ORGANIC CHEMISTRY.PART I.-~PHATIC DIVISION.H ydromrbons.A CONSIDERABLE number of papers dealing with the chemistry ofthe aliphatic hydrocarbons has appeared during the past few yearsand the position now reached would appear to warrant the inclusionof this section in the present Report.Studies of the complex decompositions undergone by simplehydrocarbons under the influence of heat have revealed that insilica vessels, which possess no appreciable catalytic activity, ethanea t 575" yields only ethylene and hydrogen. Further reactionsbetween ethylene and hydrogen then take place, accompanied bythe polymerisation of ethylene and the formation of higher hydro-carbons, propylene, methane, and ethane. It is probable thatsurface action plays a prominent part in bringing about the changes.With propane under the same conditions the following three reactionsappear to take place, (a) and (b) being rapid and (c) slow :At lower temperatures (200--400") in the presence of a nickelcatalyst a different course is followed :The latter reaction becomes prominent only at the higher temper-atures.Under these conditions hydrogen is without effect onthe decomposition but the nature of the catalyst is of greatimportance. 1The rate of dissociation of the simpler hydrocarbons in Pyrexglass tubes a t 650" under 1 atmosphere pressure has been shown todepend on the comparative complexity of the molecule, the mainreactions in the ca8es of ethane, propane, n- and iso-butane being,1 F. E. Frey and D. F. Smith, I d .Eng. Chem., 1928,20, 948; A,, 121168 ANNUAL REPORTS ON THE PROQRESS OF OHEMISTRY.under the prescribed conditions, dehydrogenation and demethan-ation.2When a mixture of hydrogen and ethylene is exposed to ultra-violet light in the presence of mercury vapour, the hydrocarbonsmethane, ethane, propane, and butane are produced. Under similarconditions ethane is unaffected, and the authors consider that theresults can be explained most readily by postulating the intermediateformation of atomic hydrogen, methylene by fission of the ethylene,and ethylene activated by loss of hydrogen.Another example of the complex nature of the rearrangementsobserved during reactions of this type is to be found in the changesundergone by propyl alcohol when it is passed over uranium oxidea t 400".The reaction products include propaldehyde, p-methyl-pentenal and ap-dimethyl-Aab-heptadienal and, by further decom-position of these substances, hydrogen, methane, unsaturatedhydrocarbons, hexane, hexene, and carbon monoxide and dioxide.4The polymerisation of acetylene under the influence of heat hasbeen studied with a variety of catalysts and at different temper-atures. The maximum yield of polymerised product (82%) wasobtained by passing a slow stream of the gas over clay in a glass tubeheated a t 650", and in this case the resulting tar by fractional dis-tillation yielded mainly benzene, naphthalene, and other aromatichydrocarbons. In it metal tube decomposition of the acetylene intocarbon, hydrogen, and other gaseous substances took place morereadily than polymerisation and the composition of the polymerisedmaterial was different, little or no naphthalene being present.5An instance of the usefulness of high-temperature reactions asa means of obtaining acetylenic hydrocarbons is to be found in thedehalogenation of aa-dichloroheptane by passage of its vapour oversoda-lime at 420°, the corresponding acetylene, C,H,,*CiCH, beingobtained in a yield ofAcetylene and carbon dioxide react when heated together in thepresence of metallic catalysts to give saturated hydrocarbons, asmaller amount of unsaturated hydrocarbons, carbon monoxide,and water.Acetylenecarbon monoxide mixtures, with nickel andcobalt as catalysts, give various aldehydes (formaldehyde, acetalde-hyde, acraldehyde) and a mixture of unsaturated hydrocarbons ofhigh molecular weight, along with carbon dioxide and water.'2 R.N. Pease, J . Amer. Chem. SOC., 1928,50, 1779; A., 988.A. R. Olson and C. H. Meyeis, aid., 1927, 49, 3131; A., 1928, 150.A. Mailhe and Renaudie, Compt. rend., 1928,186, 238; A., 268.C. Ssndonnini, Gazzetta, 1927, 57, 781; A., 1928, 43.ti C. Fujio, J . SOC. Chem. I n d . Japan, 1928, 31, 77; A., 732.7 A. J. Hill and F. Tyson, J . Amer. Chem. SOC., 1928, 50, 172 ; A., 269OWAXIC CHEMISTRY.-PART I. 69Amongst other reactions of a similar nature which have beenrecorded there may be mentioned the polymerisation and depoly-merisation of amylenes under the influence of heated silicates? whereagain a complex mixture results containing butylene, heptylene,octylene, nonylene, isopentane, and other products, both saturatedand unsaturated.8The action of the silent electric discharge on ethylenic hydro-carbons appears to be closely similar to that of high temperatureand pressure, prolonged action resulting in the production of highlypolymerised non-volatile substances.The rate of the reaction isgreatly dependent upon the structure of the original hydrocarbonand only slightly on the amount of the electric tension. In the caseof isobutylene polymerisation is accompanied by a redistribution ofhydrogen atoms, and on fractional distillation of the product mainlysaturated hydrocarbons are found in the portions of lower boilingpoint. Fission of the carbon chain and re-arrangement of alkylgroups leading to the formation of hydrocarbons containinghighly branched structures are marked features of the reaction.Naphthenes, but no aromatic substances, are found amongst thefractions of higher boiling point.gAn extensive contribution to the chemistry of the allene hydro-carbons and their derivatives has been made by N.BoU~S,~~who employs the appropriately substituted ally1 alcohol,R*CH(OH)*CK:CH,, as starting point in the preparation of thehydrocarbons. The action of phosphorus tribromide on thesealcohols is accompanied by isomerisation and leads exclusively tothe formation of R*CH:CH*CH2Br.11 Addition of bromine a t 0"now gives the tribromide, R*CHBr*CHBr*CqBr, which when fusedwith 75-80 yo potassium hydroxide yields the allene dibromide,R-CHBrGBr:CH2, from which the allene, R*CH:C:C€€,, is pre-pared by dropwise addition of the dibromide to zinc dust in boilingalcohol. For a detailed description of the various allenes and of thenumerous intermediate products and their derivatives the reader isreferred to the original paper.The addition of bromine (1 mol.)occurs mainly at the &-positions, the apPy-tetrabromo-derivativeresulting from the addition of a second molecule of bromine. Treat-ment of the allene hydrocarbons with concentrated sulphuric acida t -lo", followed by the action of water, gives rise to the ketoneCH,RCO*CH,, and sodamide reacts with the allenes to form sodiumderivatives of the isomeric acetylenes, CH,R*CiCNa.SOC., 1928, 60, 441 ; A., 732.S. V.Lebedev and I. A. Vinogrsdov-Volzynski, J . Russ. Phys. Chern.@ N. D. Prianischnikov, Ber., 1938, 61, [B], 1358; A., 866.lo Ann. Chirn., 1928, [XI, 9, 402; A., 1112.l1 See H. Burton and C. K. Ingold, J., 1928, 904; A., 63470 ANNUAL REPORTS ON THE P~OORESS OF OHEMISTRY.Much attention is being devoted to the properties of hydrocarbonspossessing conjugated double bonds and it is desirable to review insomewhat greater detail certain of the more important paperswhich have appeared in this connexion. Derivatives of py-di-methylbutadiene and of tetramethylbutadiene have been studiedby A. D. Macallum and G. S. Whitby,12 who fmd that the former ofthese substances yields two dibromides, one solid and one liquid.The solid is highly reactive and is shown by ozonolysis to be a1 : 4-dibromo-compound.Some evidence is given in favour of theview that the solid and liquid dibromides possess respectivelytrans- and &configurations. The same authors reach the con-clusion that, in general, butadienes substituted in positions 1 and 4polymerise less readily than the corresponding 2 : 3-derivatives, thusconfirming the views of earlier workers.13Results of great interest are described in a paper dealing withthe bromine addition compounds derivable from butadiene. l4 Ithas long been known that under certain specified conditions theprincipal product which can be isolated is the solid 1 : 4-dibromide l5and it has been generally assumed that the addition gives exclusivelythe 1 : 4-product.It now appears that the isomeric 1 : 2-dibromideis invariably formed during the reaction and that this is an unstablesubstance which undergoes slow spontaneous conversion into the1 : 4-isomeride. It does not follow, however, in view of the ex-perimental evidence now obtained, that the reaction can be ade-quately explained on the basis of 1 : 2-addition, followed by isomeris-ation. Another view would seek to account for the formation of thetwo products by considering that the hydrocarbon is capable ofassuming different polarised forms, C-C-C-C and C-C-C-S. Athird possibility, also suggested in the communication now underreview, is to regard the observed result as incidental to the operationsinvolved in the addition process.According to this idea the firststage of the reaction is of the normal ethylenic type and involves theformation of an addition complex of butadiene and molecularbromine. This is represented by the novel electronic formula (I).Its existence is transient and the subsequent events lead eitherto the formation of the 1 : 2-dibromide (11) (normal course), or tothe 1 : 4-dibromide (111), if the tautomeric properties of thepropene system CBr*C:C += C:C*CBr are brought into play. The+ - +l2 Tram. Roy. SOC. Canada, 1928, [iii], 22, 33, 39; A., 614.lS S. V. Lebedev and B. K. Mereahkovski, J . BUSS. Phy8. Chem. Soa., 1913,l4 E. H. Farmer, C. D. Lawrence, and J. I?. Thorpe, J., 1928, 729; A., 604.l5 G. Griner, C-t. rend., 1893, 116, 723; 117, 663; A., 1893, i, 460;46, 1249; A., 1913, i, 1286.1894, i, 02; J.Thiele, Anmlen, 1899, 808, 333; A,, lQO0, i, 2O R U m C CHEMISTRY .-PART I. 71mechanism involved is illustrated by the accompanying electronicformulae.Br Br*Br Br ..-..H H H H H H H H. . . .(1.) H:C~C:C;C:H --+ H:c:c*~c~"c:H ........ ........Br Br .... Br 4 Br .. ..(11.) H:C:C:CiC:H H:C:CiC:C:H (111.)H H H H H H H HThe additive properties of the conjugated system in hexatrienehave also been studied.16 Two forms of the parent hydrocarbonhave been found to exist, these being represented respectively ascis- and trans-isomerides,C&:CH*EH H$*CH:CH,HC*CH:CH, HC*CH:CH,It has been established that the 3 : 4 glycol (I), which has beenshown to exist in two stereoisomeric forms, gives on brominationboth the 3 : 4- and the 1 : 6-dibromide (11) and that the hexatriene(111) derived from the latter also adds bromine terminally.Migra-tion of bromine would appear to be involved in the 3 : 4-glycol+1 : 6-dibromide transformation, an observation which, if foundto be general, will be of considerable importance in arriving a tviews concerning the reactive forms of conjugated hydrocarbons,yH(OH)*CH:CH2 =- (iH:CH*CH2Br CH*CH:CH, ,,CH( OH)*CH:CH, CH:CH-CqBr CH*CH:CH,........ ........(1.1 (U.1 (111.)When addition of bromine occurs in the absence of hydrogenbromide, both the cis- and the trans-form of hexatriene yield 1 : 2-dibromides. These 1 : 2-&bromides are convertible into the 1 : 6-forms, which are obtained, directly, as mentioned above, by theuse of ordinary commercial bromine containing traces of hydrogenbromide.It is possible, therefore, to consider the chemistry ofhexatriene on the basis of bromine-migration reactions, but theauthors do not rule out altogether the possibility of direct 1 : 6-addition, which could be explained by assuming suitable polarisationof the conjugated molecule in the sense demanded by the modernextension of the Thiele hypothesis.Investigations of the additive properties of butadiene derivativesl6 E. H. Farmer, B. D. Laroia, T. M. Switz, and J. F. Thorpe, J., 1927,2937; A., 1928, 16172 ANNUAL REPORTS ON THE PROGRESS OF OECEMISTRY.have also been carried out by C. Pr6vost,17 who interprets theexperimental results by means of the bromine-migration hypothesis,and considers that the 1 : 4- and 1 : 2- (or 3 : 4)-forms of the semi-saturated derivatives are tautomeric. In the case of butadiene,if the sole primary addition product is the 1 : %compound, thenchanges in experimental conditions affecting the 1 : 2 e 1 : 4isomerisation should have a corresponding effect on the proportionof the two isomerides found in the bromination mixture.Actuallythe rate of isomerisation is found l8 to be little affected by experi-mental conditions and is much too slow to account for the largepercentages of 1 : 4-derivative which are observed. On the otherhand, these percentages are much influenced by the nature of thesolvent employed, and it was for these reasons that the altermtivehypothesis just outlined was put forward by Farmer, Lawrence,and Thorpe.The properties of conjugated systems are being investigated alsoby a study of their behaviour during catalytic hydrogenationin the presence of platinum-black.lg Four types of process aretheoretically possible according to the order in which the conjugateddouble bonds become saturated.The first considered is one whichproceeds in accordance with Thiele’s rule, where primary additionof hydrogen is exclusively in positions 1 and 4, followed by hydro-genation a t a different rate of the ethylenic derivative so produced.Substances such as diisobutenyl, CMe2:CH-CH:CR4e,, in which allthe four hydrogen atoms in the 1 : 4 positions are replaced by aliphaticradicals, belong to this group.The second type is more complexand does not follow Thiele’s rule. The addition of hydrogen takesplace in all the possible directions, 1 : 2 , 3 : 4, and 1 : 4, and is furthercomplicated by the simultaneous formation of fully saturatedmolecules. Isoprene, for example, by the addition of 1 mol. pro-portion of hydrogen yields isopentane (30y0), isopropylethylene(12 yo), as-methylethylethylene (13 yo), trimethylethylene (15%),and unchanged isoprene (30 yo). Divinyl, piperylene, and diiso-propenyl also belong to this type.The simultaneous addition of two molecules of hydrogen, withthe formation of a fully saturated substance, was suggested byC. Paa120 as the normal method of hydrogenation of conjugated1 7 C-t.rend., 1926, 183, 1292; A., 1927, 131; 1927, 184, 1460; A.,1927, 748; 1928, 186, 1209; A., 613; Bull. SOC. chim., 1928, [iv], 43, 996;A., 1212.18 E. H. Farmer, C. D. Lawrence and J. F. Thorpe, Zoc. cit.10 S. V. Lebedev and A. 0. Yakubchik, J., 1928, 823, 2190; J. Buss. Phys.2O Ber., 1912, 45, 2221; A., 1912, i, 703.C M . SOC., 1927: 59, 981; A,, 1928, 613, 1111ORGANIC CHEMISTRY .