TBE FORM OF CHANGE IN ORQANIC COMPOUNDS. 1265 CXXX1V.-The Form of Change in Organic Com- pounds, and the Furiction of the a-Meta-orientating Groups. By ARTHUR LAPWORTH. IN a paper published in the Transactions of the Society some time ago (Trans., 1898, ’73, 445), the author showed t h a t a very large number of changes and interactions in organic chemistry could be represented by forms, deducible from two simple types, which may be termed the ay- and up-rules. A critical examination of what is really essential to the applicability of these rules shows that the whole matter rests solely on the assumption that isomeric change is due, not to mere interchange of the position of atoms or groups in a molecule, but in general to a series of migrations, apparent or real, and that only one atom or group migrates at a time.With such an assumption, it is easy to show that there is no way of avoiding the conclusion that if the production of ring compounds is excluded and no change of valency occurs, the migration must be representable by a strict application of the uy-rule; if, on the other hand, change of valency does occur a t any part of the molecule, then the ap-rule must come into play at that point. It is of no consequence how the migration is effected, as the principle simply precludes a direct interchange of groups, and the representation of undissociated compounds by illegitimate formulae. The author was greatly influenced by the apparent simplicity of the stereochemical view of the mode of transference suggested in his first paper (pp. 448 and 452), but is no longer able to regard explanations based on such principles as of any real value.Occasion has already been taken to point out (Proc., 1901, 17, 2) that in applying the author’s ay- and &-rules, the intermediate positions which were assigned to a migratory group in its trans- ference must not be regarded as real resting points, but rather as representing the various possible positions of the group in the line between its initial and final positions of attachment. An explanation which appears capable of explaining, amongst other things, the very general applicability of the rulesis one based on the assumption that change in organic chemistry is largely due to the occurrence of a dissociation between singly bound groups and subservient to the following law ; the state of dissociation between, two singly bound atoms which leads to isomeric change, exists d y once in the molecule at any instant. A dissociation of the character which the above conception de- mands is found in weak electrolytes, and, in tbe author’s opinion,1266 LAPWORTH : FORM OF CHANGE IN ORGANIC COMPOUNDS, it is to electrolytic dissociation, often doubtless in extremely minute amount, that the majority of changes in organic compounds may be most probably assigned.This view, which the author has now held for a long time, and arrived at from a consideration of the prevalent forms of change, receives support in the recent ap- pearance of the papers of Euler (Bey., 1900, 33, 3202) and of Zengelis (Ber., 1901, 34, 198), who, apparently from entirely differ- ent considerations, have been led to put forward a similar sug- gestion with regard, more particularly, to the hydrolysis of esters and sugars.On this principle, the simplest possible case of isomeric change may be imagined to consist in the dissociation of a molecule X-Y at a point between two singly bound atoms into two groups, 2 and k, each of which is capable of becoming attached to one univalent atom or group to form an undissociated chemical molecule capable of free existence. The groups 2 and ? will, as a rule, be oppositely charged ions. Now the principle governing the possible modes in which X and * may again become mutually attached appears to be that the resulting, now undissociated, molecule X*Y shall be immediately re- presentable by a legitimate formula through a mere adjustment of the conventional I ‘ bindings.” * In order to avoid misconception, it must be pointed out that this adjustment takes place in general along.an already existing line between the initial and final positions of mutual attachment of one group to another, a single binding becoming a double one, a double one a treble one, or vice versd, or a single binding may altogether disappear. The converse of the last change, namely, the formation of a single binding where none existed before, is probably to be found only in cases where the formation of ring compounds is thereby in- volved; even here, however, it usually appears possible to trace exactly similar laws. It should be made clear that the dissociation of X*Y may be followed, not by a reunion of X and Y, but, in presence of other dissociated groups, by the union of these groups with others to form compounds which are not isomeric with the original substance.To illustrate the application of the preceding principles, a case of substitution, namely, the action of bromine on phenol, which appears to * Possibly the formation of the new compound is preceded by an actual re; adjustment of the bindings in the molecule. Thds it is probable, for instance, that a dissociated group i*B:C*D:E niay change to A:Bb6*D:E or A:B.C:D*k, so that the a?- and aS-rules and their stereochemical explanations may apply to the altera- tion of the position of free affinity in the dissociated group, and these changes may be successive, although no intermediate compounds are formed.b 4AND FUNCTION OF THE a-META-ORIENTATIXG GROUPS. 