519 XXXIX.-The Constitution of certain so-cdled ‘‘ Mixed Azo-compounds.” By FRANCIS R. JAPP LL.D. F.R.S. aird FELIX KLINGEMANN Ph.I). PART I. INTRODUCTORY AND THEORETICAL. UETWEEN three and four years ago Dr. Ludwig Landsberg mentioned to one of us that he had obtained by the action of sodium methoxide and methyl iodide on benzene-azo-acetone* a coiourless feebly basic cornpound melting at 64’ which he had not further examined. Dr. Landsberg stated that he did not inkend to continue the investigation of this compound and very kindly allowed us to take up the subject. Analysis showed us that as was to be expected the cornpound had been formed from benzene-azo-acetone by the substitution of a methyl-group for hydrogen. Adopting the usually accepted formula for benzene-azo-acetone, CH,*CO*CH,*N:N*C,H,, the most probable supposition was that the methyl had displaced a hydrogen-atom of the methylene-group-a view which w9 put forward in a preliminary note on the subject (Ber.20 3398). By the action of acetic anhydride on benzeue-azo-acetone we also pre-pared an acetyl-derivative to which we assigned (Zoc. cit.) an analogous constitution. We shall show in the present communication, that the above formula for benzene-azo-acetone does not correctiy express its constitution and that the formuh which we ascribed to the two derivatives just mentioned must be correspondingly modified. As the yield of the methyl-derivative obtained by Dr. Landberg’s reaction was very small,? we endeavoured t o prepare the compound by splitting off carbon dioxide from benzene-azo-methacetoacetic acid, a method which would at the same time prove its const,itution.We had hoped to obtain ethylic benzene-azo-methacetoacetate by tbe action of diazobenzene chloride on ethylic sodio-methacetoacetate, following the analogy of the reaction described by V. Meyer (Ber., 10 2075) for the preparation of ethylic benzene-azo-acetoacetate from ethylic acetoacetate and diazo-salts ; but we found that the benzene-azo-group could only be introduced into the molecule of ethylic * Benzene-azo-acetone first described by v. Richter and Munzer (Ber. 17,1928), was discovered independently by Dr. Landsberg in the Chemical Laboratory of the Noi*mal School of Science about the time referred to; but the investigation was ttbandoiied after the publication of v.Richter and Miinzer’s memoir. t A relatively advantageous method of carrying out the reaction is described later on 520 JAPP AND KLIKGERIANN THE COXSTITUTIOK OF methncetoacetate with simultaneous elimination of the acetyl-group : thus-CH,.CO*CKa(CH,>*COOC,H + C,H,*N,Cl + H,O = C11Hl,N202 + C2Ha02 + NaCl. To t'he " ester " thus formed we a t first assigned the formula-CH,*CH*COOC,H, (Ber. 20 2942) regarding it. as the ethyl salt of benzene-a-azopro-pionic acid-the first representative of a new class of benzene-azo-fatty acids. The acid CgH,,N,02 was obtained by hydrolysis; and we also prepared a number of homologous " esters '' and acids under the impression that we were dealing with a new class of compounds. The supposed benzene-azo-propionic acid CgHlnN2O2 yielded by treatment with sodium amalgam belLxene-hy~razopl.opionic acid, N:N*C6H5 CH,*(I K*COOH NH-NHCeH,' which proved to be identical with E.Fischer and Jourdan's " phenyl-hydrazinepropionic acid " (Ber. 16 2244) a fact which however, eseaped our notice at the time of our first publications on the subject (Ber. 20 2943 and 3284). Fischer and Jourdan obtained their acid by reduction of phenylhydra,zonepyrnvic acid,* C H,*C.C 0 0 H N*NHC,H * On comparing our " benzene-azo-propionic acid " with phenyl-hydrazonepyruvic acid we found that these two acids were also identical. It was therefore necessary to ascertain whether the acid had the constitution of a hydrazone as assumed by Fischei. and Jourdan and indicated by its formation from phenylhydrazine and pyruvic acid; or that of an azo-compound as we had all along supposed.The fact that in the reaction discovered by us the diazo-salt expels simultaneously sodium and acetyl from tlhc moleciile of ethylic sodio-methacetoacetate might it appeared to us be interpreted in favour of the hydrazone formula since in this process the bivalent group CH,*C.COOC,H, would be formed; on the other hand the possibility of regenerating the acid by the action of an ammoniacal solution of a cupric salt on benzene-hydrazopropionic acid as observed * Jn the nomenclature of componnds formed bv the action of phenylhydrazine on carbonyl-compounds we have followed the recently published suggestion of Emil Fischer (Ber. 21 984) CERTAIN SO-CALLED MIXED AZO-COJIPOUSDS.521 by Fiecher and Jourda.n indicated rather the constitution of a n azo-compound. The problem we pointed out was of the same order as that involved in distinguishing between an isonitroso- and a nitroso-compound ( B e y . 20 3285). Of course the quest,ion also arose as t o whether benzene-azo-acetone itself was an azo-compound or a hydrazone ; and it appeared likely that in the case of this compound a study of the alkyl-deriva-tives obtainable by Landsberg’s react’ion might furnish a ready solution of the problem. If benzene-azo-acetone was an azo-com-pouiid the alkyl-derivative would have the formula-C H3.C 0 * CHR’. N :N*C,H ; whereas if it was a hydrazone (derived from pyruvic aldehyde) the alkyl would more probably athach itself to nitrogen thus-By the action of nascent hydrogen an alkyl-derivative of the first formula ought to yield aniline ; one of the second formula an alkyl-miline.I n either case the other half of the molecule would probably yield a ketine. The method of attacking the problem was therefore analogous to that employed by Victor Meyer and Ceresole in ascertaining the constitution of the isoni troso-compounds (Bey. 15 3071). We preferred however to introduce instead of a alkylm-group the group *CH2*COOH into the molecule of benzene-azo-acetone. The resulting compound would thus yield on reduction instead of a mixture of two bases a. base and an acid in this way greatly facilitat-ing the separation of the products ; whilst in order to ascertain the constitution of the compound ik was only necessary to distinguish between aniline on the one hand and anilidoacetic acid on the other, instead of between aniline and an alkjl-aniline.Benzene-azo-acetone was accordingly heated with sodium ethoxide and ethylic chloracetate in alcoholic solution. The resulting ‘‘ ester ” was converted into the acid which pi,ored t o have the expected formula CIIH,,N,O,. On reduction this acid yielded anilidoacetic acid but no aniline. I t has therefore the formula-and “benzene-azo-acetone” is not an azo-compound at all but a h j drazone of the formula-CH3.C O.CH:N*NHC,H, 522 JAPP AND KLINGEMBKN THE CONSTITUTION OF -an “ aldehydrazone ” of pyruvic aldehyde as distinguished from the “ ketohydrazone ”* derivable from the sitme compound.This view was confirmed by the fact that “ benzene-azo-acetone ” Fielded with phenylhydraziiie a compound-CH,.C:N.NHC,H, identical with the 0sazone-f- of pyruvic aldehyde (met,hylglyoxal) pre-pared by v. Pechmann. T t appears therefore that when diazo-salts act on ethylic aceto-acetate and its analogues o r on their mono-alkylated derivatives, hydrazones are formed. R. Meyer (Ohem.