecules impede one another. A 10-triacetylcelloglucandiacetate, OAc[C,H,O,Ac,],,Ac, is stated to behave in solution likea simple substance and the gel condition is not reached until theconcentration attains 2.2% ; with a 150-triacetylcelloglucan diacetatethe transition between sol and gel lies at 1.4%.According to further indications afforded by viscosity measure-ments 44 cellulose triacetate in s-tetrachloroethane and m-cresolappears to form associations of molecules in concentrated solutionwhich become resolved when the temperature is raised.I n dilutesolution the change of the colloidal particles is uniform and is heldto be attributable to a loosening of the co-ordinative linkingsbetween the macromolecules and those of the solvent. Since withcellulose acetate, lichenin, a'nd inulin, as with the highly complexhydrocarbons, the specific viscosity of the solutions is the same atdifferent temperatures, the identity of the colloidal particles with tliemacromolecules is regarded as established.Dibenzylcelluloseexhibits abnormal behaviour in s-tetrachloroethane, whilst methyl-cellulose (OMe = 38%) shows still more marked divergence fromthe Hagen-Poiseuille law, in the same solvent. With cellulosenitrate, the measurements of viscosity are particularly difficult tointerpret, since mechanical treatment of the solution causes greatermobility.E. H. FARMER.PART II.-HOMOCYCLIC DIVISION.The Pinacol-Pinacolin and Related Rearrangements(compare Ann. Reports, 1923,20,115 ; 1925,22,116 ; 1928,25,134).DURING the last few years a considerable volume of work hasappeared dealing with the pinacol-pinacolin change and the relatedrearrangements of the substituted ethylene oxides and of thOR(XBN1C CHEMISTRY.-PART II.115a~-amino-alcohols.~-55~ *Oy61 Many data have been collected andrepeated attempts have been made to enunciate general rulesP. J. Montagne and S. A. Koopal, Rec. trav. chim., 1910, 29, 136, 150 ;A., 1910, i, 323, 324.a S . A. Koopal, ibid., 1915, 34, 115; A., 1915, i, 693.S. Danilov, Ber., 1927, 60, [B], 2390; A., 1928, 64.S . Danilov and E. Venus-Danilova, Ber., 1926, 59, [B], 377; A., 1926,519.ti Idem, ibid., p: 1032; A., 1926, 726.Idem, Ber., 1927, 60, [B], 1050; A,, 1927, 661.E. Venus-Danilova, Ber., 1928, 61, [B], 1954; A., 1928, 1244.Idem, J. Rum. Phtye. Clbem. SOC., 1929, 61, 53; A., 1929, 1070.M. Godchot and G. Cauquil, Compt. rend., 1928,186, 767 ; A., 1928,521.M.Gomberg and J. C. Bailar, jun., J. Arner. Chem. SOC., 1929,51, 2229;A., 1929, 1067.l1 R. Lagrave, Ann. Chim., 1927, [ix], 8, 363; A., 1928, 270.12 L. Leers, Bull. SOC. chim., 1926, [iv], 39, 651 ; A., 1926, 711.l3 (Mlle.) J. LBvy, ibid., 1921, [iv], 29, 820; A., 1921, i, 788.l5 (Mlle.) J. L6vy and F. Gombinska, Compt. rend., 1929,188,711 ; A., 1929,l6 (Mlle.) J. LBvy and R. Lagrave, Bull. SOC. chim., 1927, [iv], 41, 833; A.,l7 (Mlle.) J . LBvy and M. Sfiras, Compt. rend., 1927, 184, 1335; A., 1927,l8 Idem, ibid., 1928, 187, 45; A., 1928, 888.lo (Mlle.) J. LBvy and A. Tabart, ibid., 1929,188, 402; A., 1929, 448.*O (Mlle.) J. LBvy and P. Weill, ibid., 1927, 185, 135; A., 1927, 880.21 R. Locquin and L. Leers, Bull. SOC. chim., 1926, [iv], 39, 655; A., 1926,22 P.Nicolle, ibid., p. 55; A., 1926, 382.23 M. Migita, Bull. Chem. SOC. Japan, 1928, 3, 308; 1929, 4, 57; A., 1929,24 W. Madelung and M. Oberwegner, Ber., 1927 60 [B], 2469; A., 1928,25 L. Orthner, Annalen, 1927, 459, 217; A., 1928, 184.26 A. McKenzie and W. S. Dennler, J., 1926, 1596; A., 1926, 834.27 Idem, Ber., 1927, 60, [ B ] , 220; A., 1927, 243.28 A. McKenzie and M. S. Lesslie, Ber., 1929, 62, [B], 288; A., 1929, 317.2D A. McKenzie and A. K. Mills, ibid., p. 284; A., 1929, 317.3o Idem, ibid., p. 1784; A., 1929, 1066.31 A. McKenzie, A. K. Mills, and J. R. Myles, Ber., 1930, 63, [B], 90432 A. McKenzie and R. Roger, J., 1927, 671 ; A., 1927, 457.33 Idem, Ber., 1929, 62, [B], 272; A., 1929, 317.34 A. McKenzie, R.Roger, and G. 0. Wills, J., 1926, 779; A., 1926, 610.35 M. Tiffeneau and (Mlle.) J. Uvy, Bull. SOC. chirn., 1923, [iv], 33, 769;Idem, ibid., 1926, [iv], 39, 67; A., 1926, 399.555.1927, 872.662.711.448, 675.171.788.A., 1923, i, 789.Idem, ibid., p. 735; A., 1923, i, 788.37 Idem, Convpt. rend., 1926,182, 391 ; A., 1926, 383.8 8 Idem, Bull. SOC. chim., 1926, [iv], 39, 763; A., 1920, 818116 BENNETT AND CHAPMAN :regarding the changes. The interpretation of the data is, however,not it simple matter and the complications attending it will thereforebe considered in some detail, as sufficient allowance has not alwaysbeen made for them in the discussion of this subject.Although essentially similar in character, the rearrangements canbe conveniently classified into three main groups : 36 (a) the pina-colinic change proper,CRR1( OH)*C( OH)R2R3 I H $ CRR1R2*COR3(b) the semihydrobenzoinic change in which the tertiary hydroxylCRR1R2*CH0CHRRWOR2group is lost, and (c) the semipinacolinic change in which thesecondary hydroxyl group is lost,R*CO*CHR1R2CRR1(OH)*C(OH)HR2 --\ - - % O f orCRR1(OH)*C(OH)HR2 -- --$of or ' R ~ ~ O ~ C H R R ~The pinacolinic change proper presents fewest complications ininterpretation.The mode of elimination of the elements of waterfrom the two hydroxyl groups of the pinacol molecule determineswhich of the two central carbon atoms shall become the ketonicgroup and thus fixes the general direction of the rearrangement.39 M. Tiffeneau and (Mlle.) J.Ldvy, Compt. rend., 1926, 183, 969; A., 1927,40 Idem, ibid., p. 1112; A., 1927, 153.41 Idem, Compt. rend., 1927, 184, 1465; A., 1927, 769.42 Idem, Bull. SOC. chim., 1927, [iv], 41, 1351; A., 1927, 1184.43 Idem, Compt. rend., 1928, 186, 84; A., 1928, 286.44 M. Tiffeneau and A. Ordkhoff, ibid., 1920,171, 400; A., 1920, i, 672.4 5 Idern, ibid., p. 473; A., 1920, i, 673; Bull. SOC. chim., 1921, [iv], 29,429, 445; A., 1921, i, 585, 566.46 Idem, Compt. rend., 1921, 172, 387; A., 1921, i, 243; Bull. SOC. chim.,1921, [iv], 29, 809; A., 1921, i, 788.4' A. OrBkhoff and M. Tiffeneau, Bull. SOC. chim., 1922, [iv], 31, 253; A.,1922, i, 458.48 M. Tiffeneau and A. Orekhoff, ibid., 1923, [iv], 33, 195; A., 1923, i, 333.49 A. Or6khoff and M.Tiffeneau, Compt. rend., 1924,178, 1619; A., 1924,146.i, 729.Idem, Bull. SOC. chim., 1925, [iv], 37, 1410; A., 1926, 172.51 Idem, Compt. rend., 1926, 182, 67; A., 1926, 171.s2 Idem, Bull. SOC. chim., 1927, [iv], 41, 1174; A., 1027, 1076.63 S . Yamaguchi, Bull. Chem. SOC. Japan, 1926,1, 64; A., 1928, 727.64 Ber., 1929, 62, [B], 1598; A., 1929, 929.55 E. Luce, Compt. rend., 1925, 180, 145; A., 1925, i, 263ORGANIC CHEMISTRY.-PART 11. 117In compounds of the type CRR(OH)*C(OH)RIR1 the nature of thefinal product is determined by this elimination factor alone and astudy of the rearrangements of such compounds completely con-firms the view 56 that the elimination of the hydroxyl group occurspreferentially from the carbon atom to which are attached thegroups with the greatest capacity for electron release.57Further support for this view is forthcoming from the data withregard to the less symmetrical pinacols.e.g.> 79 l 2 9 143 159 1 6 9 229 383 50When all the radicals attached to the central carbon atoms arealiphatic the products of the change are usually mixtures involvingelimination in both directions.From the data available it is also possible to arrange certainpairs of groups in the order in which they facilitate the fission ofthe adjacent hydroxyl l49 1 9 9 2% 339 3 7 9 46When the substituted ethylene oxides are converted into theisomeric ketonic substances by heating, the fission of the C-0 bondfollows the same 113 1 5 5 1 8 s 37s 38 a few apparent exceptions,which occur under drastic conditions,l' being possibly explicable bya further isomerisation of the primary product.On the other hand the direction of rearrangement of the ap-amino-alcohols appears to be determined by the position of the amino-group eliminated, the oxygen of the hydroxyl group remaining inplace irrespective of the character of the adjacent groups.28323 3 4 9 399 65Thus, for example, p-amino-act-diphenyl-n-propyl alcohol (I) isconverted by nitrous acid into methyldeoxybenzoin (II),= whilst(1.1 CPh,(OH)*CHMe*NH, -+ Ph*CO*CHPhMe (11.)in the corresponding pinacol, loss of water occurs by removal of thehydroxyl group from the a-position. 68 Pinacols have, however,been obtained by deamination of some a@-amino-alcohols,32 and thepossibility of their formation as intermediate products in other casesmust not be overlooked.66 C.K. Ingold, Ann. Reports, 1928, 25, 134.67 For examples, see list included in Ref. 2 (p. 115).68 R. Stoermer, E. Schenck zu Schweinsberg, (Fr.) Sibbern-Sibbers, endP. Riebel, Ber., 1906, 39, 2288; A., 1906, i, 681118 BENNETT AND CHAPMAN :Although the general direction of a pinacolinic change is deter-mined by the mode of elimination of the elements of water or offission of the oxide linkage in an ethylene oxide, the nature of theproduct obtained depends also on the relative facility with whichthe two groups concerned will migrate from one carbon atom tothe other. For example :R\ R- C*COR2 (R1 migrates)p-K::] / RJ‘A R\R- C*COR1 ( R2 migrates) ?R dThe differences in “ migratory aptitude ’’ of various groups are bestobserved in the rearrangement of compounds of type (111) in whichR(111.) Rl>V-K;lOH OHthe two hydroxyl groups are identically situated and the eliminationfactor is therefore not involved.Investigation of the rearrangementof such compounds 1 9 2910, 609 has shown that the following groupscan be arranged in order of decreasing “migratory aptitude”thus :p-anisybp-tolyl, p-diphenyl,a-nsphthyl >p-isopropylphenyl>p -ethylphenyl>p - fluorophenyl, } phenylp-iodophenyl f >p-bromophenyl, p-chlorophenyl>o- or m-chloro- or -bromo-phenyl\ >m-tolyl>m-anisylThe ease with which these different groups migrate within themolecule is evidently not related simply to any one of their pro-perties and more data will be required before the factors governing“ migratory aptitude ” in the pinacol-pinacolin system can be fullyunderstood, especially as it has not yet been established that thisis a property of the group independent of its environment.Numerous attempts have been made to deduce the relative“ migratory aptitudes ” of various groups from observation of therearrangement of compounds other than type (111), but the inter-pretation of the results obtained is complicated by the two furtherconsiderations discussed below and it is inadvisable to draw fromthese results any definite conclusions.When one or more-of the four radicals attached to the two centralcarbon atoms are replaced by hydrogen atoms it is sometimespossible for a given product to be formed in more than one way.For example ORGANIC CHEMISTRY.-PART II.119The transformation of (IV) into (V) by the action of dilute or con-centrated sulphuric acid Z6 may occur either as a semihydrobenzoinicchange with migration of a hydrogen atom or by a vinyl dehydration,followed by ketonisation of the resultant enol. It is not permissible,therefore, to draw any conclusion from this rearrangement as to therelative ease of migration of hydrogen and the cr-naphthyl group,and similar caution must be exercised in the interpretation of thenumerous other cases of apparent semihydrobenzoinic change whichYet another complication attends the interpretation of therearrangement of ap-glycols having hydrogen atoms attached to thea- or p-carbon atom.It has been shown 3-89 2% 5l9 59 that aldehydesof the type CR3*CH0 in the presence of cold concentrated sulphuricacid, or even sometimes on boiling with the dilute acid, may undergorearrangement into isomeric ketones :involve the same dilemma.6, 14, 141% 22,44349, 50, 52CR,*CHO -+ CHR,*COR.The reagents that are known to bring about the conversion of thealdehydes into the ketones are amongst those most commonlyemployed to effect the pinacol-pinacolin change, and it is thereforeimpossible as a rule to decide whether a given ketone, obtainedapparently by a direct semipinacolinic rearrangement, is really theprimary product.R-COCHRRlSemipinacolinic change/:>?-OH OHSemihydrobenzoinic changeIt is clear, however, that the aldehyde is not necessarily an inter-s* s.Danilov, J . Russ. PhY8. Chern. soc., 1919, 61, 97120 BENNETT AND CHAPMAN :mediate product, as R. Roger and A. McKenzie 33 have shown thatthe conversion of d- p-hydroxy- p-phenyl- aa-dibenzylethanol (VI)into ay-diphenyl-y-benzylacetone (VII) by boiling dilute sulphuricacid does not involve complete loss of optical activity as would occurif the ketone were formed by way of the aldehyde (VIII) (comparealso A. McKenzie and W. S. Dennler 27 and J. L6vy and P. Weill ,O).,CH,Ph CH2Ph\ CH,PhH-C-CO*CH,Ph OHC*dCH,PhPh/ \Ph:>?%CH,Ph OH OHWI.) (VII.) (VIII.)This rearrangement of trisubstituted acetaldehydes affords anexplanation of the change in product which often accompanies achange of reagent.4, 6, 133 367 4 0 9 4* For example :PhH > VOH -KE OHPhO C H O e eI MeAldehydes are usually formed under the milder conditions andketones with the more drastic agents.The deduction of any series of " migratory aptitudes " from theresults of apparent semipinacolinic changes 43 is evidently unsafeunless the mechanism of the rearrangement can be independentlydemonstrated in each case.have described a rearrangementof the pinacol-pinacolin type involving as-migration.H. Kleinfeller and F.EckertPh Ph PhPh*CO--#-C33-(i-COPh I -- boiling + Ph*CO*(!iLC32-CO*COPhalcoholic HCl Ph OH OHOther Molecular Rearrangements.A new triad isomeric change has been discovered by A. Schonbergand L.von Vargha.62 The diarylthioncarbonates (I) were con-verted into the corresponding thiolcarbonates (11) when heated a tabout 300". The rearrangement has been employed for the prepar-A. Orkkhoff and J. Brouty, Bull. SOC. chim., 1930, [iv], 47, 621 ; A., 1179.61 J. C. Bailar, jun., J . Amer. Chern. SOC., 1930,52, 3596; A., 1438.62 Ber., 1930, 63, [B], 178; A., 320ORGANIC CHEMISTRY.-PART II. 121ation from the corresponding phenols of some disulphides otherwiseaccessible with diEculty. 63The reversible rearrangement of the triarylbenzenylamidinesreferred to in last year’s Report has now been investigated in greaterdetail a and has been shown to be intramolecular by tests similarto those applied to the imino-ether rearrangement. The relativemobilities of different migrating groups Ar (I11 and IV) were inthe order p-tolyl<phenyl<p-chlorophenyl, whilst replacementof p-tolyl as the stationary group Arl by p-chlorophenyl resultedin a retardation of interconversion.These results are exactlyparallel with those obtained for the rearrangement of imino-arylethers into arnidesG6 and it would appear that the changes areidentical in mechanism.Consideration of the positions of equilibrium of several pairs ofisomeric amidines has shown that the original electronic formulationof these rearrangements is untenable, but all the known facts are inagreement with the view 65 that the migrating group Ar first attractsone or both of the lone electrons of the nitrogen atom to which it isgoing and then, retaining its newly acquired electrons, releases theoriginal binding pair.This view has the advantage, not only offurnishing a picture of these rearrangements, but also of revealinga close connexion between them and a number of other changes, bothintra- and inter-molecular, which at first sight appear quite dis-similar. For instance, an intermolecular reaction of this kind is thedisplacement of chlorine by iodine in various alkyl chlorides whenthese are heated with an alkali iodide.67 The replace.ment may beregarded as an intermolecular migration of the radical from chlorineto iodine and there is strong evidence that the exchange takes placeby a preliminary combination of the alkyl chloride molecule and theiodide ion, followed by ejection of a chlorion.In the same classalso come reactions like the fission of diary1 sthers by piperidine,686* A. Schanberg, L. von Vargha, and W. Paul, Annulen, 1930, 483, 107.A. W. Chapman, J., 1930, 2458; A. W. Chapman and C. H. Perrott,ibid., p. 2462.61 Ann. Reportcr, 1929, 26, 123.A. W. Chapman, J., 1927, 1743; A., 1927, 874.Compare Ann. Reports, 1929, 26, 137. 6 8 Ibid., p. 133122 BENNETT AND CHAPMAN :an intermolecular migration of an aryl group from oxygen to nitrogen.Conversely the imino-ether and amidine changes may be regardedas intramolecular displacement reactions.The establishment of the constitution (V) for iso- @-naphtholsulphide 69 reveals the isomeric change of this compound under theinfluence of heat or alkalis 70 as the migration of a hydroxynaphthylgroup from oxygen to sulphur.This appears to be an intramolecular change comparable both withthe thion-thiol (and imino-ether) rearrangement and with thereplacement reactions already discussed.The migration of thep-hydroxynaphthyl radical from oxygen to sulphur is compensatedby the reverse migration of the proton. The ortho-quinonoiddehydrosulphide (VII) may be regarded as an oxidised form of thehypothetical intermediate in which both the sulphur and the oxygenatom are momentarily attached to a single nuclear carbon atom.A more detailed study of the rearrangement of phenacylbenzyl-dimethylammonium salts mentioned in the last Report has nowappeared.71 E’or the phenacyl group has been substituted p-bromo-+[Ph*CO-CH2*NMe2*CH2Ph]6H -+ Ph*CO*CH( CH,Ph)*NMe, + H,Ophenacyl and acetonyl, for the benzyl group various substitutedbenzyls, a-phenylethyl, benzhydryl, and 9-fluorenyl, and for thedimethylammonium radical piperidinium without alteration in thecharacter of the rearrangement. The change followed a unimole-cular course in alcoholic sodium ethoxide solution and no exchangeof radicals could be detected between the molecules of two differentcompounds undergoing rearrangement in a mixed solution. Therelative migration velocities of different substituted benzyl groupswere a8 follows :Substitutent ......p-M80 €I m-Br p-Br P-NO,Velocity .. ...... .... 0.77 1.0 1.6 2.52 about 30 -69 L. A. Warren and S. Smiles, J., 1930, 956; A . , 908.7O R.Henriques, Ber., 1894, 27, 3000; A., 1895, i, 103.7 1 T. S. Stevens, J., 1930, 2107; A . , 1437; T. S . Stevens, W. W. Snedden,E. T. Stiller, and T. Thornson, J., 1930, 2119; A., 1686; compare Ann,Reports, 1929, 26, 124ORCANIU CHEMISTRY.-PART II. 123The author's view of the course of the reaction may be representedthus :Ph*CO*CH,*$Me, EEe Ph*CO*8H*l!!Me2 I ---, I I -+CH2Ph (VIII.) CH2PhPh*CO*CH=NMe,0Ph-CO- H*NMe2 'i and -+-CH,Ph CH,Phthe free benzyl anion being captured again before it can escape intothe bulk of the reaction mixture. The facts could be equally wellexplained along the lines adopted for the cases discussed above.The rearrangement resembles the conversion by heat of isocyanidesinto nitriles :R-N=C- -+ -NEC-RA remarkable case of the migration of a group from a side chaininto the nucleus of an aromatic compound has been recorded.'2Met h y 1 and e t h y 1 1 : 4 - naph t haquinonepheny lhydrazone-N-carb -oxylates (IX), in the presence of cold alcoholic barium hydroxide,yielded, besides the products of hydrolysis, methyl and ethyl8-benzeneazo-5-hydroxy-l-naphthoates (X), a conversion thatinvolves the migration of the group C02R from the @-nitrogen atomin the side chain to the peri-position in the naphthalene nucleus.Phenyl ally1 aulphide is partly converted into o-allylthiophenolat temperatures above 200°,73 a rearrangement similar to that ofthe corresponding oxygen compound.74The difficulties 75 that frequently beset attempts to determinewhether a given rearrangement, dependent upon the presence of acatalyst, is or is not intramolecular are again exemplified in a studyof the rearrangement of the alkylanilines in the presence of metallic72 R.Willstatter, E. Ulbrich, L. PogBny, and C. Maimeri, Annalen, 1929,78 C . D. Hurd and H. Greengarcl, J. Amer. Chem. Soc., 1930, 62, 3356; A.,74 L. Claisen, Ber., 1912, 45, 3157; A,, 1912, i, 966.7 5 Ann. Reports, 1929, 26, 124.477, 161; A., 1930, 214.1285124 BENNETT AND CHAPMAN :salts.76 The authors conclude that the change is intramolecularand does not occur by successive fission and recombination of an--3 R -2alkyl halide, but the experimental results can be reconciled with theopposite view. The interesting observation was made that whereasisobutylaniline, heated with metallic halides, yielded p-aminoiso-butylbenzene, its hydrobromide, heated alone or in the presence of ametallic bromide, underwent rearrangement to p-amino-tert.-butyl-benzene.This was explained as due to a difference in the mechanismof rearrangement in the two cases, but a simpler explanation is thatthe hydrobromide evolved rapidly a large proportion of isobutylbromide, which isomerised to the tertiary halide in the vapour phasebefore recombination, whilst the liberation of halide from the mix-ture of base and salt was so much slower that nuclear substitutionoccurred before any considerable fraction of the halide had beenchanged.The stereochemical view of the isomerism of the chlorohydroxy-benzoyltoluic acids 77 has now been abandoned.78 The inter-convertible pairs are regarded as compounds of the type of (XI) and(XIII), and the rearrangement as proceeding through an inter-c1 c1 c1mediate compound (XII) of quinonoid structure. These conclusionsare supported by the facts that only one compound could be obtainedwhen phthalic anhydride was substituted for methylphthalicanhydride in the preparation, the two positions for the carboxylAgroup in (XIV) being equivalent, and that whilst two isomericcompounds, presumably (XV) and (XVI), were obtained by con-7 6 W.J. Hickinbottom and (in part) A. C. Waine, J . , 1930, 1558; W. J.7 7 M, Hayashi, J . , 1927, 2516; A., 1927, 1187; Ann. Reports, 1929, 26,7 8 Idem, J., 1930, 1613, 1520, 1524; A., 1183.Hickinbottom and G.H. Preston, ibid., p. 1566; A., 1174.141ORGANIC CHEMISTRY .-PART 11. 