KENNER : THE INFLUENCE OF NITRO-GROUPS, ETC. 2717CCLIV.-The InJuence of Nitro-groups on the Re-activity of Xubstituents in the Benzene N.zm?eus.By JAMES KENNER.ALTHOUGH the enhanced reactivity of substituents in the ortho-or para-position relatively t o a nitro-group has been a familiarfact since i t was first observed by Pisani in the case of picrylchloride (L4~znaZen, 1854, 92, 326), little insight appears to havebeen gained into the nature of the processes which result in thedisplacement of such substitaents by others.The ‘‘ activating ” influence of the nitro-group is in direct con-trast, with its retarding, and at times inhibitive, effect on otherreactions. Whilst, however, the nitro-group shares its power ofsteric hindrance with other substituents of large molecular volume,quite irrespective of whether they are meta- o r ortho- and para-directive, this is not the case where the opposite effect is concerned.I n general, only meta-directive groupings can render substituentsin the ortho-position mobile; for instance, Schopff and his pupilsshowed that the carboxy-, aldehydo-, sulphonic, and cyano-groupseach conferred mobility on a bromine atom in the ortho- o r para-position (Bey., 1889, 22, 900, 3281; 1890, 23, 3440, 3445, 3450;1891, 24, 3771, 3785, 3808).The presence of a nitro-group wasnot essential, but its influence was much more powerful than thatof the other groups named, and a t least two of these, as comparedwith a single nitro-group in the ortho- or para-position (compareLobry de Bruyn, Rec.trav. chim., 1894, 13, lOl), were requiredt o render a substituent mobile. Ortho-para-directive substituentsappear, however, to increase the mobility of substituents in themeta-position in some1 cases, Thus Laubenheimer showed that,whilst the conversion of o-dinitrobenzene into o-nitroaniline bymeans of alcoholic ammonia requires ten weeks a t the ordinarytemperature for its completion, the corresponding reaction in thecase of 5-chloro-1 : 2-dinitrobenzene is complete within five days, andleads to the formation of 5-chloro-2-nitroaniline (Ber., 1876, 9,1826; 1878, 11, 1156).Similarly, Tiemann found that o-chloro-pnitrobenzaldehyde iseasily converted into o-chloro-p-anisaldehyde by the action ofsodium methoxide, pnitrobenzaldehyde being stable in these, circum-stances (Ber.., 1891, 24, 709).Attention has apparently not beenpreviously directed to this phase of the problem, and i t is thereforea t present not possible to indicate how far this influence of thechlorine atom is general, or shared by other ortho-para-directivesubstituents.VOL. cv. 8 2718 BENNER: TEtE INFLUENCE Ol? NITRO-GfiOUPS ON TBEThe evidence as it stands, however, suggests a close connexionbetween the phenomena a t present under discussion and those ofdirective substitution ; and, indeed, the various facts just enumer-ated find a ready explanation in terms of the views advocated byFliirscheim (,I. p. Chem., 1902, [ii], 66, 321; 1905, 71, 497; 1907,76, 165, 185; T., 1909, 95, 718; 1910, 97, 84).Thus it is clearfrom the diagram that a nitro-group in the ortho- or para-position,and a chlorine atom in the meta-position, each weaken the attach-ment of tho substitaent X t o the nucleus, rendering it moremobile :X xand that their influence is cumulative. Conversely, i t may beexpected that the effect of, for instance, a nitro-group in the nieta-position, or a chlorine atom in the ortho- or para-position, wouldbe a detrimental one. Quantitative experiments on the velocitiesof such reactions are therefore desirable, in order to discoverwhether these influences exist, and, if so, whether they are parallelwith Lhose observed, for example, by Staudinger and Kon ( A ~ ~ n a l e ? ~ ,1912, 384, 38), which are also capable of interpretation in thelight of Flurscheim’s views.It must be remarked, however, that the attachment of thenitregroup to the benzene nucleus has been frequently assumed t obe a comparatively strong one.Thus Kaufmann (“Die Auxo-chrome,” Stuttgart, 1909, p. 76) refers t o the aliphatic nature ofthe nitro-compounds, as illustrated by the reactivity of the chloro-nitro-compounds, and later (p. 93) attributes the formula I tonitrobenzene :Aux NH21) 1)__.._ O=N=O. ..../\It II /\ . ...-\J1-*-1 I\/(1.1 (11.) (111.)I)\/(‘Ii iThe adoption of this formula, preceded as it is (pp. 80,85) by theformihe I1 and I11 for benzene with the “ ideal auxochrome ” asa substituent, and for aniline, involves certain difficulties. ThusKaufmann is a t once driven to attribute something in the natureof nu amphoteric character to the phenyl residue, and to assume“ dass die Absiittigung der Partialvalenzen bei Antiauxochromenpolar zu der bei Auxochromen stratthat.” Further, it is probablREACTIVITY OF SUBSTITUENTS IN THE BENZENE NUCLEUS.2719that the presence of raidual affinity on the ortho- and para-carbonatoms, as represented by Kaufmann, would tend to promote thesubstitution in the ortho- and para-positions which actually takesplace. I€, then, Kaufmann’s formula for nitrobenzene be accepted,substitution might be expected to take the same course in this case,but this is not so. The strongest evidence in favour of Kaufmann’sview is probably supplied by the work of Meisenheimer (Annalen,1902, 323, 222; 1907, 3155, 249), who examined very thoroughlythe action of potassium methoxide on various nitro-compounds, andexplained his results in terms of Thiele’s theory of partial valencies,making the same assumption in regard to the nitlro-group a8 thatmade by Kaufmann.It will, however, be suggested that theseresult3 are not; irreconcilable with the opposite point of view. Noris Kaufmann’s formula, in the author’s opinion, a necessary conse-quence of Thiele’s theory, which is based on a conception of thenature of the double bond uniting two atoms of the same element.Where a double bond connects two different elements, i t is conceiv-able that one of these, that of