ORGANIC CHEMISTRY.PART I.-ALIPHATIC DIVISION.01,efinic and PolyoleJinic Additions.Mono-oleJins.-The work of the past few years on the subject ofolefinic and polyolefinic addition has increasingly focused attentionon the following features of the addition process : (1) the polarisedstate of the olefin molecule, (2) the condition and manner in whichthe addendum enters reaction, (3) the point of attack in the olefinicchain (Ca, Cb, etc.), (4) the influence of substituents on the course andspeed of reaction and (5) the influence of reaction conditions (tern-perature, solvent, catalytic action, etc.). All of these enter intoconsideration in connexion with orientation problems.How profound are the effects of catalytic action on the orientationof addition products is to be gathered from the elaborate experimentsof J.P. Wibaut and his collaborators 1 on the addition of gaseoushydrogen bromide to gaseous vinyl bromide in the presence of glasswool, asbestos and of metallic bromides. I n these experiments theproportion of the isomeric dibromoethanes formed in the reactionvaried extensively with changes in the chemical composition andphysical condition of the catalyst, and quite frequently withoutassignable cause. The proportion also varied with the temperatureof interaction, but there was no interconversion of the isomeridesthemselves when heated alone, or in the presence of catalysts.Similar experiments with acetylene and hydrogen bromide indicatedthat the primary formation of vinyl bromide takes place withgreater difficulty than its subsequent conversion into dibromo-e t hanes.It has been shown by the same workers and independently byG.N. Burkhardt and W. Cocker2 that the interaction of vinylbromide with solutions of hydrogen bromide is very sensitive tochanges of temperature, concentration and solvent , and apparentlyto some extent of illumination. It is suggested by the latter authorsthat for reactions in which the molecular activations are practicallyexclusively thermal, the reactivity of vinyl bromide towards the lessconcentrated aqueous solutions of hydrogen bromide can be repre-lc-7 0sented schematically by CH,-CH-Br, in which the existence +Rec. traw. chirn., 1931, 50, 313; A., 698. Ibid., p. 837; A., 1031ORGANIC OHEMISTRY.-PART I.67of a " conjugative " effect 3 activating the @-positions for interactionwith hydrion, and a " general " effect 3 minimising the effectivenessof these polarisations, is indicated. A large fraction of the generaleffect is considered to be transmitted by paths immediately outsidethe molecule, in which circumstances it would be especially sensitiveto changes in the conditions external to the molecule such as wouldarise from changes of solvent and temperature. This effect wouldbe most important near a halogen atom. The addition in questionis a reaction which is characteristic of the ethylene bond itself; anatom or group immediately adjacent to a double bond would affectsuch a reaction by enhancing or suppressing, in some degree, one orother of the two opposite polarisations which are possible in theethylene bond.This would give a mixed product unless the polaris-ing group had a powerful effect. On the other hand, a propertywhich is not shown by the ethylene bond, e.g., addition of cyanideion to kationoid carbon, would show itself at the carbon atom moreremote from the polarising group and a single product would beexpected if the reaction proceeded a t all (compare the formation ofethylenediamine by the action of ammonia on vinyl chloride) .4Quite other considerations would enter into the discussion of orient-ation in additions promoted by light and by surface action : thechemical composition of catalytic surfaces may exert an influenceon the steric positions taken up by molecules in the adsorption layer(" directed adsorption "), but a more important influence than thisin determining the orientation of addition products is probablythe electronic distortion of the unsaturated system which takesplace in the process of adsorption.A more comprehensive hypothesis relating to the influence ofsubstituents on the orientation of olefin addition products has beenadvanced by C.K. Ingold and E. H. Ingold : 5 in this the orientationprinciples developed in connexion with aromatic substitution andtautomeric phenomena are applied with suitable modification toadditive processes. It is pointed out that when X-Y adds toR-C,=CB, the process being initiated by the orientation of( 6 +)Y to (6 -)CB (addition of Brz, HBr, etc.),6 the electron-0f hCompare J.Allan, A. E. Oxford, R. Robinson, and J. C. Smith, J.,1926, 404; A., 1926, 397; A. Lapworth and R. Robinson, Mem. Man-Chester Phil. SOC., 1927, 72, 43,ti J . , 1931, 2354; A., 1267.R. Engel, Bull. SOC. chim., 1887, [ii], 48, 96.It is admitted by the authors that addition initiated by the attack of(6 -) Hal. would also be theoretically possible, provided that the intrinsicallysmall reactivity of negative halogen could be overcome by the attachmen68 FARMER :demand made on R might have any value between a very small oneand that which corresponds to the development of a free integralcharge on C,; the actual value in a given case will depend on thatdegree of electrostriction between X and C, which is necessary todetermine the consolidation of the initial union between Y and Cg,or, in other words, on the degree to which the second stage of addition" backs up " and overlaps the first.This in turn depends, inter alia,on the degree of prior polarisation of X-Y, and the case of maximalelectron-demand will arise in the addition of previously formed ordissociated ions.The majority of those groups which activate anionotropic systemsand promote aromatic hydrogen-replacement by ordinary substitut-ing reagents will increase the additivity of an ethylene towardsreagents such as halogens, halogen halides, and nitrogen peroxide,the more active ions of which are positive, and will decrease reactivityto such addenda as hydrogen, hydrogen cyanide, nitromethane, andethyl malonate, whose more active ions are negative ; and similarlymost of those substituents which facilitate prototropy and aromatichalogen-replacement by basic reagents will deactivate the ethylenenucleus towards the former group of reagents and activate it towardsthe latter.It is recognised, however, that there are stronglyrestrictive limits to the validity of this reasoning.It should be borne in mind that when two groups are attached toan ethenoid centre, the separate groups may not compound simply.For example, if de-activation results when two substituents arepresent, it does not necessarily follow that either would act in thissense if present singly, Furthermore it is not dissymmetric sub-stitution and the consequent electrical dissymmetry of the doublelinking which constitute the primary influence of substitution ;for the electron-displacements of the addition process must besubstantially confined to the ethenoid centre and hence it is thepolarisability of the corresponding electrons, not their state ofpolarisation, which is the important factor.The classes of substituent for which distinctive effects on speedsof addition are expected to arise either when the constituent is theonly one attached to the ethylenic centre, or when if is present inconjunction with other like or unlike groups, are symbolised asfollows :of a sufficiently powerful electron-sink to the olefinic residue.It is not yetknown whether any of the available groups (GHO, GOR, 'NO,) are in factcompetent to cause this anticipated reversal of normal reactivity; the prac-tical criterion which would reveal such an occurrence would be that the rateof addition would pass through a.minimum as the efficacy of the electron-sink was progressively increasedORGANIC CHEMISTRY.-PART I. 69Class ......... (1) + I (2) - I (3) - I - T (4) - I + TClass ......... (5) + I + T (6) + I - T (7) Z t TExample ... : C + - O e :c+co,e0Example ... :C+Me :C-+NMe,@ :6%C02H :G-,BrP n 0:C---Ph u0The methyl group (Class 1) repels electrons and should thereforefacilitate the addition of a reagent such as bromine, the reactive ionof which is positive.TSI (S-1 (a+) n(&-) ........... Me+CHzzCH,. B r B rThe accumulation of methyl groups should (in the absence of groupspossessing a mechanism for electromeric change) produce, andindeed has been shown in the case of alkylethylenes to produce, aprogressive increase in the speed of addition.The carboxyl group(Class 3), the - T effect of which is weak, either alone or in thepresence of a substituent of Class 1 or 2, should invariably de-activate :this has been illustrated for acrylic and crotonic acids. Withrespect to the action of substituents of Class 4 it cannot be decideda p ~ i ~ r i whether the activating + T will fail to compensate, or willover-compensate, the de-activating - I effect; in vinyl bromideit is shown experimentally that the bromine substituent definitelyde-activates and it is to be expected that any dibromoethylene andtetrabromoethylene will exhibit strong passivity.The phenylgroup, like bromine, exhibits duplex polar action (& T) but withthe difference that both the component effects are polarisabilities :of the two electromeric effects, it is the one symbolised by + Twhich electron-demanding reaction should stimulate, wherefore thephenyl group should be an activator. The relative rates of bromineaddition to styrene and ethylene show that this is the case, and thewell-known passivity of te t rap hen ylet h ylene , and of certain otherstilbene derivatives, is probably quite wrongly attributed to de-activation by phenyl groups.The orienting influence of solvents is traced by C. K. Ingold and(Miss) E. Ramsden 7 t o that portion of the effect of a polarisinggroup which is transmitted outside the molecule itself : the latter intravelling through the medium is profoundly affected by the natureof the medium.For additions to R*CH:CH,, where R is an alkylgroup, the externally propagated orienting influence is consideredto repel electrons, anions, and the negative ends of polarised mole-cules, and to attract protons, kations, and the positive ends ofpolarised molecules. This effect must tend to orientate the adden-J., 1931, 2746; A., 139170 FARMER :dum in the sense of the expression R*CH(H)*CH,(X). Certainaspects of the problem relating to the comparison of the orientinginfluence of solvents are discussed by the same authors. I n con-nexion with the orienting influence of media, the statement that thecomposition of the product obtained by addition of hydrogenbromide to ally1 bromide in glacial acetic acid is greatly changedwhen the reaction is carried out in a magnetic field is of interest.*Information with regard to the location in the olefinic system ofthe point of initial attack does not emerge from a knowledge of thecoqstitution of the addition products.The case of the halogens is,however, amenable to practical treatment on the assumption thatthe positive end of the halogen molecule initiates reaction : thus foraddition of ICl, the location of the iodine atom in the additionproduct indicates the point of initiation of addition, and in testscarried out with propylene, styrene, and crotonic acid the resultsare found to be in accordance with the foregoing considerationsconcerning the directive influence of substituent s .gA useful review of the literature referring to the formation ofethylene chlorohydrin is contributed by E.D. GI-. Frahm.lo Thisauthor appraises the efficiency of the various practical expedientswhich have been adopted for improving the yield of chlorohydrinand depressing that of dichloride formed according to the scheme :C,H,CI, f- H,O + C1, =s H' + C1' + HOCl+ C,H,(OH)Cl.Acceleration of chlorohydrin formation and diminution of dichlorideproduction were found to result from the catalytic action of certainmetallic salts and it is suggested that chlorohydrin formation fromethylene and chlorine in aqueous media takes place by way ofethylene oxide.The degree of hydration of ethylene and propyleneproduced by heating with hydrochloric, hydrobromic, or hydriodicacid, or with aqueous solutions of various chlorides or of silvernitrate, at 100-200" does not exceed 0.6y0 even a t a pressure of800 Ib. per sq. in.; high pressures, and temperatures of 227427",are reported to favour hydration.11It is remarkable that p-chlorocrotonic acid and its sodium saltrapidly undergo catalytic reduction, although p-chlorobutyricacid does not.llU The reduction proceeds in stages-rapidly toR. H. Clarke and K. R. Gray, Trans. Roy. SOC. Canada, 1930, [iii], 24,111, 111; A., 1931, 598; see also R. H. Clarke. and E. G. Kallonquist, ibid.,p. 115; A . , 1931, 597.* C. K. Ingold and H. G. Smith, J ., 1931, 2742; A., 1391.lo Rec. trav. chim., 1931, 50, 261; A., 598.l1 R. D. Snow and D. B. Keyes, I d . Eng. Chem., 1930, 22, 1048; A , ,C. Pad, H. Schiedewitz, and K. Rauscher, Ber., 1931, 64, [B], 1521 ; A.,1562; A. V. Frost, J. Appl. Chern. Russia, 1930, 3, 1069; A., 1931, 1267.934ORGANIC CHEMISTRY.-PART I. 71crotonic acid and then more slowly to butyric acid. The differencein rate of hydrogenation between chlorocrotonic and chloroiso-crotonic acids serves to confirm the cis-tram nature of the isomerismexisting between these acids.The conclusions recently reached by A. Michael and J. Ross l2with respect to the manner in which alkylated malonic and cyano-acetic esters divide when acting as addenda in the Michael reactionare considered by N.E. Holden and A. Lapworth l3 to be erroneous.The latter authors suggest, and bring evidence in support of theview, that the apparently realisable partition of an alkylatedsodio-ester such as ethyl sodiomethylmalonate in the mannerCH,-i--CNa (C0,E t) , instead of Na-/--CMe (C0,E t)2 resultsfrom an interchange (in effect) of *CO,Et (*CN for an alkylatedcyano-ester) for the H of the CHNa group in the normal additionproduct. Thus in additions to ap-unsaturated esters of the generaltype R*CH:CH*CO,Et the normal addition product (I) may sufferextensive change to the isomeric compound (11) either intramole-cularly (e.g., via the sodio-derivative of the Dieckmann product 111)or by an intermolecular process.l4 The occurrence of such a changeunder the conditions utilised by Michael and Ross ( i e ., the employ-ment of & whole molecular proportion of sodium ethoxide for effectingcondensation) is not inconsistent with already known reactivities.R*CI-I:CH*CO,E t R*CH*CHNa*CO,E t + - 3 - 1 (1.)/ cMe(Co2Et)2CNaMe(CO,Et),R*CH*CNa( CO,Et), R*CH----- CH*CO,Et(11.1 I I >co (111.)CHMe*CO,E t CMe( C0,Et)In connexion with the characterisation of conjugated butadienesby the formation of crystalline addition products with maleicanhydride, certain olefins (e.g., styrene and indene) have beenshown to yield non-crystalline polymolecular addition products withthe same reagent; in one case (as-diphenylethylene), however, arelatively simple, although rather high-melting, crysta2Zine additionproduct has been obtained and there appears to be need for theexercise of care in applying the Diels-Alder reaction to the diagnosisof conjugated unsaturation.l6l2 J . Amer. Chern. SOC., 1930,52,4598; 1931,53,1150; A., 1931, 67; 603.13 J . , 1931, 2370; A., 1272.The transformations are represented for convenience in terms of con-l5 T. Wagner-Jauregg, Ber., 1930, 63, [B], 3213; A . , 1931, 198; Annulen,ventionally formulated sodio-derivatives.1931, 491, 172 FARMER :Polyok$ns.-A notable synthesis of an open-chain poly-ene hasbeen achieved by P. Karrer and A. Helfenstein in producing ahydrocarbon containing squalene from farnesyl bromide. l6 Theconversion is brought about in ethereal solution by the action ofpotassium or activated magnesium, and the product yields thecrystalline hexahydrochloride described by I.M. Heilbron, E. D.Kamm, and W. M. Owen.17 Squalene is therefore correctly repre-sented by the formula( CMe2:CH*CH,*CH2*CMe:CH*CH,*CH,-CMe:CH*CH2*),.Recognition that the crystalline additive compound of isoprenewith sulphur dioxide is a cyclic sulphone has been followed by theproduction of an isomeric compound by the action of ultra-violetlight on a solution of the first compound.18 The double bondappears to be retained in the original position and the possibilitythat the cause of isomerism resides in the sulphone group is rejectedon chemical grounds. The two substances yield the same dihydro-derivative and are considered to display cis- and trans- attachmentsrespectively of the *CH,*SO,*CH,* chain to the terminals of theethylenic bond, so giving planar and pleated ring configurations.An important point affecting the classification of additive reactionsin the conjugated poly-ene series turns on the mechanism of theMichael reaction.If in this reaction the addendum attacks bymeans of its anion, as in hydrogen cyanide additions,l9 then it mustbe assumed that any polarisations of the unsaturated molecule whichare capable of influencing the additive mode must exist before theanion enters the unsaturated molecule ; moreover, any plurality ofaddition products in a given instance must be explained by assumingdifferent polarisations of the butadienoid molecule, leading to theproduction of isomeric anions such as A and B, which are onlyinterconvertible via the retrograde Michael reaction (&product =Sreactants ap-product).Me*CH*gH-CH:CH*CO,Et Me*CH:CH*CH*gH*CO,Et I (*)CH(CO,Et), (4 I CH(CO,Et),(capable of prototropic change) (non-tautomeric)This is necessarily the case, since the prospect of attachment of theaddendum at the 8-carbon atom would be non-existent were it notfor the presence of a group such as carbonyl in the molecule (in otherwords, the addendum in the Michael reaction is not an additive16 Helv. Chirn.Acta, 1931, 14, 78; A., 333.17 J . , 1926, 1630; A., 1926, 816.l8 E. Eigenburger, J. pr. Chem., 1931, [ii], 129, 312; 131, 289; A., 600,la A. Lapworth, J., 1903, 83, 995; 1904, 85, 1206.1268ORGANIC CHEMISTRY.-PART I.73reagent for purely ethylenic centres) ; therefore, granting successfulattack by the anion of the reagent, it follows that the influence of theactivating group (carbonyl) is felt, and is effective in promotingreaction, at the &carbon atom-that is to say, a state of conjugationis effectively established or induced in the poly-ene chain.It has been found by E. H. Farmer and T. N. Mehta 2o that theratios in which ap- and as-ester-addition products are formed fromthe members of a series of alkyl-substituted butadiene-a-carboxylicesters are not those to be expected on the basis of an anionotropicallycontrolled incidence of addition products. Taking sorbic ester asthe standard, it is seen from the figures quoted that (a) deprivationCH,-CH:CH*CH:CHX 7-10 CH,*CH:CH*CH:CX, no additionCH,:CH*CH:CHX (2” CH,CH:CMe*CH:CHX 70-74.5CH,CH:CH*CMe:CHX (7 CH,*CH:CH*CH:CHAc 27-29CH,*CMe:CH*CMe:CHX no addition* Not exactly determined.of the S-methyl group, or introduction of a p-methyl group, repressesap-addition, (b) introduction of p- and S-methyl groups, or intro-duction of an a-carbonyl group, represses both ap- and as-addition,and ( c ) introduction of a y-methyl group, or replacement of thecarboxylic activating group by acetyl, strongly augments orp-addition.Modification of the ester addendum appears, as might be expected,only to affect the ease of addition.J.Bloom and C. K. Ingold 21 regard Farmer and Mehta’s resultsand certain supplementary results which they furnish relating toa-methylsorbic ester (not more than 2% ap-addition), cinnamylidene-acetic ester (entirely or almost entirely ap-addition), and styryl-cinnamic ester (little or no addition) as defkitely supporting the viewthat the reagent attacks in the form of its anion.They assumethat the conductivity of unsaturated ions, including dipolar ions,guarantees the distribution between CS and Cp of the positivecharge produced by the primary polarisation, whilst the greaterlability of double-bond electrons than of single-bond electronsdetermines the dominant reactivity of CS :aP (%I* aP (%I*(sorbic ester)[X = CO,R]This holds provided the conditions of substitution in the CS : C,,- andCp : C,-units, which determine the respective degrees of reversibility2o J., 1930, 1610; 1931, 1904, 2661; A., 1930, 1163; 1931, 1397.21 Ibid., 1931, 2765; A., 1391.c 74 FARMER :of the corresponding addition reactions, are comparable.The effectof modifying the existing state of substitution in the sorbic estersystem should then be of the kind observed, the phenyl groupexerting the orienting influence by virtue of its electron-releasingproperties.The formation, under suitable conditions of reaction, of ana@ -es ter - addition product of ethyl muconate, almost entirely freefrom isomerides, has confirmed the view that the extensive pro-duction of my-products formerly noted is to be attributed to @,Py-equilibration in the presence of sodium ethoxide.22I n hydrogen halide addition to conjugated compounds it might beanticipated that the hydrogen atom of the hydrogen halide molecule,probably as the ion, would initiate attack, and experiments intendedto test this point have been carried out.Here, however, thepossibility of attack a t the p- (or @'-)carbon atom of the butadienechain enters. Experiment shows that the mode of addition ofhydrogen bromide to @y and a8-dimethylbutadiene is compatiblewith reaction initiated at the terminal carbon atom, and completedin accordance with anionotropic principle^.,^The same questions as to which ion or pole of the addendummolecule initiates reaction, and which atom of the conjugatedmolecules furnishes a seat of attack, enter into the consideration ofadditions by other unsymmetrical reagents such as iodine chloride.C.K. Ingold and H. G. Smith 24 find that this reagent, when addedto butadiene, behaves in a way which might be anticipated on thebasis of the mechanism imputed to bromine addition : 25 a mixtureof CH,I*CH:CH*CH,Cl and CH,I*CHCl*CH:CH, is formed, con-formably with the initial entrance of iodine at the a-carbon atom.The results of I. E. Muskat and H. E. Northrup for the additionof chlorine to butadiene 26 confirm expectations based on knownresults for bromine addition,,' as to the simultaneous formation of@- and a6-dichlorides. Other statements of Muskat and hiscollaborators 2* relating to the 78-hydrogenation of p-vinylacrylicCompareE. H. Farmer, ;bid., 1923, 123, 3324.Furtherhydrobromination results (unpublished) obtained at the same time by theseauthors, which refer to piperylene and aa-dimethylbutadiene, can be bestexplained on the basis of hydrion attack initiated at the a- (or a'-) carbon atom.22 E.H. Farmer and T. N. Mehta, J . , 1931, 1762; A., 1037.23 E. H. Farmer and F. C. B. Marshall, ibid., 1931, 129; A., 460.24 Ibid., p. 2752; A., 1391.25 Ann. Reports, 1 9 2 8 , s . 131.26 J . Amer. Chem. SOC., 1930, 52, 4043; A., 1930, 1553.27 E. H. Farmer, C. D. Lawrence, and J. F. Thorpe, J . , 1928, 729; A . ,1928, 604.28 I. E. Muskat and B. Knapp, Ber., 1931, 64, [B], 779; A., 719; I. E.Muskat and L. Hudson, J . Amer. Chem. SOC., 1931, 63, 3178; A., 1148ORGANIC CHEMISTRY.-PART I. 75acid under catalytic conditions and the (apparently complete)formation of a @-addition product of the same acid with hypo-chlorous acid are inadequately established by the published experi-mental evidence : moreover, unpublished observations by the writerand his collaborators show that the statements only partiallyrepresent the facts.Isomeric forms (two) of apy8-tetrachlorobutadiene are obtainedas a by-product in the preparation of trichloroethylene from tetra-chloroethane by elimination of hydrogen chloride.29 The formationof both saturated and unsaturated four-carbon molecules fromhexachloroethane takes place apparently with moderate facility,but reversion to a two-carbon molecule (hexachloroethane) has onlybeen accomplished by exhaustive chlorination of the halogenateddiene .Acids.New instances of polporphism amongst carboxylic acids andesters have been recognised. Malonic, succinic, and glutaric acidsare dimorphous, the curve of transition points following a similarcourse to that of the melting p0ints.~0 Ethyl palmitate, ethylstearate, cetyl acetate, octadecyl acetate, ethyl cetylmalonate (aswell as cetyl chloride) have also been found to be dim0rphous.3~It isfound that sodium antimonyl tartrate (tartar emetic itself is tooinsoluble for freezing-point determinations to be made) is presentin the unimolecular condition in dilute aqueous solution, and itappears by analogy with a-phenylethylammonium and p-nitro-a-phenylethylammonium antimonyl tartrate, which contain nowater of crystallisation, that the water of tartar emetic is not con-stitutive.Tartar emetic is only weakly acidic, but is easily esterified :A new formula has been proposed for tartar emetic.32consequently only one carboxyl groupcan be bound to antimony.TheOH2 formula (I) is advanced which differsfrom Schiff’s formula in the addition0:y.oHy.o>Sb. . . . .(1.1 CO,K*HC*Oof a water molecule attached t o antimony by a residual valency.It represents an unstrained condition as shown by models andis supported by the fact that the water molecule in bariumantimonyl tartrate, although not in tartar emetic itself, is re-placeable by pyridine, the product probably existing as the acid29 E. Miiller and F. Huther, Ber., 1931, 64, [ B ] , 589; A . , 597.so F. D. laTour, Compt. red., 1930,191, 1348; 1931,193, 180; A., 1931,91 J.W. C. Phillips and S. A. Mumford, J., 1931, 1732; A., 1003; J. C.82 H. Reihlen and E. Hezel, Annalen, 1931, 487, 213; A., 936.198, 1036.Smith, ibid., p. 802; A., 68476 FARMER :[C,H,N . * * . SbC,H,O,]H in the solid state, and as the salt[H20 * * . - SbC,H206]H,C5H5N in solution. The existence ofthe auxiliary valency in (I) is supported by the loss of wat,er fromthe sodium salt in an evacuated vessel at room temperature (but fr0.mthe potassium salt, only a t 120") and by the existence of the freeacid only in the anhydrous form. Unlike the corresponding boronand arsenic compounds, tartar emetic is only slightly hydrolysed toHK(C,O,H,) and Sb(OH),, but its acid reaction is due to "intra-molecular hydrolysis. "Measurements of the specific conductivity of the complexesformed by tartaric and boric acids in solution have led M.Amadori 33(I) Dextro. (11) LZVO. (111) Meso.to the conclusion that the spatial disposition of the groups in thethree tartaric acids is to be represented by the models (I, 11, and 111),rather than by the well-known Fischer models in which the carboxylgroups are on the same side of the molecule. It is considered thatboric acid complexes are formed to about equal extents by the activeand inactive acids and that they contain two (*O*CH*CO,)BOHgroups which are disposed on opposite sides of the molecule; thesegroups if projected on to a plane normal to the axis of the centralcarbon atoms would have the relative dispositions indicated infigures IV and V.83 Bazzetta, 1931, 61, 216; A., 822ORGANIC CHEMISTRY .-PART I.77Various views have been put forward with regard to thespontaneous decomposition of dihydroxymaleic acid in aqueoussolution according to the equationHO,C*C( OH):C( OH)*CO,H -+ OH*CH,*CHO + 2C0,The suggestions have in turn been made that the decompositiondepends on hydration of the acid to trihydroxysuccinic acid, or theproduction of the bivalent negative ion ( C4H206)", and consequentlyon the ionising power of the medium, on the production of theunivalent negative ion (C,H,O,)', and finally on the concentrationof the undissociated acid solvated on the carboxyl group. It is nowshown from the relation existing between the velocity coefficient ofdecarboxylation and the degree of dissociation of the free acid, itslithium hydrogen salt, and lithium salt solutions of varying pH, thatthe univalent ion (C4H306)' is the actively decomposing consti-tuent.,5 The decomposition of the free acid appears to occur atthe velocity for the ion and the ratio of the ionisation coefficientsof the acid is that of a tram-acid of this series.Boric acid exerts astabilising action proportional to its own concentration and that ofthe dihydroxy-acid, and st kinetic investigation points to theformation of an equimolecular complex. Organic solvents depressthe formation of the active ion and forthe solvents water, acetonitrile, methylinverse proportionality of - log k (fordecarboxylation) is established. The catalytic influence on thedecomposition of the salts of heavy metals and of weak bases hasalso been studied.Conditions for the catalytic hydrogenation of carboxylic acids andesters to the corresponding alcohols have been investigated by anumber of workers.36 Various catalysts (copper chromite, nickel,cobalt, copper, etc.) are used in conjunction with high pressures, andtemperatures ranging from 250" to above 300"; at temperaturesabove 350" reduction can be carried to the hydrocarbon stage. It isnot essential, however, for the smooth reduction of the carboxylgroup to use high pressures, and provided a sufficiently activecatalyst is employed3' there is no hindrance to bringing about34 H.J. H. Fenton, J., 1895, 67, 774; 1905, 87, 804; S. Skinner, ibid.,1898, 73, 483; A.Locke, J. Amer. Chem. SOC., 1924, 46, 1246; A., 1924, i,708.(p2H R'o>B<oH OH+ alcohol, ethyl alcohol, and acetone theH02(JY-C*035 W. Franke and G. Brathuhn, Annalen, 1931, 487, 1; A., 936.86 W. Normann, 2. angew. Chern., 1931, 44, 714; A., 1269; H. Adkinsand K. Folkers, J . Amer. Chern. SOC., 1931,53, 1095; A., 698; W. Sohrauth,0. Schenck, and K. Stiokdorn, Ber., 1931, 64, [B], 1314; A., 932.87 0. Schmidt, Ber., 1931, 64, [B], 2061; A., 126878 FARMER :reduction a t atmospheric pressure ; thus, ethyl oleate in the presenceof copper chromate a t 270-280" yields octadecyl alcohol.Oleic acid is convertible to the extent of 50% into elaidic acid byheating with sulphur ; 38 erucic acid is likewise stated to be smoothlyconverted into brassidic acid when heated with 30% nitric acid andsome sodium nitrite.39 Oleic acid suffers oxidation by dilutepermanganate solution into the corresponding elaido- and oleo-dihydroxystearic acids; both of these substances, as well as thecorresponding dibromostearic acids, are resolvable through thestrychnine salts ; oleo-dichlorostearic acid is reported to yield withalcoholic potash stearolic acid and Aqt-octadecadienoic acid, thelatter of which is in turn convertible into stearolic acid by heatingwith the same reagent at a higher temperature.m From the parachorvalues obtained for oleic, elaidic, AB-oleic, erucic and brassidic acidsit is concluded that oleic and erucic acids have a &-configurationwhereas elaidic, M-oleic, and brassidic acids have a trans-con-f i g ~ r a t i o n .~ ~ The hydroxyoleic acid from ergot oil has beeninvestigated and found to be exactly similar in its physical con-stants and behaviour on oxidation to ricinoleic acid : the twosubstances are concluded to be identical.42 cis- and trans-Forms oferucylacetic (AE-tetracosenoic) acid have been synthesised, and theformer found to be identical with nervonic a ~ i d . 4 ~An observation which has an important bearing on currentattempts to determine the constitution of the higher unsaturatedfatty acids concerns the isomerisation of mono-olefinic esters duringcatalytic reduction. Methyl oleate, palmitoleate and erucate, whentreated a t 114-220" with insufficient hydrogen to produce completesaturation (nickel-kieselguhr catalyst), yield mixtures of esters inwhich unsaturated isomerides of the original esters occur ; thedouble bond in these isomerides occupies a position adjacent to thatit occupied in the original compounds, and both the original com-pounds and their isomerides appear to be present in cis- and trans-forms.44 Possibly the isomerisation is to be attributed to the com-38 G.Rankov, Ber., 1931, 64, [B], 619; A., 601.39 Idem, J . pr. Chem., 1931, [ii], 131, 293; A., 1272.40 Y. Inoue and B. Suzuki, Proc. Imp. Acad. Tokyo, 1931, '7, 261, 265;A., 1271.41 G. B. Semeria and G. Ribotti-Lissone, Guzzettu, 1930, 60, 862; A . , 1931,149.42 H. Matthes and 0. H. Kurschner, Arch. Phurm., 1931, 269, 88, 101;A., 602.43 J. B. Hale, W.H. Lycan, and R. Adams, J. Arner. Chem. SOC., 1930,52, 4536; A., 1931, 65.44 T. P. Hilditch and N. L. Vidyarthi, Proc. Roy. SOC., 1929, [ A ] , 122, 562;A . , 1929, 423; A. Steger and H. W. Scheffers, Chem. Umschuu, 1931, 38, 45,61; A., 711ORGANIC CHEMISTRY.