-Zeit. 1887 836) has sliown that the compound obtained from ethylic malonate and diazo-benzene chloride is identical with the phenylhydrazone of ethylic mesoxalate but draws the opposite conclusion to that arrived a t by us believing that the hydrazone undergoes transformation into an azo-compound. CH:N-NHC R,’ * These names are formed on the analogy of “ aldoxime” and “ketoxime.” ‘‘ Aldehydrazone ” is a contraction for (‘ aldehydehydrazone.” t “Osazone” is the term introduced by Fischer to denote the dihydrazone of a n a dicarbonyl-compound.$ This preliminary communication in the Chemiker-Zeitung had unfortunately escaped our notice at the time we published our first note on the identity of our ‘. benzene-azo-propionic acid ” with Fischer and Jourdan’s plieiiylliydrazonepyriiric acid . Our views on the constitution of the above “ mixed azo-compounds,” were first published together with a brief account of the experimental proofs in the Proceedings of December 15 1887 and somewhat later in the Berichte (20, 3284 and 3398). Professor Victor Meger who arrived independently a t the same conclusions published a paper later still in the Berichte (21 ll) in which he dis-cusses the theoretical question very fully but does not contribute any fresh experi-nrental material although he suggests various experiments some of which we had alrcady described in our paper.After his paper had been sent to the Berichte (date of reception 30th December 1887) Professor Victor Meyer received the number of the Proceedings containing the account of our work and despatched to the Berichte an addendum (4th January 1888) in which he says “ Ich gestatte mir darauf hinzuweisen dass ich uber das Thema am 13. December in der Gottinger chemischen Gesellschaft vorgetragen habe.” We believe that in the above sentence Professor Victor Meyer has unintentionally conveyed the impression that he read a paper before the Gottingen Chemical Society on December 13.VSTe understood the words in this sense and therefore tiwned to the account of the meeting in question given in the Chemiker-Zeitung (1837 11 1584) but found no mention of any such paper. Later however on January 8 in the number of the Chemiker-Zeitung following that i n which an account of our work had been published there appeared a second and “ supplementary ’) notice of the above meeting of the Gottingen Chemical Society in which it was stated that in a discussion which followed on the reading of a paper by P. Jacobsen, Professor Victor Meyer had made certain remarks on the constitution of mixed azo CERTXIS SO-CALLED MIXED AZO-COJIPOUKDS. 523 The bivalent hydrazone-group :N*NHC,H, thns corresponds with the bivalent isonitroso-group :N*OH.The isonitroso-compounds as Victor Meyer has shown are formed either by the action of hydroxyl-amine on a carbonyl-compound or by the action of ‘nitrous acid on a compound containing the group CH2 or CHR’ att,ached to two electro-negative radicles R’ being a readily displaceable radicle (acetyl or carboxyl). Substituting in the foregoing statement “ phenyl-hydrazine ” for “ hydroxylamine,” and “ diazo-componnd ’’ for “nitrous acid,” it describes the modes of formation of the hydr-azones. The analogy between the action of hydroxylamine and phenyl-hydrazine on carbonyl-compounds needs no illustaration. The analogy between the action of nitrous acid and diazobenzene-compounds 0x1 ethylic acetoacetate acetoacetic acid and their monalkyl-derivatives, is exemplified in the following equations in which for the sake of simplicity free diazobenzene is employed instead of its salts :-1.With ethylic acetoacelate the reactions may be expressed as follows :-2. I n the case of monalkyl-derivatives of ethylic acetoacetate the acetyl-group is expelled :-CEJ322>CH*COOC,H + HNO = CH,-C*COOC2H + C2HJ0,, N*OH compounds. The views embodied in these remarks were we niny say identical with those put forward by us and to this extent we were forestalled by Professor Victor Meyer. Two criticisms suggest themselves on the foregoing. In the first place as regards the general question although a discussion made before a learned Society may embody valuable remarks yet if this discussion is not reported in the ordinary course along with the other work of the Society i t would be hardly fail.to outsiders to regard such remarks as equivalent to publication. Secondly on the present and personal matter if we had chosen to take the same course as ProPessor Victor Mejer and publish our theoretical views first leaving the experimental confirmation for later we might have considerably anticipated the date of December 13 on which he lays so much stress 524 JAPP AXD KLINGEMANN THE COSSTITUTION OF 3. When nitzous acid acts on free acetoacetic acid or its monalkyl-derivatives instead of on their ethyl salts the carboxyl-group is eliminated and isonitroao-ketones are formed. We find that. in this case also diazobenzene behaves in an analogous manner. Thus with acetoacetic acid we have-whilst methacetoacetic acid yields with diazobenzene the compound CH,.CO*C (CH,) :N.NHC,H (the monohydrazone of d iacetyl already prepared by v.Pechmann from diacetyl and phenylhydrazine) ; a n d ethauetoacetic acid yields the corresponding compound of the formula CH,*CO.C (C,H,):N.NHC,€€,. The so-called azo-ketones are thus monohydrazones of a-dicarbonyl-compounds. We also discovered in the course of the present investigation the following curious mode of formation of osazones (dihydrazones of a-dicarbonyl-compounds). ILL heating phenylhydrazonepyruvic acid at 180-190° until it had ceased to evolve gas we obtabed along with oily matters a crystal-line substance of the formula CIGKIBN1 ( B e y . 20 2943). We after-wards found that this compound was identical wit'h the osazone of diacetyl since prepared by v.Pechmann. Its formation from plienylhydrazonepyruvic acid may be expressed by the equation-2C9K,,N20 = C,,H,,N + 2C0 + H,; but we believe that it is in reality formed in a secondary reaction. E. Fischer and Jourdan have shown (Be, 16 2242) that phenyl-hydrazonepyruvic acid breaks up on heating with evolution of carboii dioxide and formation of aldehydrazone (ethylidenephenylhydr-azine)-CSH,N,Oz = C,EI,,N + GO,. I n fact this is the chief reaction the oily subshance which we obtained consisting mainly of aldehydrazone whilst the osazone is formed only in small quantity. We therefore heated aldehydrazone a t 180-190" and found that the osazone of diacetyl was formed. The following equation shows t lie reaction : CERTAIN SO-CALLE t) MIXED AZO-COhlPOUICDS.525 PART 11. EXPERIMENTAL. 