FENTON AND JONES: RELATIONSXIPS OF OXALACETIC ACID. 91 VII I.-Re1 a tionships of Oxalacetic Acid. By HENRY J. HORSTMAN FENTON, F.R.S., and HUMPHREY OWEN JONES, B.A., B.Sc. IT has been shown by the authors in a previous communication (Trans., 1900, 77, '77) that malic acid, when oxidised by hydrogen dioxide in presence of ferrous iron, gives free oxalacetic acid, and that the latter may readily be isolated by extraction with ether under certain con- ditions, Some of the properties of the acid were then described, and the present paper gives a n account of further studies in this direction. Special attention is given to such reactions as necessitate the use of the free acid rather than its esters, since the latter have been well in- vestigated. Action of PhenyZhydraxine.-It was previously shown that on mix- ing molecular proportions of phenylhydrazine acetate and oxalacetic acid in aqueous solution, in the cold, the hydrazone is precipitated and is obtained on recrystallisation from ether in the form of coloiirless, transparent prisms." This substance when heated turns yellow and decomposes without melting at about 95O, and on crystallising the resulting product from hot alcohol pale yellow needles are obtained which melt a t 192'. It was shown in the previous paper t h a t when the hydrazone of oxalacetic acid is heated with excess of dilute sulphuric acid, it is changed almost immediately into a crystalline mass consisting of l-phenyl-5-pyrazolone-3-carboxylic acid, identical with that previously obtained from the esters by Wislicenus and by Buchner.No visible evolution of gas occurs in this reaction. It has now been observed that if pure water be substituted for the dilute acid, a brisk evolution of carbon dioxide takes place ; the whole goes into solution, and almost immediately a crystalline precipitate separates. This, when recrystal- lised from alcohol, separates in pale yellow needles which melt at 19Z0, and is evidently identical with the product obtained by the action of heat alone. It coincides exactly in melting point and other properties with the hydrazone of pyruvic acid. On analysis the following results were obtained : * That this is the hydrazone, and not the hydrazide, is proved by its behaviour as a dibasic acid. 0'3940 required 33-9 C.C. of N/10 KOH for neutralisation, the calculated amount being 35.5 C.C.A solution of the substance in dilute alcohol was precipitated with excess of silver nitrate dissolved in the same solvent, the resulting salt being washed, and dried in a vacuum desiccator in the absence of light. Ag=49*77. C,,H,O,NBAg, requires Ag= 49'54 per cent. For example : 0.2883 gave 0.1428 Ag.92 FENTON AND JONES: XELATIONSHIPS OF OXALACETIC ACID. 0.1506 gave 0.3334 CO, and 0.0761 H20. The changes which take place are therefore represented as follows : (1) When excess of dilute sulphuric acid is used : (2) when water only is employed : C = 60.37 ; H = 5.61. C9Hlo0,N, requires C = 60.67 ; H = 5.61 per cent. CIOW1,o,N, = C10H,03N2 + H,07 CloHlo0,N2 = C9Nl0O2N, + CO,. When the concentration of the acid falls below a certain value, both reactions occur simultaneously, and the nature of the change can easily be followed by measurement of the evolved carbon dioxide.Using decinormal sulphuric acid, for example, about 26 per cent. of the hydrazone was found to undergo the second change, and the proportion of carbon dioxide became less and less as more concentrated solutions of the acid were employed, until with acid of about normal strength the amount of gas was not perceptible. Experi- ments were now made with other acids, using decinormal solutions and making the conditions exactly the same in each case, and it will be seen that the stability of the hydrazone, as regards retention of the 4-carbon molecule, is evidently a function of the concentration of the hydrogen ions.The experiments were made by placing about 0.1 gram of the hydrazone in a tube having a capacity of about 15 c.c., covering it with 7.5 C.C. of the decinormal acid, and heating t o boiling for about 2 or 3 