-PART I. 79paratively high temperatures employed, since oleic acid is found tosuffer isomerisation to the A1-acid when heated at 250" in contactwith alumina ; at any rate the occurrence of isomerism presentsa complicating factor in the determination by reductive proceduresof the number and constitution of the polyolefinic acids which occurin natural products.T. P. Hilditch and N. L. Vidyarthi have deduced from the natureof the oxidation products of partially hydrogenated methyl linoleatefrom soya bean and cotton seed oil that the unsaturation is ofAoA-ty~e.~~ Y.Inoue and B. Suzuki have found by a similar methodthat a methyl isolinoleate derived from the oil of silkworm pupae isalso a A%ompound but differs from linoleic ester in having a&-configuration about the &-double bond.47CH,*[CH2I4*CH:CH*CH,*GH CH,-[CH,],*CH:CH*CH,$HCH*[CH,],*CO,H H0,C*[CH2]7*CHBy debrominating a solid tetrabromostearic acid prepared frompoppy and soya bean oils, W. C. Smit 48 has obtained an oily linoleicacid (m. p. - 12" to - 11") to which the A%onstitution is assigned ;by thermal dehydration of ricinelaidic acid, however, he obtained aproduct consisting of a solid AeK-compound (m. p. 54") admixed withliquid (probably stereoisomeric) isomerides ; likewise by the distill-ation of ricinoleic acid he obtained an oily mixture which appearedto contain both AeA- and AeK-linoleic acids, including a second form(m.p. 56") of the AeK-acid. The constitutions of the unsaturatedacids were determined by ozonisation.The conjugated AeK-linoleic acid and its ethyl ester combinenormally with maleic anhydride and by means of this reaction it hasbeen verified that the dehydration product of ricinoleic acid containsboth AOA- and AeK-holeic acids (about 15% and 75% respe~tively).~~Ethyl AeK-linoleate when partially hydrogenated is stated to yieldethyl Ah-elaidate, probably of the elaidic acid configuration. 5oEarlier work on the partial hydrogenation of ethyl linolenateindicated that, although some isomerisation occurs in the process,the unsaturated centres are probably to be assigned to 01- andAppositions and also to a point beyond the fourteenth carbon atomof the chain.46951 Later results obtained by partial hydrogenation4 5 K.H. Bauer and M. Krallis, Chem. Umschau, 1931, 38, 201; A . , 1034.4 6 Proc. Roy. Soc., 1929, [ A ] , 122, 563; A., 1929, 423.47 Proc. Imp. Acad. Tokyo, 1931, 7 , 15; A., 601.48 Rec. trav. chim., 1930, 49, 539; A., 1930, 891.49 J. Boeseken and R. Hoevers, ibid., p. 1166; A., 1931, 198.50 Idem, ibicl., p. 1161; A., 1931, 196.6 1 T. P. Hilditch, Chem. Umchau, 1930,37, 354; A., 1931, 335.Linoleic mid iSoLinoleic aci80 FARMER :methods indicated that the linolenic acid prepared from the solidhexabromide was a Aenbornpound and suffered hydrogenation instages, the OL- and go-linkings being reduced more readily than thehp-linking ; furthermore no migration of double linkings duringhydrogenation was noticed.52 It has since been concluded 53 thatthe AA-linking is first saturated, yielding AeE-octadecadienoic acid,from which the A”-isomeride is formed by shifting of the unsaturatedlinkings. On treatment with 2 mols.of hydrogen the remote doublelinkings of these two diolefinic acids are saturated, forming At- andAe-oleic acids, the latter of which yields the AT- (and probably At-)acid by isomerisation.Couepic acid, a new isomeride of elzostearic acid, which occursin the oil from Couepia grandijlora, gives oxidation products whichindicate that the acid is a geometrical isomeride of the knowna- and @-forms of elzeostearic acid,CH3*[CH,]3*[CH:CH]3*[CH2],*C02H.54Behenolic acid has been synthesised from the undecenoic acidobtainable from castorKetones.W.Bradley and R. Robinson 56 found that dibenzoylmethanes,R*CO*CH,*CO*R,, suffered fission by aqueous caustic soda in such away that the stronger of the two acids R*CO,H and R,*CO,H wasin nearly all cases produced in the greater relative amount. Thefission was considered to be a characteristic of the diketonic phasesof the tautomeric system present and the hydrolysis was thoughtprobably to result from the formation and decomposition of acomplex anion as in the scheme :R*CO*CH,*CO*R, + 8H + R*C(OH)(8)*CH2*CO*R1 -+RC8, + H*CH,*CO*R,Hence the chief factors controlling the change were taken to be( a ) the rates of formation of the two possible complex ions and ( b )the extent to which each of these breaks down with the formationof carboxylate ions.Experi-ments on the alkaline hydrolysis (0.1N-sodium hydroxide) and the1271.Several recent researches bear directly on this question.61 K.H. Bauer and E. Ermann, Chem. Umschau, 1930, 37, 241 ; A., 1930,b3 H. van der Veen, ibid., 1931, 38, 89; A., 712.54 J. van Loon and A. Steger, Rec. trav. chim., 1931, 50, 936; A., 1034;compare J. van Loon, ibid., p. 638; A., 822.55 R. Bhattacharya and J. L. Simonaen, Proc. X‘v. Indian S C ~ . Cong., 1928,153; A., 1931, 1271.6 6 J . , 1926, 2366; A., 1926, 1145ORGANIC CHEMISTRY.-PART I.81acid alcoholysis (alcoholic hydrogen chloride) of nine unsymmetricalP-diketones of the general formula CH,*CO*CHR*CO*R’ show thatthere is no relationship between the amounts of acid produced byfission and their strengths.57 The hydrolysis and the alcoholysisfigures differ considerably and it is suggested that the ratio of theamounts of CH,*CO,Et and R*CO,Et produced by alcoholysis is ameasure of the extent of enolisation of the respective carbonylgroups. Further experiments on the rate of alkaline alcoholysis(alcoholic sodium ethoxide) of six p-diketones and five p-ketonicesters 58 show that, although the ketonic esters are the more stable,the rate of alcoholysis of both diketones and ketonic esters increasesvery greatly with substitution on the carbon atom between thecarbonyl groups.The rate of alcoholysis of the diketones has beenmeasured in the presence both of hydrogen chloride and of sodiumethoxide and it is found that arrangement of the diketones in theorder of decreasing reactivity with respect to acid alcoholysispresents them (with one exception) in the order of increasing react-ivity towards alkaline alcoholysis. Diethyldibenzoylmethane, whichcannot enolise, readily suffers alcoholysis and hydrolysis.A. Michael and J. Ross 59 have formed the opinion that carbonylderivatives which do not form enolates with alkali react primarilyby addition of the alkali to the carbonyl group to form the group>C(OM)(OH). The slightly neutralised energy of the alkali in thiscomplex loosens the affinity of the C-atom for the directly attachedC-atoms so that cleavage can occur by hydrolysis.The ruptureshould take place at the carbon linking which is most weakened bythe positive influence of the alkali atom. Fission, however, is notdenied to enolic centres, and the fission of a sodium enolate byaqueous alkali is represented thus :R*CO*CH:C( ONa)R, + H,O + R-CO*CH,-C( ONa) (OH)*R, +R*CO*CH, + R,*CO$aAs regards the point of hydrolytic fission in the @-diketone molecule,it is considered that if the energy of the metal, in the alternativelyformed enolates ROC( ONa):CH*CO*R, and R*CO*CH:C( 0Na)-R,, ismore effectively neutralised in the first than in the second, fissionwill occur preferentially at the enolic centre of the second ; the pointa t which p-diketones are attacked by ordinary carbonyl reagentsdepends on whether the latter add more facilely to the carbonyl than5 7 W.M. Kutz and H. A h , J . Amer. Chem. SOC., 1930, 52, 4036; A.,s8 Idern, ibid., p. 4391; A., 1931, 69. Compare N. L. Drake and R. W.69 Ibid., 1931, 53, 2394; A,, 1035.1930, 1559.Riemenschneider, ibid., p. 5005; A., 1931, 19782 FARMER :to the *C(OH):C* group. Michael and Ross have also sought to showthat “ the principle of maximum neutralisation ” satisfactorilyexplains the course of reaction in the alkylation of mixtures of enolicesters or of a substance such as or-acetyl-or’-carbethoxylsuccinicester,60 when the enolic centres in each instance are made to competefor sodium.The experimental results are striking and are held toshow that the velocity relations applying to the action of alkylhalide upon the sodio-derivatives which are present, depend upon thefree positive energy of the sodium : thus, the proportions of thesodium enolates present a t the beginning of reaction become ofimportance only when the respective velocities of reaction are nearlyequal.All ketones which are structurally capable of enolisation, althoughnot necessarily appreciably enolised under ordinary conditions,appear to suffer displacement of the keto-enol equilibrium in favourof the enolic forms in the presence of excess of Grignard reagents :the enolic forms are thereby fixed as magnesium organo-enolates anda volatile hydrocarbon is liberated.61 Now when Grignard reagentsinteract with @-diketones, relatively normal products are obtainedonly from those compounds which are largely or entirely enolic.62Then two molecules of the reagent enter into reaction, liberating onemolecule of hydrocarbon, thus :PhMgX PhMgX Ph*CO*CH,*COPh --, Ph*H+Ph*C(OMgX):CH*COPh ~ 3Ph*C( OMgX):CH*CPh,*OMgX 4 Ph*CO*CH,*CPh,*OH.If the diketones are added to only one equivalent of the reagent,half of the material is recovered unchanged.The little-enolisedmono- and the di-substitution products of these simple diketonesalso react with two equivalents of the reagent. Addition takesplace to one of the carbonyl groups and the resulting magnesiumcompound suffers fission :PhMgX PhCO*CHPh*COPh -+ Ph*CO*CHPh*CPh,( OMgX) +PhMgX Ph*C( 0MgX):CHPh + Ph,CO --f Ph,C( OMgX).A similar kind of cleavage takes place when oxido-ketones aretreated with Grignard reagents.636o The assumption that the position which the alkyl group ultimatelyoccupies in this compound is necessarily that at which it enters duringalkylation appears to be no more justified than the similar assumption withrespect to the position taken up by the alkyl group of an alkylated addendumin the Michael reaction (compare p.71).61 Ann. Reports, 1930, 27, 100; V. Grignard and H. Blanchon, Bull. SOC.chim., 1931, [iv], 49, 23; A., 465.E. P. Kohler and J. L. E. Erickson, J . Amer. Chem. Soc., 1931,53, 2301 ;A., 1060.63 E. P. Kohler, N. K. Richtmyer, and W. F. Hester, ibid., p. 205; A., 354ORGANIC CHEMISTRY.-PART I.83The r61e of sodium in the acetoacetic ester condensationwhetheras true condensing agent or as a mere generator of sodium ethoxide-has been the subject of renewed discussion. The fact is stressed thatsodium produces two distinct types of reaction with aliphatic esters,wix., the acetoacetic ester and the acyloin types of condensation.64The metal itself is considered to be the direct cause of the acyloincondensation and to produce a t the same time sodium ethoxide :2R*CO,Et 5 ONa*CR(OEt)*CR(OEt)*ONa 2ONa*CR:CR*ONa + 2NaOEt.There are no data available to show whether the primary additionproduct reacts directly with more sodium or first loses sodiumethoxide to yield the free diketone; there is no need, however, toassume that reaction proceeds, when possible, through the enolform, or to postulate special mechanisms to account for the slownessof reaction of an ester which cannot enolise.With regard to theacetoacetic ester condensation it is deduced, from the proportions ofacetoacetic ester, alcohol and hydrogen which are actually producedby the action of sodium on ethyl acetate, that the r6le of the metal isimprobably other than that of a generator of sodium ethoxide, thelatter being the real condensing agent.It has been shown that the catalytic reduction of ethyl aceto-acetate in the liquid phase in presence of a nickel catalyst yieldsmixtures of ethyl p-hydroxybutyrate and the condensation productCHMe(OH)*CH2*C02*CHMe*CH2*C02Et. 65 The latter is the reduc-tion product of the compound which is usually supposed to be anintermediate in the formation of dehydracetic acid and would point tothe possibility that the formation of bimolecular reduction productsof the kind in question is related to the enolising capacity of theparticular acetoacetic ester employed, under the actual conditionsof its reduction.This view is supported by the results of experi-ments in which no solvent, and solvents of different ionising powerare employed, and by use of acetoacetic esters incapable of ionisation.The function of the base employed as condensing agent in theKnoevenagel and allied reactions remains obscure in spite of severalnew investigations. The first stage of the union, without condensingagent, of formaldehyde with malonic ester (to yield hydroxymethyl-malonic ester) 66 constitutes a bimolecular reaction, the velocity of64 J.M. Snell and S. M. McElvain, J . Amer. Chem. Soc., 1931,53,750,2310;A., 464, 1035. Compare B. B. Corson, W. L. Benson, and T. T. Goodwin,ibid., 1930, 52, 3988; A., 1930, 1559.66 H. Adkins, R. Connor, and H. Crarner, ibid., 1930, 52, 5192; A., 1931,197.66 K. N. Welch, J., 1931, 653; A., 603; compare A. C. 0. Hann and A.Lapworth, ibid., 1904,85,46; E. Hope and R. Robinson, ibid., 1911;99, 211784 FARMER :which varies inversely as the hydrogen-ion concentration ; ifammonia or trimethylamine is added as catalyst, the pH value of themixture not being allowed to change, little acceleration of thereaction ensues, but if piperidine is added rapid acceleration setsin.I n the reaction between cinnamaldehyde and malonic acid(in aqueous alcohol), however, the velocity of reaction increases withincrease in the hydrogen-ion concentration if the catalyst be one ofvarious amino-acids ; 67 moreover, urea and similar weak bases whendissolved in acetic acid are found to form very efficacious catalystsfor the Knoevenagel reaction. The catalytic influence in the lattercases is imputed to the kation of the amino-acid or the kation of thehighly-dissociated salt formed from the weak base and acetic acid.I n the Doebner reaction (more properly designated, it appears,the Verley reaction) the nature of the condensation product cannotbe correlated with the order of the base employed.68 Further, it isclear that the action of pyridine in yielding @-unsaturated acids isquite abnormal with respect to the general action of bases (even ofother tertiary bases, or bases of similar strength) although, remark-ably enough, the use of traces of pyridine leads to the production ofnormal py-unsaturated products.The catalytic activity of the basemight presumably here also be imputed to the kation of a salt formedwith malonic acid but could not generally be attributed to thelowering brought about by the base in the acidity of the medium.W. Cocker, A. Lapworth, and A. T. Peters 69 show that thePinner-Kotz theory with respect to the action of potassium cyanideand chloral in forming potassium dichloroacetate is capable ofextension to cover a group of reactions including the conversion ofcyanohydrins into amino-nitriles (a reaction proceeding by way ofthe free aldehyde or ketone), and the reversible formation of acetals.Some details with regard to the reactivity of ketones in acetalformation are recorded by G.F. Pfeiffer and H. Adkins.'OOxidation Processes and Peroxides.Several noteworthy attempts have been made to throw light onoxidation processes which employ or involve the formation ofperoxide compounds : these comprise oxidations by ozone, byhydrogen peroxide, and auto-oxidation processes.Ozonisation of an undiluted aldehyde gives rise to a mixture of thecorresponding acid and peracid, containing the former in excess :C7 K. C. Blanchard, D. L. Klein, and J.MacDonald, J . Amer. Chern. SOC.,1931, 53, 2809; A., 1017.6 8 S. E. Boxer and R. P. Linstead, J., 1931, 740; A., 935.69 Ibid., p. 1382; A., 1037.70 J . Amer. Chem. Soc., 1931, 53, 1043; A., 606ORGAXIC CHEMISTRY.-PART I. 85in this way the whole of the ozone is used up (2R-CHO + 0, --+R*CO,H $- R*CO,H). On the other hand, ozonisation of thealdehyde in solution causes more extensive oxidation than isaccounted for by the amount of ozone used; it is suggested in thiscase that ozone itself catalytically promotes interaction between theozonide first formed and oxygen (R*CHO,O, + 0,+ R*CO,H + 0,),and this behaviour probably also characterises the ozonisation ofolefins in solution. 7 1From the form of ozone-absorption curves relating to unsaturatedhydrocarbons treated with ozonised oxygen, G.Brus and G. Peyres-blanques 72 have concluded that the primary products of reactionare normal ozonides and that the number of double linkings can bededuced from the data. Perozonide formation appears to occursubsequently, probably by addition of an atom of oxygen at theketo-group. A recent re-examination of the well-known “ lsvulicaldehyde diperoxide ” of Harries 73 has shown that this substance isa decomposition product of caoutchouc ozonide, which is progressivelyformed when solutions of’ the latter are allowed to stand for a longtime. It can be obtained in yields representing 15-20% of thecarbon skeleton in the caoutchouc employed. In character it isacidic, peroxidic, but non-aldehydic : it does not behave as a peracidbut yields lzevulic acid (over 90% yield) on reduction with aluminiumamalgam.The formula (A) of Harries is replaced by the bimolecularformula (B).CH,*C*CH,*CH,*CH*O CH,*C*CH,*CH,*CO,H0\4 0 AA. Rieche and his collaborators have shown that aldehydes readilyunite with alkyl hydrogen peroxides to give monohydroxydialkylperoxides, R*CH( OH)*O*OR, and with hydrogen peroxide itself togive hydroxyalkyl hydrogen peroxides, RCH( OH)*O*OH, and thecorresponding di- a-hydroxyalkyl peroxides. Ethylidene peroxide,it is interesting to note, could be obtained not only by dehydrationof hydroxyethyl derivatives (mono- or di-), but also by warmingbutylene ozonide or by the auto-oxidation of ethyl ether. The modeof decomposition of these substances is de~cribed.7~71 F. G.Fischer, H. Dull, and J. L. Volz, AnmaZen, 1931,486, 80; A., 604.73 R. Pummerer, G. Ebermayer, and K. Gerlach, Ber., 1931, 64, [B], 804;A., 733.I4 A. Rieche and F. Hitz, ibid., 1930, 63, [B], 2642; A., 1930, 1554; A.Rieche and R. Meister, ibid., 1931, 64, [B], 2328, 2335; A., 1267, 1268.Compt. rend., 1930,190, 601; A., 1930, 44986 FAl3MER :The oxidation of malonic, tartaric, maleic and similar dibasic acidsby means of hydrogen peroxide at 100" proceeds through an acid-hydrogen peroxide complex but does not involve the formation of aperacid; 75 the oxidation of formaldehyde by acidified hydrogenperoxide is stated to yield either formic acid and hydrogen(2CH,O + H202 I_, BH*CO,H + H,) or formic acid and water.76Considerable interest attaches to the suggestion that the auto-oxidation of ethers proceeds by addition of molecular oxygen to theethereal oxygen atom.77 The immediate product is represented tobe of oxonium type and suffers decomposition in two different ways :(R-CH2),0 -+ (R*CH,)28*O*8 _I, R-EH, + R*CH,*O*O*8R*CH,*O*O*O*CH,R + R*CHO + H,O --F --+ R*CH,*OH+R-CH,*O*O*OH + R*CH,*OH+R-CHO+H,ODifferences which have been observed in the behaviour of peraceticand perbenzoic acid towards unsaturated substances are now foundto be only of secondary character : both reagents form primarilycyclic oxides, but fission of the oxide ring may occur subsequentlyunder the conditions of reaction.78 Solutions of peracetic acid havebeen applied to the determination of the degree of unsaturation inoils and fats 79 and i t is claimed the method gives in the cases examinedresults which are in agreement with the Wijs iodine values and aremore concordant than those obtained by the perbenzoic acid methodof Nametkin and Abakumovski. Apparently, however, oxidationat the ethylenic linking does not inevitably occur with peracids,80since styryl methyl ketone is reported to yield p-phenylvinyl acetatewith perbenzoic acid (CHPhXHAc + CHPh:CH*OAc).Oxidationof the acetylenic linking takes place less readily than is the case withthe ethylenic linking and follows, when it occurs, a more complicatedcourse of reaction.slPolymerisation.For many years opinion has been divided between the " cyclo-butane " formula of Kramer and Spilker for dicyclopentadiene, andthe cyclohexene formula of Wieland.The isolation of cydopentane-7 5 W. H. Hatcher and W. H. Mueller, Canadian J . Res., 1930, 3, 291; A.,1930, 1558.76 H. S. Fry and J. H. Payne, J. Amer. Chem. SOC., 1931, 53, 1973; A.,819.7 7 N. A. Milas, ibid., p. 221; A., 334..J. Boeseken and G. C. C. C. Schneider, J . pr. Chem., 1931, [ii], 131, 285;A., 1150.?* W. C . Smit, Rec. trav. chim., 1930, 49, 691; A., 1930, 1020.J. Boeseken and A. Kremer, ibid., 1931, 50, 827; A., 1050.81 J. Boeseken and G. Slooff, ibid., 1930, 49, 95; A., 1930, 322ORGANIC CHEMISTRY .-PART1 : 3-dicarboxylic acid from amongst the oxidation products of thedihydrogenated dimeride shows that Wieland's formula is correct(KGamer and Spilker)(Wieland)Thus dimerisation of a simple cyclic butadiene is now definitelyknown to follow the course which is characteristic of the open-chainbutadienes.S2 Further polymerisation, leading to the formation oftri-, etc., cyclopentadienes, probably occurs at the reactive ethylenelinking in ring (2).The dimeride of isoprene, obtained together with dipentene andcaoutchouc by heating the hydrocarbon alone or with aceticis a 1 : 3-menthadiene; formerly this com-pound was regarded by Harries as a CHMe A/ \ dimethyloctatriene.Similarly the dimer- Erne ide obtained by heating apyS-tetramethyl-\ / butadiene with concentrated formic acidYHMev CHMe is the met hylpropenylhexamet hylcyclo-Degradation of the triisobutylene pre-pared by the action of sulphuric acid on tert.-butyl alcoholindicates that the polymeride is produced in conformity withthe additive mechanism indicated in the scheme :CMe,:CH, CMe2*CH, CMe,:CH, CMe,*CH,(14 hexene (I). 84+ = I Y I - + CMe,: C [ HI CMe , CMe,:C* CMe , CMe,:CH[H] CMe,:CHTwo additional isomerides of the final product appear, however, tobe formed, and these could arise if double-bond movement occurreda t the dimeride, and at both dimeride and trimeride stages of thepolymerisation, respectively.85If polymerisation of this type be considered to apply to butadienoid ,hydrocarbons, it is obvious that by taking into account both aP- andas-types of addition, long polymeric chains could be formulated inwhich the same isopentene unit occurs indefinitely but the end groups82 K.Alder, G. Stein, and H. Finzenhagen, Annalen, 1931, 485, 223; A.,473.** T. Wagner-Jauregg, ibid., 1931, 488, 176; A., 1031.81 P. van Romburgh and G. van Romburgh, Proc. K . Akad. Wetensch.a5 R. J. RlcCubbin, J . Amer. Chem. SOC., 1931, 53, 356; A., 333.Amsterdam, 1931, 34, 224; A., 81988 FARMER :change according to the additive scheme adopted. Thus, if for theunsymmetrical compound, isoprene, it be assumed that the adden-dum molecule may (in effect) divide in the mannerH-CH:CH*CMe:CH, or H-CH:CMe*CH:CH,and add at the aP- or as-positions of the conjugated monomeride,dimeride, etc., hydrocarbon, then six formulations become possiblein which the unit (CH,*CH:CMe*CH,) = U recurs but the end groupschange ; in this way the polymerides (A), (B), (C), etc., are formed : 86(A) CH,:CMe*CH,*CH,*[U],*CH:CMe*CH:CH,(B) CH,:CH*CHMe*CH,*[U],*CH:CMe*CH:CH,(C) CH,*CMe:CH*CH,*[U],*CH:CMe*CH:CH,On ozonolysis such substances should yield not only large quantitiesof laevulaldehyde, laevulic acid, and lzevulic acid peroxide, but smallquantities of one or more of formaldehyde, pyruvaldehyde, acet -aldehyde or the corresponding acids, etc.The recognition of these end groups and their accurate quantit-ative estimation in relation to the quantity of laxtilaldehyde and thederivatives thereof formed at the same time would give definiteinformation concerning the non-cyclic nature of the polymeride, itsmolecular weight, and the mechanism of its formation.A newattempt to identify and estimate the ozonolysis products of caout-chouc 87 has demonstrated the formation of the laevulic series ofproducts to an extent corresponding to nearly 90% of the carbonskeleton of the polymeride; in addition, small quantities of aceticacid, acetone (regarded as an impurity), a substance fermentable byyeast, carbon dioxide, formic acid, and succinic acid, have beenrecognised. Now, although none of the minor products can bedefinitely stated to have its origin in an end group, there is littledoubt that at least 2% of the acetic acid produced is a primaryfission product of caoutchouc, and not merely a decompositionproduct of laevulic acid; the origin of the fermentable substance(possibly pyruvic acid) is still in doubt.The statement of Harrieshpopos of the cyclic formula for rubber, that no ozone fission productscontaining less than five carbon atoms were to be obtained, is plainlyill-founded. The question of terminal groups remains open and itssolution awaits further improvement of experimental technique.The polymerisation of many substances is actively catalysed byozonides and peroxides. Ozonised mixtures of two isobutyleneshave been found to accelerate the polymerisation of styrene, indene,py-dimethylbutadiene, and furfuryl alcohol but not of stilbene,R. Purnmerer, G. Ebermayer, and K. Gerlach, Ber., 1931, 64, [BJ, 809;86 G. S. Whitby, Trans. Inst. Rubber Id., 1929, 5, 190.A., 733ORGANIC CHEMISTRY.-PART I. 89o-hydroxybenzyl alcohol, and P-methyl-AP-butene.88 The velocityof polymerisation generally increases with increasing concentrationof the catalyst up to 2% ; moreover the catalytic activity is inherentin the original ozonides, and not in any decomposition products, sincethe activity diminishes rapidly on keeping. The polymerisation ofunsaturated hydrocarbons is strongly accelerated by simultaneouslyoccurring auto-oxidation processes-more effectively, indeed, thanis accomplished by addition of peroxides.89 Probably the actualcatalysts in these cases are primary peroxides which have a morepowerful accelerating action than that of the polymeric peroxidesultimately found among the polymerisation products. Polymeris-ation and peroxide-formation (polymeric) take place side by side,in proportions varying with the chemical constitution of the unsatur-ated compound and the availability of oxygen.When acrylic acid is polymerised by irradiation, non-homogeneous,amorphous products result.Polymerisation is considered to followthe scheme :. * CH(C02H)*CH2 - * - + xCO,H*CH:CH, -+*CH(C02H)*CH2.[CH(C02H)*CH2],, 1*CH(C02H)*CH,*Similarly, irradiation of vinyl acetate in chloroform is stated toyield a product of the type Me-CH( OR)*[CH,*CH( OR)],*CH,*CHCl,.The properties of these polymerides and of polyvinyl alcohols aredescribed.90 Further details are available with respect to theformation of polyoxymethylenes from formaldehyde 91 and it hasbeen shown that polymerisation of aqueous formaldehyde withcalcium hydroxide yields a solution which contains pentoses,hexoses, and possibly some heptose~.~~Nitrogen Compound8.The synthesis of the simple amino-acids and their derivatives hasattracted so large a measure of attention that it is impossible in thespace a t the writer’s disposal to do more than indicate the maingroups of compounds which have formed the subject of investigation.Improved methods of carrying out the Strecker and Kolbesynthesis of a-amino-acids have been described in detail ; 93 methods8 8 R.C. Houtz and H. Adkins, J. Amer. Chern. SOC., 1931, 53, 1058; A.,89 H. Staudinger and L. Lautenschliiger, Annakn, 1931, 488, 1 ; A,, 1031.90 H. Staudinger and H. W. Kohlschutter, Ber., 1931, 64, [B], 2091 ; A.,1270; H. Staudinger and A.Schwalbach, Annakn, 1931, 488, 8; A., 1032.91 H. Staudinger, R. Signer, and 0. Schweitzer, Ber., 1931, 64, [B], 398;A . , 465.92 P. Karrer and E. von Krams, Helv. Chim. ActGT, 1931,14, 820 ; A , , 1037.93 W. Cocker and A. Lapworth, J., 1931, 1391; A., 943.59790 FARMER :for preparing a-amino-acids and a m i d e ~ , ~ ~ N-alkylated p-amino-esters , y- amino-, w -amino -, and various dimet hylamino - andbi~dimethylamino-~~ derivatives of fatty acids have been workedout ; and syntheses of dl-aspartic acidYgg creatine and alacreatine,lserine,2 ethyl aminomalonate and its derivative^,^ derivatives ofl y ~ i n e , ~ and various aminohydroxy- and hydroxyamino-acids 6 havebeen recorded. The formylation of amino-acids,' their acetylationby beten,* and the preparation of glycerol esters therefrom have alsobeen s t ~ d i e d .~A series of investigations by A. Skita and F. Keil and theircollaborators lo deals with the production of the more complexamines by the catalytic reduction of carbonyl compounds in thepresence of simple amines. In most cases colloidal platinum is usedas catalyst. Schiff's bases, or mixtures of the aldehyde and primaryamines from which the bases are formed, yield secondary amines onreduction ; aldehydes of high molecular weight give secondaryamines with ammonia, but the simpler aldehydes yield tertiaryamines. Ketones react usually with aliphatic amines to produceN-substituted imines which are further reducible to secondaryamines : under suitable conditions, however, substantial amountsQp T.Curtius and others, J. pr. Chem., 1930, [ii], 125, 211; A., 1930, 752;H. Krause, Chem.-Ztg., 1931,55, 666; A., 1402; W. Cocker and A. Lapworth,J., 1931, 1894; A., 1402; R. Locquin and V. Cerchez, Bull. SOC. chim., 1930,[iv], 47, 1386; A., 1931, 338; D. M. Birosel, J . Amer. Chem. SOC., 1931, 53,3039; A., 1149; P. S. Yang and M. M. Rising, ibid., p. 3183; A., 1160.95 J. DQcombe, Compt. rend., 1930, 191, 945; A., 1931, 76.Q6 A. Miiller and E. Feld, Monutsh., 1931, 58, 22 ; A., 943.s7 B. Flaschentrgger with F. Halle, T. Hosoda, F. Gebhardt, and B. Blech-man, 2. physiol. Chem., 1930, 192, 245, 249, 253, 257; A., 1931, 76.Q8 V. Prelog, Coll. Czech. Chem. Comm., 1930, 2, 712; A., 1931, 204.Qs M. S. Dunn and B.W. Smart, J. Biol. Chem., 1930,89,41; A , , 1931, 76.H. King, J., 1930, 2374; A . , 1930, 563.S. K. Mitra, J . Indian Chem. SOC., 1930, 7, 799; A., 1931, 205.R. Locquin and V. Cerchez, Bull. SOC. chim., 1930, [iv], 47, 1274, 1252,1287, 1377, 1381; 1931, [iv], 49, 42, 47; A . , 1931, 205, 338, 471.R. Enger and H. Steib, 2. physiol. Chem., 1930,191,97; A., 1930, 1419;R. Enger and F. Halle, ibid., p. 103 ; A., 1930, 1420 ; R. Enger, ibid., p. 117 ;A . , 1930, 1420.W. J. N. Burch, J., 1930, 310; A . , 1930, 460.L. W. Jones and R. T. Major, J . Amer. Chem SOC., 1930,52, 1078; A . ,1030, 754; M. Tomita and J. Karashima, 2. physiol. Chem., 1930, 187, 238;A . , 1930, 585.R. E. Steiger, J. Biol. Chem., 1930, 86, 695; A., 1930, 752.M. Bergmann and F.Stern, Ber., 1930, 63, [BJ, 437; A., 1930, 459.L. Haskelberg, Compt. rend., 1930, 190, 270; A., 1930, 328.lo Ber., 1928, 61, [B], 1452, 1652; 1929, 62, [B], 1142; A., 1928, 1120,1228; 1929, 808; Monatsh., 1929, 53, 753; A., 1929, 1436; Ber., 1930, 63,[B], 34; A., 1930, 327ORGANIC CHEMISTRY.