greater residual affinity, may claimthe whole of the available affinity of the other.This is especially soas regards the nitro-group, where the single nitrogen atom isattached to two oxygen atoms, and if such were actually the case,Fliirscheim’s formula would represent nitrobenzene more accurately.Further, Werner’s explanation of the activity and inertia of therespective halogen atoms in triphenylmethyl bromide and tribenzoyl-bromoinethane (Ber., 1906, 39, 1282), symbolised by ttlie formulae :(C,H,),~C---Br, (C,H,--CO),ZE-C--Br,appears to lend support to the following formula for bromotrinitro-methane, the bromine atom of which is also inert (Meisenheimerand Schwarz, (Ber., 1906, 39, 2544):( N0,)3EC-Br.The evidence supplied by these considerations is therefore also infavour of Fliirscheim’s views.The following instance appears to show that the activating influ-ence of nitro-groups may act through a chain of atoms, which donot all form part of an aromatic nucleus.Whilst o-hydrazino-benzoic acid only passes over into 3-keto-l : 3-dihydroindazole whenit is heated alone to 220-230° o r with boiling phosphoryl chloride(E. Fischer, Ber., 1880, 13, 681; E. Fischer and Seuffert,, ibid.,1901, 34, 795), methyl 3 : 5-dinitro-2-hydrazinobenzoate cannot beisolated as a product of the interaction of methyl 1-chloro-3 : 5-di-nitrobenzoate and hydrazine hydrate, but is immediately convertedinto 5 : 7-dinitro-3-Leto-1 : 3-dihyd~oindazole. The following schemeshows that this reaction is also in accordance with the views justdiscussed :8 P 2720 KENNER: THE INFLUENCE OF NITRO-GROUPS ON THE.HNH-N'\H--+N/\0 0NH-NHI IO,Nf\ -eoIt might, however, be objected that, whereas Flurscheim hasapplied his views to explain the replacement of a hydrogen atom inthe meta-position with respect to a nitro-group, they are hereemployed to explain the replacement of substituents in the ortho-or para-positions. This consideration is of great importancebecause it shows that, whatever the nature of the substitutiveprocesses involved in the two cases may be, they must be essent!iallydifferent in character.Such a conclusion is supported by the factsthat the steric hindrance of the negative groups in the ortho-posi-tion is to a large extent, although not completeIy, overcome, andthat the nitro-group, which is the substituent most effective inexerting this hindrance, is also the most effective in promoting thecondensations of the type under consideration.It seems clear thatthe activating group is itself concerned in the reaction. I n viewof the increasing favour with which the addition theory of chemicalreaction is now received, and of the fact that the nitro-group isespecially prone to take part in the production of molecular com-pounds, it seems natural to assume t,hat the formation of sucha compound is the first, step in the process a t present under dis-cussion: and, indeed, when an aromatic amine, for example, reactswith an active chloronitro-compound in alcoholic solution, an intensecoloration is a t once produced which differs from that imparted toalcohol by t'he final condensation product, and is probably to beatltributed to the formatdon of an additive compound.This sug-gestion is not a novel one, and appears to have been originallymade by Bamberger (Ber., 1900, 33, l02), who, however, expressedno opinion as to the nature of the process by which the intermediateproduct passed over into the final condensation product. Bam-berger actually isolated a molecular compound of picryl chlorideand a-naphthylamine, from which picryl-a-naphthylamine was easilyformed ; whilst Sudborough and Picton isolated similar additivecompounds of 4 : 6-dichloro-1 : 3-dinitrobenzene and of 2 : 4-dichloro-1 : 3 : 5-trinitrobenzene, each with one' molecular proportion of a- or8-naphthylamine (T., 1906, 89, 583).Sudborough and PictoREACTIVITY OF SUBSTITUENTS IN THE BENZENE NUCLEUS. 2721suggested the following scheme to explain the mechanism of thereaction (compare also Richter, Ber., 1875, 8, 1419):C1 NH*C,,H7N H*C,,H:,02Nf‘,N02c1\/ \/NO2 NO2 NO2whilst Borsche has recently (Annalen, 1911, 386, 356; 1913, 402,81 ; compare also Meisenheimer, ibid., 1902, 323, 218) advocatedthe following expression of the reaction between 4-chloro-1 : 3-di-nitrobenzene and ethyl sodiomalonate * :Lapworth (P., 1903, 19, 123; compare Meisenheimer and Patzig,Ber., 1906, 39, 253; Lobry de Bruyn, Rec. trav. chinz., 1904, 23,60) suggested the formation of the same type of intermediateproduct’, as a result of the passage of the substituting radicle froman u- t o a y-position (T., 1898, 73, 445):Br OHBr \ / n T r O HIt is obvious that the whole question depends on that of the con-stit’ution of the molecular compounds first formed.The fact that* 4 : 6.Dichloro-1 : 3-dinitrobenzene furnishes first ethyl 5-chloro-2 : 4-dinitro-phenylmalonate, only reacting with a further molecular proportion of cthyl sodio-malonate with much greater difficulty, and this unequal reactivity is considered tobe “an experimental proof” of Kekul6’s beazene formula, as against Thiele’s or thecentric formula. I t appears exceedingly doubtful whether this evidence is asvaluable as that afforded by the work of Marckwald (Anncclesi, 1894, 279, 5 ) , for i tis hardly convincing that evidence, which is interpreted with the assistance ofThiele’s theory of partial valencies, should be used to disprove a formula whichfollows from an application of this theory to KekulB’s formula.Further, it may be remarked that the formula of ethyl sodiomalonate used byBorsche is not the one generally adopted, and to this extent his choice of anillustration for his theory is not a happy one,Also, according to Borsche, the chlorine atom in the compound nnder discussionis not loosely held.In this case i t is difficult to account for the fact that Ullmann’ssynthesis of diphenyl derivatives by heating the haloid benzenes with copper powderis only