1267 be fairly simple, may be considered. The bromine may be imagined to dissociate slightly into Br and Br, and the weakly acid phenol as usual into H and O*C,H,, the bromination consisting in the union of Br and O*C,H,, the possible products are those in which these two groups (otherwise preserving their original characteristics so far as the relative positions of atoms contained in each are concerned) are mutually attached, so that by merely altering, if necessary, the bindings in the above-mentioned sense, a legitimate formula is at once obtained for the compound C6H,0Br. This is possible at the oxygen atom, a t the carbon atoms 2- and 4-, but not a t the carbon atoms 1- or 3..The possible products are therefore : - + -k - + - OBr 0 0 H A P H / j H >< H+H HI1 lKB' Or HI1 IIH \/ HI IH H H \/ H Br unless it is supposed that ring compounds such as : 1-0 0 might be produced, a possibility which is sufficiently remote to require no further remarks. The law relating to dissociated groups may be stated more strictly than on p. 1265 as follows. The ions of organic compounds usually possess onEy one point of free valency, which is merely another way of stating that they are usually weak electrolytes, and therefore form only univalent ions. Following out the recently advanced conceptions of Abegg and Bodlander (Zeit. ccnorg. Chem., 1899, 20, 453), a complex ion may be considered to arise from the union of a simple ion with a neutral component, and, conversely, a complex ion may break down into a simpler one and a neutral component, It should follow, therefore, that a molecule, X*Y*Z, may become dissociated into ions, say, X and Y*Z, and that afterwards the ion Y*Z may become further simplified into a neutral component Y and the ion Z , the latter carry- ing away the free affinity and the charge.A law similar to that just formulated will hold for the latter process, namely, that the neutral component Y shall be at once representable by a legitimate formula through an adjustment of the bindings in the sense already explained.1268 LAPWORTH: FORM OF CHANGE IN ORGANIC COMPOUNDS, It will follow in general that a simple ion will be the more readily eliminated from a complex ion of a given type if it has a large ‘‘ electroaffinity ” (Abegg and Bodlander, loc.cit.), and it may be pointed out here that,, in most simple cases where two groups are withdrawn from a molecule, the product of their union is a recognised electrolyte. In illustration of this point, the author’s conception of the breaking down of acetone cyanohydrin, CN*CMe,*OH, under the influence of alkali may be alluded to. Here the dissociation probably occurs first between the hydrogen atom and the oxygen of the hydroxyl group, since it must in any case be supposed that all alcohols are feebly dis- sociated in this way. The ions are therefore H and CN*CMe,*b, and in the negativelycharged ion the CN group is known to have a large electroaffinity for the negative charge ; it is, moreover, in such a position that its removal will result in the formation of a molecule at once representable by the legitimate formula CMe,:O.The function of the alkali is doubtless to provide hydroxyl ions in the presence of which the concentration of the hydrogen ions of the cyanohydrin is greatly diminished, and the dissociation of the cyanohydrin con- sequently raised to a corresponding extent. With this brief introduction, it is possible to proceed to some special cases t,o which these principles apply, and, in order to be able to deal generally with some of these, the following rule, easily deducible from what has been said, may be stated. If, durifig a reaction, a group X might be expected to unite with an atom R, in the complex R,*Rp:R,, then the product may contctin, not only substances with the compkx R,X-R,:R,, but also, and, sometimes exclusiuely, those with R,:R,*R,X and thei9* tautome& forms.Some important forms of interaction may be a t once mechanically deduced from the expansion of the simplest type of addition at a double binding, A:B + X*Y = AX-BY, and it is easy to show that the general equations may be true if read in either sense. - I- . Firstly, expanding B we obtain : A:B,-B,:B, + x + ir f-f AX~I~,=B,:B, + + t-, AX~B,:B,*B,Y (R), easily seen t o be the form to which so much importance has been attached by Thiele. ,!!econdly, expanding Y only : f i AX*& + $,*Y :Y % i*BdY,-Yp:Ya + k f l A:B + X+ 3?,*Ys:Y,, @ %X*B*YY*YB Y, (S). This form is not translatable in terms of Thiele’s rule, but is oneAND FUNCTION OF THE a-META-ORIENTATING GROUPS.1269 which will probably be found t o be the basis of many reactions as yet imperfectly understood, and one or two cases where it appears directly applicable may be discussed. It is, no doubt, the type of change involved in the Claisen reaction in its simplest form, as, for example, when sodium ethoxide acts on a mixture of acetone and ethyl oxalate, a condensation which is perhaps the result of the follow- ing series of changes. The ions Na and CW,:CM& of the sodium derivative of acetone, CH,:CMe*ONa (Freer, Amer. Chem. J., 1891, i3, 322), unite with a carbonyl group of the ethyl oxalate. As sodium shows very little tendency to form complex ions, ibis perhaps the negative ion which first becomes attached to the ethyl oxalate (compare, however, Proc., 1901, 17, 95) : - + s a + CJ3,:CMe.b + O:C(OEt)*CO,Et .t-f $a + 6 * C(OE t)(CO,E t)*CH,*CMe:O.The components on the right are the ions of the sodium compound, NaO*C(OEt)(C0,Et)*CH2*CMe:0, which, by loss of alcohol, affords the stable sodium compound, NaO*C(C02Et):CN*CMe:0, as the actual product. As anything like an adequate discussion of the details of this reaction would occupy a very large amount of space, the author con- fines himself a t present to the foregoing remarks. The reaction between aldehydes and the metallic-derivatives of iso- nitro-compounds doubtless proceeds in accordance with this type. Thus with pot a ssium isoni t ropropane and f ormaldeh y de, K + bNO:CMe, + CH,:O ++ K + O:NO*CMe,*CH,*b f-f O:NO*CMe,*CH2*OE. Thirdly, expanding both B and Y, we obtain the form thus roughly expressed : X*Ya*Yp:Y, + A:Bu*Bp:BY ++ AX*Bu:Bp*B,*Yy9Yp:Ya .- (T). This doubtless includes, amongst others, the cases of interaction of +unsaturated esters and ketones with ethyl sodio-malonate, -cyano- acetate, and -acetoacetate. Thus with methyl acrylate and ethyl sodiocyanoacetate : Na*O* C( OEt) :CH*CN + 0: C(OMe)*CH :CH, + Na*O*C( OMe) : CH* CH,. CH(CN) C( OEt): 0. In his first paper, dealing with the subject of isomeric change (Zoc. cit., p. 455), the author referred briefly to the case of meta-substitution VOL. LXXIX. 