-PART I. 91of primary amines may arise. From diketones and primary amines,N-substituted amino-alcohols can be obtained ; in this way a-phenyl-propane-ap-dione, when reduced together with methylamine, givesephedrine-a method of preparation for the latter which had beenindependently employed by R. H. F. Manske and T. B. Johnson.llTertiary amines are formed by the reduction of nitriles, alone or inthe presence of ketones, and in some cases also by the reduction ofaldehydes admixed with secondary amines.Acylvinyl- and arylvin yl-amines of the general formulaR=CO*CH:CH*NR,R,may be conveniently prepared by the action of ammonium chlorideor the hydrochloride of an amine on the sodium salt of aliphatic andaromatic hydroxymethylene ketones.l2By the action of dry hydrogen sulphide on dry amines (usuallyin ether at 0") M. Achterhof, R. F. Conaway, and C. E. Boord haveprepared a series of amine hydrogen ~u1phides.l~ The salts from themore volatile amines decompose in air, giving sulphur, but thosefrom the less volatile amines suffer oxidation to the correspondingthiosulphates.The structure of sphingosine derived from cerebroside mixtureshas been the subject of renewed in~estigati0n.l~ It is now foundthat, whereas the ozonolysis of triacetylsphingosine yields tetra-decoic acid (and aldehyde) together with a-amino- py-dihydroxy-n-butyric acid, the oxidation of dihydrosphingosine yields palmiticacid; moreover the oxidation of sphingosine sulphate also givestetradecoic acid.It is concluded, therefore, that sphingosine has theconstitution CH,*[CH,],,*CH:CH*CH( NH,)*CH( OH)=CH,*OH.New derivatives of asparaginedialdehyde l5 and hydrazinoprop-aldehyde16 have been obtained and many new examples of theformation of hydrazides, azides, and carbamides of mono- andpoly-carboxylic acids and lactones are furnished by the researchesof T. Curtius and his collaborators.17Muscarine has been successfully purified by the precipitation and11 J .Amer. Chem. SOC., 1929, 61, 580; A., 1929, 441.12 E. Benary, Ber., 1930, 63, [B], 1573; A., 1930, 1026.13 J . Amer. Chem. SOC., 1931, 58, 2682; A., 1041.14 E. Klenk, 2. physiol. Chem., 1929,185,169; A., 1930, 73; E. Klenk andW. Diebold, ibid., 1931, 198, 25; A., 829.1 5 A. Wohl and E. Bernreuther, Annalen, 1930, 481, 1 ; A . , 1930, 1021.16 A. Wohl and A. Pranschke, Ber., 1931, 64, [B], 1381; A . , 937.17 T. Curtius, G. von Briining, and H. Derlon, J. pr. Chem., 1930, [ii], 125,6 3 ; A., 1930, 766; T. Curtius and W. Sandhaas, ibid., p. 90; A., 1930, 757;T. Curtius and H. Sauerberg, ibid., p. 139; A., 1930, 757; T. Curtius, W.Sieber, F. Nadenheim, 0. Hambsch, and W. Ritter, ibid., p. 152; A., 1930,766 ; T. Curtius and W. Dorr, ibid., p.425 ; A., 1930, 75692 FARMER :crystallisation of its salt with the radical [(NH,),Cr(SCN),]- ; analysisof the salt indicates the formula [C,H,,O,$]+ for the muscarine ion.Since muscarine contains a hydroxy-group, gives trimethylamineand dihydroxy-n-valeric acid by the Hofmann degradation, andbehaves as an aldehyde, it is considered to have the constitutionOH*CHEt*CH(CHO)*NMe3*OH or (less probably)CHO CII ( OH ) *CHE t*NMe,*OH. l8Cystine may be satisfactorily reduced to cysteine by the actionof sodium in liquid ammonia,lg by aluminium amalgam in aqueoussolution,20 or catalytically in the presence of spongy palladium.21Treatment of a solution of cystine in dilute sulphuric acid with silversulphate, followed by neutralisation to pH 5-6, causes precipitationof 70--85(70 of the total nitrogen in the form of a compound(C,H50,NSAg),,Ag,S0,.Decomposition of this compound withhydrochloric acid yields a solution containing more than 90% of itsnitrogen in the form of cysteine.22 The silver compound differsfrom the apparently normal silver salt of cystine and is regarded asa silver mercaptide.The question of the ketimine-enamine tautomerism of B-amino-acrylic esters has been examined by K. von Auwers and W. Susemihl 23with the aid of refractivity determinations. Ketimino-forms,because they contain no conjugated system, should display normaloptical behaviour : enamines, on the other hand, with the systemN*C:C*C:O, should exhibit marked exaltation. The optical data,particularly in the case of ethyl p-aminocrotonate and its derivatives,are shown to be in harmony with the constitutionbut cannot be explained in relation to the alternative structuresNHR-CMe:CH*CO,Et ,YHMeNR:CMe*CH,*CO,Et and NH - >CH*CO,Et.Marked differencesare observed between the keto-enolic and ketimine-enamine typesof tautomerism, since in the latter type an equilibrium mixturedependent on the nature of the compound and the external con-ditions does not appear to exist in the molten material in which theenamine form predominates. The conditions which determine thedegree of enolisation in oxygen compounds appear to be little, if atall, active with corresponding nitrogen compounds.18 F. Kog1,H. Duisberg, and H. Erxleben, Annulen, 1931,489,156 ; A., 1279.19 V.du Vigneaud, L. F. Audrieth, and H. S. Loring, J . Anzer. Chem. SOC.,20 E. Gebauer-Fiilnegg, ibid., p. 4610; A . , 1931, 76.21 M. Bergmann and G. Michalis, Ber., 1930, 63, [B], 987; A . , 1930, 754.22 H. B. Vickery and C. S. Leavenworth, J . Biol. Chem., 1930, 86, 129;2s Ber., 1030, 63, [B], 1072; A., 1930, 897.1930, 52, 4500; A., 1931, 76.A., 1930, 754ORaANIC CHEMISTRY.-PBRT I. 93According to A. Hantzsch, a comparison of the ultra-violetabsorption of t ric hloroace t amide , its pi peridine - derivative,CCl,*CO*NC,H,,, and the iminoether CCI,*C( 0Me):NH in methylalcohol, water, and chloroform indicates that the equilibriumCCI,*CO*NH, CCl,*C( 0H):NH is displaced extensively to theright in chloroform and further in water; similarly, benzamideappears to be almost completely dissolved as the iminohydrinPh*C(OH):NH.M Comparison of the boiling points of the amides,methylamides, and iminoethers shows that amides and methyl-amides exist mainly in the iminohydrin form R*C(OH):NH andR-C(OH):MMe.The anomalies in the boiling points of the amidesand methylamides are attributed to association caused by thepresence of the hydroxyl groups and consequent difficulty in passinginto the unimolecular condition. In contrast to the acid amides,thioacetamide appears to be present as the true thioamideCH,*CS*NH, in chloroform or ether and only to a minor extent asthe iminothioether in water.Hantzsch further considers that the two additive productsderivable from hydrogen halides and nitriles are respectively identicalwith the products obtained by the interaction of amides and phos-phorus pentachloride and hitherto regarded as imide chlorides,R*CCI:NH, and amide chlorides, R*CC&-NH,.25 They are actuallynitrilium salts, the former being monohalides [R*C"NH]Cl, and thelatter dihalides [R*CNH]Cl,HCl, in which the second acid moleculeis not structurally combined.Nitriles dissolved in concentratedsulphuric acid are present as completely dissociated sulphates[R*C_IUH]SO,H ; amides also when dissolved by sulphuric acidyield completely dissociated sulphates. Aromatic amides are con-sidered to yield iminohydrinium salts [Ar*C( OH):NH,]SO,H inconformity with the view that the free amides are preponderatinglyiminohydrins, ROC( 0H):NH ; the true tertiary amides, Ar*CO*NR,,however, also yield similarly constituted iminohydrinium salts.The transformation of nitrilium sulphate into iminohydriniumsulphate takes place slowly but completely at the ordinary tem-perature. The conversion of nitrile into amide in aqueous solutiondepends on the formation of nitrilium salts, which pass by additionof water into iminohydrinium salts: these are hydrolysed toamides.A ketenimino-structure is also supposed to be assumed by aceto-nittile in reacting with ethylmagnesium bromide, the product,24 Ber., 1931, 64, [B], 661 ; A., 608.2s Ibid., p. 667; A., 608; compare J.Verhulst, Bull. SOC. chim. Belg.,1930, 39, 563; A., 1931, 47194 FBMER :CH,:C:N*MgBr, reacting with water (addition at the ethyleniclinkage) to yield some aceta'mide.26Saturated nitriles are obtainable in good yield by passing many ofthe members of the fatty acid series, together with ammonia, overheated silica gel.27 New details relating to the preparation of anumber of clp-(cis- and trans-) and py-unsaturated nitriles 28 arerecorded.It is a matter of considerable interest that hydrogen chloride hasbeen found to combine with liquid (or dissolved) hydrogen cyanide,to yield the sesquichloride 2HCN,3HC1; hydrogen bromide andiodide also yield sesquihalides with hydrogen cyanide, but theformation of no such compound as HCN,HCl has been observed.29The sesquichloride, when heated alone or with liquid hydrogencyanide, or when kept in a vacuum over sodium hydroxide, wasfound to yield chloromethyleneformamidine, NH:CH-N:CHCl, ahygroscopic solid: this when heated with quinoline yielded abimolecular hydrogen cyanide-evidently imidoformylcarbylamine,NH:CH*N:C.Mono-, Di- and Tri-saccharides.In last year's Report reference was made to C.S. Hudson'scontention that change of ring form may and does occur during themethylation of sugars, certain of W. N. Haworth's conclusionsrespecting the structural representation of the sugars thereby beingvitiated. The validity of the principle of optical superposition, asutilised by Hudson in arriving at his suggested formulze for certainof the mono- and di-saccharides, has been examined by W. N.Haworth and E. L. E r s t in considerable detail and it is shown thatreliance on the principle in classifying the sugars leads to seriousinconsistencies and necessitates the postulation of structures a tvariance with the bulk of the experimental evidence.30In particular, Hudson ascribes to a-mannose and a-methyl-mannoside a different ring structure from that of p-mannose anda- and p-glucose and their glucosides and tetra-acetates, and assumes26 G.Mignonac and C. Hoffmann, Compt. r e d . , 1930, 191, 718; A., 1930,1564.27 J. A. Mitchell and E. E. Reid, J . Amer. Ch,em. Soc., 1931, 53, 321; A . ,339.28 R. Breckpot, Bull. SOC. chirn. Belg., 1930, 39, 462; A . , 1931, 194; G .Heim, ibid., p. 458 ; A . , 1931, 205 ; P. Colmant, ibid., p. 568 ; A., 1931, 472 ;P. Bruylants and H. Minetti, BuEl. Acad. TOY.Belg., 1930, [v], 16, 1116; A.,1931, 205; P. Bruylants, ibid., 1931, [v], 17, 1008; A . , 1403; P. Bruylantsand L. Ernould, ibid., p. 1027; A., 1403.L. E. Hinkel and R. T. Dunn, J . , 1930, 1834; A . , 1930, 1421.30 J., 1930, 2615; A . , 1931, 200; compare K. Freudenberg, Sitzungsber.Heidelberger Akad. Wiss., 1931, No. 2, 3 ; A . , 825ORGANIC CHEMISTRY .-PART I. 95that the true structural isomeride of p-mannose (and a- and p-glucose,etc.) should be an unknown form of or-mannose having [.ID 4- 77"'since such a value would bring the rotational difference between thisu- and the known p-form into line with those that are found for thetwo glucoses. Hudson's scheme moreover requires that this varietyof u-mannose, and not the normal variety (+ 30")' occurs as acombined residue in 4-glucosido- a-mannose and in 4-galactosido-a-mannose.It has now been conclusively proved by hydrolysis with emulsinthat the mannoside residue in 4-glucosido- and 4-galactosido-a-methylmannoside is the normal pyranoside variety of a-methyl-mannoside and not the hypothetical form of Hudson; 31 moreoverit has been demonstrated that both forms (a- and p-) of methyl-mannoside which are obtainable in the normal manner from mannosehave an identical six-atom structure 32 and that derivatives of4-glucosido- and 4-ga~ac~osido-ma~ose retain unimpaired theirstructural rigidity during methylation.33 From the manner inwhich the optical rotatory power of different methylated lactonesvaries from solvent to solvent there can be little doubt that con-figurational relationships, e.g., the cis- relationship of methoxylgroups in the methylated molecules, have a profound influence onthe optical rotatory powers of substances which cannot alter theirring structure under the infiuence of solvent and are to be lookedto as the source of some of the observed discrepancies in correlatingrotational data.Other demonstrations of the presence of a pyranose structurerefer to penta-acetyl p-d-mann~se,~~ P-methylgala~toside,~~ gly-curonic diisopropylidenegalac t ~ s e , ~ * and 5-met hoxyke tose .39The phenomenon of acyl migration in sugars and polyhydric31 W. N. Haworth, E. L. Hirst, H. R. L. Streight, H. A. Thomas, and J. I.Webb, J., 1930, 2636; A., 1931, 200; W. N.Haworth, E. L. Hirst, (Miss)M. M. T. Plant, and R. J. W. Reynolds, ibid., p. 2644; A., 1931, 200. Com-pare H. S. Isbell, Bur. Stand. J. Res., 1930, 5, 1179; A., 1931, 201.32 W. N. Haworth, E. L. Hirst, H. G. Bott, and R. S. Tipson, J., 1930,2653; A., 1931, 200.s3 W. N. Haworth, E. L. Hirst, and H. R. L. Streight, ibid., 1931, 1349;A., 939; W. N. Haworth, E. L. Hirst, and (Miss) M. M. T. Plant, ibid., 1931,1354; A., 939.34 W. N. Haworth, E. L. Hirst, and J. A. B. Smith, ibid., 1930, 2659; A.,1931, 200.35 P. A. Levene and R. S. Tipson, J. Biol. Chem., 1931,90, 89; A . , 337.36 A. Muller, Ber., 1931, 64, [B], 1820; A., 1039.9 7 J. Pryde and R. T. Williams, Nature, 1931, 128, 187; A., 1036.88 P. A. Levene and G. M. Meyer, J. Biol. Chem., 1931, 92, 257; A.,88 E.F. Hersant and W. H. Linnell, Quart. J. Pharrn., 1931, 4, 52;1276.A , , 82596 FARMER :alcohols has attracted further attention. Certain changes in thepolarity of the acyl group appear to increase the tendency towardsthe formation of cyclic dioxolane derivatives by attracting themobile hydrogen to the CO-group. p-Hydroxyethyl trichloro-acetate, for instance, is unobtainable in the open-chain form : allattempts to prepare it yield the ortho-ester isomerideF*,*O OHIn other instances the strong tendency towards hydrogen migrationfacilitates fission of the acetate molecule into two molecules ofaldehyde on heating, e.g., benzyl trichloroacetate yields benzaldehydeand chloral (CH,Ph.O*CO*CCI, -+ Ph-CHO + CCI,*CHO) .40New examples of the appearance of " y "- or ortho-ester-forms ofsugar acetates have been reported for triacetyl methyl-d-riboside 41and hepta-acetyl 4-glucosidomethylmannoside 42 and migration ofacetyl occurs during the methylation of 2 : 3 : 4-triacetyl a-methyl-glu~oside.~~ The kinetics of acyl migration have been studied forsolutions of 3-acetylisopropylideneglucose by K.Josephson,44 whofinds that a proportionality exists between the rate of acyl migrationand the concentration of hydroxyl ion over a considerable range ofacidity. The '' y "-form of monoacetyl methylrhamnoside, how-ever, is unstable in acid solution and the change whereby theglycosidic methyl group of this compound is hydrolysed, and theacetyl group transferred wholly to Cz, appears to be catalysed byhydrogen ions .45Several interesting observations have been made with regard tothe properties of the free aldehyde forms of sugars and the occurrenceof the aldehyde phase in certain ordinary sugar reactions.Appar-ently the free aldehyde form of galactose is produced in the form-ation bf the phenylhydrazone, since after acetylation the product isidentical with the phenylhydrazone generated from penta-acetylaldehydogala~tose.~~ Also the ring form of glucose oxime wouldseem to open to the aldehydo-form when submitted to the Wohlprocedure for the production of glucononitrile : a t least, hexa-acetylaldehydo-glucose oxime (I) is formed in the process and yieldspenta-acetyl glucononitrile by loss of acetic acid, whereas the cyclic40 H.Hibbert and M. E. Greig, Canadian J . Res., 1931, 4, 254; A., 820.41 P. A. Levene and R. S. Tipson, J . Biol. Chem., 1931, 92, 109;43 H. S. Isbell, J. Amer. Chem. SOC., 1930, 52, 5298; A., 1931, 201.43 W. N. Haworth, E. L. Hirst, and (Miss) E. G. Teece, J., 1931, 2858.44 Ber., 1930, 63, [Bl, 3089; A., 1931, 199.45 w. N. Haworth, E. L. Hirst, and H. Samuels, J., 1931, 2861.46 M. L. WoIfrom and C. C . Christman, J . Amer. Chem. Soc., 1931,53,3413 ;C€€,*O>C<CC1,~A., 938.A,, 1276ORGA.NIC! CHEMISTRY .-PART I. 97hexa-acetyl @-glucose oxime (11) which is produced at the same timedoes not yield the nitrile when heated.47HQ*OAc 1 Ac ( Ac O)N*QHAcsO+ AcO*QH bHQ*OAc QHZNOAc Ho*7JqAcsO+ HO*$?H %GG-AcO*VHH Hw :FJ= CH,*OH CH,*OAcPenta-acetyl aldehydogalactose forms well-defined crystallinecompounds with alcohol and with water.These, unlike the freepenta-acetyl aldehydo-sugars themselves, show mutarotation inalcohol-free chl~roform.~~ The curve for the alcoholate falls rapidlyto a minimum and then increases rather less rapidly to a value higherthan the original ; its form points to the presence in solution of twosemi-acetal individuals in equilibrium with the aldehyde and alcohol :(11.)HT*OAcHQ*OAc(1.)CH,*OAcOH OEtThe aldehydrol, on the other hand, under like conditions shows onlya steady fall in rotation along an exponential curve indicating asimple unimolecular decomposition into free aldehyde and water.Lactose phenylosazone also exhibits mutarotation in methyl alcoholbut not in anhydrous alcohol-free solvents ; the mutarotation ofosazones does not appear to arise from tautomerism betweendihydrazone and azo-forms, since fructose methylosazone, which doesnot possess a labile hydrogen atom, exhibits mutarotation.48~Constitutions have been established, as the result of synthesis,of two natural glucosides : aesculin is formulated as 6-glucosidoxy-7-hydroxyco~rnarin~~ and euxanthic acid as the compound (I) inwhich the sugar residue is glycuronic acid, probably attached to thenucleus by a p-glucosidic linking.50 In addition, the formuhtion ofmonotropitoside (gaultherin) as a primeveroside of methyl salicylate(111) has been synthetically demonstrated.51The aldobionic acid from gum arabic, referred to in last year'sReport, is now shown to be a glycuronosido-galactose in which the4 7 M.L. Wolfrom and A. Thompson, J . Amer. Chem. SOC., 1931,53, 622 ;A., 467.48 M. L. Wolfrom, ibid., p. 2275; A . , 1039.4 h E. VotoEek and F. Vdentin, Arhiv Hemiju, 1931,5, 155 ; A., 1274.49 F. S. Head and A. Robertson, J . , 1930, 2434; A., 1931, 73.50 A, Robertson and R. B. Waters, ibid., 1931, 1709; A., 1040.5 1 Idem, ibid., p. 1881; A., 1400.REP.-VOL. XXVm. 98 FARMER :reducing group of the glycuronic acid residue is joined to the 6-posi-tion of the galactose residue. 52 Sedosan (anhydrosedoheptose) isformulated as the 2 : 7-anhydropyranose (II).538-(R = o-C,H,*CO,Me.)The hexosemonophosphate obtained by R. Robison throughfermentation of glucose or fructose in presence of sodium phosphateis reported to be glucose-6-phosphate, which has now been synthesisedfrom isopropylideneglucose ; 54 methylation of Robison’s productis reported to yield two distinct forms of the corresponding methyl-glycoside.55The saccharals of epimeric sugars prove to be identical, as wouldbe expected : d-glucal is identical with d-mannal, and d-xylal withd-lyxa1.56 It has also been found that tetra-acetyl hydroxyglucalyields the same osazone as tetra-acetyl hydroxygalactal,57 and, byaddition of chlorine, followed by hydrolysis, yields both tetra-acetylglucosone hydrate and triacetyl glucosone hydrate. 58 Bothglucal 59 and +-glucal 6o have pyranose structures, and the latter,which has the double bond in the 2 : 3-position7 constitutes theunsaturated unit in +-cellobial ; the reduction products of +-glucaland $-cellobial, vix., 2 : 3-dideoxyglucose and 2 : 3-dideoxycellobiose,are also pyranose forms.It is suggested as the result of oxidation(hydroxylation) experiments with galactal that the group attachedto the C,-atom in glucals (e.g., OH, OMe, or OAc) may have a direct-ing influence on the spatial disposition of the hydroxyl group which5 2 S. W. Challinor, W. N. Haworth, and E. L. Hirst, J . , 1931, 258; A . ,53 H. Hibbert and C. G. Anderson, Canadian J . Res., 1930,3,306 ; A . , 1930,54 P. A. Levene and A. L. Raymond, J. Biol. Chem., 1930, 89, 479; 1931,5 5 E. J. King, R. R. McLaughlin, and W. T. J. Morgan, Biochem. J., 1931,56 M.Gehrke and F. Obst, Ber., 1931, 64, [ B ] , 1724; A., 1038.57 M. Bergmann and L. Zervas, ibid., pp. 1434, 2032; A., 939, 1275.5 8 K. Maurer and W. Petsch, ibid., p. 2011; A., 1275.59 E. L. Hirst and C. S. Woolvin, J . , 1931 1131; A., 1399.6o 3%. Bergmann and W. Freudenberg, Ber. 1931, 64, [ B ] , 158; A . , 336.465; see Ann. Reports, 1930, 27, 109.1560.91, 751;25, 310; A., 600.1931, 92, 757, 765; A . , 1931, 63, 933, 1148ORGANIC CHEMISTRY .-PART I. 99enters a t the C,-atom, thus being responsible for the predominatingcis- or tmns-dispositions at the C,- and C,-atoms produced inaccordance with the scheme :Alcohols react with acetohalogeno-sugars (mono- or di-saccharides)in the presence of mercuric acetate to produce glycosides of the sugaracetates, the a- or p-configuration of which is to some extentgoverned by the proportions of the reactants.62 By replacing thealcohol with a sugar derivative possessing one free hydroxyl group,the formation of the acetyl derivatives of di- and tri-saccharides canbe effected with some control over the nature (a- or p-) of the glyco-sidic linking.63 A somewhat similar method of coupling, employingsilver carbonate, calcium chloride and iodine in place of mercurysalts, has been applied to the synthesis of tri- and tetra-saccharideacetates.64Recent observations in relation to turanose and melezitose 6 5indicate that these sugars are more correctly represented by thestructures originally assigned to them by G. Zemplkn and G.Braun G6 than by those subsequently assigned by G.C. Anew disaccharide, neotrehalose, synthesised from triacetyl glucose1 : %anhydride and tetra-acetylglucose, appears to be the hithertounknown a-glucosidyl- p-glucoside. 6*Polysaccharides.Indin.-Following the suggestion that glucose may occur as aproduct of hydrolysis of i n ~ l i n , ~ ~ it has been pointed out that the61 P. A. Levene and R. S. Tipson, J . Biol. Chem., 1931,93, 631 ; A., 1400.62 G. ZemplBn, Ber., 1929,62, [B], 990; A., 1929, 683; G. Zemplh andZ. S. Nagy, ibid., 1930,63, [B], 368 ; A., 1930,456 ; G. Zemplh, ibid.,p. 1820;A., 1930, 1167; G. Zemplh and A. Gerecs, ibid., p. 2720; A., 1931, 72.63 G. Zemplbn, Z. Bruckner, and A. Gerecs, ibid., 1931, 64, [ B ] , 744;A., 716; G.Zempldn and A. Gerecs, ibid., p. 1545; A., 1039; G. Zemplbnand Z. Bruckner, ibid., p. 1852; A . , 1040.64 B. Helferich and R. Gootz, ibid., p. 109; A . , 337.6 5 E. Pacsu, J . Amer. Chem. SOC., 1931, 53, 3099; A., 1149.66 Ber., 1926, 59, 2230; A . , 1926, 1229.6 7 J . , 1927, 588; A., 1927, 450.68 W. N. Haworth and W. J. Hickinbottom, ibid., 1931, 2847.60 Ann. Reports, 1929, 28, 102100 FARMER :iodine value of the acid hydrolysis product of inulin is compatiblewith a glucose content in inulin of 5y0.7* Recent experiments ofR. F. Jackson and E. McDonald, however, show that inulin fromdifferent sources (members of the family Composita?) gives onhydrolysis with acid about 92% of fructose, 3% of glucose, and 5%of a group of refractory difructose anhydride^.^^ This group ofdisaccharides is stated to be invariably a constituent of inulinregardless of the plant from which the inulin is extracted or of thedegree of purification to which the inulin is subjected.Sincepolysaccharides related to inulin do not yield the disaccharides, it isconcluded that the latter form an integral part of the inulin moleculeand are not produced by side reactions occurring during hydrolysis.The occurrence of the disaccharides in such constant proportionsregardless of the source of the inulin or of its degree of purificationis taken to indicate that inulin is essentially a homogeneous sub-stance; in other words, it is composed of similarly constructedmolecules, each of which contains the refractory difructose residues.The whole molecule must then contain not less than 110 hexoseresidues, which upon condensation and loss of 110 mols.of water,would produce a polysaccharide having the molecular weight 18,000(compare E. Berner’s value of 3,500 and H. D. K. Drew and W. N.Haworth’s value of not less than about 4,000).Three crystalline difructose anhydrides have been isolated byJackson and McDonald which differ from the anhydrofructoseisolated by J . C. Irvine and J. W. Stevenson.72 A new difructoseanhydride acetate has been derived by the action of inulinase oninulin,v3 but the corresponding difructose anhydride is not as yetknown to be different from all the compounds of Jackson andMcDonald.It is suggested that inulin-bearing plants contain a series ofpolysaccharides of varying complexity which culminate in inulinitself.71 As the series is ascended there is a diminishing ratio ofglucose to fructose until inulin is reached.In the dahlia the develop-ment of the more complex members of the series has proceeded tosuch an extent that inulin is the predominating polysaccharide : amore nearly complete elimination of glucose is therefore to beexpected in the dahlia inulin than in that of plants which appear tohave less ability to synthesise inulin.‘O 13. Pringsheim, J. Reilly, and others, Ber., 1930, 63, [B], 2636; A . , 1930,7 1 Bur. Stand. J. Res., 1930, 5, 1151; 1931, 6, 709; A., 1931, 202,72 J . Amer. Chem. Soc., 1929, 51, 2197; A., 1929, 1046.1562.827.H. Pringsheim and W.G . Hensel, Ber., 1931, 64, [B], 1431 ; A., 941ORGANIC CHEMISTRY.-PART I. 101Lc~vun.-H. Hibbert and his collaborators 74 accept the view thatin the growing plant glucose units which later unite to form starchand cellulose are convertible, to a considerable extent, into fructosewith the resulting formation of l z v u l ~ s a n s . ~ ~ The wheat plant,after growing spikes, contains in its stem an alcohol-insolublesubstance which yields a large amount .of fructose; immaturewheat grains, however, contain 6% of laevulosan and an equalpercentage of starch; as maturity approaches, starch is formed inincreasing quantities (50-60 yo) while lzvulosan gradually falls toabout 0.4%. The presence of these polymerised anhydrofructoses(lzemlosans) is held not to be confined to any group of plants : onthe contrary, their occurrence is widespread and the existence of anumber of polymerides intermediate between fructose anhydride andinulin is to be expected.Although various observers agree that bacteria can synthesiselaevulosans from sucrose, and, indeed, new preparations of lzevanfrom sucrose have been obtained by the action of B.mesentericusand B. ~ubtilis,~~, 76 no case appears to have been recorded in whichbacterial action has produced a lzevulosan from glucose, as appearsto occur in the plant. Lzevan is apparently produced only fromthose sugars containing the furanose ring of fructose. Theproduct obtained by employment of B. mesentericus gives a 97%yield of fructose; it also gives a triacetate differing markedly fromthat of inulin, and a trimethyl derivative which gives 1 : 3 : 4-tri-methylfructofuranose on hydrolysis.The polysaccharide appearsto be built up from the anhydro-fructofuranose unit with linkingsat positions 2 and 6 of the fructose chain and is probably to berepresented : 74HO H HO H HO HCH,-0The hydrolysis of laevan with 0-1N-oxalic acid at 65" proceeds a tapproximately the same rate as that of sucrose and inulin.77Starch.-The view that natural starch is composed of amylopectinand amylose is not generally accepted. These substances arereported to be formed by the successive depolymerisation and74 H. Hibbert, R. S. Tipson, and F. Brauns, Canadian J. Res., 1931,4,221;75 H. Colin, Bull. SOC. Chirn. biol., 1925, 7, 173; A., 1925, i, 618.7 6 H.Hibbert and F. Brauns, Canadian J . Res., 1931,4, 596; A., 1276.7 7 H. Hibbert and E. G. V. Percival, J. Amr. Chem. SOC., 1930, 52, 3996;A., 827.A . , 1930, 1661102 FARMER :repolymerisation of natural starch, and their distinctive character-istics to arise from differences in physical character or from admixturewith other substances. '8 A substance containing both phos-phorus and nitrogen is reported to be present in potato- and wheat8-starch, combined with the polysaccharide by means of a radicalcontaining phosphorus ; the bulk of both phosphorus and nitrogencan be extracted, the characteristics of the starch becoming some-what changed in the process. The phosphorus content of certainpotato-starch degradation products has been found to attain a valuebetween that corresponding to a disaccharide-H3P0, and a tri-saccharide-H3P0, aggregate.79W. N. Haworth and E. G. V. Percival have shown that the biosederivable by degradation of trimethylamylose with acetyl bromidecontains two glucopyranose units and conclude that the essentialstructure of the original polysaccharide comprises maltose residuesjoined to form a chain of a-glucopyranose units thus : 80Treatment of anhydrous maltose with hydrofluoric acid is foundto give a polymeride (maltan) which is indistinguishable from asimilar substance (amylan) obtainable from anhydrous potato-starch by the same treatment.81 The isolation of various newamylosans is reported.S2The glycogens obtained from different sources are stated to beidentical, observable distinctions being attributed to differences inthe degree of polymerisation and of esterification with phosphoricacid.83 Degradation experiments carried out with trimethylglycogen show that glycogen is undoubtedly constituted on the samegeneral plan as starch and consists largely of a-glucopyranoseresidues joined in a continuous chain as shown for starch above.8078 J. Effront, Ann. Xoc. Zymol., 1930, 2, 1; A., 1931, 469; S . von Naray-Szab6,Z. physikal. Chem., 1930,151,420; A., 1931,202. Compare M. Samec,E. Pehani, and J. Stojkovib, Kolloidchern. Beih., 1931, 33, 103; A., 941. '' M. Samec and W. Bendger, Kolloidchem. Beih., 1931, 33, 95; A . , 941;M. Samec, S. Seligkar, and V.Zitko, ibid., p. 449; A., 1277.8o J . , 1931, 1342; A . , 941.B. Helferich, A. Starker, and 0. Peters, Annalen, 1930, 482, 183; A . ,1931, 74.82 H. Pringsheim, A. Wiener, and A. Weidinger, Ber., 1930, 63, [B], 2628;A., 1930, 1561 ; H. Pringsheim, A. Weidinger, and P. Ohlmeyer, ibid., 1931,64, [B], 2125; A . , 1277.83 K. M. Daoud and A. R. Ling, J . SOC. Chem. Id., 1931, 50, 365, 379;A., 1277ORGANIC CHEMISTRY.-PART I. 103Cellulose.-Further considerations relating to the molecular sizeof cellulose as deduced from viscosity measurements have been putforward by H. Staudinger and 0. Schweit~er.~~ The viscosity ofsolutions of purified cotton in Schweitzer’s reagent diminishes withtime without becoming constant after four weeks, the change depend-ing on the oxidative degradation of the macromolecules of thecellulose by copper oxide.Calculation of the molecular weight ofcellulose from the viscosity of its solutions in Schweitzer’s reagentshows that a more or less profound degradation of the moleculeoccurs according to the method of purification. The highestobserved value (190,000) is for purified cotton ; slight degradationaccompanies mercerisation. Sulphite-cellulose and copper-silk aremore appreciably affected (mol. wt. 80,000-65,000 and 35,000respectively).The general trustworthiness of Staudinger’s procedure in applyingthe measurement of viscosity to the determination of the molecularweight of cellulose and its derivatives is, however, questioned byK. Hess and his co11aborators,*5 who point out that the viscosimetricproperties appear to be dependent on fibre purification (and accom-panying disorganisation), during which foreign matter having agreat influence on the viscosity is removed.Treatment of the fibrewith 2% caustic soda and dilute hypochlorite solution, for example,is stated to remove a, non-cellulose material localised on the outerlayers, and a further component is removed by chlorine dioxide.Other workers as a result of a study of the action of dilute acids oncellulose have regarded the latter as a mixture of a crystalline phase(the source of polysaccharides) enveloped in an amorphous skinwhich is destroyed under more or less drastic conditions.s6It is contended, however, that the viscosity method is justified bythe complete analogy which exists between the behaviour of thesesubstances and that of synthetic, highly polymerised compounds.87With both synthetic and natural products, deviations from theHagen-Poiseuille law become more marked with increase in mole-cular weight and more noticeable in concentrated than in dilutesolution. They are not observed in very dilute solution (sol solution)in which the molecules have freedom of movement. It is main-tained that diminution of the viscosity of cellulose solutions withincreasing purification of the preparations is not due to the removal*4 Ber., 1930, 63, [B], 3132; A., 1931, 202.85 K. Hess, C. Trogus, L. Akim, and I. Sakurada, ibid., 1931,64, [B], 408;A., 469; K. Hess and I. Sakurada, ibid., pp.1174, 1183; A., 828; compareK. H. Meyer and H. Mark, ibid., p. 1999; A., 1276.86 Bouchonnet, Jacquet, and Mathieu, Bull. SOC. chim., 1930, [iv], 47,1266; A., 1931, 202.8’ H. Staudinger, Bet-., 1931, 64, [B], 1688; A., 1040104 ORGANIC CHEMISTRY.