125densation of 4-methylphthalic anhydride with benzene, they werenot interconvertible, the essential substituent for the productionof quinonoid structure being absent.Free Radicals(compare Ann. Reports, 1924, 21, 115; 1928, 25, 150).An important event of the last two years has been the demonstra-tion of the real though transitory existence of the free methyl andethyl radicals.79 When the vapour of lead tetramethyl was passedat low pressure (1-5-2 mm.) through a heated tube, decompositionoccurred and lead was deposited. Passage of the gaseous productsof decomposition over a second lead mirror further along the tubecaused the disappearance of this mirror with the production of acompound that could be condensed, volatilised, and decomposedby passage through a hot tube togiveanother deposit of lead.Theactive material was also capable of attacking metallic zinc andantimony, producing in the former case a substance recognised aszinc methyl. The observed activity fell off rapidly as the distancealong the tube increased and disappeared when the material wascooled by means of liquid air. It was therefore concluded that thefree methyl radical had been formed by the decomposition of thelead alkyl, and the difliculty of its isolation was accounted for bythe very short, though measurable, life period, estimated at 0.006sec. for half disappearance under the conditions of the experiment.The preparation of the ethyl radical has since been announced.Itscompound with zinc has been isolated and found to be convertibleinto ethyl alcohol and to be free from lead.Whilst the free methyl and ethyl radicals represent an extremedivergence both in type and in stability from the well-knowntriarylmethyls, a number of tervalent carbon compounds have beendetected which differ to some extent from this class. Such arediphenyl-tert . - butylmet hyl80 and diphenylphenylethinylmethyl 81obtained on warming the corresponding ethanes and detected bythe reversible change of colour with change of temperature. Thelatter is a companion case to that of tetraphenylallyLa2A detailed study of another class of tervalent carbon compound79 F. Paneth and W.Hofeditz, Ber., 1929, 62, [B], 1335; A., 1929, 788;F. Paneth and W. Lautsch, Nature, 1930, 125, 564; A., 735. CompareG. Schultze and E. Muller, 2. physikal. Chem., 1930, [B], 6, 267; A., 302;F. Paneth, ibid., 7, 155; A , , 721.J. B. Conant and N. M. Bigelow, J . Amer. ClLem. SOC., 1928, 50, 2041;A . , 1928, 994.81 H. Wieland and H. Kloss, AnnuZen, 1929, 470, 201 ; A . , 1929, 1053.8a K. Ziegler, ibid., 1923, 434, 34; A., 1924, i, 308; Ann. Reports, 1928, 25,164126 BENNETT AND CHAPMAN :has been carried out by A. Lowenbein and his collaborators 83 on thecyano-, acyl, and lactone derivatives of tetraphenylethane (I to I11respectively).[CPh,(COPh)-1, (OGP) (1.1 (11.) \ 0 1 2[CPh,(~)-J,(111.)These compounds were all colourless and undissociated in the solidstate or in cold solution, but became deeply coloured on fusion orwhen heated in solvents.The dissociation constants were measuredby determination of the apparent molecular weights a t variousdilutions in boiling toluene or ethylene dibromide and the followingresults were obtained :Dissociation Dissociationconstant, const ant,in in in intoluene. C,H,Br,. toluene. C,H4Br,.0.0088 0.037 (jf:0*00066 0.0043Q&C 0 0dissociationcommenceda t 140"0.0011 - ph27- co CN0(Me0.C,H4)2. j! - dissociationcommencedCN at 60-80"Reassociation with corresponding loss of colour occurred when thesolutions were cooled. Similar dissociation has been qualitativelyobserved in the case of benzpinacol dibenzoate in molten naphthal-ene.84 The 2 : 3 : 4-trisubstituted dichromenyls 85 also dissociate8.9 A.Ltswenbein and H. Simonis (with H. Lang and W. Jacobus), Ber., 1924,57, [B], 2040; A., 1926, i, 147; A. Lawenbein (with W. Folberth), Ber., 1926,58, [B], 601; A., 1926, i, 662; A. Lowenbein and R. F. Gagarin, ibid., p.2646; A., 1926, 168; A. LSwenbein and H. Schmidt, Ber., 1927, 60, [B],1861; A., 1927, 1072; A. Lijwenbein and L. Schuster, Annalen, 1930, 481,106; A , , 1184.** F. F. Blicke, J. Amer. Chern. SOC., 1926, 47, 1477; A., 1925, i, 811.86 A. Lowenbein and B. Roaenbaum, Annalen, 1926, 448, 223; A., 1926,955ORGANIC CHEMISTRY .-PART 11. 127on solution or fusion, but this case is complicated by the possibilityof the free valency occurring on either the 2- or the 4-carbon atom.I R \(R = Ph, a-naphthyl, benzyl.)The possibility of free radicals being formed as intermediateproducts in the course of chemical reactions, whilst always admissible,is only occasionally supported by the experimental evidence.Investigations by Wieland and his collaborators 86 of the thermaldecomposition of various azo-compounds, however, illustrate thispossibility and show that some less well-known types of free radicalsare capable of at least momentary existence.When azo-compoundsof the type ArN:N*CPh, were heated in inert solvents, nitrogen wasrapidly eliminated and trip henylmethyl was recognised among theproducts. The aryl radical underwent reduction to the correspond-ing hydrocarbon. Certain aryl compounds R*CO-N:N*CPh, alsogave off nitrogen when heated and formed deeply coloured solutions.The colour faded even in t'he absence of air and the ketoneR*CO*CPh, was obtained from the resulting solution.In thepresence of oxygen, however, triphenylmethyl peroxide was formedtogether, in some cases, with the corresponding acyl peroxide.Evidently the decomposition resulted in the temporary formationof the two free radicals.R*CO*N,*CPh, + RCO- + --CPh, + N2 -+ R*CO*CPh, - (RCOO-), + (CPh,*O),Determinations of the velocity of dissociation of hexaphenyl-ethane 87 and of various dialkylxanthyls 88 have been made, iodineand oxygen respectively being used to combine with the free radicalsas formed. The conclusion of M. Gomberg and F. W. Sullivan 89that the colour and degree of dissociation of free radicals in solutionare not simply related has been criticised for theoretical reasons andon the ground of insufficient accuracy of the experimental evidence.g086 H.Wieland, E. Popper, and H. Seefried, Ber., 1922, 55, [B], 1816; A.,1922, i, 772 ; H. Wieland, H. vom Hove, and K. Borner, Annalen, 1926, 446,31; A., 1926, 61; H. Wieland, A. Hintermaier, and I. Dennstedt (with J.Lorenzo), Annalen, 1927, 452, 1 ; A., 1927, 237.K. Ziegler, L. Eweld, and P. Orth, Annalen, 1930, 479, 277; A., 711.8 8 J. B. Conant and M. W. Evans, J . Arner. Chern. SOC., 1929,51,1926; A.,1929, 934.Ibid., 1922, 44, 1810; A., 1922, i, 929.C. B. Wooster, ibid., 1929, 51, 1163; A., 1929, 648128 BENNETT AND CHAPMAN :A further set of optical measurements 91 on solutions of hexaphenyl-ethane, tetraphenyldi- -naphthylethane, and bis-2 : 3 : 4-triphenyl-6-methylchromenyl in various solvents has shown that, if theratio E,/E, of the extinction coefficients at dilution v and a t infinitedilution be taken as a measure of the degree of dissociation of thediradical, Ostwald’s dilution law is obeyed.The heat of dissociationof hexaphenylethane calculated from the temperature coefficientof dissociation was found to be about 11.5 kg.-cals. per mol. and tobe practically independent of the solvent in which dissociation tookplace. The kinetics of the oxidation of hexaphenylethane by freeoxygen have also been in~estigated.~2In the Annual Report for 1928 it was suggested 93 that the degreeof dissociation of substituted hexa-arylethanes was determinedlargely by the effects of substitution on the stability of the anionAr,C and that these effects were not necessarily the converse ofthose of similar substitution in the kationic radical.The dissoci-ation constants of a number of para-substituted triphenylmethylchlorides have recently been determinedg4 and follow the orderC6H4*OMe,C6H4*NO,,Ph > Ph, > C6H4*OMe, (C6H4*N02)2 >C6H4*N0,,Ph2> (C,H,-NO,),,Ph, which should be, and is, that ofkationic stability. The dissociation of the corresponding hexa-phenylethanes could not be studied quantitatively owing to theinstability of the compounds, but qualitative observations sufficedto show that the order of radical dissociation was not parallel withthe determined order of kationic stability.It is not, however, soeasy to reconcile the comparative stability of the undissociateddiacyltetraphenylethanes 83 with this suggestion, as the presenceof the acyl group should tend to stabilise the anionic radical and sopromote dissociation.On the other hand the stability of free hydrazyl radicals, as waspointed out in the same Report,93 is not determined by the stabilityof the anionic form, but is favoured by the simultaneous presence ofgroups capable of attracting and repelling electrons. A series of1 : 1 : 4 : 4-tetra-aryl-2 : 3-dibenzoyltetrazanes studied by S. Gold-Schmidt and J. Bader 95 illustrates still more clearly the validity ofthis generalisation.91 K.Zieglsr and L. Ewald, Annalen, 1929, 473, 163; A., 1929, 1010.92 R. C. Mithoff and G. E. K. Branch, J. Amer. Chent. SOC., 1930, 52, 255;Ann. Reports, 1928, 25, 154; compare also H. Burton and C. K. Ingold,A., 301.Proc. Leeds Phil. SOC., 1929, 1, 421 ; A., 1929, 1052.a4 K. Ziegler and W. Mathes, Annalen, 1930, 479, 111 ; A., 762.@j Ibid., 1929, 473, 137; A., 1929, 1173ORGANIC CHEMISTRY .-PART 11. 129In these compounds a constant electron-restraining influence ofconsiderable magnitude is provided by the benzoyl group on theP-nitrogen atom. The dissociation should therefore be favouredin proportion as the groups R and R1 have electron-releasing capaci-ties of either inductive or tautomeric kind. This is borne out bythe order of the dissociation constants of these compounds, whichwas found to be (for R and R1), di-p-nitrophenyl<phenyl, p-nitro-phenyl< di-p- bromophenyl < phenyl, p - bromophenyl <diphenyl <p henyl, p - tolyl < di-p- tolyl < phenyl, p - anisyl < di-p - anisyl.The lastcompound was completely dissociated in acetone solution.isocyanides and Related Compounds.An important addition to the knowledge of this class of compoundshas been made by D. L. Hammick, R. C. A. New, N. V. Sidgwick, andL. E. Sutton,96 who have obtained conclusive evidence in favour ofthe formula (I) for the isocyanides in place of the bivalent carbonstructure (11) by measurement of the parachors and dipole momentaof a number of isonitriles.(I). R-NZC R-N=C (11).The calculated parachor of the group -NfC is 62.3, whilst themost probable value for -N=C is 40.5.Examination of p-tolyl,p-anisyl, and ethyl isocyanides gave values of 66, 66, and 69respectively for the -NC group.The dipole moment of the group -NC was found from measure-ments with p-tolyl and p-chlorophenyl isocyanides to be 3.6 (E.S.U.x 10-l8), the carbon being tho negative pole. The Nef formula (11)is untenable, since it leads to a value of 2 in the opposite direction,but the experimental result is in good agreement with a computedvalue for -NSC. The structure thus established is in agreementwith the chemical properties of the isonitriles and reveals theidentity of the ions (111) derived from both tautomeric forms ofhydrogen cyanide. Similar structures can be ascribed to carbon(111.) [-N--C-]- [-c-o-] (IV.)monoxide (IV)97 and presumably to compounds such as carbonmonoxide diethyla~etal.~~9 6 J., 1930, 1876; A., 1239.g7 S.Sugden, " The Parachor and Valency," 1930, 171.g * H. Scheibler, Ber., 1926, 59, [ B ] , 1022; 1927, 60, [B], 554; A., 1926,711; 1927, 338; H. Scheibler and E. Baumann, Ber., 1929, 62, [ B ] , 2057;A., 1929, 1296.REP .-VOL . XXVII . 130 BENNETT AND CHAPMAN :Aromtic Substitution(continued from Ann. Reports, 1928, 25, 137).The two years now under review have seen the publication ofmuch systematic work on the substitution of aromatic compounds,particularly in the diphenyl and other polynuclear series. Severalother investigations are worthy of special note, including thosedealing with the nitration of phenylboric acid ; the directive influenceof the nitroso-group; and the manner in which stereochemicalfactors may control the direction of substitution.General Theory.-A new orientation rule has been proposed whichis at the same time simple and based on the fundamental classi-fication of the elements.99 It states that, if in a benzene derivativePh-X-Y, Y is in a higher group of the periodic table than X, or if,being in the same group, Y is of lower atomic weight than X, thegroup XY is meta-directive, whilst in other cases [including thosein which (a) X = Y and ( b ) Y is absent] the direction is of theortho-para type.It must also be borne in mind that ionic chargeson XY of positive or negative sign will cause meta- or ortho-para-direction respectively.Nitrosobenzene seemed to be the only exception to this rule, asbromination of the substance takes place in the para-position incarbon disulphide at - 5O.1 However, experiment shows that thenitroso-compound is considerably associated under these conditions,and that in acetic acid, in which it is not associated and with whichit does not combine, reaction is very slow and no para-substitutioncan be detected.That the nitroso-group belongs to the meta-directive category is, moreover, clear from the observation that o-and p-bromonitrosobenzenes react with silver nitrate in glacialacetic acid to yield silver bromide, whereas the m-isomeride isunaffected.A calculation of the net charge on the substituent atom of variousgroups, assuming the electrons to be shared in the ratio of thenuclear charges of the atoms they unite,2 places the various substi-tuents approximately in the order of their directing powers.Thereare, however, some marked divergences, such as the position of thefree poles NH, and NR,, which should be found at the extreme ofm-directing groups.The angle between the plane of a double bond and the adjacentsingle bonds is according to the tetrahedral theory of the order 125".9s D. L. Hammick and W. S. Illingworth, J . , 1930, 2358; A., 1666.@ c3C. K. Ingold, J., 1925, 513; A., 1925, i, 646.W. M. Latimer and C. W. Porter, J . Amer. Chem. SOC., 1930, 52, 206; A.,331ORGANIC CHEMISTRY .-PART II. 131The external valencies of a benzene nucleus of the Kekul6 formulashould therefore not be directed as from the centre of the hexagon,but should diverge at both sides of each double bond so that theexternal angle p exceeds the adjacent angle cc by an amount whichowing to the closure of the ring w i l l be even more than in open-chainolefins (I).3 Consequently it may be concluded that of the tKopossible forms of hydrindene (11) and (111), (111) should be the lessstrained and should constitute the major constituent of the equili-brium mixture of forms.In confirmation of this it is found that 5-hydroxyhydrindene issubstituted by bromine and by diazonium salts in position 6 ratherthan 4 : for the analogy between phenols and enols in such reactionsmakes it reasonable to expect that substitution occurs at the carbonatom joined to the C*OH by means of a double bond, so that theattack on the two possible forms of 5-hydroxyhydrindene (IV and3 W.H. Mills and I. G. Nixon, J., 1930, 2610132 BENNETT AND CHAPMAN :V) would be as indicated by the arrows : structure (V) is thereforeJ. .L CH,~0//4\/2, (+(,/ H O P 0 +v Ho?y):2 \A(vr.) CH, (IV.) (V.1indicated for the hydroxyhydrindene and (111) for the parenthydrocarbon. In the same way 5-acetamidohydrindene is bromin-ated in position 6.5On the other hand, ar-tetrahydro-@-naphthol is known to beattacked as shown in (VI) and the corresponding acetamidotetra-hydronaphthalene reacts in the same ( a - ) position.6 Of the twopossible structures, formula (VI) consequently represents thearrangement of bonds in the tetrahydronaphthalene nucleus.It isshown by a careful consideration of both the angles and the inter-atomic distances (different for the aromatic and the alicyclic atoms)that the saturated part of the system would most probably attain agreater freedom from strain by departure from the single plane inthis structure (VI) than would be the case for the alternative formin which the common bond between the two rings is a single one.These arguments reveal important factors which must not be lostsight of in the orientation problem, and they also constitute a remark-able tribute to the value of KekulB's formula for benzene.The importance of nitrous acid in certain nitrations has recentlybeen emphasised.' As a result of a detailed study of the nitrosationand nitration of phenols, S.Veibe18 considers that the nitrophenolsresult from the oxidation by nitric acid of intermediate addition-compounds of phenol and nitrous acid, which are also the precursorsof o- and p-nitrosophenols. Reaction velocity data and the o / pratios in nitration and nitrosation are accounted for if the postulatedaddition compounds are equally readily oxidised but the dehydrationto p-nitrosophenol is faster than that to the o-isomeride.The directive effects of a series of groups have been determinedin recent years by F. Challenger and his associates. The resultsof the nitration of a number of aromatic thiocyanates and seleno-W. Borsche and A. Bodenstein, Ber., 1926,59, [B], 1910; A., 1926, 1133.G.Shroeter, Annalen, 1922, 426, 83; C. Smith, J., 1904, 85, 730; A.,1922, i, 126.7 L. A. Pinck, J. Arner. Chem. SOC., 1927, 49, 2536; 3'. H. Cohen, Proc. K.Akad. Wetensch. Amsterdam, 1928, 31, 692; M. Battegay, Bull. SOC. chim.,1928, [iv], 423, 109; H. H. Hodgson and A. Kershaw, J., 1930, 277; A., 1927,1177 ; 1928, 272,402 ; 1930,466.a Ber., 1930,63, [B], 1577, 2074; 2. physikal. Chem., 1930, [B), 10, 22; A.,1033, 1429, 1673.4 Compare the coupling of diazonium salts with jl-naphtholORGANIC CHEMISTRY .-PART 11. 133cyanatesg show that the SCN and SeCN groups have far greaterop-directive powers than the halogens but less than the NHAcgroup. An interesting point in connexion with the nitration ofo- and p-tolyl thiocyanates is discussed below (see o/p Ratio, p.138).Another result of importance is the production of 72% of them-substitution product in the nitration of phenylboric acid,Yh-B(OH),,10 at - 20". The boron atom has in this compound onlysix electrons in the outer shell. Conjugation with the nucleus(+T effect) is impossible and the tendency is conversely to attractelectrons from the nucleus in order to complete the octet. Con-sequently m-direction prevails. Tribenzylphosphine oxide anddibenzylphosphinic acid are nitrated almost entirely in the p-position.A study of the nitration of a long series of alkyl benzoates gavethe following proportions of m-nitration product : 11AZkyZ ...... Methyl Ethyl Propyl Butyl Amy1 Hexyl Heptyl OctyiAZkyE ......Cetyl Chloromethyl /3-Chloroethyl y-Chloropropy]AZkyZ ...... 6-Bromoethyl y-Bromopropyl woButyl sec.-Butylyo m ......... 71.5 72.8 69-4 65.2y0m ......... 72.6 69.9 71.8 67.9 68.3 63.7 62-8 60.2y0m ......... ca. 52 81.9 75.8 77.3AlkyZ ...... isoPropyl 8ec.-Octyl te.rt.-Butylyo m......... 64-1 59.4 59.4The progressive depression of the m-directing power of the estergroup as the normal carbon chain increases is clearly seen, but thereis a definite alternation in the first three and possibly the first fiveof the series. The effects of branched chains and halogen substitu-tion are fully in accord with the known inductive effects.The surprising tendency for certain alkyl groups to enter them-position in toluene during its condensation with alkyl halidesin presence of aluminium chloride has been confirmed by J.B.Shoesmith and J. F. McGechen.12 From tert.-butyl chloride inpresence of either aluminium or ferric chloride, tert. -butyltolueneresulted in the ratio m : p = 65-70 : 35-30. With n-butylchloride the sec.-butyltoluenss were obtained, m : p = 75 : 25.Positive Poles.-The m-orienting influence of oxonium oxygen andquinolinium nitrogen has been observed by R. J. W. Le FBvre 13S F. Challenger and A. D. Collins, J., 1924, 125, 1377; F. Challenger andA. T. Peters, J., 1928, 1364; F. Challenger, (Miss) C. Higginbottom, and A.Huntingdon, J., 1930,26; A., 1924, i, 953; 1928, 750; 1930, 332; compareAnn. Reporb, 1924, 21, 110.10 A. D. Ainley and F. Challenger, J., 1930, 2171 ; A., 1457.11 A.Zaki, J., 1928, 983; 1930, 2269; A., 1928, 636; 1930, 1578.12 J., 1930, 2231.13 J . , 1929, 2771; R. J. W. Le Fhre and F. C. Mathur, J., 1930,2236; A.,271, 1696134 BENNETT AND CHAPMAN :with the 2-phenylbenzopyrylium and 2-phenylquinolinium salts (I)and (11), nitration of which attacks exclusively the positions indic-ated. No isomerides were detected and the yields were 86% and1 0 0 ~ o respectively. 2-Phenylquinoline itself gave 66 yo of thecorresponding nitro-compound.x r”? i5 NMelXThe effects of the sulphonium and selenonium groups in aromaticsubstitution have been determined. A. Pollard and R. Robinson 14find that benzyldiethylsulphonium picrate yields on nitration 28 yoof the m- and about 61% of the p-isomeride.These figures shouldbe compared with those for benzyltriethylammonium picrate(85% m) l5 and benzyldiethylamine sulphate (4&-52y0 m ; l661 yo m).17 A direct comparison of phenyl and benzyl sulphoniumand selenonium salts in nitration has been made by J. W. Baker andW. G. Moffitt,lg the percentages of m-products being :Ph-iMe, . . 100 Ph*CH,.iMe, . . 52Ph-ieMe, . . 100 Ph*CH,*SeMe, . . 16These and previous results all show that for similar compounds theproportion of meta-substitution (a) decreases with increasing atomicnumber of the directing atoms in a single group of the periodicsystem, ( b ) increases with increasing atomic number in any oneperiod. Thus for m-directive power we have the series N<O andg>ie, whereas for op-direction we have N>O>F.19The fact that a basic group loses its predominant position in thecontrol of substitution when it is converted into a salt, as a resultof the relative slowness of the reaction in presence of the positivepole, is illustrated by several recently recorded instances.Thusnitration of o- and m-bromoanilines in concentrated sulphuric acid *O1 4 J . , 1930, 1765; A . , 1302.15 13’. R. GOSS, C. K. Ingold, and I. S. Wilson, J., 1026, 2440; &4., 1926,l6 B. Fliirscheim and E. L. Holmes, ibid., p. 1562; A., 1926, 830.17 Baker and Ingold, Ref. 27 (p. 136).la J . , 1930, 1722; A., 1302.2o R. Luke; and J. Fragner, J. Czech. Chem. Comrn., 1929,1, 294; A., 1929,4-+ +1133.Ann. Reporte, 1926, 23, 138.804ORGANIC CHEMISTRY.-PART II.135takes place in positions 5 and 4 respectively, as shown in (111) and(IV), the bromine atom having the greater influence. Further(111.) ?Tso4observations of the nitration of phenylbenzylamine derivatives 21are fully in accord with modern views. When phenylbenzyl-n-butylamine and phenyldibenzylamine are nitrated in sulphuricacid, the isolated products (c. 50% and 21% respectively) &rephenyl-m-nitrobenzyl-n-butylaniline and phenyldi-m-nitrobemyl-amine. On the other hand, phenyldibenzylamine nitrated in aceticacid, and phenylbenzylnitrosoamine with ordinary nitric acid, aresubstituted according to the op law in the phenyl nucleus as wouldbe expected.22+Acids and t,h-Bcc;ses.-It has been pointed out 23 that the anionof a +acid such as phenylnitromethane, in which the negativecharge may be regarded as distributing itself between the possiblePh*CR:NO*O.tautomeric positions, should undergo op-substitution :J.W. Baker has made a study of the nitration of such +acids undervarious c0nditions.