1 . Introductim of Momd Radicle.? into the Molecule of Yyruualdehydr-azone (L' Beraene-azo-acetone "). The pyruvaldehydrazone (to give the compound the name denotinq the coustitution to which the following experiments point) was prepared by a modification of v. Richter and Miinzer's method (Bey., 17,1928). The crude product of the action of diazobenzene chloride on ethylic sodacetoacetate wa9 first hydrolysed by warming for a few minutes with alcoholic caustic soda. Instead however of continuing the heating until the sodium salt thus formed was converted by the action of the caustic alkali into pyruvaldehydrazone and sodium carbonate we added watei. acidified with hydrochloric acid so as partially to precipitate the organic acid added sodium carbonate so as to redissolve the precipitate and then heated the liquid on a water-bath for several hours.I n this way the pyruvaldehydrazone was not exposed to %he deoomposing action of the caustic alkali and was obtained in a purer condition. It separated from the 'hot liquid in large plates which were filtered off and the liquid was heated afresh until 1 1 0 further separation occurred. The latter portions were less pure. The compound is most convenientlg purified by recrystallisat,ion from hot benzene or methyl alcohol. As regards its physical properties, we have nothing to add to v. Richter and Munzer's description. We have described the foregoing method because it was trhat8 by which the material employed in this investigation was obtained but later on we g5ve a method by which the compound can be much more readily prepared namely by the action of diazobenzene chloride on acetoacetic acid.Action of Acetic Anhyd&ie.-PFruvaldehydrazone was boiled with excess of acetic anhydride for several hours after which the solution was poured into water- An oil separated which speedily became crystalline. By recrJ-stallisation from boiling light petroleum the compound was obtained in colourless needles melting at 93". From a benzene solution it is deposited in thick well-furmed crystals appa-rently quadratic. Analysis agreed with the forniuls CllH12N202. Substance. co,. HzO. . I. 0.2 150 0.5105 0.1151 11 0-2265 0.5375 0.1202 111. 0.1110 gram burnt with oxide of copper in a vacuum gave 22-63 C.C.of a mixture of nitrogen and nitric oxide measured dry at 19.5" and under 459 mm pressure. After absorption of the nitric oxide there remained 22.63 c.c of dry nitrogen at 19.5" and under 434 mm. pressure 52 6 JXPP AND KLISGEMZANN THE COSSTlTUTION OF Found. 7 Calculittctd for r-h-C,IH1,N20,. I. 11. xIr. C 64-70 64.i5 64.72 H . . . . . . 5.88 5.91 5.89 -N 13.73 - 14.05 0 15.69 10o.c 10 --- - I -The compound is therefore acetylp!jri~valiZehydI-~xone, CH,.CO.CH:N.N(C,HI,O)CGH,. At the time when we first prepared this compound we were under the impression that the acetyl-group had displaced hjdrogen in thp '' methylene-group of benzene-azo-acetone " and that the compouri'l was therefore a diketone.We consequently heated i t with an excess of phenylhydrazine in alcoholic solution biit even at 200" only oy,e oxygen-atom was displaced by the phenylhydrazone-group. The phenylhydrazone which in reality has the formula, CH:N-?L'(C,H,O)CGH, C H,.C:N*NHC,H 9 crystallises in pale-yellow needles melting at 329" readily soluble in hot sparingly soluble in cold alcohol. 0.1013 gram burnt with copper oxide in a vacuum gave 22.63 C.C. of a niixtnre of nitrogen and nitric oxide measured dry a t 15.5" and under a pressure of 548 mm. After absorption of the nitric oxide, there remained 22.63 C.C. of dry nitrogen a t 15.5" and under 543 mm. pressure. Calculated for C17"18N 0 Found. N in 100 parts 19.04 19.06 A c h n of Methyl Iodide.-By heating together in methyl alcoholic solution equimolecular proportions of pyruvaldehydrazone sodium methoxide and methyl iodide only a very small proportion of the pyruvaldehydrazone was converted into the methyl-derivat,ive.This is due to the fact that the sodium methoxide and methyl iodide inter-act in great part by themselves as if no pyruvaldehJdrnzone were present. We then tried whether an increase in the proportion of sodium methoxide and methyl iodide would render the conversion more complete ; but the result was still unsatisfactory. The follow-ing slight modification of the process however by which the presence of any but a very small proportion of sodium methoxide during the progress of the reaction is avoided gave an excellent yield. A solution of 1 mol.proportion of pyruvaldehydrazone with 4 mol CERTAIN SO-CALLED MIXED AZO-COMPOUN DS. 527 proportions of methyl iodide in methyl alcohol was boiled with a reflux condenser ; a solution of 4 atomic proportions of sodium in methll alcohol was allowed to drop gradually into the boiling liquid so that the process extended over several hours; and the boiling was continued for a short time after the whole of the methoxide had been added. On pouring the contents of the flask into excess of water an oil separated which speedily solidified ; this substance was purified by recrystallisation from boiling light) petroleum. It forms colourless, flat needles melting at 64" very soluble in alcohol ether and benzene, somewhat less soluble in light petroleum. It is a weak base dis-solving i a concentrated hydrochloric acid but is reprecipitated by water.Analysis showed that it had the formula C,,H,,N,O :-Substance. co,. HsO. I-. . 0.2180 0.5450 0.1344 IL 0-2039 0.5098 0.1260 111. 0.1088 gram burnt with copper oxide in a vacuum gave 22.63 C.C. of nitrogen measured dry at 18.2" and under 500 nim. pressure Nitric oxide was not present. Found. -3 Calculated for r-h-ClOH,,"A). I. 11. 111. C 6818 68.18 68.18 H . . . . . . 6.82 6.85 6.86 -N 15.91 - 16-12 0 9\09 --- - --100~00 The compound has the constitution CH,*CO*CH:N*N<:g. That the methyl-group has not displaced hydrogen attached to carbon is shown by the fact that the compound is distinct from the hydrazone of diacetyl ( u i d e infka). When heated with phenylhydrazine in alcoholic solution it yields a compound, CH:N*X(CH,)C6H, which crystallises in faint-yellow needles melting a t 151-1 52", sparingly soluble in alcohol readily soluble in ether.0.0770 gram burnt with copper oxide in a vacuum gave 22.63 C.C. of a mixture of nitrogen and nitric oxide measured dry at 15.4" and under a pressure of 471 mm. After absorption of the nitric oxide, there remained 22.63 C.C. of dry nitrogen at 15.6" and under 460 mn]. pressure. C H,. C N N €l C H 528 JAPP AND KLINGEMAXN THE CONSTITUTJON OF Calciilxted for Cltal8N.l. Found. N in 100 parts 21.05 21.4 Action qf Ethyl Iodide.-Pyruvalde hydrazone reacted in an analogous manner with sodium ethoxide and ethyl iodide yielding the compound C H3.CO*CH:N.N(C2H5) C6H5 which crystallised from light petroleum in thin prisms melting at 5.5".The substance had a faint reddish tinge doubtless due to impurity. 0,1942 gram gave 0.4923 gram CO and 0.1277 gram H,O. Calculated for CIIH,,N,O. Found. c . 69.47 69-1 3 H . 7-37 7.30 N. 14.74 0 - 8.42 100~00 -Action qf Ethylic Ch1omcetata.