-PART I.of incrusting matter, but to the degradation of the sensitivemacromolecules.Formaldehyde appears to have an important influence on cellulosewhen applied in mercerisation operations with sulphuric acid or inthe action of acetic anhydride-acetic acid mixture at 100°.88 I nthe former case the fibres are protected and an enhanced affinity forsubstantive dyes is conferred, and in the latter the action of the acidmixture is retarded, the fibres are rendered immune to substantivedyes, and their property of linear swelling on treatment with sodiumhydroxide solution is modified. It is suggested that the formal-dehyde can so modify (immunise) the cellulose that the specificaction of sodium hydroxide (cellulosate formation) is excluded.The action of the formaldehyde is conceived to result in the form-ation of hydroxymethyl ethers of cellulose : I_',C*OH + CH20 +h3OCH,*OH I or + (zC=O),CH,.Questions relating to the adsorption of sodium hydroxide bycellulose are considered by S.Liepatov and N. Sokolova 89 and byG. Champetier.gO The last author demonstrates by a titrationmethod the formation of the compounds (C6H,005)2,NaOH,(C6HI0O 5) ,,2NaOH, ( C6H lo0,),,3NaOH, and C6H ,,,o ,,NaOH insolutions containing 10 to 60% of caustic soda.The interactionof sodium with cellulose in dry liquid ammonia a t its boiling pointyields rapidly the monosodio-derivative (1Na : lC6H1005), thetheoretical amount of hydrogen being evolved ; further reaction isthen slow, but yields eventually the Na,-deri~ative.~lIt is possible by use of mixtures of phosphoric acid and nitric acidto obtain cellulose nitrate preparations (free from phosphate-nitrateesters) corresponding in nitrogen content to the pure trinitrate( 14.14y0 N).92 The process is accompanied by swelling of the fibres,which facilitates access of acid to the interior of the material andthus ensures uniform action. Nitration effected in this way isrelatively slow, the nitric acid mercerising the cellulose beforeesterification commences.With other nitrating mixtures, only thenitric acid is active as an esterifying agent, but the maximum degreeof nitration always falls short of the pure trinitrate.93Degradation of cellulose to the triose, tetraose, and hexaose stageshas been accomplished by partial hydrolysis with very concentrated88 M. Schenk, Helv. Chim. Acta, 1931, 14, 520; A., 717.89 2. anorg. Chern., 1930, 192, 383; A., 1930, 1562.Compt. rend., 1931, 192, 1593; A . , 941.91 P. C. Scherer, jun., and R. E. Husser, J . Amer. Chem. SOC., 1931, 53,2344; A . , 1041.s2 E. Bed and G. Rueff, Ber., 1930, 63, [B], 3212; A., 1931, 203; Cellulose-chem., 1931, 12, 53; A., 606; C. Trogus, Ber., 1931, 64, [B], 406; A., 409.93 R. C. Farmer, J . SOC. Chem. Id., 1931, 60, 75; A., 717ORGANIC CHEMISTRY.-PART II.105hydrochloric acid,w and to the tetraose and hexaose stages byhydrolysis of appropriate acetate fractions by Zemplbn’s method.95Similarly, methylation of acetone-soluble cellulose acetate, followedby the acetolysis, deacetylation, and further methylation of thederived trimethyl cellulose, has yielded decamethyl P-methylcello-trioside and a methylated cellodextrin which is possibly a cello-tetraose deri~ative.~6E. H. BARMER.PART II.-HOMOCYCLIC DIVISION.Tautomerism.(Continuedfrom Ann. Reports, 1929, 26, 116.)Prototropy.-Triad systems. The first clear case of three-carbonprototropic change in a system of the type NR,GH:CH*CH,*NR’,,in which covalency changes beyond the limits of the system itselfare structurally excluded, has been investigated by C.K. Ingoldand E. R0thstein.l The previous case examined by these authorswas complicated by superimposed anionotropic changes. The iso-meric bis-quaternary salts (I) and (11) have now been prepared andtheir individuality as separate, structural isomerides has beenproved. In the presence of moderate concentrations of hydroxideor alkoxide ions, however, they undergo facile prototropic inter-conversion :8 @8 e 8 { NEt,*CH,*CJXCK*h!teEt,}X, e { NEt3*CH:CH*CH2*beEt,} g2(1.1 (11.)Since the hydroxy- and methoxy-derivatives of the type{ $R3*CH2*CH (OR”) *CH,-gR ’, } g2(R” = H or Me)‘could not be converted by ordinary reagents intothe unsaturated derivatives, the possibility that interconversion ofthe latter is preceded by addition of water or alcohol to the doublelinking (as in the case of the diisobutylenes) is excluded.Finally,it should be noted that, whereas the unsymmetrical glutaconicesters often do not maintain their individuality as separate isomer-ides,3 the bis-quaternary propene derivatives have separate indi-94 L. Zechmeister and G. Toth, Ber., 1931, 64, [B], 854; A., 716.86 F. Klages, ibid., p. 1193; A., 827.96 W. N. Haworth, E. L. Hirst, and H. A. Thomas, J . , 1931, 824; A., 941.1 J., 1931, 1666; A., 1042.8 F. Feiat, Anlzalen, 1922, 428, 26, 61, 71; A., 1922, i, 621, 653.Ann. Reports, 1929, 26, 119.D 106 BAKER AND BENNETT :vidual existence. This is in agreement with the view4 that themore strongly m-directive NR, group is less effective in promotingprototropic mobility than is the more weakly m-directive GORgroup because, unlike the latter, it cannot provide a seat for theanionic charge.isfurnished by the conversion of 1 : 2- into 2 : 3-diphenylindene underthe influence of ethoxide ions.,Important light on the conditions governing mobility andequilibrium in the ay-diphenylmethyleneazomethine (A) and di-phenylpropene (B) systems has been obtained by C.W. Shoppee.78A further example of the mobility of the indene structurekl(111.1 p-C,H,R*CH:N*CH,Ph G C,H,R*CH,*N:CHPh (IV. Akap C ,H,R*CH:C( CO,E t)*CH,Ph s C,H,R*CH,*C( CO,E t):CHPh B(V.) WI.)The following table summarises the data obtained :@- 0.054 0.557 6.82 - 7.09 7.83 0.322 Ak,+k, (hr.-') { - <<0.060 0-058 0.39 - 0.64 1.02 - B0.149 0.370 0.887 (1.0) 1.08 1.13 1.23 A rZl - 0.77 - - 0.85 1.00 1-08 BR ...............NR,. NMe,. OhIe. I. (H.) Br. GI. Me........................................... K87.0 73.0 63.0 (50.0) 48.0 47.0 45.0 A - 56.5 - - 54.0 50.0 48.0 B ...... Equilibrium % of {p ,;;Dipole moment of C,H,R,e.8.u. x 10-1* ............ - +1*39 -0.80 -1.25 - -1.51 -1.56 +0.50Mobility clearly runs parallel to the inductive (-1) effect, increasingin the order p-NMe,<Me<OMe<Hal., and is thus obviously de-pendent on the ease of ionisation of the proton. On the other handthe simple view that equilibrium is controlled solely by the relativestability of the electromeric forms of the ion * cannot be main-tained, since, interpenetrating a series of groups of the polar typeI (NR, and Me) which in themselves are correctly arranged, isa series of groups (NMe,, OMe, Hals.) of the polar type - I + T,the expected order of which is completely inverted.Analysis ofthe system (111) s ion e (IV), in which the concentrationof the intermediate ion is presumed to be small, shows that theequilibrium constant K = v1/v2/v3/v4 = ~ 1 / ~ 4 / v z / v 3 , that is, it is the@2'1 %Va 2'4H. Burton and C. K. Ingold, J . , 1928, 904; A., 1928, 634.Ann. Reports, 1928, 25, 119.A. Garcia Banh and F. Calvet, Anal. Pis. Quim., 1929, 27, 49; A . , 1929,688. ' J., 1930, 968; 1931, 1226; A., 1930, 912; 1931, 834.Ann. Reports, 1927, 84, 106ORGANIC CHEMISTRY.-PART II.107ratio of the ratios of the rates of ionic dissociation (facilitated by- I and - T and retarded only by + I effects) and ionic associa-tion (facilitated by + I and + T and retarded only by - I effects).Hence equilibrium, as opposed to mobility, involves two rate-affecting phases which partake of the nature of side-chain reactionsof the opposite types. By analysis of his results the author showsthat, as in cyanohydrin f~rmation,~ the co-existence of two antagon-istic processes of control must be assumed, and, since it is shownthat it is the + T effect which is important, this process of duplexcontrol must apply to a stage in which the external reagent iselectron-seeking, that is, to the reassociation stage, which is thusof fundamental importance in controlling the equilibrium.Thus inethe electromeric ion C,H,R*CH*N*CHPh in which the anionic chargeis distributed between C, and C,, recombination of the proton willoccur preferentially at C, either if R strongly attracts electrons(- I , absorption of charge on C,) or if it strongly repels or releasesthem easily (strong + I + T ; the charge developed on the carbonatom para, to R repels electrons from Ca to C,,) ; but if R repels orreleases electrons only slightly (weak + I + T), the tendency forthe aromatic ring to absorb the charge on C, may be frustrated,causing ppeferential recombination a t this point.In making use of data derived from the orient-ing effects of groups in aromatic substitution as a practical criterionof their polar effect in prototropy, care must be taken not to deduceI effects, which constitute a permanent polarisation of the mole-cule, by the purely mechanical inversion of & T effects representingmainly a polarisability called into play only at the demand of thereagent.Thus, whereas benzene hydrogen substitution demandsavailability of electrons a t the point of reaction and is thereforefacilitated by + I and + T effects and retarded only by - 1effects, prototropy depends on the positivity of the field at the seatof ionisation and is therefore facilitated by - I and - T effectsbut retarded only by $- I effects. Hence, retardation of benzenehydrogen substitution does not run parallel with facilitation ofprototropy unless electromeric effects of either sign are structurallyimpossible.As C. W. Shoppee l1 has pointed out, it does notfollow that, because a p-carbethoxy-group (- 2') tends to stabilisethe py-phase in the system citraconic-itaconic esters,l2 a P-methylgroup (+ I ) should therefore stabilise the ap-phase. ConsiderationA. Lapworth and R. H. F. Manske, Ann. Reports, 1928, 25, 147, andref. 27.a YPentad systems.l1 J., 1930, 968; A., 1930, 912.l a Ann. Reporb, 1927,24, 113108 BAKER AND BENNETT:of the two enolide ions in the critical stage shows that the - Teffect of the p-carbethoxy-group (b b,) acts on displaceable double-bond electrons, inhibiting a3. The + I effect of the p-methyl group(c) acts on the p-carbon, charging it negatively and directing theQR bcr, OR4.A- &I >c==c-c-c-0Y l V S VJ IORelectrons away from a2 into the alternative route a ; since a3 dependson initiation from a2 (which in turn depends on a), the + I effect(c) will also indirectly inhibit a3. Hence, in accord with experimentalresults,13 the methyl group, like the carbethoxy-group, should favourrecombination of the proton at C, with consequent stabilisation ofthe py-phase.Further results of G. A. R. Kon and R. P. Linstead and theirco-workers indicate that some factor other than the purely polarmust affect the position of equilibrium in such systems.14 The resultsobtained for a series of acids of the type CHR1R2*CR3:CR4*C02Hin N-sodium ethoxide at 25" are as follows : l5% aP-Form at yo a#l-Fom atRIR2.R3. R4. equilibrium. R1R2. R3. R4. equilibrium.H, H H 100 EtMe H H 23Me H 100 MeH Et H 21 ZiH Me H 33 MeH Et Me 50E t a Me Et 57 EtH PP H 33Me, H H approx. 10 ** In the case of the isohexenoic acid the quantitative results are onlyapproximate owing to the complication introduced by the addition of alcoholto the aP-ester, which reaction is appreciably faster than the tautomericchange.Here the qualitative effect of a-, P-, and y-alkyl substituents followsthe lines expected from theory.16 More difficult to interpret is theeffect of a substituent methyl group on the position of equilibriumin cyclohexylidene- and cyclopentylidene-acetic acids and the corre-sponding methyl ketones. Pure dry cyclohexylidene- and cyclo-l3 A. A. Goldberg and R.P. Linstead, J., 1928, 2343; A., 1928, 1214.14 Compare Ann. Reports, 1927, 24, 112.1 5 Ann. Reports, 1927, 24, 111; R. P. Linstead, J., 1930, 1603 ; 1929, 2498;A,, 1930, 1162, 64; R. P. Linstead and J. T. W. Mann, J . , 1,930, 2064; A.,1930, 1405; G. A. R. Kon, E. Luton, R. P. Linstead, and L. G. B. Parsons,J., 1931, 1411 ; A., 934.Ann. Reports, 1927, 24, 111ORGANIC CHEMISTRY .-PAR,T II. 109hexenyl-acetic acids are interconvertible in carbon dioxide at 150°,17the proportion of the py-form (83%) in the equilibrium mixtureclosely approximating to that (88%) observed when equilibrationis effected in sodium ethoxide. The position of equilibrium issimilar in the case of the corresponding cyclopentylideneacetic acid,although mobility is much greater.l8 The effect of a- and ringsubstituent methyl groups in the systems (CH,), >C:CH*COR issummarised in the following table : l9Acids (R = OH).Ketones (R = Me).r- A \ -7 9; CCB-FOW % aP-FonnSubstituent. at equili- Mobility, at equili- Mobility,Ring. a*. brium. 10 ( k , + k , ) hr.-l. brium. lo4 (k,+k,) min.-l.n = 5- H 12 1 43 700- Me 32 0.0075 02-Me H 11.5 0-15 03-Me H 9 0.27 25 -4-Me H 7 0.42 13.5 600-700--n=4- H 14 22 84 ca. 3000- Me 38 0.58 64 3000Thus, although, as would be expected, introduction of a methylgroup in the a-position causes a marked shift in the equilibriumtowards the ap-form and a strong retardation of mobility in bothacid series, the reverse effect is observed in the ketone series, similarsubstitution here favouring the py-phase, the effect on mobilitybeing much less marked.In the cyclohexene acid series nuclearmethylation causes a slight displacement (4>3>2) in favour of thepy-phase, and a similar small decrease in mobility, whilst in thecorresponding ketone system introduction of a methyl group inposition 2 stabilises the &-form to the almost complete exclusionof the a@-form, similar substitution in position 3 or 4 exerting nopronounced effect .The more facile mobility of the system >CH*CO*OR =+> C:C( OH)*OR compared with that of its extended pentad-enolform >CH*C:C-CO*OR =+ >C:C*C:C(OH)*OR 2o is illustrated byresults obtained by W. E. Hugh and G. A. R. Kon.21 Ethyl cyclo-pentylidenemalonate is converted by sodium ethoxide into thesodio-enol.Weak acids such as benzoic are able to catalyse only17 R. P. Linstesd, J., 1930, 1603; A., 1930, 1163.l* Compare Ann. Reports, 1929, 26, 118.19 G. A. R. Kon and R. S. Thkur, J., 1930, 2217; A., 1930, 1582.20 Ann. Reports, 1927, 24, 109.2 1 J . , 1930, 776; A., 1930, 773110 BAKER AND BENNETT :the keto-enol change and not the less facile three-carbon tauto-merism and the product contains 30--50% of the py-ester.OEt OEtHCH,*CH,I \C*CHdH,-CHNContrary to the authors’ previous suggestion22 that the cyclo-hexane ring exerts a special attraction for the double linking, similartreatment of ethyl cyclohexylidenecyanoacetate causes no conversioninto the py-form, nor was any such conversion observed in the caseof ethyl a-c yano - p - et hylcinnamat e or of ethyl a-cyano- y-phenyl-P-methyl-Aa-butenoate.I n these cases, however, the system isterminated by the combination <gz2R, which is probably moreeffective in promoting mobility than is the group <COzR.23 Thecyano-group will tend to absorb the charge on Ca,24 thus promotingreassociation of the proton a t C, with a consequent stabihation ofthe crp-form. Moreover, in ethyl a-cyano-p-ethylcinnamate thephenyl group in the p-position will activate Cy towards a proton25and thus further displace equilibrium in favour of the ap-phase.R. P. Linstead26 attributes the effect of p- and y-phenyl sub-stituents to an element of stability associated with conjugatedsystems, termed the “ Thiele factor ” by A.Lapworth and R. H. F.Man~ke.~’ The latter authors regarded this as not necessarily ofa polar nature : R. P. Linstead asserts that it is non-polar, but theview that the phenomenon is essentially of a polar nature seemsequally tenable and undoubtedly of more value in correlating alarge number of experimental facts.The Glutaconic Problem-Some years ago a study of the additionof cyanoacetic esters to glutaconic esters led E. P. Kohler andCO R22 G. A. R. Kon and E. A. Speight, Ann. Reporb, 1927,24, 113.23 Ibid., 1927, 24, 110.24 J. W. Baker, K. E. Cooper, and C . K. Ingold, dbid., 1928, 25, 138.25 Ibid.,p. 146.26 J., 1929, 2498; A., 1930, 64.27 J., 1928, 2535; A., 1928, 1246ORGANIC CHEMISTRY.-PART 11. 111G. H. Reid28 to the conclusion that these reactions proceed aswould be expected with ap-unsaturated esters, the results lendingno support to the conception of a " normal " form of the glutaconicesters having a symmetrical formula and a capacity for additionin the 1 : 3-positions.The whole subject is now being re-examined in the South Ken-sington laborat~ries.~~ The resolution of " normal '' ocy-dimethyl-glutaconic acid into optically active forms 30 points directly to theunsymmetrical structure CO,H*CMe:CH*CHMe*CO,H.A carefulexamination of the tautomerism of some cyanoglutaconic estersby G. A. R. Kon and H. R. Nanji has led to the isolation of both theAa- and the AB-isomeride of ethyl a-cyano- py-dimethylglutaconate,C0,Et *C( CN):CMe*CHMe*CO,Etand CO,Et*CH( CN)*CMe:CMe*CO,Etand the proof that both the " normal " and the labile form describedbefore 31 consist of the equilibrium mixture arising from the catalyticaction of sodium ethoxide on the Aa-ester.These results are in fact adequately explained without the specialassumption of a " normal " form, and the view that the isomerismof the glutaconic acids and their derivatives is due to the occurrenceof three-carbon tautomerism together with cis-trans isomerism will,no doubt, now be generally accepted.This was substantially thecontention of F. F e i ~ t , ~ , who, however, seems to have over-emphasised the single factor of cis-trans isomerism, regarding thethree-carbon systems as always dynamic. In actual fact, where acrystalline substance is isolated, a static formula must be assignedto it, and both structural and steric isomerism may be expectedto occur.In an investigation of some substituted glutaconic esters fromthis point of view G.A. R. Kon and E. M. Watson find that twoforms of the ester of a-benzyl-p-methylglutaconic acid arise, shownby ozonolysis to be structurally isomeric. These are found to bethe cis-ap- and the trans- py-isomeride : the remaining stereoisomer-ides (trans-ap and cis-py) have not yet been isolated.33Two anomalous cases still remain which have hitherto beenregarded as evidence of the " normal, " form of glutaconic acids.28 J . Amer. Chem. SOC., 1925, 47, 2803; A., 1926,48.29 J . , 1931, 1011.30 T. H. McCombs, J. Packer, and J. F. Thorpe, ibid., p.547; A., 604.31 J., 1931, 660; A., 608; compare N. Bland and J. F. Thorpe, J., 1912,82 LOC. oit., ref. 3.33 J., 1932, 1.101, 871 ; Ann. Reports, 1912,9, 86-89112 BAKER AND BENNETT :The first of these is the isolation,3* from the products of the reactionof ethyl sodioc yanoacet ate with ethyl p- hydroxyglutarate (whichunder the experimental conditions gives “ nascent ” ethyl glutacon-ate), in addition to the expected 1 : 2-additive product, of a smallproportion of ethyl cyclopentanone-2 : 4-dicarboxylate (hydrolysed tocyclopentanone-3-carboxylic a~id).~5 This must be produced bycyclisation of ethyl cc-cyano- p-carbethoxyadipate apparently formedby 1 : 3-addition :-yH*CO,Et TH( CO,Et)*CH( CN)*CO,Et,-CH*CO,Et+ 7%CH,*CO,EtCH(CO,Et)*CH, 1 \C:NH --+CH,-CH( C0,Et )/TH2CH( C0,H)-CH,I \co I CH,- CH,/It should bb noted, however, that in ethyl glutaconate a stronglypolar carbethoxy-group is present on each side of the double linking,with a methylene group intervening on one side only.Hence, justas the addition of iodine chloride to a double bond occurs in bothsenses, only the major product being in accord with Markovnikov’srule,36 the by-product in the above reaction may possibly be due,not to 1 : 3-addition, but to reversed 1 : 2-addition :GH*CO,EtCH*CH,*CO,Et CH,*CH,*CO,Et - QH (C02E t ) *CH ( CN)*CO,E tIf this explanation is correct, there only remains the second anomaly,namely, the reported isomeric forms of the ester of 3-methylcyclo-propene-1 : 2-dicarboxylic acid.37 Some explanation of this willno doubt soon be found.Transannular Tautmerism.--The isolation of two transannular(also termed transannelar) isomeric hydrocarbons is described byE.Bergmann.3s Benzhydrylideneanthrone is converted by phenyl-magnesium bromide into 10- hydroxy- 1 O-phenyl-9- benzhydrylidene-9 : 10-dihydroanthracene, which is reduced by sodium formate andformic acid to a mixture of 10-phenyl-9-benzhydrylidene-9 : 10-dihydroanthracene and 10-phenyl-9-benzhydrylanthracene. Thelatter is converted into the former through its sodium derivative.3 p C. K. Ingold and J. F. Thorpe, J., 1921, 119, 492.3s F. W. Kay and W. H. Perkin, J., 1906,89, 1645.36 A. Michael, J. pr. Chmn., 1899, 60, 450; A., 1900, i, 321.s7 Ann.Reports, 1923,20, 107-110.38 Be?., 1930, 63, [B], 1037; A., 1930, 902ORUANIC CHEMISTRY.-PART II. 113C 6 H 4 < b 6 H 4@‘Ph1I - C6H4<7)C,H,2e CPhc6H4<c>6H4CHPhThe possibility of both prototropic and anionotropic changes in9-arylidene-9 : 10-dihydroanthracene derivatives of type (11) mustbe borne in mind 39 in the interpretation of the results obtained bybromination of 9-alkyl- or 9-aryl-anthracenes, the investigation ofwhich has been continued by E. de Barry Barnett and his collabor-ators. It will be noticed that the intermediate dibromide (IV),which is presumably the initial product in the bromination, can passdirectly by elimination of hydrogen bromide into derivatives oftype (11) or (111), and, when such derivatives are formed, con-ditions governing the mobility and equilibrium of prototropic systemsare obviously involved.Derivatives of type (I), however, cannot[HI’(1.1“l;I?y+ /+ (111.) \BrCHR,\ /Brc6H4</ YC 6 H 4 UV.1H’ ‘Brbe thus directly obtained and their formation indicates that con-ditions governing anionotropic changes must be considered. Inthis respect the effect of cc-chlorine substitution on the brominationof 9-substituted anthracenes is instructive. Bromhation of 4-chloro-9-benzyl- and 4 : 5-dichloro-9-methyl-anthracene affords, respec-tively, 4-chloro-w-bromo-9-benzyl- and 4 : 5-dichloro-9-bromomethyl-anthra-cene 40 (both of type I), which are the stable forms anticipatedin the anionotropic system involved.39Ann. Rep&, 1928,%, 130.40 J.W. Cook, J., 1928, 2802; A., 1928, 1365114 BAKER AND BENNETT 1The following table summarises the effect of nuclear substitutionon the relative proportions of derivatives of types (11) and (111) inthe bromination of various 9-arylanthracenes :Predominant form ofNuclear substituent. RZ. bromination product. .................. Unsubstituted 41 Ph, H I112-Chloro 42 Ph, H I11l-Chloro 43 ........................... Ph, H 70% (11) 30% (111)Unsubstituted 4 5 p-CI*C,H,-, H I111 : 5-DichIoro 45 p-CI.C,H,-, H I12 : 3-Dimethyl 46 Ph, H I111 : 4- ,, 47 Ph, H I1In these cases prototropic change is involved and the effect ofnuclear chlorine-substitution is readily explained by the - I effectof the chlorophenyl nucleus, which, by withdrawing electrons fromthe central nucleus, will tend to stabilise the negative charge ofthe electromeric anion on the nuclear atom Cl0 with the resultant..............................1 : 8- or 1 : 5-Dichloro 44 Ph, H I1 ........................................................................@HR ICHR I1 c:> @ CBrIon-Type (111). Ion-Type (11).preferential (1-chloro-) or exclusive (1 : 5- or 1 : 8-dichloro-) form-ation of derivatives of type (11). The case of 1 : 4-dimethyl sub-stitution is, from this point of view, an0malous.4~Anionotropy.-A study of further examples confirms the con-clusions previously reported 49 relating to the factors affecting thernobiEity of anionotropic systems. Thus in boiling acetic anhydrideboth methylvinylcarbinyl and crotyl trichloroacetates afford thesame equilibrium mixture containing 55% of the former, whilsta-m-4-xylylallyl alcohol is converted into 2 : 4-dimethylcinnamylalcohol by acetylation and subsequent hydrolysis.5O41 J. W. Cook, J., 1926, 2161; A., 1926, 1131.42 E. de B. Barnett and J. L. Wiltshire, J., 1928, 1822; A., 1928, 995.43 J. W. Cook, J., 1928, 2802; A., 1928, 1365.44 E. de B. Barnett, N. F. Goodway, and J. L. Wiltshire, Ber., 1930, 63,46 E. de B. Barnett and J. L. Wiltshire, ibid., 1929, 62, [B], 3072; A., 1930,46 E. de B. Barnett and F. C. Harrison, ibid., 1931,64, [B], ‘535 ; A., 612.47 E. de B. Barnett and L. A. Low, ;bid., p. 49; A., 341.48 Compare E. de B. Barnett and C. L. Hewett, ibid., p. 1572; A., 1058.4D Ann.Roports, 1928, 25, 127.50 H. Burton, J., 1930, 248; A., 1930, 460.[B], 472 ; A., 1930, 465.202ORUANIC CHEMISTRY.-PART II. 115The increased activation of a three-carbon anionotropic systemby polynuclear aryl groups is illustrated by the conversion of a-and P-naphthylvinylcarbinols into y- 1- and y-2-naphthylallylp-nitrobenzoates, respectively, by the action of p-nitrobenzoylchloride and pyridine, under which conditions the correspondingphenyl derivatives each give the ester of the original alcohol.51The position of equilibrium of such systems, however, islargely affected by experimental conditions. Thus, whereas theaction of anhydrous hydrogen halides on either cinnamyl ora-phenylallyl alcohol gives only cinnamyl halides, treatment ofcinnamyl chloride with 70 yo aqueous-alcoholic potassium hydroxideaffords 20-25%, and with silver oxide in absolute alcohol, 35-45%,of a-phenylallyl ethyl ether.52CHPh:CH*CH,-Cl *- CHPh.CH CH, =+ p. @ne mCHPh*CH:CH, Ph*CHCl*CH:CH,The formation of the ether is readily explained by the greaterco-ordinating power of the ethoxide, relative to that of the chlorideion. In the halides recombination of the anion occurs exclusivelyin the y-position to the phenyl group, but the ethoxide ion combinesin the a-position, indicating a considerable displacement of theequilibriu‘m under these conditions in favour of the other electro-meric form of the kation.Differences in experimental conditions also explain C. Prhvost’s 53failure to confirm H.Burton’s results (above) on the interconversionof crotyl and mefhylvinylcarbinyl esters, since the low dielectricconstant of the solvents (toluene and trichloroacetic acid), the lowertemperatures, and the shorter periods of heating employed mayprevent a system of such small mobility from attaining completeequilibrium.Aromatic Substitution.(Continued from Ann. Reports, 1930,27, 130.)General Theory. Dynamics of Aromatic Substitution.-Themodern electronic view of aromatic substitution has been elab-orated almost exclusively from observations of the proportions of62 J. Meisenheimer and J. Link, Annalen, 1930, 479, 211; A., 1930, 769 :compare H. Gilmm and S. A. Harris, J . Amr. Chem. Soc, 1931, 53, 3541;A., 1290.E. Burton, J., 1931, 759; A., 723.53 Bull.SOC. chim., 1931, [iv], 49, 261; A., 601116 BAKER AND BENNETT:isomerides simultaneously formed, but four years ago the importancewas emphasized of considering not only the proportions of iso-merides, but also the relative velocities of substitution, and pre-liminary data of the latter kind were rep0rted.5~ Accurate measure-ments of the reduced velocity of substitution of toluene relativeto benzene have now been furnished by C. K. Ingold, A. Lapworth,E. Rothstein, and D. Ward,55 and the effect of certain modificationsin experimental conditions has been ascertained. Under standardconditions (nitration by acetyl nitrate at 30") the values of the partialrate factor F (previously designated as the coefficient of activation,oix., the factor by which the introduction of the directing groupincreases the probability of substitution at a given nuclear carbonatom) for the o-, rn-, and p-positions are respectively 40, 3.0, and51.Hence, in agreement with the earlier findings of C . K. Ingoldand F. R. Shaw,= the methyl group not only causes a large increasein op-reactivity but also a small increase in m-reactivity. Altera-tion of the reagent and solvent to nitric acid in nitromethane gaveidentical values for F , whilst changing the temperature to 0"resulted in a slightly higher reduced velocity and generally highercoefficients of activation.It was shown by A. E. Bradfield and B. Jones 56 that in o- andp-substituted alkoxybenzenes the two substituents contributecharacteristically and additively to the logarithm of the velocityof nuclear halogenation, on the assumption that in the equationfor the critical energy increment k = ae-E'RT, variations in k aredue entirely to variations in E, a remaining constant.The halo-genation of a further series of phenyl ethers and of p-tolyl ethersXof the general formula RO O M e at 20" and 35" has now beenstudied.57 There is little difference between the rates of chlorin-ation of corresponding chloro- and bromo-ethers, and the groupcontributions for the new phenyl ethers examined are in agree-ment with the earlier results. The data also serve to confirm theoriginal hypothesis above mentioned and to discredit three other(more complicated) possible assumptions.At the same time someof the figures reveal a definite tendency to departure from thesimple additive relation in directive effects of groups. Thus therelative effects in the series of tolyl ethers agree with the resultsfor phenyl ethers when X is NO, but not when X is a halogen;64 Ann. Reports, 1927, 24, 152.55 J., 1931, 1969; A., 1405.56 Ann. Reports, 1928, 25, 143.5 7 A. E. Bradfield and B. Jones, J . , 1931, 2903; A. E. Bradfield, W. 0.Jones, and F. Spencer, ibid., p. 2907ORUAHIU CHBMISTRY .-PART II. 117and in general the relative directive effects are 5-15% lower inthe tolyl ether series than for the corresponding phenyl ethers.The conclusion that the nitroso-group is m-directive 68 was basedmainly on the slow velocity of bromination and absence of p -substitution of nitrosobenzene in acetic acid, in which solvent it isnot associated, and on the reaction of the bromine in p- but not inm-bromonitrosobenzene with silver nitrate in acetic acid.Cautionshould be observed, however, in accepting this conclusion, sinceR. J. W. Le F h r e 59 has found that p-bromonitrosobenzene reactswith silver nitrate under conditions in which even picryl bromide(activation by three nitro-groups) is unattacked. Moreover,nitrosobenzene can be mono-p-brominated in benzene, a solventin which it is also unimolecular. The most satisfactory view ofthe properties of the nitroso-group would seem to be that it is capableof electromeric effects of either sign,60 aromatic hydrogen substi-tution or replacement of halogen being controlled by displacements,of type (I), whereas hydrolytic reactions (e.g., the hydrolysis ofp-nitrosodimethylaniline) are initiated by those of type (11).Ionogenic Systems.-By analogy, with Schiff ’s bases, which havepreviously been proved 61 to exist, in concentrated sulphuric acidsolution, as salts of the type PhCH:NHA.r}%, J.W. Baker 62 hassuggested that the intensely coloured solutions formed by benzalde-hyde and acetophenone in the same medium are due to the formationof oxonium salts of the type PhCR:OH}HSO,, the ions of whichare in equilibrium with the colourless, pseudo-salt form. Partition(between sulphuric acid and ligroin) and colorimetric data indicatethat the equilibria@ -0- 0f - - - __ - - -- - - - -- - colourless - - _- - - _- - - - _- - ____ - - ___ - - - 3 coloured87+ PhCR:O + HO*S03H + PhCR(OH)*O*SO,H T- O*803H + PhCR:OH +are displaced largely towards the right even in SOY0 sulphuric acid.As in the case of Schiff’s bases, the distribution of the positivecharge on the carbonium-oxonium kation causes a considerableincrease in the proportion of m-nitration observed under these5% Ann.Reports, 1930,27, 130.59 J., 1931, 810; A., 945 : but compere D. L. Hammick, ibid., p. 3105.6O R. Robimon, Chem. and I d . , 1925, 44, 456.61 Ann. Rep&, 1930,27, 136.soluble f -- ----- --- ---- insoluble in ligrob _________ - __________ __-62 J., 1931, 307; A. 486118 BAKER AND BENNETT :conditions.63 Thus in 7.3% oleum, benzaldehyde and acetophe-none afford 90.8 and 90.0% of the m-isomeride, respectively. Thisproportion decreases gradually as the concentration of the sul-phuric anhydride is diminished until in 8076 sulphuric acid thevalues are 83.9 and 83*1%, and in nitric acid (d 1.53) alone, 72.1and Toy0, respectively.Moreover, in all cases, addition of ammon-ium sulphate to the sulphuric acid nitrating medium causes 5t de-pression in the amount of m-isomeride formed, comparable bothin magnitude and in type with that observed in the case of Schiff’sbases, an indication that substitution occurs mainly through thekation of the oxonium salt.MisceZkcneous.