minutes in a water-bath. The evolved carbon dioxide was collected in a Lunge's nitrometer, the residual gas swept out of the tube by a current of about 40 C.C. of purified air, and the estimation made by absorption with potash in the usual manner. The accompanying sketch will give an idea of the apparatus employed. A preliminary experiment was made with pure Iceland spar in order t o test the apparatus, and the result showed that the accuracy was well within the limit desired for the purpose, as 0.1065 gram gave 23.6 C.C. of carbon dioxide (cow.), the calculated volume being 23% C.C.The order obtained in the table on p. 93 very closely approximates to that of the relative strengths or ' affinities ' of the acids as measured by the well-known methods. In the case of trichloroacetic acid (with regard to which there was a t one time a discrepancy), it is probableFENTON AND JONES : RELATIONSHIPS OF OXALACETIC ACID. 93 Corrected volume of co,. that the high result is due to hydrolytic action a t the temperature employed. It would appear, therefore, that the process described affords a very simple method of comparing the affinities of acids a t looo, that is, of placing them in ' order of merit.' The exact interpretation of the numbers themselves can scarcely be arrived at from the few and relatively rough determinations here recorded, but the authors hope to make more extended observations in this direction.If it be admit- ted, as above conjectured, that the results depend upon the con- centration of the hydrogen ions, it is easy t o understand why the full proportion of carbon dioxide is not obtained even in the case of pure water, since the resulting product is also an acid, and therefore tends to increase the stability of the still undecomposed hydrazone. CO, for 1 gram substance. Weight of substance. Nature of acid. I- Nitric.. .......................... Hydrochloric .................. Suiphuric ....................... Sulphuric.. ..................... TrichIoroacetic .............. Tartaric ........................ Malic ............................ Succinic.. ......................Citric .......................... Acetic ........................... [Pure water .................... 0.0992 0.0993 0.0990 0*1050 0'0980 0.0990 0-0990 0-0998 0.0995 G.0984 0.1200 I 1.65 1'73 2.56 2.84 3-96 6 4 8 6 *76 7 -21 7 -30 7.72 10.90 16.6 17.4 25 '8 27-0 50.8 66'9 68.2 72 *I 73 '3 78-6 90'01 Action of Hydrazine.-When well cooled alcoholic solutions of oxalacetic acid and hydrazine hydrate are mixed in molecular propor- tion, the mixture being surrounded by ice and salt, an immediate turbidity is produced and a heavy oil separates which is quickly changed to a hard, white, apparently amorphous mass. This was separated from the liquid, the examination of which will be described below, thoroughly ground with several changes of absolute alcohol and dried in a vacuum desiccator.The yield was about half the weight of acid taken. It dissolves easily in cold water, the solution gives a red colour with ferric chloride, and quickly reduces Fehling's solution in the cold. It melts sharply and with violent decomposition a t 99'. On analysis, the following results were obtained : 0.1656 gave 0.1614 CO, and 0.0835 H,O. C = 26.58 ; H = 5.60. 0.1041 ,, 28.6 C.C. nitrogen a t 18.5' and 734 mm. N=31*17. C4Hlo0,N, requires C= 26.97 ; H = 5.61 ; N = 31.46 per cent. The substance is therefore the Aydrazine salt of the hydrazone, or of94 PENTON AND JONES : RELATIONSHIPS OF OXALACETIC ACID. the hydmzide of oxalacetic acid, C0,H*C(N,H2)~CH2*C02H,N2H, or C0,H G O CH,* CO*N,H,,N,H,.