~4 There was considerable meta-substitutionin all cases, but the results fully bear out the author’s view, thealkali salts showing a much lower proportion of m-direction, unlessnitric acid of density 1.529 was used. Thus the free phenylnitro-methane yielded 67% m- + 33% op-, but the potassium salt (withnitric acid of density 1.497) gave 42% m- + 58% op. The contrastwas even more striking in t)he case of ethyl a-nitrophenylacetate,NO,*CHPh*CO,Et. These results emphasise the large differenceswhich may arise from small differences in the composition of thenitrating acid, a point shown even more clearly in another paperfrom the same source describing the nitration of phenylbromo-~yanonitromethane,~~ where a similar variation in the concentrationof the nitric acid raised tho proportion of m-products from 32%0 -to 94%.21 J.Reilly, P. J. D r u m , and T. V. Creedon, J., 1929, 641 ; Sci. Proc. Roy.Compare Ann. Reports,p2 R. D. Desai, J . Indian Chern. SOC., 1928, 5, 425; A., 1928, 1237.25 Ann. Reports, 1927, 24, 115; 1928, 25, 140.24 J., 1929, 2257; A., 1929, 1447; compare B. Fliirscheim and E. L.26 J. W. Baker and C. K. Ingold, J., 1929, 423; A., 1929, 546. See alsoDublin SOC., 1930,19, 377 ; A., 1029, 691 ; 1930, 904.1928, 25, 138.Holmes, J., 1928, 453 ; A., 1928, 403.Ann. Reporta, 1929, 26, 124136 BENNETT AND CHAPMAN :Just as +acids with phenyl attached to the $-atom show op-direc-tion, so $-bases tend to direct m.B. Flurscheim and E. L. Holmes z6obtained 87.5% of m-nitro-derivative from benzylidene-m-nitro-aniline, C,H,*CH:N*C,H,-NO,, by nitration in concentrated sulphuricacid. They held that this represented a nitration of the free base,and the result was important, for Fliirscheim’s theory requires thatPh*CH:NR shall show higher m-direction than Ph*CH:O (81% m-),whereas the electronic theory leads to the opposite prediction for@ nitration of the free base. Clear evidence }g has been produced, however, by J. W. Baker mcH’NH*Ar and C.K. Ingold27 that the Schiff’s base ispresent in the sulphuric acid as a salt (V),(V-1 which was actually isolated. The proportionof m-nitro-derivative was moreover reduced by addition of arnmon-ium sulphate from 89% to 84%, the magnitude of which effect isin accordance with expectation, since the salt of which the dis-sociation is being depressed is it $-salt with a neutral form intowhich it can pass. The situation is therefore reversed and thequestion may be regarded as decided in favour of the electronictheory.An observation of considerable interest both from a theoreticaland a practical point of view is that the p-toluenesulphonamido-group has a markedly stronger op-directing power than the acet-amido-group.28 For example, 2-p-toluenesulphonamidodiphenyl isreadily nitrated first in the 5- and then in the 3-positionY yielding(I) and (11), whereas 2-acetamidodiphenyl is nitrated in position 4’to give (III).29O2N 0,N7 1 -NH*S0,*C7H7 NHAc(1.1 (11.) (111.)The corresponding methylated acyl derivative, 2-p-toluenesulphon-methylamidodiphenyl, is nitrated with much greater difficulty a tposition 5 .Bearing in mind the solubilityof the unmethylated sulphonamido-compoundin alkali, it appears that the nitrogen atoma partial negative charge on it, the systempressed by methylation.in it has a special directive power owing to- being a +acidic one. This is of course sup-2 6 J . , 1928, 2230; A . , 1928, 1126. 27 J . , 1930, 431 ; A . , 594.28 F. Bell, J., 1928, 2770; 1929, 2784, 2787; A ., 1928, 1367; 1929, 204.29 H. A. Scarborough and W. A. Waters, J., 1927, 89; A., 1927, 236ORGANIC CHEMISTRY.-PART 11. 137The Ortho-Puru Rutio.--In an interesting discussion of theinfluences which determine the ratio of the ortho- and para-directedproducts in aromatic substitution, A. Lapworth and R. Robinsonconclude that, as a result of the presence in the benzene nucleus of agroup A having a smaller attraction for electrons than has thehydrogen atom, electron-availability and therefore reactivity shouldtend to be greater in the 0- than in the p-position :0With a substituent B which attracts electrons, the potential gradientin the molecular field will be reversed and consequently reactionshould occur pf>o’. This leads to the expectation of a high o/pratio for groups such as Me but a low one for m-directive groupssuch as NR,, NO, and CO,H, and also for op-directive groups havinga natural electron attraction such as C1, OMe.The considerable o-orientation caused by NO,, CHO, and CO*CH,is regarded as an anomaly due to the attraction of the reagent tosuch groups, which are actually known to exert basic function^.^^The tendency of the nitro-group to enter 0- to a m-directive grouphas been emphasised by J.Obermiller,32 who terms this phenomenon“ auto-orientation.” In this respect there is a sharp contrast betweenthe positions of entry of the NO, and SO,H groups (possibly owingto the greater acidic strength of nitric acid as compared with sul-phuric acid).An appreciable proportion of o-product is formedin the nitration of nitrobenzene, benzenesulphonic acid, and benzoicacid, but none in the corresponding sulphonations, and the dis-similar products of nitration and sulphonation of metanilic acid alsoshow this effect :+30 Mern. Mancheeter Phil. SOC., 1927, 72, 43; A., 546; compare Ann.Reports, 1926, 23, 140.31 It should be noted t o avoid confusion that this view involves a reversalof the explanation given by J. Allan, A. E. Oxford, R. Robinson, and J. C.Smith (J., 1926, 409; A., 1926, 397) where these o-substitutions were takento be normal and the influence of the electrical field was stated in the oppositesense.32 J . pr. Chem., 1914, [ii], 89, 70; 1930,126, 257; A., 1914, i, 513; 1930,1028.E138 BENNETT AND CHAPMAN :The former authors 30 emphasise the large influence which con-ditions of reaction may have on the o/p ratio, and a further dis-turbing factor which may operate is the steric inhibition of o-sub-stitution, nitration of the higher alkylbenzenes being a possiblecase in point.Apart from these difficulties this view of the o/p ratio 33 gives asatisfactory explanation of a large number of facts.Some newcases of this kind will now be referred to.The contrast between the orientations of the principal productsof nitration of o- and p-tolyl thiocyanates (I) and (11) (comparep. 133), suggesting at first sight that Me has a greater directive powerthan SCN in one case and a smaller in the other, is clearly a naturalconsequence of the opposite o / p ratios which the two groups possess(nitration of toluene, o : p = 56 : 41 ; of phenyl thiocyanate, o : p =20 : 80).SCN SCNA study of the nitration of cyclohexylbenzene and its p-halogeno-derivatives by H.A. Mayes and E. E. Turner 34 is of interest fromthe same point of view. The percentages of mononitro-isomeridesobtained are indicated by the figures in the following formuls, datafor toluene compounds being added for comparison :The cyclohexyl group has thus a lower o/p ratio than methyl inagreement with the behaviour of the higher alkyl groups. Thefigure of the ratio p/*o is 3.3, which may be compared with thosefor chlorine (4.6) and bromine (3.3). The explanation of theresult for p-bromocyclohexylbenzene may lie in a higher level ofpromotion of reactivity due to alkyl as compared with that due tobromine.s3 The interpretation given by C.K. Ingold (Rec. trav. chim., 1929, 48,807 ; A., 1929, 1289 ; Ann. Reports, 1926, 23, 140) more closely resembles thescheme of J. Allan, A. E. Oxford, R. Robinson, and J. C. Smith, Eoc. cit.34 J., 1929, 500; A., 1929, 550, and for comparative data A. F. Hollemanand J. P. Wibaut, Proc. K. Akad. Wetensch. Amsterdam, 1912, 15, 694; L.Gindraux, HeEv. Chim. Acta, 1929, 12, 921 ; A., 1913, i, 169; 1929, 1433ORGANIC CHEMISTRY .-PART 11. 139The ratios of o/p-hydroxy-aldehydes formed in the Reimer-Tiemann reaction from a series of phenols C6H4X*OH have beenfound to be as follows : 35X .................. H o-Me m-Me o-C1 o-Br 0-1O/P ..................0.6 0.48 0.46 1.6 1.25 1.0701p .................. 0.87 0.71 0.72 0.78 0.06X .................. m-3’ m-C1 m-Br m-I o-CO~HThe figures for m-C1, Br, I may be regarded as identical, but they arehigher than that for phenol and lower than that for m-fluorophenol.These facts are consistent with the directing power of the halogens,which in general promote substitution para with respect to theirown position and particularly so in the case of fluorine. Thehigh o/p value found for o-halogenophenols appears to constituteanother instance of the tendency for substituents to enter thenucleus in the ortho-position to a strongly op-directive group whena second (weaker) op-directive atom is next to it (in the 0’-position),and the relative effect of the halogens (Cl>Br>I) in the presentcase is in accordance with expectation.36Orientation in Diphenyl Compounds.-Although pp’-dibromo-diphenyl had been shown to yield on nitration the o-nitro-compound(I),37 pp’-difluorodiphenyl was stated 38 to be substituted in them-position to give (11).R.J. W. Le E’Bvre and E. E. Turner find 39 that the latter productdoes not react with piperidine and must be, not (11), but the fluorineanalogue of (I). A second nitro-group enters to give an unsym-metrical product (111), as in the case of the dibromo- and dichloro-diphenyls : one fluorine atom in (111) is removed by piperidine.Diphenyl and its mono- and di-nitro-derivatives are nitrated inthe o- and p-positions to the extent shown by the figures attached : 4oH.H. Hodgson and J. A. Jenkinson, J., 1929,469,1639; H. H. Hodgsonand J. Nixon, J., 1929, 1632; A., 1929, 669, 1177.s6 Ann. Reports, 1926, 23, 138; E. L. Holmes, C. K. Ingold, and (Mrs.)E. H. Ingold, J., 1926, 1684; A,, 1926, 947.57 Ann. Reports, 1926, 23, 137.38 G. Schiemann and W. Roselius, Ber., 1929, 62, [B], 1805; A., 1929, 1052.as J., 1930, 1168; A., 901.40 H. C. Gull and E. E. Turner, J., 1929, 491; A., 1929, 647; compareF. Bell, J. Kenyon, and P. H. Robinson, J., 1926, 1239, 2706; A., 1926, 830,1241 ; W. Blakey and H. A. Scarborough, J., 1927, 3000; A,, 1928, 166140 BENNETT AND CHAPMAN :68The o/p ratio is approximately the same in the nitrodiphenyls andis definitely lower than that for diphenyl itself, as it should be.Further nitration gives the 2 : 4 : 4’-trinitro- and the 2 : 4 : 2’ : 4’-tetranitro-compound.The study of the chemistry of halogeno-, hydroxy-, and amino-diphenyls and their acyl derivatives has engaged the attention ofa number of workers, and the resulting orientation data are for themost part normal.Two general conclusions of importance are :that the phenyl or substituted phenyl nucleus exerts a strongop-directive effect; and that there is no evidence of conjugationbetween the nuclei, their behaviour being independent so far asconjugative (tautomeric) effects are ~oncerned.~l This is in con-trast with the behaviour of the phenylpyridine~.~~The sequence of directing powers OH> OMe> O*SO,*C,H, isclearly indicated in the results obtained by F.Bell and J. Kenyon 43in which it is found that 4-hydroxydiphenyl is substituted first inposition 3, the methoxy-compound in positions 3 and 4‘, and theacyloxy-compound in position 4’ :There are frequent examples of the well-known tendency forsubstitution to take place in the second nucleus when a deactivatinggroup is present in the first.44 Two anomalous cases of substitutionwere reported,45 namely, the entry of bromine into the 4’-positionin 4-acetamidodiphenyl and of the nitro-group into the 4’-positionin 2-acetamidodiphenyl (yield 50%). However, J. Kenyon and41 H. A. Scarborough and W. A. Waters, J., 1926, 557; R. J. W. Le FBvre42 Ann. Reports, 1926, 23, 136.43 J., 1927, 3044; A., 1928, 145.44 Compare Ann.Reports, 1926, 23, 134; 1928, 25, 138.4 5 H. A. Scarborough and W. A. Waters, J., 1926, 5 6 e 1927, 89, 1133;A . , 1926, 612; 1927, 236, 666.and E. E. Turner, J., 1928, 245, 963; A., 1926, 512; 1928, 283, G30ORGANIC CHEMISTRY.-PART 11. 141P. H. Robinson 46 have found that chlorine substitutes the 4-acet-amido-compound exclusively in position 3- to give (I) and thereaction with bromine yields the analogous product (I) to the extentof 50% with only 30% of (11). The nitration referred to is notsurprising, as the sulphurio acid which was present must tend toconvert the acetamido-group into a deactivating rather than anactivating agent .47Other PoZynuclear Types.-The examination of a number of casesof substitution of benzophenones has given straightforward results,48but an interesting problem is presented by the proportions of mono-nitration products, as shown by the figures attached to the annexedformuke, of pp’-chlorobromo-benzophenone and -diphenylsulph-one : 49Nitration of diphenyl ether has given 24% and 44% of 0- andp-nitro-compounds respectively : but in presence of acetic anhydride46% 0- and 54% p - have beenpp’-Dichlorodiphenyl ether is nitrated in the 00’-positions(exclusive direction by the oxygen atom), but on the other hand aseries of compounds of the type (I), in which A is either an acyl or anitrated phenyl radical, are nitrated in position 5 as shown, thedirective power of the oxygen atom being no doubt reduced by thegroup A.514 6 J., 1926, 1242, 3050; A., 1926, 830; 1027, 142.4 7 Compare the nitration of aceto-p-toluidide and that of aceto-sn-4-xylidide ;See also Ref.2848 W. Blakey, W. I. Jones, and H. A. Scarborough, J . , 1927, 2865; W.Blakey and H. A. Scarborough, J., 1928, 2489; W. A. Waters, J., 1020,2106; L. Chardonnens, Helv. Chim. Acta, 1929,12, 649; J . van Alphen, Rec.trav. chim., 1930, 49, 153, 383; A . , 1928, 66, 1246; 1929, 928, 1299; 1030,476, 603.H. E. Dadswell and J. Kenner, J . , 1927,1102; A., 1927,656.(p. 136).49 L. G. Groves and E. E. Tumor, J . , 1929, 509; A., 1929, 561.50 C. M. Suter, J . Arner. Chem. Soc., 1929, 51, 2581; G. Lock, Monatsh.,1930,55, 167; A., 1929, 1174; 1930, 767.5 1 R. J. W. Le Fhvre, S. L. M. Saunders, and E. E. Turner, J ., 1927,1168; L. G. Groves, E. E. Turner, and G. I. Sharp, J., 1929, 512; A , , 1927,660; 1929, 551142 BENNETT AND CHAPMAN :A tendency for substitution in one nucleus of a diphenyl etherto stop when one out of two available o-positions in a single nucleushas become occupied has been noticed 52 in certain brominations.A0 Br(JA 0-0-0 NHA~~if NHAcc1(1.1 (11.) (111.)Bromination of o- and p-acetamidodiphenyl ethers also givesproducts (11) and (111) in which substitution does not occur o- tothe acetamido-group. The acetamido-group, it should be remem-bered, has a low o/p ratio.Several other studies of diphenyl ethers have been published.%Some of the earlier work on the nitration of azobenzene wasfaulty. Nitration or bromination of this substance occurs inpositions 4 and 4', and this result is unaltered by the presence a tposition 2, 3, or 4 of either Me or C1.The group C,H,*N:N- is thusmore strongly directive than either chlorine or methyl. On theother hand, methoxyl, amino-, and acetamido-groups supersede theazo-group in controlling substitution in such molecules.54Heterocyclic Compounds.-The nitration of the three isomericbenzylpyridines is of interest in comparison with that of the phenyl-pyridines 55 previously reported. The amounts of 2-, 3-, and 4-m-nitrobenzylpyridines found are 10.4y0, trace, and 4.8% respectively.There is thus a reduction of m-direction (corresponding amounts ofthe m-nitrophenylpyridines : 2-, 34.9% ; 3-, trace; 4-, 286%), aswould be expected from the presence of an extra carbon atombetween the pole and the seat of reaction.There is nevertheless areal direction to the m-position.The nitro-group entered chiefly the pposition of 2-phenyl-l-methylglyoxaline, 1 - and 2-phenylglyoxalines, and 4-phenyl-piperidine. The m-nitro-compound predominated, however, when4-hydroxy-2-phenyl-6-methylpyrimidine 56 was nitrated.52 H. A. Scarborough, J., 1929, 2361 ; H. McCombie, W. G. Macmillan, andH. A. Scarborough, J., 1930, 1202; A . , 1929, 1439; 1930, 1034.63 (Miss) R. V. Henley and E. E. Turner, J., 1930, 928; (Miss) D. L. Foxand E. E. Turner, ibid., pp. 1115, 1853; L. C. Raiford and I. J. Wernert,J. Amer. Chem. SOL, 1930,52, 1205; A., 907, 909, 1283, 767.64 J. Burns, H. McCombie, and H.A. Scarborough, J., 1928, 2928; A.,1929, 58.6 5 F. Bryans and F. L. Pyman, J., 1929, 649; A., 1929, 577. CompareAnn. Reports, 1926, 23, 136.66 R. Forsyth and F. L. Pyman, J., 1930, 397; A., 618. Compare Ann.RepOTt8, 1924, 21, 110ORCANIC CHEMISTRY .-PART 11. 143Polar In@ences on Properties and Rectctions.Elimination Reactions. Exha;ustive Methylation.--In a series ofpapers on the modes of' decomposition of quaternary ammoniumand phosphonium compounds and of sulphones, C. K. Ingold andhis assistants have elucidated the mechanism of the four reactionswhich may occur and the structural conditions which determinetheir relative imp~rtance.~' The first of these (A) is the olefin-elimination familiar in the Hofmann degradation, which takes placeaccording to the scheme :Wide variations of the structure of the quaternary base affect thisreaction in the expected manner.Substitution in the c(-CH2 haslittle effect and, if complete (absence of a-H), does not preventreaction A. On the other hand, the character of R hastens orretards it by promoting or suppressing the incipient ionisation ofthe p-H. Elimination of ethylene (R = H) is thus more ready thanthat of higher alkylenes (R == alkyl), but styrene (R = Ph) and stillmore p-nitrostyrene are liberated with great ease in accordance withthe known repulsion of electrons by alkyl and attraction by phenyland nitrophenyl.When reaction A is difficult, the second reaction B occurs, whichmay be written :This reaction, as the theory requires, becomes increasingly importantif R and R1 contain chains of carbon atoms which branch or extend-the latter effect becoming constant for chains of more than fouratoms.Reaction C, characteristic of phosphonium hydroxides, involvesthe elimination of a paraffin thus :e e OH 8 @e 8 fH,OE3R,P + OH T+= R4P*OH ---+ R4P0 -+ R,PO + R --+ RH + Ok67 C.K. Ingold with W. Hanhart, J., 1927, 997; C. C. N. Vass, J., 1928,3126; G. W. Fenton, J., 1928, 3127; 1929, 2338, 2342; 1930, 706; J. A.Jessop, J., 1929, 2357; 1930, 708, 713; A., 1927, 650; 1929, 171, 176, 1423,1431 ; 1930, 73, 739, 759. Compare J. v. Braun, W. Teuf€ert, and K. Weiss-bach, Annalen, 1929,472, 121 ; A., 1929, 1046144 BENNETT AND CHAPMAN :The formation of the undissociated form of the base must involvean increase to ten in the number of electrons in the outer shell ofthe phosphorus atom.As the nitrogen atom is unable t o increaseits outer shell, reaction C is impossible with ammonium compounds.The radical R eliminated as paraffin should be that which besttolerates a negative charge, and this is verified by a long series ofcomparative decompositions. Readiness of elimination is in theorder benzyl > phenyl > methyl > p-phenylethyl > ethyl > higheralkyls : carbethoxymethyl is eliminated with great ease.Reaction C is very facile for phosphonium hydroxides and con-sequently reaction A is then not normally observed, but by selectinga case where A should be favoured it was found in fact to take place :the hydroxide CHPh,*CH,$Bu,} 0"H gave the olefin CPh,:CH2 andthe phosphine as principal products.Decomposition of sulphones tends to follow a reaction similart o A :H13 E d R-CH-CH,-SO,*Alk + 0"H -+ R*CH:CH, + ALk*S8, -t- H,O,an alkylene and a sulphinate being produced.Substitution of other negative ions for hydroxyl should affectthe extent to which reaction A takes place in any instance.It isfound in fact that in the decomposition of the hydroxide, phenoxide,and m-nitrophenoxide of a quaternary ammonium salt reaction Bis progressively increased a t the expense of A, the negative ionsinvolved being in order of diminishing proton-a,ffinity. On theother hand, the ethoside ion has a greater proton-affinity thanhydroxyl. Consequently the sulphone disruption is brought aboutby sodium ethoxide in many cases where it fails with sodiumhydroxide.Sulphones may also exhibit reaction C, the order in which theradicals tend to this type of elimination being, as the theory requires,the same as for the phosphonium hydroxides.Finally, a fourth reaction (D) involving 1 : 1-elimination some-times occurs when @-H atoms are absent :n I f3@H"0 + H-CRR1--NR, -+ H,O + CRR1+ NR,The methylene, CRR1, having only a sextet of electrons, immediatelypolymeriaes or isomerises.This has been observed with fluorenyl-trialkylammonium hydroxides (which yield bis-00'-diphenylene-ethylene), with the analogous sulphonium hydroxide, and withbenzylmethylsulphone. The order of facility of methylenic extruORGANIC CHEMISTRY .-PART 11.145sion (D) is : g-fluorenyl> benzyl>methyl, which correctly followsthe order of anionic ~tability.5~Strengths of Carboxylic Acids.-Dibasic acids. According to thetheory of N. Bjerrum 59 the free pole of one carboxyl group of twopresent in a molecule will affect the active mass of hydrogen ionsnear the other (about to dissociate), so that the following relation willhold : log KJK, - 0.6 = 3.1 x 10-8/r, where K, and Kz are thefirst and second dissociation constants of the dibasic acid and r isthe distance between the ionic centres.The subject has been taken up by R. Gane and C. K. IngoldYGowho from measured values of K , and K2 find the following valuesof r in A. for the normal acids CO,H*[CH,],*CO,H :r, A ..........1.5 5.0 9.2 11.6 13.2 14-5 16.8Various effects which must tend to influence these figures are dis-cussed. A large distortion of the results must arise from polareffects, of which the chief will be the inductive effect of one carboxylgroup on the ionisation of the other-tending to make the apparentvalue of r too small.From the third to the seventh acid in the above list the values of rare fairly regular with an average increase of 1.73 8. per CH,. Thedifference between this and the average distance in the crystal(1 *26 A.)61 is regarded as representing the systematic error due tosolvation and electrostriction. The large deviation from the seriesof values of r shown by the first two members represents the polareffects mentioned and is vanishingly small when three carbon atomsare interposed between the pole and the carboxyl.n = 1 2 3 4 5 6 7This result and those obtained in nitrating the saltsPh*[CH,],*NMe,)X 62justify the use of comparative values of r found for p-substitutedglutaric acids in connexion with the valency-deflexion hypothesiswithout fear of error due to the polar effects of the alkyl groups.63A series of determinations with mono- and di-substituted malonicacids, CRR1(CO,H),, gave values of the apparent distances r betweenthe carboxyl groups.The order of these is dissected in a mostconvincing manner and shown to be not only qualitatively butsemi-quantitatively such as should result from the combinedoperation of the valency-defiexion effect (confirmed in the p-substi-pare Ann.Reports, 1928, 25, 121.58 Benzybmethyl in order of either anionic or kationic stability. Com-2. physikal. Chem., 1923, 106, 219; A., 1923, i, 1059.6o J., 1928, 1594, 2267; 1929, 1691; A., 1928, 846; 1929, 1144. Compare61 W. A. Caspari, J., 1928, 3235; A,, 1929, 126.62 Ann. Reports, 1926, 23, 131.also A. I. Vogel, J . , 1929, 1476; A., 1929, 1009.63 This vol., p. 153146 BENNETT AND CIEAPMAN :tuted glutaric acids) and the known internal polar effects of thealkyl groups.The enhanced strength of all o-substi-tuted benzoic acids was ascribed by B. Flurscheim 64 to a sterichindrance effect. But a pure steric hindrance, although it mightaffect the speed of dissociation, should not affect the equilibrium-position.This particular effect of ortho-substituents is betterascribed t o a direct polar effect of the substituent on the carboxylThe effect of the methyl group in the o-position is also toincrease the strength of benzoic acid, although its familiar polareffect is to weaken an acid. This indicates the existence of a widelyspread electrical field outside the methyl group of an electron-attracting kind : a similarly reversed field must exist outside anypolar group at a point on its axis produced into space.66Attention may here be directed to recent determinations of thestrengths of the halogenobenzoic acids, the phthelic p-cyano-benzoic acid, 68 and a- and p-selenocyanopropionic acidsY6O fromwhich the cyano- and selenocyano-groups are seen to approach thenitro-group in degree of polarity.HydroZysis of Esters.