4 s1270 LAPWORTH: FORM OF CHANGE IN ORGANIC COMPOUNDS, in the benzene series, using the case of the sulphonation of nitro- benzene by way of illustration, and showed that the most probable mode in which the change might be supposed to occur was that the acting agent united in the first instance with the side group, and that elimination of a substance then took placo by removal of a hydrogen atom in the nucleus (necessarily the rneta-atom, unless it is supposed that ring compounds like those depicted on p.1267 are produced) with a group attached to the side chain. The initial stages were mostly similar to those usually assumed to take place when nitromethane, for example, is converted into a salt of its iso-form by a metallic hydr- oxide; subsequently, in the benzene derivative, the hydrogen atom is removed, not from the a-position, but from the related y-position. That is to say, the benzene compound was supposed to be converted into a derivative of its tautomeric form corresponding with the iso-forms of fatty nitro-compounds, and this was then converted into the rneta- substitution derivative by isomeric change in accordance with the ay-rule.The first actions assumed, written in the usual form, may be compared thus : (1) Caustic potash acting on a fatty nitro-compound, :CIPNO:O + HOK + : c H ~ N o < ~ ~ --+ :C:NO~OK + HOH. (2) Sulphuric .acid acting on nitrobenzene (S = SO,H), -CH-~X~*NO:O + HOS -+ :CH*~:C.NO<E,H ---f :C:b*6:NO*OS + H OH It was not perceived at the time when this view was put forward that, if thers be any truth in.the conception, then those groups which, in the aromatic series, are the characteristic meta-orientating groups should be capable of reacting on the tautomeric normal and iso-forms >CH*R and >C:RH when in attachment to a >CH group. The fact that the groups, NO,, CO*R, and CN, do possess both properties is a t least very striking (compare Hantzsch and Veit, Ber., 1899, 32, 607, and Hantzsch and Osmund, ihid., 34l), and as regards the only other meta-orientating group, namely, SO,*R, the experiments of K8tz (Bev., 1900, 33, 1120) are sufficient to show that its properties are similar to those of the other three groups, more especially as it was found that the compound Ph*SO,*CH(SO,Et), behaved as a strong acid.It may be remarked that the foregoing groups in the fatty series render the a-hydrogen atom, and in the benzene series the related 7-hydrogen atom, replaceable, so that it appears probable that the formation of large quantities of meta-di-derivatives is due in general to the occurrence of changes very similar to those occurring in the fatty series, the relationship being expressed by the ay-rule.AND FUNCTION OF THE a-META-ORIENTATING GROUPS. 1271 It may be argued that substitution by direct attack in the nucleus is possible, and that therefore meta-derivatives may be formed in this way.Doubtless they are thus produced to a certain extent in all cases; nevertheless, it seems impossible to believe that such widely different groups as N(CH&, CH,, C1, and CH,CI, should render the ortho- and para-hydrogen atoms directly replaceable, and that the others, NO,, CN, &c., should have precisely the opposite effect by a mere exertion of an attractive or repulsive influence. I n the author's opinion, so far as direct substitution in the nucleus is concerned, all groups may be presumed to have an ovtho-paw orientating influence, due, perhaps, io considerations like those sug- gested by Thiele, and the formation of msta-substitution derivatives in quantities preponderating over those of the ortho- and par-deri- vatives combined, must be attributed to an entirely different pheno- menon.The formation of ortho- and para-substitution products simul- taneously from an ordinary benzene mono-substitution derivative C,H,*X by an agent PQ is perhaps due to the formation of inter- mediate compounds of the types X P and H Q related by the ay-rule, and yielding by loss of HI? the benzenoid o~tho- and pwa-derivatives, C,H,XQ.In consequence, were X of a nature capable of exerting a steric hindrance on the formation of such an intermediate compound, the action would be hindered, but without, of course, preventing the para-hydrogen atom from being removed as easily as the meta-atom. If, therefore, in the fatty series, the group X*b:Q forms additive products with difficulty, then it is to be expected that substitution in the compound C,H,*X will also be difficult. Such an example is found in tert.buty1- benzene. Since, then, it appears that the function of the characteristic meta-orientating groups in the benzene series is probably quite similar to that which they exercise in the fatty series, it should follow that the similarity could be traced, not only in the effect which they exercise in the simplest cases of substitution, but in more complex reactions.One of these, applicable to many fatty compounds in which they occur, is the Claisen reaction, the course 4 s 212’72 LAPWORTH: FORM OF CHANGE IN ORGAKIC COMPOUNDS, of which has already been discussed (p. 1869). Now, in order that this action may occur, it is necessary that two hydrogen atoms should be present a t the a-carbon atom, and it happens that in benzenoid ketones, esters, &c., where the carbonyl group is in direct attachment to the nucleus there is no hydrogen at the a-position; proceeding, therefore, t o the related point, the y-position, only one hydrogen atom is present as a rule, namely, the meta-hydrogen atom, but in some cases there is a stable CH, group a t the y- or y’- position outside the ring.Such a substance is found in 0- or p - methylbenzophenone or ethyl 0- or p-toluate, but not in the corre- sponding meta-derivatives. Whilst these compounds react only with the greatest difficulty, if at all, those containing the more reactive nitro-group, as, for example, 0- and p-nitrotoluenes, are capable of taking part in the Claisen reaction. Thus Reissert found (Be?