-The pyruvaldehydTazone together with four times its molecular porportion of ethylic chloracetate was dissolved in alcohol and a solution of 4 atoms of sodium in alcohol was gradually added t o the boiling liquid conducting the operation as in the preparation of the methyl-derivative. Caustic soda was then added and the solution was boiled so as t o hydrolyse the "ester" which had been formed in the reaction. After expellinq part of the alcohol by heating water was added the liquid was filtered and the new acid was precipitated with hydrochloric acid.I n order to purify the substance it was dissolved in a cold solution of sodium carbonate, reprecipitated with hydrochloric acid and repeatedly recrystallised from boiling water. It formed yellow needles which melted with decomposition about 161-1 62". Analysis agreed with the formula CH3.C CbCH:N*N( CGH,) C HZ*COOH. I. 0.2128 gram gavet0.4671 gram CO and 0.1070 gram H,O. 11. 0.1067 gram burnt with coppel. oxide in a vacuum gave 22.63 c.c- of a mixture of nitrogen and nitric oxide measured dry a t 16" and under 405 mm. pressure. After absorption of the nitric oxide there remained 22.63 C.C. of dry nitrogen at 16' and under 395 mm. pressure.Found. Calculated for rL- 7 C,,H*2N2O3* I. 11. C 60.00 59.86 H 5-58 - 245 N 12-73 - 13.25 0 21.82 -- -100~0 CERTAIN SO-CALLED MlXED AZO-COMPOUNDS. 529 Action of Nascent Hydrogen on the Acid CIIH,,N203.-The acid was dissolved in alcohol and treated with excess of tin and hydrochloric acid. The tin was then precipitated with sulphuretted hydrogen the solution evaporated to dryness the residue extracted with ether,* and the substance remaining on evaporation of the ether recrystallised from hot water. The aqueous solution on cooling deposited small crystals with a slight brownish tinge melting at 126-1227'. (M. p. of anilidoacetic acid 126-127"; P. Meyer). I n all its properties, excepting the slight brown colour it was indistinguishable from ft specimen of anilidoacetic acid prepared from bromacetic acid and aniline.As the quantity was too small for further purification we analysed it direct. 0.0926 gram burnt with copper oxide in a vacuum gave 22-63 C.C. of a mixture of nitrogen and nitric oxide measured dry at 15" and under 253 mm. pressure. After absorption of the nitric oxide there remained 22-63 C.C. of dry nitrogen at 15" and under 246 mm. pressure. Calculated for 9.27 NH(C,HJ *CH,.COOH. Found. N in 100 parts. . . . . . . 9-54 Anilidoacetic acid was also obtained by reducing the acid CllHI2N2O3 with sodium amalgam in aqueous solution. No aniline was formed; but a strong smell of a ketine was perceived. We reduced simultaneously isonitrosoacetone with sodium amalgam and perceived the same smell.The acid CH,*CO*CH:N*N( CsH5)*CH2*COOH is therefore broken up by the nascent hydrogen with separation of the nitrogen-atoms, one-half of the molecule yielding milidoacetic acid the other a ketine. The formation of anilidoacetic acid from this acid proves that the latter has the constitution here assigned to it and thus indirectly, that the so-called '' benzene-azo-acetone " is pymvaldehydrazone. 2. Action of Pheny lhydrazine o n $ome Ketonic Compounds containing the Phenylhydraxone-group. Action on Ethylic Pheyiy lhydrazone-acetoglyozylate (" Benzene-azo-ucetoacetafe") .-lo grams of ethylic yhenylhydrazone-acetoglyoxylate and 5.5 grams of phenylhydrazine were heated with glacial acetic acid in a sealed tube for an hour at 120-130". On cooling the tube * As the substance here extracted was anilidoacetic acid we would point out that Sellwebel's statement (Be?.10 2046) that this acid is insoluble in ether is incorrect. Michaelson and Lipprnann the discoverers of anilidoacetic acid say (Annulen 139, 236) c c It is fairly soluble in water less soluble in ei her." This is correct. The statement as to its insolubility in ether has passed into Beilstein's Handbuck VOL. LIII. 2 530 JAPP AND ELINGEMANN THE CONSTITUTION OF was found to be filled with red needles with a characteristic blue reflex ; these after crystallisation from glacial acetic acid melted at 155". These properties pointed to the identity of the compound with Knorr's (l)-phenyl-(3)-methy7pyrazoZone-(4)-azobentene ( A n n a l e n , 238 183) which this author obtained by the action of diazobenzene salts on phenylmethylpyrazolone and to which he assigns the consti-tution-CGH5.N A N CO II I CH3.C -CH*N:N*C,H5.In the present case it is formed according to the equation-CH,*CO*C*COOC,H + C,Hj*NH*NH2 = Ci,H,,N,O + C2HGO + H2O. N*NHC,H, A nitrogen determination gave the following result :-0.0967 gram burnt with copper oxide in a vacuum gare 22.63 C.C. of a mixture of nitrogen and nitric oxide measured dry at 18" and under 556 mm. pressure. After absorption of the nitric oxide there remained 22.63 C.C. of dry nitrogen at 18" and under 550 mm. pressure. Calculated for CMHl,N,O. Found. N in 100 parts . . . . 20.14 20.06 Although ethylic phenylhydrazone-acetoglyoxylate must be regarded as a hydrazone we should not like positively t o affirm the same of the pyrazolone-derivative obtained from it in the foregoing reaction.It is quite conceivable that in the formation of this compound the hydrazone-group is transformed into an axo-group ; and indeed the extremely pronounced colour of Knorr's compound would rather indi-cate the occurrence of such a change.* p he n y 1-hydrazone-acetoglyoxylic acid is warmed with phenylhydrazine in alcoholic solution osazone-acetoglyozy lic acid-Action on P h en y Ih y drazone- acet og ly oxy 1 ic Acid .-When *CH3* C :N zHC,H, C:NzH CsH5, COOH separates almost immediately in slender yellow needles melting at 209". This acid was obtained by Knorr (Annden 238 195) by the action of phenylhydrazine on rubazonic acid.Knorr gives the melting point at 212". * Compare also Bernthsen Bev. 21 743 CERTAIN SO-CALLED MIXED AZO-COMPOUNDS. 531 0.1048 gram burnt with copper oxide in a vacuum gave 22.63 C.C. of a mixture of nitrogen and nitric oxide measured dry at 15.8" and under 572 mm. pressure. Aftel. absorption of the nitric oxide there remained 22.63 C.C. of dry nitrogen at 16" and under 555 mm. pressure. Calculated for CltJH16N402* Found. N in 100 parts 18.92 19.00 When heated with acetic acid this compound part@ with the ele-ments of 1 mol. H20 and yields the foregoing phenylmethylpyrazo-lone-azobenzene melting at 155O.* Action o n Pyrucalde lay d raz one.-Pyruvaldeh y drazone and phen yl-hydrazine when heated together in alcoholic solution react yielding the compound-C H3*C :N,HC,H, CH:N2HCsH; I t is deposited from hot alcohol in which it is only sparingly soluble, as a yellowish crystalline powder melting a t 145".It dissolves in concentrated sulphuric acid in the cold with an olive-green colour, which speedily changes to slate-blue and finally after standing for some hours to violet. A drop of the blue or violet solution poured into a porcelain basin and breathed upon becomes first green and afterwards yellow. 