-It has for some time been clear that thevelocities of hydrolysis and esterification must be influenced by bothpolar and steric hindrance factors.It was essential to find amethod of observing the effects separately or of discriminatingbetween the two effects when superposed. The speeds of alkalinehydrolysis of m- and p-substituted benzoic esters determined byK. Kindler 7O illustrate the separate polar effect, for steric hindrancedue to the substituents must here be negligible. The attack is byhydroxyl and consequently the reaction is facilitated by electron-attracting groups .An important method for revealing the polar influences in esterhydrolysis independently of steric effects has been developed byC.K. IngoldY7l starting from the views of H. M. Dawson and ofSubstituted benzoic acids.64 J., 1909,95,718; 1910,97,84; Chem. andInd., 1925,44, 246.66 A. Lapworth and R. H. F. Manske, J., 1928, 2533; A., 1928, 1245.66 G. M. Bennett and A. N. Mosses, J., 1930, 2364; A., 1555.6 7 R. Kuhn and A. Wassermann, Helv. Chim. Ada, 1928, 11, 31, 44; A.,6 8 E. P. Valby and H. J. Liicas, J. Amer. Chem. SOC., 1929, 51, 2718; A.,69 A. Fredga, J. pr. Chem., 1929, [ii], 121, 56; A., 1929, 426.'O Annalen, 1926, 450, 1 ; A., 1927, 55.1928, 240.1929, 1384.Compare Ann. Reports, 1928, 25,147, where these results were, owing to an error in abstracting, stated in theinverted order. The speeds of hydrolysis fall in the order of diminishingstrengths of the substituted benzoic acids.71 J., 1930, 1032, 1375 ; with (Miss) C.M. Groocock and A. Jackson, ibid.,p. 1039; A., 868, 869, 1131ORUANIC CHEMISTRY.-PART 11. 147J. N. Bronsted on reaction catalysis. Whatever may be the detailedmechanism of the reactions,72 the alkaline hydrolysis depends onthe action of hydroxyl and the acid hydrolysis on the action ofhydrogen ions. It is demonstrated that the ratio of the velocitiesof alkaline and acid hydrolysis, koH/kH, is probably within widelimits a function of polar factors only (the steric effects cancellingout).in which pH* is the hydrolytic stability maximum orpoint of minimumhydrolysis and K , is the ionic product for water. The progressivepolar effects of methyl groups in the hydrolysis of alkyl acetates isshown by the following figures calculated from data available :Alkyl Me CH,Me CHMe, CMe,lo-' ( k 0 ~ I h 4 16.1 9.9 4.7 1.5Use is made of Dawson's relation : 73- - log KV = log ( k o ~ / k ~ )An extended series of alkyl groups has been studied by using thewater-soluble glyceric esters in solutions buffered with sodiumglycerate. The order of falling k,,=/kH ratio was Me>Et>Pr>Bu>Am>iso-Am>iso-Bu> iso-Pr, which is clearly in accord withthe usual inductive effects of the groups.In the third memoir the author discusses the various factorscontrolling the speeds of hydrolysis of esters.The statistical factorleads to the expectation that the ratio of the velocities of the firstand second stages in the hydrolysis of a symmetrical dicarboxylicester, k1/k2, should be 2.0, and this is true if the ester groups areseparated by at least two carbon atoms.Otherwise the pohrfactor is also important. Consideration of the influence of one poleon the reaction at the second carbethoxyl leads for symmetricaldicarboxylic esters to k J k 2 z 2 exp(7/108r) and the values of rdetermined in this way are in general agreement with those foundfrom the dissociation constants. 74 A systematic consideration isalso given to steric hindrance and the influence of the medium.Types of Polar Eflect.-It has recently been emphasised by A.Lapmorth and R. H. F. Manske 75 that the effect ef a substituent onthe reactivity of adjacent atoms observed experimentally mayresult from several superposed influences difficult or impossible toseparate. The terms " quantitative " and '' electropolar " factorsused by B.Flurscheim 76 are advocated ; the former being indicatedby the op-directive action of a group, and the latter by a comparison72 T. M. Lowry, J . , 1925, 127, 1371; J. W. Baker, J., 1928, 1583; A , ,1925, i, 886; 1928, 870.73 J., 1927, 1146; A., 1927, 632.74 This vol., p. 145.7 5 J., 1928, 2533; A,, 1928, 1246. 76 L O G . cit148 BENNETT AND CHAPMAB :of the strengths of the m- and p-substituted benzoic acids with thatof benzoic acid.The increase in availability of electrons at A, when H-A, isconverted by substitution into A-A, is termed the “primaryinternal effect ” of A and is written Z--Al.Examination of the stabilities of a further series of ketone cyano-hydrins shows that the introduction of an alkyl group has (a) 8stabilising “ steric effect ” a t close quarters and ( b ) a destabilisingelectropolar influence at points more remote.77The effect of the oxygen atom in MeO*CH,Cl is to increase theincipient anionisation of the chlorine, which consequently showshigh reactivity.In a similar way, although the cyanogen group hasa tendency to particularly stable covalent attachment to carbon, theaddition of alkali to HO*CR,*CN produces O*CR,*CN and here thecyanogen becomes rapidly ionised : the aminonitrile NR,*CHMe*CNis highly unstable. These are cases of an effect related to that ofop-direction in which the order of efficiency is O>NR,>OH andBoth a- and p-chloro-sulphides (and the analogous hydroxy-compounds) are reactive in a similar manner,79 but the y-substitutedsulphides are entirely unactivated.The effect thus dies out at adistance of the same order as that found in the dicarboxylic acids(p. 145). The 8-hydroxy-sulphidey however, shows a remarkablereactivity which may be attributed to a large activation at momentswhen the CH,*OH group is near the sulphur atom in space, an eventstereochemically probable.80The possibility should be borne in mind of interpreting theknown facts of both aliphatic and aromatic chemistry in terms oftwo polar effects only : namely, the general polar or electropolareffect (to be regarded as of the nature of both I and D) and theconjugative or tautomeric effect (5!’).81 It is at least significantthat the effects D and I have been described as similar “ in sign andmagnitude in .. . variation as between one directing group andanother,” 82 and there appears to be no definite evidence that theinductive effect is propagated along saturated carbon chains morethan through space. On the other hand the inability of internalz-- A€3eoMe.787 7 A. Lapworth and R. H. F. Manske, J . , 1930, 1076; A . , 1251 ; compare7 8 W. Cocker, A. Lapworth, and A. Walton, J., 1930, 446; A . , 571.‘9 G. M. Bennett and A. L. Hock, J . , 1927, 477; A . , 1927, 365.80 G. M. Bennett and A. N. Mosses, Zoc. cit. (Ref. 66; p. 14G).81 For symbols I, T, and D, see Ann. Reports, 1927, 24, 151 ; 1928,25, 140.82 C.K. Ingold, Zoc. cit. (Ref. 33; p. 138).Ann. Reports, 1928, 25, 147ORGANIC CHEMISTRY.-PART 11. 149polar effects in aliphatic compounds to penetrate beyond thesecond carbon atom and the parallelism with op-directive effectssuggest that they are essentially of the tautomeric kind.Stereochemistry .A valuable general discussion of optical rotatory power has beenheld during the year.83 The investigations of W. Kuhn 84 haveprovided us with a relatively simple theory of the mechanism of theoptical activity of dissymmetric molecules, and are of the greatestimportance in connexion with the study of rotatory dispersion andof the relation between rotatory power and chemical constitution.Stereochemistry of Elements of the Xulphur Goup (continued fromAnn.Reports, l927,102).-The following results and others describedbelow under " cis-trans isomerism of ring compounds " affordabundant evidence in support of the newer view of the steric dis-position of the atoms or groups round the sulphur atom in sulphoxidesand sulphilimines .The resolution of the sulphoxide and the sulphilimine derived fromm-carboxyphenyl ethyl sulphide, namely, the compounds (I) and(11), into their optical antipodes by means of their alkaloidal saltsconfirms the earlier results. The occurrence of aromatic disul-phoxides as pairs of diastereoisomerides has been shown in eightCO,HCGH4*C0,H + CGH**CO,H -(1.1 (11.1"-"<, 2 5 C7H7*S02*%--S G , H 5 ss()g<e, (111.)instances, in four of which a single dioxide had been recorded in theliterature.86 Of the two dioxides (111) of 3 : 5-dimethylthiolbenzoicacid, one was found to be resolvable into optically active forms asrequired by theory whilst the other resisted resolution and is there-fore internally compensated.An anomaly of long standing has been removed 87 by the re-examination of the mercuri-iodides of optically active sulphoniumsalts.It is now' found that the mercuri-tri-iodide and -tetraiodide,83 Trans. Faraday Soc., 1930, 26, 266; A., 980.84 2. phyeikal. Chem., 1929, [B], 4, 14; Ber., 1930, 63, [B], 190; A., 1929,W. Kuhn and E. Knopf, 2. physikal. Chem., 1930, [B], 7, 981; 1930, 276.292; A., 717.85 J. Holloway, J. Kenyon, end H. Phillips, J., 1928, 3000; A., 1929, 65.86 E.V. Bell and G. M. Bennett, J., 1928, 3189; 1930, 1 ; A., 1929, 179;Compare1930, 340.W. J. Pope and A. Neville, J., 1902, 81, 1652.M. P. Balfe, J. Kenyon, end H. Phillips, J., 1930, 2554150 BENNETT AND CHAPMAN :B[HgI,] and B,[Hg14] (where B is phenacylmethylethyhulphonium),and also the cadmi-iodides, B2[Cd14] and B,[CdI,], are obtainablewith considerable rotatory powers but racemise somewhat readilyin aqueous solution in presence of iodides.The isolation of enantiomorphous selenonium salts followedrapidly after that of the sulphonium salts, but correspondingevidence in the case of telluronium compounds has hitherto beenlacking. The revision last year 88 of our views as to the nature ofR. H. Vernon's telluronium di-iodides prepared the way for theannouncement 89 of the isolation of phenyl-p-tolylmethyltelluroniumsalts in optically active forms having [M],,,, of about 70".Theactivity is fugitive, but it indicates that the attached radicals intelluronium compounds are arranged in a similar way to those insulphonium compounds.Optically Active aci-Nitroparafim and Diaxo-compounds. Semi-polar Bond to Carbon.-The remarkable results of R. Kuhn andH. Albrecht in 1927 have been completely confirmed by an examin-ation of optically active (3-nitro-octane. The nitro-compoundrecovered from the solution of its alkali salt by acidification at- 70" retains more than 70% of its activity. The formula (I)is necessary for the optically active ion, possibly stabilised as asolvate such as (11).The fact that the rotation of the alkalinesolution is not affected by keeping for 24 hours seems to exclude theidea that the active ion is in tautomeric equilibrium with the form[RR1C&(?1-. A substance of the structure RR1C-N-OH mayL \oJpossibly be involved.g1CH,-CMe*CO,Me(111.1 I >CMe,CH,-CH-NH,'0'A similar problem is presented by the occurrence of opticalactivity in aliphatic diazo-compounds. The original formulaAnn. Reports, 1929, 26, 80.T. M. Lowry and F. L. Gilbert, J., 1929, 2867; A., 1930, 232.R. L. Shriner and J. H. Young, J. Amer. Chem. SOC., 1930, 52, 3332;O 1 G. E. K. Branch and J . Jaxon-Deelman, J . Amer. Chem. Soc., 1927, 49,A., 1269 : compare Ann. Reports, 1927, 24, 103, 107.1766; A., 1927, 852ORGANIC CHEMISTRY.-PART 11.151NNR,C<I I for these substances had been largely abandoned in favourof the open-chain formula, which, according to modern views, maybe either R2C=N=Nx or R2Y--EN- ; 92 but recent discussionof the physical properties of these substances and the closely relatedazides has led to the revival of the earlier formula and they are nowregarded as partly or wholly cyclic.93On the other hand, the possibility of the occurrence of opticallyactive diazo-compounds was indicated in 1920 and succeedingbut the observed rotations of the products from which theexistence of the active diazo-compounds was inferred were small.The dissymmetry of the carbon atom bearing a diazo-group has nowbeen shown in a more convincing manner by the work of F.E. Ray,g5who has isolated and compared the modes of decomposition of thediazo-esters from methyl cis- and trans-aminocamphonanates (111),in the molecule of which there is one asymmetric carbon atom apartfrom that carrying the amino-group. Decomposition of the cis-diazo-compound at low temperatures with dilute acid gave 39.5%of hydroxy-esters and 60.5% of unsaturated esters as compared with68% and 27% of these products respectively from the trans-diazo-compound (of 90%. purity). This difference of behaviour points tothe existence of diastereoisomeric diazo-compounds and an opticallyactive form RR1-$+B-- which may perhaps be in tautomericequilibrium with the cyclic form.Both the aci-nitro-paraffin and the aliphatic diazo-compoundthus appear to have in their molecules a semi-polar bond betweencarbon and nitrogen.The isolation of dimethylsulphoniumfluorenylidide 96 (IV) provides a case of such a bond between carbonand sulphur and strengthens the case for the existence of theanalogous nitrogen compound. The decomposition of the latterwith production of bisdiphenylene-ethylene is similar to the form-ation of tetraphenylethylene from diphenyldiazomethane.Ring Pormation and Stability.-R,eduction of the large ringmono- and di-ketones (Ann. Reports, 1928, 25, 112) by the Clem-92 Ann. Reports, 1922,19, 86.93 Ann. Reporb, 1929, 26, 183; H. Lindemann, A. Wolter, and R. Groger,Ber., 1930, 63, [B], 702; A., 686.94 C. S. Marvel and W. A. Noyes, J . Amer.Chem. SOC., 1920,42, 2269; A,,1921, i, 15; H. M. Chiles and W. A. Noyes, ibid., 1922, 44, 1798; A., 1922, i,924; P. A. Levene and L. A. Mikeska, J . Biol. Chem., 1921, 45, 693; 1922,52, 486; A., 1921, i, 233; 1922, i, 818.Compare F. E. Kendalland W. A. Noyes, ibid., 1926, 48, 2404; A., 1926, 1134.~e 0 @0 @s5 J . Amer. Chem. SOC., 1930,52, 3004; A., 1281.96 C . K. Ingold and J. A. Jessop, J., 1929, 2367; 1930, 713; A., 73, 769152 BENNETT AND CHAPMAN :mensen method has now been described 97 and the resulting cyclo-paraffins having rings of 12-16 and 22-30 carbon atoms have beenexamined together with some new diketones as shown in the tableof m. p.'s below :Mono-Di-ketones - - - - 41" 32 36 42 45 - - -ketones - - - - - - 65 69 73 - - 78"The production of analogous five- and six-membered (hetero-cyclic) rings by internal sulphonium salt formation from phenyl8-chlorobutyl and c-chloroamyl sulphides proceeds as a measurablefirst-order reaction and at 80" the five-membered ring is producedseventy-six times as fast as the six-membered.This ratio is remark-ably close t o that (70 : 1) found for the similar ring-closure whichoccurs in 8- and c-chloro-amines in presence of alkali.98 Theauthors regard this ratio as being a measure of the statistical proba-bilities of approach within atomic distance of the two ends of thefive- or six-atom chain during the course of molecular vibrationswithout appreciable strain.The Valency-dejexion Hypothesis.-A number of further applic-ations of the Thorpe-Ingold hypothesis have been made in recentyears, of which one or two will be selected for mention.The large effect on ring stability caused by the presence of twogemdimethyl groups is well illustrated 99 by the fact that, whereasas-diacetylbutane is a normal diketone (V), the correspondingtetramethyl compound reacts exclusively in the cyclic form (VI).CH,-CH,-COMe(V.) ICH2-CH2*COMeCMe2-CH, I >CMe*OH (I7I.)CMe2-CH*COMeA similar progressive influence is evident among the three com-pounds (I), (11), and (III),l the introduction of two successivemethyl groups into the open-chain compound (I) converting it firstinto a ring-chain tautomeric mixture (11) and then into a stablering compound (111) :9 7 L.Ruzicka, M.Stoll, and others, Helv. Clhim. Ada, 1930, 13, 1152; P I . ,1422.9* G. M. Bennett, F. Heathcoat, and A. N. Mosses, J., 1929, 2567; A.,61 ; H. Freundlich and A. Krestovnikov, 2. physikal. ClLern., 1911, 76, 79;A., 1911, ii, 266.9a I. Vogel, J., 1927, 594; A., 1927, 449.1 E. Rothstein and C. W. Shoppee, J., 1927, 531 ; A . , 1927, 447ORGANIC CHNMISTRY . -PART 11. 153CHMe*CO,H CHMeCOhMe2* co *CO,H CMe, I *CO*C02H &Me2-,$H)*C0,HCH,*CO,HL Y (I, stable.) (11, tautomeric mixture.)CMe,--COCMe,--C( OH)*CO,H(111, stable.)I >oThe speeds of alkaline hydrolysis of substituted malonic esterwere found by R. Gane and C. K. Ingold2 to be in the followingdescending order : cyclopropane- 1 : l-dicarboxylic ester > cyclo-butane-1 : l-dicarboxylic ester>gem-dimethylmalonic ester>cycZo-hexane- 1 : 1-dicarboxylic ester>gem-diethylmalonic ester.Thisorder agrees in the main with the series of valency angles 8 computedfor groups R,C4 and C,>Cd.Other recent investigations of the valency-deflexion effect haveconcerned keto-lactol tautomeric mixtures, cyclic anhydrides,imides, and lac tone^.^The valency-deflexion hypothesis has been criticised by W.Huckel and defended by its a ~ t h o r s . ~ There shouId now be nodoubt that the work carried out with its aid has abundantly showna graduated alteration of the ease of closure and the stability ofrings by the presence, in the chain of carbon atoms from which therings are formed, of a group of the type R2C< or Cn>C<.Thecomputed valency angles do in general agree in respect of theirorder with the order deduced from experiments, but they cannot,as the authors admit, be regarded as having more than a qualitativesignificance.The contention that polar effects should not be admitted tosimultaneous consideration is unwarranted, for cases can be selectedwhich demonstrate the valency deflexion and polar effects inde-pendently, and their combined effects in more complicated caseshave been analysed with remarkable success. In the psubstitutedglutaric acids the polar effects of the substituents on the carboxylgroups are in general negligible, and the distances between the twoionising groups calculated from the first and second dissociationconstants are in the order : glutaric acid,>p-methylglutaric acid>J., 1926, 10; A., 1926, 249.M.Qudrat-i-Khuda, J., 1929, 1913; 1930, 206; A., 1929, 1273; 1930,471; E. H. Farmer and J. Kracovski, J., 1927, 680; A., 1927, 447; S. S. G.Sircar, J., 1927, 600, 1252, 1257; 1928, 898; A., 1927, 451, 756; 1928, 618.Fortschritte der Chernie, Physik und p?tysikalischen Chemie, 1927, 19, 4;J . , 1928, 1318; A., 1928, 1173.This VOI., p. 145154 BENNETT AND CHAPMAN :p -n-propylglutaric acid > p p'- dimethylglut aric acid > cyclopentane-1 : 1-diacetic acid> cycloheptane-1 : 1 -diacetic acid>cyclohexane-1 : l-diacetic acid> pp'-diethyl- and pp'-dipropyl-glutaric acids.6This is not only in general agreement with the valency-deflexionhypothesis, but in detail reproduces minor peculiarities previouslydiscovered by experiments on the closure and stability of rings.The influence of the cyclohexane ring on the valency angle appears,a t first sight, to constitute it serious anomaly.That ring wasoriginally shown to confer added stability on a three-memberedring sharing a spiran carbon atom with it.' Such an effect had beenexpected, for the angle of 120" of a phne hexagonal ring shouldcause the external valency angle to fall to lO7a". But from ourpresent knowledge of strainless forms of rings of six or more carbonatoms, in which the angles presumably approximate to the tetra-hedral value of this result evidently needs furtherconsideration.An experimental discovery of the failure of a large ring to exertthe effect originally anticipated was announced for the case of theseven-membered ring by J.W. Baker and C. K. Ingold in 1923 8and the suggestion that the ring relieved its strain by becomingnon-planar was then made, independently of the announcement ofthe isolation of the fist static isomerides with fused non-planarrings.g Yet the influence of the six-carbon ring originally observedhas been repeatedly confirmed and the cycloheptane ring has con-sistently exerted an effect lying between those of the cyclopentaneand cyclohexane rings. This may be justified as follows : 10 somestrain is necessarily involved in the interconversion of the twopossible strainless forms of the cyclohexane ring, and, as isomericforms have not been isolated, it must be assumed that such aninterconversion is continually taking place in the course of molecularagitation and vibration, so that the average effect of the ring is thatof a partly strained and not a strainless form.If the cycloheptanering becomes strained by molecular vibration, this may be of lessduration than in the six-membered ring, so that the average strainmay be in the order 6-ring>7-ring>5-ring. This seems the morecredible from a consideration of the fact that movements involvingrotation round a bond may be more easy in the cycloheptane ring,so that an almost strainless interconversion of forms should occur.The valency-deflexion effect of the trans-decalin ring has been6 R. Gane and C. K. Ingold, J., 1928,2267; A., 1928,1083; compare C.H.Spiers and J. F. Thorpe, J., 1926,127, 638; A., 1926, ii, 395.Ann. Reporta, 1916, 12, 109.Ibid., 1923, 20, 103. Ibid., 1924, 21, 92.10 J. W. Baker, J., 1926,lZ7, 1680; A., 1926, i, 1277ORUA'NIC CHEMISTRY.-PART 11. 