*., 1897, 30, 1030) that the methyl group in these two compounds was attacked, but that no reaction occiirred in the case of the meta-nitrotoluene; he asserts, also, that ethyl oxalate is the only ester which can be used. The author has con- firmed Reissert’s statement regarding the inactivity of the meta- compound, but has found that amyl nitrite may be used in place of ethyl oxalate, and yields the oximes of 0- and p-nitrobenzaldehydes ; no doubt other esters might also be used, provided that the product were a sufficiently stable sodium derivative.The reaction is evidently a particular case of y- or meta-substitution, in which the meta-position is outside the ring, that in the ring being excluded from the type of reaction. It may be supposed, as in other cases, that a, compound of the type CH,:F*F:NO*ONa is produced, having a constitution analogous to that of the quinoneoximes on the one hand, and to the isosulphate of nitro- benzene (p. 1270) on the other. The iso-derivative then reacts with ethyl oxalate, as follows (compare p. 1269) : C0,Et *C](OE t)*ONa -+ C0,Et *C(OEt):O CH,: ?* <:NO*ONa CH, *?: -NO:O -+ that is, in accordance with the reaction expressed by A:B+X*Y, in which Y is expanded twice.As such a reaction should, in turn, be applicable to similarly con- stituted fatty compounds, it seemed that it might be possible to con- dense ethyl oxalate, hc., with open chain substances of the type *CH,*CR:CR*X (X = NO,, CN, SO,R, CO*R) at the y-position instead of at the a-position, as usual.AND FUNCTION OF THE a-META-ORIENTATING GROUPS. 1273 Compounds of the above type and containing no a-hydrogen atom are difficult to obtain in large quantities. Experiments were most successful with ethyl crotonate, and the results justify the prediction, as ethyl y-oxalocrotonate may be obtained, under the proper conditions, in fairly large amount. It is a matter of little difficulty to trace the course of the reaction in the above manner.The number of possible condensation products in such cases, and the fact that some of them may be formed with greater ease than the y-oxalo-derivatives, may obviously render the isolation of the latter a matter of great difficulty. That the ethyl oxalocrotonate is not the a-derivative is shown by the marked difference between its properties and those of other a-oxalo- esters and ketones. It affords a black instead of the red to violet coloration with ferric chloride which a-derivatives usually give, and its stability towards hydrolytic agents is incomparably greater than that of such compounds ; whilst the a-oxalo-esters decompose into oxalic acid on the one hand, and pyruvic acids and carbon dioxide on the other, the new substance does neither ; when boiled with alkalis, it gives no oxalic acid, and with acids it suffers hydrolysis and the product loses water, being converted into a substance which is evi- dently a ring compound, CH<~E$~>co2H, a change which would be impossible with an a-oxalocrotonic ester.Finally, the idea that the substance might be a P-oxalo-ester may be at once set aside, as it is impossible to find a suitable formula for such a substance. The condensation of ethyl crotonate with ethyl oxalate in the above way might, at first sight, be attributed to the presence of the group CJ&*bC*, but that this is not the case is shown by several facts. Thus ethyl a-methylacrylate, which is isomeric with ethyl crotonate and contains that group, gives no indication of forming a similar compound.Again, other unsaturated esters containing the grouping *CH,*CR:CR, should behave in a similar way, bnt examination of the ester of camphorenic acid, which also contains this grouping, showed that no condensation occurred, The power of two such pairs of doubly-bound carbon atoms in attachment to a CH, group to render the reaction a possible one is unquestionable (Thiele, Ber,, 1900, 33, 666), but with those substances which contain only one such pair, there is almost certainly a sharp line of demarcation, both in the fatty and in the aromatic series, between the two classes represenced respectively by the formulse *CH,*6:6* and *CH,*d:c'*X (where x = co, &c.).1274 LAPWORTH: FORM OF CHANGE IN ORGlANlC COMPOUNDS, EXPERIMENTAL.Action of Sodium Ethoxide and Amyl Nitrite on 0- and p-Xitrotoluenes. Amyl nitrite has little or no action on 0- or p-nitrotoluene in presence of alcoholic sodium ethoxide, but if the anhydrous ethoxide is used there is little difficulty in isolating the oxime of 0- or p-nitro- benzaldehyde respectively from the product. The mode of treatment found most suitable was as follows. Sodium (2-3 grams) was di&olved in 12 times its weight of alcohol, and the excess of the latter afterwards got rid of by distillation in a stream of hydrogen at about 180°, the removal being completed in a vacuum at this temperature, The residue was covered with about 200 C.C. of thoroughly dried and purified ether, and a mixture of amyl nitrite and o-nitrotoluene was then slowly poured in, the temperature being at first kept down by means of a freezing mixture.The whole, after remaining a few days at the ordinary temperature, was poured on crushed ice, the resulting aqueous solution being separated, extracted twice with pure ether, and freed from the latter by a current of air. On adding dilute hydrochloric acid to the solution, a fairly copious precipitate of solid matter separated. This was collected, dried, and crystallised from benzene. 0.2097 gave 0.3878 CO, and 0-0710 H,O. The substance melted at 95-96O and had all the appearance and properties of o-nitrobenzaldoxime. I n order to complete its indentifi- cation, it was heated for some time with concentrated hydrochloric acid, the liquid being then extracted with ether and the latter evaporated.A residue of solid matter remained, which had the properties of o-nitrobenzaldehyde, and on dissolving it in dilute acetone and adding a few drops of dilute sodium hydroxide, rapid darkening ensued, followed by a deposition of indigo. With p-nitrotoluene, the yield of oxime is not so good, but no difficulty was experienced in isolating it. The crystalline substance obtained melted at 128--12Q0, and when heated with pure p-nitro benzaldoxime, its melting point was not depressed, 0 1505 gave 0-2'780 CO, and 0.0515 H,O. C = 50.3; H = 3.8 per cent. With m-nitrotoluene, no evidence of the formation of an oxime could be obtained, and no positive result ensued, even when in this case ethyl oxalate was substituted for amyl nitrite.As condensation with ethyl oxalate almost invariably takes place more readily than with C = 50.4 ; H = 3.8. CIH,O,N, requires C = 50.7 ; H = 3.7 per cent. On analysis :AND FUNCTION OF THE WMETA-ORIENTATING GROUPS. 