* When an aqueous solution of this product is heated for a short time in a water-bath it undergoes a marked change.The solution now gives with ferric chloride a very intense brownish-violet colour, and after concentration the solution sets to a soft mass which is seen under the microscope to consist of long, transparent prisms. These, after washing with alcohol and drying in a vacuum desiccator, melted at 196'. On analysis, the following result WAS obtained : 0*1696 gave 58.6 C.C. nitrogen at 20' and 737 mm. C,H,O,N, requires N = 39 -43 per cent. Rothenburg (Bey., 1892, 25, 344l), by the action of hydrazine hydrate on ethyl oxalacetate, obtained (1) a product easily soluble in alcohol, which is the ester of pyrazolone-3-carboxylic acid, and (2) one nearly insoluble in alcohol which melted at 2 3 8 O , and had the composition C,H,O,N,.The latter he considers to be the hydrazide of py razolone-3-carbox y lic acid (Pyrazolon-3-carbonyl h y drstzin), N= 39.16. It appears highly probable that this second product is identical with the one obtained in the present case, although there is considerable discrepancy as to the melting point. The alcoholic solution which remained after separating the hydrazine salt above described was allowed to evaporate in a vacuum desiccator, and the partly solid residue was treated with cold water, In this way a nearly white, granular residue was left which is very aparingly soluble in cold water, and separates from boiling water in minute crystalline aggregates. For analysis, it was purified by dissolving i n alcohol, precipitating with ether and light petroleum, and drying in a vacuum.It turned dark brown, or nearly black, without melting, at about 260-270'. On analysis : 0*1110 gave 0.1514 CO, and 0.0353 H20. C=37*20; H=3.53. 0.1102 ,, 20.9 C.C. nitrogen at 20' and 735 mm. N = 21.46. C,H,O,N, requires C = 37-50 ; H = 3.12 ; N = 21-87 per cent. This product therefore corresponds entirely in composition and pro- perties w it h the pyazolonecarboxy Zic acid which Rot henburg obtained from the ester above mentioned by hydrolysis with alkalis or con- centrated hydrochloric acid. Action. of Hydroxy~amirze.-when an alcoholic solution of hydroxyl- amine, prepared by Wohl's method (Ber., 1893, 26, 730) is mixed * This composition explains the comparatively small yield, since only one mol. of hydrazine was employed.l?ENTON AND JONES : RELATIONSHIPS OF OXAT,ACETIC ACID, 95 with oxalacetic acid, also dissolved in alcohol, in molecular proportion, a turbidity is produced which almost immediately disappears, and the solution after standing for some hours and evaporation in a vacuum desiccator leaves a thick syrup which sets to a white, crystalline mass on stirring or on long standing. This when recrystallised from anhydrous ether melts and decomposes with violent frothing at 133' when gradually heated, or 142-143' if quickly heated. It is very soluble in water or alcohol, but less so in ether ; with ferric chloride, the solution gives a, brownish or yellow colour.On analysis : 0.1758 gave 0.2135 GO, and 0.0550 H20. C,H50,N requires C = 32.65 ; H = 3.40 per cent. When the substance is mixed with excess of acetic anhydride or acetyl chloride, i t dissolves, and if the resulting solution be allowed to stand for a day or two and then kept in a vacuum desiccator over solid potash and sulphuric acid, the product on treatment with water gives an intense indigo-blue colour with ferric chloride, Piutti (Chem.Centr., 1888, 68; Gaxx., 1888, 18, 457), by the action of the sodium derivative of ethyl oxalacetate upon hydroxylamine hydrochloride, obtained the compound (CO,Et),*CH,*G:NOH, and the compound CO,Et*CH,*G(NOH)*CO,H by partial hydrolysis. From the latter, Cramer (Be?., 1891, 24, 1206) obtained the free acid P-oximidosuccinic acid or the 6 oxime of oxalacetic acid.' This melted at 88' and gave a characteristic 6Zue colour with ferric chloride." Ebert (Annalen, 1885,229, 76), by the action of nitrous acid on ethyl succinosuccinate, had obtained an acid having the same composition, but this decomposed with frothing a t 126O, and gave, with ferric chloride, a brown o r yellow colour.Cramer (Zoc. cit.) showed that these two acids are stereoisomerides, and gives reasons for assigning to them the following constitutions (see also Hantzsch, Ber., 1891, 24, 1192) : C=33*12; Hz3.47. a-acid (Ebert's) c02"* 8 *CH2* (labile), HO*N @acid (Cramer's) Co2H*8 *CH2* (stable). N*OH By the action of concentrated sulphuric acid, acetyl chloride, or acetic anhydride