--PART I. 73compounds. This, however, (Lebedev’s Type 111) has not yet beenobserved experimentally and a further investigation of the ca8esdescribed by Paal reveals that they do not in fact belong to this type.The classification is completed by consideration of a fourth typein which addition of hydrogen occurs exclusively in the 1 : 2- and3 : 4-positionsY no 1 : 4-addition products being formed.Here againthe process does not follow Thiele’s rule. In all cases when onemolecule of hydrogen has been added the number of unchanged(conjugated) molecules is equal to the number of fully saturatedmolecules. A characteristic point during the hydrogenation, andone which determines fully the course of the reaction, is reached whenthe whole of the original conjugated system is consumed. Theauthors term this the “critical point of hydrogenation of a con-jugated system.”In this connexion it is of interest to refer to some hydrogenationexperiments performed with the aid of sodium amalgam.21 Thepreparation of pure amalgam is described and this is found to differwidely in properties from that prepared in the usual way.Forexample, according to von Baeyer, terephthalic acid is reducedexclusively to A2:6-dihydroterephthalic acid, which does notundergo further hydrogenation. Under similar conditions withpure amalgam, A2-tetrahydrophthalic acid is formed, and, contraryboth to the hitherto accepted evidence and to the postulates ofThiele’s theory, the A2-tetrahydrophthalic acid hydrogenates muchmore rapidly than the corresponding A1-acid. The isomerisinginfluence of the alkali hydroxide exerted for very different lengthsof time may possibly help to explain the observed differences.Consideration of the mechanism of these reductions indicatesthat the evolution of nascent hydrogen, which is then added at thedouble bond of the organic molecule, must be regarded as highlyimprobable, and instead of this the addition of sodium a t the doublelinking is postulated, the resulting product being afterwards decom-posed by the water present.22These views may be compared with those advanced by A.Gillet 23in interpreting the hydrogenation of conjugated compounds. Thisauthor regards the process as due to the addition of two sodiumatoms in the 1 : 2-positions followed by isomerisation :>CNa.CIINa*CH:C< + >CINa*CNa:CH*CH<. In a more recentpaper 24 this interpretation is objected to on the ground that itaffords no explanation of the hydrogenation of conjugated com-21 R. Willstiitter, F. Seitz, and E. Bumm, Bep., 1928, 61, [B], 871 ; A,, 766.pa Compare W. Schlenk and E. Bergmann, Annalen, 1928,468, i ; A., 1031 ;this Report, p.96.Bull. SOC. chirn., 1927, [iv], 41, 927; A,, 1927, 921.z4 G. Vavon, ibid., p. 1598; A., 1928, 1SO.0 74 ANNUAL REPORTS ON W E PROGRESS OF OHEMISTRY.pounds by agents, other than sodium, which act by virtue of theproduction of nascent hydrogen.Aldehydes and Ketones.The fact that the p-hydroxypropane- p-sulphonic acid obtainedby hydrolysis of phenyl propane- p p-disulphonate is not identicalwith “ acetone bisulphite ” 25 is at variance with the views of F.Raschig and W. Prahl on the structure of the bisulphite additioncompounds.26 The experiments with phenyl propane- pp-disul-phonate have therefore been repeated by the latter authors, whohave failed to confirm Schroeter’s results and maintain their formu-lation of the bisulphite addition compounds as hydroxy-sulphonicacids, R,C(OH)(S0,H).27 In a vigorous reply by Schroeter newevidence is given in support of his original observations.Since thebisulphite compounds cannot be regarded as sulphurous esters, andSchroeter regards the hydroxy-sulphonic acid structure as disproved,a unitary formulation becomes impossible and the structure(R,C:O)(SO,)(HOH) is now postulated. In this the known labileadditive compound of the aldehyde or ketone with sulphur dioxideis regarded as uniting with water to give a more stable substance,which can behave as a monobasic acid.% On the other hand, thestrong evidence adduced by Raschig and Prahl in favour of thehydroxy-sulphonic acid structure has been supplemented andconfirmed by studies of the Rontgen-ray and absorption spectra ofsuch s~bstances.2~The preparation of aldehydes by Rosenmund’s method has beenfound to proceed more rapidly when freshly precipitated bariumsulphate is used in preparing the palladium catalyst, and theavailability of the method is exemplified in the reduction of stearylchloride.In this case the stearaldehyde produced is accompaniedby a solid dimeride, (C18H360)2, which possesses no aldehydicproperties but yields stearaldehyde on distillation under diminishedpressure. 30A comparison of the reactivities of various alcohols for acetalformation has been made by a study of the rates of reaction ofacetaldehyde with methyl, ethyl, isopropyl, and n-butyl alcohols,25 G.Schroeter, Ber., 1926, 59, [B], 2341; A., 1926, 1226.26 F. Raschig and W. Prahl, Annalen, 1926, 448, 266; A., 1926, 939.27 Idem, Bm., 1928, 61, [B], 179; A., 273.2* G. Schroeter (with M. Sulzbacher), iw., p. 1616; A., 1216.*9 0. Stelling, “ Zusammenhang zwiechen Chemischer Konstitution undK-Rontgen-Absorptions Spektra,” Lund, 1927, p. 168 ; CeUulose Chem.,1928, 9, 100; A., 1217.so R. Feulgen and M. Behrem, 2. phydol. Chem., 1928,177,221; A,, 1117.SeeAnn. Reports, 1927, 24, 66ORUANIC CHEMISTRY .-PA€T I. 75apecially purified materials being used in the presence of a traceof hydrogen chloride. The water eliminated is shown to have amarked effect on the progress of the reaction, the hemi-acetalfirst formed undergoing hydrolysis, with the result that the apparentvalue of the bimolecular coefficient for the reaction between thehemi-acetal and the alcohol diminishes with time.The fouralcohols show only slight differences in behaviour.31 The converseproblem, involving the relation of the structure of ketones to theirreactivity in acetal formation, has also been inve~tigated.~~ In thiscase the extent of acetal formation between ethyl orthoformate andvarious ketones was measured by a specially devised analyticalmethod. Of the following eight ketones-acetone, methyl ethyl,diethyl, methyl n-hexyl, methyl tert.-butyl ketones, acetophenone,propiophenone, and benzophenone-acetone showed most reactivity(00% acetal formation) and methyl tert.-butyl ketone least (12%).In general, substitution of higher n-alkyl radicals for the methylgroup in acetone redrices the reactivity slightly, greater reductionsbeing observed with the aromatic ketones.Compounds possessing the general skeleton (I) or (11), where Xis an electro-negative group, are usually tautomeric , the mobilityand point of equilibrium being the resultant of structural factors.I I > C=C-CHX (11.1 1 1 (1.1 >CH--C=CXInvestigations into the properties of such systems have now beenextended to include cyclohexylideneacetone and cyclohexylidene-methyl ethyl ketone.= Each of these substances reacts in two forms,(111) and (IV), both of which have been isolated.These showcharacteristic differences in physical and chemical properties, andunder the influence of catalysts each form undergoes partial con-version into the other, the equilibrium mixture containing about70% of (IV).It was found that the freshly formed cyclohexylidene-acetone is much more labile than the purified material and it issuggested that the greatly enhanced reactivity is to be attributedto the changed properties of substances in statu nascendi.a Theproof thus supplied of the existence of pairs of substances which81 H. Adkins and A. E. Broderick, J . Amer. C h . SOC., 1928, 50, 178; A.,s2 H. E. Carswell and H. Adkins, ibid., p. 236 ; A., 274.58 A. H. Dickins, W. E. Hugh, and G. A. R. Kon, J., 1928, 1630; A., 887.84 Compare F. R. Goss and C. K. Ingold, J., 1926,127,2776; A., 1926, 289.27476 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.form an equilibrium mixture under the influence of catalystsemphasises the analogy existing between tautomeric substances ofthe keto-enol class and those containing the " three-carbon system."Further evidence in this connexion is being obtained from a studyof the condensation of ketones with ethyl acetoacet'ate to givetautomeric substances of such a type asA more complicated system containing the three-carbon skeletonis found amongst the substances obtained by the catalytic inter-molecular condensation of methyl ethyl ketone.The four productstheoretically probable may be divided into two pairs (I) and (11),(111) and (IV), each pair consisting of substances which should(1.1 CH,Me*CMe:CH*COEt CHMe:CMe*CH,*COEt. (11.)(111.1 CH,Me*CMe:CMe- COMe CHMe:CMeCHMe*COMe.(IV. 1exhibit tautomerism of the type now under consideration. As theresult of recent work in this difficult field all the above four sub-stances have been synthesised and the relationships existing betweenthe various homomesitones have been succehsfully established.36It appears that alkaline condensing agents acting .on methyl ethylketone lead to mixtures of (I) and (11) containing the former inexcess. Acid condensing agents, on the other hand, give ketoneswith a branched chain, (111) and (IV), sulphuric acid giving mainly(IV) and hydrochloric acid (111). Definite indications have beenobtained that the homomesitones display cis-trans isomerism ;for example, (IV) gives two different semicarbazones, m. p.203-204" and 163". It has also been proved that the homomesitonesexhibit tautomerism of the expected kind. For instance, (I) or (11)under the influence of catalysts yields an equilibrium mixture whichcontains some 68% of the a@-compound. In this case the mobilityis high, but with (111) and (IV) a very low mobility is encounteredand the final equilibrium mixture contains only 17% of the a@-com-pound. The difference between the two pairs of ketones is probablyattributable to the effect produced by the a-methyl group in (111)and (IV), which has already been found to favour the @y-phase incertain cases.3'The catalytic condensation of methyl ethyl ketone has been35 L. G. Jupp, G. A. R. Kon, and E. H. Lockton, J., 1928, 1638; A., 885.313 (Miss) A.E. Abbott, G. A. R. Kon, and R. D. Satchell, ibid., p. 2514; A.,See G. A. R. Kon and B. T. Narayanan, J., 1927, 1836; A., 1927, 873;1218.A. A Goldberg and R. P. Linstead, J., 1928, 2343; A., 1214OBGLWIC CHEMISTRY.