1275 amyl nitrite, it may be concluded that m-nitrotoluene behaves alto- gether differently from the 0- and pcompounds in this regard. Experiments were made in which amyl and ethyl formates and nitrates were substituted for oxalates and nitrites. With the formates, reaction certainly occurs, but on treating- the product from o-nitro- toluene with water, a crystalline, neutral compound separates, and only a small quantity of any acid substance could be detected. Theneutral substance crystallised from alcohol in nearly colourless prisms melting at 121°, and gave the following results on analysis : 0.1148 gave 0.2577 CO, and 0.0492 H,O.The substance was identified as 2 : 2'-dinitrodibenzyI, C = 61.2 ; H= 4.7. C,,H,,0,N2 requires C = 61-7 ; H = 4.4 per cent. which, it is interesting to note, has been obtained only on one previous occasion, namely, on treating o-nitrophenylpyruvic acid, the product from o-nitrotoluene, ethyl oxalate and sodium ethoxide, with alkali (Reissert, Ber., 1897, 30, 1039). As the substance is not formed when ethyl formate is omitted in the above reaction, it appears highly probable that it owes its origin to a condensation product of o-nitro- toluene and ethyl formate, which, judging from the properties of p-nitropheng lacetaldehyde, would be a highly unstable substance (Lipp, Ber., 1886, 19, 2647).Experiments with other o- and p-Substituted 5!'oluemes.--In endeavour- ing to extend the above observations, ethyl 0- and p-toluates were treated with ethyl oxalate and sodium ethoxide under several condi- tions. Here it was found that small quantities of unstable, acidic substances were produced which were ketonic in their nature and gave blue to violet colorations with ferric chloride. These substances were possibly the condensation products sought for, but good numbers were not obtained when they were analysed, owing t o the small quantities hitherto isolated, and the consequent difficulty of ensuring their purity. With 0- and p-toluonitriles, violent reactions took place, but rapid decomposition ensued, black, amorphous masses being deposited, and this occurred also in absence of ethyl oxalate. It is hoped that further experiments with 0- and p-methylsulphones and ketones will give more definite results.I n none of these cases, however, is it to be expected that such definite results will be obtained as with the compounds containing the highly reactive nitro-group.1276 LAPWORTH: FORM OF CHANGE IN ORGANIC COMPOUNDS, Condensation of Ethyl Crotonata with Ethyl Oxalate. Formation of Ethyl crotonate and ethyl oxalate condense in presence of anhydrous sodium ethoxide suspended in ether or toluene. The liquid slowly becomes bright yellow in colour as the ethoxide dissolves, and after the lapse of a week, a yellow, microcrystalline deposit settles on the walls of the containing vessel. A considerable quantity of the product remains dissolved, however, and the whole should be poured into ice- water containing an excess of dilute acetic acid, the ethereal solution separated, washed, and extracted repeatedly with dilute sodium car- bonate.The crude ester may be isolated by acidifying the alkaline extract with acetic acid and extracting with ether, which should after- wards be allowed to evaporate spontaneously. The oxalo-derivative may also be obtained by adding a mixture of the two esters to the calculated quantity of sodium, cut into thin sheets, and covered with a considerable quantity of absolute ether. In all cases, the presence of a very small quantity of water is sufficient to completely prevent the formation of the desired product, and the yield never exceeds 40 per cent.of the theoretical amount. The ester is best purified by dissolution in cold dilute sodium car- bonate, powdered sodium acetate being afterwards added to the soh- tion. The sodium compound then separates as a voluminous mass of needles which may be separated by filtration, washed with dilute sodium acetate solution, and afterwards decomposed by dilute hydro- chloric acid. The nearly white ester which separates is dried, and crystallised from a mixture of benzene and light petroleum. A speci- men of the ester was analysed. Ethyl y- Oxdocrotonccte, C0,Et *C( OH) : CH* CH : CH* C0,Et. 0.2265 gave 0.4684 CO, and 0.1336 H,O. C = 56.4 ; H = 6.6. C,,H1,O, requires C = 66.0 ; H = 6.6 per cent. Ethyl y-oxalocrotonate is a nearly colourless solid. It dissolves very readily in all organic media with the exception of light petroleum in which it is sparingly soluble, even when hot; it is nearly insoluble in water.It crystallises from a mixture of ethyl acetate and light petroleum in small, transparent prisms, and from light petroleum in fan-shaped aggregates of plates or isolated, short, compact prisms or tablets. When these are examined in convergent polarised light, an optic axis of a biaxial interference figure is seen to emerge obliquely with regard to the field. When melted on a glass slip beneath a cover-glass, it solidifies sud- denly on cooling in bundles of fine needles interspersed with irregularly arranged patches of indistinct crystalline structure.AND FUNCTION OF THE U-META-ORIENTATING GROUPS. 