0.0889 gram burnt with copper oxide in a vacuum gave 22-63 C.C. of a mixture of nitrogen and nitric oxide measured dry a t 16.2" and under 561.5 mm. pressure. After absorption of the nitric oxide there remained 22.63 C.C. of dry nitrogen at 16.5" and under 551 mm. pressure. Calculated for C15H16N4.Found. N in 100 parts 22.22 22-08 Assuming that the above formula is correct the compound should be identical with the osazone of pyruvaldehyde (methylglyoxal) obtained by v. Pechmann by the act<ion of phenylhydrazine on pyruv-aldehyde and also by heating isonitrosoacetone with phenylhydrazine in acetic acid solution (Ber. 20 2543). We therefore prepared the * After the above was alceady in type we received No. 6 of the Berichte 1888 in which Knorr describes the formation of phenylmethylpyrazolone-azobenzene from osazone-acetoglyoxy1ic acid and points out its identity with phenylhydrazone-phenylmethylketopyrazolone (Ber. 21 1203). Knorr is of opinion that the compound is a hydrazone. Knorr converts the osazone into the pyrazolone-compound by dissolving it in caustic alkali and precipitating with acetic acid.2 0 532 JAPP AND KLINQEMANN THE CONSTITUTION OF compound by the latter method and found that it agreed in all its properties with that above described. 3. Action of Diazo-salts on Monalkyl-deriratives of Eth ylic Acetoacetate : Form ation of Hydrazones of u- Ketonic Acids. Ethylic Methacetoacstate a n d Diazobenzene Chloride.-An aqueous solution of diazobenzene chloride was gradually added in equimole-cular proportion to an alcoholic solution of ethylic sodiomethaceto-acetate cooling during the process. The liquid became deep red, and towards the end of the operation a dark-red oil i;eparated. Water was then added to complete the precipitation after which the oil generally solidified. When this did not occur the oil was washed with water in a separating funnel and then mixed with alcohol which caused it to become solid and crystalline.The solid substance was iwxystallised several times first from alcohol afterwards from light petroleum by which means it was obtained in yellow needles melting at 117". Analysis led to the formula CI,H,,N20 :-I. 0.2050 gram gave 0.4820 gram CO and 0.1255 gram H,O. 11. 0.1123 gram burnt with copper oxide in a vacuum gave 22.63 C.C. of nitrogen measured dry a t 17*S0 and under 438 mm. pressure. Nitric oxide was not present. Po nnd . Calculated for -7 CllH14N202. I. 11. - C 64-08 64.12 H 6.80 6.80 N 13.59 - 1370 0 15.53 100~00 -- -The acid was readily obtained by warming the '' ester " with an alcoholic solution of caustic soda and precipitating with hydrochloric acid.It crystallises from hot benzene in yellowish needles melting with evolution of gas at 185". It dissolves in cold concentrated sul-phuric acid with a yellow colour which soon changes to a deep-red. On allowing a drop of the concentrated red solution to flow about in a porcelain baoin and breathing on it the colour changes t o blue. Analysis gave figures agreeing with the formula C,H,,N,O :-I. 0.2202 gram gave 0.4893 gram CO and 0.1126 gram H,O. 11. 0*1102 gram burnt with oxide of copper in a vaciium gave 22-63 C.C. of nitrogen measured dry a t 23.4" and under 501 mm. pressure. Nitric oxide was not present CERTAIN SO-CALLED MIXED AZO-COMPOUNDS. 533 Found. Calculated for r-- 7 CISH1ON302.I. 11. C 60.67 60.60 H 5.68 I 5-62 N 15.73 - 15.68 0 17-98 -- -This acid proved as already mentioned to be identical with E. Fischer and Jourdan's phen ylhydyazonepyruvic acid, CH,*C*COOH N*N HC8HI' (Ber. 16 2241). Fischer and Jourdan give the melting point of their acid at 169" ; but later Fischer (Ber. 17 578) stated that this was a misprint for 192" which he gives as the true melting point of the acid.* We therefore carefully purified specimens of acid prepared both by our method and by that of Fischer and Jourdan but in * In an earlier communication on this subject (Ber. 20 3283) me unfortunately overlooked Fischer's correction and stated that Fischer and Jourdan had given the melting point " much too low." Our attention was called to the oversight by V.Meyer (Ber. 21 18). Fischer's correction appears to have escaped notice in another quarter thus Sandmeyer in a paper emanating from the laboratory of the Zurich Polytechnic (Ber. 20 641) identifies pyruvic acid by means of its phenyl-hydrazone which according to him " after dissolving once in dilute ammonia and precipitating with hydrochloric acid showed the correct melting point of 169"." Since the foregoing was written Professor Fischer has alluded to the subject of the melting point of phenylhydrazonepyruvic acid (Ber. 21 987). After pointing out that the acid as a substance which decomposes in melting will show too low a melting point unless rapidly heated he adds :-" It was evidently from this cause that Messrs. Japp and Klingemann who have overlooked both my correct,ed statement as to the melting point " [i.e .that 169" was a misprint for 192"] " and my remarks on the mode of heating found the melting point of the compound first at 182" and afterwards a t 185"." We stated (Ber. 20 3285) that in our first communication the melting point of the compound " was erro-neously given a t 182" instead of a t 1 8 5 O . " Professor Fischer somewhat hastily concludes that " erroneously given " is synonymous with '' erroneously found.'' As a matter of fact the error in question was an error in transcribing although we did not specify this at the time imagining that the cause of such an error-whether il misprint as in Professor Fischer's case or otherwise-could be of interest to no one save the authors.We have already expressed our regret that we overlooked Professor Fischer's correction. As regards the other oversight-of Professor Fischer's " remarks on the mode of heating"-we would point out that we expressly stated that the com-pound decomposed in melting and we are under the impression that the precaution to be taken in determining the melting point of substances of this character was There is a point here which calls for correction 534 JAPP AND KLISGEXANN THE COSSTITUTION OF neither case could we succeed in raising the melting point above 185" which we must therefore regard as the correct melting point. Both acids on heating yielded a crystalline compound of the formula C,F,H,,N1-diacetylosazone (vide supra)-together with oily matters. We also found that Fischer and Jourdan's acid gave the above characteristic colour reaction with concentrated sulphuric acid.Fischer and Jourdan give the melting point of the ethyl salt of their acid a t 114-115". We found f o r the ethyl salt prepared as above described the melting point 117". The metallic salts appear to be very unstable and could not be obtained in a state fit for anadysis. We reduced the acid obtained by the above method with sodium amalgam. By moderate reduction of the sodium salt in aqueous solution wiih sodium amalgam a hy drazo-acid was obtained which was precipitated on acidifying with hydrochloric acid avoiding an excess which was found to redissolve the hydrazo-acid. It crystnllised from methyl alcohol in slender colonrless needles which on exposure to air become yellow especially when moist.We found the melting point at 170° but had not sufEcient substance for further recrystalli-sation; the true melting point is 172'* (see later on). The sub-stance evolves gas in melting. Analysis gave figures agreeing with the formula C9H12N202 :-I. 0.1910 gram gave 0.4207 gram CO and 0.1174 gram H,O. 11. 