the a-acid (or its compounds) can be transformed into the p-form, the product then giving the characteristic blue or violet colour with ferric chloride. It would naturally have been expected that the action of hydroxyl- * From the somewhat high percentage of carbon in the product obtained in the present case, it is possible that it is not absolutely pure j the numbers, however, are practically identical with those obtained by Cramer, who found C=33'16 and 33.22 ; H=3*10 and 3'16 per cent.96 FENTON AND JONES: RELATIONSHIPS OF OXALACETIC ACID.amine upon free oxalacetic acid would give rise to the P-acid, but this evidently is not the case. The properties of the product obtained in the present case closely resemble those of Ebert's a-acid, and the action of acetic anhydride transforms it into a derivative of the P-form. This difference between the behaviour of the free acid and the ethyl ester may possibly be due to the tautomeric difference previously suggested (Michael and Bucher, Ber., 1896, 29, 1792).Action of Ammonia.-By mixing together alcoholic solutions of oxal- acetic acid and ammonia in different proportions, white precipitates are immediately produced which, after washing with alcohol and drying in a vacuum desiccator, have a composition varying between that required for a normal and an acid salt. If, however, a fairly dilute solution of the acid in absolute alcohol be poured into an excess (slightly more than 3 mols.) of alcoholic ammonia with continuous and vigorous shak- ing, the precipitate obtained, after washing and drying as before, corresponds in composition to the normal salt. 0.1638 gave 24.9 C.C. nitrogen a t 22' and 750 mm.The salt melts and decomposes at 75-77', giving a yellow liquid which solidifies t o a crystalline mass on cooling. It has a neutral re- action, and gives a deep red colour with ferric chloride. When the ammonium salt is left exposed t o the air, it is transformed into a yellow, gummy mass which no longer gives the red colour with ferric chloride. I n order to ascertain whether a further action of ammonia could be induced, the ammonium salt was heated in a stoppered bottle for some hours at 100' with alcoholic ammonia in large excess. A white, crys- talline crust of ammonium carbonate was obtained, and a yellow, gummy mass separated at the bottom of the vessel (the alcoholic liquid left only a trace of residue on evaporation). This gummy mass was dissolved in water and acidified with dilute sulphuric acid.A white precipitate was thus obtained which, after washing and drying, melted and decom- posed at 273O, gave a reddish-yellow colour with ferrous sulphate and appeared to agree in all respects with the methylpyridinecarboxy~~c acid (m. p. 2'74') which Bottinger (Annalen, 1877, 188, 330) obtained by the action of alcoholic ammonia on pyruvic acid. Action of Urea.-On mixing alcoholic solutions of oxalacetic acid (1 mol.) and urea (rather over 2 mols.) and allowing the mixture to stand in a vessel surrounded by ice, a crystalline precipitate is formed on stirring. This, when washed with alcohol and dried in a va,cuum, melts and decomposes at 124' and has the composition of a normal urea salt. N= 17.27. C,H40,,2NH, requires N= 16.86 per cent.0.1866 gave 23.5 C.C. nitrogen a t 17' and 747 mm. N= 14.62. C,H,0,,2CON,H4 requires N = 14.58 per cent,FENTON AND JONES: RELATIONSHIPS OF OXALACETIC ACID. 97 The salt is fairly easily soluble in water and gives a deep red colour with ferric chloride. When heated alone or with phosphorus oxy- chloride, it yields crystalline products which have not yet been examined, Action of Aniline.-When ethereal solutions of the acid and aniline are mixed in molecular proportion, a white precipitate is formed which melts a t 65-66'. An aqueous solution of this salt remained clear for over a week and gave no precipitate on boiling. If the enolic formula be ascribed to oxalacetic acid, this observation would further confirm the hydroxyfumaric constitution as suggested by Nef and by Michael (compare Michael, Bey., 1886, 19, 1372; Michael and Bucher, Zoc.Action of Benzylphenylhydi.axine.