12'77 That the ester exists in the enolic form, C0,Et *C(OH):CH*CH:CH*CO,Et, is shown by the fact that it is a fairly strong acid; it expels carbon dioxide rapidly from dilute solutions of sodium carbonate, and if a moderately strong solution is made, the yellow sodium derivative crys- tallises out.From its solution in alkali, the ester is precipitated on addition of acetic acid. The presence of the enolic grouping in the substance is also shown by the fact that its dilute alcoholic solution is coloured an intense brownish-black on addition of ferric chloride, and this coloration is not discharged by the addition of a very large excess of strong mineral acid ; with ferrous sulphate, no marked coloration is produced. Attempts to prepare the acetyl and benzoyl derivatives of the ester resulted in the formation of white, amorphous masses, which did not dissolve at once in dilute sodium carbonate solution.They were slowly decomposed by water and moist solvents, however, and could cot be made to crystallise. The ester cannot be purified by distillation, even under reduced pressure, as it rapidly decomposes ; it boils at 180' under the ordinary pressure, losing a certain amount of alcohol, and affording, for the most part, a carbonaceous mass. A small quantity of an oily substance distils over and a phenolic odour becomes apparent, but the majority of the distillate is probably the ester of the coumalincarboxylic acid described later. On warming the compound with soda, it dissolves rapidly, and if the alkali be strong, separation of a red, oily sodium salt ensues.The salt dissolves on shaking, however, and the solution becomes intensely yellow. On acidifying the diluted liquid after boiling for a consider- able time, little or no oxalic acid could be detected when the highly purified ester was used. When hot hydrochloric acid is used in the hydrolysis, the liquid, a t first slightly yellow, becomes colourless as the substance dissolves, and here again no carbon dioxide or oxalic acid could be detected as a product of decomposition. When the substance is warmed with solutions of aniline, or hydr- mines in dilute acetic acid or in alcohol, yellow precipitates are slowly produced j the aniline derivative is readily obtained in a crystalline form, but has not as yet been closely examined.The molecular weight of ethyl oxalocrotouate in acetic acid was determined by the cryoscopic method. The mean of three concordant measurements was 225, the number calculated for a substance having the formula C,,H,,O, being 214.1278 LAPWORTH: FORM OF CHANGE IN ORGANIC COMPOUNDS, Action of #odium Ethoxide on a Mixture of Ethyl P-Bthoxybzctyrocte and Ethyl Oxakate. CH,*CH(OEt)*CH:C(OEt)*ONa might be regarded as a probable step in the formation of ethyl oxalocrotonate, sodium ethoxide was allowed to act on a mixture of ethyl /3-ethoxybutyrate and ethyl oxalate under the conditions found to be most advantageous for the production of that substance. It was found that it was eady to obtain some quantity of ethyl oxalo- crotonate in this way, but the yield, a5 compared with that obtained directly from ethyl crotonate, was small, and it is probable that the compound CH,*CH:CH-C(OEt),*ONa is the true intermediate sub- stance if any such is produced a t all, although it is manifestly im- possible to determine this with any degree of certainty.I n order to obtain evidence as to whether the compound The sodium derivative of ethyl oxalocrotonate, C0,Et C( ONa): CH* CH* CO,Et, is formed when the ester is dissolved in sodium carbonate or sodium hydroxide solution. It is not very readily soluble in cold water, and quite sparingly so in solutions of more soluble sodium salts. It forms long, yellow needles which have straight extinctions in polarised light, their directions of greatest length and elasticity being coincident.A specimen crystallised from a solution of sodium acetate was washed with water and analysed after drying over sulphuric acid in a vacuum. C,oHl,O,Na requires Na = 10.0 per cent. 0,3962 gave 0.1143 Na,SO,. Na=9*4. When this substance is washed with ether which has not been specially dried, it slowly decomposes and the ether on evaporation yields the nearly pure ester, indicating that the substance is hydro- lytically decomposed by water to a very considerable extent. The coppev derivative, (C,oHl,O,),Cu, is produced when an aqueous solution of copper acetate is added to the ester dissolved in alcohol; the blue solution of the copper salt becomes yellowish-green, and finally an amorphous greenish-brown precipitate is formed, which may be separated and washed with dilute alcohol.It is soluble in alcohol or ether, and could not be obtained in a crystalline condition ; on treat- ment with acids, it yields the pure ester. A specimen, after drying at looo, was analysed : 0,3635 gave 0.634 CuO. C20H,60,,Cu requires Cu = 12.9 per cent. With dilute aqueous soIutions of the sodium derivative, solutiona of soluble calcium, barium, nickel, or cobalt salts give no precipitates ; C = 12.8.AND FUNCTION OF THE U-META-ORIENTATING GROUPS. 1270 mercuric chloride throws down a bulky precipitate insoluble in ether. In somewhat concentrated solutions, calcium chloride forms a white, flocculent precipitate insoluble in ether j the silver compound is precipi- tated on adding silver nitrate, but rapidly darkens in the light, deposit- ing metallic silver.y- Oxalocrotonic Acid, C0,H C( OH) : CH- OH: CH. C0,H. This substance may be obtained in small quantities by the following process. The sodium derivative of the diethgl ester is finely powdered and triturated in a mortar with a 30 per cent. solution of sodium hydroxide, the liquid being added in small quantities and the temper- ature kept low. The solution, which is reddish-yellow in colour, is allowed to remain at the ordinary temperature for several hours until a small quantity, when diluted and acidified with acetic acid, gives no precipitate. The whole is then diluted with several times its bulk of water, cooled in ice, acidified with a large excess of sulphuric acid, and extracted about 20 times with ether, the latter being dried over calcium chloride and evaporated.The substance may be purified by crystallisation from a large bulk of ethyl acetate. 