applicable to chloro-compounds mhcn they contain nitro-groups in theortho- or para-position to the halogen atom (Ullmann and Bielecki, B&. , 1907, 34,2177)2722 KENNER: THE INFLUENCE OF NITRO-GROUPS ON THEtetranitromethane and other aliphatic nitro-derivatives may formsuch compounds (Werner, Ber., 1909, 42, 4328) appears to bedecisive evidence against a formula of the type postulated bySudborough and Picton, and by Borsche, whilst Lapworth’s formulasupplies no explanation of the intense colour phenomena observed.The formula (IV) suggested by Werner (Zoc.c i t . ) is seen to be(IV.) (V.)closely related to that (V) adopted by Hantzsch for the nitro-anilines (Ber., 1910, 43, 1669), and is in harmony with theobservations of Sudborough and Beard (T., 1910, 97, 779), thatsubstituents which tend to increase the auxochromic effect of theamino-group always tend to increase the depth of colour of theadditive compounds of trinitrobenzene with amines (compare alsoGreen and Rowe, T., 1912, 101, 2446; Meldola and Hewitt, T.,1913, 103, 884; Meldola and Hollely, this vol., p.413). No attemptappears to have been made to institute a quantitative comparisonby an examination of the absorption spectra of the two series ofcompounds, but i t seems to the present author that this point needsattention before tho alternative formula (VI), hinted a t by Pfeiffer(Ammlen, 1914, 404, 13), is accepted:R-NO, C,H,NH,.Meanwhile, Werner’s formula has been reta’ined in the followingattempt to depict in outline the course of the reaction by whichan arylamino-radicle displaces a mobile substituent X :VI.)NHRAs a consequence of the formation of the molecular compound, acertain amount of unsatisfied valency may be expected momentarilyto exist, as shown, on the carbon atom to which X is attached, anREACTIVITY o ~ ’ SUBSTITUENTS IN THE BENZENE NUCLEUS.2723may either be absorbed by X or engage a portion of the residualaffinity of the spatially proximate nitrogen atom. Of the twoalternatives, the latter, represented by the third formula in thescheme, seems the more probable because it enables the system, bya subsequent rearrangement in the manner indicated, to attain acondition of smaller potential energy.The process by which the nitro-group facilitates the condensationis thus considered to consist in rendering i t intra- rather thaninter-molecular by bringing the amine molecule into a positionfavourable to the formation of a fivemembered ring (or, if Pfeiffer’sformula were adopted, of a six-membered ring).It is this featurethat differentiates the views now put forward from those ofLapworth ; whilst, in the author’s opinion, the supporters of theremaining theory are confronted with the dilemma that the sterichindrance of groups is probably explicable in a manner substan-tially similar to that by which they seek to explain the oppositeeffect. Thus Henry’s theory, that esterification depends on theformation of the intermediate compound (VII), has been extendedHO OR/OH RC-OH\OEt(VII.)by Wegscheider to the case of tho phenols, i t being assumed thatthe sterically active group X prevents the formation of the inter-mediate additive compound (VIII) (Monatslt., 1895, 16, 140 ; com-pare also Davies, T., 1900, 77, 33).The theory, that this is theusual mode of formation of substitution derivatives of phenol, hasbeen current for a considerable time, and is strongly supportedby experimental evidence recently adduced (Meyer and Lenhardt,A?c?tnle?t, 1913, 398, 51 ; Meyer, Irschich and Schlosser, Ber., 1914,47, 1741 ; compare Auwers and Michaelis, ibid., p. 1275).I n this connexion attention may be drawn to the behaviour ofpicramide. This aniine furnishes a reddish-brown explosive potass-ium salt (Green and Rowe, T., 1913, 103, 513), but is attacked byacetic anhydride only in the presence of sulphuric acid * (Witt andWitte, Ber., 1908, 41, 3092; compare Paal and Benker, Ber., 1899,32, 1251; Paal and Hartel, ibid., p.2051; Meldola and Hollely,this vol., p. 410). Similarly, 5 : 7-dinitro-3-keto-1 : 3-dihydroindazole* It is noteworthy that sulphuric acid has also been found to be effective infacilitating the esterificdtion of acids which remain unaffected under the nsualconditions (Wegscheider, Ber,, 1895, 28, 3128 ; Kenner and Mathews, this POI.,p. 2474)2724 RENNER: THE INFLUENCE OF NITRO-GROUPS ON THEfurnished a rnoiioacetyl derivative (IX), and a deep brown explo-sive disodium salt (X), whereas its 2-phenyl derivative could not beacetylated, and only yielded a rnoimsodium salt% (XI). It thusappears that the hydrogen atom in position (1) cannot' be acetylatedunder ordinary conditions, but can take part in salbformation.Since a diacetyl derivative and a monosodium salt are obtainedfrom 3-keto-1 : 3-dihydroindazole itself, it is clear that its relation-ship in these respects t o its dinitro-derivative corresponds exactlywith t h a t of aniline to picramide :(IX ) (X.) (XI.)These examples illustrate the fact t,hat a nitro-group can protectan amino-group by steric hindrance, and yet render its hydrogenatoms active towards alkali.I n other words, the two processesinvolved, of which one is a particular example of those with whichthis paper is concerned, are fundamentally different.A group of large molecular volume occupying the second ortho-position to the substituent X may, of course, be expected toexercise the steric hindrance usually observed in such cases of ring-formation, and we therefore ficd that 2-chloro-1 : 3-dinitrobenzenereacts much less rapidly than 4-chloro-1 : 3-dinitrobenzene (Borscheand Rantscheff, AnnuZen, 1911, 379, 152).Steric influencesmust further be considered in regard t o their effect on theformation of the primary additive compound. Thus Sudboroughand Picton (Zoc. cit.) showed that the introduction of three methylo r two methoxyl groups, or three bromine atoms into the moleculeof trinitrobenzenei completely inhibited the formation of molecularcompounds with a- or j3-naphthylamine, whilst Hofmann and Kirm-reuther found that alcoholic solutions of trinitromesitylene and oftrinitro-m-xylene, in contrast with trinitrobenzene, developed nocolour when hydrazine hydrate was added (Ber., 1910, 33, 1765).Similarly, Jackson and Boos showed that di- and tri-nitromesi-tylenes did not give coloured additive products with metallic alkyl-oxides (Amer.Chem. J., 1898, 20, 444). The importance of thisfactor is illustrated by the case of 4 : 6-dibromo-l : 2-dinitrobenzene,from which the 1-nitro-group is displaced by the amino-group(Blanksma, Rec. trav. ckirn., 1908, 27, 50):* The formulae given to these salts in the text are of course alternative to para-quinoiioid ones, or t o Hantzsch's nci-formul?REACTIVITY OF SUBSTITUEN'TS IN THE BENZENE NUCLEUS. 