-Further examples of the effect of experimentalconditions in altering the proportion of isomerides formed havebeen recorded.Thus, although the occurrence of a large amount(85%) of m-substitution when phenylboric acid is nitrated withnitric acid (d 1.53) has been confirmed, in acetic anhydride a 65%yield of a mixture containing 95% of o- and 5% of p-nitrophenyl-,boric acid is obtained.64 Pending further investigation, this resultsuggests that the explanation of the m-directive power previouslygiven65 should be accepted with reserve.That the nature of the substituting reagent also affects tlre posi-tion of substitution is illustrated by results obtained with diphenylderivatives and diphenyl ethers. Thus p-acetamidodiphenyl etheris nitrated 66 and brominated 67 respectively in the positions shown :these results forming a parallel with those obtained by J. Kenyonand P.H. Robinson in the case of 4-acetamidodiphenyl. Nitra-tion of 4-acetamido-4’-methoxydiphenyl ether also occurs in posi-tion 3, whilst p-methoxydiphenyl ether is brominated in the 4’-position. Similar differences are found by F. Bell 69 in the caseof 3-bromo-4-p-toluenesulponamidodiphenyl, which is nitrated inposition 5, but brominated in position 4’. The explanation of theseresults may lie in the lower degree but higher frequency of activa-tion in position 4’ than in position 3 or 5,’* so that the more ener-G3 J. W. Baker and W. G. Moffitt, J., 1931, 314; A., 485.64 W. Seaman and J. R. Johnson, J . Amer. Chem. SOC., 1931,53,711; A., 502.6 5 Ann. Reports, 1930, 27, 133.6 6 M. Oesterlin, Monutsh., 1931, 57, 31 ; A., 479.6 7 ,Ann.Reports, 1930, 27, 142.70 A. Lapworth and R. Robinson, Mem. Munchester Phil. SOC., 1927, 72,J . , 1926, 3050; A., 1927, 142. J . , 1931, 2338; A., 1284.43; A., 1927, 546ORGAN10 CHEMISTRY.-PART II. 119getically negative-centre seeking reagent, bromine, enters position4’. When bromination of 3- bromo-4-p- t oluenesulphonamidodi-phenyl is carried out in pyridine, the activating power of the p -toluenesulphonamido-group is increased by removal of its incip-iently ionised hydrogen by salt formation, and the second bromineenters at position 5.In conclusion, attention may be directed to investigations on theaction of bromine on phenols,71 and on the chlorination of iodo-phenols. 72Polycyclic A r m t i c Compounds.No systematic report on this subject has been made for severalyears, but limitation of space restricts the following summary to abrief outline of a few of the more important lines of general develop-ment. Attention may be directed to the monograph on ‘‘HigherCoal Tar Hydrocarbons ’’ by A.E. Everest which gives an accountof the chemistry of higher polycyclic aromatic hydrocarbons up tothe end of 1926.After removal of 1 : 6-, 2 : 6-,and 2 : 7-dimethylnaphthalenes from low-temperature coal tar,2 : 3-dimethylnaphthalene has been isolated as its 6-sulphonic acidand converted into various derivatives. 73The ct-directive catalytic infiuence of mercury in the sulphon-ation of naphthalene finds a parallel in the exclusive formation of@-derivatives when tetrahydronaphthalene is condensed with alkylbromides or acyl chlorides in the presence of aluminium halides,other metallic halides affording mixtures of a- and p-derivatives.74Catalytic dehydrogenation of the products affords S-alkylnaphth-alenes and p-naphthyl ketones, respectively. Similar condens-ation of tert.-butyl chloride with naphthalene itself, however, givesa mixture of di-tert. - butylnaphthalenes.75By modifications of the general method involving ring-closureof substituted y-phenylbutyryl chlorides with aluminium chloride,followed by reduction and subsequent catalytic dehydrogenation,I. M. Heilbron and his co-workers have synthesised l-methyl-5-,-6-, and -7-eth~lnaphthalene,’~ and 2 : 3 : 5-, 1 : 3 : 5, and 1 : 3 : 8-t~imethylnaphthalene.~~ . Neither the 1 : 3 : 5- nor the 1 : 3 : 8-derivative is identical with the trimethylnaphthalene obtained bydehydrogenation of tetracyclosqualene.78 Since the oxidation productsNaphthalene.-Alkylnaphthalenes,71 G. Heller (with others), J. pr. chem., 1931, [ii], 129, 211; A., 477.72 S. Buchan and H. McCombie, J., 1931, 137; A., 346.73 0. Kruber, Ber., 1929, 82, [B], 3044; A., 1930, 202.74 A. Barbot, Bull. SOC. chim., 1930, [iv], 47, 1314; A., 208.75 W. Gump, J . Amer. Chem. SOC., 1931, 53, 380; A., 341.76 J. Harvey, I. M. Heilbron, and D. G. Wilkimon,J., 1930,423 ; A., 1930,693.7 7 D. G. Wilkinson, J., 1931,1333; A., 948. 78 Ann. Reports, 1927,24,128120 BAKER AND BENNETT:of the latter could arise only from the 1 : 3 : 8- or the 1 : 2 : 5-tri-methyl derivative, it must be the latter,79 a conclusion subsequentlyconfirmed by the synthesis of this hydrocarbon by L.Ruzickaand J. R. Hosking.80 The same hydrocarbon results from thecatalytic dehydrogenation of Congo copal oil.81 The syntheses of1 : 5-, 1 : 8-,82 and 2 : 8-83dimethylnaphthalene have also beeneffected. The conversion of 2-methylnaphthalenes into benz-anthracenes is discussed on p. 126.Substitution of naphthalene derivatives. The problem of orient-ation in the naphthalene nucleus is complex from both the experi-mental and the theoretical point of view. The study of this fieldon modern lines has not yet made much progress, but a few recentinvestigations may be mentioned.The nitration of cc-chloro- or a-bromo-naphthalene occurs SUC-cessively at positions 4, 5 , and 8, and of a-alkoxynaphthalenes a tpositions 2, 4, and 5.85 Nitration of 2-chloro-, bromo-, or alkoxy-naphthalene yields the 1 : 6 : 8-trinitro-derivati~e.~~ The halogenatom in such compounds is readily replaced by positive-centreseeking groups such as OH, OR, and S.The influence of a nitro-group in producing high reactivity in a halogen atom is found byJ. Salkind 87 not to extend from one ring to the other of the naphth-alene nucleus. On the other hand it is well known that inhalogen substitution a group often deactivates one nucleus asa whole relative to the other.84Chlorosulphonic acid converts l-methylnaphthalene into the4-sulphonic acid,88 although 2-methylnaphthalene is stated to givethe 8-sulphonic acid.89 Further nitration of a mixture of 4- and5-nitro-l-methylnaphthalene gives the 4 : 5-dinitro-derivative,whilst the presence of some 2 : 4-dinitro-compound is indicatedby the formation of 2-nitro-4-amino- and 4-nitro-2-amino-l-methyl-naphthalene on reducti~n.~OI. M.Heilbron and D. G. Wilkinson, J., 1930, 2546; A., 80.8o Helv. Chim. Acta, 1930, 13, 1402; A., 231.81 L. Wertenburg and J. P. Wibaut, Rec. trav. chim., 1931, 50, 188; A . , 484.82 V. Veselg and F. &ursa, Coll. Czech. Chm. Comm., 1931, 3, 430; A.,83 V. Vesely and A. Medvedeva, ibid., p. 440 ; A., 1282.84 Ann. Reports, 1926, 23, 134; 1930,27, 140.85 H. W. Talen, Rec. trav. chim., 1928, 47, 329, 346; A , , 1928, 405.86 E. J. van der Kam, ibid., 1926, 45, 564, 722 ; A., 1926, 1029, 1239.87 Ber., 1931, 64, [B], 289; A., 474; compare Ann.Reports, 1929, 26, 132.R. E. Steiger, Helv. Chim. Acta, 1930, 13, 173; A., 1930, 593.K. Dziewoliski and A. Wulffsohn, Bull. Acad. Polonaise, 1929, A , 143;*O V. Vesely, F. &ursa, H. OlejnSek, and E. Rein, Coll. Czech. Chm. Comm.,1282.A., 1929, 803.1929, 1, 493; 1930,2, 145; A., 1929, 1288; 1930, 593ORQANIU CHEMISTRY.-PBRT II. 121An investigation of the bromination of p-naphthol by K. Friesand K. Schimmelschmidt 91 has explained various discrepanciesin the literature. Such bromination affords, successively, thel-mono-, 1 : 6&-, and 1 : 4 : 6-tri-bromo-deri~atives.~~ The tetra-bromo- p-naphthols, formed by bromination in acetic acid, consistof the 1 : 3 : 4 : 6-, 1 : 3 : 5 : 6-, and 1 : 3 : 6 : 8-derivativesY thelast two compounds not being formed by simple substitutionprocesses.The reaction C,,H,*OH(p) + Br, CIoH,Br*OH +HBr is reversible, since treatment of l-bromo-p-naphthol with asaturated solution of hydrogen bromide in acetic acid in the presenceof potassium iodide (to remove bromine) affords p-naphthol. Thecourse of bromination is largely influenced by the conversion ofp-naphthol into quinonoid forms. Bromination of keto-bromides(the formation of which is favoured by the addition of sodiumacetate to remove hydrogen bromide) and subsequent reductionof the products gives rise to different derivatives from thoseobtained when bromo-p-naphthols are brominated. Thus 1 : 3 : 6-and not 1 : 4 : 6-tribromo-p-naphthol is produced when 1 : 1-dibromo-2-keto-1 : 2-dihydronaphthalene is brominated and theproduct reduced.The mechanism of the formation of the 1 : 6-dibromo-compound and its successive reconversion into the6-bromo-derivative and into p-naphthol is indicated in the followingscheme :Br Br Br,-HBr 1slow2 7HBrSimilar ketonisation of 2 : 4-dibromo- a-naphthol is suggested as9 1 Annulen, 1930, 484, 245; A., 216.92 Compare H. Franzen and G. Stauble, J . p r . Chern., 1921, [ii], 103, 362;A., 1922, i, 450122 BAKER AND BENNETT :the mechanism of its conversion into the indigotin type of deriv-ative (I) under the influence of alkali : 93OH 0 0Bromination of 1 : 5-dihydroxynaphthalene yields the 2 : Ci-dibromo-derivative, which, on oxidation, affords 2 : 6-dibromo-5-hydroxy-1 : 4-naphthaquinonq only one of the rings being converted intothe quinonoid f0rm.~4An interesting new re-action of various aryl diazosulphonates of p-napht hol- 1 -sulphonicacid has been investigated by F.M. Rowe and his ~ollaborators.~5Whilst acidification of a solution of 4’-nitrobenzene-2-naphthol-l-diazosulphonate (11) in a molecular quantity of sodium hydroxidegives a quantitative yield of Para-red (p-nitrobenzeneazo-fi-naphth-ol), addition of the diazosulphonate in sodium carbonate solutionto an excess of sodium hydroxide and subsequent acidificationgives only a trace of Para-red, the main product being the sodiumsalt of 3 - (4’- nit rop hen y 1) - 1 : 3 - di h y dr op ht halazine - 4 -acetic - 1 - sul-phonic acid.This compound, to which structure (V) is provisionallyassigned, is converted by boiling hydrochloric acid into the corre-sponding 1 -hydroxy-derivative (VI).Although the condensation products of diazotised p-nitroanilinewith 1 -halogeno- or 1 -methyl- p-naphthol are true diazo-oxide~,~~the initial stage in the above reaction is the formation, not of adiazo-oxide (XII) as postulated by H. T. Bu~herer,~’ but of sodium4’-nitrobenzene-l-azo-fi-naphthaquinone-l-sulphonate (111), ringfission and subsequent cyclisation of the acid (IV) giving thedihydrophthalazine derivative (V). The corresponding hydroxy-compound (VI) is reduced to l-hydroxy-3-(4’-aminophenyl)tetra-hydrophthalazine-4-acetic acid (VII).The possible alternativeisoquinoline structure (XIII) is shown to be untenable by a com-parison of its properties with those of ~V-4’-nitrophenylaminoiso-93 R. Willstiitter and L. Schuler, Ber., 1928, 61, [B], 362; compare S.94 A. S.WheelerandD. R. Ergle, J. Amer. Chem. SOC., 1930,52,4872; A., 217.98 F. M. Rowe, E. Levin, A. C. Buri~s, J. S. H. Davies, and W. Tepper,J., 1926, 690; A., 1926, 625.g6 F. M. Rowe and A. T. Peters, J., 1931, 1065; A., 835; compare J.Pollak and E. Gebauer-Fiilnegg, Monutsh., 1928, 50, 310; A., 1929, 58.9 7 Ber., 1909,42,47; A., 1909, i, 193; H. T. Bucherer and C. Tama, J. pr.Chem.. 1930,127,39; A., 1930, 1280; compare A. Wahl and R. Lantz, Bull.SOC. chirn., 1923, 33, 97; A . , 1923, i, 209.Naphthalene-azo- and diazo-compounds.Goldschmidt and H.Wessbecher, ibid., p. 372 ; A., 1928, 408, 409ORGANIC CHEMISTRY .-PART 11. 123carbostyril-3-carboxylic acid (XIV), a rational synthesis of whichmas effected.SO,*N:NR NaO,S\/N:NR CH*N:NR ~CH:CH*CO,Na J7 S03NaI _ _ , i(11.1 (111.) (IV.1CH*SO,Na CH-OH CH*OHHC1 -+Qg - + / ‘($C-CH,*CO,H CH*CH,*CO,H // C*CH,.CO,Na , I’CHI 4(XII.)(VIII.) (IX.)Zn + HCl I oxidation 1 Me,SO,C=OMe phthdic acid +benzoquinone $pRf1XJS}8H p HMeC*OMe co(R = p-NO,*C,H,* ; R’ = p-NH,*C6H,*)Boiling with dilute mineral acids converts (VI), with eliminationof acetic acid, into 4’-nitro-3-phenylphthalaz- 1-one 98 (VIII).98 F. M. Rowe and E. Levin, J., 1928, 2658; A,, 1928, 1262124 BAKER AND BENNETT :The latter, in agreement with the betaine structure assigned,exhibits both acidic and basic properties and is reduced to thecorresponding amino-compound (IX), similarly obtained from(VII). The isomeric 4’-nitro-3-phenylphthalaz-4-onej which wassynthesised for comparison, is insoluble in alkali and does not yieldsalts with mineral acids. On the other hand, cold acid potassiumdichromate merely decarboxylates the acetic acid side chain in(VI), giving 4’-nitro-3-phenyl-4-methylphthalaz- 1 -one (XI).99A similar series of derivatives is obtained from 3’-nitrobenzene-land 2’ : 6‘-dichloro- and 2’ : 6’-dibromo-4’-nitrobenzene-2-naphthol-l-diazosulphonates,2 but it is perhaps significant that in this seriesof decreasing - I and increasing + T effects there is an increasingtendency towards the formation of the corresponding benzeneazo-p-naphthols.This, together with the fact that, with the diazo-sulphonates derived from such components as aniline, toluidine,2 : 5-dichloroanilineY and anthranilic acid, no formation of phthal-azine derivatives was observed, the presence of a nitro-group being,in fact, essential, suggests that the latter reaction is favoured by apowerful - I effect, whilst conversion into the benzeneazo-p-naphthol is facilitated by a + T effect. The observation that thenew reaction does not occur with the derivative from diazotisedsulphanilic acid may possibly be explained by the presence of the-++/- 0b anion S-0 in the alkaline medium, but the close polar similarityof the NO, and S0,R groups suggests that an examination of thederivatives from diazotised sulphanilic esters or p-aminophenyl-sulphones would be of interest.The scheme given on p.123 summarises the main results (includingthose obtained on methylation of the phthalaz-l-ones, which cannotbe dealt with in detail in this Report), the constitutions being thoseprovisionally assigned by the authors. The structure (VI), however,is open to the criticism that a nitrogen atom with two unsharedelectron pairs could scarcely be stable in juxtaposition to an am-monium nitrogen atom. A more probable constitution for thissubstance is the dihydrophthalaz-l-one structure (XV). 3 : 4-Elimination of acetic acid from its betaine form (XVI) would thenreadily explain its conversion into the phthalaz-l-one (VIII).The formation from the diazosulphonate would be represented bythe following scheme, which also suggests a reason for the facilit-ating effect of - I groups in the substituent R (see above).09 F.M. Rowe, E. Levin, and A. T. Peters, J., 1931, 1067; A., 835.1 F. M. Rowe, M. A. Himmat, and E. Levin, J., 1928,2556 ; A., 1928,1262.2 F. M. Rowe, C. Dunbar, and N. H. Williams, J., 1931, 1073; A . , 835ORGANIC CHEMISTRY.-PART TI. 12570,Naacids --+H*CH2*C02NaC*OH co--+CH*CH,*CO,H CH*CH,* C0,H- AcOH - 3 (VIII.) (XVI.)CH ;\H\ '\CH,*CO,HThis view of the structure (XV) receives support from thefact that preliminary experiments have recently indicated thatN-methyl derivatives of substances of this type can be prepared.,Anthracene.-Certain evidence has been adduced in favour ofthe Armstrong rather than the Graebe-Liebermann structure foranthracene.Thus reduction of active benzyloxanthrone-2-carb-oxylic acids affords inactive 9-benzylanthracene-2-carboxylic acid,4whilst the presence of two double linkings at the 9- and 10-positionsagrees well with the observation that anthra- /\/I\,\ cene reacts additively with ap-unsaturatedacids and esters to form stable derivatives 1 11 :IaoIl I of the type anthracene-9 : 10-endomaleic \/\I/\/ anhydride (I). Such arguments are based,however, on a static conception, and itseems unlikely that the 9 : 10-bridged an-thracene form would be any more permanent than is the Dewarphase in benzene.g(1.)J Private communication from Professor F.M. Rowe.4 H. Goudet, Helu. Chim. Acta, 1931,14, 379; A., 619.6 Compare E. Clar, Ber., 1931, 64, [B], 1676, 2193; A., 1044, 1292.0. Diels and K. Alder, Annalen, 1931, 486,191 ; A., 848126 BAKER AND BENNETT :Alkylanthracenes. The presence of 2 : 6- and 2 : 7-dimethyl-,*2 : 3 : 6-trimethyl-, and 2 : 3 : 6 : 7-tetramethyl-anthracene in low-temperature tar has been codrmed by direct synthesis of thesehydrocarbons.’ The appropriate methyl-substituted benzophenone(prepared by condensation of the Grignard compound of a halogeno-methylbenzene with p-toluonitrile) is slowly converted into theanthracene on heating :Me MeM e o r + C N O M e M e ) C O O M e ---+Aromatic hydrocarbons may be conveniently alkylated by treat -ment with the appropriate alcohol and 80% sulphuric acid a t70-80”.* Thus, with benzyl alcohol, toluene and o-xylene afford,in addition to phenyl-p-tolylmcthane and 4-benzyl-o-xylene,anthracene and 1 -methylanthracene, respectively, the latter result-ing from oxidation of the intermediate 9 : 10-dihydro-derivatives.Owing to their carcinogenetic properties the synthesis of benz-and dibenz-anthracenes has recently assumed considerable import-ance.Distillation of l-benzyl-2-methylnaphthalene with zinc dustaffords mainly 1 : 2-benzanthracene together with a little of the2 : 3-compound, whilst the corresponding .ketone gives the 1 : 2-derivative exclusively.9 By the condensation of 2-methylnaphth-alene with p-naphthoyl chloride, E.Clar lo obtained Z-methyl-1 : 2’-dinaphthyl ketone, pyrogenic decomposition of which gives1 : 2 : 5 : 6-dibenzanthracene. I n the similar decomposition of2-methyl-1 : 1’-dinaphthyl ketone J. W. Cook l1 has shown thatisomerisation to the 1 : 2’-ketone probably precedes cyclisation,since 1 : 2 : 5 : 6-dibenzanthracene is again formed and not the1 : 2 : 7 : 8-compound (as stated by Clar). The same compound isobtained by pyrolysis of 1 : 4’-dimethyl-1 : 1’-dinaphthyl ketonewith elimination of the methyl group in the a-position to a con-densed ring.* Described by J. Lavaux and &.I. Lombard (BulE. SOC. chirn., 1910, [iv], 7,539; A., 1910, i, 548) as the 2 : 7-, and 1 : 6- or 1 : 7-derivatives, respectively.G. T. Morgan and E. A. Coulson, J . , 1929, 2203, 2551; 1931, 2323; A .,1929, 1436; 1930, 80; 1931, 1282.H. Meyer and K. Bernhauer, Monatsh., 1929, 53 and 54, 721 ; A., 1929,1441.K. Dziewoliski and E. Ritt, Bull. Acad. Polowise, 1927, 3, [ A ] , 181 ; A.,1928, 52.lo Ber., 1929, 62, [B], 350; A., 1929, 435.l1 J . , 1931, 487, 489; A., 612ORUANIC CHEMISTRY.-PART II. 127By the action of Grignard reagents on 1 : 2 : 5 : 6-dibenzanthra-quinone and dehydration of the resulting diols a series of 9 : 10-dialkyl-1 : 2 : 5 : 6-dibenzanthracenes has been prepared : l2 thespectrum of the dibenzyl derivative is almost identical with thatof the carcinogenetic mixture obtained by the action of aluminiumchloride on tetralin.13 Similar spectra are shown by phenanthra-naphthene and phenanthrafluorene. Pyrogenic decomposition of1 : 5-dibenzoyl-2 : 6-dimethylnaphthalene affords 2’ : 1’-anthraceno-1 : 2-anthracene l 4 (11), 1 : 2 : 3 : 4 : 5 : 6-tribenzanthracene(111) is obtained by similar methods.15andAnthraquinone derivatives.According to A. Meyer l6 conversionof anthraquinone-a-sulphonic acid into the p-acid by heating withsulphuric acid occurs only in the presence of metallic mercury.Since the p-acid is stable under the same conditions, the almostexclusive a-substitution observed l7 when anthraquinone is sul-phonated in the presence of mercury salts cannot be due to theprimary formation of the p-acid and subsequent isomerisation.Disulphonation of anthraquinone in the presence of mercury saltsgives 1 : 5-(4O-&y0), 1 : 8-(23y0), and traces of the 1 : 6- and 1 : 7-disulphonic acids.The a-directive catalytic influence of mercuricsalts is also emphasised in a systematic re-investigation of thesulphonation of or-chloro- and dichloro-anthraquinones by A. A.Goldberg,ls in the course of which 1 : 2 : 3-, 1 : 2 : 5-, 1 : 2 : 6-,1 : 2 : 7-, 1 : 3 : 6-, 1 : 3 : 7-, 1 : 4 : 5-, and 1 : 4 : G-trichloroanthra-quinones are synthesised and oriented.An interesting series of A2-tetrahydroanthraquinones hydro-genated in the “ twin ” positions (designated as y in naphthaleneand 6 in anthracene), obtained by addition of hydrocarbons withl2 J. W. Cook, loc. cit.lS I. Hieger, Biochem. J., 1930, 24, 505; A., 1930, 807.l4 E. Clar, F. John, and B. Hawran, Ber., 1929,62, [B], 940, 950; A., 1929,l5 L.F. Fieser and E. M. Dietz, ibid., p. 1827; A . , 1929, 1055.l6 Compt. rend., 1926, 183, 519; Bull. SOC. chim., 1927, [iv], 41, 1627; A.,1926, 1146; 1928, 781.1’ H. E. Fierz-David, Helv. Chim. Acta, 1927, 10, 197; 1928, 11, 197; A.,1927, 463 ; 1928, 293.1s J., 1931, 1771 ; A . , 1062.689128 BAKER AND BENNETT:conjugated double linkings to benzo- and naphtha-quinones wasmentioned in Part I of last year’s Report.19Bolecular Compounds.Under this heading some reference will also be made to therelated subjects of association and chelation. If these phenomenaare approached from the merely physical point of view, theirexplanation may be attempted exclusively in terms of dipoleassociation and intermolecular forces due to the residual valencyfields of the molecules or van der Waals forces.Many chemists,however, although admitting the importance of the considerationsmentioned, prefer to regard the formation of new chemical mole-cules as a, principal factor. Thus, N. V. Sidgwick20 in particularhas represented association as involving the formation of newmolecules by means of co-ordinate linkages, and the extension ofthis view to molecular compounds in general would seem reason-able.21 Molecular compound formation and association will thenbe regarded as similar intermolecular co-ordination phenomenabetween different and like molecules respectively, whilst chelationmay be regarded as the analogous intramolecular process. It maybe pointed out that this view follows the trend of modern chemicaltheory, which has been increasingly to attempt the representationby definite structural formula? of complexes a t first regarded asindefinite assemblages, the most important instance of this beingWerner’s theory, and another recent one the polymerisation theoryof H.Staudinger.22The isolation of a crystalline specimen is definite evidence of achemical fact, but it is necessary to bear in mind that forces of thesecond order may contribute largely to the stability of a crystalassemblage. A true molecular compound will survive, to someextent at least, in the liquid or dissolved condition, and proof ofthis survival in the liquid state is very desirable. To take a recentexample, compounds have been isolated in which either deoxycholicacid or apocholic acid (dihydroxy-carboxylic acids, saturated andcontaining one double bond, respectively) is associated with 8 mols.of either stearic or palmitic acid or of cetyl It is difficultto believe that these are united by chemical bonds.The recent appearance of a new edition of the important mono-lb Ann.Reports, 1930, 27, 89.2o “ The Electronic Theory of Valency,” 1927, p. 132.21 T. M. Lowry, Chem. and Id., 1924, 43, 218 ; G. M. Bennett and G. H.Willis, J., 1929, 256 ; A., 1929, 436.22 Ann. Reports, 1929, 26, 105-115.23 H. Rheinboldt, 2. physiol. Chem., 1929,180, 180; A., 1929, 443ORQANIC CHEMISTRY.-PART 11. 129graph on molecular compounds by P. Pfeiffer 24 makes availablean admirably classified survey of this field.Chelate Compounds.-The chelate ring in o-substituted aromaticcompounds, first suggested by P.Pfeiffer 25 for l-hydroxyanthra-quinone, was closely studied by N. V. Sidgwick and his colleagues.26An examination has now been made 27 of the parachors of certainaromatic compounds having the chelate ring. Four substances ofthe type o-C,H,(OH)X (where X is NO,, CHO, or C02Me), such as(I), consistently show a defect in their parachor values of 14.4below that calculated (due allowance being made for the presenceof a six-membered ring), whereas the corresponding methyl ethersgive normal values. Of this defect, 1.6 is regarded by the authorsas due to the transferred electron and 12.8 as associated with therise in the covalency of hydrogen from 1 to 2.In passing from ter-to quinque-valent phosphorus in the chlorides there is a similarcontraction of 25 units. These figures can equally well be numeric-ally explained by means of formulae involving singlet linkages.28S. Sugden regards the hydrogen atom with four shared electronsas inconsistent with Pauli’s principle. N. V. Sidgwick29 contendsthat this is not so, a second quantum group being permissible,but that the odd electron of the singlet linkage would on the otherhand be definitely anomalous.Marked differences of boiling point and solubility have beenascribed to self-co-ordination in the o-nitrobenzyldialkylamines (II),30but the melting points of o-nitrodialkylanilines31 should not betaken as evidence of chelation, and the four-membered ring whichwould be formed is improbable on stereochemical grounds.Itseems unjustifiable, moreover, to postulate chelate rings involvinghydrogen attached to carbon.3224 “ Organische Molekulverbindungen,” 2nd Ed., 1927.25 Annalen, 1913,398,137; A., 1913, i, 879.26 Ann. Repwts, 1924, 21, 104.27 N. V. Sidgwick and N. S. Bayliss, J., 1930, 2027; A., 1240.28 S . Sugden, “ The Parachor and Valency,” 1930, p. 150; and J., 1927,29 LOC. cit. and J., 1924, 125, 532; A., 1924, i, 506.30 G. M. Bennett and G. H. Willis, Zoc. cit.31 H. H. Hodgson and A. Kershaw, J., 1930, 497; A., 1930, 595.32 W. Kistiakowski, 2. physikal. Chem., 1928, 137, 383; E. de B. Barnettand J. A. Low, Ber., 1931, 64, [B], 49; T. M. Lowry and W.V. Lloyd, J.,1929, 1784; A., 1929, 22, 1186; 1931, 341.1173; A., 1927, 714.REP.-VOL. XXVIII. 130 BAKER AND BENNETT:A number of chelate metallic compounds are referred to in thefollowing section.The subject of association cannot be discussed here in detail,but reference may be made to reviews of the evidence from observ-ations of molecular polarisation by J. Errara 33 and C. P. Smyth.34Metal Co-ordination Cmpounds.-The lithium halides form withaliphatic amines a series of compounds decomposing a t or belowlaboratory temperature and containing 1 - 4 and sometimes 5molecules of amine per atom ofCompounds of the type MB,HB [where HB is +indoxyl-spiro-cyclopentane and M is lithium, sodium or potassium], soluble intoluene, are formed even in presence of a large excess of the alkali.36They are satisfactorily represented as co-ordination compounds,c/c=o,\C-”Ithe metal being 4-covalent and combined,MgoE‘c\c as in the annexed formula.Alkali-metalderivatives of diketonic substances andpounds with a second molecule of the parent substance; and alkalio-nitrophenoxides combine with one molecule of ~alicylaldehyde.~’All these products are bichelate compounds with 4-covalent metalNH-c// I of salicylaldehyde tend to form com-atoms.A number of addition compounds of calcium, strontium, andbarium halides with one, two, or three molecules of mono- anddi-methylaminoacetic acid and betaine have been des~ribed,3~ suchas CaBr,,X,; BaI,,X,,4H20; SrC12,X,4H,0, where X is a mole-cule of the N-methylated glycine.Following Pfeiffer,39 the authorsformulate them with the carboxyl group co-ordinated to the metal,the alkylamino-groups being left free (I). The survival of a freeamino-group is, however, improbable: the formula (11) is to bepreferred, as it overcomes t,his difficulty and a t the same time showsthe analogy of the betaine complexes with the others.{ [Me,N*CH,*C( OH) :O],Ba( H20),)12 [( Me,lfH-CH,*C~),Ba( H20) ,]IT,(1.1 (11.133 “ The Dipole Moment and Chemical Structure,” P. Debye, 1931, p. 101.3 4 “Dielectric Constant and Molecular Structure,” 1931, p. 169; C. P.Smyth and others, J. Amer. Chern. Soc., 1929,51, 1736, 3312; 1930, 52, 1824;A., 1929, 994; 1930, 135, 841. See also N. V. Sidgwick, Zoc. c i t . and 2.EZeEtrochem., 1928, 34, 445; C.T. Zahn, Physical Rev., 1931, [ii], 37, 1516;A , , 1928, 13; 1931, 895.35 A. Simon and R. Glauner, 2, unorg. Chem., 1929,178,177; A . , 1929, 431.36 N. V. Sidpick and S. G. P. Plant, J . , 1925, 127, 209; A., 1925, i, 298.37 N. V. Sidgwick and F. M. Brewer, ibid., p. 2379; F. M. Brewer, J., 1931,38 W. K. Anslow and H. King, Biochem. J . , 1928,22, 1253; A., 1928, 1362.39 Op. cit., p. 140.361 ; A . , 1926, 71 ; 1931, 443ORGANIU CHEMISTRY.-PART II. 131A series of compounds of calcium chloride with four and of calciumiodide with six molecules of various amides such as urea, thiourea,urethane, phenylurea, and antipyrine have been prepared.40o-Phenylenediamine and o-tolylenediamine form a number ofcomplexes with cadmium and zinc chlorides and bromides contain-ing either two, three, or four molecules of amine per atom of themetal,41 such as Br2, and there are indicationsof the occurrence of isomerism.The octammine type is unusual andtherefore of particular interest : a cubic configuration is suggested.One or two complexes with m- and p-phenylenediamines are described,but evidence as to their molecular complexity would be desirable.A study of the conductivities in acetone and of the parachors ofsome sulphonium and ammonium mercuri-iodides leads to the con-clusion that they should be formulated as [R,N]+[HgI,]- andI *I [R,S]+[HgI,]- and that the negative ion may be I-Hg<I or I-HgeIThe compound of mercuric iodide and dibenzyl sulphide is a non-electrolyte and may be written R2S-Hg<I or alternatively withsinglet linkages.42Remarkably stable chelate compounds of dimethylthallium withP-diketones and salicylaldehyde have been prepared, a typicalstructure being Me T1’ b H .4 3 Parachor determinations onthese and other derivatives of P-diketones ‘with beryllium andaluminium are held by S. Sugden to point to the formula of thetype No It is here againequally possible to reconcile them with the alternative formula.The free diketones may contain some chelate form and some open-chain diketone.The hydrocarbons stilbene, dimethylbutadiene, and caroteneyield definite compounds with antimony and ferric chlorides ofthe composition X,2MC13.4540 F. R. Greenbaum, J . Amer. Phrrn. ASSOC., 1929,18, 784; A., 1929, 1169.41 W.WaM, F6rh. 111 nord. KernistmBtet, 1928, 172; A., 1929, 167.d2 H. J. Cave11 and S. Sugden, J., 1930, 2672; A., 26.43 R. C. Menzies, N. V. Sidgwick, E. F. Cutcliffe, and J. M. C. Fox, J., 1928,44 J., 1929,316; A., 1929,427; compare N. V. Sidgwick and L. E. Sutton,45 H. von Euler and H. Willstaedt, Arkiv Kerni, Min. Geol., 1929,10, B, 9 ;Zn NH2 C H [ ( N H h 4 ) JIO-CR‘O=CR/OTCRb H with singlet linkages.44-1288 ; A., 1928, 746.J . , 1930, 1472; A., 1930, 1062.A., 1930, 333132 BAKER AND BENNETT :A stable pink complex of nickel chloride with three molecules ofad-dipyridyl has been resolved into optically active forms throughthe tartrate. The salt [Ni,3dipy]C12,6H20 had [a]5461 + 550", theactivity vanishing rapidly in aqueous solution.46Nickel cyanide combines with aliphatic and aromatic amines,the complexes containing one or two molecules of the base.*' Thelack of further co-ordinating power in the compoundis held to show that the nickel is 4-covalentY the phenyl nucleusproviding one point of attachment.The molybdenum derivatives of several P-diketones prove tohave the composition MoO,[R*CO*CH*CO*R], and are typical co-ordination compounds.48The discovery of bivalent silver salts 49 has been followed by thepreparation of several complexes of the type [AgB,]X2, where Bis a-phenanthroline 50 or ad-dipyridyl.51I n forming co-ordination compounds with copper salts, themolecule of thiourea (tu) tends to form a, bridge between meta,latoms so that, besides the simple compound [Cu,2tu]N03,H20,there result others such as [Cu2,5tu](N03),,2H,0 ; [C~,,7tul(NO~)~ ;[Cu5,11tu](N03)5,8H20.The cyclic ethylenethiourea ( = etu) doesnot cause this complexity, and a series of compounds results suchas [ Cuy4etu]X, [Ag,4etu]X, [ Cd ,4e tu]X2 , [ Hg , 3etu]X2 , [ Pd ,4etu]X2together with one or two complexes with lower co-ordinationnumbers such as [Ag,3et~]C1.~1"The co-ordination -of chloroplatinic and chloroplatinous acidswith spy-triaminopropane, CH2(NH2)*CH(NH2)*CH,-NH2, yieldsthe interesting compounds[Ni( CN),,CH,Ph*CH,-NH,]which were each resolved into their optically active components byF. G. Mann.52 The central carbon atoms (*) are asymmetric. Acomplex with the same base of the formula [PtB2]12 was also pre-pared in which all the amino-groups are definitely engaged, makingthe metal 6-covalent.A number of other co-ordination compounds46 G. T. Morgan and F. H. Burstall, J., 1931, 2213; A., 1168.4 7 E. Hertel, 2. anorg. Chern., 1929, 178, 202; A., 1929, 380.4 8 G. T. Morgan and R. A. S. Castell, J., 1928, 3252.49 Ann. Reports, 1928, 25, 44.50 W. Hieber and F. Miihlbauer, Ber., 1928, 61, 2149; A., 1381.51 G. T. Morgan and F. H. Burstall, J., 1930, 2694; A., 234.51u Idem, J., 1928, 143; A., 1928, 278.52 J., 1927, 1224, 2904; 1928, 890, 1261; 1929, 409, 651; A., 1927, 754;1928, 157, 622, 745; 1929, 545, 678. Compare Ann. Reports, 1826, 23, 109OWANIC CIfEM3STRY.