155investigated recently by K. A. N. Rao,ll who concludes from astudy of tram-decalin- p-spirocyclopropane derivatives that thedecalin ring exerts less influence on the valency angle than the freecyclohexane ring. The yield in ring-closure was the same as withcyclohexane and cycloheptane derivatives : the difference is deducedfrom observations of the disruption of the compounds by acid ofvarious strengths. The facts relating to a series of these spirancompounds and caronic acid are here tabulated for comparison :Mildea t conditionsAcid. for disruption.5% HCl at 200"5 yo HCl a t 200"10 yo HC1 at 240"20% HCl at 240"Stable to 20% HClCaronic acid ..... .. . ... .. . .. ... . ... .. . ... ..... .... .. . ... . . . . . . .. . ... .cycZoPentaneq+ocycZopropanedicarboxylic acid . . . . . .. . .Decalin-&vpirocycZopropanedicarboxylic acid . . . . . . . . . . . .cycZoHeptane8p~roc~cZopropanedicarboxyl~c acid . . .. . . . . .cycZoHex~neep~rocycZopropanedicarboxyl~c acid . . . . . . . . .a t 240'This evidence, as it stands, certainly seems to support the ideathat cyclohexane is more strained than decalin, but it is not decisive.The argument is weakened by the fact that only in the case ofcaronic acid was a definite product of disruption (terebic acid)isolated from these reactions.The fact that tram-decalin-2 : 2'-di-acetic acid was found to be stable under the conditions of the dis-ruption experiment cannot be held to exclude entirely the possibilitythat the decalin nucleus of the spiran is broken, for a condition ofstrain is no doubt imposed upon it by the other rings.Such experiments would in any case afford very indirect evidenceof the valency-deflexion effects, whereas the calculation from thetwo dissociation constants of the substituted glutaric acids is asdirect a confirmation as could be desired.It may be pointed out, in conclusion, that many chemists willconsider it simpler to explain a number of the effects which havebeen studied under this heading as due to steric hindrance ratherthan to a secondary consequence of a valency deflexion.Thisapplies particularly to the experiments on the hydrolysis of estersand the disruption of lactone and imide rings, in which the generalscreening effects of the substituents may be the principal factor.The poly-merisation of mono- and di-chloroacetaldehydes under the influenceof acids yields, in addition to amorphous meta-aldehydes, probablyof high molecular weight, crystalline trimeric para-aldehydes towhich a six-atom ring structure may be given. From chloral, onlypolymerides of the meta type were known, but F. D. Chattaway andE. G. Kellett 12 have obtained by the action of cold sulphuric acid11 J . , 1929, 1964; 1930, 1164; A., 1929, 1297; 1930, 914.11 J., 1928, 2709; 1929, 2908; A., 1928, 1367; 1930, 194.'Stereoisomerism of Ring Compounds.-Parachlorals156 BENNETT AND CHAPMAN :a mixture of two crystalline parachlorals.They were separatedand are no doubt cis- and trans-isomerides, (I) and (11), as arethe two parabutylchlorals similarly prepared from butyl chloralCH,*CHCL*CCl,*CHO. R ? 7cjc13s vr3 R o H O R A\ R O H O H /?\\I / \I g-o--F Q-O-t! HI / F-0--QH H H R HThe analogous trichloralimides (CCl,*CH:NH), have long beenknown.13 The action of sulphuric acid on chloralsulphydrateyielded a mixture of trithioparachloral (one of the two possibleisomerides) and two dithioparachlorals. Three isomerides arepossible (111-V), of which (V) is potentially resolvable. The(1.) (11.) (111.)HAH H CCl, H H H1(IV.) (V.1 (VI.)behaviour of these substances gives some information as to theirconfiguration.Alcoholic potassium acetate removes the elements of hydrogenchloride from trithioparachloral t o give 2 : 4 : 6-trisdichloromethyl-ene-1 : 3 : 5-trithian, CC1,:C >S , and the two dithiopnra-chlorals furnish two distinct bistrichloromethylmonodichloro-methylene compounds, the trichloromethyl group adjacent to bothsulphur atoms being attacked preferentially.The latter productsmust therefore be (VI) and (VII), from which it is clear that either/s-c:cCl2\s--c:ccL,c y,s' '"p cCl,:c(-y:l' CCI,.CCl< S-CCI*CCI, >oT--o----q s-c:cc1, S-CCl*CCI,CCI, H(VII.) (VIII.) (IX.)13 A. BBhal and E. Choay, Ann. Chim. Phys., 1892, 26, 34; Compt.rend.,1890,110, 1270; A., 1890, 1093ORGANIC CHEMT3TRY.-PART 11. 157(111) or (IV) must be the missing third isomeride, as these two wouldyield the same dichloromethylene derivative (VI). One of the twodithioparachlorals isolated is therefore the dl-compound (V).These substances are converted by alcoholic potassium cyanideinto the trisdichloromethylene compound (VIII), the isomerismvanishing. Chlorine combines with this compound to produce thetrichlorotristrichloromethyl derivative (IX), but once again twoisomerides were found in place of the three theoretically possible.In continuation of the studyof the cis-trans isomerism arising from the non-planar configurationof the sulphoxide and related groups l4 several further pairs ofderivatives of 1 : 4-dithian have been described.15 Cautiousoxidation of dithian yielded a monosulphoxide, from which twoisomeric dithian monoxide methylsulphonium salts (I) and twoSulphoxides and elated substances.CH,*CH CH,*CH, 06< '>gMe}X Oi< .y*%*SO,*C,H,(1.) (11.1CH,GH, CH, CH,cis and trans. ci8 and trans.monoxidesulphilimines (11) were prepared. Moreover the bis-sulphilimine of dithian (111) l6 was separated into two distinctisomerides by crystallisation from solvents of high boiling point.CH,. CH,Cq*CH,cis and trans.(111.1 C,H7-SOz*%*6< y*N*S0,*C7H,A closely analogous case is that of the sulphoxides of penthianols,which occur in isomeric pairs of the type (IV).17The oxidation of phenylpenthianol (R = Ph), benzylpenthianol(R = CKPh), and penthianolcarboxylic acid (R = CO,H) gave ineach case two distinct isomerides.The methylsulphonium salt andsulphilimine of phenylpenthianol were also examined : the formerwas separated into cis- and trans-forms which are similar in typel4 Ann. Reporb, 1927, 24, 103.l6 E. V. Bell and G. M. Bennett, J., 1928, 86; A., 1928, 299.16 F. G. Mann and (Sir) W. J. Pope, J., 1922,121, 1052.l7 G. M. Bennett and W. B. Waddington, J., 1999,2832; A , , 219158 BENNETT AND CHAPMAN :to the isomeric 4-substituted piperidinium salts; l8 but a secondisomeric sulphilimine could not be found.to give rise to a single dioxide and a single crystalline trioxide(disregarding a further amorphous oxidation product having thecharacteristics of a substance of high molecular weight).A studyof the regulated oxidation of trithian has now not only led to theisolation of all the theoretically possible mono-, di-, and tri-sulph-oxides derived from it but also provided convincing evidence as totheir configurations.2° The results are summarised in the annexedscheme of oxidations :Trithian monoxideHitherto trimethylene trisulphide (trithian) has been known-~ I- 8-dioxide (3 times asa-dioxide (dimorphous). soluble as a-dioxide).LG I /’ 4 0 -0 CH, CH, 6 C d i C H 2 0S-CH,-S S-CH,---Y- 8-trioxide (10 times as 0I/ \+ I/ \I+ + fa-trioxide. soluble as a-trioxide).The compounds previously known were the @-dioxide and thea-trioxide. The configurations of the dioxides are fixed by thefact that one is oxidised to a single trioxide and the other to bothtrioxides. Moreover the common product of oxidation of bothdioxides is necessarily the tram-trioxide, and the other trioxide thecis-isomeride.The same configurations would have been assignedto these isomerides on the ground of their relative solubilities inwater, and this justifies the configurations previously allocated tothe dithian dioxides.,lIs Ann. Reports, 1927, 24, 102.lD 0. Hinsberg, J . p r . Chem., 1912, [ii], 85, 337; 1913, [ii], 88, 49; A.,1912, i, 546; 1913, i, 818.E. V. Bell and G. M. Bennett, J., 1929, 16; A., 1929, 293.41 Ann. Report8, 1927, 24, 103ORGANIC CHEMISTRY. -PART 11. 159The view that the supposed third isomeride of trithioacetaldehydeis a mixture of the a- and b-compounds has been confirmed.22 Theoxidation of these trisulphides has been re-examined Z3 and, althoughthe numerous possible isomeric sulphoxides were not all isolated,the results supply a definite proof that the configurations allocatedto the trithioacetaldehydes (011 physical grounds) must be reversed.The results are summarieed in the following table :Isomeride.Configuration.a-, m. p. 101" trans (hitherto cis),!I-, m. p. 126' cis (hitherto trans)MeIMe II C-Me Me Me Me ITwo tramkulphones.Oxidation products,expected. found.4 Monoxides 2 Monoxides2 Monosulphones 2 Sulphones2 Monoxides 1 Monoxide1 Monosulphone 1 SulphoneMe Ic--so2--4I I H HOne cis-sulphone.All three possible monosulphones were thus found, and the trithio-acetaldehyde which yielded only one of them must have the cis-configuration.cis-trans-Fwed rings.(1) Octahydronaphthtzlenes. Five isomericoctahydronaphthalenes should be possible (II-IV). Of theseA' cis and trans. Aa cis and tram.(11.1 (111.)A@.hydrocarbons, the trans-A2-isomeride was prepared and examinedby H. Leroux in 1910z4 and the cis-A2-isomeride by W. Borsche22 E. V. Bell, G. M. Bennett, and F. G. Maw, J., 1929,1462 ; A., 1929,1042.23 F. D. Chattaway and E. G. Kellett, J., 1930, 1362; A., 1022.s4 Ann. Chim., 1910, 21, 458; Compt. rend., 1910, 157, 384; A., 1910,ii,828160 BENNETT AND CHAPMAN :and E. Lange in 1923 25 by the removal of hydrogen chloride fromcis- P-chlorodecalin obtained from the decalol.These substancesare oxidised to trans- and cis-cyclohexanediacetic acids (V). Byheating the chlorination product of decalin with aniline, however,the same authors obtained a slightly impure specimen of the A1-octa-hydronaphthalene, presumably a mixture of isomerides, givingas principal oxidation product 1-carboxycyclohexane-2-propionicacid (I).Both cis- and tram-A*-isomerides may be obtained by thedehydration of the respective @-decalols with potassium bisulphate.26Attempts had been made to prepare the A9-isomeride by nitratingdecalin with boiling dilute nitric acid, reducing the resulting 9-nitro-decalin to the amino-compound, and decomposing this with nitrousacid.27 The researches of W. Hiickel and his assistants indicate thatthis method yields a mixture.28 The pure substance has now beenprepared and its structure carefully established.It may beobtained, by a complex reaction involving more than one rearrange-ment, by the dehydration of 2-cycZopentylcycZopentanol (VI) (cisor tram) with zinc chloride. It has also been separated from themixture of isomeric octahydronaphthalenes (produced by thedehydration of decalol with zinc chloride) in the form of its bluecrystalline nitrosochloride, and is regenerated from this by theaction of sodium methoxide.The structure of this octahydronaphthalene is confirmed by itsconversion by ozone in acetic acid into cyclodecane-1 : 6-dione (VII)and S-ketosebacic acid (VIII).In the course of a study of spiran form-ation in this series K.A. N. Rao 29 has realised a number of cases of(2) Demlin-p-spirans.25 Annalen, 1923, 434, 219; A., 1924, i, 32.es W. Huckel and H. Friedrich, Annalen, 1926, 451, 132; A,, 1927, 239;compare F. Eisenlohr and R. Polenske, Ber., 1924, 57, [ B ] , 1639; A., 1924, i,1291.27 S. Nametkin and others, Ber., 1926, 59, [B], 370; 1929, 62, [BJ, 1570;A., 1926,508; 1929, 921.28 W. Huckel, R. Danneel, A. Schwartz, and A. Gercke, Annalen, 1929,474,121 ; 477, 99; A , , 76,206.29 LOC. cit. (Ref. 11 ; p. 155)ORGANIC CHEMISTRY.-PBRT 11. 161X complicated isomerism. Substances of the type CgH,,>C<yderived from cis- and trans-decalins should occur in four isomericforms, two from each of the isomeric decalones. Two p-decalin-2 : 2’-diacetic acids (X = Y) were prepared, a single substance fromeach ketone, and each gave an anhydride.Their anilic acids, how-ever, (X and Y different) involve the additional isomerism and all fourforms of C g H 1 , > C < ~ ~ : ~ ~ & p h were separated, passing by lossof water into two anils, C,H16>C<g2:gg>NHPh. The decalin-p-spirocyclopropane-1 : 2-dicarboxylic acids and the derived lactonicCH-CO,Hwere also studied, and of the latter all acids CgH,,>four possible isomerides from one decalone were isolated.(3) Octcthydrohe~taquinolines. The variation in the proportionsof two stereoisomerides produced in different methods of reductionof heterocyclic bases was referred to in the last Report (Ann.Reports, 1929, 26, 165).In the light of this fact the reductionof tetrahydroheptaquinoline (I) has been repeated,m and it is nowfound that the use of sodium and alcohol produces a second isomeric<Ci*octahydro-base (11) inresulting knowledge ofaddition to that -&ready de~cribed.~~ Thethe properties of these substances has led tothe detection of this second isomeride in the product of reductionby tin and hydrochloric acid and in the reduction product of theketo-compound (111).Adirect demonstration of the enantiomorphism of a single allene asforeseen by J. H. van ’t Hoff is still lacking, and the view that thefour terminal valencies in allene are co-planar was recently advanced32on the ground of the homogeneity of a substance described as diethyldi-I-menthyl allenetetracarboxylate.It has been shown, however,that the substance is not an allene derivative at all.33Several new instances have been reported of the separation intoso S. G. P. Plant and R. J. Rosser, J., 1930, 1840; A., 1297.31 Ann. Reports, 1928, 25, 184.32 F. Faltis, J. Pirsch, and L. Bermann, Ber., 1930, 63, [B], 691 ; A., 678.3* C. K. Ingold and C. W. Shoppee, J., 1930, 1619; A., 1163.REP.-VOL. XXVII. PMolecular Dissymmetry.-Allene, spiran, and related types162 BENNETT AND CHAPMAN :optically active forms of a compound belonging to the fascinatinglysymmetrical class of spirans. The resolution of cyclobutanespiro-cyclobutane-1 : l-dicarboxylic acid (I) is due to H. J. Backer andH. B. J. Schurink,34 and the compounds (II-V) have been shownto be enantiomorphous by J.Boeseken and his assistant^.^^The potassium borotartrate isolated by T. M. Lowry is probablyof the same type.36The recorded rotatory powers of those of the above compoundswith spiran carbon as centre are very small. Apart from theprobability of ready racemisation, it may be noted that the opticalrotatory powers of substances of such complete symmetry andsaturation in the immediate vicinity of the central atom are to beexpected to be exceptionally small.A heterocyclic spiran of remarkably simple type has beenresolved by Sir W. J. Pope and J. B. Whit~orth.~7 The rotationof this spiro-5 : 5-dihydantoin (VI) is reversed in sign in ammoniacalsolution, probably owing to enolisation in salt formation.The announcement has been made of the resolution of the o-carb-oxyphenylhydrazone of methyltrimethylenedithiocarbonate (VII)by W.H. Mills and B. C. S a ~ n d e r s . ~ ~NH*COp/NH-TO CO*NH I \CO-NH C H 7 5 E 2 S (VII.)(VI-1 N\NH*C,H4*C02HThis new and clear evidence of the non-planar disposition of thevalencies of the doubly linked nitrogen atom is of particular34 Proc. K . Aicad. Wetenech. Amsterdam, 1928, 31, 370; A., 1928, 1134,36 Rec. trav. chim., 1926, 45, 919; Bey., 1929, 62, [ E l , 1310; A., 1927, 132;s6 J., 1929, 2863; A., 136. Compare E. Damnois, J . Chim. physique, 1930,87 Chem. and Ind., 1930, 49, 748.s8 Trans. Paraday SOC., 1930, 26, 431 ; A., 1096.1929, 791 ; compare Ann. Repwta, 1929, 26, 74.27, 179; A., 864OWANIC CHEMISTRY.-PART II.163importance owing to the possible alternative explanation of thedissymmetry of the pyridylhydrazone of cyclohexylene dithiocarbon-ate resolved in 1923.39Dissymmetry due to restricted rotation. This field of work con-tinues to attract considerable attention. The past year has seenthe extension of the phenomenon to a quinoline derivative and therealisation of the diastereoisomerism consequent on its occurrencetwice in the same molecule.A quaternary salt derived from 8- benzenesulphonylethylamino-quinoline has been resolved with the aid of bromocamphorsulphonicacid by W. H. Mills and J. G. Bre~kenridge,~~ the optically activeiodide (I) being obtained.In the diphenyl series the following have been resolved : 2 : 4-di-nitro-2’-methyIdiphenyl-6-carboxylic acid (corresponding acids withthe methyl group absent or in position 3’ could not be resolved),4O2 : 4 : 6 : 2’ : 4’ : 6’-hexanitrodiphenyl-3 : 3’-dicarboxylic acid, and2 : 4 : 6 : 2’ : 4‘-pentanitrodiphenyl-3-carboxylic acid.41 A generalreview of this type of isomerism has been given with an analysis ofthe possibilities of interference of various atoms and groups judgedfrom their dimensions as ascertained from X-ray crystal data.The inactive diastereoisomerides (11) and (111) have been isolatedby E.Browning and R. Adarn~,~2 which are oxidised to a commonquinone. The obstruction of rotation thus disappears with theremoval of two hydrogen atoms and the change of the centralnucleus to the quinonoid form.Stereoisomerism of Co-ordination Compounds.-Optical activitydependent on 6-co-ordinated copper has been realised by W.Wahl 43in the diethylenediaminediaquocupric salts, of which the active ionis estimated to have [MI = - 190”. Indications of a similaroptically active compound of 6-co-ordinated nickel were obtained.30 Ann. Reports, 1927, 24, 99.40 (Miss) M. S. Lesslie and E. E. Turner, J., 1930, 1768; A., 1287.4 1 R. Adam and others, J . Amer. Chem. Soc., 1930, 52, 1200, 2064, 2070,42 Ibicl., 1930, 52, 4098.‘8 Acta Sci. Fennicce Comrn. Phg8. Math., 1927, 4, 1 ; A., 1928, 396.2959, 447, 4628; A., 762, 911, 914, 1180164 BENNETT AND CHAPMAN :Structure of m p m n d s of 4-w-mdimted platinum. The questionof the planar or tetrahedral disposition of the valencies of 4-co-ordinated platinum was mentioned in the Report for 1927.44 Sincethat time the idea that the arrangement of the four valencies oftellurium is planar has been abandoned.45 F.G. Angell, H. D. K.Drew, and W. Wardlaw have now reinvestigated the two isomericcomplexes of platinous chloride with diethyl sulphide, described byBlomstrand in 1888, and regarded as of cis- and trans-planar con-figurations by A. Werner, and conclude that these are structuralisomerides and that there is consequently no need for a specialstereochemical hypothesis to account for them.46 Formulae (I) and(11) are now proposed :Et2Sh HpKcJ; ~ <sEt2****c1Et,S SEt, . . . . C1(I.) a-Dichloride. (11.) p-Dichloride.The a-dichloride is less polar, being insoluble in water; the p-iso-meride is appreciably soluble in water, the solution being conducting.By the action of moist silver oxide, the p-dichloride yields a strongbase, Pt(SEt,*OH),, from which an oxalate and other salts areobtained.The a-dichloride, on the other hand, is slowly butcompletely decomposed by silver oxide.The tetrahedral configuration for the or-dichloride is supported byexperiments on the addition of halogens to these substances, theresults of which are regarded as definitely inconsistent with Werner'sformulation. By implication doubt is also cast on the nature of theisomerism of the platinosammines.On the other hand the problem why the ionisable chlorine in thep-dichloride (11) does not rapidly co-ordinate with the platinumatom to yield the or-isomeride (I) remains to be elucidated.Moreover, a review by T.M. Lowry4' of the results of X-rayinvestigations of crystals of complex compounds from this point ofview suggests that caution is necessary in accepting the new formulae.This author contends that the anions in the tetragonal crystals ofthe salt K,PtCl, are undoubtedly of planar configuration as com-pared with the tetrahedral anions in the cubic K,Zn(CN),.48 Con-sequently the planar configuration for other complexes of 4-co-ordinated platinum should not be regarded as improbable.In view of the Stereochemistry of the atom w-ordimted to a metal.44 Ann. Reports, 1927, 24, 104.46 Ibid., 1929, 26, 80; this vol., p. 150.47 Proc. C a d . Phil. SOC., 1929, ZS, 219; A., 1929, 629.4a R.G. Dickinson, J . Arner. Chem. SOC., 1922, 44, 774, 2404; A., 1922,J . , 1930, 349; A., 669.i, 632 ; 1923, ii, 26ORGANIC CHEMISTRY .-PART II. 165discovery of optically active complexes of the metals by A. Werner,and of the fact that the distinction between principal and auxiliaryvalencies has been abandoned for some time, the stereochemicalreality of the co-ordination bond might be regarded as obvious.Yet it is only in recent years that attention has been paid to thesteric environment of the aon-metallic atoms such as nitrogen andsulphur through which co-ordination to a metal takes place.J. Meisenheimer in 1924a9 obtained evidence of the existence oftwo optically active dias tereoisomerides of the sarc osinedie t hy lene -diaminecobalt salts (I) and accounted for their existence by thesupposition that in this case the nitrogen atom forms a second centreof dissymmetry in addition to the cobalt atom. It may be pointedout, incidentally, that if this explanation is correct the compound isan unusual one in another respect, for a hydrogen atom is attachedto nitrogen, whereas the optical resolution of compounds of the type[NHR1R2R3]X has not been observed.The stereochemistry of thenitrogen and sulphur atoms in certain complexes has also beendiscussed by H. Reihlex~.~~J . xzThe first case of optical activity solely due t o this cause has beendescribed recently by F. G. Mann.61 The complex (11) resistedresolution, but when the platinum was oxidised to the 6-co-ordinatedcondition the resulting complex was successfully separated with theaid of camphor-10-sulphonic acid and the salt (111) was obtainedhaving [M],,,, + 1110’.49 Annden, 1924,438,217 ; A,, 1924, i, 1036.60 Z.a w g . Chem., 1926,151, 71; Annalen, 1926,447,211; 448,312; A.,61 J., 1930,1746; A., 1404.1926, 467, 699, 888166 BENNETT AND CHAPMAN :The dissymmetry is here centred in the co-ordinated sulphur atomand arises from the non-planar distribution of its three bonds,which are disposed as in sulphonium salts and sulphoxides. Anexamination of some chelate complexes of the type of (IV) for thepresence of the expected pairs of diastereoisomerides (here cis andtrans with respect to the heterocyclic ring) was unsuccessful.52This corresponds with the failure of the resolution of the platinouscomplex (11) and may be attributed to the instability of the co-ordination bond, with consequent ready interconversion of theisomerides.Natural Products ,53Owing to pressure of space an account will be given under thisheading of three subjects only, namely, the chemistry of angustione,carotene, and santonin.Other topics are held over until next year,when it is hoped to deal with recent work on the sterols and bileacids.Angzlstione.-A naturally occurring p-diketone has been foundfor the first time in angustione from the essential oil of Backhousiaangustifolia. 