-On mixing oxalacetic acid, dis- solved in a small quantity of water, with benzylphenylhydrazine dis- solved in acetic acid, in molecular proportion, and cooling the mixture with ice, a certain quantity of gummy substance separates at first, but after standing for several hours a nearly white precipitate is obtained. This was collected and washed, first with strong and then with dilute acetic acid, dried in a vacuum desiccator, and crystallised from ether. It melted at 105-106'. cit.). 0.1207 gave 10.4 C.C. nitrogen a t 18Oand 757 mm. N = 10.20 per cent. The substance is therefore probably not an oxalacetic acid derivative (the benzylphenylhydraxone of which would require N = 8.9 per cent.), but the corresponding hydrazone of pywvic acid, C16H2602N2, which requires N = 10.44 per cent.On using alcoholic solutions of the acid and the hydrazine, a yellow gum is obtained, which never quite solidifies, even on long standing in a desiccator. By treating this with light petroleum and dissolving the residue in ether, a certain amount of yellow, crystalline substance was obtained, but the quantity was too small for examination. Oxidation of the Acid.-When an aqueous solution of the acid is acted upon by bromine in molecular proportion, a product is obtained, as was previously shown, which gives an intense bluish-violet colour with ferric chloride in presence of caustic alkalis, whilst this on acidification with dilute sulphuric acid is changed to a transient emerald green.The colour-changes, in fact, exactly resemble those given by hydroxymaleic acid with the same reagents (Trans., 1894, 65, 904). If other solvents, such as acetic acid or ethyl acetate, be employed instead of water, a similar result is obtained, but the violet colour appears more slowly. It is probable that a bromo-acid (bromoxalacetic or bromohydroxyfumaric acid) is first produced and then hydrolysed. On evaporating the acetic acid or ethyl acetate solution in a vacuum, or by extracting the aqueous product with VOL. LXXIX. H98 FENTON AND JONES: RELATIONSHIPS OF OXALACETIC ACID. ether, a thick syrup is obtained which does not crystallise but still gives the ferric chloride reaction." A product giving similar colour reaction is o6tained when oxalacetic acid is oxidised with the calculated quantity of hydrogen dioxide in presence of ferrous iron, or when malic acid is so oxidised, using nearly two atoms of oxygen to one molecule of the latter acid.All attempts to isolate this colour-giving acid have so far been unsuc- cessful. It is evidently very unstable and does not separate under the same conditions as those employed in the isolation of dihydroxymaleic acid nor can i t be extracted with ether. The authors are, therefore, inclined to the opinion that it is isomeric and not identical with the latter acid, being probably the ketonic modification, CO,H*CO*CH(OH)*CO,H. Further light is thrown upon this question by the action of phenyl- hydrazine. It was shown that the solution obtained by oxidising malic acid in presence of iron gives with phenylhydrazine in the cold a voluminous orange precipitate, which by recrystallisation from hot chloroform is obtained in brilliant orange needles melting at 217-219'.This product gave very constant results on analysis, the mean of four concordant analyses made with distinct specimens, prepared on different occasions, giving, C = 63.20 ; I3 = 5.17 ; N = 20.16 per cent.? These numbers do not correspond to those required for a derivative of oxalacetic acid, but t o a more oxidised product, and since no such compound can be obtained by the direct action of phenylhydrazine on oxalacetic acid, it is probable that it is the result of further oxidation by the hydrogen dioxide, which attacks a portion of the oxalacetic acid as fast as i t is formed.It can, in fact, be prepared by oxidising oxalacetic acid in presence of ferrous iron, or by adding a further quantity of the dioxide in the oxidation of malic acid, and subsequent addition of phenylhydrazine