-PART II. 133of nickel, cobalt, and platinum with various aliphatic diamines suchas mono- and tri-aminotriethylamine, p-methyl-, p-hydroxy-, andp-bromo-trimethylamine have been described by the same author.The volatile acetylacetonate of trimethylplatinum 53 and asimilar derivative of diethylgold are evidently chelate compounds.The latter was isolated in the course of experiments with diethyl-gold bromide in which a number of other compounds of 4-covalentgoldwere describedsuchas the amines [Et2AuBr,NH3], [Et&uBr,Py],[Et2Au en]Br, and [Pra&u en]Br.From gold bromide there werealso prepared deep-coloured complexes [AuBr,,B] (where B ispyridine, quinoline, isoquinoline, 2-aminopyridine, or di-2-pyridyl-amine) and the compounds [AuBr,,Py,]Br and [Au en2]Br,.Compounds with Co-ordinated Oxygen.--Aliphatic alcohols yieldcrystalline hexacyanocobaltiates of the types [ROH,],[Co(CN),]and [ROH,],H[Co( CN),] which are evidently normal alkoxoniumderivatives.55 More complicated compounds are formed fromalcohols, ethers, and acetone with hydrogen chloride or bromideat low temperatures,s6 such as Et20,2HX and Et20,5HC1.Tri-phenylcarbinol yields with sulphuric acid the compoundCPh ,*OH,4H2S 0,. 57The coloration shown by cholesterol in chloroform with sulphuricacid is due to a red equimolecular compound, cholesterol sulphate.58The shape of the solubility curves of a number of aliphatic andcyclic ethers,59 of the isomeric butyl alcohols,60 and of cyclohexanolin water 61 indicates that there is some combination between thesesubstances and the water.There are several factors involved, butthe figures for the isomeric methyl butyl ethers show the expectedincrease of solubility arising from the increased co-ordinatingpower of an oxygen atom under the influence of the polar effectsof the adjacent methyl groups. That dibenzyl ether causes thepartial dissociation of benzoic and acetic acids dissolved in it mayalso be attributed to co-ordination of the acids with the ether.6263636s R. C. Menzies, J . , 1928, 565; A , , 1928, 609.54 C. S. Gibson and J. L. Sirnoneen, J., 1930,2531 ; C. S. Gibson and W. M.Colles, J., 1931, 2407 ; A., 78, 1316.5 5 F. Holzl, T. Meier-Mohar, and F. Viditz, Monutsh., 1929, 52, 73; A.,1929, 898.56 D.McIntosh,Proc.NovaScotiun Inst. Sci., 1928,17,112,116; A., 1929,292.5 7 H.R. Dittmar, J . Physical Chem., 1929,33, 633; A., 1929, 656.58 V. J. Nikolaev and S. A. Krastelevskaja, J . Rum. Phys. Chem. SOC., 1928,59 G. M. Bennett and W. G. Philip, J., 1928, 1930; A., 1928, 944.60 D. C. Jones, J . , 1929, 799; A., 1929, 638.61 N. V. Sidgwick and L. E. Sutton, J., 1930, 1323; A., 1930, 988.62 G. M. Bennett and G. H. Willis, J., 1928,2306; A., 1928, 1089.60, 1211 ; A., 1929, 61134 BAKER AND BENNETT :A qualitative test for weak bases, depending upon the formationof ferrichlorides in solution, is given by most oxygen, nitrogen, andsulphur c0mpounds.6~A crystalline sulphate of nitrobenzene, [Ph*NO,H]HSO,, has beendescribed, which is an ele~trolyte.~~ The increase in the conduc-tivity of sulphuric acid caused by addition of p-nitrotoluene is greaterthan with nitrobenzene, whilst dinitrobenzene has no such effect.Nitrobenzene forms an equimolecular compound with stannicchloride and the existence of the compounds 2PhN02,SnCI,(Br,)is deduced from the viscosity curves of mixtures of these suh-stances. 65A study of the freezing points of mixtures of carbon tetra-chloride and chloroform with acetone and ether66 shows theexistence at low temperatures of a number of compounds such asCCl,,Et,O ; CC14,2Et,0 ; CCl,,COMe, ; CHCl,,3Et20 ; CHCI3,2Et2O ;CHCl,,Et,O ; and 2CHC13,Et20.Cmpounds of Polynitro-aromatic Substances.-The largest groupof molecular compounds (apart from co-ordination compounds ofthe metals) comprises those formed by the union of substancessuch as di- and tri-nitrobenzene or picric acid with bases, aromatichydrocarbons or their derivatives.In the development of ourknowledge of this class of complex J. J. Sudborough,67 A. Werner,68and P. Pfeiffer 69 took prominent parts.The problem of their structure is a difficult one : but it is clearfrom a general review of the facts that, if they are formed by mutualco-ordination of the components, the hydrocarbon or base is thedonor molecule and the nitro-compound the acceptor. 7O Theirdepth of colour is in general roughly parallel with their stability 7land the influence of polar groups substituted in either componentis consistent with this view.A further illustration of this point is provided by the series ofcompounds described by C. A.Buehler and his colleagues fromm-dinitrobenzene, 2 : 4-dinitrotolueneJ 2 : 4-dinitrophenol, andchloro- and bromo-2 : 4-dinitrobenzenes with a series of bases,63 R. Robinson, J., 1925,127, 768; A., 1925, ii, 606.61 E. Cherbuliez, Helv. Chim. Acta, 1923,6, 281 ; A., 1923, i, 452. Comparc6 5 F. de Carli, Atti R. Accad. Lincei, 1929, [vi], 10, 186, 250, 372; A . , 1930,66 W. F. Wyatt, Trans. Faraday SOC., 1929, 25¶ 43; A., 1929, 254.G 7 J . , 1901, 79, 622; 1910, 97, 773; 1916,109, 1339.6 8 Ber., 1909, 42, 4324; A., 1910, i, 20.70 G. M. Bennett and G. H. Willis, Zoc. cit.7 1 P. Pfeiffer, op. cit., p. 348.I. Masson, J., 1931, 3200.26, 284.60 op. citORGANIC CHEMISTRY .-PART 11. 135hydrocarbons and naphthols. 72 The stability of the productsdiminishes with substitution in the dinitro-compound in the orderCl>Br > OH>H >Me >NH2.p-Chloronitrobenzene also combineswith cc-naphthol.Other recent additions to the list of such compounds are severalfrom tetryl with aromatic hydrocarbon^,^^ a compound of 1 : 3 : 4 : 5-tetranitrobenzene with benzene (1 : 1),74 and of 4 : 4'-dithiolarseno-benzene with picric acid (1 : 2).75 Others are mentioned in thediscussion below.The stability of such complexes in solution has been the subjectof repeated investigation. R. Behrend in 1894,76 by measuringthe joint solubilities of anthracene and picric acid in alcohol a t25", found a dissociation constant for the compound of 4-7-57;and the union of picric acid and p-naphthol was studied in a similarmanner.77 Determinat'ions of freezing points of solutions of thecomponents in nitrobenzene by F.S. Browsl 78 show a dissociationconstant for naphthalene picrate of 0.23 and for naphthalene-s-trinitrobenzene of 0.46. The free energy of formation of the formercompound was calculated to be 2080 cals. per mol., which agreeswell with the value 2150 cals. found by J. N. Bronsted fromdeterminations of E.M.F. in the system naphthalene-picric acid-potassium chloride-hydrochloric acid-~ater.7~The same question has been examined by H. von Halban andE. Zimpelmann 80 by the photoelectric determination of the absorp-tion of light in solutions of various concentrations at more thanone temperature. The following values of the heats of formation(&) and the equilibrium constants ( K ) at 25" were found :Compound.Solvent. Q (cals.). K.Acenaphthene-s-trinitrobenzene ... Tetrachloroethane 2450 0.41Acenaphthenem-dinitrobenzene ... Y f 1350 3-15Acenaphthene-picric acid ......... 9 , 950 0.50Anthracene-picric acid ............... Chloroform 3990 0.2272 C. A. Buehler with A. G. Heap, J . Amer. Chem. SOC., 1926,48,3168; withA. Hisey and J. H. Wood, ibid., 1930, 52, 1939; with C. R. Alexander andG. Stratton, ibid., 1931, 53, 4094; A., 1927, 141; 1930, 905.N. N. Efremov and A. M. Tichomirova, Ann. Inst. Anal. Phys. Chem.,1928, 4, 65, 92.74 A. F. Holleman, Rec. trav. chirn., 1930, 49, 112, 501; A., 1930, 333, 900.75 S. Krishna and R. Krishna, J . Indian Chem.Soc., 1929,6,665; A., 1929,76 2. physikal. Chem., 1894, 15, 183; A., 1895, ii, 71.' 7 B. B. Kuriloff, ibid., 1897, 23, 90, 673; A., 1897, ii, 397, 484.78 J., 1925, 127, 345; A., 1925, ii, 296.7 9 2. physikal. Chem., 1911, 78, 284; A., 1912, ii, 20.80 Ibid., 1925, 117, 461; A., 1926, 25.1320136 BAKER AND BENNETT :A precise investigation of such equilibria by the partition methodhas recently been published by T. S. Moore, F. Shepherd, and E.Goodall,sl who have examined a range of substances wide enoughto reveal the effects of substitution. Compound formation betweenpicric acid and a series of aromatic substances in chloroform a t18" is calculatled from indirect determinations of the free picricacid present. The latter is found by observing the concentrationin an aqueous phase in contact with the chloroform.Similarlywith aniline and a series of aromatic substances in chloroform a t25", the aniline in the water phase being determined bromometric-ally. The assumption of (1 : 1)-combination gives good equilibriumconstants K . A relatively small effect which the added substancesexert on the partition curves is studied and is expressed by the''. depression constants " below.K .Benzene ... 0.09Toluene ... 0.12o-Xylene ... 0.16p - ,, ... 0.16Mesitylene 0.18Tetrahydro -naphthalene 0.26?I%- ,, ... 0.14Picric Acid Cmpounds.K .Bromobenzene ...... 0.09 Nitrobenzene ......o-Dichlorobenzene 0.10 m-Dinitrobenzene ...... P- 9 7 0.03 o-Nitrotoluenea-Bromonaphthalene 0-23 p- ......a-Methylnaphthalene 2.76 s-Trinitrotoluene ...8- ?, 3-44 NitronaphthaleneY YNaphthalene .........2.17 2 : 4-Dinitrotoluene...Aniline Cmpounds.K .Nitrobenzene ..................... 0.01 m-Dinitrobenzene ........................ ............... Y , a-Nitronaphthalene 0.02 P-p-Chloronitrobenzene ......... 0-03 2 : 4-Dinitrotoluene ............m-Bromonitrobenzene ......... 0.06 1 : 6-Dinitronaphthalene ......s-Trinitrotoluene ...............K.0.561.030.410.570.671.021.02K.0.170.180.150.570.56Depression Constants.Picric acid. Aniline.Hexane ... 0-54 o-Nitrotoluene ...... 0.01 a-Bromonaphthalene 0.06Hexachloro- Toluene ............... 0.12 Naphthalene ......... 0.07benzene 0.38 p-Xylene.. .............0.15 o-Dichlorobenzene 0.06Carbon tetra- a-Methylnaphthalene 0.09 p-Dichlorobenzene 0-06chloride 0.27 j3- 9 9 0.09 Bromobenzene ...... 0.06The depression constants may be regarded as minor influenceson mutual solubility arising from differences of polarity among thesubstances, and the smaller values of K in the aniline series maybe attributed to increases of solubility arising in a similar way,although some would regard the colour produced as evidence ofcombination. With di- and tri-nitro-compounds the authorsregard the K values as representing combination of the aniline.cycloHexane 0.48 p - ...... 0-02 Carbon tetrachloride 0- 10 9 7*1 J., 1931, 1447; A , , 949ORGANIC CHEMISTRY .-PART If. 137In the picrate series the naphthalenes combine more than thebenzenes and the union is increased by methyl groups but dimin-ished by chlorine atoms in agreement with the general conclusion(p. 134) as to the function of the components.There is, however, a large increase of combination evident whenone or more nitro-groups are introduced into the molecule, althoughthe converse result might have been expected from the oppositepolar nature of the methyl and the nitro-group.The solutionsin these cases, nevertheless, show relatively little colour. A paleyellow compound of picric acid with nitronaphthalene was isolated(previously detected from the freezing-point curve).82The authors ascribe the different type of union here involvedto mutual dipole association of nitro-groups in the two substances. 83It is equally possible to attribute the union to the basic propertyof the nitro-group as exhibited in nitrobenzene sulphate (p.134).The stability and conditions for isolation of solid molecularcompounds have been discussed by 0. Dimroth 84 and E. Bamberger.For the reaction A + B e C, they write the equilibrium constantK = [C]/[A][B] = G . c/ab, in which a, b, and c are the solubilitiesof A, B, and C in a given solvent, and G is a stability constant forthe crystalline molecular compound C which is independent of thesolvent used. a! may also be written as Xc/XA. SB, where Xc, XA,and SB are the degrees of saturation of the compounds C, A, and Brespectively. This law is verified for anthracene picrate with thesolvents alcohol, chloroform, ether, petroleum, and carbon tetra-chloride.It leads to a precise statement of the conditions underwhich a solvent causes the decomposition of a molecular compound.No decomposition occurs when the quotient (solubility of the moresoluble component)/(solubility of the less soluble component) < G.By determining the free picric acid in a small volume of waterin equilibrium with partly decomposed compounds, the followingrelative order of increasing stabilities is found for a series of picrates :benzene < fluorene <anthracene <indene < phenanthrene < naphthal-ene <acenaphthene < a-methylnaphthalene < p-methylnaphthalene.The problem of formulating these substances has attractedconsiderable attention. Werner 85 argued that the nitro-group is82 P.L. Jovinet, Mdm. Poudres, 1928,23, 36; A., 1928, 1085.83 The freezing-point curve of nitrobenzene with m-dinitrobenzene has mean-while revealed the existence of an unstable (1 : 1) compound, and that ofnitrobenzene and 8-trinitrobenzene a (2 : 1) compound (D. L. Hammick,L. W. Andrew, and J. Hampson, J., 1932, 171).84 Annalen, 1924, 438, 67; A., 1925, ii, 36.85 LOC. cit. The argument from colour development would indicate combin-ation of tetranitromethane with a wide range of unsaturated or basic sub-stances; see E. M. Harper and A. K. Macbeth, J., 1915, 107, 87, 1824.E 138 BAKER AND BENNETTa point of attachment, since aliphatic nitro-compounds such astctranitrometbane and hexanitroethane yield similar complexes.Definite structural formuh for the base-nitro-compound complexeswere discussed by Sudborough in his earlier papers, but in 1915 86he adopted the residual valency conception of P.Pfeiffer, whichmay be represented by the typical formuh C,H,(NO,), . . . C&,and C,H3(N02), . . . . C,,H,*NH,. This was also supported byJ. Kenner.87The idea that a more precise formulation should be possiblehas been recently revived 88 and it is suggested that the unionmust be through a single nitro-group, any others present in thesame molecule having an indirect reinforcing effect. In support ofthis it was shown that, although molecular compounds of mono-nitrobenzenes (excepting nitrophenols) have hitherto been prac-tically unknown, the substances (I) and (11), in which the nitro-group is similarly reinforced by other electron-attracting groups,do yield such complexes.0Ar-N<NO2$ 0NR3(11.) (111.)The nitro-compound is regarded as united to an amine throughthe basic nitrogen atom as suggested by Werner and the union isnow formulated as in (111).Combination with hydrocarbons isregarded as occurring with one ethylene bond in the polarisedform - 1.. 'H-CH-. The occurrence of 1 mol. of ethylene or ofbenzene in metal co-ordination compounds such as [PtCl,,NH3,C,H,],[Ni(CN),,NH3,C6H6], and the nickel cyanide complex mentionedon p. 132, points definitely to the provision of one covalency by onemolecule of ethylene or one benzene nucleus.The behaviour of dinitrobenzenes with ammonia should havesome bearing on the question of structure here involved.W. E.Garner and his colleagues have found 89 that m-dinitrobenzenecombines slowly with (anhydrous) ammonia, the complex thenrapidly ionising to give a coloured negative ion containing theorganic substance. The nature of the positive ion is less clear,but it is regarded as NH, or N,H,. . .@86 L O C . cit.8 7 J. Kenner and M. Parkin, J., 1920,117, 855.88 G. M. Bennett and G. H. Willis, Zoc. cit.M. J. Field, W. E. Garner, and C. C. Smith, J., 1925, 127, 1227; W. E.Garner and H. F. Gillbe, J., 1928, 2889; A., 1925, ii, 792; 1929, 29ORUANIC CHElKIS”RY.-PruzIT 11. 139The question of the structure in the crystals of these molecularcompounds has, been attacked by the X-ray method 90 and it hasbeen shown that they have crystal lattices distinct from those oftheir components.Suggestions are made as to the arrangement ofthe units, but the positions of the atoms have not been determined.The formation of distinctf yellow and red isomeric compoundsof picric acid or dinitrophenols with halogenated amines is termedby E. Hertel ‘‘ complex isomerism ” and the substances are formul-ated as [(NO,),C,~*O][NH,Ar] (yellow) and HO*C,H,(NO,), . . . .ArNH, (red).s1The production of colour in the formation of molecular compoundsis a circumstance which has been of considerable importance intheir study. Mention may be made of some observations on halo-chromism in chalkone derivatives 92 and in styrylquinoxalines 93and of a recent theory of the probable origin of colour in quin-h ydrones .94The Bile Acids.(Continued from ,4nn. Reports, 1928, 25, 163.)9 QImportant modScations in the formulze of all the cholane deriv-atives must follow from the recent developments in this field. Thenew evidence indicates that ring I11 and probably also ring I areeach seven- instead of six-membered (compare cholanic acidformula, p. 140).The structure (111) originally assigned g5 to the tetrabasic acidC1,H,,O, obtained from pyrodeoxybilianic acid (I) through thedibasic diketo-acid C23H3406 (11) has been found to require re-vision.96 The passage from the diketo-acid (11) to (111) involvedthe shortening of the side chain by two carbon atoms, and as relatedcompounds such as choloidanic acid s7 retained their complete sidechain under similar conditions, it seemed doubtful whether (11)was actually attacked in this way.The dibasic keto-acid, C1SH220s(IV), obtained by heating the tetrabasic acid (111), was reduced byso E. Hertel, 2. physihd. Chem., 1930, [B], 7, 188; 1930, [B], 11, 69, 77;91 E. Hertel and I(. Schneider, {bid., 1930, 151, 413; L931, [B], 13, 387;O8 W. Dilthey, L. Neuhaus, and W. Schommer, J . pr. Chmn., 1929, [ii], 123,D3 G. M. Bennett and G. H. Willis, loc. cit.s4 E. Weitz, 2. Elektrochem., 1928, 54, 638; A., 1929, 190.96 H. Wieland and F. Vocke, 2. physiol. Chent., 1928, 177, 68; A., 1928,O * Idem, &id., 1930,191, 69; A., 1930, 1436.9 7 Ann. Reports, 1927,24, 137.A., 1930, 668; 1931, 153.A., 210, 1114.235; 1930, [GI, 124, 81; A., 1929, 1300; 1930, 604.1007. Compare Ann.Reports, 1928,2!5, 160140 BAKER AND BENNET*:Me Me0- & ~ C ~ & f e * ( C H , ) , * C O , HL-- Lo\/Cliolanic acid (1928) (I.) Pyrodeoxybilianic acid(11.) Dibasic C ,,-diketo-acid(V.) Tribasic C,,-acid (IV.) Dibasic CIS-ke to -acid (111.) Tetrabasic C,,-acitlMe(VIII.) [ = (V) corrected]k 2 H \ '2E:CO2H(vI.) [= (IV) corrected]J. (VII.) [ = (111) corrected]MeI /'IA, (IX.) I ' ICO,H CHMe*(CH,),-CO,H?U Me1Me /rI ICHMeC0,H (CH,),*CPh,*OH HO=CPh, (CH,),*CPh,.OHJ.Mep32-q-CH2~ x I I . l ~ ! - > 3 CH, \ / CH-CH>CH2 \ +- f d; (XIII.)QH CHMe*CO,H CHMe8FG02H CPh,-OH (XIV.) OH*CPh, CH,*CO,ORGANIC CHEMISTRY.-PART n.141Clemmensen’s method to the dibasic acid CI5Hz4O4 (IX), and thelatter converted by the action of phenylmagnesium bromide on itsester into an acid tertiary alcohol (X) and a ditertiary alcohol (XI).The acid carbinol was oxidised to the dibasic acid C,4H2204 (XII),whilst the ditertiary alcohol yielded the diphenylcarbinol deriv-ative (XIII) of this acid. In either case a repetition of the Grignardand oxidation processes led finally to the carbinol acid (XIV)(derived from an acid C13H2,04). This shortening of the side chainby one carbon atom at a time would be impossible if the dibasicketo-acid had structure (IV), but the alternative structure (VI)with side chain longer by 2CH, groups is in complete agreementwith these results. If structure (VI) be accepted for the acidC15H2205, the tetrabasic C,, acid must be (VII), not (111), and thetribasic acid C,,HzoO6 must be (VIII) instead of (V).The occurrenceof n-butane-aryy-tricarboxplic acid together with (VII) in theoxidation products from the dibasic diketo-acid (11) is equallywell accounted for by the new formulae, which explain more readilythe simultaneous production of a-methylglutaric acid.Since the acids (VI), (VII), and (VIII) have two more carbonatoms in the side chain than was originally supposed, there mustbe a hydrogen atom and not an ethyl group in position 10 in thesecompounds. This, however, leaves C2H4 unaccounted for in theformule of the dibasic diketo-acid C,H,,O, and all related sub-stances. Two possibilities are discussed by the authors to meetthis difficulty.The less probable assumption is that in the courseof the oxidation of (11) a 10-ethyl group migrates to some otherpart of the molecule and is then oxidised away. The alternativeconclusion seems almost inevitable, namely, that the missing atomsform part of ring 111 of the cholane skeleton. It has not beenrigidly proved that this ring is six-membered : the availableevidence is consistent with its being one of 7 carbon atoms.In the first place the suggestion adopted was that a CHMe groupwas inserted between C,, and CIz so that (I) should bebut later work has pointed to the CHMe group being between C,and C,, whereby rings I and I11 will both be seven-membered.The first indication of this comes from a detailed study 98 of the98 H.Wielasd and V. Deulofeu, Z. physid. Chern., 1931,198,127; A., 841142 BAKER AND BENNETT:two dihydroxycholenic acids, C24H3804, obtained by dehydrationof cholic acid.99 Of these, one is catalytically reduced to deoxy-cholic acid, and consequently has the normal bile acid skeleton.The second (apocholic acid) is not so hydrogenated and a morecomplicated relationship seemed possible. However, the two acidsby the action of bromine yield with loss of hydrogen bromide thesame dihydroxycholadienic acid, C2,H3,04. Catalytic reductionsaturates only one of the two double bonds in this substance andthe product is very similar to apocholic acid but differs from itin rotatory power.These results are in any case inconsistent with the earlier cholicacid formula with a six-membered ring 111.For apocholic acidcannot have its double bond between C,, and C13, as it would thenbe readily reducible just as is the same bond in the cholatrienicacid from cholic acid. Nor has the other dihydroxycholenic acidan ethylenic linkage a t C12-C13, since it does not yield the chola-trienic acid by loss of water. That a double bond may have wanderedis probable, sulphuric acid and zinc chloride being the reagentsused in the reaction, but it is unlikely that it will have migratedinto another ring and consequently there must be seven membersin ring 111. suggest that the inert W. Borsche and A. R. ToddI\CH,/double bond in these acids is in fact a bridge, and they represent ringI11 in apocholic acid as (XV) with the CMe between C,, and C12.(XVII.)HO-CHOn the other hand, H. Wieland and V. Deulofeu prefer the formuh(XVI), (XVII), and (XVIII) for apocholic acid, dihydroxycholenicacid, and dihydroxycholadienic acid respectively. The environ-ment of the central double bond explains its inertness.55, 2302; A., 1921, i, 865; 1922, i, 1027.9g F. Boedecker and H. Volk, Ber., 1920, 53, 1852; 1921, 54, 2489; 1952,1 2. physiol. Chem., 1931,197, 173; A., 841ORGANIC CHEMISTRY .-PART II. 143upoCholic acid yields by loss of water on heating a cholatrienicacid in which, again, one tlouble bond resists hydrogenation. Alittle cholanic acid was, however, detected which suffices to confirmthe normal carbon skeleton of these substances.A closer examination of pyrocholoidanic acid has, moreover,2provided evidence that ring I is seven-membered. It is not a keto-anhydride acid as hitherto assumed (XIX) but a keto-lactonic acid.(XIX.) WX.) (XXI.)Titration with cold alkali reveals a second carboxyl group (be-sides that in the side chain), and hot alkali three : diazomethane,moreover, yields a dimethyl ester.The dehydration thus involvesonly one carboxyl group and if the keto-acid (XX) is first formed,the keto-lactone structure (XXI) seems probable. But this formulais contrary to Bredt's rule that a double bond is not stable adjacentto the bridge head of a bridged six-membered ring.3This difficulty is overcome by the use of the new formula so thatcholoidanic acid is (XXII) and the pyro-acid is (XXIII).1 CHMel C Me 1A (XXII.) 1 \ I / CO,H CH-C*CH,-CO,H ? F*CH2*Co2H91 V=zQ (XXIII. )0 CH2\ /€I02C*CH2---6H*C02Hv coThe formation of two stereoisomerides is also accounted for.These corrections are fundamental and further evidence will soon,it may be hoped, allow of a permanent revision of all the formulaein the cholane group.Ergosterol.(Continued from Ann.Reports, 1927, 24, 148.)Chemical investigation of the phytosterols has been principallyconcerned with ergosterol 4 on account of the connection betweenH. Wieland, L. Ertel, and W. Schonberger, 2. physiol. Chem., 1931, 197,31 ; A., 841.J. Bredt, H. Thouet, and J. Schmitz, Annalen, 1924, 437, 1 ; A. Windausand A. Bohne, ibid., 1925, 442, 7; A., 1924, i, 643; 1925, i, 552.4 Other sterols : see Ann.Repwts, 1929, 26, 211, and H. Wieland, withM. Asano, Annalen, 1929,473, 300; with G. A. C. Gough, ibid., 1930,482, 36;with W. M. Stanley, ibid., 1931, 489, 31; F. Reindel and A. Weichmann,ibid., 1930,482, 120; H. Sandqvist and J. Gorton, Ber., 1930, 63, [B], 1935;A., 1929, 1200; 1930, 1431; 1931, 1164; 1930, 1578, 1431144 BAKER AND BENNETT:this compound and vitamin-D, the latter having been isolated as acrystalline isomeride of the steroL5Ergosterol, obtained from yeast, is purified as the acetate or thebenzoate.6 Its molecule, C2,H420, contains four rings, a hydroxylgroup, and three double bonds. Complete reduction yields ergo-stane, C27H48, a hydrocarbon isomeric but not identical with thesimilar products from cholesterol, coprosterol, and sitosterol.Thecarbon framework is shown to be similar to that of cholesterol bythe observation of 0. Diels and A. Karstens that the main productsof dehydrogenation of ergosterol with selenium are the same twocrystalline hydrocarbons, CIsH16 and C25H24, previously found asdehydrogenation products of cholesteryl chloride. It is thereforeprobable that ergostane is a stereoisomeride of cholest ane. Thereare, indeed, indications that cholesterol may be converted intoergosterol by heat and the saturated derivative of sitosterol isequally closely related.lOOxidative degradation of ergosterol has yielded few resultsowing to the extreme sensitiveness of the sterol to all reagents :it absorbs oxygen (5 atomic proportions) from t'he air.6 F.Reindeland K. Niederlander 11 have, however, isolated a tribasic acid,C,H,(C02H)3, from the products of oxidation with nitric acid.This appears t o be a cyclopentadiene derivative arising from a&membered ring in the nucleus. An aldehyde, C,H,,*CHO, wasalso obtained by ozonisation, which yields on oxidation a methyl-glutaric acid, and points to the group >C:CH*CH,*CH,*CHMe, inthe side chain of ergosterol.When the sterol is catalytically reduced, two molecular pro-portions of hydrogen are added to give ergostenol, C2,H4,0, ortlhree with formation of the completely saturated ergostanol,C27H480.127 l3, 14Compare Biochemistry section of this report, p. 212.R. K. Callow, Biochem J ., 1931, 25, 79; A., 618.A. Windaus and W. Grosskopf, 2. physiol. Chem., 1922,124, 8 ; A., 1923,i, 75; F. Reindel and E. Walter, Annalen, 1928, 460, 212; A., 1928, 295.Compare 0. Diels, W. Glidke,and P. Kording, ibid., 1927, 459, 1 ; A., 1928, 169.F. C. Koch, E. M. Koch, and I. K. Ragins, J. BioZ. Chem., 1929, 85, 141 ;H. B. Lemon, ibid., p. 159; A., 1930, 256, 257; R. Schonheimer, Naturwiss.,1930,18, 881 ; A., 1930, 1577.lo F. S. Spring, J., 1930, 2664; A . , 219.l1 Annalen, 1930, 482, 264; A., 1930, 1578.l2 F. Reindel and E. Walter, Annalen, 1928, 460, 212 ; I. M. Heilbron andW. A. Sexton, J., 1929, 921; A., 1928, 295; 1929, 809.l3 F. Reindel, Annalen, 1928, 460, 131; A., 1928, 295.l4 F. Reindel, E. Walter, and H. Rauch, Annulen, 1927, 452, 34; I.M.Heilbron and F. S. Spring, J., 1929, 2806; A., 1927, 241 ; 1930, 210.* Annalen, 1930, 478, 129; A., 1930, 470ORGANIC CHEMISTRY .-PART 11. 145A significant observation is the production of a dibasic acid,C25H44( CO,H),, when ergostanol (allo-a-ergostanol = hexahydro-ergosterol) is oxidised with chromic acid 13 to the ketone ergostanone.This acid when heated gives a ketone, C26H440, by loss of carbondioxide and water, which points t o the rupture in the first placeof a six- or seven-membered ring.In connexion with vitamin-D the study of various isomeridesof the sterol is of the greatest importance, and a large number ofexperiments in this direction have been recorded of which a shortaccount will now be given. The changes in structure no doubtinvolve movements in position of the double bonds and changes insteric arrangement of more than one kind.The location of thepoints of unsaturation is a matter of difficulty, and the conclusionsso far reached must be regarded as tentative.Isomeric change in ergosterol was fist brought about by theaction of hydrogen chloride upon its acetate,l* the product, ergo-sterol-B (isoergosterol), being obtained in several stereoisomeric forms(ergosterols B,, B,, and B3). The only effect here is an alterationin position of the double bonds or possibly conversion of a cis-into a trans-configuration with respect to a double bond in the sidechain ; 15 for the products all yield the same perhydro-derivative.The addition of two atoms of hydrogen (sodium and alcohol) l6and their removal at another point by means of mercuric acetateor perbenzoic acid 173 1% 2o yields ergosterol-D : these operations ininverse order furnish ergosterol-P ; 18 and yet another isomeride,ergosterol-E, has been described as resulting from the combinationof these processes with the hydrogen chloride treatment.19 Adifferent series of compounds arises when the sterol is convertedby mercuric acetate into its dehydro-derivative and this is subjectedto the catalytic action of finely divided nickel at 220°.20 Theproducts, ergostatrienone-D and u-ergostatrienone, yield onreduction ergosterol-D and the new isomeride u-ergostatrienol.In the study of the irradiation of ergosterol with ultra-violetl5 A.Windaus, K.Dithmar, H. Murke, and F. Suckfull, Annakn, 1931,1* A. Windaus and J. Brunken, ibid., 1928, 460, 225; A., 1928, 424.l7 A. Windaus and A. Luttringhaus, ibid., 1930, 481, 119; A., 1930,1178.l* A. Windaus, W. Bergmami, and A. Luttringhaus, ibid., 1929, 472, 195 ;A. Windaus, W. Bergmann, and H. Butte, ibid., 1930, 477, 268; A., 1929,1065; 1930, 338.Is I. M. Heilbron, F. Johnstone, and F. S. Spring, J., 1929, 2248; A., 1929,1442.2O A. Windaus and E. Auhagen, Annalen, 1929, 472, 185; A., 1929,1065.488, 91 ; A., 1061146 ORGANIC CHEMISTRY.-PART 11.light, two physiologically inactive sterols, suprasterols I and 11,have been isolated besides the vitamin-D.21Certain sterols undergo isomerisation when heated with sodiumethoxide in ethyl alcohol at ZOO": this results in the productionof epi-derivatives (in which the configuration of the >CH*O€I groupis inverted), hydrogenation of a double bond occurring to some extentat the same time.15 Finally, the production of isomerides by suc-cessive dehydration and hydration l5 and by the action of benzoylchloride, followed by hydrolysis (ergosterol-C),2O has not yet beenthoroughly investigated.The isomerides B, D, E and F all retain their three double bondsand are reducible to ergostanol.Like ergosterol itself, they giveinsoluble addition compounds with digitonin. A second group ofsubstances is not precipitated by digitonin, but is converted byheating with sodium ethoxide into members of the first group.They are reduced catalytically to epiergostanol, the relationshipof which to ergostanol is evident from their oxidation to the com-mon ketone ergostanone.A third group, including u-ergosteroland the products of irradiation, is not precipitated by digitonineither at once or after heating with sodium ethoxide. These iso-merides also still have three double bonds but on complete hydro-genation they yield neither ergostanol nor epiergostanol. It is asyet uncertain whether the difference here is of a steric kind or oneinvolving a change in the structure of the carbon skeleton.A number of attempts have been made to define the exact posi-tions of the double bonds in ergosterol and its isomerides, but thisproblem has not yet been solved. One argument 22 used is basedon the similarity of the absorption spectra of ergosterol and ofcholesterilene, C27H44, a dehydration product of cholesterol, butapart from the uncertainty of such evidence it cannot yet beclaimed that the structure of this hydrocarbon has been deter-mined beyond doubt. It may be taken as certain, however, thatconjugated double bonds are present in ergosterol, dehydroergo-sterol, ergosterol-B,, -D, and -F, dehydroergosterol, the two supra-sterols, and vitamin-D-all of which combine with maleic anhydride.23The isolation of the vitamin will no doubt lead to a still morerapid development of t'he chemistry of ergosterol and the furtherresults will be awaited with great interest.J.W. BAKER.G . M. BENNETT.21 A. Windaus, J. Gaede, J. Koser, and G.Stein, Annalen, 1930, 483, 17;22 I. M. Heilbron, R. A. Morton, and W. A. Sexton, J., 1928,48, 347; A. C.z3 A. Windam and A. Luttringhaus, Ber., 1931, 64, [B], 860; A., 840.A., 1577.Bose and W. Doran, J., 1929, 2244; A., 1928, 219, 410; 1929, 1432ORGANIC CHEMISTRY.