54 The substance, which has the composition C,,H,,O,,shows an intense ferric chloride coloration, readily yields a copperderivative, and by the action of ammonia an amino-compound.The action of alcoholic potassium hydroxide at 150" causes fissioninto acetic acid and a diketone, C,H1,O,.This is of the substituteddihydroresorcinol class, yields my-trimethylglutaric acid on treat-ment with sodium hypobromite, and is therefore 1 : 1 : 3-trimethyl-cyclohexane-4 : 6-dione (I). The unsaturated compound (11),obtained by oxidation of (I) or of angustione with ferric chloride,is itself oxidised to dimethylmalonic acid and is converted byphosphorus trichloride into 4 : 6-dichloro-1 : 2 : 3-trimethylbenzene.55Thereis found with it a second diketonic substance, dehydroangustione(IV), which is transformed into the unsaturated diketone (11) by theaction of 50% sulphuric acid.This unusual structure for dehydro-angustione has been confirmed by its oxidation to acty-trimethyl-glutaconic acid 55a and a p-hydroxy-my-trimethylglutaric acid (V).m G. M. Bennett, A. N. Mosses, and F. S. Statham, J . , 1930, 1668; A . ,1432.63 Mention should have been made in the last Report (p. 146) of the im-portant addition to C. A. Kern's synthesis of norpinic acid made by C. W.Shoppee and J. L. Simonsen (Chem. and Ind., 1929, 48, 730), who convertedthe synthetic trane-acid into the cis-acid identical with that from pinene.64 A. R. Penfold, J. Proc. Roy. SOC. New South Walee, 1923, 57, 300; C. S.Gibson, A. R. Penfold, and J. L. Simonsen, J., 1930, 1184; A,, 1924, i, 1328;1930, 921.Angustione is therefore a triketone of the formula (111).5 5 Compare Ann.Reports, 1906, 3, 122. 66a Priva?e communicationORGAXIC CHEMISTRY.-PART II. 167These substances thus contain the “ionone” ring which hasrecently been found in carotene (see below). They possess a carbonskeleton which evidently is not a simple multiple of the isopreneunit, but the problem of their genesis is an interesting one and itmay perhaps be assumed that at least two molecules of isoprene areinvolved.HF/CO,HCO CH-0 C0,H CH*OH\ /( 3 5 3 2CO CO-H (IV.) (V.)Carotene and Lycopene.-Recent biochemical work in connexionwith vitamin4 56 having shown the importance of carotene, theelaboration of a formula for this substance by P. Karrer and hisassociates is of exceptional interest.Carotene is a yellow hydrocarbon, C40H56, found in the carrot andin the leaves and seeds of many plants ; it is the colouring matter ofbutter.The substance was isolated with chlorophyll from stingingnettles by R. Willstatter and W. Mieg in 1907.57 The isomericlycopene (which also gives the colour reaction for vitamin-A) wasobtained from tomatoes and rose hips.58 The latter hydrocarbonis aliphatic, for complete hydrogenation converts it into the paraffinC40H82.59 The formulaCMe,:CH*CH,.CH,.CMe:CH[CH:CH*CMe:CH],CH,*CH,*CMe:CHMewas proposed for lycopene 60 and received some justification fromoxidation experiments.m Ann. RepoTt8, 1929, M, 245; H. N. Green and E. Mellanby, Brit. J .Exp. Path., 1930, 11, 81.67 Anrtalen, 1907, 355, 1 ; A., 1907, i, 865.6 8 R.Willstiitter and H. H. Escher, 2. phy8i0l. Chem., 1910, 64, 52 ; H. H.6B P. Karrer and R. Widmer, Helv. Chim. Acta, 1928, 11, 751; A., 1928,60 P. Karrer and W. E. Bachmann, ibid., 1929, 12, 286; A., 1929, 669.Escher, Helv. Chim. Acta, 1928, 11, 762; A., 1910, i, 330; 1928, 1016.1016.P. Karrer, A. Helfenstein, and H. Wehrli, ibid., 1930, 13, 87; A., 333168 BENNETT AND CHAPMAN :Catalytic hydrogenation and other tests showed that the carotenemolecule contained eleven double bonds with three (or two) lessreactive than the others,61 the ultimate reduction product being ofcomposition C,,H7,. It is therefore dicyclic.The nature of one ring in carotene was indicated by the productionfrom it by the action of cold permanganate of a substance identicalwith ionone and yielding, like ionone, as-dimethylsuccinic acid onoxidation.62 It follows that the molecule contains the trimethyl-cyclohexene ring of ionone, and the partial formula (I) was adopted.,CH,-CMe, ,CH,-CMe,CH, >C[ CH : CH CMe : CHI qC 11 H 1, CH, \CO,H\CH2-CMe \CH,-COMe(1.1 (11.)In the latest memoir 63 the isolation of other oxidation products isdescribed, including geronic acid (11), confirming the presence of theionone ring.As the second ring of carotene, contained in the residueC11H17 in formula (I), is likely to have resulted from ring-closurein the lycopene molecule, this residue would (from the structure - -CH -CH assumed for lycopene) be of the form -CH:CH*CMe<cM~:CM~>cH2.But none of the expected oxidation products of such a ring isdetected, nor is carotene optically active as it would then pre-sumably be.A careful study of the proportions of acetic acidresulting from oxidation of lycopene and carotene with chromicacid moreover shows that each yields 6 mols. of acetic acid per mol.,whereas these formulse require 8 and 7 mols. respectively (one foreach C-CMe:C group).These difficulties are now surmounted by adopting the formula(111) and (IV) for the two hydrocarbons :EH[ CH:CMe*CH:CH],CH:CMe[CH,],CH:CMe,(111.) CH[CH:CMeCH:CH],CH:CMe[CH,I2CH:CMe,,CELyCMe, CMe,*CH,%€&ie UV. 1 \CMe-CH2C[CH:CH*CMe:CH],CH:CH[CH:CMe*CH:CH],d >CH2The second ring in (IV), being equally an ionone ring, gives rise tono other oxidation products.It is noticeable that the formula no longer represent products of6 1 R.Pummerer and L. Rebmann, Ber., 1928, 61, [B], 1099 ; W. Reindel,ibM., 1929, 62, [B], 1411; L. Zechmeister and others, ibid., 1928, 61, [B],666, 1634; 1929, 62, [B], 2232; A., 1928, 624, 766, 1016; 1929, 906, 1306.62 P. Karrer and A. Helfenstein, Hdv. Chim. Acta, 1929, 12, 1142; A,, 76.68 P. Karrer, A. Helfenstein, H. Wehrli, and A. Wettstein, ibid., 1930,13,1084 ; A., 1422ORGANIC CHEMISTRY.-PART II. 169continuous polymerisation of isoprene, but it is suggested that thelycopene carbon chain may arise by the union of two molecules ofphytol aldehyde. Phytol, the alcohol of chlorophyll, has recentlybeen identified as 3 : 7 : 11 : 15-tetramethyl-A2-hexadecen-1-01, theconstitution having been codrmed by a synthesis from +ionone.84The corresponding aldehyde,CHMe,[CH,],CHMe[CH,],CHMe[CH,],CMe:CH*CHO,by a benzoin condensation or a pinacol reduction would give therequired carbon chain and dehydrogenation would lead to lycopeneThis argument has led the authors to suggest a corresponding(111).alteration in the formula for ~ q u a l e n e , ~ ~ which would be{CMe,:CH[CH,],CMe:CH*CH,*CH,*CMe:CH*CH,~},.The ketone C19H,,0 obtained by oxidation of partlyreduced squaleneshould therefore be 2 : 6 : 10-trimethylhexadecan-15-one and notthe 3 : 7 : ll-isomeride.The former has been synthesised andappears to be identical with the ketone from squalene.8antonin.-This substance, which is the chief active constituentof wormseed, has long been known to be a lactone derived froma reduced naphthalene molecule.Its chemistry was developedparticularly by S. Cannizzaro and his pupils and the formula (I) wasgenerally accepted.Santonin is ketonic but is converted by concentrated hydrochloricacid into a phenol, desmotroposantonin, hitherto written as (11), anddrastic reduction yields santonous acid (111).Me CH,CH-CHM HO(~$Ef-CHMeyo $JyJ CH-aco Y5tCH; Me 0Me CH, (I.)The question of the constitution of santonin has been reopenedby G. R. Clemo, R. D. Haworth, and E. Walton,66 who consideredthat the conversion into desmotroposantonin could not be merelya keto-enolic change, and that the santonin skeleton would moreprobably be such as could be derived by the union of isoprene units.The structure (IV) was suggested as meeting these and other points-in particular the observations of A.Angeli and L. mar in^,^^ whoobtained by oxidation a heptanetetracarboxylic acid, apparentlypossessing a quaternary carbon atom, to which formula (XX) may64 F. H. Fischer and K. Lowenberg, Annalen, 1928, 464, 69; 1929, 476,183; A., 1928, 989; 1929, 1421.66 J., 1929,2368; 1930,1110,2679; A., 1929,1464; 1930,919.67 Atti R. Accad. Lincei, 1907, 16, i, 159; Mem. A d . Lincei, 1908, 6,6s Ann. Reports, 1929,26, 90.386 ; A., 1907, i, 321 ; 1908, i, 643.F 170 ORGANIC CHEMISTRY.-PART 11.be assigned. The change to desmotroposantonin must on this viewinvolve the migration of a methyl group-for which, however, thereMe CH, CO,H C0,HHOTy&*CHMe*CO,H 70,H qH-CHMe \ IH02dfhH,/cH2\/\/Me CH,(111.) Me (XX.)is a close analogy in the conversion of 2 : 4-dimethylchinol (V) into2 : 5-dimethylquinol (VI).68CMe CH, Me MeHO/\ 11 IOH\/Me (IV.) (V.1 (VI-)MeMe06co/CH:CH*C02H M e 0 6 yH-1 (p2Et\CO*CH,Me (VII.) Me (VIII.)The position of the a-propionic acid side-chain was first fixed by asynthesis of santonous acid. Condensation of p-xylyl methyl etherwith maleic anhydride in presence of aluminium chloride gave theunsaturated acid (VII), which was converted by hydrogen chlorideand ethyl alcohol into the ester (VIII). The latter by couplingwith diethyl methylmalonate and hydrolysis yielded a mixture ofstereoisomeric acids (IX), which were reduced by Clemmensen’smethod to two acids of structure (X).The action of sulphuric acidcaused ring closure, and the product (XI) (obtained in an enol-lactonic form) by further reduction furnished dl-santonous acidmethyl ether, which was demethylated to the acid itself (111).MeO/) Me yH-CHMe*CO,H YO,H Me06,d!r FH--CHMe*CO,HbyCH. Me 0 Me H,(IX.) (X-)6* E. Bamberger and F. Brady, Ber., 1900,33,3642 ; A., 1901, i, 142ORGANIC CHEMISTRY.-PART m. 171This confirmed the accepted formula, for santonous acid, and thatfor desmotroposantonin was next proved by synthesis. One of theacids (X) yielded, when heated with hydriodic acid, the enol-lactone(XII), and sodium amalgam reduced this to desmotroposantonin(XIII). This formula differs from (11) only in the position of thelactone ring.Such an oblique arrangement of rings had beensuggested by S. Cannizzaro.69 It entails a readjustment of thesantonin formula to (XIV).Evidence was still needed of the position of the methyl groupregarded as migrating, and it was obtained as follows. Tetra-hydrosantonin, resulting from catalytic hydrogenation, was reducedby Clemmensen’s method to the fully saturated lactone, from whichby the action of selenium a hydrocarbon was formed which provedto be l-methyl-7-ethylnaphthalene ‘0 (XV).’O--T0 CH-CHMe f\/\\F,t IMe CH Me Me CH QqQH2 \A/HO,(~@~CX&CHMe ’T\./ \/Me CH,(XIII.) Me (XIV.) F V - 1The formula (XIV) for santonin has been accepted by L.R~zicka,~l who announced the degradation to l-methyl-7-ethyl-naphthalene a little before the other authors.Santonin is thus placed in the eudesmol (selinene) group 72 ofterpene compounds.G.M. BENNETT.A. W. CHAPMAN.PART III.-HETEROCYCLIC DIVISION.Oxygen Ring Compounds.AN account was given in last year’s Report of the preparation ofvarious flavones, flavonols, etc., by the acylation of o-hydroxy-acetophenone derivatives. This process has now been furtherapplied by A. Lovecy, R. Robinson, and S. Sugasawa2 to thesynthesis of luteolin 3’- and 4’-methyl ether (I and I1 respectively)by acting upon phloracetophenone with sodium O-benzylvanillateBer., 1893, 26, 786; A., 1893, i, 364.70 J. Harvey, I. M. Heilbron, and D. G. Wilkinson, J., 1930, 423; A., 593.L. Ruzicka and E. Eichenberger, Helv. Chim. Acta, 1930,13, 1117; A.,Ann. Reports, 1923, 20, 100; 1924, 21, 103.Ann.Reports, 1929, €36, 162.1442.a J., 1930, 817172 W T :and O-benzylvanillic anhydride, on the one hand, and sodiumO-benzylisovanillate and O-benzylisovanillic anhydride on theother, followed by the hydrolysis of the primary products in each0 OMe 0 OHcase. The latter flavone (11) was found to be identical with dios-metin, the rhamnoglucoside of which is diosmin, recently isolatedby 0. A. Oesterle and G. Wander.3 An examination of the reactionsof primetin, from PrimuZa modesta, has led W. Nagai and S. Hattori *to the conclusion that it is 5 : 6-dihydroxyflavone (111). Thepresence of an unsubstituted 2-phenyl group is indicated by theformation of benzoic acid on alkali fission, and the vicinal characterof the two hydroxyl groups is suggested by a green coloration withferric chloride and the readiness with which oxidation can occur.Primetin can give a diacetyl derivative, but methylation withdiazomethane or methyl iodide and alkali leads to a monomethylether, which gives a violet-brown colour with ferric chloride andcan be acetylated.The formula (111) for primetin is further supportedby an investigation of the absorption spectrum.c1F%H o e s HodzaoH HO CO (111.) RO (IV.)Some of the earlier syntheses of well-known anthocyanidinsthrough their methyl ethers have proved to be unsatisfactory, sincethe subsequent demethylation process led to impure products, andin more recent times they have been replaced in certain cases bymuch improved methods which avoid demethylation.These laterreactions have now been applied by W. Bradley, R. Robinson, andG. Schwarzenbach 6 to the preparation of delphinidin chloride(IV ; R = H). The interaction of 2-O-benzoylphloroglucinaldehydeH O P CHO A c t 6 AcO CO*CH,*OAc Ph2c<:o CO*CH,*OAcOBz Me0(V. ) W.) (VII.)8 Helv. Chim. Acta, 1926, 8, 619.ti See Ann. Reports, 1928,25, 163.Acta Phytochim., 1930, 5, 1; A., 704.J., 1930, 793ORGANIC CHEMISTRY.-PART III. 173(V) and w : 3 : 4 : 5-tetra-acetoxyacetophenone (VI) by means ofhydrogen chloride in alcohol-ethyl acetate yielded 5-O-benzoyl-delphinidin chloride (IV ; R = Bz), the acetyl groups being removedduring the reaction. Hydrolysis with aqueous-alcoholic sodiumhydroxide and subsequent t’reatment with hydrochloric acid thenc1m-\ ;~Qo~cH~~oA~ HO dzQp OMe M e O PCHOMe0 HO OH(VIII.) (X.1gave a product which was identical with delphinidin chloride fromnatural sources.A similar series of reactions with (V) and either(VII) or (VIII) led ultimately to 3’-O-methyldelphinidin chloride(IX), which very closely resembled natural petunidin chloride,although complete identity was not established, possibly owing tothe disturbing effects of traces of impurity in the natural product.The same authors have condensed 2-Q-benzoyl-4-O-methylphloro-glucinaldehyde (X) with w-acetoxy-4-benzyloxy-3 : Ei-dimethoxy-acetophenone (XI) by an analogous process, and, after hydrolysisof the resulting 5-O-benzoyl derivative, 7 : 3’ : 5’-O-trimethyl-OMec10 OMe r4delphinidin chloride (XII) was obtained.This substance proved tobe identical with llirsutidin chloride, recently isolated from Primulahirsuta,’ so that the structure of this anthocyanidin is now definitelyestablished.One of the most interesting problems associated with the naturallyoccurring oxygen ring compounds is the interconversion of thevarious types. Several processes of this nature have alreadyreceived considerable attention,8 and R. Robinson and G. Schwar-zenbachg have now made a study of the conversion of flavyliumsalts into flavones. Although this change is reminiscent of thefacile transformation of an alkylpyridinium salt to a pyridone, nosuccessful general procedure has hitherto been developed for it.7 P.Karrer and R. Widmer, Helv. Chim. Acta, 1927,10, 758.* See Ann. Repor& 1928,25, 169. 9 J., 1930,822174 PLANT :7-Hydroxy-4-carboxyflavylium betaine (XIII) has been oxidisedto 7-hydroxyflavone (XIV) by the use of chromic acid in acetic acid(XIII.) (XIV.) (XV.1solution,1° but the method cannot be applied in general to related4-carboxyflavylium salts or the corresponding betaines. It hasbeen found, however, that scutellarein tetramethyl ether (XV) canbe synthesised by an application of the Hofmann reaction to theappropriate acid amide of this type. The 4-carbamyl-5 : 6 : 7 : 4'-tetramethoxyflavylium chloride (XVI) resulting from the interactionof anisoylpyruvamide (XVII) , antiarol (3 : 4 : 5-trimethoxyphenol) ,and hydrogen chloride was submitted to a Hofmann reaction andyielded a pseudo-base (XVIII) which was subsequently convertedinto the tetramethoxyflavone (XV) by treatment with boilingdilute aqueous sodium hydroxide.The authors point out that, ifan easily accessible route can be found for the preparation of thearoylpyruvamides, this might become a valuable method for thesynthesis of flavones. The precedifig preparation of scutellareintetramethyl ether is of added interest in view of the fact that theearlier synthesis of this flavone by G. Bargellini l1 involved a processwhich was not quite unambiguous. An alternative synthesis hasmore recently been described by F. Wessely and G. H. Moser.12The action of anisic anhydride and potassium anisate on 2 : 4-di-hydroxy-3 : 6-dimethoxyacetophenone (XIX) at 180-185", followedM e O @ - ~ O M e Ho(JL CO*CH,OMe PH Me0Me0 C:NH(XVIII.) (XIX.)lo C.Biilow and H. Wagner, Ber., 1903, 36, 1941.l1 Gazzetta, 1915, 45, 69. l2 Monatsh., 1930, 56, 97; A., 1295ORGANIC CHEMISTRY.-PART III. 175by hydrolysis of the primary product, resulted in the unexpectedformation of 5 : 7-dihydroxy-6 : 4’-dimethoxyflavone (XX), whichinvolved partial demethylation. The tetra-acetate of the5 : 6 : 7 : 4’-tetrahydroxyflavone obtained by the demethylation ofthe latter product proved to be identical with scutellarein tetra-acetate.There now appears to be some doubt whether the reduction offlavonols in acid solution can actually lead to anthocyanidins aspreviously believed.13 The product obtained from quercetin (XXI ;R = H), although it closely resembles cyanidin chloride, is regardedby T.Malkin and M. Nierenstein14 as having the constitution(XXII), and an analogous structure is assigned to the product ofthe reduction of rhamnetin (XXI; R = Me).n OHHO CO (XXI.)c10 OH rc?It was mentioned in last year’s Report l5 that +baptigenin hadbeen recognised as an isoflavone of the formula (XXIII). Theconstitution assigned to this product has now been confirmedsynthetically in two ways by E. Spath and E. Lederer.16 In thefirst route, the crude cyanohydrin of the substance (XXIV), whichwas obtained by condensing resorcinol and w-bromoacetopiperonein the presence of sodium hydroxide, was submitted to a Hoesch(XXIV.) (XXV.)13 R.Willstiitter and H. Mallison, Sitzungsber. K. Akad. Wiss. Berlin, 1914,l4 J . Amer. Chem. SOC., 1930, 52, 2864; A., 1189.l6 Ann, Reports, 1929,26, 156.769; A., 1914, i, 1081; A. Robertson and R. Robinson, J . , 1927, 2196.l6 Ber., 1930, 63, [B], 743; A , , 611176 PLANT :reaction, and the product, which undoubtedly contained the com-pound (XXV), was subsequently sublimed in a vacuum. Besidesthe unchanged ketone (XXIV), $-baptigenin, identical with thesubstance from natural sources, was obtained in this way. In thesecond method, $- baptigenetin (XXVI) , previously synthesisedfrom resorcinol and 3 : 4-methylenedioxyphenylacetonitrile by aHoesch reaction, was treated with ethyl formate and sodium, andthe product (XXVII) was heated with alcohol and fuming hydro-chloric acid ; after subsequent sublimation, $-baptigenin was againisolated.(XXVI.) (XXVII.)Using synthetical processes already well known in the chalkoneand flavanone series,17 J. Shinoda, S. Sato, and M. Kawagoe l8have prepared butein (XXVIII) by first condensing resorcinol and3 : 4-diethylcarbonatocinnamoyl chloride in nitrobenzene in thepresence of aluminium chloride, and then heating the product withaqueous potassium hydroxide. On treatment with boiling alcoholichydrochloric acid, butein was converted into the correspondingflavanone, butin (XXIX).OH 0 OH HOP . OH H O ~ ~ Z - ~ O W C H , CO*CH.CH(XXVIII.) co (XXIX.)Considerable attention has recently been directed towards theposition of the sugar residue in certain well-known glucosides.Thewidely occurring product phloridzin yields phloretin and glucose onhydrolysis, but, although the structure of phloretin has long beenestablished and it has been known that the glucose residue is attachedto the phloroglucinol nucleus, its exact location has been uncertain, y)&.H yyH CO R*H Meo()ggR*MeOH OH OMe(1.1 (11.) (111.)(X = C,HllO, ; R = *CH,*CH,C,H,*O*)See, e.g., J. Shinoda and S. Sato, J . Pharrn. SOC. Japan, 1928, 48, 109;Ann. Reports, 1928, 25, 169.l8 J . Pharm. SOC. Japan, 1929,49, 123; A., 1930, 93ORGANIC CHEMISTRY .-PBRT 111. 177the formula? (I) and (11) both being possible. This point has nowbeen settled by an examination of the product obtained by hydrolys-ing the fully methylated gluc0side.1~ This product must have thestructure (111), since, on heating with acetic anhydride and sodiumacetate, acetylation is accompanied by ring closure to give a sub-stance which is either a coumarin (Tv) or a chromone (V), a processwhich is dependent upon the presence of a free hydroxyl group inthe 2-position with respect to the carbonyl group.Furthermore,the compound (111) has also been prepared by the catalytic hydro-genation of 2-hydroxy-4 : 6-dimethoxyphenyl p-methoxystyrylketone. Phloridzin, therefore, has the constitution (I). Previousexperience with analogous reactions indicates that the acylationproduct of the substance (111) is almost certainly the chromone (V),and this has recently been codkmed by J.Shinoda and T. Sato 2ofrom a study of the chemical behaviour of the compound.The glucoside obtained from p-acetobromoglucose anddaphnetin21 (VI) has been shown to have the glucose residue inthe 8-position,22 and S. Hattori23 has now proved that it is notidentical with daphnin, the naturally occurring glucoside ofdaphnetin. Furthermore, the methyldaphnetin obtained from thesynthetic glucoside by methylation and subsequent hydrolysis wasfound to be different from the methyldaphnetin obtained by asimilar process from d a ~ h n i n . ~ ~ It follows that dnphnin must havethe glucose residue in the 7-position.HO 0 HO 0 0(VI.1 (VII.) (VIII.)19 F. Wessely and K. Sturm, Monatsh., 1929, 53 and 54, 554; A., 1929,1452; F. R. Johnson end A.Robertson, J., 1930, 21.2o J . Pharrn. SOC. Japan, 1930, 50, 32; Chem. Zentr., 1930, ii, 404.a1 P. Leone, Gazzetta, 1925, 55, 674; A., 1926, 75.2a F. Wessely and I(. Sturm, Ber., 1929, 62, [B], 115; A., 1929, 298.98 J . Pharrn. Soc. Japan, 1930, 50, 82.24 See also F, Wessely and K. Sturm, Ber., 1930, 68, [B], 1299178 PLANT :Fraxin, the naturally occurring glucoside of fraxetin (VII), hasbeen shown by F. Wessely and E. Demmer 25 from a series of alkyl-ation experiments to have the glucose residue in the 8-position,for it is converted by successive methylation, hydrolysis, andethylation into 6 : 7-dimethoxy-8-ethoxycoumarin, identical withthe product obtained from 7-methoxy-8-ethoxycoumarin byoxidation 26 and subsequent methylation.Although it has long been known that mculin yields glucose andaesculetin on hydrolysis, and that the latter is 6 : 7-dihydroxy-coumarin (VIII; R = H), the exact location of the glucose residuehas again hitherto remained uncertain.F. S. H. Head and A.Robertson 27 have now shown that the glucosidoxy-group is in the6-position (VIII ; R = C,HI10,). By successive methylation,hydrolysis, and ethylation, the glucoside yielded a methoxyethoxy-coumarin, which was converted into the methyl ester of 2 : 4-di-methoxy-5-ethoxycinnamic acid by the subsequent action of methylsulphate and sodium hydroxide. The structure of the acid derivedfrom this ester on hydrolysis was confirmed by synthesis.Mention may be made of several interesting observations in thechemistry of other oxygen ring compounds not directly related tonatural products.An investigation by R. E. Lutz 28 into thereduction products of several unsaturated 1 : 4-diketones containingthe group O:C*C:C*C:O has shown that they may consist of thecorresponding saturated diketone, the corresponding furan deriv-ative, or a mixture of these two, according to the conditionsemployed. The fact that the saturated 1 : 4-diketones are them-selves unchanged by the conditions which lead to the furans suggeststhat the course of the formation of the furan ring involves first1 : 6-addition of hydrogen to give the group HO*C:C*C:C-OH, fol-lowed by the elimination of water.An interesting isomeric change has been brought to light byT. Rei~hstein,~~ who has shown that the action of aqueous potassiumcyanide on 2-chloromethylfuran leads essentially to 5-cyano-2-methylfuran (I), only a small quantity of the corresponding 2-cyano-methylfuran (11) being formed. Hydrolysis of the mixed cyanidesgave the analogous 2-methylfuran-5-carboxylic acid and furan-2-acetic acid.Zb Ber., 1929, 62, [ B ] , 120; A., 1929, 298.2 6 Compare G.Bargellini, Guzzettu, 1916, 46, 249; A., 1916, i, 489.2' J., 1930, 2434.t8 J . Amer. Chern. SOC., 1929,51,3008; A., 1929, 1459.28 Ber., 1930, 63, [B], 749; A., 611; see also M. M. Runde, E. W. Scott,and J. R. Johnson, J . Arner. Chern. SOC., 1930, 62, 1284; A., 783ORaANIC CHEMISTRY.