0,3721 gave 0.6264 CO, and 0,1194 H,O. The basicity of the acid was determined by titration with N/10 caustic soda, with phenolphthalein as indicator; the end point is difficult to determine precisely, owing to the deep colour of the result- ing sodium salt. The numbers obtained as the equivalent varied from 82-85, that required for a dibasic acid, CGH606, being 79. y-Oxalocrotonic acid is sparingly soluble in water and organic media in general, and is nearly insoluble in benzene, chloroform, or petroleum; it is most abundantly dissolved by acetic or formic acid or methyl or ethyl alcohol, none of which, however, deposits the substance in a well crystallised form.As usually obtained from aqueous or alcoholic solutions, it forms a bright yellow, microcrystal- line mass, and it is not easy to ascertain whether it is a uniform sub- stance or a mixture of tautomeric forms, but the author inclines towards the latter view. The cold aqueous solution of the acid is coloured intensely brownish- black by a solution of ferric chloride, indicating the presence of an enolic form, whilst on addition of hydrazine acetate immediate yellow precipitates are obtained, indicating that a ketonic form may be present. Hydroxylamine and semicarbazide, however, do not afford sparingly soluble derivatives ; with a solution of the former, the yellow colour of the solution becomes much less intense. C=45*8; H = 3 5 CGHGO, requires C = 45.6 ; H = 3.8 per cent, It melts and decomposes at about 190'.1280 LAPWORTH: FORM OF CHANGE IN ORGANIC COMPOUNDS, Cournalin-karboxylic Acid (a-P~rone-d-curbox~~~c Acid), cO>O.CH- CHKCH:C( C0,Et) When diethyl oxalocrotonate is heated with acids, or when the free acid is heated at its melting point, ring formation ensues and a coumalin- carboxylic acid is formed, The most advantageous mode of preparation consists in boiling the ester with excess of fuming hydrochloric acid until complete dissolution has taken place, the liquid being then heated with animal charcoal, filtered, and evaporated to dryness on the water-bath. For further purification, the residue is triturated with enough strong hydrochloric acid to form a paste, transferred to porous earthenware to drain, and then recrystallised two or three times from dilute acetic acid. On analysis : 0.2337 gave 0.4410 CO, and 0.0592 H,O.C,H,O, requires C-51.43; H= 2.8 per cent. The equivalent of the acid as determined by titration with caustic soda was 140, or exactly that required for a monobasic acid having the formula C,H,O,. The acid dissolves somewhat sparingly in hot water, alcohol, acetic acid, acetone, or ethyl acetate, and was prmtical1y insoluble in benzene, chloroform, or light petroleum. It crystallises best from hot, strong hydrochloric acid, When heated, it melts and decomposes slightly a t 227-22S3. The crystals from hydrochloric acid are rectangular, transparent plates, or long, transversely striated needles which show a tendency to become twinned.Heated beneath slips of glass, it sublimes slightly before melting, and on cooling solidifies rapidly, forming for the most part a microscopic mass of indistinct structure, but here and there, especially among the sublimed parts, large, well-formed crystals may be seen ; in these crystals, the extinction in polarised light is straight, and the directions of greatest elasticity and length are at right angles. On slow sublimation, as during its analysis by the combustion method, it sublimes in brilliant needles. A dilute solution of the sodium salt gave no precipitates with salts of barium or calcium. A nearly white precipitate was produced on addition of lead acetate, and this dissolved readily in acetic acid ; with ferric chloride, a copious, flocculent, brown precipitate was formed, and with mercuric chloride a small quantity of an insoluble salt was precipitated. A cold aqueous or alcoholic solution of the acid gave no coloration with ferric chloride.On warming the aqueous solution of the acid with silver nitrate and a drop of ammonia, rapid blackening ensues and a brilliant mirror is formed on the walls of the tube. This is doubtless due to the C= 51.5 ; H e 2.8.AND FUNCTION OF THE a-META-ORIENTATING GROUPS. 1281 hydrolysis of the acid to oxalocrotonic acid, as the solution of the acid, when made alkaline with ammonia, warmed, and afterwards acidified gives a very distinct brown coloration with ferric chloride. When the acid is heated rapidly in a test-tube, a distinct odour of coumalin becomes noticeable, and this substance appears to be formed in small quantity when the calcium salt of the acid is heated strongly in a stream of hydrogen.The aqueous distillato in the latter case, on addition of ammonium sulphate, becomes milky, and a small quantity of a nearly colourless oil is deposited. This, after extraction with ether, had the properties of coumalin so far as the small quantity obtained would permit them to be observed. Thus, i t gave a bluish- red coloration with ferric chloride, and evolved an odour resembling that of crotonaldehyde when warmed with alkalis. It may be observed that. the odour and colorations were directly compared with those afforded by coumalin made from mslic acid in accordance with the directions given by Pechmann (Anncden, 1891, 264, 305), and no difference could be detected. The acid shows little or no tendency to yield a pyridone derivative when treated with ammonia, but when its ammonium salt is strongly heated with lime, or even alone, a powerful odour of pyridine is evolved, and the distillate affords a perbromide closely resembling that of this base; the large quantity of ammonia and the small quantity of the basic substance in the distillate have hitherto prevented the isolation of any pure pyridine derivative.