-PART III. 147PART 111 .-HETEROCYCLIC DIVISION.Oxygen Ring Compounds.THE past year has produced another crop of important develop-ments in the chemistry of those natural products which containan oxygen atom as part of a heterocyclic system.Although previous investigations have made it clear that scopo-letin, which has been isolated from several different natural sources,has the structure (I), its synthesis from 2 : 4-dihydroxyanisole isnot without interest. The latter compound was transformed firstinto 2 : 4-dihydroxy-5-methoxybenzaldehyde (11) by condensationwit'h hydrogen cyanide in ether in Dhe presence of zinc cyanide andhydrogen chloride, and then into O-acetylscopoletin by heating with(1.) (11.) (111.)sodium acetate and acetic anhydride.Hydrolysis led to scopoletin,identical with the naturally occurring substance. A similar seriesof reactions starting from 2 : 5-dihydroxyanisole led to 6-hydroxy-7-methoxycoumarin (aesculetin monomethyl ether) .2 The latterproved to be identical with the product derived from the hydrolysisof aesculin methyl ether, so that the position assigned to the glucoseresidue in axculin (111) is thus confirmed.The problem of determining the exact positions occupied by thehydroxyl or methoxyl groups in natural products belonging t o theflavone and related types is frequently one of great difficulty,largely because the synthetical methods employed are not alwaysfree from ambiguity.Considerable doubt has recently existedconcerning the validity of the formulae previously assigned tobaicalein and wogonin, two flavones which have been isolated fromthe roots of Scutelluria bairnlensis. The view that baicalein is5 : 6 : 7-trihydroxyflavone (IV) rested upon the fact that it isapparently identical with the compound obtained by G. Bargellini 51 F. S. H. Head and A. Robertson, J., 1931, 1241 ; R. Seka and P. Kallir,Ber., 1931, 64, [I?], 909; A., 737.2 R. Seka and P. Kallir, ibid., p. 622; A., 606.a F . S. H. Head and A.Robertson, J., 1930, 2434; Ann. Reports, 1930, 27,4 K. Shibata, S. Iwata, and M. Nakamura, Acta Phytochim., 1923, 1, 105;178.L4., 1923, i, 591.Gazzetta, 1919, 49, ii, 47; it., 1919, i, 545148 PLANT :by the action of hydriodic acid on 2 : 3 : 4 : 6-tetramethoxy-w-benzoylacetophenone. The 'constitution accorded to Bargellini's0 RO 0 Me0(W. ) (V.) (VI. 1product, for which there were at first two alternatives, depended,however, upon the fact that it differed from a hydroxychrysin whichhad been regarded as 5 : 7 : 8-trihydroxyflavone (V; R = H).Wogonin (a dihydroxymethoxyflavone) has been demethylated toa trihydroxyflavone which is not identical with baicalein, and itsdimethyl ether was found to be different from baicalein trimethylether.' Furthermore, by heating 2 : 3 : 4 : 6-tetramethoxyaceto-phenone with aluminium chloride and treating the product-supposed to be 2-hydroxy-3 : 4 : 6-trimethoxyacetophenone (V1)-with benzoic anhydride and sodium benzoate, wogonin dimethylether was obtained.Wogonin has been regarded, therefore, as5 : 7-dihydroxy-8-methoxyflavone (V; R = Me). It has becomeapparent, however, that hydroxychrysin cannot be 5 : 7 : 8-tri-hydroxyflavone, and, furthermore, the evidence for the structuresassigned to baicalein and wogonin became less sound when it wasreported a few months ago * that the substance (VI) was in reality6-hydroxy-2 : 3 : 4-trimethoxyacetophenone. More recently, how-ever, it has transpired that the latter observation was unsound, andfurther work9 has now confirmed the nature of (VI) and estab-lished definitely the structures of the natural products.Antiarol(3 : 4 : 5-trimethoxyphenol) was transformed into its ethyl etherand then converted into 2-hydroxy-3 : 4-dimethoxy-6-ethoxy-acetophenone (VII) by the action of acetyl chloride and aluminiumchloride. The position of the hydroxyl group was confirmed bythe coloration which is developed with ferric chloride. After treat-ment with benzoic anhydride and sodium benzoate, followed byde-alkylation, 5 : 7 : 8-trihydroxyflavone was obtained. It provedto be identical with the demethylation product of wogonin, and itnecessarily follows that baicalein is correctly represented by theformula (IV). A re-investigation of Bargellini's synthesis of thelatter substance has indicated, however, that it is by no means as6 M.Nierenstein, Ber., 1912, 45, [B], 499.* K. Shibata and S. Hattori, J . Pharm. Xoc. Japan, 1931, 51, 15; Chem.0 S. Hattori, Acta Phytochim., 1931, 5, 219.S. Hattori, Acta Phytochim., 1930, 5, 99; A., 493.Zentr., 1931, i, 3358ORGANIC CHEMISTRY.-PUT fII. 149satisfactory as it appeared to be, for it has been reported thatt,he reaction leads more frequently to the isomeric 5 : 7 : 8-trihydr-oxyflavone. Similarly the action of hydriodic acid on 2 : 3 : 4 : 6-Me0 0EtOt etramet hoxy - a-anisoylacet ophenone, which Bargellini l1 has em -ployed for the synthesis of scutellarein (VIII; R = H), is statedto yield mainly 5 : 7 : 8 : 4'-tetrahydroxyflavone, the structure ofwhich was confirmed by the anisoylation and subsequent demethyl-ation of (VI).Further conhmation for the structure assigned tobaicalein is derived from the fact that its trimethyl ether yieldsantiarol, in addition to benzoic acid and acetophenone, when boiledwith aqueous potassium hydroxide for several hours.It has recently been reported l2 that the anisoylation of 2 : 4-di-hydroxy-3 : 6-dimethoxyacet ophenone (IX) proceeds normally andthat the product obtained on subsequent demethylation is5 : 7 : 8 : 4'-tetrahydroxyflavone (X). These observations are in-compatible with earlier statements 13 to the effect that the reactiontakes an unexpected course resulting in partial demethylation andthe formation of (VIII; R = Me), and the ultimate synthesis, onMe0 HO 0further demethylation, of scutellarein.Portunately the constitu-tion assigned to scutellarein is not involved in this ambiguity, sincealternative syntheses have been developed.The structure allotted to citronetin, a compound obtained bythe hydrolysis of the glucoside citronin, has been confirmed14 bythe use of a reaction which has previously been employed forsynthesising fla~an0nes.l~ The action of o-methoxycinnamoylchloride on phloroglucinol in nitrobenzene in the presence ofaluminium chloride yielded, among other substances, 5 : 7-di-lo S. Hattori, loc. cit.l1 Gazzetta, 1915, 45, 69; A., 1915, i, 84.l2 S. Furukawa and H. Tamaki, Chem. Zentr., 1931, ii, 2161.l3 F. Wessely and G. H. Moser, Monatsh., 1930,56, 97; Ann.Reports, 1930,14 3. Shinoda and S. Sato, Chem. Zentr., 1931, ii, 2326.l5 See Ann. Reports, 1928, 25, 169.27, 174150 PLANT :hydroxy-2’-metlhoxyflavanone (XI), which proved to be identicalwith the natural product.Further interesting observations have been made regarding theconversion of quercetin (XII) into cyanidin (XIII). Thus, whilethe action of sodium amalgam in alkaline solution on quercetinpent amethyl ether leads readily to the pseudo-base of pentamethyl-H O O A V H a Me0 €IO&-&Hcyanidin, quercetin itself gives a negligible quantity of cyanidin.That the failure of the process under these conditions is due to thepresence of the free hydroxyl group in the 3-position is madeprobable by the fact that a good yield of cyanidin chloride is obtainedwhen similar treatment is applied to rutin (a rhamnoglucoside ofquercetin with the sugar residue in the 3-position) and the reactionmixture is heated with hydrochloric acid.16 It has also beenreported l7 that a small yield of pentamethylcyanidin chloride canbe obtained by the reduction of quercetin pentamethyl ether withtitanium trichloride in methyl-alcoholic ammonia.Of considerable interest is the isolation of several isoflavoneglucosides from the soya bean.ls One of these, which has beencalled “ genistin,” yielded, on hydrolysis, glucose and the previouslyknown genistein (XIV; R = OH).The constitution of theglucose-free product was indicated not only by the properties ofits simple derivatives, but also by alkali fission into formic acid,phloroglucinol, and p-hydroxyphenylacetic acid.An examinationof the product derived from the hydrolysis of the methylatedglucoside showed that it was 7-hydroxy-5 : 4’-dimethoxyisoflavone,from which it follows that the glucose residue in genistin is in the7-position. A second glucoside, daidzin, has given glucose and a@H2 C*OHHO co (XI.) HO CO (XII.)c1r”? 0 (XIV.)HomHQ*Ho@doH H (XIII.) R COhitherto unknown isoflavone (XIV; R = H) t o which the name“ daidzein ” has been given. Alkaline hydrolysis of the latterl6 Y . Asahina and M. Inubuse, Ber., 1931,64, [B], 1256; A., 940.1’ P. Karrer, Y . Yen, and I. Reichstein, Helu. Chim. Acta, 1930, 13, 1308;l8 E. Walz, Annalen, 1931, 489, 118.A., 233ORGANICY CHEMISTRY .-PART IU.151yielded formic acid and 2 : 4-dihydroxyphenyl p-hydroxybenzylketone, the identity of which was established by its synthesis fromresorcinol and p-hydroxyphenylacetonitrile by means of a Hoeschreaction. The glucose residue is again attached in the 7-position,since the methylated glucoside, on hydrolysis in two stages, gave2 : 4-dihydroxyphenyl p-methoxybenzyl ketone. At least one otherisoflavone glucoside has been isolated from the same source, but theexact positions of the substituents in the glucose-free material havenot yet been determined.The earlier applications of the anthocyanidin syntheses of Prattand Robinson to substances containing the glucose residue, culmin-ating in the synthesis of cczllistephin chloride,1g foreshadowed thepreparation of other flower pigments, and further extensive work inthis field has now been described.The acetylated p-glucoside (XV),from the action of dry silver carbonate on O-tetra-acetyl-or-glucosidylbromide and w-hydroxy-3 : 4-diacetoxyacetophenone (XVI) inchloroform solution, has been treated in chloroform-ether withO-benzoylphloroglucinaldehyde (XVII) and dry hydrogen chloride.OAcC 0 C ) O A c I ~ ~ ~ ( ) C 0 * C H 2 * 0 HCH,*O*C,H,O( OAc),(XV. 1 (XVI.) (XVII. )The resulting flavylium salt, on hydrolysis with aqueous sodiumhydroxide and subsequent acidification with hydrochloric acid,yielded 3-@-glucosidylcyanidin chloride 2O (XVIII ; R = H), whichproved to be identical with a specimen of chrysanthemin chloridefrom the flowers of the deep-red chrysanthemum.By a similarseries of reactions O-benzoylphloroglucinaldehyde and the tetra-acetylglucoside of the 6)-hydroxy-ketone (XIX) ultimately gave3-~-glucosidylmalvidin chloride 21 (XX), which was shown t o bec10 OR r”lidentical with cenin chloride, an anthocyaninOMec o e I OMe I (XIX.)CH2*OHfrom the skins ofpurple-black grapes. Furthermore, 3- p-glucosidylpeonidin chlor-ide 22 (XVIII; R = Me), from the tetra-acetylglucoside of o-hydr-See Ann. Reports, 1928, 25, 166.2o S. Murakami, A. Robertson, and R. Robinson, J., 1931, 2665.21 L. F. Levy, T. Posternack, and R. Robinson, ibid., p. 2701.2z L. F. Levy and R. Robinson, ibid., p. 2715152 PLANT :oxy-4-acetoxy-3-methoxyacetophenone and O-benzoylphloroglucin-aldehyde, was found to be identical with oxycoccicyanin chloride,a pigment of American ~ranberries.~~ The galactoside obtainedby the action of dry silver carbonate on w-hydroxy-3 : 4-diacetoxy-acetophenone and O-tetra-acetyl-a-galactosidyl bromide has beenconvert,ed by a similar process into 3-p-galactosidylcyanidin chlor-ide 24 (XVIII; R = H; C6Hl1O, = galactose residue), which isapparently identical with natural idaein chloride.These synthetical methods have been further extended to thepreparation of several flavylium salts with a sugar residue in ot’herc10 OMer”?than the 3-po~ition.~~ For example, 7- p-glucosidylpelargonidinchloride (XXI) was obtained from O) : 4-dihydroxyacetophenonec10r”?and the tetra-acetyl- p-glucoside (XXII), while a related seriesof reactions gave ultimately 5- @-glucosidy Ipelargonidin chloride(XXIII), which was found t o be identical with pelargoneninchloride derived from the partial hydrolysis of pelargonin chloride.c1r”7H O & E O H(XXIII.)()EO(AcO),C,H,O*O(=I*) OBZ c6H1105*0The definite characterisation of many of these glucosides hasrevealed several very interesting generalisations. For instance,the presence of a free hydroxyl group in the 3-position renders the23 (Miss) K. E. Grove and R. Robinson, Biochem. J., 1931, 25, 1706.24 (Miss) K. E. Grove and R. Robinson, J., 1931, 2722.25 A. Leh, A. Robertson, R. Robinson, and T. R. Seshadri, ibid., p. 2672;L. F. Levy and R. Robinson, ibid., p. 2738ORGANIC CHEMISTRY.-PART III.153molecule very susceptible to attack by oxidising agents,26 so thatsuch compounds rapidly decolourise dilute solutions of ferric chloride.Thus, for example, the yellow-brown colour of a very dilute ferricchloride solution was comparatively rapidly discharged by pelar-gonidin chloride and its 4'-, 5-, and 7-p-glucosides, but was unchangedby callistephin chloride (3-p-glucosidylpelargonidin chloride), 3- p-glucosidylpeonidin chloride (XVIII ; R = Me), and luteolinidinchloride 27 (XXIV) after a much greater period of time. Analogousimportant generalisations can be drawn from a study of the dis-tribution of these substances between amyl alcohol and dilutehydrochloric acid, and also from a careful examination of theircolour reactions in a range of buffered solutions of graded pH.2*With the aid of these characteristic features it is possible to makefar-reaching deductions concerning the structures of other naturallyoccurring anthocyanins.For instance, cyanin, pelargonin, peonin,and malvin have hitherto been considered t'o be Ei-biosides, but,since there are now clear indications that there are sugar residuesin both the 3- and the 5-position, these anthocyanins must be regardedas 3 : 5-diglucosides. On the other hand, mecocyanin, prunicyanin,and keracyanin are biosides with the sugar group in the 3-position.29Furthermore, the colour reactions, etc., of the anthocyanins andanthocyanidins have been made the basis of methods for examiningand detecting the nature of the anthocyanins present in numerousc10 OH h 0flPh (XXV.)H (XXIV.) HO COflowers.30 A very interesting outcome of the latter investigationis the observation that the colours of these salts can be modifiedto a remarkable degree by the presence of other substances (co-pigments) which apparently form with them weak additive com-plexes.The actual colours of the flowers themselves may dependto a considerable extent upon this factor.Other investigations among natural products belonging to theoxygen ring types have included studies into the position of thesugar group in a few miscellaneous glycosides. Baicalin, from the26 Compare P. Karrer, R. Widmer, A. Helfenstein, W. Hiirliman, 0. Niever-gelt, and P. Monsarrat-Thomas, Helv. Chim.Acta, 1927, 10, 729.27 A. Le6n and R. Robinson, J., 1931,2732.2 8 Compare A. Robertson and R. Robinson, Biochem. J., 1929, 23, 35.29 (Mrs.) G. M. Robinson and R. Robinson, Nature, 1931,128, 413.30 (Mrs.) G. M. Robinson and R. Robinson, Biochem. J., 1931,25, 1687154 PLANT :roots of Xcutellaria baicalensis, yields glycuronic acid and baicalein(5 : 6 : 7-trihydroxyflavone) on hydrolysis, and it has now beenshown31 that the glycuronic acid residue is in the 7-position(formula XXV) from a comparison of the reactions of baicalinwith those of primetin (5 : 6-dihydroxyflavone). Thus it gives amonomethyl ether which develops a violet-brown colour with ferricchloride, indicative of a free hydroxyl group in the 5-position.Furthermore, baicalin itself , like primetin, gives a green colour withferric chloride, and is very readily oxidised.It has been con-cluded 32 that the product obtained from myricitrin, a rhamnosideof myricetin (5 : 7 : 3’ : 4’ : 5’-pentahydroxyflavonol), by methyl-ation and subsequent hydrolysis is myricetin 5 : 7 : 3’ : 4’ : 6’-pentamethyl ether (XXVI), from which it follows that the sugarresidue, as in several other flavonol glycosides, is in the 3-position.The bioside hesperidin, which yields rhamnose, glucose, and hes-0 OMe 0Me0 CO (XXVI.) HO CO (XXVII.)(XXVIII.) Ho()g%H : C H e M eMe OMeperitin (XXVII ; R = H) on hydrolysis, has been shown 33 to havethe biose residue in the 7-position (XXVII; R = C,,H2,09). Onmethylation and subsequent hydrolysis it gave 4-hydroxy-2 : 6-di-methoxyphenyl3 : 4-dimethoxystyryl ketone (XXVIII) (the flavan-one structure readily isomerises to the chalkone), which has beensynthesised by the interaction of veratraldehyde and 4-hydroxy-2 : 6-dimet hoxyacetophenone.Many int,eresting and important investigations of the methodsused for the synthesis of chromones and coumarins have recentlybeen described.The extremely valuable reaction whereby theacylation of certain polyhydroxyacetophenone derivabives leads tothe production of several chromone types3* cannot be used witho-hydroxyacetophenone itself. It has now been found,35 however,that the process can be applied to 2-acetyl-l-naphthol. 1 : 4-a-Naphthapyrones (I) result from the action of a mixture of the31 K.Shibata and S. Hattori, Acta Phytochim., 1930, 5, 117; A . , 493.32 S. Hattori and K. Hayashi, ibid., 1931, 5, 213.33 F. E. King and A. Robertson, J., 1931, 1704.34 See Ann. Reports, 1929, 26, 152.35 A. S. Bhullar and K. Venkataraman, J., 1931, 1165ORGANIC CHEMISTRY.-PART III. 155anhydride and alkali salt of the appropriate acid upon this com-pound, but the simultaneous formation of the 3-acyl derivatives (11) eo O/\ &yo O/\COR &,"(1.) (11.) (111.)is a marked feature of the process, whereas analogous carbon-acylation is very rare in the aforementioned benzene series. Thereaction can also be applied to l-acetyl-2-naphthol with the form-ation of 1 : 4-pa-naphthapyrones 36 (111) ; simultaneous carbon-acylation again occurs when acetic anhydride is used.Furtherexamples of this reaction involving the use of simpler hydroxy-ketones have been described.37 By acylation of 2-hydroxy-4 : 6-dimet hoxy propiophenone and 2 - hydroxy-4 : 6-dimethoxy-n-butyro-phenone, for example, chromones of the type (IV) have beenprepared. Acetylation of 2-hydroxy-4 : 6-dimethoxyacetophenoneunder vigorous conditions gave a product, believed to be (IV;R = Me; R' = Ac), in which acetylation in the 3-position hasapparently taken place. A further extension38 of the reaction tothe acylation of 4 : 6-diacetylresorcinol (V) has yielded the di-chromones (VI) and the chromones (VII).R R 7?The condensation of phenols with p-ketonic esters has beenextensively employed for the synthesis of both coumarins andchromones.It has previously been considered that, in general, theformer result when sulphuric acid is used as the condensing agent,and the latter in the presence of phosphoric oxide. More recentstudies 37, 39 of the reaction have shown, however, that while -the36 B. K. Menon and K. Venkataraman, J., 1931, 2591.37 F. W. Canter, F. H. Curd, and A. Robertson, ibid., pp. 1245, 1255;F. W. Canter, A. R. Martin, and A. Robertson, ibid., p. 1877.38 K. C. Gulati and K. Venkataraman, ibid., p. 2376.39 D. Chakravarti, J. Indian Chem. SOC., 1931, 8, 129, 407; A., 962, 1304;A. Robertson, W. F. Sandrock, and Miss C. B. Hendry, J., 1931,2426156 PLANT :use of sulphuric acid always leads to the formation of coumarins,when this is replaced by phosphoric oxide, the nature of the productdepends upon the particular phenol which is employed.Thus,although simple monohydric phenols and quinol yield chromones,resorcinol, orcinol, pyrogallol, phloroglucinol, phlorogl ucinol di -methyl ether, and a-naphthol give coumarins. The course of thereaction is apparently independent of the nature of the 8-ketonicester.An investigation 40 of the condensation of a-formylphenylaceto-nitrile, CHO.CHPh.CN, and benzoyloxymethylenephenylaceto-nitrile, CH(O*COPh):CPh*CN, with polyhydric phenols in dry etherin the presence of zinc chloride and hydrogen chloride has made itclear that the reaction leads, in general, t o the correspondingderivatives of 3-phenylcoumarin and not t o the isomeric isoflavones(3-phenylchromones). For instance, the use of resorcinol resultsin.the formation of 7-hydroxy-3-phenylcoumarin (VIII). Thecondensation, which involves the interaction of the cyano-groupand a phenolic hydroxyl group, is interesting, therefore, because itdoes not follow the normal course of the Hoesch reaction.It is possible in the space available to mention only a very fewof the more outstanding of the remaining points in the year's out-put of work on oxygen ring compounds. Coumarins substituted inthe 3-position can readily be converted by treatment with phenyl-magnesium bromide in dilute solution a t room temperature intothe corresponding 2-phenylbenzopyrylium salts (e.g., X from IX),but, when the reaction is carried out in hot concentrated solution,the 2 : 4-diphenyl-A2-chromens (e.g., XI) are formed.41 The formerc1(IX.) (X.> (XI.)reaction, however, gives only poor yields of the benzopyryliumsalts when applied to coumarins substituted in the 4-position, butthe more vigorous conditions lead readily t,o the production of2 : 2-diphenyl-A3-chromens (XII).Of some interest is the observation 42 that coumarone and diphenyl-40 I.0. Badhwar, W. Baker, B. K. Menon, and K. Venkataraman, J., 1931,4 1 I. M. Heilbron and D. W. Hill, J., 1927, 2005; I. M. Heilbron, D. W.42 N. A. Orlov and V. V. Tiatschenko, Ber., 1930,63, [B], 2948; A., 233.1541.Hill, and H. N. Walls, J., 1931, 1701ORGANIC CHEMISTRY.-PART IIT.. 157ene oxide are formed in good yields when the vapours of coumarinand xanthone respectively are submitted at atmospheric pressureto temperatures of 860-880".Open-chain ketones do not losecarbon monoxide to give simple products under similar conditions.The recent preparation of derivatives of acetylene oxide, inwhich there is an unsaturated 3-membered oxygen ring, is a note-worthy feature in this field. These substances resulted in pooryields by reactions analogous to those which lead to the ethyleneoxides. The action of powdered potassium hydroxide on desylPhC---CPh(CH,,,<"H2>CH2(XII.) (XIII.) (XIV.)chloride, CHPhClCOPh, in toluene has given the diphenyl deriv-ative (XIII), while the monophenyl compound has been obtainedfrom phenacyl bromide by the use of silver oxide.43 The productswere found to be comparatively stable, but gave desyl chloride andphenacyl chloride respectively when heated for two hours withconcentrated hydrochloric acid at 120".The preparation of cyclotellurobutane (tetrahydrotellurophen)(XIV; n = 2) by a process essentially similar to that previouslyused for the synthesis of cyclotelluropentane 44 (XIV; n = 3) hasbeen de~cribed.,~ The action of aluminium telluride on &tetra-[ CH,],Br C,H 8:Te*[CH,] ,*Te:C,H s(xv.) c,H,:T/ I I (XVI.)Br \Br Brmethylene dibromide led to a mixture of the telluronium salts (XV)and (XVI), which dissociated, on heating, to give tetramethylenedibromide and cyclotellurobutane.A more convenient procedurewas found by dissolving amorphous tellurium in as-tetramethylenedi-iodide at 130".The cyclotelluributane 1 : 1 &-iodide, C4Hs:Te12,which was formed was reduced by passing sulphur dioxide throughits suspension. in boiling water. cycZoTellurobutane exhibited theadditive properties characteristic of the saturated cyclic telluro-and seleno-hydrocarbons.5-Membered Nitrogen Ring Compounds.A noteworthy feature of the chemistry of pyrrole derivatives isthe recent resolution of 1 - o - carb oxy p heny 1 - 2 : 5 - dimet hy lp y rr ole -43 W. Madelung and M. E. Oberwegner, AnnaZen, 1931,480,201.44 See Ann. Reports, 1928, 25, 196.4 5 G. T. Morgan and F. H. Burstall, J., 1931, 180158 PLANT :3-carboxylic acid 46 (I) and 2 : 5 : 2’ : 5’-tetramethyl-1 : 1‘-dipyrryl-3 : S’-dicarboxylic acid 47 (11) into optically active components.ItH0,C H 0 ,C*C=CMeHCZCMe(1.1 (11.)is probable that the activity of these compounds is due to restrictedrotation similar to that observed in the diphenyl series, although itis just possible that its origin may lie in the asymmetric nitrogenatoms.The action of aliphatic diazo-compounds on pyrroles may pursuevarious courses.48 The use of diazoacetic ester has been found toresult in the introduction of the *CH,*CO,Et group into the nucleusat an a-carbon atom, if this is available, but diazo-ketones anddiazo-ketonic esters react differently. For instance, benzoyl-phenyldiazomet hane and 2 : 4-dimethylpyrrole yielded the product(111), which has apparently resulted from the intermediate formationof diphenylketen. In fact, the latter substance has been shown toadd directly to 2 : 4-‘dimethylpyrrole with the formation of (111).MeC C*CO*CHPh, MeC C*CO*CHMe*CO,Et,Hs-EMe Hg-gMev NH (IV.)v NH (111.)Similarly, diazoacetoacetic ester and 2 : 4-dimethylpyrrole reactedtogether to give the compound (IV).Ketens and isocyanic estersare stated t o combine additively with pyrroles with the formationof ketones and amides respectively.Very interesting results have also been obtained by 0. Diels,K. Alder, and their collaborators in an extension of their study ofadditive processes to certain heterocyclic systems. Furan behavedCOae(VII.)in a manner similar to that observed in unsaturated carbocyclicsystems, and with maleic anhydride yielded the product *9 (V).46 L. H. Bock and R.Adams, J. Amer. Chem. Soc., 1931, 53, 374; A., 362.4 7 C. Chang and R. Adams, ibid., p. 2353; A., 1074.4 8 C. D. Nenitzescu and E. Solomonica, Ber., 1931, 64, [B], 1924; A., 1305.49 0. Diels, K. Alder, andE. Naujoks,BeT., 1929,02, [B], 654; A., 1929, 570ORGANIC CHEMISTRY.-PART III. 159With dimethyl acetylenedicarboxylate it gave a similar substance(VI) besides a compound (VII) derived from the further additionof a second molecule of f~ran.~O In the a-pyrone series similaradditive reactions have been realised. Thus methyl coumalategave with maleic anhydride in boiling toluene the product (VIII),but in boiling xylene the reaction proved to be rather more com-plex and led to (IX).51 In the pyrrole series, however, the additive4 H L ,MeO2C.G HTCO>O CH*COyMeO,C*f(o TH \FHCO>OHC $0 CH*CO HC HYCO CHCO‘CH’ \ C H Y(VIII.) PX-1processes are of a different character and do not result in the form-ation of polycyclic derivatives. The main reactions are illustratedby the addition of maleic acid to 2-methylpyrrole with the formationof (X), or of acetylenedicarboxylic acid to 1-methylpyrrole with theproduction of (XI) .52 Certain indole derivatives closely resemblethe pyrroles in this respect, but the reactions may also be moreHc-GHHcj/c.C( C02H) :CH*CO,HHFl -P MeCQ*CH (C0,H) *CH,*C02HNH (X.) (XI.)Me C0,H MeC ‘co NHMe CO CONH NH(=I*) (XIII.)complex. For example, 2-methylindole and maleic anhydridegave not only (XII) but a, further reaction product (XIII).53 Inthe case of indole and 3-methylindole the course of the reaction wascomplicated by the primary polymerisation of the indole.The tendency for the 2 : 3-linkage in the indole skeleton to com-bine additively with OH and NO,, or 20H, which is such a markedfeature of the action of nitric acid on the N-acyl derivatives oftetrahydrocarbazole (XIV ; n = 2) and dihydropentindole (XIV ;n = l), has been found to be enormously diminished when thisso 0.Diels, K. Alder, H. Nianburg, and (Frl.) 0. Schmalbeck, Annalen,1931, 490, 243.51 0. Diels, K. Alder, and K. Niiller, ibid., p. 257.s2 0. Diels, K. Alder, and D. Winter, ibid., 1931, 486, 211; A., 849;0. Diels, K. Alder, and €3. Winckler, ibk?., 1931, 490, 267.53 0. Diels, K. Alder, and W. Lubbert, ibid., p.277160 PLANT :system is contained in more complex polycyclic types.54 Thus,of many compounds investigated, only one, viz. , 7-benzoylbenzo-pentindole (XV), was found to give an addition product (XVI) ofthe kind mentioned.(XW.) (XV.) (XVI.)The synthesis of certain compounds in which a nitrogen atom iscommon to two fused ring systems is of considerable interest inview of the probable occurrence of similar structures in somealkaloids, e.g., lupinine. When the addition product obtained from2-methylpyridine and bromoacetone was distilled in steam in thepresence of sodium hydrogen carbonate, 2-methylpyrrocoline (XVII)was obtained,55 and the reaction has been applied to other analogouscases. Pyrrocolines have also been reported from the action ofacid anhydrides on Z-met!hylpyridine, e.g., l-propionyl-3-methyl-f) =====QH f)======y*CO*CH,Me HR-9H HvH N*[CH,],;CN\\,N\//CM" N A P H CH CMe(XVII.) (XVIII.) (XIX.)pyrrocoline (XVIII) when propionic anhydride was used.56 Inaddition, potassium pyrrole has been treated 57 with P-cyanoethyltoluene-p-sulphonate, and the 1-p-cyanoethylpyrrole (XIX ; n = 2)so obtained made to undergo a Hoesch reaction with the formationof 3-keto-4 : 5-dihydrodi-(l : 2)-pyrrole (XX).Similarly, y-l-pyrryl-butyronitrile (XIX; n = 3), from the interaction of potassiumpyrrole and y-bromobutyronitrile, yielded 8-keto-5 : 6 : 7 : %tetra-hydropyrrocoline (XXI).coHY==(i-(iO H 2 ~ ~ = 7 H Hg-(iHMeH C v C ONMe(XXII.)/ HHC\/NyCH, CH H, H2cx, bii?(XX.1 (XXI.)54 S.H. Oakeshott and S. G. P. Plant, J., 1928, 1840; S. A. Bryant and5 5 A. E. Tschitschibabin, Ber., 1927,60, [B], 1607; A., 1927,885.6 6 A. E. Tschitschibabin and F. N. Stepanov, Ber., 1930, 63, [BJ, 470 ; A .5 7 G. R. Clemo and G. R. Ramage, J . , 1931, 49.S. G. P. Plant, J., 1931, 93.1930, 619ORGANIC CHEMISTRY .-PART III. 161Turning to natural products of the indole type, it is of interest torecord that an examination 5* of the basic substances which can beextracted from toad secretions goes to show that bufotenine, forwhich the structure (XXII) has previously been suggested,59 is anindole derivative of the formula (XXIII), while bufotenidine is therelated betaine (XXIV). When calycanthine, an alkaloid from-G*CH2*CH(NMe,)*C02H G-CH2-yH-70/CH Me3N-0NMe(XXIV.)o v , H NMe OT(XXIII.)CaZycanthus glaucus, was first benzoylated, and then either oxidisedwith potassium permanganate or treated with alkali, it yieldedbenzoyl-N-methyltryptamine, the constitution of which was estab-lished by its preparation from tryptamine (3-p-aminoethylindole)by the action of an excess of inethyl iodide and subsequent benzoyl-ation.On the basis of these results 6o it has been suggested thatthe alkaloid is a p-carboline derivative containing the group (XXV).It may also be mentioned that tryptophol (XXVI), derived fromthe action of yeast on tryptophan, has recently been synthesisedby the reduction of ethyl indolyl-3-acetate with sodium and alcohol.Work dealing with glyoxaline derivatives and published duringthe past two years includes the synthesis 62 of I-2-thiolhistidine (IV).This is of special interest in view of the close relationship betweenthis compound and several natural substances, and of the prob-ability that it may itself be found among the products of thehydrolysis of proteins. Natural I-histidine (I) was converted firstinto its methyl ester dihydrochloride and then, by benzoylation,into methyl ay8-tribenzamido- Av-pentenoate (11), which, withmethyl-alcoholic hydrogen chloride, gave methyl a6-dibenzamido-y-ketovalerate (111).The dihydrochloride of the correspondinga6-diamino-y-ketovaleric acid, obtained by boiling with 20%s8 H. Wieland, G. Hesse, and H. Mittasch, BeT., 1931, 64, [B], 2099; A .,1310.6B H. Handovsky, A., 1920, i, 495.6o R. H. F. Manske, Canadian J . Res., 1931, 4, 275; A., 855..st R. W. Jackson, J. Biol. Chem., 1930, 88, 659; A., 97.J. N. Ashley and C. R. Harington, J., 1930, 2586.REP.