acetate. Prepared in the latter manner, the recrystallised product melted at 223-224'. On analysis, the following result was obtained : 0.1259 gave 0,2909 CO, and 0.0561 H,O. * When excess of bromine was employed and the mixture heated, crystalline plates were obtained which no longer gave the above reaction. They melted at about 92" and gave, on analysis, Br=64.61 per cent. This product is evidently dibromopyruvic acid, which requires Br = 65.04 per cent. .t. The properties of this substance are, in some respects, very similar to those of the osaaone of hydroxypyruvic acid discovered by Nastvogel, and in composition the two nubstances do not differ widely.C,,H,,N40, requires C = 63'82, H = 4-96, N=19*85 per cent. The two conipounds are, however, at once distinguished by the melting points and by other properties which will be described below. C=63*01; H=4*94.FENTON AND JONES : RELATIONSHIPS OF OXALACETIC ACID. 99 These numbers correspond to those required for a derivative ob- tained by the addition of three molecules of phenylhydrazine to one molecule of either dihydroxymaleic or dioxosuccinic acid,* with elim- ination of three molecules of water, C,H,O, + 3N,H,Ph - 3H,O requires C = 63.15 ; H = 5-26 ; N = 20.09 C,H,O, + 3N2H,Ph - 3H20 requires C = 6346 ; H = 4.80 ; N = 20.19 I n order to throw light upon the constitution of the compound, the following experiments were made, The well dried substance was mixed with excess (about ten times its weight) of acetic anhydride and heated on a water-bath, when a considerable evolution of gas took place and a clear, dark coloured solution resulted.The latter, on standing, deposited a mass of orange crystals, which when recrystallised from hot alcohol appeared in the form of beautiful, bright, orange-red needles melting a t 149-1 50'. On analysis : per cent. per cent. 0.1502 gave 0.3729 CO, and 0.0616 H,O. C1,H,,ON, requires C = 68.18 ; H.= 4-54 per cent. The composition, melting point, and other properties of this product show that it is identical with that which Knorr obtained by the action of acetic anhydride upon the osazone of hydroxypyruvic acid (Bey., 1888, 21, 1201), and which Will prepared from the same source by tho action of hydrogen chloride in alcoholic solution (Bey., 1891, 24, 3831).It was also obtained by Knorr (Zoc. c k ) by heating the corre- sponding pyrazolonecarboxylic acid at its melting point, and was shown by him to be phenylhydrazineketophenylpyrazolone, C = 67-72 ; H = 4.55. r P h * C O N= CH >C:N*NHPh. This being a derivative from a 3-carbon acid, it was important to ascertain the nature of the other products of the reaction; and the experiment was therefore repeated in such a manner that the gaseous product could be displaced by a current of inert gas, the substance being heated with acetic anhydride as before, a current of purified nitrogen passed through the flask, and the product collected in an excess of lime-water.In this way, abundant evidence of the presence of carbow dioxide was obtained. After the resulting mixture had been allowed to stand for a day or two, the mother liquor from the crystals was mixed with excess of sodium hydroxide solution and * That is ' anhydrous ' clihpdroxytartnric acid, C02H*CO*CO*C0,H. Compare Anschiita and Parlato (Ber., 1892, 25, 1977). l I 2100 FENTON AND JONES : RELATIONSHIPS OF OXALACETIC ACID. distilled in steam, when a notable quantity of aniline was found in the distillate. These experiments fairly well establish the fact that the substance in question is a 4-carbon derivative, and it must therefore be regarded either as FO*CO,H CH(OH)* CO,H’ (1) The dihydrazide-hydrazone of the ketonic acid, tautomeric with dihydroxymaleic acid, or as (2) The monohydrazide-dihydrazone of dioxosuccinic acid.