0.0978 gmm burnt with copper oxide in a vacuum gave 22-63 C.C. of a mixture of nitrogen and nitric oxide measured dry a t 16.8" and under 433 mm. pressure. After absorption of the nitric oxide there remained 22.63 C.C. of dry nitrogen at 16.9" and under 430 mm. pressure. Found. Calculated for r-A--7 C9H,2N202* I. 11. c 60.00 60.07 H 6.67 6.83 -N 15-55 - 15.58 0 17.78 -. . . . . . . . . . - -known before Professor Fischer illustrated it in the case of the hydrazones.We made a point of heating as rapidly as is practicable. Professor Fischer's explanation will hardly account €or the fact that in the case of the corresponding hydrazo-acid-also a compound which decomposes in melting -we found the melting point 20" higher than he did. * The melting point (162") given in our first publication on this subject (Ber., 20 2943) is too low. The substance had not been sufficiently purified CERTAIN SO-CALLED MIXED AZO-COMPOUNDS. 535 This acid would be benzene-3-hydrazopropionic acid, CH,. CH-COOH N H-NH C,H,' Fischer and Jourdan (Ber. 16 2244) obtained this acid by the re-duction of phenylhydrazonepyruvic acid but assign to it the melting point 152-154".We therefore prepared phenylhyrazonepyruvic acid from phenylhydrazine and pyruvic acid and reduced it with sodium amalgam. The hydrazo-acid thus obtained was identical with the foregoing and after repeated recrystallisation from methyl alcohol had a constant melting point of 172". Aniline is always formed in this reduction as observed by Fischer and Jourdan. We therefore endeavoured to isolate the other product which would be formed by the breaking up of the niolecule of the hydrszo-acid by reduction-alanine. A quantity of phenylhydr-azonepyruvic acid was dissolved in dilute caustic soda and treated with excess of sodium anialgam on the water-bath. The aniline which was formed was then extracted with ether; the alkaline aqueous solution was slightly acidified with hydrochloric acid thus precipitating a small quantity of the hydrazo-acid which was removed by filtration ; the filtrate was evaporated to dryness and the residue treated with boiling absolute alcohol to extract the alanine hydrochloride which remained as a crystalline residue on evaporating the alcohol; t h i s residue was boiled with water and litharge; the filtered solution containing the lead salt of alanine, was precipitated with sulphuretted hydrogen ; and the filtrate from the lead sulphide was evaporated to a small bulk and then mixed with an excess of alcohol.Alanine separated in slender white needles. The alanine thus obtained had a sweet taste; sublimed above 200" and gave on analysis :-C. H. N. Calculated for C,E,NO? 40.45 7.86 15.73 per cent.Found 40.38 8.03 15-61 ,, E t h y l i c Methacetoacetate and o-Diazotoluene Chloride.-The re-action was conducted as in the former case. The " ester " was ob-tained as a red oil which did not solidify even in a freezing mixture : it was therefore hydrolysed by heating with caustic soda and alcohol. The acid was precipitated with hydrochloric acid and purified by recrystallisation from benzene. It formed small yellowish laminae, melting with evolution of gas a t 156". It is identical with Raschen's o-tolylh ydyazorzepyr.uz;ic acid (Annalen 239 228)' which we prepared from pyruvic acid and 0 . tolylhydrazine hydrochloride for the purpose of comparing it with our compound. Raschen give 536 JAPP AND KL~NGEMANN THE CONSTITUTION OF the melting-point at 158-159'.gave the following figures on analysis :-The acid prepared by our method I. 0.2214 gram gave 0.5051 gram COz and 0.1221 gram H,O. 11. 0.0988 gram burnt with copper oxide in a vacuum gave 22.63 C.C. of a mixture of nitrogen and nitric oxide measured dry at 15.7" and under 412 mm. pressure After absorption of the nitric oxide there remained 22.63 C.C. dry nitrogen at 16" and under 408 mm. pressure. Found. Calculated for +- -7 C,oH,&z02. I. 11. C 62 22 - 62.50 H 6.13 - 6-25 N 14-58 - 14.67 0 16.67 - -100~00 The hydrazo-acid which had not hitherto been prepared was obtained by reducing tbe acid from 0-diazotoluene and ethylic meth-acetoacetate with sodium amalgam. The acid was precipitated by careful addition of hydrochloric acid and recrystallised from methyl alcohol.It formed tufts of small colourless needles meltiiig at 143". This melting point is possibly too low as the quantity of substance was iiisufficien t for further recryatallisation. Substance. co,. H20. I 0.1911 0.4506 0.1262 11 0.2048 0.4617 0.1331 IIT. 0.0998 gram burnt with copper oxide in a vacuum gave 22-63 C.C. of nitrogen measured dry at 20" and under 412 mm. pressure. Nitric oxide was not present. Found. Calculated for I---- 7 CloHI,N?O?. I. 11. 111. C 61-86 61.45 61.48 -IT. 7.33 7.22 - 7-22 N . . 14.43 - - 14-39 0 . 16-49 100~000 - - -The acid is consequently the expected o-toluenehydmzopropionic acid, CH,*C H* C 0 OH NH.NHCsHa*CH,( 1 2). o-Toluidina was regenerated during the reduction CERTAIN SO-CALLED MIXED AZO-COVPOUNDS.537 Ethylic Jfethacetoacetate and p-Diazotoluene Chloride.-In this case, the " ester," when precipitated with water solidified. It was deposited in yellowish laminse from light petroleum and in needles from alcohol melting at 106". C. H. N. Calculated for C12H,,N,02 . . 65.45 7-27 12.73 per cent. B'ound 65.64 7.33 13.05 ,, The " ester " was hydrolysed with alcoholic caustic soda. The pre-cipitated acid was recrystallised from benzene from which it was deposited in yellow laminae melting at 162" with evolution of gas C. H. N. Calculated for C,,H12N,02 . . 62.50 6.25 14.58 per cent. Found. 62.48 6.28 14.54 ,, This acid is identical with Raschen's p-tolylhydraaonepyrucic acid (Annalen 239 224) which we prepared for comparison.Raschen gives the melting point of the acid at 158-160" ; that of the ethjl salt at 106-107". The hydrazo-acid was not prepared. Ethylic Ethacetoacetate and Diazobenzene Chloride.-In this case, diazobenzene chloride was added to the molecular proportion of ethylic ~~ sodethacetoacetate dissolved in alcohol. The new " ester," when precipitated with water did not solidify and was therefore hydro-lysed with alkali. The acid precipitated with hydrochloric acid, was purified by recrystallisation from benzene which deposited it in yellow needles with a silky lustre melting with decomposition at 152'. I. 0,1984 gram gave 0.4543 gram CO and 0,1139 gram H,O. 11. 