-PART I. 77examined also by A. Petrov,w who finds that, at the ordinary temper-ature and under the duence of agents such as hydrogen chloride,sulphuric acid and sodium ethoxide, products analogous to thosederivable from acetone are obtained, but that the yield of triethylbenz-ene is remarkably small. At about 400"/100 atm., in the presence ofaluminium oxide, the total yield of hydrocarbons from methyl ethylketone is very low (8%), homomesitylene oxide and homoiso-phorone being obtained. The action of sodamide at 0" results in theformation of a mixture containing the four possible homoisophorones,C,,H,O, but the isomerides (A) and (B) seem to predominate :Acids.Differences of opinion continue to exist concerning the mechanismof Kolbe's electro-synthesis of hydrocarbons.The view of Schall,sthat acid peroxides are the intermediate products formed at theanode during the electrolysis of fatty acids, has been stronglysupported by the work of F. Fichter,m who extends the peroxidetheory to electro-chemical oxidation in general. Direct evidencein support of the peroxide theory is claimed as the result of experi-ments with potassium hexoate, which, when electrolysed at lowtemperatures, was found to yield small quantities of the acidperoxide. From a comparison of the course of decomposition ofthe peroxide and the per-acid with the behaviour of the acid duringelectrolysis it appeared that the latter reaction proceeded via theperoxide.On the other hand, 0.J. Walker 41 has obtained evidence whichis regarded as incompatible with the peroxide theory. This authorholds that the detection of peroxides during electrolysis furnishesno proof that the Kolbe reaction itself proceeds through theseintermediate compounds, and comments are made on the greatdifficulties encountered in isolating from the electro-chemicalproducts any peroxide, even at temperatures where the peroxide isknown to be stable. E'urthermore, a study of the thermal decom-position of acetyl peroxide, in the pure state or in solution, showsthat only small amounts of ethane are produced, although com-paratively large quantities of methane are formed.Since methaneis never found amongst %he anode gases during the electrolysis of8s Ber., 1927, 60, [B], 2548; A., 1928, 166.89 2. Elektrochem., 1896, 3, 83; A., 1897, i, 317.4* Tram. Amer. EZectrochern. Soc., 1924, 45, 131; A., 1924, i, 829; J .Chim. ph@que, 1926, 23, 481 ; A., 1926,912; F. Fichter and R. Zumbrunn,Hetv. Chim. Acta, 1927, 10, 869; A,, 1928, 45.41 J., 1928, 2040; A,, 111478 ANNUAL REPORTS ON THE PROGRESS OF UHEMISTRY.acetates, whereas ethane is still formed a t temperatures (100') farabove the decomposition point of acetyl peroxide, the availableexperimental evidence is held to be decidedly against the peroxidetheory.Treatment of a solution of oxalic acid a t 60" with a deficiencyof potassium permanganate in an inert atmosphere results in theformation of an activated variety of oxalic acid, which in nearlyneutral solution is capable of reducing chromic acid, bromine,bromate, and nitrate, as well as salts of silver, platinum, andmercury. The activated form was not isolated and yielded quan-titatively the ordinary calcium oxalate.The activated acid regainedits normal properties slowly at the ordinary temperature and morerapidly when heated.42A study of the action of thionyl chloride on organic acids revealssurprising differences from acid to acid. For example, oxalic,tartaric, and fumaric acids are unattacked, chloroacetic acid (butnot trichloroacetic acid or glycine), malonic, suberic, and sebacicacids form acid chlorides, and succinic, glutaric, and maleic acidsyield anhydrides.In the aromatic series, acid chloride formationis more general, but here again certain acids (e.g., terephthalic andp-hydroxybenzoic) remain unaffected and others (phthalic acid) givethe anhydride.43It has been found that the action of acetic anhydride on simplemonocarboxylic acids yields only the simple acid anhydride and nota mixed anhydride as is generally supposed. The simple anhydrideoften crystallises with one molecule of acetic anhydride of crystal-lisation, giving a substance whose empirical formula is identicalwith that of the mixed anhydride.44The action of bromine water on olefinic acids has been investigatedfrom the point of view of the effect of the concentration of the acid.45By maleic acid in 0-05N-aqueous solution 89.3% of the reactingbromine was converted into the bromohydrin, and this proportionfell to 62.5y0 in a 0-33N-solution.Maleic acid therefore behavessimilarly to ethylene and ally1 alcohol.** Sodium maleate reacts42 F. Oberhauser and W. Hensinger, Ber., 1928, 61, [B], 621; A., 605;A. E. Tschitschibabin, J. pr. Chem., 1928, 120, 214; A., 1929, 48.43 L. McMaster and F. F. Ahman, J . Amer. Chem. SOC., 1928,50, 146; A,,271.44 A. W. Van der Haar, Rec. trav. chim., 1928, 47,321 ; A,, 393. ContrastW. Autenrieth, Ber., 1887, 20, 3187; A., 1888, 260, and P. Askenmy andV. Meyer, Ber., 1893, 26, 1364; A., 1893, 607.45 J. Read and W.G. Reid, J., 1928, 746; A., 606. Compare E. Biilmann,Rec. trav. chim., 1917, 36, 313; A., 1917, i, 378.4* J. Read and R. G. Hook, J., 1920, 117, 1214; J. Read end E. Hurst,J., 1922, 121, 989OWANIC CHEMISTRY .-PUT I. 79much more rapidly than the free acid with increase in the proportionof bromohydrin (95.6% at 0.05N: 7506% at 0.33N). Undersimilar conditions, fumaric acid reacted with extreme slowness, andthe sodium salt gave, with a velocity about half that of the maleate,a slightly enhanced proportion of bromohydrin. With oleic acid(O.lh'-mixture), only some 51% of the bromine was found to beeffective. Although it was formerly considered essential to carryout experiments of this type a t a low temperature and in theabsence of bright light, it is now known that in certain cases lightpromotes the reaction; and it is shown that with oleic acid, raisingthe temperature to 90" results in increased yields of the bromo-hydrin.A comparison of the behaviour of acid anhydrides and of acidswhen passed over heated thoria, shows that the former are convertedmore easily into the corresponding ketones ; and it is suggested that,in 'the synthesis from acids, ketones arise rather from the inter-mediate formation and decomposition of the anhydrides thanthrough the thorium salts.Experiments on the formation of aceticanhydride from acetic acid by passage over titanic oxide at 300"are described in support of this ~ i e w . 4 ~The catalytic decomposition of suberic acid under the influenceof heat is found to proceed most satisfactorily in the presence of anequal weight of iron filings along with 5% by weight of crystallisedbaryta.Inthis case also, the mechanism of reaction which is considered mostprobable involves the intermediate formation of the acid anhydride.48A continuation of researches on the action of heat on salts ofpolymethylenedicarboxylic acids, the earlier stages of which havealready received mention in these Reports,49 has now led to theformation of rings containing as many as twenty-two carbon atoms.For instance, the action of heat on the yttrium salt of nonane-1 : 9-dicarboxylic acid furnishes some cycloeicosane- 1 : 11 -dione, which,when reduced by Clemmensen's method, yields cycloeicosanone.The same cyclic ketone is obtained from the thorium salt of japanicacid (from Japan wax), the constitution of which is now proved to benonadecane-1 : 19-dicarboxylic acid.A reaction similar to thatgiven by the nonane homologue leads to the form'ation from yttriumdecane- 1 : 10-dicarboxylate of a small quantity of cyclodocosane-1 : 12-dione, containing a ring of 22 carbon atoms. Only slighttraces of cyclic ketones are reported from the action of heat on thethorium or yttrium salts of tetradecane-1 : 13- or -2 : 13-dicarboxylicacid , Z-methyldodecane- 1 : 12-dicarboxylic acid, and other pents-Yields up to 40% of suberone are thus obtainable.4 7 J. Campardou and M. %on, Compt. rend., 1928,186, 691 ; A., 393.I s I. Vogel, J., 1928, 2032; A., 1136. 4e Ann. Reports, 1926, 23, 11280 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.decane-, hexadecane-, and tetradecane-dicarboxylic acids, whichsometimes give aliphatic ketones in small amount.For instance,from 2 : 13-dimethyltetradecane-1 : 14-dicarboxylic acid, smallquantities of methyl 2 : 13-dimethyltetradecyl ketone were obtained,and similar aliphatic ketones have been observed as accompanyingthe cyclic ketones in cases where these are f0rmed.mSeveral communications by R. Adams and his collaborators dealwith acids synthesised in the course of their search for substancestoxic towards B. l e p m . The number of new individual substancesobtained is so large that little more can be done than to indicate thenature of the various series investigated. These include cyclo-hexylalkylacetic acids, cyclohexylmethyl-alkylacetic acids, cyclo-pentylalkylacetic acids, (1-cyclopentylethyl-ethylalkylacetic acids,corresponding p - A2 - c y clopent enyl acids, c yclopr o p ylme t hyl- alk y 1 -acetic acids, and di(cycZohexylalky1)-acetic acids.In general itmay be said that in all series maximum toxicity towards B. lepmis shown by acids containing 16-18 carbon atoms. Bactericidalpower appears to depend little on the position of the carboxyl groupin the hydrocarbon chain, and again cycloalkylalkylacetic acids ofequal molecular weight or with equally long side chains differ onlyslightly in toxicity whether the cycloalkyl group is cyclohexyl, cyclo-pentyl, cyclopentenyl, or cyclopropyl. In the last series mentionedabove, no increased toxicity was developed by the introductionof a second cyclohexyl group, giving compounds of the typeC6H11*[CH2]z*CH(C02H)*[CH2],*C,H,,, despite the fact that thepresence of one such group at the end of certain straight-chainaliphatic acids induces toxicity towards the bacillus.51 The interestat present evoked by acids of this type is further shown by a longpaper on the reactions involved in the degradation of chaulmoogricacid, 12-A2-cyclopentenyldodecane- 1 -carboxylic acid, a synthesisof which was mentioned in last year’s Report (p.88), to homo-hydnocarpylamine, C6H,-[CH2]11*NH2, a modified form of Curtiusreaction being employed.52That the substance C8HI2O4 obtained by oxidation of a-A3-carenecannot be either a- or p-isopropylglutaconic acid has been shown inthe following way : 53 The synthesis of a-isopropylglutaconic acid10 L.Rmicka, M. Stoll, H. Schinz, Helv. Chim. Acta, 1928, 11, 670; A.,887; L. Ruzicka, H. Schinz, and M. Pfeiffer, ibid., p. 686; A,, 887,51 R. Adams, W. M. Stanley, and H. A. Stearns, J . Arner. Chem. SOC.,1928,50, 1475; A,, 764; G. R. Yohe and R. Adams, ibid., p. 1503; A., 754;J. A. Arvin and R. Adams, ibid., pp. 1790, 1983; A., 1003, 1003; L. A.Davies and R. Adams, ibid., p. 2297; A., 1132.O* C. Naegeliand G. Stefanovitsch, Helv. Chim. Acta, 1928,11,609; A., 881.68 K. V. Harihsran, IC. N. Menon, and J. L. Simonsen, J., 1928, 431; A.,396ORGANIC CHEMISTRY .-PART I. 81has been achieved by removing hydrogen chloride from ethyl B-chloro-a-isopropylglutarate (A), the preparation of the latter sub-stance being effected by the following series of reactions :CO,Et*CHPrWO*CH,*CO,Et --+ CO,Et*CHPrs*CH( OH)*CH,*CO,Et--+ C02Et*CHPrfl*CHC1*CH,*C02Et (A)The resulting a-isopropylglutaconic acid was separated into cis-and trans-forms by means of acetyl chloride.The correspondingp-isopropyl acid has also been prepared, but neither of these acidswas identical with t’he above substance C8HI2O4, for which the cyclicstructure /’ has been tentatively advanced.54yMe,;CH*C 0,HCH-CH,*CO,HThe difficult problem of the constitution of Balbiano’s acidhas been resolved by a synthesis which serves to establish itsstructure.55 This dibasic acid, C,H,,O,, was obtained by Balbiano 543in the course of oxidation experiments carried out with cam-phoric acid.When reduced it gives a monobasic lactonic acid,C,H,,04, which, on further reduction, yields app-trimethylglutaricacid. Of the possible alternative formulze (I) and (11) for thelactonic acid, the second is ruled out by synthesis,57 and (I) isCHMe-CO CH,-COtherefore Balbiano’s lactoiiic acid. The structure of the parentacid could not, however, be settled by this observation, and formany years the rival formulz of Balbiano (111) and of Mahla andTiemann (IV) 58 held the field. In more recent years the latter wasmodified by Kon, Stevenson, and T h ~ r p e , ~ ~ who preferred torepresent the acid in the liquid state or in solution, as an equili-brium mixture of the tautomeric forms (IV) and (V).It wasapparent by this time that Balbiano’s oxide structure of the acidcould not satisfy the experimental requirements, and syntheticproof of the keto-structure has now been supplied.(111.) (IV.1 (V.1s4 J. L. Simonsen and M. G. Rau, J., 1923,123, 553.5s J. C. Bsrdhan, J., 1928, 2591, 2604; A., 1243, 1215.56 Rend. R. Accad. Lincei, 1892, i, 278; A., 1893, i, 174.5 7 G. Blanc, BuEE. SOC. chim., 1901, 25, 68; A., 1901, i, 119.s* Ber., 1895, 28, 2161; A., 1896, i, 678.6s J . , 1922, 121, 65682 ANNUAL REPORTS ON THE PROURESS OF CHEMISTRY.The synthesis of ccpp-trisubstituted ketoglutaric acids such as (IV)is a matter of considerable difficulty, success being achieved in thepresent case by oxidising the acetyl group of app-trimethyl-laevulicacid (VI) to the CO-CO,H group.This necessitated a synthesis ofapp-trimethyl-laevulic acid, which was prepared, along with itsisomeride (VII), from the acid chloride (VIII) by the action ofFMe,*CO-CH, QRle2*C02H FMe,*COClCHMe *CO,H CHMe*CO *CH, CHNe*CO,Et(VI. 1 (VII.) (VIII.)zinc methyl iodide. The acid chloride was obtained from the halfester COzH~CMe,~CHMe*COzEt prepared by the action of sodiumethoxide on trimethylsuccinic anhydride. Since the last reaction mayresult in the formation to some extent of CO,Et*CMe,*CHMe*CO,H,the presence of isomerides is readily explicable. Separation of theapp- and the a@-trimethyl-lmwlic acids was accomplished byfractional crystallisation of the semicarbazones and complete proofof their identity was furnished by an independent synthesis of theaaP-compomd.Balbiano's acid was readily obtained by oxidation of app-trimethyl-laevulic acid, and since the synthesis just described is consistent onlywith the keto-formula the solid acid most probably exists in thisform.The proof now given of the structure of the acid serves inno way to diminish the remarkable nature of the changes by whichit is formed from camphoric acid, and the author now advances thefollowing scheme as providing the simplest explanation of thetransformation.a+5lO*CO2H p 2 Hp 2 H p e z -++ F02H 9Me2CH,-CMe*CO,H CH,-CMe*CO,HThe number of papers which have appeared during the year bearsample testimony to the continued interest taken in the natural'O J.C. Bardhan, Zoc. cit. compare J. Bredt, Ber., 1893, 26, 3060; A.,1894, i, 141ORGAN10 CHEmTRY.