2725NO2 NH2Brf)N02ESr\/ \/BrA comparison with the behaviour, already quoted, of 4-chloro-(or bromo-)-1 : 2-dinitrobenzene a t once suggests that the nitro-group in position 2 would be the more easily displaced were itnot that the other nitro-group is sterically prevented from formingthe necessary additive compound.Thus,the reaction between potassium methoxide and trinitrophenetole(Meisenheimer, Zoc.c i t . ) may be represented in the followingmanner ;The above considerations are applicable in other cases.MeOK MeOKOEt EtO i I E tO ,**"'*-,.\ /''\/ \/'NO2&feO __._- I.-.-NO, NO2&!+.U--KEtO ,*,' EtO / ,,/'\,/" O.....- \/ (j02pJ / \z?s //I I No -+\NO,\/NO,Me()----- I<The formation of coloured salts of nitrophenols, to which Hantzschhas assigned the formula (XII) (Ber., 1912, 45, go), can beexpressed in the same manner:I< (-J-- - -_------I I1I I - \\//L/(XII.)and it is no longer remarkable that "echte Nitrokorper mit einerzweiten negativen Gruppe isomerieren sich als Pseudosauren (durchWasser oder Alkalien) niemals nur zur ersten Stufe der einfache2726 KENNER: THE INFLUENCE OF NITRO-GROUPS ON THEaci-Nitrokorper, sondern stets sofort zu der zweiten Stufe der kon-jugierten aci-Nitrokorper, so dass in dem f olgenden Umwandlungs-schema das Mittelglied nicht existiat " (Hantzsch, Zoc.cit.) :The course of the condensation of 2 : 4-dinitrotoluene with benz-aldehyde in presence of secondary bases (Thiele and Escales, Ber.,1901, 34, 2842) may perhaps be considered to take place in asimilar manner. This suggestion a t once accounts for certainobservations inade by Borsche (AnnaZen, 1911, 386, 351), whichhe was unable to explain. Whilst 4 : 6-dinitro-(XIII)- and2 : 4 : 6-trinitro-(XIV)-m-xylenes respectively give yields of 35 and43 per cent.of the corresponding dist,yrylbenzenes when condensedwith benzaldehyde, 2 : 4-dinitro-(XV) and 2 : 4 : 6-trinitro-(XVI)-mesitylenes are unchanged.Me Me Me Me()Me0 2 7 \/Me (NO2 Me(,!Me Me(,)Me/\ /\NO, 02N/\N0, O,N/'\NO2(XVI.)NO2(XIII.) (XIV.) (XV. 1NO, NO2I n the last two cases the nitro-groups are sterically preventedfrom forming the required additive compound, but in the first twocases both the 4- and the 6-nitro-groups are free to do so. If, how-ever, the condensation took place by direct interaction of the methylgroups with the aldehydes, tsinitro-m-xylene should be unchanged,as trinitromesitylene is, whilst dinitromesitylene should react withease.Also, a comparison of the behaviour of the two reactive com-pounds is interesting as illustrating the resultant influence of, onOMeOMREACTIVITY OF SUBSTITUENTS IN THE BENZENE NUCLEUS. 2727tho one hand the activating influence, and on the other the sterichindrance to the later stages of the reaction, of the 2-nitro-group.Further, i t may be observed that Hope and Robinson, when con-sidering the mechanism of the condensation of nitromeconine withcotarnine (shown on p. 2726), attributed two functions t o theimino-group : (‘ first, in attracting the nitro-compound and informing a l’oose combination of the type of the compounds obtainedfrom amines and trinitrobenzene, and secondly, in effecting theCondensation between the aldehyde and methylene groups ” (T.,1911, 99, 1153, 2114).The views thus expressed approximate veryclosely to those cdvocated above.The reactivity induced by other meta-directive groups maypossibly be explained in a similar manner. Thus there is abundantevidence that the carbonyl group is able to take part in the forma-tion of molecular compounds (compare, for example, Pfeiff er,Annnlen, 1910, 376, 285; 1911, 383, 92; Ber., 1914, 47, ISSO),and i t is probable that the same will be found to be true of thesulphonyl and the cyanogen groups.It is possible that similar considerations might lead to anexplanation of the activity of the chlorine atom and the methylgroup in 2-chloro- and 2-methyl-pyridines.At any rate, theseexamples serve to expose the inaccuracy of Vorliindes’s statement(Annaleih, 1910, 320, 66), according to which the ‘( reactive ” un-saturated group, *X:Y*, causes mobility of the hydrogen atom insystem XVII, but not in XVIII or XIX:H*A*X:Y* H*X:Y* H*A*B*X:Y*(XVII.) (XVII 1.) (XIX.)Whilst, however, the views just developed suffice to explain anumber of observed facts, they do not embrace another reaction,which was observed during the course of the experiments aboutto be described. Thus i t was found that 3-chloro-5 : 7dinitroindazoZe(XX), formed when 5 : 7-dinitro-3-keto-1 : 3-dihydroindazole isheated under pressure with phosphoryl chloride a t 140°, is convertedinto 3 : 5 : 7-trichloroindaso7e (XXI) if the reaction is carried outa t 1 8 0 O ; and the 2-phenyl derivative showed a similar behaviour :NH--NH N-JSH N-NHI \NO21 A /--uo 1 -+ NO,(\i>CCl I‘ I -+ Cl()dCl\/C1(XXI.)