-VOL. XXVIII. 162 PLANT :hydrochloric acid, yielded, on treatment with one molecular equi-valent of sodium thiocyanate, the monohydrochloride of Z-Z-thiol-H02C*CH(NH2)*CH2*$-NH>CH MeO2C.QH--CH2*~=CH*NHBzHC-N NHBz NHBz(1.) (11.1Me02C*~H*CH2*CO*CH2*NHBz H02CoCH(NH2)~CH,o~-NH~c.~HNHBz HC-N(111.) (IV.)histidine (IV).interact ion of cc8-diamino- p - ket obut ane dihydr oc hloride,HCl,NH2*CH2*CO*CH2*CH2*NH2,HC1,and sodium thiocyanate is closely related to the above synthesis.It is noteworthy that this substance does not show to any appreciableextent the physiological properties of histamine.The preparation of 2-thiolhistamine 63 (V) by theAmong the naturally occurring glyoxalines, anserine, recentlyisolated from bird muscle-ti~sue,~~ has been shown to have theconstitution (VI).65 Thus it yielded, on hydrolysis with baryta,p-aminopropionic acid and an N-methyl derivative of dl-histidine.The relative positions of the side-chain and methyl group wereindicated by the fact that anserine gave 1 : 5-dimethylglyoxalinewhen distilled with soda-lime.Anserine differs from the closelyrelated carnosine only in having the methyl group attached to thenitrogen atom of the glyoxaline ring. When the hydrolysis iseffected by the use of 20% sulphuric acid for 6-8 hours, theseproducts yield E-methylhistidine and Z-histidine respectively.66Of interest is the preparation of 1 : Z-diphenylglyoxalino-4 : 5(3' : 2')-quinoline 67 (VII). This was accomplished by firstheating the azlactone (VIII), from the condensation of o-nitro-benzaldehyde with hippuric acid, with aniline in the presence ofcopper powder to obtain o-nitrobenzylidenebenzamidoacetanilide63 F. L. Pyman, J., 1930, 98.64 D. Ackermann, 0. Timpe, and K. Poller, 2. physiol. Chem., 1929, 183,1 ; A . , 1929, 944.e 5 W. Linneweh, A. W. Keil, and F. A. Hoppe-Seyler, ibid., p. 11 ; A.,1929, 944; F. L. Pyman, J., 1930, 183; W. Keil, 2. physiol. Chem., 1930,187, 1 ; A . , 1930, 617.G 6 W. Linneweh and F. Linneweh, 2. physiol. Chcm., 1930, 189, 80; A.,1930, 1049.6' K.S. Narang and J. N. RBy, J., 1931, 976ORGANIC CHEMISTRY.-PUT III. 163(TX). The latter gave 5-keto-1 : 2-diphenyl-4-o-nitrobenzylidene-4 : 5-dihydroglyoxaline (X) on treatment with phosphoryl chloride,N(VII.) (VIII.) (IX.)and the formation of the glyoxalinoquinoline (VII) was theneffected by reduction with zinc dust and acetic acid. A relatedNPoq--r. (XI.) 4 c o NH3duct (XI) has been obtained by the condensation of hydantoiwith o-nitrobenzaldehyde, and the reduction of the 5-o-nitrobenzyl-idenehydantoin thus formed with hydriodic acid and phosphorus.6sCyanine Dyes.-Of very considerable interest is the recentinvestigation 6g of the structure of indolenine-yellow, a dye obtainedby the action of nitrous acid on 2-methylene-1 : 3 : 3-trimethyl-indoline (I) in acetic anh~dride.~O It was originally thought thatthe substance possessed the formula (11), but it has now beenR Xshown that the reaction results in the elimination of hydrogencyanide and that the constitution of the dye is, in reality, that ofCMe, Me,Can indocyanine (111).The reaction is, therefore, very closelyrelated t o one already described 71f~r the preparation of thiocyanines.6 8 J. Kozak and L. Musiak, Bull. Awd. Polonaiae, 1930, [A], 432; A., 366.as R. Kuhn, A. Winterstein, and G. Balser, Ber., 1930, 63, [B], 3176; A.,70 D.R.-P. 459,616; Chem. Zen.fr., 1928, i, 3119.71 Miss N. I. Fisher and Miss F. M. Hamer, J., 1930, 2502 ; see Ann. Reports,238.1930, 27, 187164 PLANT :When the methylene base (I) was treated in glacial acetic acid withsodium nitrite and then with perchloric acid, the perchlorate (IV)was isolated, and this was subsequently converted into indolenine-yellow (111) by heating with a further quantity of the base (I) inacetic anhydride.The latter stage involved the loss of water andhydrogen cyanide.Another surprising new met hod for the production of thiocyanines(V) has been discovered; 72 it requires the interaction of methyl-magnesium iodide and an N-alkylbenzthiazolone (VI) in benzenesolution, and apparently involves the evolution of methane by amechanism which is not at present obvious.S S SR IThe scope of the +-cyanine reaction (the action of alkali on amixture of the alkyliodides of 2-iodoquinoline and quinaldine) ,which has already been applied to the preparation of numerousFrom 2-iodopyridinealkyliodide and quinaldine alkyliodide, the 4-cyanines (VII) wereobtained, while the analogous dyes (VIII) resulted from a similarreaction with 2 -iodo- p-napht haquinoline alkyliodide and p -napht ha-quinaldine alkyliodide.The +-cyanines (IX) were prepared eitherfrom 2-iodoquinoline alkyliodide and p-naphthaquinaldine alkyl-has now been further extended.74R' X (VIII.)iodide or from quinaldine alkyliodide and 2 -iodo- p-napht haquinolinealkyliodide by the action of alkali. Analogous thio-$-cyanines were72 W. Kanig, W. Kleist, and J. Gotze, Ber., 1931, 64, [B], 1664; A., 1076.73 See Ann. Reports, 1928, 25, 181 ; 1930,27, 186.7' Miss F. M.Hamer and Miss M. I. Kelly, J., 1931, 777ORGANIC CHEMISTRY .-PART In. 165obtained by a similar process from 2-iodo-p-naphthaquinolinealkyliodide and 1 -met hylbenzt hiazole alkyliodide.Alkaloids.Pyridine Group-An extension of the recent work 75 on thenature of the products from Lobelia inflata has indicated the presenceof several minor alkaloids, of which one-lobinine, C18H2702N-has been the subject of a detailed The interesting con-clusion is reached that this alkaloid may have the structure (I), inwhich there is a 7-membered ring. Its simple reactions clearlydemonstrate its nature as an alcohol, a ketone, and a tertiary amine,and much of the evidence for the size of the ring system is derivedfrom a study of the degradation of the corresponding diketone,Cl8H2,O2N.When the methiodide of the last substance wasp 2 - 7 J 3 2\/QH2 p 2 ;\". COPh*CH,-CH CH*CH,-CHMe.OHQH2 ( p 2COPh*CH,*CH CH*CH,-CHPh*OH NMe(1.) v NMe (11.)treated with aqueous sodium bicarbonate, it gave dimethylamineand an unsaturated diketone, C,,H,O,, which was transformedultimately into the corresponding saturated compound, C17H240,.The latter yielded, on oxidation, benzoic acid and suberic acid,CO,H-[CH,],*CO,H, from which it is concluded that it has theformula COPh*[CH,],*COMe. It is admitted, however, that theseresults could be accounted for on the basis of the structureCOPh.[CH,],*COCH,Me for the saturated diketone, in which caseit would be possible to formulate the alkaloid as a piperidine deriv-ative similar to the better-known members of this group.Never-theless, the 7-membered ring structure is favoured on the groundsthat, unlike the other Lobelia alkaloids, lobinine gave neitherl-methylpiperidine-2 : 6-dicarboxylic acid nor 1 -methyl-2-carboxy-piperidine-6-acetic acid on oxidation. The relative positions ofthe carbonyl and hydroxyl groups are indicated by the fact thatthe alkaloid, like lobeline (11), readily yields acetophenone.Preliminary work on the nature of the bases contained in AnabasisaphgZh has been de~cribed,~' and it is noteworthy that lupinine hasbeen isolated from the lower-boiling fraction. In addition, a newalkaloid, anabasine, believed to be 3-(2'-piperidyl)pyridine (111) has7 5 See Ann. Reports, 1929, 26, 169.76 H.Wieland, M. Iahimasa, and W. Koachara, Annalen, 1931,491, 14.7 7 A. Orechov and G. Menschikov, Ber., 1931,64, [B], 266; A., 498166 PLANT :been obtained. In support of this structure it has been observedthat the compound can be benzoylated and nitrosated. Oxidationwith potassium permanganate gave nicotinic acid, and dehydro-genation resulted in the loss of six atoms of hydrogen and theformation of a product which is apparently 2 : 3’-dipyridyl (IV).The structure (111) has previously been advanced 78 for nicotimine,CH,a minor alkaloid from tobacco, but tlhe latter appears not to beidentical with anabasine.Of considerable interest is the observation 79 that a by-product,C,H,Op,, always accompanies the nicotinic acid obtained by theoxidation of nicotine with nitric acid.The reactions of this com-pound suggest that it is 4-nitro-5-(3’-pyridyl)pyrazole (V), and itsformation clearly involves very remarkable changes in the molecularstructure .isoQuinoZine Group.-A description has appeared *O of the applic-ation to the synthesis of hydrastine (I; R = H) of two routespreviously used for the preparation of the closely related a-gnosco-pine (dZ-narcotine). The condensation of hydrastinine (11) andnitromeconine (111) led, on subsequent reduction, to a mixture ofOMeR e M eCH CO/ \H2C NO,O a O M e (111.)OC OMethe two stereoisomeric, inactive aminohydrastines (I ; R = NH,).After separation of the mixtaure and removal of the amino-groupvia the hydrazino-compounds, the corresponding stereoisomeric7 8 A.Pictet, Amh. Pham., 1906, 244, 388.79 G. A. C. GoughandH. King, J., 1931, 2968.80 E. Hope, F. L. Pyman, F. G. P. Remfry, and R. Robinson, J., 1931,81 E. Hope and R. Robinson, J., 1914,105, 2086.236ORGANIC CHEMISTRY.-PART m. 167dZ-hydrastines were obtained. Similarly, the condensation ofhydrastinine with iodomeconine gave a mixture of the two stereo-isomeric iodohydrastines ( I ; R = I). The exact relationship ofthese inactive forms to the natural Z-hydrastine is not yet clear.Earlier degradative experiments have indicated the main featuresin the constitution of oxyacanthine, and have enabled two formulie,differing only in minor details, to be advanced for theFurther work 83 has now provided confirmation for these structuralviews.The product obtained from oxyacanthine methyl ether bythe Hofmann degradation process gave two well-defined substanceson oxidation with ozone. One of these proved to be 2’-methoxy-diphenyl ether-4 : 5’-dialdehyde (IV) and yielded the correspondingdicarboxylic acid (already known) on further oxidation. The otherwas a base, C,,H,,O,N,, which contained three methoxyl groupsand gave a dimethiodide. The latter salt was easily broken downby alkali to give trimethylamine and a nitrogen-free compound,C,1H,,06, which was found to be a dialdehyde containing twoethylenic linkages. These reactions are satisfactorily explainedwith the aid of formula (V) for the base C25H3406N2, but there isno evidence yet to indicate the exact locations of the methoxylgroups and the ether linkage.Until such information is forth-coming, several alternatives of the type (VI) exist as possibilitiesfor oxyacanthine. It is probable that one of these represents theclosely related alkaloid berbamine.MeN HzY&oMe A OMe Me0 o,egeMe0 (IV.) b o € € (v;dMe2N*CH2*CH, OMe Me0 CH,*CH,*NMe,(V.1 OHCOOMe o&klHOIn last year’s Report 84 itl was mentioned that the formula (VII)had been tentatively suggested 85 for chelidonine ; valuable experi-mental evidence in support of this structure is now forthcoming.86Ann. Reports, 1929, 26, 173.83 F. von Bruchhausen and P. H. Gericke, Arch. Phccm., 1931, 269, 115;84 Ann. Reports, 1930, 2’7, 195.85 F. von Bruchhausen and H. W. Bersch, Ber., 1930,63, [B], 2520.86 E.Spath and F. Kuffner, Ber., 1931,64, [B], 370; A., 500.A., 636168 PLANT :The oxidation of N-acetylanhydrochelidonine 87 with nit,ric acid hasbeen found to yield benzene-1 : 2 : 4-tricarboxylic acid, while theoxidation of chelidonine with potassium permanganate gavehydrastic acid (VIII) and 3 : 4-methylenedioxybenzene-1 : 2-dicarb-oxylic acid. Confirmation of the nature of the fundamental ringsystems in chelidonine and the closely related alkaloid sanguinarineis derived from the fact that the latter substance yields, on zincdust distillation, a base which is now identified as a-naphthaphen-anthridine (IX). The position of the alcoholic hydroxyl group isCH,indicated by the result's obtained from an application of the Hofmannand Emde degradation processes to chelidonine.This structure for chelidonine being accepted, the known factsconcerning sanguinarine and its relationship to the former alkaloidsuggest for it the formula (X).The alkaloids homochelidonine,Cz1H2,0,N, and chelerythrine, C2,H,@,N*OH, are very closelyrelated to chelidonine and sanguinarine respectively, and probablydiffer from them only in having one of the two methylenedioxy-groups replaced by two methoxyl groups. On this assumption aninvestigation has been made 88 of some of the reactions of chel-erythrine chloride in order to determine which methylenedioxy-group is involved. When oxidised with potassium permanganateunder various conditions it yielded the methylimide (XI) of hemi-pinic acid and also hydrastic acid.It was observed, furthermore,that chelerythrine chloride gave a-naphthaphenanthridine (IX) onzinc dust distillation. The indications are, therefore, that homo-87 Derived from chelidonine by acetylation as described by J. Gadamer,8 8 E. Spath and F. Kuffner, Ber., 1931, 64, [B], 1123.Arch. Phamz., 1924, 282, 265ORGANICchelidonine is representedCHEMISTRY.-PART III. 169by the formula (XII) and chelerythrineby (XIII).CH,HO-c/. g>CH, "oO>CH2' H Me0 (&e*OHM e O R c > m e Me0 (XIII.)Me0 CH, (XII.)The assumption regarding the nature of the relationship betweenchelerythrine and sanguinczrine has been shown to be valid byremoving the met hylenedioxy-groups from dihydrochelerythrineand dihydrosanguinarine by treating these products with phloro-glucinol and boiling 50 yo sulphuric acid.The hydroxy-compoundsso obtained, on subsequent methylation, both yielded the samederivative 89 (XIV) .Me0 @ EEoA??H\T/V CH CH, CHCH, CH2/ \ / \H2Y p QHMeOMe M>d CH, \/CH, H&\/(XIV.) CH, (xv.) CH,Reference has been made in a previous Report to the preparationof substances which might, in view of the known facts concerningthe structure of emetine, possess amcebicidal properties. Since it isthought that emetine can be represented by the formula (XV),glthe work has now been extendedg2 to the synthesis of the closelyrelated 10 : ll-dimethoxy-1 : 2 : 3 : 4 : 6 : 7-hexahydrobenzpyrido-(XVI.) (XVII.) (XVIII.)coline (XVI; n = 4).8-chlorovalero- p-veratrylet hylamide,This was accomplished by first convertingCH,Cl*CH,*CH,*CH,*CO*NH*CH,*CH,*C,H,( OMe),,*Q E.Spath and F. Kuffner, Ber., 1931, 64, [B], 3034.QO Arm. Reports, 1929, 26, 172.81 W. H. Brindley and F. L. Pyman, J., 1927, 1067.Q2 R. Child and F. L. Pyman, J., 1931, 36.F 170 PLANT :into 1 - 8- chlorobutyl- 6 : 7 - dimethoxy - 3 : 4 - dihydroisoquinoline(XVII) by the action of phosphorus oxychloride. The latter gavethe salt (XVIII) on warming, and the desired base (XVI; n = 4)was then obtained by reduction. A similar series of reactionsstarting from y-bromobutyro- p-veratrylethylamide led to theanalogous base (XVI; n = 3), but, when these products and someof the intermediates were examined, they were found not to possesspronounced physiological properties.The view has been expressed 93 that trilobine and homotrilobine,alkaloids from species of CoccuZus, belong to the l-benzyltetrahydro-isoquinoline group, but the evidence so far available does not dis-close the complete structures, The isolation of two additionalalkaloids from Cocculus trilobus has also been described.94 One ofthese, trilobamine, is apparently closely related to oxyacanthineand berbamine.Aporphine Group.-It has been demonstrated 95 that three well-defined alkaloids of the aporphine type can be extracted from thebark of LuureZia Novce Zealandice.One of these, pukateine (I),contains a methylenedioxy- and a phenolic hydroxyl group, andthe formula assigned to it is based upon an examination of theproducts derived from an application of the Hofmann method of MeOO\ili,degradation and from various processes of oxidation.The secondalkaloid, laureline (11), contains a methylenedioxy- and a methoxy-group, but is different from pukateine methyl ether.ative reactions support the formula (11).Its degrad-The third, laurepukine,93 H. Kondo and M. Tomita, Chem. Zentr., 1931, i, 1114.94 H. Kondo and M. Tomita, Arch. Pham., 1931, 269, 433.s5 G. Barger and A. Girardet, HeZv. Chim. Acta, 1931, 14, 481, 604; A.,749, 750ORGANIC CHEMISTRY.-PART IlI. 171which is now characterised for the fist time, contains a methylene-dioxy- and two phenolic hydroxyl groups. The evidence for itsstructure is less conclusive than in the two previous cases, but thealkaloid may be represented by the formula (111), or possibly (IV).Confirmation of these structural views from synthetical experimentswill be awaited with interest.Considerable confusion has existed concerning the exact relativepositions of the hydroxyl and methoxyl groups in certain well-known members of the aporphine class, although the locations ofthe oxygen atoms have been definitely established by the synthesisof the fully methylated derivatives. For some time past theformula (V) has been accepted for corytuberine, while the structures(VI) and (VII) respectively represent.the prevailing views 96regarding corydine and isocorydine (the two alternative mono-methylcorytuberines). More recent work,97 however, indicates thenecessity for a further revision of these formula Corytuberinediethyl ether, on oxidation with potassium permanganate, gave4-methoxy-3-ethoxybenzene- 1 : 2-dicarboxylic acid, a result incom-patible with the structure (V).Furthermore, by partial ethylationof corytuberine, a mixture of the two possible monoethyl etherswas obtained, and this, on oxidation with potassium permanganate,gave the two products , 4- met hoxy- 3-et hoxybenzene- 1 : 2 -dicarb-oxylic acid and 5-methoxy-4-ethoxybenzene-1 : 2 : 3-tricarboxylicacid. It thus follows that corytuberine must be represented bythe formula (VIII).HOMe0Me0CH2(VII.1 (VIII.) (IX.)e6 J. Go, Chern. Zentr., 1930, i, 234; Ann. Reports, 1930,27, 196.07 E. Spath and F. Berger, Be?., 1931, 64, [B], 2038172 PLANT :The structures of corydine and isocorydine were revealed in thefollowing way. The methylenedioxy-group was removed frombulbocapnine methyl ether (IX) by treatment with phloroglucinoland sulphuric acid, and the resulting dihydric phenol gave corydineon partial methylation.Corydine must therefore have the struc-ture (X), a fact which is confirmed by the observation that bothcorydine and its ethyl ether give 3 : 4-dimethoxybenzene-1 : 2-di-carboxylic acid on oxidation, and, in view of these results, the onlypossible formula for isocorydine is (VI).AcoThe earlier aporphine syntheses based on the Bischler-Napieralskireaction98 have involved the formation of methyl ethers and sohave thrown no light upon problems of the above type.Theprobability that synthetical methods may ultimately assist insettling some of these difficult points is now foreshadowed by astudy 99 of the possibility of employing in these reactions materialsin which phenolic hydroxyl groups are protected by carbethoxy orbenzyl radicals. So far the experiments have not advanced beyonda preliminary exploratory stage, and further results will be awaitedwith interest.Previous experience has shown 98 that the application of theBischler-Napieralski reaction to the appropriate derivative of theamide (XI) proceeds satisfactorily only when a strongly p-directinggroup in the 3-position promotes the formation of the isoquinolinering. I n other cases the main reaction results in an alternativedehydration process with the production of neutral substances t owhich the structure (XII) has been assigned.A further study ofthe conditions which promote or prevent the formation of anisoquinoline derivative in such cases has recently been made,l andit is clear that the effect of the 2’-nitro-group on the neighbouringmethylene group is such as greatly to assist the alternative process,since it has been found that 4’-methoxyphenylaceto-p-phenylethyl-amide (XIII) readily yields 1 -p-methoxybenzyl-3 : 4-dihydroiso-See Ann. Reports, 1928,25, 185; 1929,26, 174.H. Kondo and S. Ishiwata, Ber., 1931, 64, [B], 1533.99 J. M. Gullsnd and his collabor~tors, J., 1931, 2872, 2881, 2885, 2893ORGANIC CHEMISTRY .-PART III. 173quinoline when treated with phosphorus oxychloride in boilingtoluene.The 2'-nitro-group is, however, necessary for the com-pletion of the aporphine synthesis, and its effect is overcome onlywhen such groups as methoxy and benzyloxy are present in the3-position.MeoQ ( 7 3 2(XIII.)Anotherreaction isco o;,,0-9 (XIV.)interesting application of the Bisc hler-Napieralskiobserved in the facile conversion of numerous acyl-o-xenylamines (XIV) into khe corresponding phenanthridine deriv-atives (XV) by heating with phosphorus oxychloride.2DiisoquinoEine Group.-Further confirmation of the fact thatsinactine, an alkaloid from Sinomenium acutum, must be regardedas Z-tetrahydroepiberberine (I) has been obtained by the resolutionof it synthetical specimen of dZ-tetrahydr~epiberberine.~ Theproperties of the Z-form of the latter base were found to be inagreement with those described for the natural product.OMe O-CH,Some time ago evidence was produced4 to show that nandinine,an alkaloid from Nandina: domesticu, is d-tetrahydroberberrubine(11), but considerable doubt is now cast upon its real nature as theresult of a recent investigation 5 of the purification and resolutionof tetrahydroberberrubine. It is found that the physical constantsof d-tetrahydroberberrubine are in reality quite different from thosegiven for nandinine .Lupin AZ7caZoids.-The past year has been one of great activityin the field of the chemistry of the lupin alkaloids.Some of the2 G. T. Morgan and L. P. Walls, J., 1931,2447.3 E.Spiith and E. Mosettig, Ber., 1931,64, [B], 2048.Z. Kitssato, Acta Phytochh., 1927, 3, 175; Ann. Reprts, 1927, 24, 172.E. Spzith and W. Leithe, Ber., 1930, 63, [B], 3007; A., 242174 PLANT :published statements are polemical in character, so that muchconfusion exists, but, since most of the work has been directedtowards testing recent structural formulze, the results are of con-siderable general interest.Although the homogeneity of lupinine has been doubted duringthe year, it now seems certain that the substance which melts at68-69' is a single compound.6 Interesting results, however, havebeen obtained 7 from the degradation of a less pure specimen(m. p. 63-65') derived from the seeds of the yellow lupin, and itis claimed that these support Karrer's formula (I) for the alkaloid.For instance, lupinane, Cl,HlgN, obtained by the dehydration andsubsequent hydrogenation of the alkaloid, was converted intobromolupinanecyanoamide, CI1Hl9N2Br, on treatment with cyanogenbromide.The lupinanecyanoamide, C,,H,,N,, obtained from thelatter by removal of the bromine, yielded a secondary piperidinebase, CloH,lN, on hydrolysis. Dehydrogenation of the latter withsilver acetate in acetic acid at 180" led to the corresponding tertiarypyridine base, C1,HISN, which was shown by oxidation to be amixture of 3-methyl- and 6-methyl-2-n-butylpyidine. Thus, bythe action of potassium permanganate, the following products wereobtained : 2 -n- butylpyridine- 6 -carboxylic acid, 2 -met hy lpyr idine -6-carboxylic acid, pyridine-2 : 3-dicarboxylic acid, and 3-methyl-pyridine-2-carboxylic acid. These results can be explained withthe aid of Karrer's formula, but require the assumption that thelupinine used contained an isomeride (11), to which the name'' aZZolupinine " has been assigned; the relative amounts of theoxidation products indicated that the base (I) preponderated. It hasbeen more recently stated * that the pyridine derivatives whichsuggest the presence of allolupinine are not obtained from a specimenof lupinine melting at 68-69'.CH, CH*CH,*OH CH, CH,H2+9d\p2 H2CAFH\CH2 (11.)(I.) H,C'\/NvCH, H , q , p y C H 2CH, CH, CH, H*CH,.OHOf considerable importance in connexion with the structure tobe assigned to lupinine is the recent synthesis of octahydropyrido-coline (111). This has been achieved by first condensing ethylpiperidine-2-carboxylate with y-bromobutyronitrile to give theC. Schopf, E. Schmidt, and W. Braun, Ber., 1931, 64, [B], 683; A., 635;P. Karrer, ibid., p. 942; A . , 750. ' K. Winterfeld and F. W-. Holschneider, ibid., p. 137; A,, 370.K. Winterfeld, ibid., p. 692 ; A., 635.G. R. Clemo end G. R. Ramage, J., 1931,437ORGANIC CHEMISTRY.-PART III. 175nitrile (IV). The corresponding dicarboxylic ester, prepared fromthis by the action of alcoholic hydrogen chloride, was submitted tothe Dieckmann react ion, and ethyl 1 - ket o-octahydropyridocoline-2-carboxylate was obDained. l-Keto-octahydropyridocoline, pre-pared from the latter by hydrolysis with dilute sulphuric acid,yielded octahy&opyridocoline when reduced with amalgamatedzinc and concentrated hydrochloric acid.Although the decarboxylation of lupininic acid (V) with soda-lime leads to a complex mixture, there has been isolated from it,after reduction with palladium and hydrogen, a base, CgH,,N,which has the same molecular formula as, but is not identical with,octahydropyridocoline. This result does not of necessity invalidateKarrer’s formula for lupinine, since fundamental changes of structuremay occur during the formation of the new base.The fact that lupanine, C1,H2,0N2, and sparteine, C,5H26N2, arestructurally closely related is made probable by the interestingobservation 10 that d-lupanine can be reduced with hydriodic acidand red phosphorus to a base which appears to be identical withnatural Z-sparteine. In an attempt to establish one of the fouralternative formulze previously advanced by Karrer and his col-laborators for sparteine, the alkaloid has now been degraded toa saturated hydrocarbon, Cl5H,,, by repeated applications of theHofmann process and catalytic hydrogenation of the intermediateproducts.12 On the basis of these four formulze this hydrocarbonmight be one of three isomeric pentadecanes, all of which have beenprepared synthetically, but the physical properties are such as tolead to no really definite conclusions.Interesting alternative views concerning the structures of lupanineand sparteine have been advanced. It has been reported thatlupanine has now been made to yield p-lupinane, C,,H,,N, and anoily product which gives the pyrrole pine-shaving reaction, byheating with fuming hydriodic acid and red phosphorus for fifty10 G. R. Clerno, R. Raper, and C. R. S. Tenniswood, J., 1931, 429.11 See Ann. Reports, 1929, 26, 182.12 P. Kmer, B. Shibita, A. Wettstein, and L. Jacubowicz, Hdv. Chim.Acta, 1930,13, 1292 ; A., 241176 PLANThours at 240". In view of this result and the fact that sparteinecan be degraded to 2-methylpyrrolidine, the formula (VI), or someclosely related alternative such as (VII), may represent 1~panine.l~The sparteine structure is derived by replacing the >CO groupby >CH,. Hydroxylupanine, C,5H,,0,N,, would have the samefundamental structure, differing from lupanine in having a hydroxylgroup in the place of a hydrogen atom. Lupanidine, which hasbeen isolated l4 from the mother-liquors of lupinine and is isomericwith lupanine, may have the structure (VIII) in view of the possiblepresence of allolupinine in crude lupinine. These views must atpresent be regarded as rather speculative. In fact, the soundnessof the experimental work upon which they are partly based hasbeen q~esti0ned.l~ Thus it has been asserted that p-lupinane can-not be obtained from pure lupanine by the action of hydriodic acidand red phosphorus, and that its isolation in the work describedabove may be explained by the presence of lupinine in the sampleH27-p32 CH, CH,H2cvc? H2flH'p32 70 H,C N CH,CH CH \/\/H2C,,NQH2CH, CH,(VI.) (VII.) (VIII. )of lupanine employed. This criticism has, however, been resisted,16and it has been re-affirmed that p-lupinane can be obtained fromlupanine .Cytisine, an alkaloid from the seeds of Cytisus laburnum, mayultimately prove to be closely related to the lupin bases. Althoughcomparatively little evidence is at present available regarding itsconstitution, the few known reactions of the substance are not allsatisfactorily explained by the formuh hitherto put forward, andsome recent structural views l7 are accordingly of interest. Thealternative formulz (IX) and (X) have been suggested partly asthe result of an examination of the products of the exhaustivemethylation of the alkaloid. If these views are sound, the pro-l3 K. Winterfeld and A. Kneuer, Ber., 1931, 64, [B], 150; A., 371.l4 A. Kneuer, Diss., Freiburg i. Br., 1929.l5 G. R. Clemo, Miss G. C. Leitch, and R. Raper, Ber., 1931, 64, [B], 1520.l6 K. Winterfeld, A. Kneuer, and F. W. Holschneider, ibid., p. 2415.l7 H. R. Ing, J., 1931, 2195ORGANIC CHEMISTRY .-PART III. 177duction of 6 : 8-dimethylquinoline, 2-hydroxy-6 : 8-dimethylquhol-ine, and ammonia when cytisine is reduced with hydriodic acid andMe iuMered phosphorus must involve primarily a rupture of the pyrrocolineskeleton, followed by the formation of the quinoline nucleus. Otherreactions of the alkaloid can be accounted for with the aid of theseformulB, but the direct evidence for the pyrrocoline structure is atpresent scanty.Morphine Group.-The production of thebainone (I) during thecomplex transformations of the morphine alkaloids has not hithertobeen observed,18 although some simple derivatives of it have beenisolated. It has now been reported,lg however, that it accom-panies metathebainone (11) in the product of the reduction ofthebaine by means of stannous chloride and concentrated hydro-Me0H0Q\p2 C CHchloric acid, and, under the most favourable conditions, may formthe main constituent. It is also produced during the reduction ofcodeinone under similar conditions. These results have enabledinteresting deductions to be made concerning the complex changeswhich ensue when thebaine is treated with concentrated hydro-chloric acid, but the latter cannot be adequat,ely dealt with in thescope of this Report.Strychnine.-Although much published work dealing withstrychnine and brucine has again appeared during the year, thel8 The “ thebainone ” prepared by Pschorr by reducing thebaine withstannous chloride and concentrated hydrochloric acid is now called “ meta-thebainone ” in accordance with the revised nomenclature of this group (seeAnn. Reports, 1930, 27, 197).1s C. Schopf and H. Hirsch, Anmlen, 1931, 489, 224178 ORGANIC CHEMISTRY.-PUT III.chemistry of this group is so complex that only that which has avery direct bearing on suggested structural formulae can excitewidespread interest, and only those aspects will be surveyed inthis review. I n the development of the recent formula for strych-nine 2o much has depended upon the nature of dinitrostrychol-carboxylic acid. Last year, from the evidence then available, thisacid was regarded 21 as possessing one of the formula (I) and (11).Subsequent investigations 22 have shown, however, that dinitro-strychol can be transformed by the usual Curtius reactions into aurethane, so that it mustCO-Hnow be considered not as a dihydroxy-OHquinoline but as an indolecarboxylic acid, probably of the struc-ture (111). Dinitrostrycholcarboxylic acid, therefore, is 5 : 7-di-nitroindole-2 : 3-dicarboxylic acid (IV). The stability of dinitro-strychol is very surprising and is largely responsible for the failuret o recognise its true nature at an earlier date. The latest strychninet*C02HC*CO,H (Iv')(111.)NO, NH NO, NHformula, containing a quinoline skeleton, now rests on a less solidfoundation, but it is not necessarily definitely excluded by thisrecent development. It is important to observe, however, that thedoubt which has existed regarding the presence of the quinolinesystem in dinitrostrycholcarboxylic acid has been coupled withthe opinion 23 that the strychnine molecule contains a dihydroindoleskeleton, which was a feature of the earlier formula of Fawcett,Perkin, and Robinson.Of some interest is the isolation24 of three new alkaloids fromthe strychnine mother-liquors. These are a-colubrine, C2,H2403N,,the isomeric p-colubrine, and $-strychnine, probably Cz,H2,03N2 orS. G. P. PLANT. C21H2403N2'2o See Ann. Reports, 1928, 25, 193; 1929, 26, 181; 1930, 27, 199.21 K. N. Menon, W. H. Perkin, and R. Robinson, J., 1930, 830.22 K. N. Menon and R. Robinson, J., 1931, 773.p3 H. Leuchs, Ber., 1931, 64, [B], 461.24 K. Warnat, Helv. Ohirn. Acta, 1931,14, 997; A., 1312