0.0927 gram burnt with copper oxide in a vacuum gave 22.63 C.C. of nitrogen measured dry at ltjO and under 384 mm.pressure. Nitric oxide was not present. Found. Calculated for r---7 C,oH*2pJ,~2* I. 11. C 62.44 - 62-50 H 6-37 - 6.25 N 14-58 - 14-63 0 16.67 100-00 - --The acid is phen,ylhydrc~xonepr~p~onylformic acid, CH,*CH,.C*COOH N.NHC,H, 538 JAPP AND KLIKGEM-4NN THE CONSTITUTION OF I n the preliminary account of these reactious which we published (Bey. 20 2943) we described this acid regarding it however as " benzene-a-axobutyric acid." Later W. Wislicenus and E. Arnold (Ber. 20 3395) described an acid obtained from propionylformic acid and phenylhydrazine which ought to be identical with the fore-going. They state that it crystallises from dilute alcohol in lamine melting a t 144-145". By reduction of the sodium salt of the above acid in aqueous solu-tion with sodium amalgam benz ene-a- hy draxo but yric acid, CH,*CH,*CH.COOH NH-NHGH: was obtained.The hydrazo-acid precipitated by hydrochloric acid, avoiding an excess and recrystallised from methyl alcohol formed small colourless needles. The compound exhibits a peculiar be-haviour on heating. At 165" it softens without distinctly melting, and a substance which colours the sulphuric acid in the melting-point bulb violet distils out of the capillary tube; then the contents of the tube again become solid; and before the temperature of 245" is reached a t which a slight residue still left in the capillary tube melts, nearly everything has sublimed out of the tube. The substance was dried over sulphuric acid for analysis.Substance. cop H,O. I . 0.2098 0.4763 0.1395 I1 . 0-2047 0.4645 0.1337 111. 0.0970 gram burnt with copper oxide in a vacuum gave 22.63 C.C. of a mixture of nitrogen and nitrous oxide measured dry at 20" and under 406 mm. pressure. After absorption of the nitric oxide there i-ernained 22.63 C.C. of dry nitrogen at 20",and under 402 mm. pressure :-Found. Calculated for f--- -7 C10H14N202. I. 11. 111. C 61.86 61-91 61.88 -7-39 7-25 - H 7.22 N . . 14.43 - - 14-51 0 16-49 100.00 - - --4. Action of Diazo-salts on Ketonic Acids Formation of Monohydrazones of a-Dicarbonyl-conapou~d~. Acetoacetic Acid atad Diazobenzene Chloride.-65 grams of ethylic acetoacetate were dissolved in a solution of 30 grams of causti CERTAIN SO-CALLED MIXED AZO-COMPOUNDS.539 potash in 1120 grams of water and allowed to stand for 24 hours, after which the solution was acidified with hydrochloric acid in order to liberate the acetoacetic acid. A solution of the molecular propor-tion of diazobenzene chloride (from 47 grams of aniline and 35 grams of sodium nitrite) containing a slight excess of hydrochloric acid was then added. The acid solutions remain clear when first mixed; then a separation of pjruvaldehydrazone (" benzene-azo-acetone "), C H3.C 0.C H :N.NH*C6H5 cornm ences accompanied by an evolution of carbou dioxide ; but the reaction takes place very slowly and is not quite complete even at the end of eight hours. It is therefore better to add to the acid solution after mixing an excess of a solution of sodium acetate ; the reaction then takes place rapidly and an almost quantitative yield of a practically pure product is obtained.I n both cases the pyruvaldehydrazone separates in a crystalline form. It is probably advisable to use rather less diazo-salt than the theory requires. After recrystallising twice from benzene the pyruvaldehydrazone showed the proper melting point 149-150". It yielded a phenyl-hydrazine-derivative melting a t 14.5" which gave with concentrated sulphuric acid the characteristic colour reaction (p. 531). The above is both as regards yield and purity of the product, by far the best method as yet known of preparing pyruvaldehydr-azone. Methacetoacetic Acid and Dinxobenzene Chlo~ide.-The operation was conducted as in the foregoing experiment substituting 72 grams of ethylic methacetoacetnte for the ethylic acetoacetate.On adding sodium acetate to the hydrochloric acid solution of methaceto-acetic acid and diazobenzene chloride diacety Zkydrazone, CH,*C:N*NHCsH, CH,*CO 3 separated in minute needles with evolution of carbon dioxide. The crude product was treated with sodium carbonate to remove any yhenylhydrazonepyruvic acid that might have been formed from unhydrolysed ethylic methacetoacetate and then reci-ystallised from hot benzene which deposited it in yellow tabular crystals mostly grouped staircase-fashion melting a t 133". 0.2350 gram gave 0.5847 gram CO and 0.1414 gram H,O. Calculal ed for ~lOH,,N,O. Found. C . 68-18 67.85 per cent. H 6.82 6.68 ,, The monophenylhydrazone of diacetyl has been prepared from diacetyl and phenylhydrazine by v.Pechmann (Ber. 20 3164) wh 540 JAPP AND KLINGEJIANN THE CONSTITUTION OF gives the melting point at 133" with which our determination exactly agrees. On pouring a concentrated aqueous solution of phenylhydrazine hydrochloride into a hot alcoholic solution of the above nionohydr-line powder melting a t 243". v. Pechmann (Zoc. &.) gives 242" as the melting point of diacetylosazone which we also prepared from diacetyl and phenylhydrazine and found identical in all its proper-ties with the compound obtained by us. A nitrogen determination was made with our compound :-0.1074 gram burnt with copper oxide in a vacuum gave 22.63 C.C. of a mixture of nitrogen and nitric oxide measured dry a t 11" and under 627 mm.pressure. After absorption of the nitric oxide, there remained 22.63 C.C. of dry nitrogen a t 11" and under 621 mm. pressure. Calculated for C,,H,*NI. Found. N in 100 parts 21.05 20.88 Diacetylosazone gives a colour-reaction with concentrat'ed sulphuric acid. It dissolves in the cold with a brown colour ; but this speedily changes to a claret-red which in thin layers appears green. E'tlzacetoacetic Acid and Diazobenzene Chloride.-The experiment was performed as in the preceding cases exceptl that the ethylic ethnceto-acetate required longer shaking with the caustic potash solution to make it dissolve and owing to the greater difficulty with which it is hydrolysed it was allowed to stand in alkaline solution for 48 hours instead of for 24.In this case the propionylacetylhydrazone C2H5'c:N'NHCsH5 separated as a red oil on the addition of sodium CH3.CO acetate to the hydrochloric acid solution of ethacetoacetic acid and diazobenzene chloride. On dissolving this oil in a small quantity of alcohol the new compound was deposited in a solid form. The product was treated with sodium carbonate to remove any acid. From concentrated solutions in hot beiizene it crystallised in yellow needles or prisms radiating from a point ; from dilute solutions in large transparent plates. I n spite of the apparent purity of the substance the carbon was found to be 1 per cent. too low on analysis and the hydrogen was also too low. Only after repeated recrgstallisatior~ from benzene were numbers obtained agreeing fairly well with the theory :-It melted a t 116-117".8 ubatance. cop HZO. I . 02669 0-6762 0.1748 11. 02333 0.5908 0.150 CERTAIN SO-CALLED MIXED AZO-COMPOUNDS. 541 Found. 