\/ \/NO2 NO2(XX.)Analogous observations have been made by Lobry de Bruyn(Rec.trav, chirn., 1896, 15, 84), who found that, among others2728 KENNER: THE INFJAUENCE OF NITRO-GROUPS ON THEs-trinitxobenzene and the three dinitrobenzenes were converted intothe corresponding chloro-derivatives when heated with saturatedhydrogen chloride solution at 200-300°. Similarly Zincke andKuchenbecker (AnizaZeiz, 1904, 330, 50) showed that 2 : 2’-dinitro-azobenzene-4 : 4/-disulphonic acid, when heated with concentratedhydrochloric acid a t 160°, furnished tetrachloroazobenzene :NO, NO, c1 c1\-/ \-/ \-/ \-/ *HO,S/-\N : N/-\SO,H --+ CI/-\N :N/-\CIThe facts that all the nitro-groups are displaced, and that in thecases of the dinitrobenzenes the reaction is independent of theirorientation, would appear t o indicate that the cause of the reactionis t o be sought in the relationship of the’ nitro-groups to thebenzene nucleus rather than in any direct mutual influence of thesegroups.Since, however, nitrobenzene itself does not undergo thisreaction (Lobry de Bruyn and van Leent, Zoc. cit.), i t is evidentthat the second nitro-group does play some part, probably bysuitably modifying t-he condition of the benzene nucleus. Thissurmise receives some measure of support from the fact that) thedisplacement of nitro-groups by chlorine atoms takes place muchmore readily in the case of naphthalene derivatives (Atterberg,Ber., 1876, 9, 1187, 1732; 1877, 10, 1843), in which the conditionof the nuclei is generally acknowledged t o be somewhat differentfrom that of the ordinary benzene nucleus.This idea is not novel,f o r it is implied in the term “negativity,” by which the chapterof organic chemistry dealing with such cases of mobility is gome-times designated. It is therefore also probable that this effect ofnitro-groups on the benzene nucleus must bel taken into accountin connexion with the reactions to which the main part of thepresent discussion has been devoted. Similarly, the nature of thesubstlituent t o be replaced also requires consideration, Thus Sud-borough and Picton were unable t’o prepare picryl-a-naphthylaminefrom picrylaniline, although the necessary additive compound wasisolated (Zoc.cit., p. 587). The fact that s-Lrinitrobenzene is con-verted into 3 : 5-dinitroanisole by the action of sodium methoxide,but is not affected by ammonia (Lobry de Bruyn, Rec. trav. chim.,1890, 9, 214, 218; 1894, 13, 148), shows that the nature of theentrant group is also of importance. It must, however, be observedthat these factors operate by predisposing the system to reactionrather than by playing any active part’ in the change.Thus thedisplacement of a nitro-group in the course of the synthesis of3 : 5-dinitro-oxazine from 2 : 4 : 6-trinitro-2~-hydroxydiphenylamine,Still other influences are noticeable in certain casesREACTIVITY OF SUBSTITUENTS IN THE BENZENE NUCLEUS.2729observed by Turpin (T., 1891, 59, 722; see also Kehrmann, Ber.,1899, 32, 2605):NH /\/ \Ais comparable with the conversion, just cited, of s-trinitrobenzeneinto 3 : 5-dinitroaniso1e7 and is probably a result of the negativity ofthe benzene nucleus. This, however, affords no explanation of thefact, observed by Ullmann, that whilst Turpin’s reaction is applic-able to 2 : 6-dinitro-2’-hydroxydiphenylamine, i t fails in the case ofthe 2 : 4-isomeride (Awnalen, 1909, 366, 79).Further, the result of the action of potassium cyanide on alcoholicsolutions of the bromonitrobenzenes and similar compounds a t 180°,first studied by Richter, is remarkable.Para- and meta-bromo-nitrobenzenes are respectively converted into meta- and ortho-bromobenzoic acids, but the ortho-isomeride is recovered un-changed (Ber., 1871, 4, 21, 461, 553; 1874, 7, 1145; 1875, 8,1418 ; compare Zincke, Ber., 1874, 7 , 1503). Whilst these reactionsare normal in the sense that the hydrogen atom in the ortho-posi-tion to the nitro-group is atbacked in each case, they disclose astrong tendency, not observed in any of the cases previouslyreferred to, on the past of the substitut’ing group (in this casecyanogen, which is subsequently hydrolysed), t o enter the nucleusas close to the other negative group as possible. So strong is thistendency that i t prevents the conversion of o-bromonitrobenzeneinto m-bromobenzoic acid, although both the bromine atom and thenitro-group in this case conduce to the mobility of the meta-hydro-gen atom; whilst in the second case quoted above the reaction takesplace irrespective of considerations of steric hindrance, which wouldsuggest the para-hydrogen atom as more likely to be attacked.* Ifdue allowance is made for this factor, the action of potassiumcyanide on various nitro-compounds, apparently so bewildering inthe variety of results observed (collated by Lobry de Bruyn, Rec.trau.chim., 1904, 23, 47), affords an excellent illustration of theeffect of the influences previously discussed, namely :(a) the loosening influeace of meta-directive groupings on ortho-and para-substituents ;( b ) the corresponding effect of ortho-para-directive groups onmeta-substituents ;* Probably bromonitrobenzonitriles are first prodwed in this reaction. Forevidence of the mobility of the nitro-group in such compounds, compare the cases ofo-nitrobenzonitrile, m-dinitrobenzene, and other compounds referred to later2730 KZNNER: THE INFLUENCE OF NITRO-GROUPS ON THE( c ) steric hindrance to the formation of the initial additive(d) the, negative condition of the benzene nucleus:( e ) the nature of the substituent to be replaced and of theentrant group.It will be observed that, in the following instances,nitro-groups attached to a negative nucleus and hydrogen atomsalone suffer displacement. Halogen atoms are not affected bypotassium cyanide.The following table shows the more notable of the resultsobtained in this direction, accompanied by references to the abovecompound and to the attack of otherwise mobile groups;summary :Nitro-compound.o - Dinitro -benzene.p - Dinitro -benzene.in - Dinitro -benzene.Dinitrated6 - nitro-2-ethoxy-b e n z o -nitrile.1 : 2 : 4-Tri-nitrobeii-zene.4 - C h l o r o -1 : 3 - d i -n i t r o -beuzene.Conditionsemployed.Alcoholic so-lution a t170".Boiling alco-ho!ic solu-tion.Boiling alco-holic solu-tion.Warm alco-holic solu-tion.Roiling alco-tion.Boiling alco-holic solu-tion.holic solu-Product.