-PART III. 179T. Reichstein30 has also made a detailed study of the ease withwhich well-known aldehyde and ketone syntheses proceed in thefuran series. Furfuraldehyde was obtained from furan by theaction of anhydrous hydrocyanic acid and hydrogen chloride inether at -15" without any additional condensing agent, the mixturebeing subsequently warmed to room temperature and the productEH-GH\/CH C*CH,*OMe RH-GH CH C*CH,.CN R H - Pv CNG CMe0 Y 0Ydecomposed with water.Similar experimental conditions resultedin the introduction of the a-aldehydo-group into several 2-alkyl-furans, but the procedure failed when applied to 2-acetylfuran, ethylpyromucate, coumarone, furfuryl methyl ether (111), and difurfurylether. These coiiditions also constituted a satisfactory method forthe conversion of 1-alkylpyrroles into the corresponding l-alkyl-pyrrole-%aldehydes, but failed in the case of pyrrole itself, 2-acetyl-pyrrole, and pyrrole-2-carboxylic acid.The method also could notbe applied to the preparation of thiophen-2-aldehyde, although thiscould be accomplished when an additional condensing agent waspresent, as, for example, when a mixture of thiophen, anhydroushydrocyanic acid, and benzene was treated with aluminium chlorideand hydrogen chloride. The 2-acetyl derivative of furan wasprepared by the interaction of furan and acetyl chloride in benzeneat 0' in the presence of stannic chloride ; 2-acetyl-5-methylfuran alsowas obtained under similar conditions, but a better yield resultedfrom the use of zinc chloride and ether. The author has discussedthe applicability of these experimental conditions and of thosepreviously described by other workers in this field.It has been observed that 7-hydroxy-2 : 3-diphenylbenzo-y-pyrone (IV) and some closely related derivatives dissolve in hotaqueous alkaline solutions and give gels on cooling, but only smallchanges in the nature of the substituents can be made withoutdestroying this chara~teristic.~1 Owing to the possible value of afluorescent gel-forming substance in the examination of the structureof gels, W.Baker 32 has prepared 10-hydroxyphenanthraxanthone(V), which, although closely related to (IV), contains a phenanthrenenucleus and so might be expected to exhibit fluorescence. Theso Helv. Chim. Acta, 1930, 13, 345, 349, 356; A., 783, 787.W. Baker and R.Robinson, J., 1925, 127,1981; W. Baker and (Miss)F. M. Eastwood, J . , 1929, 2897.52 J., 1930, 261180 PLANT :preparation was accomplished by an application of Pschorr's phen-anthrene synthesis to the methyl ether of 7-hydroxy-3-phenyl-2-o-nitrophenylbenzo-y-pyrone (VI), but the product gave a gelwhich showed no marked fluorescence and was not particularlystable.Xulphur and Selenium Ring Compounds.The configurations of the two stereoisomeric forms of 2 : 6-di-phenylpenthian-4-one (I), isolated by F. Arndt, P. Nachtwey, andJ. P u s c ~ , ~ ~ have now been established by F. Arndt and E. Scha~der.~*The modification (A), m. p. 113-114", must be the cis-, since itgave two varieties of 2 : 6-diphenyl-4-methylpenthian-4-01 (11) ontreatment with magnesium methyl bromide.These latter wereco MeC-OH CH2/\QHz p 2A5 3 3 2 p 3 2A7H2 Q H 2PhCH CHPh Ph*CH CHPh PhCH CHPh\/ S v S v S(1.) (11.) (111.)dehydrated to the same 2 : 6-diphenyl-4-methyl-A3-penthiene (or2 : 6-diphenyl-4-methylenepenthian). The modification (B), m. p.87-88", on similar treatment, gave first an amorphous product andthen an isomeric diphenylmethylpenthiene. The two forms of (I)have been reduced by amalgamated zinc and hydrochloric acid tothe corresponding cis- and trans-modifications of 2 : 6-diphenyl-penthian (111). Penthian-4-one itself has been prepared by G. M.Bennett and L. V. D. Scorah 35 by the application of a Dieckmannreaction to ethyl p- thiodipropionate, S (CH,*CH,-CO,Et), , andhydrolysis of the resulting ester.Other interesting sulphur-containing rings are found in 1 : 3-di-thiolan (IV ; n = 2) and 1 : 3-dithian (IV ; n = 3).D. T. Gibson 36has described the preparation of the former, by distilling a mixtureof formaldehyde, sodium ethylene thiosulphate, and hydrochloric33 Ber., 1925, 58, [B], 1633; A., 1926, i, 1307.34 Ibid., 1930, 68, [ B ] , 313; A,, 612. 35 J., 1927, 194. 36 J . , 1930, 12ORGCBNIC C€CEMISTRY.-PART IlX. 181acid, and of the latter from trimethylene dibromide, sodium thio-sulphate, formaldehyde, and hydrochloric acid.(IV.) (VI.)In addition to cycloselenobutane and cycloselenopentane, men-tioned in last year's ReportY3' G. T. Morgan and F. H. Burstal138have now described the preparation of cycloselenopropane (V ;n = 1) and cycloselenohexane (V; n = 4) in small yield by theinteraction of sodium selenide with trimethylene dibromide andhexame th ylene di bromide re spec tively .These sub stances are notonly more difficult to prepare than their analogues, but they are alsocharacterised by a greater degree of instability and a tendency topolymerise. In fact, the main product formed during the prepar-ation of the former compound is a six-fold polymeride, and cyclo-selenohexane also is accompanied by a dimeride and a complexpolymeride. The new monomeric cyclic selenium compounds showthe property, common to those previously described, of combiningadditively with such reagents as the halogens and mercuric chloride.By an extension of other reactions used in the earlier work, hexa-methylene dibromide has been converted by the action of potassiumselenocyanate into hexamethylene diselenocyanate (VI), which wastransformed into cyclohexamethylene 1 : 8-diselenide (VII) ontreatment with alcoholic alkali. The latter decomposed, whenheated, to give 2-methylcycloselenopentane (VIII) .C€&*CH,*CH,*Se CH2*CHMe CH,*CH*CO,HCH2*CH2-CH2*ke I C.2< c)se CH, I ,c)."" HC0,H(IX.)CH2* -H2(VII.) (VIII .)A. Fredga 39 has prepared cis-tetrahydroselenophen-2 : 5-dicarb-oxylic acid (IX) by the action of potassium diselenide on sodiummeso-aa'dibromoadipate, and the corresponding trans-modificationby the action of potassium selenide on sodium dl-ad-dibromo-adipate. The trans-form, as expected, was found to be resolvableby brucine. Also of interest is the preparation by C.S. Gibson andCH2*CH, J. D. A. Johnson of 1 : 4-selenoxanY qCH eCH >Se, by the2 2action of sodium selenide on pp'-dichlorodiethyl -ether ; *O thissubstance, like those mentioned above, readily forms additionproducts.31 Ann. Reports, 1929, 26, 160.'0 J., 1931,266.38 J., 1930, 1497.J . pr. Chem., 1930, [ii], 127, 103; A,, 1196182 PLANT :Indole Derivatives.Some interesting observations in the isatin group have beenrecorded by J. M. Gulland, R. Robinson, J. Scott, and S. Thornley.*l5 : 6-Methylenedioxyisatin (I) undergoes ring fission, when treatedwith nitric acid, to give ox-6-nitro-3 : 4-methylenedioxyanilic acid(11). A similar reaction has been previously observed42 betweennitric acid and di(methy1enedioxy)indigotin. A new route for theproduction of isatin derivatives is found in the prolonged hydrolysis,with boiling dilute aqueous-alcoholic sodium hydroxide, of theazlactone (111) derived from the interaction of 2-nitroveratraldehydeand hippuric acid.The formation of 6 : 7-dimethoxyisatin in thisway provides another example of an interesting type of intra-molecular oxidation-reduction.S. G. P. Plant 43 has prepared 9-methyl- and 9-ethyl-1 : 2 : 3 : 4 : 5 : 6 : 7 : 8-octahydrocarbazole (IV) by the action ofthe appropriate alkylamines on 2 : 2'-diketodicycbhexyl (V) inglacial acetic acid. These substances proved to be identical withcompounds obtained by J. von Braun and H. Ritter44 by thecatalytic hydrogenation of the corresponding 9-alkylcarbazoles, anddifferent from the alkylation products of the base obtained by theremoval of ammonia from cyclohexylideneazinc 45 (VI).The lastreaction, which is reminiscent of the preparation of tetraphenyl-pyrrole from phenyl benzyl k e t a ~ i n e , ~ ~ does not, therefore, yield(IV; R = H), but an isomeric ocfahydrocarbazole, probably ofthe structure (VII).42 T. G. Jones and R. Robinson, J., 1917, 111, 908. 41 J . , 1929, 2924.43 J., 1930, 1595.45 W. H. Perkin and S. G. P. Plant, J., 1924, 125, 1603.4~3 (Mrs.) G. M. Robinson and R. Robinson, J., 1918,118, 639.44 Ber., 1922, 55, [B], 3792ORGANIC CHEMISTRY.-PART III. 183The system of nomenclature and numbering hitherto employedfor the various carbolines has been anomalous and unsatisfactory.It has now been suggested:' in consequence, that 3-, 4-, 5-, and6-carboline should be re-named a-, p-, y-, and 8-carboline respectively,and that the accompanying numbering (formula VIII ; p-carbolinebeing used for illustration), which is more conventional, should beapplied to all the four groups. In some recent papers these proposalshave been adopted.The reaction previously observed by R. H. F.H HManske and R. Robinson4* in which the decomposition of theazide (IX) of (3-3-indolylpropionic mid by hydrogen chloride inbenzene is accompanied by intramolecular condensation with theformation of 2-keto-2 : 3 : 4 : 5-tetrahydro-p-carboline (X; R = H)has now been applied to the preparation of 2-keto-7-methoxy- and4 2 - CON,NH.CH.,, CH,(IX.) (X-) (XI.12-keto-8-methoxy-tetrahydro-~-carboline.49 The latter (X ; R =MeO) was found to be identical with one of the substances obtainedby the hydrolysis of the product of the oxidation (by permanganate)of acetylharmaline, and conbation is consequently provided forthe constitution (XI) assigned to acetylharmaline.wAlthough the structures of harman, harmalipe, and harmine havebeen already amply confirmed by synthesis, 51 additional methodsdeveloped by E. Sprith and E. Lederer 52 are of interest, and areclosely related to well-known reactions in the isoquinoline series.The acetyl derivative of 3- p-aminoethylindole has been convertedby phosphoric oxide in boiling xylene into dihydroharman, whichwas subsequently dehydrogenated with spongy palladium a t 200"to give harman (XII).A similar process applied to 6-methoxy-3- (3-acetamidoethylindole led first to harmaline (XIII) and then to47 J. M. Gulland, R. Robinson, J. Scott, and S. Thornley, Eoc. cit.*@ IT. S. B. Barrett, (the late) W. H. Perkin, and R. Robinson, J., 1929,2942.6o H. Nishikawrt, W. H. Perkin, and R. Robinson, J., 1924,125, 657.61 See, e-g., Ann. Reports, 1927, 24, 160.63 Ber., 1930, 63, [B], 120; A,, 363.J., 1927, 240; Ann. Reports, 1927, 24, 161184 PLANT :harmine (XIV). The same authors 53 have extended these reactionsto other acyl derivatives of 3-P-aminoethylindole and certain of itssubstitution products for the purpose of preparing a large number ofp-carbolines.They have also investigated the condensation of(XII.) (XIII.) (XIV.)formaldehyde with some 3-p-aminoethylindoles, and the dehydro-genation of the resulting 2 : 3 : 4 : 5-tetrahydro-p-carbolines top-carbolines by palladium at 160-170". Further reactions of asimilar type have been studied by S. Akabori and K. S a i t ~ . ~ ~Tetrahydroharman (XV) was prepared by the condensation of3-p-aminoethylindole with a~etaldehyde,~~ and was dehydrogenatedto harman on being boiled in aqueous solution with maleic acid andpalladium-black for five hours ; an analogous procedure resulted inthe conversion of 6-methoxy-3-~-aminoethylindole into harmine.This method of dehydrogenation, which involves the catalytic trans-ference of hydrogen to an unsaturated compound, is of considerableinterest, and has been applied by S.Akabori and his co-workers 56to a number of other products, several different unsaturatedsubstances being used.Me(XV.) (XVI.) (XVII. )By reactions analogous to those which have been used for thesynthesis of y-~arboline,~~ W. 0. Kermack and J. F. Smith 68 haveobtained derivatives of 2 : 3-benz-y-carboline (XVI). For example,4-o-aminophenylamino-2-methylquinoline, from the interaction of4-chloro-2-methylquinoline and o-phenylenediamine, was convertedinto a triazole derivative (XVII), which subsequently gave 5-methyl-63 Ber., 1930, 63, [ B ] , 2102. 64 Ibid., p. 2245.66 Compare G. Trttsui, J . Pharm. SOC. Japan, 1928, 48, 92.Proc. Imp. Acad. Tokyo, 1929, 5, 255; 1930, 6, 236; A., 1929, 1170;R.Robinson and S. Thornley, J., 1924,125, 2169.1930, 1192.68 J., 1930, 1999ORGANIC CHEMIS!I’RY.-PBRT III. 1852 : 3-benz-y-carboline by loss of nitrogen, on heating in syrupyphosphoric acid. The same authors have also prepared 1 : 54%-methyl-2 : 3-benz-y-carboline by treating o-acetamidoacetophenonephenylmethylhydrazone (XVIII) with phosphoryl chloride in boilingtoluene. The methosulphate of this carboline is apparentlyidentical with the methosulphate of the anhydronium base (XIX)derived from the action of alkali on the methosulphate of &methyl-2 : 3-benz-y-carboline. This fact, together with the fluorescentproperties of these carbolines, confirms the structures assigned tothem. The anhydronium base (XIX) is analogous to theanhydronium bases derived from certain p-ca;rboline~,~g and itspreparation has an added interest in view of the behaviour of otherclosely rela%ed substances.Thus (Mrs.) G. M. Robinson 6O hasprepared 2 : 3-pyrrolo(4‘ : 5’)-quinolines (XX) by the dehydrationof 3-acylamidoquinaldines (XXI), a n extension of a reaction pre-viously used in the pyrindole series,61 and has found that the metho-sulphate of 2 : 3-(2’-phenylpyrrolo)(4‘ : 5’)-quinoline (XX; R = Ph)Me NHreadily yields an anhydronium base (XXII) with aqueous sodiumhydroxide. It has been noted, however, that quindoline metho-sulphate (XXIII) gives no analogous anhydronium base.62N NH-wS0,Me(=I*) (=I*) (XXIII.)Mentian may be made in this section of an interesting applicationof Fischer’s indole synthesis to the phenylhydrezone of penthian-A S &(XXIV.) p z p% CH, (XXV.1 CH, CH, NR CH,(* J., 1929, 2948.b6See Ann.Reports, 1928, 25, 176.61 E. Koenigs and A. Fulde, Ber., 1927, 60, [B], 2106. ‘* J. W. Armit and R. Robinson, J., 1922, 121, 827186 PLAXT :4-one (XXIV), which resulted in the formation of penthienoindoleQuinoline Derivatives.(XXV).Further interesting developments in the chemistry of the cyaninedyes have recently been recorded. With a view to the study ofsome physical properties, Miss F. M. Hamer 64 has investigated thepossibility of preparing examples of known classes of these dyes fromcertain more complex compounds of the quinaldine or lepidine type ,containing a reactive methyl group.From the interaction of themethiodide of 2-methylacenaphthppidine (I) and the quinolinealkyliodides, isocyanines of the formula (11) resulted, and, with the2-iodoquinoline alkyliodides, +cyanines of the constitution (111)(1.1 (11.) R"-Iwere obtained by the usual methods, but attempts to prepare acarbocyanine from it by the well-known procedure with ethyl ortho-formate and pyridine were unsuccessful. Although &methyl-acridine (IV) contains a reactive methyl group, attempts to preparea carbocyanine from its methiodide or to effect condensation withquinoline methiodide by the usual methods met with failure. How-ever, from the condensation of 5-methylacridine methiodide with2-iodoquinoline alkyliodides in aqueous potassium hydroxide dyeswere obtained to which the formula (V) has been attributed, buttheir physical properties were somewhat abnormal.An interesting and surprising new route for the production ofthiocyanines, which in several respects is superior to the older6s G.M. Bennett and W. B. Waddington, J., 1929, 2829. 64 J., 1930, 995ORGANIC CHEMISTRY.-PART III. 187methods, has been discovered by Miss N. I. Fisher and Miss F. M.Hamer.65 The process came to light during an attempt to applythe reaction by which 2-methylene-1 : 3 : 3-trialkylindolines (VI),or the corresponding indoleninium salts (VII), are converted intodyes of the type (VIII; A := CR,) 66 to the preparation of thecorresponding sulphur compounds (VIII; A = S). The action ofamyl nitrite, in the presence of acetic anhydride, on the alkylchloridesof 1 -methylbenzthiazole (IX) led unexpectedly to the thiocyanines(X). The course of this reaction is not obvious, although theauthors have put forward tentative suggestions.It is well knownthat it is very difficult to prepare cyanine and carbocyanine dyes(VIII. )(X.1from monocyclic compounds, although certain carbopyridine-cyanines have recently been prepared.67 It is not surprising,therefore, that the new method for the preparation of thiocyaninesfails when applied to the alkylchlorides of 2 : 4dimethylthiazole(XI). Our knowledge of the simpler types has, however, beenextended by these authors by the preparation of thiocarbocyaninesof the formula (XII) from the action of ethyl orthoformate andpyridine on the 2 : 4-dimethylthiazole alkyliodides.Some interesting observations have been made by G.R. Clemoand H. J. Johnson G8 during the course of experiments designed forthe purpose of synthesising 12 : 13-dimethoxyisoindenoquinoline(XIIIa and XIIIb). The condensation of 4-keto-1 : 2 : 3 : 4-tetra-hydroquinoline with veratraldehyde in alcoholic sodium hydroxidett6 J., 1930, 2602. 66 D.R.-P. 459,616; B.P. 291,888.67 See Ann. Reports, 1929, 26, 164. 68 J., 1930, 2133188 PLANT :led to 4-hydroxy-3-veratrylquinolinealternative routes which were explored(XIV), but the variousin order to effect conversionof this product into (XIII) failed to give the desired result. It wasfound, however, that, when these two substances were condensed inglacial acetic acid in the presence of hydrogen chloride, the reactionyielded 4-keto-3-veratrylidene-1 : 2 : 3 : 4-tetrahydroquinoline (XV ;R = H), which was easily transformed into (XIV) by the action ofOH coalcoholic alkali.By the use of 4-keto-l-acetyl-1 : 2 : 3 : 4-tetra-hydroquinoline this latter possibility was avoided and the product(XV; R = Ac) then gave 4-keto-l-acetyl-3-veratryl-1 : 2 : 3 : 4-tetrahydroquinoline (XVI) on catalytic reduction. The substance(XVI) was converted into (XIII) by the action of warm sulphuricacid (SO%), a process which involved deacetylation and oxidationin addition to ring closure. The product was obtained in two inter-convertible forms (one faintly yellow and the other reddish-brown)which apparently are represented by the structures (XIIIct) and(XIIIb) respectively.co(XVI.)T.R. Seshadri,G9 during an investigation of some aminoalkyl-quinolinium salts, has made the interesting observation that theaction of aqueous sodium hydroxide on p-aminoethylquinoliniumbromide hydrobromide (XVII) leads to a base for which the structure(XVIII) is proposed. S. Gabriel 7O suggested the constitution(XIX) for an analogous pyridine derivative, but this is very unlikely.Investigations by S. G. P. Plant and R. J. Roaser 71 into the6@ J., 1929, 2962. 70 Ber., 1920, 68, [B], 1986.7 1 J., 1929, 1861; 1930, 2444ORGANIC OHEMISTRY.-PART III. 189reduction of certain simple quinolines under various conditionshave shown that, although 2 : 3-dimethyl- and 2 : 4-dimethyl-quinozine readily give the two stereoisomeric modifications (cis- andtrans-) of 2 : 3-dimethyl- and 2 : 4-dimethyl-1 : 2 : 3 : 4-tetrahydro-quinoline respectively, the 3 : 4-dimethyl compound yields only oneof the two possible forms of its tetrahydro-derivative.Anexplanation of this fact can be found in the steric effect of the4-methyl group in the intermediate 3 : 4-dimethyl-1 : 4-&hydro-quinoline, which is to displace the 3-methyl group from its sym-metrical position with respect to the double bond. As a result ofthis, the subsequent addition of a hydrogen atom in the 3-positiontakes place in one only of the two possible directions. The actionCHMe EP Y ‘YH 0 V-YH NHCH,-CH,(XVIII.) (XIX.) (XX.1HCH,-CH,of sodium amalgam upon 2-keto-3 : 4-dimethyl-1 : 2-&hydro-quinoline in alcoholic solution led to the two forms of (XX), butfurther reduction of this mipture by sodium and alcohol resultedin the single modiiication of 3 : 4-dimethyl-1 : 2 : 3 : 4-tetrahydro-quinoline.This fact is almost certainly explained by the inter-mediate formation in the latter reaction of 3 : 4-dimethylquinoline,traces of which were found in the product. A similar explanationwill account for the formation of only one modification of2 : 3 : 4 : 5 : 6 : 13-hexahydro-a-quinindene (XXI) during the reduc-tion of a mixture of the two stereoisomeric forms of 5-keto-2 : 3 : 4 : 5 : 6 : 13-hexahydro-a-quinindene 72 (XXII).H,C-?H,IH,C--YH,CHMe*CHMeHN<CHMe*CHMe>NH(=I.) (XXII.) (XXIII.)Five geometrical isomerides are possible in the case of 2 : 3 : 5 : 6-tetramethylpiperazine (XXIII), and it is of interest to record thatB’.B. Kipping 73 has isolated four of these by reducing 2 : 3 : 5 : 6-Reports, 1929,28, 164.72 B. K. Blount, W. H. Perkin, and S. G. P. Plant, J., 1929, 1975; Ann.J., 1929, 28891 90 PLANT :tetramethylpyrazine under various conditions. Previously onlytwo had been definitely identified.Alkaloids.In a study of the de-alkylation of tertiary amines by heatingwith organic acids J. von Braun and K. Weissbach 74 have includedan investigation of some cyclic bases. 2-Methyltetrahydroiso-quinoline was converted into 2-benzoyl- and 2- p-phenylpropionyl-tetrahydroisoquinoline by the action of the appropriate acid, andtropane yielded benzoyl- and @-phenylpropionyl-nortropane undersimilar conditions.Nicotine, on like treatment, gave productscontaining acylnornicotines, from which nornicotine (I) was obtainedon hydrolysis.Quinoline Group .-Further investigations by E. Spath andJ. Pikl 75 into the nature of the products obtained from angosturabark have resulted in the separation of four additional bases fromthe low-boiling fraction, vix., quinoline, 2-methylquinoline, 2-keto-l-methyl-1 : 2-dihydroquinoline, and 2-n-amylquinoline. The isol-ation of such simple bases from this source is of considerable interest.