-FAR'X: I. 83fats and the acids derived from them. Studies of the decompositionof oleic acid in the presence of 30% of aluminium chloride showthat the reaction, which begins at the ordinary temperature, becomesvery vigorous at 150" and results in the formation of carbon dioxide,idammable unsaturated gases, and liquid products. Hydrocarbonsof the paraffin and eyeloparaifin types are formed, but neither hydro-aromatic derivatives nor solid p a r a f f i could be obtained. Thereaction proceeds even more readily with palmitic and steasic acids,yielding unsaturated gases, a small liquid distillate, and considerablequantities of a solid p a r a w c hydrocarbon, It is suggested that thelatter two acids may be the parent substances of petroleums richin solid paraffins.61As a result of a quantitative study62 of the oxidation of methyloleate and methyl elaidate by means of hydrogen peroxide in thepresence of acetic acid, it is again suggested that under these con-ditions oleates yield the dihydroxystearic acid of m.p. 95" andelaidates give the isomeride, m. p. 132", neither reaction involvingany change in the molecular configuration. This conclusion is inconflict with the general views of Lapworth and Mottram 63 and ofBoeseken and Belk~fante,~Q who consider that the oxidation withpermanganate, which produces the acid of 111. p. 132" from oIeicacid, does not involve cor@uratioml transformation.It i~ pointedout that the experimental conditions of the permanganate oxidationnecessitate in this particular cam the use of a large excess of alkali,which, in the opinion of Hilditch, may promote komeric change.On the other hand, it must be urged that experimental evidence insupport of the opposite view has been contrib~ted.~5 It is shownthat, in a series of alicycIic compounds containing an ethyleniclinking, permanganate gives invariably cis-glycols whereas per-acids give xhe to trans-isomerides.The pyrogenic decomposition of methyl ricinoleate has beeninvestigated in order to determine the position of the double bond,and its behaviour under the influence of heat, the ester being usedowing to the tendency of the free acid to polymerise.Heptrtlde-hyde and methyl undecenoate were obtained in good yieId by choiceof suitable conditions of heating and no sign of either racemisationor migration of the double bond could be detected.86An mid of common occurrence in marine animal oils is namedN. D. Zelinski and K. P. Lavrovski, Ber., 1928,61, [B], 1064; A,, 731.6* T. P. Hilditch and C. H. Lea, J., 1928, 1678; A., 868,68 M m . Mancheater La. Phil. SOL, 1927, 71, 63; A., 1928, 221,61 Rec. trav. chim., 1926, 45, 917; A., 1927, 132.65 See J. Bikeken, ibid., 1928, 47, 683; A., 734.66 P. S. Psniutin, J . Buss. Phy8. Chem. Soc., 1928, 80, 1 ; A,, 61784 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.cetoleic acid, C,,H&O,. It is isomeric with erucic acid and althoughthe latter has been frequently recorded as occurring in marineanimal oils, its presence is now considered doubtful.The sug-gestion is made that the supposed erucic acid was in reality cetoleicacid which, when oxidised by potassium permanganate in acetone,yields n-undecoic acid and nonane-l : 9-dicarboxylic acid. Aftertreatment with ozone the same two acids and also n-undecaldehydewere identsed, the constitution of cetoleic acid being thus provedto be CR,*[CH,],*CH:CH*[CH,],*CO,H. The same author 67 hasinvestigated the structure of zoomaric acid, a hexadecenoic acidfound in cod-liver oil, and various whale oils. This proves to beCH,*[CH,],*CH:CH*[CH,],*CO,H, identical with an acid namedpalmitoleic and isolated from a specimen of South Georgia whale oil.68Amongst other acids, obtainable from natural source8, whosestructures have now been elucidated may be mentioned jalapinolicacid, which has been shown to be d-10-hydroxyhexadecoic acid.In the course of the work various hydroxy- and keto-fatty acids wereprepared and both 11-hydroxypentadecoic and 1 l-hydroxyhexa-decoic acids were synthesised, The former of these differs struc-turally from convolvulinolic acid, which occurs along with jala-pinolic acid in the jalapin gluco~ides.~~The use of triphenylmethyl ethers, which have proved so advan-tageous in synthetic work in the carbohydrate group, has now beenextended to the synthesis of partly acylated glycerides.Thepreparation of glycerol ap-dibenzoate may be cited as an exampleof the method. Glycerol or-monotriphenylmethyl ether is firstprepared and the dibenzoate of this substance when treated withhydrobromic acid in acetic acid at 0" readily yields glycerol orp-di-benzoate.By suitable modifications of the process glycerol p-benz-oate, glycerol p-p-nitrobenzoate, and other similar compounds maybe prepared.,OIt is only within recent years that methods have become availablefor determining quantitatively the composition of natural fats, andfigures are available for only a few of the fats in common use.Information concerning the way in which the fatty acids are com-bined with glycerol in the fats is still more lacking owing to the veryconsiderable experimental di%culties which have to be faced. In arecent paper71 results are given which have been obtained by the6 7 Y.Toyama, J . SOC. Chem. Ind. Japan, 1927,30, 597, 603; A., 1928,164.68 E. F. Armstrong and T. P. Hilditch, J . SOC. Chem. Ind., 1925, 44, 1801.;6B L.A. DaviesandR. Adams, J. Amer. Chem. Xoc., 1928,50,1749; A., 990.' 0 B. Helferich, P. E. Speidel, and W. Toeldte, Ber., 1923, 56, 766; A.,1923, i, 331 ; B. Helferich and H. Sieber, 2. physiol. Chem., 1927, 170, 31 ;1928, 175, 311; A., 44, 734.A., 1925, i, 778.'l T. P. Hilditch and C. H. Lea, J., 1927, 3106; A., 1928, 162ORGANIC CHEMISTRY.-PART I. 85application of new principles in an attempt to elucidate this complexproblem. The underlying idea has been to alter the chemicalcharacteristics of the glycerides in such a manner that new glycerides,more readily separable, may be obtained without disturbance of thecombined glyceryl radical.Two such processes are described :(a) controlled oxidation by potassium permanganate in acetone,which does not affect saturated glycerides, but converts the mono-,di-, and tri-oleins into the corresponding acid azelaic esters; (b)controlled oxidation with an acetic acid solution of hydrogenperoxide, which is again without effect on the saturated glyceridesbut gives dihydroxystearic esters with oleins. In so far as theresults obtained can be summarised in the space now available, itappears, contrary to a widespread impression, that in vegetableglycerides the fatty acids are distributed impartially rather thanselectively and that in consequence the occurrence of simple tri-glycerides and mixed saturated glycerides is reduced to a minimum.Thus the physical properties of such a fat must depend primarilyupon the particular mixture of fatty acids from which it takes itsorigin.For instance, Cacao butter, which contains palmitic,stearic, and oleic acids in approximately equal proportions, consistsmainly of mono-oleic disaturated glycerides and this fat more,closely resembles an individual glyceride than one in which theproportions of the component acids are very different. Again, incotton-seed oil entirely saturated glycerides are present only in verysmall proportion, and the palmitic acid is uniformly combined withthe unsaturated acids, the proportion of palmitic and unsaturatedacids being approximately 1 to 3.In mutbon tallow, taken as arepresentative of the animal fats, a very different state of affairsprevails and a much greater proportion of fully saturabed glyceridesis present.Another method which is being employed with success in thestudy of natural glycerides consists in the bromination of the fatin light petroleum, followed by separation of the bromides bydissolution in various solvents. The constitution of the individualbrominated glycerides may then be determined by means ofhydrolysis with hydrochloric acid. Linseed oil has in this wayyielded dilinoleolinolenin bromide, two linoleodilinolein bromidesand dilinoleo-olein bromide. Soya-bean oil, train oil, oil of silk-worm pupa, and cod-liver oil have also been investigated and thelarge number of individual glycerides separated and identified bySuzuki and his collaborators provides a tribute both to the skill ofthe workers and to the power of the experimental method.7272 B.Suzuki and Y. Yokoyama, PTOC. Imp. Acad. Tokyo, 1927,3, 526,529;1928, 4, 161; A., 152, 736; B. Suzuki and Y. Masuda, ibid., 1927, 3, 531;1928, 4, 161, 165; A., 153, 736. See also (for soya-bean oil) K. Hashi, J .SOC. Chem. Ind. Japan? 1927,30, 849, 866; A., 1928, 73686 ANNUAL REPORTS ON THE PROGRESS OF OHEMISTRY.NitTqen CO?T&'pCYUnds.Important developments have taken place during the year inthe chemistry of the aliphatic diazo-compounds. It has beenknown for some time that certain aldehydes react with diazomethaneto form methyl ketones,73R*CHO + CH,N, + R-CO-CH, + N,,and an analogous reaction was later applied to certain acid chloridesand expressed as follows : 74R-COCL + CH,N2 + R*CO*CH,Cl + N2.This observation has now been followed by detailed investigationswhich have revealed'5 that the course of the reaction with acylchlorides is in point of fact very different from that indicated in theabove equation, and that the primary product is a diazo-ketone.The chloro-ketone, if formed at all, is the result of subsequentdecomposition of the diazo-ketone in the presence of halogen acid.The general nature of the reaction for the acylation of diazomethanehas thus been established and may be summarised by equations(A), (B), and (C), of which under all conditions (A) and (B) pre-dominate over the secondary reaction (C) :R*COCl+ CH,N, --+ R-CO-CHN, + HCI .. (A)R*CO*CHN2 + HC1+ R*CO*CH,Cl + N, . . (C)CH2N2+HCl+ CH3Cl+N, . . . . (B)The best conditions foI: the preparation of the diazo-ketones involvethe addition of the acid chloride (1 mol.) to an ethereal solution ofdiazomethane (2 mols.), and Nierenstein 7f3 has recently claimed thatby adding the reagents in the reverse order the course of the re-action may be altered with the production of the chloro-ketone inhigh yields. Other workers,?' however, have failed to co&m thisview and maintain that under all conditions (A) is the primaryreaction. If a chloro-ketone is required, the best procedure is tomake the diazo-ketone and submit it to the action of hydrogenchloride.73 H.Meyer, Monateh., 1906,26, 1300; A., 1906, i, 87.74 D. A. Clibbena and M. Nierenstein, J., 1916,107, 1491 ; A., 1916, i, 1062.See also R. T. Dale and M. Nierenatein, Ber., 1927, 60, [BJ, 1026; A,, 1927,664.7 6 W. Bradley and R. Robinson, J., 1928, 1310, 1646; A., 769; F. Arndtand J. Amende, Ber., 1928, 61, [B], 1122; A.* 769. Compare F. Arndt,B. Eistert, and J. Partale, Ber., 1927, 60, [B], 1364; A,, 1927, 774.70 Nature, 1928, 121, 940; A., 739.77 W. Bradley and G. Schwarzenbach, J., 1928, 2904; A., 1929, 68ORGAN10 UHEWSTRY .-PART I. 87Since the diazenes R,*CN2 react vigorously both with electronseeking agents such as halogen molecules and w&h electron-donatingagents (for instance, organo-metallic compounds), theee compoundspresent interesting problems from the point of view of the electrontheory of ~alency,~8 on which basis the Angeli-Thiele formula maybe written CR,=&=N=, where a line represents two electrons.Different positions in the diazomethane molecule are involved inthe two types of reactions referred to and it is found that the anionoidcentre is situated on the carbon atom next to the positively chargednitrogen atom, and the negatively charged nitrogen is the kationoidcentre.The apparent anomaly is explained by assuming that themaintenance of atomic electron configurations is more importantthan neutralkation of the charges and it is shown that in each casethe ultimate result of the electronic displacements postulated in theanionoid (A) and in the kationoid type of reactivity (B) is a moreeven distribution of the charge :0 l o - (A) CH2=N=N= CH2-N-N=-&/The reaction with magnesium phenyl bromide may beas an example of kationoid reactivity,(B)considered+ - ++ - ICR2=N=N- + Fh{Mg}Br --+ CR,=N-&PhYMg Br ++I -bht the action of acids on diazomethane is probably representedby the scheme (A).+X-@ + CH2=h=k + ~(H--CH,--N=_N-- -+P - +X CH3 + -"- + XCH, + N,In this case the process depends upon the intermediate formationof a diazonium salt, which then decomposes in the usual way.Rather more complex cases are to be found in the reactions betweenthe diazenes and aldehydes and acyl chlorides.A possible mechanism for the former is given in the annexedequations, although it is necessary to remember that the f;rst stagemay proceed only to a minute extent, instead of involving the com-plete transfer of charge here represented.Once the carbon atomaare joined, the main changes leading to the separation of nitrogen' 8 W. Bradley and R. Robinson, loc. cit88 ANNUAL REPORTS ON THE PBOGRESS OF CHEMISTRY.and the wandering of a proton may proceed by the more direct routeindicated in the expression (B) I 1-0-) - +- ff - I R*CH=O< + Cq=NzN= --+ RCH-CH2-N5-As an example of the latter type of reaction the formation of diazo-acetophenone from benzoyl chloride and diazomethane may berepresented by the following two stages, the explanation given beingquoted from the paper by Bradley and Robinson.0 tf - A-!! I I Ph-Q--CH='N-N- P h y - - - < i H N ,C1 H Cl @" At this stage the terminal nitrogen has succeeded in transferringa fraction of its negative charge to the oxygen atom and a bondbetween the carbon atoms has been initiated. The repulsionexerted by the central nitrogen on the proton indicated is nowreinforced by the oxygen and chlorine nuclei and the followingchanges may be almost synchronous."This conception of a reaction mechanism, in which the direction isinitiated and controlled by a process involving fractional valenciesbut the course is completed by a different and more direct route,was put forward in explanation of m-substitution in the benzeneseries some years ag0,79 and is now shown to be of wide generalapplication.79 R, Robinson, M m .Xa&ter P?d. SOC., 1920, 64, No. 4; A., 1921,ii, 646ORQANIC CHEMISTRY .