~ No change.p-Nitroanisole.12 - 6; Dinitro -benzonitrile]-+ 6-nitro-2 - niethoxy -benzonitrile.A d i n i t r i l e ,the 6-nitro-group beingdisplaced.2 :4- Dinitroani-sole.5 - C h 1 o r o- 6 -n i t r o - 2 -m e t h o x y-benzonitrilc.Referencest o literature.Bruyn and van Geuns,Xec.trav. chim., 1904,23, 32.Bruyn and van Geuns,Eoc. c i l .Rruyn and van Geuns,Eoc. cit. ; other refer-ences in this paper.van Geuns, quoted byBruyn, REC. trau.chim., 1904, 23, 52.Bruyn, Xec. trav. chinz.,1890, 9, 193.van Heteren, Ihc. trav.chim., 1901, 20, 107 ;Blanksma, ibid., 1902,21, 424.* The cyanogen group renders a nitro-group i n the ortho-position mobile(influences (a) + ( d ) ) . Thus o- and p-nitrobenzonitriles, when boiled with a solutionof sodium methoxide, are converted into o- and p-methoxybenzonitrileel (Ringer,Rec. trav.chim., 1899, 18, 330 ; Reinders and Ringer, ibid., 326). Lulofs showedthat some o-nitroanisole is also produced in the former case (Rec. trav. chim., 1901,20, 321). Further, Lobry de Bruyn showed that 2-methoxy-6-nitrobenzonitrile wasconverted into 2 : 6-dimethoxpbenzonitrile by ruethyl-alcoholic potassium hydIoxide(Rec. trav. chim., 1883, 2, 205). t The l-nitro-group diminishes the mobility of the 3-hydrogen atom, and thusprevents the reaction from following a similar coiirse to that observed in the case ofm-dinitrobenzene. The influence of the chlorine atom and the amino-group in thecases next quoted is, however, in favour of such a result. This influence of thechlorine atoni is clearly responsible for the partial conversion of 5-chloroREAC'fIViTlf OF SUBSTITUENTS IN THE BENZZNE NUCLEUS.2731Lippniann and Fleiss-ner, Nonatsh., 1885,6, 807 ; 1886, 7, 95.Nietzki and Petri, Ber.,1900, 33, 1788;Rorsche, Ber., 1900,33, 2718, 2995;Horsche and Locatelli,Ber., 1902, 35, 569 ;Horsche and Biicker,Ber., 1903,36, 4357 ;Bruyn, Bec. trav.chirn., 1904, 23, 56.Blanksmn, Rec. trav.c7~ina., 1901, 20, 411,423.Nitro-compound.Alcoholic so-lution a t60".Aqueous solu-tion at 35"( N i e t z k iand Petii).2:4-Dinitro-aniline.2: 6 - Dinitro-3- (a) i (b) + ( d )Potassium iso- ( a ) + ( b ) + (c)aminophenol. + (c).purpurate.* + ( e ) .2 : 4 :6-Tri-nitrophe-nol.Di- and tri-n i trome-sity lenes.Referencesto literature.Product. 1 Influences. I employed.1 : 3-dinitrobenzene into 3-chloro-5-nitroanisole by the action of sodium methoxide,whilst 1 : 3-dinitrobenzene is simply reduced (Kock, Bee. trav. chim., 1901, 20, 111 ;Steger, ibid., 1899, 18, 13).* The authoritics quoted agree that isopurpuric acid contains cyanogen groups inthe 3- and 5-positions, but differ in regard to the fate of the 2- aud 6-uitro-groups.According to Nietzki and Petri, one of them is converted into an amiuo-group, whilstBorsche suggests that the rcduction only reaches the stage of a hydroxylamino-group. Lobry de Bruyn considers that the colour of the compound would bebetter explaiued by the presence of a nitroso- and a hydroxylamino-group in placeof the two nitro-groups.The question as to the extent t o which the nitro-groupssuffer reduction has, however, no bearing on the present discussion.Two points in connexion with the above table call for explana-tion. No attempt is made to explain the introduction a t one timeof a cyanogen group, a t another of an alkyloxy-group, as a resultof the action of potassium cyanide. This problem, which lies out?sido the scope of the present discussion, was considered by Lobryde Bruyn, who, however, was unable t o arrive a t any very definiteconclusion (Rec. trav. chirn., 1904, 23, 47). Also, the term" influences " a t the head of one column of the table is only meantto comprehend those which would appear t o play a decisive partin determining the final result2932 KENNER: THE INFLUENCE OF NITRO-GROUPS ON THEEXPERIMENTAL.[With RAYMOND CURTIS.]5 : 7-Dinitro-3-keto-1 : 3-dihydroiizdnzole,NH*NHOpN(\;--bO\/NO2An alcoholic solution of hydrazine hydrate (6 grams) wascautiously added to a Golution of methyl 2-chloro-3 : 5-dinitro-benzoate (10 grams) in warm alcohol (100 c.c.).Each additioncaused a vigorous reaction, which became violent if the initialtemperature was too high, and a dark red hydrazine salt wasimmediately precipitated. After dilute hydrochloric acid had beenadded t o the mixture, the free indazole derivative was collected,and purified by crystallisation from glacial acetic acid. Itseparated from the solution in yellow, hexagonal plates, melting anddecomposing a t about 300O:0.1900 gave 0.2596 CO, and 0.0368 H20.0.1736 ,, 38.2 C.C.N, at 17O and 734 m p N=25*06.C7H40,N4 requires C = 37.50 ; H = 1.78 ; N = 25.00 per cent.The compound was not appreciably soluble in cold alcohol, water,benzene, or chloroform; i t was fairly soluble in boiling alcohol, andreadily 80 in glacial acetic acid or acetic anhydride.The1 d’isodium salt was prepared by triturating 5 : 7;dinitro-3-keto-1 : 3-dihydroindazole (2 grams) with 5N-sodium hydroxidesolution (7 c.c.). After the addition of absolute alcohol (5-10 c.c.)t o the resulting deep brown paste, the salt was collected andwashed with absolute alcohol until the washings were free fromalkali. For analysis two different. specimens were prepared, thesecond by the use of a considerable excess of sodium hydroxidesolution; although the products were dried a t 130°, the salt stillretained two molecular proportions of water :0.1288 gave 0.0603 Na2S0,.Na = 15.17.0.2938 ,, 0.1380 Na2S0,. Na = 15.22.0.134 required 9.2 C.C. N/lO-H,SO,.C,H20,NENa,,2H,0 requires Na = 15.13 per cent. ; M.W. = 304.The salt exploded when heated, and this prevented determina-tions of its other elements being made. It was readily soluble inOwing to the difficulty indetermining the exact end-point, no claim can be made for great accuracy in theresult.C=37.26; H=2*15.M.W. =291.** The salt itself served as indicator in the titrationREACTIVITY Ol? SUBSTITUENTS IN THE BENZENE NUCLEUS. 