(1.) (11.) (111.)Y. Asahina, T. Ohta, and M. Inubuse 76 have isolated fromSkimmia repens a base, C,,H,O,N, which was found to be identicalwith dictamnine, an alkaloid obtained from Dictamnus a t l b ~ ~ , ~ ~and, from a study of its reactions, the formula (11) has been sug-gested. Thus it was found to contain one methoxyl group, thesecond oxygen atom appeared to be of the ether type, and, onoxidation with potassium permanganate, the alkaloid gave analdehyde (dictamnal), C1,H,O,N, together with the correspondingacid (dictamnic acid), C,,H,O,N.The latter, on heating withconcentrated hydrochloric or hydrobromic acid, yielded 2 : 4-di-hydroxyquinoline with the loss of a methyl group and carbondioxide ; it was found, however, not to be identical with a syntheticalspecimen of 4-hydroxy-2-methoxyquinoline-3-carboxylic acid,78 soMonatsh., 1930, 55, 352; A., 1049; for earlier work, see Ann.Reports,76 Ber., 1930, 63, [B], 2045; A., 1454.7 7 H. Thorns, A., 1923, i, 639.74 Ber., 1930, 63, [B], 489, 2018; A., 458, 1444.1924, 21, 131 ; 1929, 26, 170.Prepared by a method analogous to that described by C. A. BLchoff,Annalen, 1889, 251, 360ORGANIC CHEMISTRY.-PART III. 191that it appears to be 2-hydroxy-4-methoxyquinoline-3-carboxylicacid (111).It has also been shown, by Y. Asahina and M. Inub~se,'~ thatthe alkaloid skimmianine, isolated by J. Honda 8O from SEimmiajaponica, has the molecular formula C,,H 1304N and is closely relatedto dictamnine. It contains three methoxyl groups, and, on oxid-ation with potassium permanganate, it yielded an aldehyde,C13Hl,0,N, and an acid, C13H13OGN.The acid, on heating withconcentrated hydrochloric acid, lost a methyl group and carbondioxide with the formation of a product which was shown syn-thetically to be 2 : 4-dihydroxy-7 : 8-dimethoxyquinoline (IV).OHThis synthesis was accomplished *l by condensing 2-nitro-3 : 4-di-methoxybenzoyl chloride with malonic ester and heating the productwith tin and alcoholic hydrochloric acid. Skimmianine, therefore,appears to be a dimethoxydictamnine of the constitution (V).isoQuinidine, which does not occur naturally, is obtained whenits isomeride, quinidine (itself a naturally occurring isomeride ofquinine), is dissolved in warm sulphuric acid. A. Konopnicki andJ. Suszko 82 amert that several of the reactions of isoquinidine areexplicable with the aid of formula (VI).It is not acted upon byacid chlorides, acetic anhydride, the Grignard reagent, phenyl-hydrazine, or semicarbazide. The bwe also forms a mono- and adi-methiodide, an oxide (with hydrogen peroxide), and two per-bromides, both of which give isoquinidine with dilute alkalis. Byheating isoquinidine sulphate alone, or the free base in 25% aceticacid, another isomeride, iaoquinicine, is produced. The latter is notketonic in characfer, but appears to contain an >NH group, whichcan be acetylated, nitrosated, and methylated.isoQuinoZine Croup.-A considerable amount of work dealingwith the chemistry of substances of the papaverine type has beenBer., 1930, 03, [B], 2062; A., 1464.81 Y. Asahina and S. Nakanishi, Ber., 1930,63, [B], 2067; A., 1445.Bull.Acad. Polonaise, 1929, A , 340; A,, 1930, 97.no A., 1906, i, 152192 PLANT :recorded during the year. By suitable modification of the well-known synthesis of Pictet and G a m ~ , ~ ~ J. s. Buck has preparedthe hitherto unknown 1 : 2-dihydropapaverine (IV). w-Homo-veratroylaminoacetoveratrone (I) was dehydrated by means ofphosphoryl chloride to give (11) ; this was catalytically reduced to(111), which was treated with phosphorus pentachloride in coldco M e O o z ( i ' H 2 CH-OHMeo \pCH2X Q CH2X(1.) (11.) (111.)CH CH CHTo CH2XCH2X co*x co*x(IV.) (V.) (VI.)(X = 3 : 4-dimethoxyphenyl)chloroform with the production of (IV). 1 : 2-Dihydropapaverinewas reduced catalytically to tetrahydropapaverine, and dehydrogen-ated to give papaverine by heating with palladium-black. Itreadily underwent oxidation by air to the ketone (V), and, whenheated with methyl-alcoholic potassium hydroxide, it yieldedpapaveraldine (VI). In the latter reactions 1 : 2-dihydropapaverineclosely resembles its 3 : 4-is0meride.8~I.men and J. S. Buck86 have investigated the possibility ofpreparing substances of the papaverine type by the application ofa reaction already used in the isoquinoline series,87 vix., fromderivatives of benzylaminoacetal by ring closure and oxidation.The oxime of deoxyveratroin (VII) was reduced with sodiumamalgam and alcoholic acetic acid to @-di-(3 : 4-dimethoxypheny1)-ethylamine (VIII), but the product derived from the condensationof this substance and bromoacetal was completely decomposed bytreatment with sulphuric acid and arsenic oxide in an attempt toprepare papaverine.Similar results were obtained with theanalogous di-methylenedioxy-derivative from deoxypiperoin.8s Ber., 1909, 42, 2943.84 J . Amer. Chem. SOC., 1930, 52, 3610; A., 1455.8 5 See J. S . Buck, R. D. Haworth, and W. H. Perkin, J., 1924, 125, 2176.86 J . Amer. Chem. SOC., 1930, 52, 310; A., 353.87 See, e.g., L. Rugheher and P. Schon, Ber., 1909, 42, 2374ORGANIC CHEMISTRY.-PART III. 193CH2E:EgH2 CH(VII.) (VIII.)Having in mind the possible route by which the papaverinealkaloids are formed in nature,88 E. Spiith and F. Berger 89 havestudied the condensation of homoveratrylamine with 3 : 4-dimethoxy -phenylacetaldehyde, and the conversion of the product (IX) intodl-tetrahydropapaverine. Ring closure was accomplished with19% hydrochloric acid, but the yield of the tetrahydro-base wassmall (8%).of the reactionsof coclaurine, an alkaloid from CocczcZus Zaurifoliw, has shown thatit is represented by the formula (X).The following were the mainobservations which led to this result and established the relativeAn investigation by H. Kondo and T. Kondopositions of the6 OHmethoxyl and hydroxyl groups. The metho-(XI-) (XII.)sulphate of the triethyl derivative of coclaurine gave a product(XI) by the Hofmann degradation process, which, on oxidation,yielded p-ethoxybenzoic acid and the acid (XII). The structureof the latter was established by a further application of the Hof-mann process to give 4-methoxy-3-ethoxy-6-vinylbenzoic acid, and,on subsequent reduction, 4-met~hoxy-3-ethoxy-6-ethylbenzoic acid,the identity of which was confirmed by synthesis.Furthermore,the dehydration of p-methoxyphenylaceto-P-3 : Q-dimethoxyphenyl-88 See R. Robinson, J . , 1917,111, 876.88 Ber., 1930, 63, [B], 2098; A., 1454.00 J. pr. Chem., 1930, [ii], 126, 24; A., 794.REP.-VOL. XXVII. 194 PLANT :ethylamide (XIII) gave an isoquinoline derivative, which, onreduction and subsequent treatment with methyl sulphate, gave themethosulphate of trimethylcoclaurine ; the latter was identical withthe product obtained directly from the alkaloid by the action ofmethyl sulphate and alkali. The formula (XIV) has been assignedby H.Kondo and Z. Narita 91 to dauricine, an alkaloid from Meni-spermum dauricurn, as a result of an investigation of the oxidationproducts derived from the nitrogen-free compounds obtained by anapplication of the Hofmann degradation process to methyldauricinemethiodide and ethyldauricine ethobromide. The alkaloid thusappears to be closely related to coclaurine, although F. Faltis andH. Frauendorfer92 have expressed the view that it has a morecomplex structure (XV) of the bimolecular type with an oxygenbridge.In the formulae previously advanced for isochondodendrine methylether 93 the exact points of attachment of the two methoxyl groupsand the ether-oxygen to the isoquinoline system have been indoubt. I n solving this particular problem much has dependedupon the positions of the snbstituent groups in a tricarboxydi-methoxydiphenyl ether which was obtained by the oxidation of theproduct derived from the Hofmann degradation of the methyl ether.F.Faltis and H. Frauendorfer 94 have now established syntheticallythat this acid is 2 : 3-dimethoxydiphenyl ether 5 : 6 : 4’-tricarboxylicO4 Loc.cit.g1 Chem. Zentr., 1927, ii, 264; 1929, ii, 1926; Ber., 1930, 63, [B], 2420.Ber., 1930,63, [B], 806. O8 See Ann. Reporta, 1928,25,190ORGANIC CHEMISTRY.-PART III. 195acid (XVI), a fact which necessitates a revision of the earlier formulsin favour of one of the two alternatives (XVII) and (XVIII). 9 C02H gg$:g2eCH2HO2C($0Me H02C OMe lc\l(XVI.) (XVII.) (XVIII.)The formuls previously assigned to chelidonine cannot be regardedas satisfactory, and, as a result of a review of the reactions of thisalkaloid, together with a consideration of its relationship to proto-pine and a re-investigation of the structure of a degradation product,F.von Bruchhausen and H. W. Bersch 95 have now advanced theformula (XIX). Confirmation of these views from degradative andsynthetical experiments will be awaited with interest, since thestructures to be assigned to related alkaloids, such as chelerythrine,are involved.(XIX.)HOMe0Me0Me082Me0Me0 H2(111.)9 5 Ber., 1930, 65, [BJ, 2620196 PLANT :Aporphine Group.-After considerable activity during the pre-ceding two years, there is comparatively little to record on thisoccasion concerning members of the aporphine group.One interest-ing development, however, has been the isolation and investigationby J. Gog6 of Z-corydine and d-isocorydine, together with otheralkaloids, from CorydaZis ternata. Previous work 97 has settled thestructures of the closely related alkaloids bulbocapnine (I) andcorytuberine (11), and J. Gadamerg* has suggested the formulae(111) and (IV) for corydine and isocorydine respectively, a mixtureof these two products being obtained by the monomethylation ofcorytuberine. The new work indicates, however, that corydinehas the structure (IV) and isocorydine (111). Thus bulbocapnineethyl ether was converted into the substance (V; R = H) andsubsequently methylated, the resulting derivative (V; R = Me)being found to be identical with d-corydine ethyl ether.DiisoquinoZine Group.-K.Goto and H. Sudzuki g9 recentlydescribed the isolation of a new alkaloid, sinactine, in small quanti-ties from Sinomenium acutum, and it has now been identified byK. Goto and Z. Kitasato 1 as Z-tetrahydroepiberberine (VI). Oxid-ation with iodine in alcoholic solution led to epiberberinium salts,which, on subsequent reduction of the chloride, yielded dl-sinactine ;this proved to be identical withh ydroepi berberine .2OMea specimen of iynthetical dZ-tetra-O-YH,IO-FH, I(VIII.)Me0 CH96 J . Pharm. Xoc. Japan, 1929,49, 125, 126, 129; Chem. Zentr., 1930, i, 234.9 7 See Ann. Reports, 1928, 25, 187.98 Arch. Pharm., 1911, 249, 503.99 Bull.Chem. SOC. Japan, 1929, 4, 220; Ann. Reports, 1929, 26, 181.1 J., 1930, 1234.2 W. H. Perkin, J., 1918,113, 512 ; R. D. Hsworth and W. H. Perkin, J.,1926, 1777ORGAXIC CHEMISTRY.-PART I". 197The supposed synthesis of tetrahydroberberine by Pictet andGams3 was shown to be unsound by the work of R. D. Haworth,W. H. Perkin, and J. Railkin,4 and further investigations of thereactions concerned have now been described by J. S. Buck andR. M. Davis,5 who have shown that the substance which was believedby Pictet and Gams to be veratrylnorhydrohydrastinine (VII)is, in reality, 9-keto-6 : 7-methylenedioxy-3' : 4'-dimethoxyproto-papaverine (VIII). In actual fact, a mixture of these two productswas found to result during the earlier stages of the synthesis.Morphine Croup.-The " thebainone " prepared by R.Pschorr 6by reducing thebaine with stannous chloride and concentratedhydrochloric acid is regarded by C. Schopf and P. Borkowsky 7 ashaving the structure (I). Since dihydrothebainone, hydroxythe-bainone, and hydroxydihydrothebainone do not have the samering system as Pschorr's product, it is now suggested by C . Schopfand (Frl.) H. Perrey 8 that the latter should be called " metathe-bainone," the name " thebainone " being reserved for the substance(11)-The conversion of derivatives of dihydrocodeinone (111; R = H)into compounds of the dihydrothebainone (IV ; R = H) type, withthe rupture of the oxygen bridge and consequent appearance of aphenolic hydroxyl group in the 4-position, is well known as a featureof hydrogenation processes in the morphine series.The conversetransformation of dihydrothebainone into dihydrocodeinone hasnow been accomplished by C. Schopf and T. Pfeifer by fist con-verting the former substance into its l-bromo-derivative. Furtherbromination gave a product (V), which, by treatment with alkali,Ber., 1911, 44, 2480.J., 1924,125, 1686; Ann. Reports, 1924,21, 134.Annalen, 1927,458,148 ; Ann. Reports, 1928,25, 192.Anden, 1930,483, 169.li J . Amr. Chem. SOC., 1930, 52, 660; A., 485. Ber., 1905, 38, 3160.Ibid., p. 157198 PLANT :yielded 1 -bromodihydrocodeinone. The removal of the halogenatom with the formation of dihydrocodeinone was accomplishedby catalytic hydrogenation. The reduction of 1 -bromodihydro-%QI FH2 C CH%()\C CHI QH2 C CHcodeinone with zinc dust in alcohol gave l-bromodihydrothebainone.By a similar series of reactions hydroxydihydrothebainone (IV ;R = OH) has been converted into hydroxydihydrocodeinone (111;R = OH), and the bromosinomeneine obtained by the successiveaction of bromine and alkali on sinomenine lo (VI) is apparentlyformed by an analogous process and can be represented by theformula (VII), whereby it appears as an optical isomeride of l-bromo-7-methoxycodeinone.By bromination of dihydrometathebainone(VIII) (previously called " thebainol "), followed by the action ofalkali, 1-bromodihydrometacodeinone has been obtained, and thishas been converted by catalytic hydrogenation into dihydrometa-( 7 3 2CH bH-NMe v\/ I HC\/ \/OMe CH2 CH,(VII.) (VIII.) (IX.)codeinone (IX), which can be transformed into dihydrometa-thebainone again by the action of sodium amalgam. Dihydro-metacodeinone is readily converted by acids or alkalis into theisomeric metathebainone (I).lo K. Goto, J . Chem. SOC. Japan, 1923,44, 815 ; K. Goto and T. Nakamura,Bull. Chm. SOC. Japan, 1929, 4, 195; K. Goto and T. Nambo, ibid., 1930,5, 165.11 C. Schopf and (Frl.) H. Perrey, loc. citORGANIC CHEMISTRY .-PART 111. 199The spatial configurations of several members of the morphinegroup have recently been discussed by H. Emde.12Strychnine.-Further investigations into the chemistry ofstrychnine derivatives have necessitated the abandonment offormula (I), which was advanced for this alkaloid two years ago,lSand, in its place, a new formula (11) has been put forward by K. N.Menon, (the late) W. H. Perkin, and R. Robinson.l4 In thedevelopment of these structural views much has depended on thenature of dinitrostrycholcarboxylic acid, CSNH,(OH)2(N0,),*C02H,which is obtained from the alkaloid by heating it first with diluteand then with concentrated nitric acid. A fundamental postulateC,H,in the development of the earlier structure was that dinitrostrycholis a dinitrodihydroxyisoquinoline, but it has now been conclusivelyshown that this substance is a quinoline derivative, as originallybelieved. Thus dinitrostrycholcarboxylic acid was converted duringesterification into ethyl dinitro-0-ethylstrycholcarboxylate,CSNH2(NO2),( OH)(OEt)*CO,Et, which was transformed via thehydrazide and the usual Curtius reactions into the correspondingurethane, C,NH2(N0,),(OH)(OEt)*NH*C0,Et. The urethane, unlikedinitrostrycholcarboxylic acid, was degraded by hot nitric acid toyield picric acid, a result which is consistent with a quinoline, butnot an isoquinoline, structure for strychol. Furthermore, theurethane, by hydrolysis and subsequent oxidation with permangan-ate, was converted into 5 : 7-dinitroisatin (111). From the evidenceCO,H OH No2wcz N o W g g HNO, NH NO, N NO, N(111.) (IV.) (V.)la Helv. Chim. Acta, 1930, 13, 1035.1s R. C. Fawcett, W. H. Perkin, and R. Robinson, J., 1928, 3082; W. H.Perkin and R. Robinson, J., 1929, 964; Ann. Reports, 1928, 25, 193; 1929,26, 181. l4 J., 1930, 830200 PLANT :now available it is possible to say that dinitrostrycholcarboxylicacid must have one of the structures (IV) and (V).It will be noticed that the carbazole nucleus, a feature of formula(I) as well as of the original strychnine formula of W. H. Perkinand R. Robinson,15 is absent from the new structure. The evidencefor its presence, vix., the isolation of carbazole from the degradationof the strychnine molecule, was always uncertain on account of thevigorous nature of the experimental conditions and the ease withwhich the carbazole system can be formed from other types. Severalcharacteristics of formula (I) must, however, be regarded as estab-lished, e.g., the existence and arrangement of the ether oxygen andthe ethylenic linkage, and these are embodied in the new formula.A fundamental point in the development of formula (11) is thatthe hydroxyl group in position 3 in the quinoline ring of dinitro-strycholcarboxylic acid fixes the position of the second nitrogen atomin strychnine. Only in this way can the formation of the 3-hydroxylgroup during the degradation of strychnine be accounted for. Thenew formula provides an excellent explanation of many of thecomplex changes observed in the chemistry of this alkaloid andthereby obtains further justification.p 2 - q H Y C H 2 Thus, for example, the presenceCK I A,/ of the grouping :N*CH,*y:CH%H,*O*,\?a/ ,A?H, and its successive transformation to:NCO*q( OH)*CO*CH,*O*,N /yg/c\\, :N*CO*yH( OH) + CO,H*CH,*O* I I I (by hydrolysis), and :N*CO*YO +c* CH CH, CO,H*CH,*O*, accounts for the \d \o/ formation of dihydrostrychninonicacid, and then strychninonic acid,by the oxidation of strychnine withpermanganate. In formula (11) the exact arrangement of certaincarbon and hydrogen atoms, comprising C2H,, still remains t o bedecided. There are several possibilities, and the authors suggestprovisionally a bridge formulation of the type (VI), which isclosely related to the cinchonine structure.MisceZZuneous AZkuZoids.-Of the many possible stereoisomericmodifications of ephedrine and related substances, six have so farbeen recognised amongst the constituents of the Chinese drug MaHuang, wix., Z-ephedrine (I), d-$-ephedrine (I), Z-methylephedrine(11), d-methyl-+ephedrine (11), nor-&#-ephedrine (111), and nor-Ph YH-7 HMe PhyH-YHMe Php-yHMeOH NHMe OH NMe, OH NH2(1.1 (11.1 (111.)l 6 J . , 1910, 97, 305.(VI.ORGANIC CHEMISTRY.-PART III. 201Z-ephedrine (111) .I6 Considerable work has recently been devotedto the synthesis of the two stereoisomeric inactive bases, nor-dl-ephedrine and nor-dZ-+ephedrine, of the formula (111). D. H. Hey l7has found that the direct reduction of isonitrosopropiophenone (IV)in the desired direction is difficult, although it has apparently beenachieved catalytically,l* but the action of sodium amalgam onphenylacetylcarbinol oxime (V) in dilute acetic acid solution ledmore readily to a mixture of the two stereoisomerides. In addition,W. N. Nagai and S. Kanao l9 found that the reduction of p-nitro-a-phenylpropyl alcohol (VI) with iron and aqueous alcoholic sulphuricacid, or with tin and hydrochloric acid, led to a mixture of the twoinactive bases (111), which were subsequently resolved with the aidof d- and Z-tartaric acids. C. S. Gibson and B. Levin20 havePhE-Me PhYH-e Ph YH-THMe0 NOH OH NOH OH NO,(IV.) (V.1 (VI.)employed the natural nor-&+-ephedrine as a convenient base forthe resolution of some externally compensated acids.Previous investigations into the alkaloids of ergot have shownthat ergotoxine and ergotinine are interconvertible, as also areergotamine and ergotaminine. Furthermore, the properties andpharmacological action of ergotoxine and ergotamine hithertodescribed are so similar as to have promoted suspicion that thesetwo alkaloids are really identical and that small observed differenceswere due to impurities. This possibility has now been removed byS. Smith and G. M. Timmis,21 who have prepared all four alkaloidsin a state of purity from different specimens of ergot. A carefulexamination of their physical properties clearly indicates thatthey are definite and distinct substances. Little evidence is, as yet,forthcoming regarding the constitutions and inter-relationships ofthese alkaloids.Mention may be made in this section of interesting developmentswhich have recently occurred in the chemistry of certain pungentacid amides. The structure (VII; R = H) assigned to capsaicin,the pungent principle of Capsicurn annuurn, which has been basedl6 S . Smith, J., 1928, 51; W. Nagai and S. Kanao, J . Pharm. SOC. Japan,1928,48, 101; Chem. Zentr., 1928, ii, 2553; S. Kanao, Ber., 1930, 63, [B], 95;A., 362.l7 J . , 1930, 18, 1232.P. Rabe, Ber., 1912, 45, 2166; W. H. Hartung and J. C. Munch, J .Amer. Chem. SOC., 1929, 51, 2262.lo Annalen, 1929,470, 167; A., 1929, 807.2o J., 1929, 2754.21 J . , 1930, 1390202 ORGANIC CHEMISTRY .-PART 111.upon the degradative reactions of this substance,22 hasOMe OR CHMe,*CH:CH*[CH,],*CO*NH*CH,confirmation from a synthesis by E. Spath and S.now received(VII.)F. Darling.23The keto-acid, CHMe2*CH,*CO*[CH2],*C02H, from the action ofzinc isobutyl iodide on the chloride of ethyl hydrogen adipate, wasreduced with sodium and alcohol to the corresponding secondaryalcohol, from which the bromide, CHMe,*CH,*CHBrfCH,],*CO,H,was obtained on heating with fuming hydrobromic acid. Thecorresponding ester was then distilled with quinoline, and the productwas treated successively with sodium hydroxide, thionyl chloride,and veratrylamine to give capsaicin methyl ether (VII; R = Me).Hydrolysis of the latter gave the unsaturated acid,from the chloride of which capsaicin was obtained on interactionwith vanillylamine. The position of the double linkage in thealiphatic section of the molecule has been established by oxidationexperiments. Spilanthol, the pungent principle of para cress, hashitherto been regarded as the isobutylamide of a n-nonenecarboxylicacid, yielding n-decoisobutylamide on the addition of two atoms ofhydrogen,24 but M. Asano and T. Kanematsu 25 have made a furtherpurification of this substance and now believe that it has two atomsof hydrogen less than was originally supposed. Thus it is stated toabsorb four atoms of hydrogen when it is converted into n-decoiso-butylamide. Since oxidation with ozone gave succinic acid,n-valeric acid, and formic acid, the structureis assigned to it. A product closely related to spilanthol is pelli-torine, the pungent principle of Anacyclus pyrethrum, which hasrecently been investigated by J. M. Gulland and G. U. Hopton,26who came to the conclusion that it is the isobutylamide of an-nonadienecarboxylic acid, C9H,,*C0,H, absorbing four atoms ofhydrogen on catalytic reduction to give n-decoisobutylamide. Theseviews are not in entire agreement with those of E. Ott and A. Behr,,'who have stated that pellitorine is the isobutylamide of a n-decadiene-carboxylic acid, C,,H1,*CO,H.CHMe,*CH:CH*[CH,],-CO,H,CH,*[CH,]3*CH:C:CH*[CH2]2*CO*NH*CH2.CRMe2S. G. P . PLANT.22 See, e.g., E. K. Nelson and L. E. Dawson, J . Amer. Chem. SOC., 1923, 45,23 Ber., 1930, 63, [B], 737; A., 599.24 Y. Asahina and M. Asano, J. Pharm. SOC. Japan, 1920, 503 ; 1922, 85.26 Ibid., 1927, No. 644, 521.2 6 J . , 1930, 6.2179; A., 1923, i, 1108.27 Ber., 1927, 60, [B], 2284; A., 1928, 50