-PART I. 89As the result of studies on the action of diazomethane on aldehydesit is concluded 80 that the course of the reaction is governed principallyby the nature of the aldehyde and t o a lesser extent by experimentalconditions. Amongst the examples quoted in support of this viewthe following may be mentioned as illustrative. Although m-nitro-benzaldehyde affords exclusively m-nitroacetophenone, the p -isomeride gives a mixture of p-nitroacetophenone and p-nitro-phenylethylene oxide. Chloral is converted into aaa-trichloro-propylene &-oxide, and the earlier claim 81 that trichloroacetoneis obtainable in this reaction is not substantiated.The availableevidence appears to indicate that substituted ethylene oxides areformed most readily from those aldehydes which show a pro-nounced tendency towards the formation of hydrates.The behaviour of diazomethane has been examined from yetanother point of view by H. Meerwein and W. Burneleit.82 Theseauthors have investigated the possibility of activating atomicgroups, in this case the carbonyl group, by means of complexformation. At low temperatures a solution of diazomethane inacetone undergoes little or no change, but on addition of some10% of water nitrogen is freely evolved, and the reaction thusinitiated continues as diazomethane is added, until about 80% of theequivalent amount has been used. The products include as-dimethylethylene oxide, methyl ethyl ketone, and in all probabilitydiethyl ketone and methyl n-propyl ketone.The authors representthe course of the change by the accompanying equations, theactivity of the diazomethane being anionoid in character and pro-ceeding in accordance with the scheme outlined above.Further reaction with diazomethane yields methyl ethyl ketoneand then higher ketones. During the reaction the added waterremains unmethylated and in a similar way alcohols, which mayreplace the water as catalysts, remain unaffected. Since dilutesodium hydroxide solution also promotes the reaction, the activatingeffect cannot be attributed to hydrogen ions.The interaction of nitroguanidine with hydrazine in dilute aqueoussolution a t 50-60" yields the interesting nitroaminoguanidine,NO,*NH*C( :NH)*NH*NH,, probably via an intermediate additive*O F.Arndt and B. Eistert, Ber., 1928, 61, [B], 1118; A,, 739; F. Arndt,B. Eistert, and J. Amende, {bid., p. 1949; A,, 1240.81 F. Schlotterbeck, Bsr., 1909, 42, 2669; A., 1909, i, 663.8% Bw., 1928, 61, [B], 1840; A,, 121790 ANNUAL REPORTS ON THE PROGRESS OF OHEMISTRY.compound. This new substance decomposes explosively at itsmelting point, reduces copper and silver solutions with formationof explosive metallic derivatives, and may be used in the quantitativeestimation of nickel. In addition, the blue colour of the nickel saltin alkaline solution provides a delicate qualitative test for thatmet al.83In the course of work on the synthesis of creatinol84 [N-methyl-N-( 8-hydroxyethy1)guanidinel it was found that the most satisfactorymethod for the preparation of amino-alcohols of the aminoethanoltype is to treat ethylene chlorohydrin with liquid carbonyl chlorideand to allow the resulting P-chloroethyl chloroformate to react withamines in benzene solution.The product is now a p-chloroethylalkylcarbamate, CH,Cl*CH,*O*CO*NHR, which can be decomposedby excess of alkali hydroxide to give the corresponding amino-alcohol, HO*Cl&*CH,*NHR. These aminoethanols may then betransformed into the guanido-alcohol either by the Erlenmeyermethod, in which a salt of the amine is allowed to react withcyanamide, or by Rathke’s method, in which the amino-alcohol istreated with a salt of S-alkyl isothiocarbamide.The mechanism ofthe latter reaction is still by no means clear and in the author’spaper the earlier addition theories85 are rejected and a decom-position theory is advocated in which the formation of cyanamidefrom the alkyl isothiocarbamide is postulated as a first stage, whichis then followed by the usual Erlenmeyer reaction. The synthesisof creatinol hydrobromide, which is cited as an example of thistype of reaction, is achieved by allowing 8-ethyl isothiocarbamidehydrobromide to react with methyl-P-hydroxyethylamine in thepresence of a little water. The free base is sensitive to the action ofalkali, but the salts are stable towards acids.Pure crystalline alanine may be obtained from pyruvic acid bysubmitting the latter to the simultaneous action of hydrogen andammonia in the presence of colloidal palladium stabilised by starchpaste .*6A more convenient method for the preparation of arginine fromgelatin has been described 8’ and in the s m e paper the transformationof arginine into glycocyamidine hydrochloride is reported.This is88 R. Phillip and J. F. Williams, J . Amer. Chem. Soc., 1928, lio, 2466; A.,84 H. Schotte, H. Priewe, and H. Roescheisen, 2. physiol. Chem., 1928,85 H. Lecher and F. Graf, Ber., 1923,58,1326; A., 1923, i, 761 ; H. Lecher,86 Aubel and Bourguel, Cornpt. rend., 1928, 188, 1844; A., 868.M. Berpann and L. Zervas, 2. physiol. (%ern., 1927, 172, 277; A.,1229.174, 119; A,, 1122.Z. physiol. Chem., 1928, 176, 43; A., 1123.1928, 612OBOANIC CHEiMISTRY.-PART I.91carried out by careful treatment of triacetyl anhydro-arginine withglycine ester, the resulting ethyl ester of diacetyl glycocyamine beingheated with hydrochloric acid a t 100". The hydrochloride ofglycocyamidine (I) thus obtained has also been prepared by heatingglycocyamine (11) in a sealed tube with fuming hydrochloric acida t 140".Another method 88 for the preparation of glycocyamidine consistsin the interaction of guanidine and ethyl aminoacetate a t 0", areaction which suggested a possible method for the detection ofesters from polypeptides. That this expectation was not realisedmay perhaps be due to the delicate nature of the reaction, which isinhibited completely by the presence of ammonia. The reactionmechanism suggested by the authors involves the loss of ammoniafrom the guanidine, followed by the union of the amino-acid withnascent cyanamide.When a-amino-acids are warmed with acetic anhydride andpyridine, carbon dioxide is eliminated and an a-acetamido-ketone isproduced by the attMhment of acetyl groups to the nitrogen andto the a-carbon atom :Certain other acids undergo a similar change, benzyl methyl ketonebeing formed in this way from phenylacetic acid, but alkylamino-acids, @-amino-acids, and amino-acids which contain no replaceablehydrogen in the a-position undergo simple acetylation withoutfurther change.Furthermore, the azlactones formed from leucine,phenylalanine and aspartic acid by the action of acetic anhydride,yield with acetic anhydride and pyridine the same acetamido-ketones as do the corresponding amino-acids.The suggestion istherefore made that the azlactones are formed as intermediatecompounds during the reaction, which may involve also the formationof a @-ketonic acid at one stage.89Experiments carried out by L. Zervas and M. BergmanngO haveserved to show that the substance produced by the auto-condens-ation of arginine ester and considered by Fischer and Suzuki to bearginylarginine is in reality optically inactive as-diguanido-n-valeric anhydride, NH2*C( :NH)*NH*[ CH2],*CH<NH-c .NH. Thelatter substance has been synthesised by the condensation of arginine8 8 E. Abderhdden and H. Sickel, 2. physiol. Chem., 1928, 173, 61 ; A,, 611.89 H. D. DakinandR. West, J .Biol. Chem., 1928,78,91,745; A., 874,1120.90 Ber., 1928, 61, [B], 1106; A,, 874.R.CH(NHJ.CO2H + (CH,*CO),O -+ R*CH(NH*CO*CHJ*CO*CH.SCO-YH 92 ANNUAL REPORTS ON THE PaOGRlSS OF CHEMISTRY.with S-ethyl isothiocarbamicle and found to be identical with theso-called “ arginylarginine.” It appears that, on the ‘liberation ofarginine ester from its hydrochloride, interaction takes place betweenthe ester group of one molecule and the guanido-group of another,giving a very unstable ester of arginylarginine, from which as-diguanidovaleric anhydride is produced by loss of ornithine methylester, followed by internal ring closure in the residue.Hydantoin-3-acetic acid yields on hydrolysis a glycylg1ycine-N-carboxylic acid which has been formulated variously asCO,H*NH*CH,*CO*NH*CH,*CO,H 91and as CO,H*NH*CH,*C( OH):N*CH,-C02H.92 It is now found thathydantoin-3-acetic acid is transformed by ammonia into a diamideidentical with that formed from ammonia and carbonyl bisglycineester, and that the ester prepared by the action of carbonyl chlorideon glycine ester is identical with the ester of glycylg1ycine-N-carboxylic acid. For these reasons it is now suggested that thecorrect structure of the last-named acid is given by the formulaCO (NH* CH2*C02H)2.93Further investigation of oxidation and reduction reactions withcystine and cysteine indicates that, in the absence of some thirdcomponent, neither reduced indigo nor reduced indigo-carmine canbe oxidised by cystine.The activating agent is not iron, whichcan, however, influence the reaction velocity, but seems to be anunstable oxygen or sulphur additive product of indigo-carmine whoseessential function is to activate the sulphur atom, whereupon thelatent reducing or oxidising power of the -SH or IS, groups isbrought into play.9*An improved method has been employed in the isolation ofphosphocreatine, which has been obtained as the calcium derivative,to which is assigned the composition C4H,0,N,PCa,4H,0. Thestructure suggested is PO( OH),*NH*C(:NH)*NMe~CH,*C02H, whichrenders this substance the first compound containing phosphorusattached to nitrogen to be isolated from natural sources.1 Thcquestion of its importance in the series of chemical changes involvedin muscular contraction is as yet undecided.2Organa -metallic Compounds.Beryllium alkyl derivatives are obtained by allowing anhydrousberyllium chloride to react with magnesium alkyl halides in the9l E.Fischer and E. Fourneau, Ber., 1901, 34, 2868; A., 1901, i, 676.92 H. Leuchs and P. Sander, Ber., 1925, 58, 1528; A., 1926, i, 1248.93 F. Wessely and E. Komm, 2. physiol. Chem., 1928, 174, 306; A., 623.94 E. C. Kendall and D. F. Loewen, Biochem. J., 1928, 22, 649; A., 1122.1 C. H. Fiske and Y . Subbarow, Science, 1928, 67, 169; A., 744.D. Ferdmann (with 0. Feinschmidt), 2. phy8ioZ. Chem., 1928, 178, 62 ;A,, 1267ORGAKW CHEMISTRY.-PmT I. 93complete absence of oxygen. The interaction of the metal withmercury alkyls fails to give the beryllium compounds, althoughthis method may be used to prepare the aryl derivatives.Thealkyl derivatives combine very readily with oxygen. The dimethyland diethyl compounds are spontaneously inflammable, but beryl-lium di-n-butyl oxidises more smoothly, yielding butyl alcohol andprobably beryllium n-butyl oxide. The beryllium alkyls reactviolently with water to form the corresponding hydrocarbon.Carbon dioxide reacts with beryllium dimethyl and yields aceticacid, but with the diethyl derivative triethylcarbinol is formed.Differences in reactivity between the dimethyl and the diethylcompound are revealed also in certain other reactions, such astheir interactmion with benzophenone, the former giving in this casediphenylmethylcarbinol and the latter, by reduction of the ketone,benzhydrol.3 Beryllium alkyl halides, BeRX, are also described.They are prepared by heating beryllium with alkyl iodides in thepresence of mercuric chloride, and in general they are less reactivethan either the beryllium dialkyls or the Grignard compounds.4Some organo-metallic compounds of thallium and of platinumhave been prepared, and have special interest from the point of viewof the co-ordination theory of valency.They are obtainable eitherby the double. decomposition of thallium dialkyl halides withcompounds such as thallium acetylacetone, or by the action ofthallium dialkyl ethoxide or dialkyl carbonate on the appropriatediketone. In this way thallium dimethyl acetylacetone and thecorresponding diethyl derivative were obtained as well as thalliumdimethyl benzoylacetone and analogous compounds.They are allcrystalline compounds with unusual properties. When heated underdiminished pressure they sublime, and in benzene and hexane theydissolve readily, giving non-ionised solutions. Their properties,therefore, are exactly those to be expected of chelate co-ordinatedcompounds with the structure (I) :CMe-0\clle:Of‘PtMe, (11.1 (1.) CHH \TlMe,In aqueous solution the covalent state gives place to an ionisedcondition, the solutions are now alkaline in reaction, and the thalliumdialkyl can be titrated quantitatively.5The analogous platinum compound (11) is obtained from thallousH. Gilman and F. Schulze, J., 1927, 2663; A,, 1928, 50.Idem, J . Amer.Chem. SOC., 1927, 49, 2904; A,, 1928, 60.R. C. Menzies, N. V. Sidgwick, E. F. Cutcliffe, and J. M. C. Fox, J., 1928,Compare F. Feigl and E. Backer, Monntsh., 1928, 49, 401 ; 1288 ; A., 746.A. 112694 ANNUAL REPORTS ON THE PROBRESS OF CHEMISTRY.acetylacetone and trimethyl p3atinic iodide. Like the correspondingthallium compounds, it may be sublimed unchanged.6A further examination of the reaction between sodium andpropionamide and butyramide in benzene shows that the mainproduct is the derivative R*CO*"a and that there is no markeddifference between the action of sodium and potassium. Evidencein favour