2733water, soluble in alcohol or acetone, but insoluble in benzene o rclilorof orin.By treating its aqueous solution with suitable re-agents, a reddish ferric salt, an insoluble dark brown silver salt,and a moderately soluble dark brown copper salt were obtained.The’ monoctcetyl derivative separated in greenish-yellow crystalson cooling the solution obtained by boiling the compound (2 grams)with acetic anhydride (5 grams) for eight hours. It crystallisedfrom glacial acetic acid in rectangular prisms, melting a t195-200° :0.1701 gave 0.2548 CO, and 0.0340 H,O.0.1548 ,, 28.6 C.C. N, a t 1 8 O and 741 mm. N=21*2.The solubility relationships of the acetyl derivative correspondedIt was soluble in alkali, form-All attempts to prepare a diacetyl derivative were fruitless, theC=40*85; H=2*22.C,H,O,N, requires C = 40.60 ; H = 2.25 ; N = 21.05 per cent.with those of the parent substance.ing a deep red solution.monoacetyl derivative being produced in every case.5 : 7-Dicsrnino-3-keto-l : 3-dihydroindazole.N H*KHBH,The dinitro-derivative (4 grams) was added to a solution ofstannous chloride in glacial acetic acid (80 grams), saturated withhydrogen chloride.After twelve hours, the reduction product wasisolated from the precipitate by treatment with hydrogen sulphidein the csual manner. It was analysed in the form of its oxalate,which was precipitated when ammonium oxalate was added t o anaqueous solution of the hydrochloride of the base :0.1062 gave 20.8 C.C. N, at 1 8 O and 740 mm. N=21*95.C,H,ON,,C,H,O, requires N = 22-20 per cent.5 : 7-Didro-3-ke to-2-phenyl-l : 3-dih ydroiir dazol e,N H NPhO,N/j-&O\/NO2An alcoholic solution of plienylhydrazine (12 grams) was slowlyadded to a warm solution of methyl 2-chloro-3 : 5-dinitrobenzoate(10 grams) in alcohol (100 c.c.).After the mixture had beenheated on the water-bath for fifteen minutes, it was worked up asVOL. cv. 8 2734 KENNER: THE INFLUENCE OF NITRO-GROUPS ON THEin the previous case. The product consisted of two compounds,one of which was extracted by boiling alcohol. The residue(90-95 per cent. of €he yield) crystallised from glacial acetic acidin flat needles, which melted and decomposed between 220° and250° :0.1814 gave 0.3464 CO, and 0-0477 H,O.0.1552 ,, 25.2 C.C. N, a t 16.5O and 740 mm. N = 15-72.CI3H8O,N4 requires C =52.00 ; H = 2.66 ; N = 18.66 per eent.The compound was therefore either the 1- or the 2-phenyl deriv-ative of 5 : 7-dinitro-3-keto-1 : 3-dihydroindazole.It was onlysparingly soluble in boiling alcohol, benzene, or chloroform. Itwas moderately soluble in glacial acetic acid and in epichlorohydrin.Tho monosodium salt was obtained as an amorphous powder inthe manner already described in the case of dinitroketodihydro-indazole, and exhibited a similar tendency t o explode when heated.It was readily soluble in water, soluble in alcohol or acetone, butinsoluble in chloroform or benzene :C=52.08; H=2*91.0.2162 gave 0.0491 Na2S04.0.161 0 required 5.0 C.C. N/lO-H,SO,.Na = 7-35,M.W. = 322.C,,H70,N4Na requires Na = 7.1 per cent.; M.W. = 322.By double deconiposition with its aqueous solution, an insoluble,deep red silver salt, and a moderately soluble, brown coppw saltwere obtained. The compound is insufficiently acidic to permit ofthe formation of a ferric salt.Attempts to prepare) an acetyl derivative by various methodswere unsuccessful, the original compound being recovered in eachcase. A colourless substance separated from the solution obtainedby adding sulphuric acid to- a mixture of the compound with aceticanhydride, but it decomposed and turned yellow immediately itcame in contact with the moisture of the atmosphere.As explained in the introduction, these reactions agree with thebehaviour of picramide, and i t is therefore considered that thecompound is a 2-phenyl derivative.This conclusion is in harmonywith the results of E. Fischer (Aiz~inleiz, 1878, 190, 67), whoshowed that the condensation of picryl chloride with phenyl-hydrazine results in thel formation of trinitrohydrazobenzene. Itappears that other products may result under slightly different con-ditions (Willgerodt and Ferko, J. p. Chem., 1888, lii], 37, 346;Fischcr, AIznaleu, 1889, 253, l), and Fischer himself showed thata i-nixture of symmetrical and unsymmetrical derivatives is pro-duced by the action of alkyl bromides on phenylhydrazine (com-pare also lticliaelis and Schmidti Ber., 1887, 20, 43). These factsmay throw some light on the formation in the present case of REACTIVITY OF SUBSTITUENTS IN THE BENZENE NUCLEUS. 2735second compound, which was deposited from the above-mentionedalcoholic solution in yellow, rhomboidal plates.After repeatedcrystallisation from alcohol, the compound melted a t 175O :0.1420 gave 0.2897 CO, and 0.0502 H20.0.1450 ,, 25.6 C.C. N, a t 1 6 O and 737 mm. N=20*51.ClSH,,0,N6 requires C= 55.89 ; H = 3.92 ; N = 20.60 per cent.C=55*64; H=3*93.The compound therefore resulted from the condensation of onemolecular proportion of the ester with two of phenylhydrazine :CGH,(N0,)2C1*C0,Me + 2C,H3*NH-NH, +We were unable to oxidise i t with mercuric oxide, to condense i twith aldehydes, to convert it into the indazole derivative producedsimultaneoixsly with it4, o r to prepare it from 2 : 4-dinitro-6-carbo-methoxyhydrazobenzene. Unless steric influences be assumed, thisevidmce is a t variance with the possible formulae XXII andXXIII, whilst forniula XXIV, representing the product of thesemidined inversion of XXIIC, is rendered improbable by the factthat the compound in question may be isolated without the aid ofacid by removing the phenylhydrazine salt of dinitroketophenyl-dihydroindazole by solution in cold alcohol :NPh-NH, NH*NHPhHCl + MeOH + CGH,(N02)2*C',,H,40N4./\CO*NH*NHPhO,N{)CO*NKNHPh 02Nl I\/ \/NO2(XXII I.NO2(XXII.)N H*C,H,*NH,O,N(\CO*N H ON H Ph(XXI v.)2 : 4-Dir~itro-6-car bomethoxyhydrazobenzene,C0,Me/ - \NH*NH-C,H,.02N\- -/NO2After plienylhydrazine (8 grams) had been added to a solutionof methyl ' 2-chloro-3 : 5-dinitrobenzoata (10 grams) in alcohol(100 c.c.), the mixture was heated to the boiling point, and thencooled.The precipitate thus obtained was warmed with alcohol,and, after filtration, the solut8ion deposited orange crystals of thehydrazo-derivative. After further crystallisation from warmalcohol, it melted a t 144-145O:S Q 2'736 KENNER: THE INFLUENCE OF NITRO-GROUPS ON THE0.1634 gave 0.3016 CO, and 0.0523 H,O.0.1430 ,, 21.7 C.C. N, a t 20° and 732 mm. N=17.11.Cl4I6CI3O6N4 requires C = 50.60 ; H = 3.61 ; N = 16.87 per cent.When its alcoholic solution was boiled f o r a few minutes, internalcondensation took place, and the corresponding dihydroindazolederivative was precipitated. The same product was obtained wiienattempts were' made t o condense the hydrazo-derivative withphenylhydrazine or benzaldehyde.C=50*45; H=3.55.s-2 : 4 : 2f : 4f-Tetrai~itro-6 : 6 ~ - d i c a r b o m e t h o x y t e t r a ~ h e ~ ~ ~ ~ -NO2 NO9h ydrasiize, O,N/-\N Ph* N P b <-)NO2.\-/C0,Me C0,McAn alcoholic solution of hydrazobenzene (2 grams) was heatedwith a solution of methyl 2-chloro-3 : 5-dinitrobenzoate (6 grams)in alcohol (60 ex.) on the water-bath for one hour. The brick-redprecipitate, after filtration and crystallisation from nitrobenzene,melted a t above 340O:0.1777 gave 0.3480 CO, and 0.0526 H,O.0.1918 ,, 21.6 C.C. N, a t 20° and 747 mm. N=12.94.C2,H2,012N, requires C = 53-16 ; H = 3.1 6 ; N = 13-29 per cent.It was insoluble in alkalis, and in most organic solvents exceptnitrobenzene and epichlorohydrin. Attempts t'o hydrolyse the com-pound with acid or with alkali were unsuccessful.I f the coiistitution assigned to the phenylindazole derivative,already described, is accepted, it is remarkable that the secondphenylamho-group of hydrazobenzene reacts with a second mole-cule of ester rather than with the carbomethoxyl group of the firstmolecule, and it, might be suggested that this is evidence in favourof the alternative constitution.On the other hand, however, itmay bO argued that the velocity of reaction in the present case ismuch slower than when phenylhydrazine is used, and that this isdue to the greater facility with which the ester condenses with theamino-group of phenylhydrazine.C=53*40; H=3.29.N-NHO,N()-CCI "\I3-Chlo Fo-5 : 7-dirtit roindaaol e,5 : '7-Dinitro-3-keto-1 : 3-dihydroindazole (5 grams) was heatedwith phosphoryl chloride (35 grams) for four hours unde'r pressureat 120-140O.After the excess of phosphoryl chloride had beenremoved by warming t.he product under diminished pressure, thREACTIVITY OF SUBSTI'I'UENTS IN THE BENZENE NUCLEUS. 2737residue was stirred with water, collected, and repeatedly crystal-lised from glacial acetic acid. I n this way, yellow needles(3 grams) were obtained, which melted a t 179-180°:0.1994 gave 0.250s CO, and 0.0197 H,O.0.2020 ,, 40.6 C.C. N, a t 18O and 752 mm. N=23*39.0.3615 ,, 0.2112 AgCl. C1=14.46.C7H30,N,C1 requires C = 34.66 ; H = 1.19 ; N= 23.09 ; C1= 14.64.per cent.C=34.30; H=1*09.N-NH5 : 7-Dinitro-3-keto-1 : 3-dihydroindazole was heated underpressure with 10 parts by weight of pliosphoryl chloride for tenhours a t 160-180°.The product was isolated in the same manneras in the previous preparation, and separated from glacial aceticacid in microscopic, white twinned needles, which melted a t190-190*5O :0.1085 gave 11.8 C.C. N, at 18O and 740 mm. N=12*47.C,H,N,C13 requires N = 12.64 per cent.N-NPhAfter a sealed tabe containing a mixture of 5 : 7-dinitro-3-keto-2-phenyl-l : 3-dihydroindazole (3 grams) and phosphoryl chloride(24 grams) had been heated for four hours a t 120-140°, andcooled, it; was found that the above chloroindazole derivative hadcrystallised. After separation from the liquor, and recrystallisa-tion from glacial acetic acid, somewhat indefinite small, greenish-yellow needles were obtained :0*2000 gave 30.7 C.C. N, at 17O and 725 mm. N=17.3.C,,H70,N,Cl requires N = 17.6 per cent.N-NPh ' \\I3 : 5 : 7-Trichloro-2-p~eizylilzdaaole,\Cl5 : 7-Dinitro-3-keto-1-phenyl-1 : 3-dihydroindazole (5 grams) washeated with phosphoryl chloride (40 grams) for fourteen hour2738 EWINS: THE ALKALOIDS OF QUEBRACHO BARK. PART I.under prwsure a t 160-170°. The product, isolated in the Bamemanner as in the previous cases, separated from glacial acetic acidin silky, yellow needles, which melted a t 208-210° :0.1779 gave 13.8 C.C. N2 a t 18.5O and 754 mm.0'3783 ,, 0.5429 AgCl. C1=35*53.N=9*06.C1,€I,N,CI, requires N = 9-41 ; C1= 35.79 per cent.Action of Phosphoryl Chtoride on $2 : 4 : 21 : 4'-Tetranitro-6 : 61-dicarbomethoxytetraphenylhydrazine.A mixture of the hydrazine derivative (3 grams) with phosphorylchloride (40 grams) was heat'ed for ten hours under pressure a t170-180°. The product of the reaction could not be crystallisedfrom the ordinary solvents, and was therefore purified by digestionwibh sodium hydroxide solution and reprecipitation from thefiltered solution :0.1815 gave 0.3572 CO, and 0.0436 H,O.0.1672 ,, 11.0 C.C. N, a t 20° and 731 mm. N=7.4.C,,H,,0,N4C1, requires C = 53.51 ; H = 2.74 ; N = 9.90 per cent.C,,H,,O,N,Cl, ,, C=55*51; H=2*84; N=4-99 ,, ,,The product was therefore a mixture! of two acids, derived fromthe original est4er by hydrolysis and replacement of respectivelytwo and four of its nitro-groups by chlorine. The incompletenature of the relaction was probably due to the high molecularweight and sparing solubility of the hydrazine derivative.This resulb is of value, since it shows that the pyrazole ring wasnot the predisposing cause' of the mobility of the nitro-groups inthe previous instances, although this mobility is specially notice-able in t'he naphthalene series (see the introduction), and Frieshas shown (Amtalen, 1912, 389, 313) that 6-hydroxyindazole iscomparable in its reactions with &naphthol.C=53*67; H=2*66.THE UNIVERSITY,SHEFFIELD