7 Calculated for T-h-C,,H,4N,O. J. 11. C 69-47 69-09 69.06 H 7-36 7.28 7.17 We suspect, therefore that the substance was contaminated with a small quantity of pyruvaldehydrazone which would be formed if the ethylic eth-acetoacetate contained as it very easily might unchanged ethylic acetoacetate. We find that propionylacetylhydrazone and pyruv-aldehydrazone crystallise together. Possiblya small quantity of the same impurity formed in the same way was present in the specimen of diacetylhydrazone which we analyaed (ante) in which the hydrogen was also found too low. Even in these analyses the hydrogen is too low.C!2H5.C:N'NHC6H5 was obtained The osazone of propionylacetyl CH,.C:N.NHC,H5, by heating an alcoholic solution of the foregoing compound with phenylhydrazine hydrochloride a i d sodium acetate. It crystallises from benzene in yellow needles melting at 162" which give with concentrated sulphnric acid exactly the same colour-reaction as the osazone of diacetyl (p. 540) .* 0.1012 gram burnt with copper oxide in a vacuum gave 22.63 C.C. of a mixture of nitrogen and nitric oxide measured dry at 14" and under 557 mm. pressure. After absorption of the nitric oxide there remained 22.63 C.C. of dry nitrogen at 14" and under 553 mm. pressure. Calculated for C17H20N4' Found. N in 100 parts 20.00 19.51 5. Action of Heat o n Hydrazones of Ketonic Acids and o n Alclehydrazorbes Formation of Osaxones.Action of Heat on Phenylhydrazonepyruvic Acid.-A quantity of phenylhydrazonepyruvic acid was heated in an oil-bat]h at 180-190" until gas ceased to be evolved. From the dark-coloured oil thus obtained a crystalline substance separated on adding alcohol. It was * No. 6 of the Berichte received after the above was in type contains an account of the hjdrazones of propionylncetyl by v. Pechmann. The author prepared these compounds by the direct action of phenylhydrazine on propionylacetyl. He finds that the monohydrazone obtained in this way is distinct from our compound and he therefore assigns to i t the formula-C2HS * C 0 CH3.C N*NHC,jH,' For the osazone he finds the melting point a t 161-162.5' agreeing wit11 our deter-mination 542 JAPP AND KLINGEMANN THE CONSTITUTION OF washed with hot alcohol in which it is only sparingly soluble and dried for analysis.It formed a yellow crystalline powder melting at 238". Analysis led t o the formula C,,H,,N,. I. 0,1717 gram gave 0.4536 grain CO and 0.1041 gram H,O. 11. 0.1047 gram burnt with copper oxide in a vacuum gave 22.63 C.C. of a mixture of nitrogen and nitric oxide measured dry at 14.7" and under 617 mm. pressure. After absorption of the nitric oxide there remained 22-63 C.C. of dry nitrogen at 15" and under 610 mm. pressure. Found. Calculated for r---7 C16H1RN4. I. 11. C 72.18 72.04 -H 6.77 6.73 N 21.05 - 20.78 -100*00 The compound was soluble in cold concentrated sulphuric acid with a brown colour changing afterwards t o a claret-red which in thin layers amppeared green.The formula and the colour-reaction with sulphuric acid are there-fore those of diacetylosnzone (p. 540) with which the compound is indeed identical. The difference in the melting point-238" instead of 242"-is due to the fact that the substitiice owing to the smallness of the quantity available was not recrystallised. The yield is very small. The formation of the osazone of diacetyl in the above reaction may be expressed thus :-CH3*C:N,HC,Hj I CH,*C:N2HC,H, Action of Heat on AZdehydrazone.-As already mentioned Fischer and Jourdan have shown (Zoc. cit.) that the oil which is obtained on heating phenylhydrazonepyruvic acid mainly consists of aldehydr-azone (ethylidenepheny1hydrazine)-CH3*C:N,HC,Hj = CH3*CH:N,H.CsH + CO,.COOH It therefore seemed possible that the osazone of diacetyl had been formed from the aldehydrazone in a secondary reaction according to the equation CERTAIN SO-CALLED MIXED AZO-COMPOUNDS. 54 3 We therefore heated aldehydrazone wit,h a reflax condensing- tube for three hours. On adding alcohol to the liquid product the ex-pected osazone separated as a crystalline powder. I t was identified by its melting point (237-238") and by the reaction with sulphuric acid. The yield is very small but may be somewhat increased by passing air through the aldehydrazone during the heating process. The osazone obtained by the above methods was carefully compared with a specimen prepared from diacetyl arid phenylhydrazine.Action of Heat on the Tolylh~clrcnzolzepyr~~vic Acids.-o-Tolylhydr-azonepyruvic acid when heated f o r some time at its melting point, yielded in like manner diacetyl-o-toZy ZosazolLe, CE3~C:N2H*C6H4~CH3 CH,-C:N,H.C,H,*CH (17 2)9 which separated from the oily product of the reaction on adding alcohol. It formed a yellow crystalline powder melting at 198", sparingly soluble in alcohol. The sulphuric acid reaction is not characteristic; the compound dissolves in the concentrated acid with a brown colour which afterwards changes to a reddish-brown. The same compound was obtained by heating an aqueous solution of diacetyl with o-tolylhydrazine hydrochloride and sodium acetate. I. 0.0827 gram burnt with copper oxide in a vacuum gave 22-63 C.C.of a mixture of nitrogen and nitric oxide measured dry at 15.5" and under 448 mm. pressure. After absorption of the nitric oxide there remained 22.63 C.C. of dry nitrogen at 15.5" and under 444 mm. pressure. 11. 0.0754 gram burnt with copper oxide in a vacuum gave 22.63 C.C. of a mixture of nitrogen and nitric oxide measured dry at 14" and under 402 mm. pressure. After absorption of nitric oxide there remained 22.63 C.C. of dry nitrogen at 14" and under 400 mm. pressure. Found. r. Calculated for C18H22N4. I. N in 100 parts . . . . . . . 19.05 19-08 18.92 Analysis I was made with a preparation obtained from 0-tolyl-hydrazoriepyruvic acid ; analysis I1 with a preparation from diacetyl and o-tolylhydrazine. p-Tolylhydrazonepyruvic acid yields when heated diacety Z-p-toty Zos- -azone " ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ (1 4).It was separated like the foregoing compounds of this class and formed a yellow crystalline powde 5-14 SHENSTONE AND CUNDALL INFLUENCE OF TEMPERATURE melting a t 229-230". It dissolves in cold concentrated sulphuric acid with a brown colour which changes through yellow and greenish-yellow to green. 0*1101 gram burnt with copper oxide in a vacuum gave 22.63 C.C. OE a mixture of nitrogen and nitric oxide measured dry a t 11" and under 585 mm. pressure. After absorption of the nitric oxide there remained 22-63 C.C. of dry nitrogen a t 11" and under 580 mm. pressure. Calculated for c 1 8H22N4. Found. N in 100 parts . . . . . . . . 19.05 19-02 The same compound was obtained by the action of free p-tolyl-hydrazine on diacetyl in ethereal solution. Curiously enough when we attempted to prepare the compound like its ortho-isomeiide, by heating an aqueous solution of diacetyl with p-tolylhydrazine hydrochloride and sodium acetate only the mono-y-toZyZhydrazone, CH3*c:N2H'C6H4*CH3 (l' was formed. It crystallises from benzene CH,*CO in slender yellow needles melting at 161". A nitrogen determination gave 14-91 per cent. Calculated N 14.73 per cent. Normal School of Xcience, South Kensington