of this view is furnished by the observation that propionylchloride and the product formed by the action of sodium on pro-pionamide yield dipropionamide.7The interest taken in Grignard's reaction and its applicationscontinues unabated and it will be possible to indicate only a few ofthe new observations which have recently been described.Un-certainty still remains concerning the structure and even the mole-cular weight of the magnesium alkyl halides. As the result ofevidence gained by determinations of molecular weight in etherealsolution, Terentiev 8 advocated the formula Mg(Alk,MgI,,2Et20),but it now appears that such experiments are complicated by thestrong tendency towards association shown by these compounds.Under suitable conditions the molecular weight corresponds tothat required by the formula Alk.MgI1 z (OEt,),. Thus in dilutesolution magnesium ethyl bromide exists mainly as the complexEtMgBr f 15 (OEt,),, only small amounts being ionised or poly-merised.Exactly analogous observations have been made withmagnesium iodide dietherate and there appears to be no validreason for postulating the existence of bimolecular compoundsduring the reaction between magnesium methyl iodide and water.gThe magnesium alkyl and aryl compounds in ethereal solutionundergo autoxidation unless oxygen is rigidly excluded. A detailedstudy of this reaction has now been contributed in the course ofwhich the authors make certain suggestions concerning themechanism of the reaction between magnesium methyl iodide andoxygen. The first stage of this is considered to lead to the formationof the complex (MeMgI 9 - 3 0 - - MeMgI,Et,O) which may decom-pose in a t least two ways.In the one case, by means of a uni-molecular reaction, the two iodine atoms are replaced by oxygen,and the liberated iodine reacts in the known manner with un-changed magnesium methyl iodide. The newly formed complex6 R. C. Menzies, J., 1928, 565; A., 609.7 A. Parts, Ber., 1927, 60, [B], 2520; A., 1928, 168.8 2. anorg. Chem., 1926, 156, 73; A., 1926, 1130.0 J. Meisenheimer, Ber., 1928, 61, [B], 708; A., 624; J. Meisenheimer andW. Schlichenmsier, ibid., p. 720; A,, 626. See also Q. Mingoia, Gazzettu,1928, 58, 532; A., 1266. Contrast L. Eierzek, Bull. Soc. chim., 1927, [iv],41, 1299; A,, 1927, 1176; andD. Ivanov, Corrvpt. rend., 1927, la, 605; A,,1927, 961ORUANIC CHEMISTRY.-PART I. 95(MeMg - - 3 0 MgMe,Et,O) then decomposes, giving MgO andMg(OMe),.In the second case, a bimolecular reaction is responsiblefor the oxidation, by the complex, of one molecule of MeMgI toMeOOMgI. The oxidation potential of the residue(MeMgI - - 2 0 MeMgI,Et,O)is thereby so much diminished that expulsion of the iodine no longeroccurs, and the oxidation now results in the formation of methoxy-residues. Both these reactions occur simultaneously in con-centrated solutions of the magnesium alkyl iodides, but the bimole-cular reaction is excluded in the case of iodides in diIute solu-tion. The unimolecular reaction invariably predominates withbromides .loThe interaction of magnesium alkyl halides and alkyl sulphonatesis usually represented by the equation2RS0,Alk + 2R’MgX --+ 2R’Alk + (RSO,),Mg + MgX,,which, however, does not agree with the results of more recentinvestigations.It now seems 11 that the product R’Alk is obtainedin yields never greater than 50% and that an equivalent amountof alkyl halide is formed simultaneously.(a) RS0,Alk + R’MgX --+ R’Alk + RS0,MgXand (b) RS0,Alk + RS0,MgX --+ AlkX + (RSO,),Mgis now suggested and experimental evidence in its favour is providedby the separation of n-butyl iodide and magnesium naphthalene-2-sulphonate from the products of reaction of iodo-magnesiumnaphthalene-2-sulphonate and n-butyl p-toluenesulphonate. Im-proved yields of the derivative R’Alk are, however, claimed byS. S. Rossander and C. S. Marvel 12 for the particular case in whichone molecule of a Grignard reagent containing six or more carbonatoms reacts with two molecules of y-chloropropyl p-toluene-sulphonate.This provides a method whereby a carbon chain maybe lengthened by three carbon atoms. A different reaction mechan-ism, not favoured by Gilman and Heck, is advocated by theseauthors.The utilisation of mixed magnesium alkyloxyhalides, RO-MgX,which are prepared by the action of magnesium alkyloxides onetherated magnesium halides, has been studied by V. Grignard andM. Fluchaire.ls Neither the etherated halides nor the magnesiuml o 5. Meisenheher and W. Schlichenmsier, Bw., 1928, 61, [BJ, 2029; A.,1232.l1 H. Gilmsn and L. L. Heck, J . Arner. Chem. Soc., 1928, 50, 2223; A,,1124.le Ibid., p. 1491; A., 732.la Ann.Chh., 1928,[x], 9, 6 ; A., 396.The schem96 BNNUAL REPORTS ON THE PROQRESS OB CHEMISTRY.alkyloxides act as condensing agents and a mixture of the two isinactive until the reaction MgX, + Mg(OR), o_, 2ROaMgX hastaken place. With aldehydes in the presence of mixed magnesiumalkyloxyhalides, two simultaneous reactions are observed : (I) aldolformation, and (11) ester formation, SRCHO + R*CO,*CH,R. Forinstance, acetaldehyde and magnesium butoxyiodide in the presenceof ether yield ethyl acetate, aldol, and butyl acetate, the formationof the last substance being due to the interaction of ethyl acetateand the magnesium derivative. The production of esters is explainedby the intermediate formation of a semi-acetal,l4 the subsequentcondensation of which yields the ester :RGHO + BuO*MgI + BuO*CHR*OMgI2BuO*CHR*OMgI --+ BBuO*MgI + HR.0.HR LJ HR.0. HR --+ R*CO,*CH,R. LFAldol formation is represented by the equationCHO*CH,R + BuO*CH(OMgI)*CH,R +BuOH + CHO*CHR*CH(OMgI)*CH,RSometimes, as with furfuraldehyde, reduction to the alcohol takesplace, R*CHO and BuO-MgBr giving CH,Pr*O*CHR*OMgBr, whichthen decomposes into PrCHO and CH,R*OMgBr.Methyl alkyl ketones react in accordance with the scheme2RoCOMe --+ R*CO*CH,*CMeR*OH and from the alcohols so formedunsaturated ketones of the type R*CO*CH:CMeR are readilyobtainable. In certain cases, e.g., when methyl isobutyl ketonereacts with magnesium butoxybromide and benzoyl chloride, theenolising effect of the organo-metallic compound becomes evidentand an ester of the enolic form of the original ketone is producedt o some extent.In this case the first stage is represented byBuO*CR’(CH,R)*OMgBr .+ BuOH + CHR:CR’*OMgBr, and in thepresence of benzoyl chloride the ester CHR:CR’-0,CPh is thenformed.The addition of sodium to ethylenic substances takes place onlywith those compounds which have aryl groups attached to thecarbon atoms concerned in the ethylenic bond. Numerous examplesin support of this conclusion are cited by W. Schlenk and E. Berg-mann 15 in the course of their lengthy and important communication,comprising more than 350 pages, on the chemistry of the alkali1‘ Compare A. Verley, Bull. SOC. chirn., 1926, 37, 837; A., 1926, i, 783;Ann. Report%, 1926, 22, 72; 1927, 24, 92.An&, 1928, 463, 1; 464, 1; A., 1031Oa&BM(3 CHEMISTRY.-PART I.97metal-organic compounds. One exception only to this rule has beendiscovered and in this case the substance in question (I) so closelyresembles the fulvenes (11) in structure as to render the exceptionalbehaviour more apparent than real. I n general, addition may takeplace in one of two ways : (a) represented by CPh,:CPh2~+CPhPa*CPh,Na or (b) represented by 2CPh,:Cq + 2Na +CPh,Na*CH,*CH,*CPh,Na. In both cases the obvious tendency isfor sodium to attach itself only to arylated carbon atoms, and it isinteresting to find that in the fulvene series only C, combines withsodium [i.e., the reaction follows course (b)] unless C, is arylated.Lithium enters into additive combination more rapidly thansodium, and sometimes leads to compounds differing in configurationfrom those derived from the corresponding sodium compounds.For an account of a systematic survey of this problem and of thebehaviour of the alkali compounds towards water, carbon dioxide,alkyl halides, and many other reagents, reference must be directedto the original paper.Carbohydrates.Monosacchides and G1ucosides.-In studying the optical rotationof sugars in solution Hudson has relied to a considerable extent onthe classification of the sugars into a- and @-varieties.This hasbeen subjected to criticism 16 and it is pointed out that contradictoryresults are obtained when other physical constants are considered.It is necessary to determine more than one additive property and,when only two forms are present during mutarotation, the ratiokl/k2 = K must be constant.From observations of specificrotation and molecular volume the value K is 0-50 and 0.47 formannose, whereas from the index of refraction the value K = 0.64is deduced. It is thus concluded that more than two forms ofmannose are present in solution, although the a- and p-varietiespreponderate. A similar conclusion is also reached from the studyof the optical behaviour of galactose17 in aqueous solution. Themutarotation curves of a-galactose do not obey the unimolecularlaw, and the early stages of the changes are rapid. With P-galactosethere is no change of rotation during 4 minutes, after which time themutarotation proceeds rapidly, giving an inflected curve with al6 C.N. Riiber and J. Mimaas, Ber., 1927, 60, [BJ, 2402; A., 1928, 47.G. F. Smith and T. M. Lowry, J., 1928, 666; A., 610.REP.-VOL. XXV. 98 ANNUAL REPORTS ON THE PROGRESS OF OHEMISTRY.maximum a t about 10 minutes. Assuming the third variety ofgalactose to be a form p, the mutarotation is represented bya- =+ p @-, and the rotatory power of p is of the same order asfor @-galactose. It is calculated that at equilibrium the proportionsof each form present are : a-, 286% ; 8-, 59.5% ; p, 12%.Almost general unanimity of opinion has been reached in regardto the formulation of simple sugars as pyranoses when these occuras free hexoses or pentoses or in the form of their normal glucosides.C. S.Hudson is exceptional, however, in opposing 18 this view andprefers to represent glucose and its a- and p-glucosides as five-atomring structures. He included in this category also p-mannose,lyxose, and a-methyl-lyxoside. F’rom optical rotation values,combined with an extension of the additive principle of opticalsuperposition, Hudson has computed rotational effects which aresingular in all the above cases. Since they do not yield expectedvalues on the above statistical basis, the above-mentioned sugarsand their glucosides are given a five-atom ring structure as distinctfrom the six-atom ring structure which is conceded for galactose,arabinose, xylose and a-mannose.In a further reply to this argument other authors l9 have reviewedthe statistical method of Hudson and have demonstrated thatin its application to the ring-structure of sugars it gives rise tofallacious deductions which are contrary to the chemical facts.It isshown that, whilst Hudson’s value a (for the rotational effect ofthat part of a sugar which involves the carbon atom of the reducinggroup) is approximately a constant for the glucose, galactose, xyloseseries yet quite another value for the constant is furnished by theseries mannose, rhamnose, lyxose. Indeed, if the latter series hadbeen the commoner or more accessible sugars, this singularity wouldhave been credited to the former series and statistical conformitywould probably have been conceded to the latter.It is suggested, however, that the lack of statistical agreementin the two series is associated with the cis-configuration of thehydroxyl groups in mannose, rhamnose, and lyxose.The new@-form of lyxose has been isolated, and this is in every respect com-parable with 8-mannose.Moreover, as a final argument, a-methyl-lyxoside (I) has beenmethylated to give the crystalline trimethyl lyxose (11). Thisyields a S-lactone (111), which undergoes oxidation to d-arabo-trimethoxyglutaric acid (IV) .2018 F. P. Phelps and C. S. Hudson, J . Amer. Chem. Soc., 1928, 50, 2049;A,, 991; 8ee also Ann. Repor&, 1926, 23, 78.W. N. Haworth and E. L. Hirst, J., 1928, 1221; A., 740.*O E. L. Erst and J. A. B. Smith, M., p. 3147ORUANIO CHEMISTRY.-PART I. 99(1.1 (11.1 (111.) (Iv*)Two isomeric triacetyl methyl-lyxosides are known, and fromthe properties of these it is evident that the normal methyl-lyxosideis a lyxo-pyranoside.21Crystalline g-methylfructoside has long been known and haslatterly been recognised as 13-methylfructopyranoside.The recentisolation of the crystalline a-form is a matter of interest and im-portance.22 The a- and p-methylfructosides are seen to stand in thesame relation to each other as a- and p-methylglucosides and allof them are six-atom ring forms or pyranosides. It is very satis-factory that the missing form has now been discovered and thatfructose in its general relationships is brought completely into linewith other sugars.The interconversion of a- and p-hexosides has hitherto beenpossible by digestion with methyl-alcoholic hydrogen chloride.Anovel method of some utility is now devised whereby @-forms areconverted with facility into a-forms in non-ionising solvents con-taining stannic chl0ride.