-This ester is most easily obtained by the method used by Pechmann in preparing the ester of the isomeric acid, as the properties of the acids and their esters are very similar (Zoc. cit., 279). It is purified by crystallisation from light petroleum. Bthyl CoumaZi.n-6-carl1oxyZate, CH<cH:C(Co,Et)>O.CH- CO 0.3808 gave 0.7948 CO, and 0.1672 H,O. C = 56-8 ; H = 4.9. C,H,O, requires C = 5'7.1 ; H = 4% per cent, The ester is readily soluble in water and in most organic media with the exception of light petroleum, in which it dissolves only sparingly when cold. It crystallises from hot light petroleum in beauti- ful, thin, rectangular plates melting at 59-60', The crystals, in polarised light, show interference colours of high orders, and have straight extinction, their directions of greatest length and elasticity being at right angles. After melting, the substance solidifies to opaque masses of long, flattened needles or thin plates. When covered with 15 per cent. aqueous ammonia, the substance first dissolves, and on stirring, a mass of beautiful, white crystals separates.These disappear on warming, however, and on treatment1282 LAPWORTH: FORM OF CHANGE IN ORGANIC COMPOUNDS, with potash, yield a yellow solution which gives the tests for the salt of oxalocrotonic acid; the ester, therefore, is not the ethyl pyridone- carboxylate which it was hoped would be produced. No other method of treatment tried gave any basic substance whatever. Action of Amyl Formate and Nitrite on Ethyl Crotmunte. Amyl formate and nitrite both react with ethyl crotonate in pre- sence of sodium ethoxide or metallic sodium suspended in dry ether, yielding acidic products in considerable quantity. These are possibly the y-substituted derivatives sought for, but no attempts to obtain them in a pure form have as yet been successful, nor has any better result been achieved by endeavouring t o isolate the corresponding acids.The isonitroso-compound is converted into a basic substance on reduction with sodium amalgam, and it is hoped that by means of this prodnct the nature of the original reaction may be finally ascertained. Action of Ethyl Oxakate on EtlTlyZ a-Methykacrykate, Ethyl a-methylacrylate was made by heating ethyl a-bromoiso- butyrate with diethylaniline, as recommended by Howles, Thorpe, and Udall (Trans., 1900, "7, 947), but was isolated by distilling the resulting mixture and fractionating the portion which passed over below 160°, a method which appears t o afford a somewhat better yield than that recommended by these authors. The ester was treated with ethyl oxalate in presence of sodium ethoxide and also of sodium and ether, and even under the conditions most advantageous for the formation of ethyl oxalocrotonate, no corre- sponding oxalyl derivative was obtained, even in an impure condition.This was shown by the fact that, after shaking out the product with water and acidifying the aqueous extract with acetic or hydrochloric acid, no precipitate was produced, and no black coloration was developed on addition of ferric chloride, It is thus established with some degree of certainty that the forma- tion of ethyl oxalocrotonate is not to be accounted for by the proximity of the negative double binding CH,*I):V, for this is also present in ethyl a-methy lacrylate. Negative results were also obtained on endeavouring to condense ethyl oxalate with ethyl camphorenate, which contains the grouping CH,*CH:CH.AND FUNCTION OF THE a-META-OHLENTATING GROUPS.1283 Applicution of the Claisen Beaction to othe~ Compounds containing the G;rozcping C H, 6 : 6- C 0. On treating ethyl dimethylacrylate, CMe,:CHeCO,Et, with ethyl oxalate in presence of sodium and ether or of sodium ethoxide, the bright yellow colour which is characteristic of the early stages in the preparation of ethyl oxalocrotonate a t once appears, and if, after a few seconds, the mass is treated with water and the aqueous extract acidified, a colourless oil separates in small quantity, and, on adding ferric chloride to a solution of this in ether, an intense black colora- tion ensues. If, however, the treatment with sodium or with sodium ethoxide is allowed to continue, the liquid becomes green, and, on adding water, a dark-red, aqueous solution is obtained, which deposits a comparatively small amount of an oil on addition of acetic acid. I n both cases, a small quantity of a brown copper compound, resem- bling that obtained from ethyl oxalocrotonate, may be obtained, but no crystalline substance is produced on decomposing this. Phorone, CMe,:CH*CO*CH:CMe,, reacts very readily with ethyl oxalate in presence of sodium ethoxide, and a large yield of an oily or resinous product is obtained. This affords a black coloration with ferric chloride, and gives a brown, copper salt ; it does not yield oxalic acid on alkaline hydrolysis, but gives a sparingly soluble, pulverulent acid which forms an orange-red solution in alkalis. Benzylidenemesityl oxide, CMe,: CH* CO*CH: CHP h, also reacts, al- though somewhat less readily, with ethyl oxalate under the conditions described, and a similar compound is produced. This, as well as the preceding ketone, also reacts with amyl formate or nitrite, but the products are equally uninviting, Camphorone gives exactly similar derivatives. Although the products have not yet been obtained in a pure form, it appears probable from these results that there is a line of demarca- tion between those esters and ketones which contain the grouping *CH2*&6*CO* and those which do not. The difficulty of obtaining pure substances is due, in part, to the fact that the products are readily decomposed by distillation, even in a vacuum, and also, doubtless, to the readiness with which the up-unsaturated esters and ketones undergo decompo- sition and polymerisation under the conditions of experiment, as has been shown by Pechmann and others (Be?., 1900, 33, 3329, &c.), so that, unless the condensation with the oxalate is sufficiently rapid the product is likely to be of a very mixed character. The investigation of the foregoing compounds is being continued,1284 LAPWORTH AND LENTON: THE CONSTITUTION OF and it is intended that the behaviour of ethyl tetrolate, ethyl a-ethyl- crotonate, &c., towards ethyl oxalate should be examined. The author desires to express his best thanks to the Research Fund Committee of the Chemical Society for a grant which has largely de- frayed the cost of the preceding investigation, and with the help of which it is being continued. CREMICAL DEPARTMENT, SCHOOL OF PHARMACY, BLOOMSBURY SQUARE, LONDON, W.C.