* F( N,HPh)*CO*N,H,Ph S(N,HPh)*CO,H (l) CH(OH)*CO*N2H2Ph Or (2) C(N,HPh)*CO*N,H,Ph I n the first case, the formation of aniline can easily be explained by the oxidation of the CH*OH group by phenylhydrazine, and the change might be expressed by the equation The second formula, however, perhaps better explains the acidic character of the substance (for example, the formation of a crystalline sodium salt when heated with sodium carbonate solution), and in this case the production of aniline must be accounted for in the decompo- sition of the phenylhydrazine residue by heat (compare Reissert and Kayser, Ber., 1890, 23, 3703).I n either case, the initial product would be the ketonic acid, CO,H*CO*CH(OH)*CO,H. It has already been suggested (Fenton, Trans., 1896, 69, 547) that dihydroxymaleic acid may undergo tautomeric change into this form, and further evidence in this direc- tion mill be afforded in a subsequent communication. Experiments were therefore made in order to ascertain whether the product at present under discussion could be produced from dihydroxy- maleic acid. It has previously been shown (ibid., 548) that by the action of phenylhydrazine on this acid in acetic acid or alcoholic solution, the normal phenylhydrazine salt is produced, C,H40,,2N,H3Ph. This salt was now heated on a water-bath with excess (about 2 mols.) of phenylhydrazine acetate and water for about 2 hours.The result was an orange-coloured product, which, when recrystallised from hot chloroform, melted at 222O and had all the properties of the malic acid derivative under discussion. 0.1210 gave 20.8 C.C. nitrogen a t 20° and 756 mm. N = 19.98 per cent. .It It might, of course, be the phenylhydrazine salt of the corresponding pyrazolone- carboxylic acid ; but this is improbable from the fact that the sodium salt, when decomposed by hydrogen chloride, gives back a product which melts practically at the same temperature, and appears to be identical with the parent substance.FENTON AND JONES : RELATIONSHIPS OF OXALACETIC ACID. 101 Reviewing the whole of this evidence, the authors are inclined to prefer the second formula suggested for the substance, that is, to re- gard it as the monohydrazide-dihydrazone of dioxosuccinic acid, since it seems improbable that the CH*OH group represented in the first formula would resist the action of phenylhydrazine in the experiment last recorded; this formula also agrees somewhat bebter with the results of analysis.Surnmayy. The principal results obtained in the present investigation may briefly be summed up as follows. Oxalacetic acid by interaction with phenylhydrazine yields the phenyZ- hydrazone, CO,H-CH,*C( N,HPh)*CO,H, which readily loses carbon dioxide when heated with pure water giving the hydrazone of pyruvic acid. This decomposition is prevented by acids if the latter are of suffi- cient concentration, the result in this case being phenylpyrazolone-3-carb- oxylic acid. By comparing the influence of different acids in this respect, i t is easy to compare their relative ‘ strengths ’ or affinities.’ Hydrazine hydrate gives the hydrazine salt of the hydrazone CO,H*CH,*C(N,H,)- CO,H,N,H, together with Rot henburg’s pyyazo- Zonecarboxylic acid. The first named salt when heated with water gives a white, crystalline compound which is probably the jiydrcczide of the latter acid. Hydroxylaminegivesrise to the compound CO,H*CH,*C(NOH)*CO,H which contrary t o expectation appears to be identical with Ebert’s a-oximidosuccinic acid and not the /3-compound. It yields, however, a derivative of the latter by action of acetic anhydride or acetyl chloride. Ammonia and urea give the respective normal salts and the former salt, when heated with excess of alcoholic ammonia, gives a substance corresponding with Bottinger’s methylpyridinecarboxylic acid. When oxalacetic acid is oxidised with hydrogen dioxide in presence of iron, a product is obtained closely resembling dihgdroxymaleic acid in properties but which has not yet been isolated. The authors give reasons for considering that this acid has tbe tautomeric formula CO,H-CO*CH(OH)*CO,H. When treated with phenylhydrazine, this product gives a bright orange compound melting at 222-224O. The latter has been the subject of an exhaustive examination and the evi- dence leaves little room for doubt that it is the hydrazide-dihydrazone of dioxosuccinic acid.