~3 In the presence of titanium tetra-chloride, tetra-acetyl @-methylglucoside is smoothly converted intothe a-variety ; the completely acetylated sugars also undergo asimilar change, but this is accompanied by the introduction ofchlorine : for example, p-glucose penta-acetate is transformed intoa-chloroglucose tetra-acetate. The reaction appears to be a generalone.By the use of a very active form of silver chloride the inversion 24of a-chloroglucose tetra-acetate to the p-compound occurs in 8-10minutes. The specific rotation of the latter differs by about 30"from that calculated by Hudson.It is suggested that the principleof optical superposition is not valid in these cases and that Hudson'smethods of calculation should not be applied.The remarkable way in which hydrogen of the hydroxyl groupsin sugars may be replaced by thallium when employed as thalloushydroxide has again been illustrated.25 This react.ion is likely21 P. A. Levene and M. L. Wolfrom, J. Biol. Chem., 1928, '78, 525; A., 991.22 H. H. Schlubach and G. A. Schroter, Ber., 1928, 61, [B], 1216; A., 873.23 E. Pacsu, ibid., pp. 137, 1508; A,, 275, 1118.24 H. H. Schlubach, P. Stadler, and I. Wolf, ibid., p. 287; A., 398.25 R. C. Menzies and (Miss) M. E. Kieser, J., 1928, 186; A., 275100 ANNUAL REPORTS ON W E PROURESS OF CHEMISTRY.to prove of great value in effecting the protection of hydroxylgroups in difficult cases.By treatment with dilute alkali methylated augars are shown toundergo the Lobry de Bruyn and von Ekenstein transformation inmuch the same way as the unsubstituted sugars. For instance,tetramethyl glucose is convertible 26 into tetramethyl mannose,thus confirming the work reported last year on the identity of thering structure in each of these sugars.Elimination of phosphoric acid residues in sugars is accomplishedby the agency of bone phosphatase.By this procedure the a- andp-methylhexosidediphosphoric acids give rise to a- and p-methyl-hexosides and it is suggested by the authors that these are recog-nisable as a- and P-methylfructofuranosides. Thus it wouldappear that the sugar residue occurring in these compounds isy-fructose.27 Deductions based on magnitudes of specific rotationshould, however, be accepted with caution, and in the case underreview the authors would doubtless be well advised to confirm thisimportant conclusion by applying other and more certain methodsof diagnosis.Compounds having unsaturated sugar chains have long beenrecognised in the series known as the glucals.A sugar derivative,having an ethylene bond in the side chain of the ring, is nowavailable 28 in the compound a-tetra-acetyl glucoseen (V). Theproperties of this substance suggest a relationship to the transitionproducts between glucose and lignin or the compound from whichlignin arises biologically.A general study of the use of triacetyl glucose-1 : 2-anhydride (VI)has illustrated the facility with which this reagent 29 combines withhydroxylated compounds with the format'ion chiefly of p-glucosides.Ha *OAc 4',1 H * ? ! qHO*Y*H H*TGl H.7Ac0.F-H AcO$7*HH*(j*OAc I H*F*OAc 1 H*Q*OHH.7 - He?--CH,-OAc CH,*OH F(VII.)CH2(V.1 FI.1With phenol the reagent yields, however, a-phenylglucoside.A., 509.1214.A26 M. L. Wolfrom and W. L. Lewis, J . Amer. Chem. Soc., 1928, 50, 837;27 W. T. J. Morgan and R. Robison, Biochem. J., 1928, 22, 1270; A.,28 B. Helferich and E. Himmen, Bet-., 1928, 61, [B], 1826; A,, 1221.29 W. J. Hickinbottom, J., 1928, 3140ORGANIC CHEMISTRY .-PART I. 101comparison of the properties of the anhydride with those of Pictet’sa-glucosan (VII), which should be the parent substance of the above1 : 2-anhydrideY shows a certain disparity, since a-glucosan is saidto be capable of crystallisation from methyl alcohol, whereas thesubstance (VI) combines readily with methyl alcohol to yield theglucoside. The structure allocated to (VI) is, however, supportedby the observation that methylation of the glucoside to which itgives rise, followed by elimination of the glucosidic group and ofthe acetyl residues, leads to the recognition of a %methyl glucose.If a-glucosan has the constitution (VII), its properties present someunaccountable difficulties.More than usual interest attaches to Fischer and Zach’s anhydro-glucose, which is recognised as the 3 : 6-glucose anhydrideVHY ?HQHCH0.C- C -C--C-CH,H I A B IL-0-Jand, unlike glucose, has the property of restoring the colour ofSchif€’s reagent. The constitution here assigned is supported by anew method30 of preparation of the anhydro-sugar from glucosemonoacetone : the latter yields a di-p-toluenesulphonyl derivativewhich, on treatment with 1 mol.of alkali, is converted into acrystalline mono-p-toluenesulphonyl derivative of the above.Further treatment with alkali effects the removal of the remainingtoluenesulphonyl residue, yielding anhydroglucose-monoacetone , acrystalline substance. Elimination of the acetone residue is broughtabout by dilute acids and the free anhydro-sugar is then isolated inthe crystalline state.Among several unexpected results is the observation 31 thata-methylmannoside, which is known to have the pyranose structure,is converted into the diacetone compound of a-methylmanno-furanoside by the agency of 1% of hydrogen chloride in presence ofacetone.It would appear that under these conditions the six-atomring form is transformed into a five-atom ring form in order toaccommodate the two acetone residues which simultaneously con-dense with the four exposed hydroxyl groups. It is known that thiatype of change occurs readily in the free sugar, and the a-glucosideform is now seen to respond to the same structural change. Theacetone derivatives of xylose have long defied constitutional in-3O H. Ohle, L. von Vargha, and H. Erlbach, Ber., 1928, 61, [B], 1211; A.,871; K.Freudenberg, H. Toepffer, and C. C. Andersen, ibid., p. 1760; A,,1223.31 P. A. Levene and G. M. Meyer, J . Biol. Ohern., 1928, 78, 363; A,, 992.Compare also ibid., pp. 1208, 1870; A., 871, 1220102 ANNUAL REPORTS ON THB PROGRESS OF UHEMISTRY.vestigation, and speculation as to the possible occurrence in thesecompounds of a four-atom ring type of xylose has been current. Itis now demonstrated 3, that xylose-monoacetone is a derivative ofxylofuranose, since methylation yields a dimethyl derivative which,on hydrolysis and oxidation, passes into a 3 : 5-dimethyl yxylono-lactone (X). This is converted into the trimethyl y-xylonolactone(XI) which had previously been identified, and is now shown toyield a crystalline phenylhydrazide. The hydration curves of thelactone and its acid are given and are compared with those from theisomeric trimethyl S-xylonolactone.Moreover, oxidation of the7-lactone with nitric acid yielded d- dimet hox y succinic acid.It is shown that xylose-monoacetone must have the constitution(IX) and consequently that the diacetone is preferably representedby (VIII).Di-, Tri-, and Tetra-saccharides.-F'irst in importance among thepublications of the year on this topic is the communication byPictet and Vogel on the synthesis of sucrose. It will be remembered(Ann. Reports, 1916,13,90 ; 1920,17,65) that twelve years ago theadvance made in the study of sucrose led to the overthrow of theconstitution which had, up to that time, been accepted. Up to1916 it had been tacitly assumed that, since invert sugar consists ofa mixture of glucose and fructose, the forms of the sugars thusisolated were of necessity the forms in which they are combined inthe disaccharide.It was then founds that methylated sucroseshowed no inversion of sign on hydrolysis, and that, whilst the usualform of tetramethyl glucose could be isolated from this product, themethylated ketose displayed a rotatory power and a reactiontowards permanganate which at once implied its structural relation-*' W. N. Haworth and C. R. Porter, J., 1928, 611 ; A,, 609.*I W. N. Haworth and J. Law, J., 1916,109, 1314OWANIC OHEMISTRY .-PART I. 103ship to y-glucose, and the methylated ketose component was there-fore recognised as tetramethyl y-fructose.Although only a pro-visional formula could then be allocated, sucrose was thus shown tobe a disaccharide in which the normal form of glucose was linkedwith a y-form of fructose. It was reported that “ a new formula forsucrose is thus indicated which accounts not only for the extremeease with which the disaccharide is hydrolysed, but also for theprevious failure to effect its synthesis.” Four years later theisolation of the actual fructose component as a dextrorotatorytetramethyl y-fructose was announced,a and preliminary attemptsto ascertain its ring structure were made. At that time all suchattempts to elucidate the structure of y-sugars were doomed tofailure so long as the normal sugars were rigidly accepted as butylene-oxide forms. This fallacy underlying the whole of sugar chemistrywas exposed and glucose and fructose were shown to have a six-atom ring structure.36 It was then made abundantly clear that theolder formulations of sugars as five-atom ring compounds had beenerroneously applied to these substances, and that this discardedformula actually applied to the series of y-sugars.The six-atomring sugars, related to pyran, were described as pyranoses: they- or five-atom ring sugars were related to furan and designatedfuranoses.Sucrose was then definitely a complex containing each of thetwo kinds of sugar rir1g.~6 The glucose component was present asglucopyranose, and the fructose component as fructofuranose, andin sucrose they were combined by mutual linking through theirreducing groups.It became evident that if a suitably substituted derivative ofthe dextrorotatory variety of fructose could be prepared (that is,a fructofuranose as distinct from the lavorotatory fructopyranose),a synthesis of sucrose might be attempted with definite prospect ofsuccess.After an exhaustive search such a derivative has now beenprovided in the dextrorotatory tetra-acetyl y-fructose which occursas a by-product to the extent of 3% in the preparation of the normalor pyranose form of laevorotatory tetra-acetyl fructose. It is nowcommunicated by Pictet and Vogel37 that, when this tetra-acetyly-fructose (XII) is condensed with tetra-acetyl glucose (XIII) in achloroform suspension of phosphoric oxide, the genuine octa-acetylsucrose (XIV) is formed.By the elimination of the acetyl residuesW. N. Haworth, J., 1920,117, 199.?,ti Ann. Reports, 1925, 22, 83; 1926, 23, 74; 1927, 24, 66.88 Avery, Haworth, and Hirst, J., 1927, 2308; A., 1927, 1057.IW C-t. rend., 1928,186, 724; A., 510; Helv. Ch,h. Ack, 1928,11, 436;A., 741104 AXN'UAL REPORTS ON THE PROGRESS OF OHEMISTRY.the latter substance gives rise to sucrose identical in melting pointand optical properties with the natural product.It appears that sucrose, when crystallised from aqueous methylalcohol,% gives a modification (B), m. p. 171", whereas from mostother organic solvents the sucrose (A) separates, m. p. 185". Allthe properties of A and B are identical except the melting point, andthe conversion of B into A is instantaneous in water.The differencesin the two varieties seem to be purely physical.In another communication by the same authors 39 an account isgiven of the condensation of the usual lzevorotatory variety oftetra-acetyl fructose with tetra-acetyl glucose, yielding a disac-charide in which both the glucose and the fructose component arepyranose forms. The product, an isomeric form of sucrose describedas sucrose C, is crystalline and lsvorotatory and differs widely inproperties from natural sucrose. It is evident, therefore, that theearlier work on the constitution of natural sucrose is supported bythe synthetic experiments of Pictet and Vogel.The constitution assigned in previous work40 to melibiose hasnow been confirmed by a structural synthesis 41 of this disaccharide,which derives its importance from the circumstance that it occursas an essential component of raffiose.It was noted as a preliminarythat a-bromoacetogalactose, when heated with phenol and quinoline,yielded a mixture of a- and p-phenylgalactosides. The formationof this or-galactoside gave promise that, since melibiose has alreadybeen assigned an a-configuration, condensation of the abovea-bromoacetogalactose (XVI) with 1 : 2 : 3 : 4-tetra-acetyl p-glucose(XV) would yield a mixture of octa-acetyl glucose-a- and -p-galactosides from which octa-acetyl melibiose might be isolated.This has indeed proved to be the case. The octa-acetyl glucose-38 Gompt. mnd., 1928,186, 901; A., 1223.39 Idem, ibid., p. 905; A,, 1223.40 W.N. Haworth, J. V. Loach, and C. W. Long, J., 1927, 3146; A., 1928,166.R. Helferich m d H. Bredereck, Annalen, 1928, 466, 166ORUANIU (33HEMISTEY .-PART I. 105a-galactoside (XVII) was found to be identical with a specimen ofb%J0 (XVII.)The application of a similar procedure involving the condensationof (XV) with acetobromoarabinose led to the isolation of a hepta-acetyl derivative of the naturally occurring vicianose. Again, thelinking of cellobiose with gentiobiose has been accomplished with theformation of a synthetic tetrasaccharide of definite constitution,although this tetraglucose complex has not been identified with aknown natural product.A method of synthesis of raffinose is reported *2 depending on thefusion of sucrose and galactose, a procedure which is said to give a1% yield of the trisaccharide.A new tetrasaccharide, which isdesignated maltotetrose, has also been obtained by condensation ofhepta-acetyl maltose, and other new sugars of the trehalose type 43are obtainable by the similar treatment of p-tetra-acetyl hexoses intoluene containing zinc chloride to which phosphoric oxide issubsequently added.From the products of fermentation of glucose or fructose, inpresence of phosphate, a phosphoric ester of a disaccharide has beenisolated. This loses a phosphoric acid residue by the agency ofbone phosphatase and liberates trehalose,M thus affording a synthesisof thh non-reducing disaccharide.Other disaccharides which are entirely synthetic in origin have beenprepared by Freudenberg45 and his co-workers.They have4a H. Vogel and A. Pictet, Hdv. Chim. Acta, 1928, 11, 898; A., 1224.4* H. Vogel and H. Debrowska-Eurnicka, ibid., p. 910.44 R. Robison and W. T. J. Morgan, Biochem. J., 1928,22,1277; A., 1286.K. Freudenberg, H. Toepffer, and C. C. Andersen, loc. cit. ; K. Freuden-berg, A. Wolf, E. Knopf, and 8. H. Zaheer, Ber., 1928,61, [B}, 1743 ; A., 1222.D 106 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.utilised for this purpose glucose-6-bromohydrin (XVIII). Bysuitably protecting two hydroxyl residues in this compound, e.g.,by an acetone group or two acetyl groups, two hydroxyl positionsremain available for condensation, as in the compound (XIX).This has been combined with acetylglucosido-halides of severaltypes, and derivatives of new dihexoses are made available.Other similar experiments have led to the isolation of synthetictrisaccharides of the type (XX).H,Brr- 1Experiments of great interest and value have also been com-municated by the same author 46 dealing with the rates of hydrolysisof disaccharides, glucosides, and sugar-acetones.The figures forthe inversion constants of most of the known compounds are thusmade available for comparison.PoZysacchccrides.-Experiments on the constitution of inulin haverevealed 4’ the points of junction of contiguous residues of fructo-furanose (y-fructose). Methylation of inulin yielded a ‘‘ trimethylinulin,” which was hydrolysed to give 3 : 4 : 6-trimethyl fructo-furanose (XXI). This yielded a crystalline osazone and underwentoxidation to a 3 : 4 : 6-trimethyl fructuronic acid (XXII).Thelatter was converted, after methylation, into the same crystallineamide as that which had previously been obtained from tetramethylfructofuranose derived from sucrose. The presence of the sameform of fructose in both sucrose and inulin was thus confirmed.Oxidation of the trimethyl fructuronic acid with acid permanganateled to the isolation of the crystalline d-2 : 3 : 5-trimethyl y-arabono-46 I(. Freudenberg and others, Ber., 1928,61, [B], 1735; A,, 1222.4 7 W. N. Haworth and A. Learner, J., 1928, 619; A, 610ORQANIC aHEM1STRY.-PART I. 107lactone (XXIII).to Z-dimethoxysuccinic acid (XXIV,.This is characterised by its oxidative degradationQ02H Hi@?----H.F-Ap W H+ - M e O * F y bH*$?*OMe ICH,*OMe(=.I (=I.) (XXIII.)From these results it is shown that the union of the y-fructoseresidues in inulin occurs through the positions 1 and 2 in each sugarchain. This structural scheme is indicated below.On this view, the simplest expression for inulin is the followingformula,O-------CHCH,*OHCH(OH)-(!H-OHwherein two fructose anhydride units are mutually linked.Valuesfor molecular weight determinations of triacetyl inulin have beenreported which correspond to hexa-acetyl difructose anhydride?and moreover the complexity of inulin in liquid ammonia isseen to be in agreement with this simple expression. An endeavourto synthesise a compound having this simplified structure hasrecently been made.48On the other hand, direct determinations of the molecular weightof inulin in water indicate the presence of a more complex molecule,containing at least 24 anhydro-fructose ~nits.4~ Hydrolysis of theinulin in warm water is shown, however, to be progressive andafter 28 minutes the above value is diminished to half and thereducing power of the solution increases as the molecular weightdiminishes.Under other conditions inulin breaks down readily togive a “laevulin” corresponding to six or eight C6Hlo05 units. EarlierH. H. Schlubach and H. Elmer, N U t U ? ~ i 8 8 . , 1928, 16, 772; Ber., 1928,81, [ B ] , 2358; A., 1221.H. D. K. Drew and W. N. Haworth, J., 1928, 2690; A,, 1360108 ANNUAL REPORTS ON THE PROGRESS OF UHEWSTRY.determinations of molecular weight have been commented on inprevious Reports.Others are now communicated 50 correspondingto complexes varying from (C,H,,O,), to (C,H,,O,),. The formervalue is obtained by the cryoscopic method with aqueoussolutions of inulin. But by heating inulin acetate (at temperaturesvarying from 250" to 290") in presence of tetrahydronaphthalenethe values diminish progressively to the above minimum. Whetherthis diminution corresponds to dissociation or decomposition is aproblem remaining for future decision.Inulin itself, heated with glycerol a t 120"/15 mm., is found61to undergo definite structural change to a trilaevulosan, and at aslightly higher temperature a dilaevulosan which reduces Fehling'ssolution is reported to be the product. It is doubtful, therefore,whether the transformations of inulin acetate already discussed canbe interpreted as a simple dissociation or depolymerisation of inulinto its " structural unit."Recent experiments 5, have shown that the methylation ofcellulose a t 20" with methyl sulphate gives rise to a trimethylcellulose which differs essentially from the specimens previouslyisolated. This is soluble in chloroform, tetrachloroethane, andacetic acid, giving clear viscous solutions, and is quite insolublein water.The yield is 93% of trimethyl cellul0se,5~ and this onhydrolysis furnishes 2 : 3 : 6-trimethyl glucose in a yield of 91%,but in no case could a trace of tetramethyl glucose be isolated.This establishes the structural identity of all the glucose units incellulose.When it was heated with hydrogen chloride in ethersolution, the trimethyl cellulose yielded 1 -chloro-2 : 3 : 6-trimethylglucose, which, in contact with sodium, was transformed into acompound described as 2 : 3 : 6-trimethyl glucose anhydride.According t o 6ome earlier theories such a compound would beexpected to be the '' structural unit " of trimethyl cellulose. Butthe complete dissimilarity of the anhydride with the latter provesthat cellulose cannot be regarded as a unimolecular glucose anhydride.There appeared to be no grounds for the suggestion that trimethylglucose anhydride would undergo reversible interconversion intotrimethyl cellulose. The author thus falls back on the older tradi-tional view that cellobiose is the true breakdown product of celluloseand is not a reversion product. It is held to be probable that cello-biose residues comprise the essential units of cellulose and that these60 H.Pringsheim and I. Fellner, Annalen, 1928, 482, 231; A,, 742; H.Pringsheim and J. Reilly, Ber., 1928, 61, [B], 2018; A., 1226.61 H. Vogel and A. Pictet, Helv. Chim. Acta, 1928, 11, 216; A., 276.62 H. Urban, Cellulosechem., 1926, 7, 73; B., 1926, 631.63 K. Freudenberg and E. Braun, Annalen, 1928, 480, 288; A,, 399; I(.Freudenberg, ibid., 461, 130 ; A,, 743ORGANIC CHEMISTRY .-PART I. 109are joined in a chain which is united throughout by ordinary co-valency links. A similar view has also been independently expressedby another worker,a and since the cellobiose structure has been finallydetermined it follows that this picture of the cellulose constitutionis given by the scheme :CH,.OH H OH CH,*OH H OHA re-interpretation of the crystal lattice of cellulose is held tosupport this view entirely.55 The elementary cell has the dimension10.3 8.along the fibre axis in ramie cellulose, and the other dimen-sions of the cell are given as 7.9 and 8 - 7 8 . This accommodatesfour glucose residues, and the first dimension is in close agreementwith two units of the six-atom ring form of glucose as arranged incellobiose. It is deduced that the expression formulated abovereceives strong support, the cellobiose residues being oriented in thedirection of the fibre axis and mutually linked by glucosidic oxygen.Forty or more glucopyranose residues are regarded as being joinedby p-glucosidic linkings in this way.It is suggested that thisstructural scheme is in harmony with the established chemicalproperties of cellulose and is compatible with and affords a readyexplanation of the behaviour of cellulose during esterification andswelling.Other interpretations of the X-ray data commented on in lastyear's Report66 anticipated in the main these features of thecellulose structure. In one essential detail, however, they differed,inasmuch as the occurrence of the 3.4 A. line offered difficulties inaccepting the cellobiose type of linking of the glucopyranose units.Instead of the union being 1 : 4-1 : 4 between contiguous glucoseresidues, the earlier authors preferred to arrange the linkings in theorder 1 : 4 4 : 1.But the ordered linking of glucopyranose unitsin long chains arranged side by side is the main conclusion of bothinterpretations of the X-ray diagrams of cellulose. Doubtless thesedifferences will be adjusted by further experimental work, whichshould be largely chemical in its scope.The experiments and conclusions of Hess 57 are entirely opposed6 6 K. H. Meyer and H. Mark, Ber., 1928, 61, [B], 593, 1936; A., 621; 2.W. N. Haworth, Helv. Chim. Acta, 1928, 11, 534.angew. Chem., 1928, 34, 935.Ann. Reports, 1927,24, 82.67 K. Hess and C. Trogus, Bey., 1928, 61, [B], 1982110 d14NUAL REPORTS ON THE PROGRESS OF CHEMISTRY.to these conceptions of the cellulose structure. This author con-siders that the above interpretations of the X-ray diagrams areinvalid in so far as they assume the presence of cellobiose residuesor of glucosidic disaccharide linkings. He urges that the dis-entanglement of the structural factors of cellulose in the X-raydiagram is rendered difficult by ignorance of the correspondingdiagrams of carbohydrates of known structure. The diagrams of anumber of such products are now given, and it is claimed that oneof these, a biosan (C12H20010) which the author has isolated fromcellulose, differs but little from cellulose in the mass distributionof its constituent atoms. The effect of substitution of groups onmass distribution is intimately considered. On this view a moreappropriate conception of the cellulose structure is that of anarrangement of masses of the biosan or glucosan complex united bydirected associative forces.In the past, difficulty has been encountered in the preparation oftriacetyl and trimethyl starch. Small yields of products wereobtained, in some cases after very prolonged and repeated treatmentswith the substituting agents. This led to serious doubts whetherthe final products were genuine derivatives of the original starchor of a portion only of the polysaccharide which had been segregatedduring these observations.In a new series of experiments * starch has been brought into acondition in which it is more susceptible t o reagents. Starch pastewas quantitatively precipitated with alcohol and gave a powderwhich retained the properties of the substance. On acetylationthis gave a 96% yield of a triacetyl starch. De-acetylationfurnished a product having the properties of regenerated amylose.Methylation of the triacetyl starch under new conditions gaverise to a trimethyl starch in 89y0 yield after only six treatmentswith the reagents. This is compared with the results of earlierworkers, who obtained a 25% yield after 24 methylations. It isshown that a high yield of 2 : 3 : 6-trimethyl glucose was obtainablefrom the methylated starch. The authors were unable to confirmearlier claims that starch undergoes preferential methylation to ahomogeneous dimethyl starch, and the structural views based on suchan observation are rendered doubtful. The probability that starchis largely composed of conjugated a-glucopyranose units is dis-cussed and alternative formula are considered.W. N. HAWORTH.E. L. HIRST.66 W. N. Haworth, E. L. Hirst, and J. I. Webb, J., 1928, 2681; A., 1360ORGANIC CHEMISTRY .-PART 11. 111PART II.-HOMOCYCLIC DMSION.IN this contribution the aim will be to bring up to date the accountof those subjects which, having been discussed in one or more of thecorresponding Reports since 1923, have subsequently made markedprogress. Other topics will be left aside that they may be dealt withby a future Reporter in a more connected manner than would bepossible if they were discussed at the present stage.Large Carbon Rings.(Continued from Ann. Reports, 1926, 23, 112-119.)Since the date of the Report referred to, a detailed examination byRuzicka and his collaborators of the distillates obtained fromthorium azelate and sebacate, from yttrium nonane- and decane-aw-dicarboxylate, from thorium pentadecane-, octadecane-, andnonadecane- aw-dicarboxylate, and from yttrium eicosane- andoctacosane-aw-dicarboxylate, has led to the isolation of a large numberof new cyclic and open-chain compounds. The cyclic substancesinclude hydrocarbons with rings of 16, 18, and 30 methylene groups,monoketones with rings containing 19, 20, 21, 29, and 30 carbonatoms, and a series of symmetrical cyclic diketones with 16,18,20,22,and 30 carbon atoms in their rings. The last group of substances areCO of the type (CH,)n<co>(CH2)n, of which the only previously knownexamples are the cyclobutane- and cyclohexane-diones.The following table contains a list, complete from C,, to Cm, of thesaturated, unbranched, cyclic hydrocarbons, monoketones, anddiketones which have been described up to the time of writing. Thefigures represent the m. p.'s of the compounds ; those in Roman typerelate to compounds the preparation of which is referred to in theReport for 1926, and those in italics to those which have beenprepared since that time.etc. etc.