72 FORSTER AND FIERZ: THE THIAZO-GROUP. PART I.VIII.-The Ft*iazo-G.i*oup. Part I. Triuzoacetic Acidand Triazoacet one (Ace t ony lazoimid e) .By MARTIN ONSLOW FORSTER and HANS EDUARD FIERZ.DESPITE the variety of triazo derivatives which have been investigated,principally by Griess, Curtius, Noel ting, and their collaborators duringthe past forty pars, one class of these compounds would appear tohave escaped examination; we refer to those types in which the triazo-group has replaced hydrogen in carboxylic acids, ketones, aldehydes,and alcohols of the aliphatic series. So far as we have been able toascertain, the only triazo-ketone to be met with in the literature iscamphorylazoimide (Trans., 1905, 87, 836), and it is chiefly inconnexion with this compound that our attention has been drawn tothe subject.The " triazoacetic acid '' described by Cartius and Lang(J. p . C'hem., 1888, [ii], 38, 532), at first regarded as having theconstitution FOHSTER AND FIERZ: THE TRIAZO-GROUP. PART I. 73is now recognised as a bimolecular polymeride of diazoacetic acid;entitled bisdiazoacetic acid, it is variously represented asIn the present communication, we describe triazoacetic acid,N3*CH2*C0,H, its amide, and ethyl ester, along with the simplesttriazo-ketone, acetonylazoimide or triazoacetone, CH,*CO*CH,*N,.The last-named substance is the true analogue of camphorylazoimide,which it recalls in respect of its behaviour towards potash; theaction of alkali on the simpler molecule is very vigorous, alcoholicpotash liberating nitrogen from acetonylazoimide almost explosively,whilst the aqueous alkali, even when very dilute (2 per cent.), sets upa brisk effervescence. There is a distinction, however, betweencamphorylazoimide and triazoacetone in regard t o the quantitativeaspect of the change.We have shown that exactly two-thirds of thenitrogen is liberated from the camphor derivative, which passes intothe imine of camphorquinone :but the gasometric study of the alteration which acetonylazoimideundergoes with alkali shows that subsidiary to the change,CH,*CO*CH2*N, -+ CH,*CO*CH:NH --+ CH,*CO*CH:O,there occurs, t o the extent of, roughly, 15 per cent., elimination ofhydrazoic acid,CH,*CO*CH2*N, + H,O = CH,*CO*CH,*OH + HN,,which does not take place when camphorylazoimide is treated withboiling alkali.This difference in behaviour probably owes its originto the absence of the cycloid structure, which renders the camphorderivative more stable, and also to the more hydrcgenised condition ofthe carbon atom to which the triazo-group is attached in acetonyl-azoimide, as well as to the comparative freedom of the unsubstitutedmethyl group, because a preliminary experiment with 2-triazocyclo-hexanone has indicated that in this case, also, both changes proceedsimultaneously :Yields nitrogen and Yields nitrogen and Yields nitrogen only.hydrazoic acid. hydrazoic acid.The facility with which nitrogen is eliminated from these compound74 FORSTER AND FIERZ: THE TRIAZO-GROUP. PART 1.appears to depend on the immediate neighbourhood of t h e ketonic andtriazo-groups, and leads us to suggest, as a possible explanation, t h a tthe action of alkali is such as to bring about the formation of anunstable ring-system from which nitrogen is forthwith liberated :-CH*N<R -C:NH*R -C:NH-co -f ]/ --+ N2+ I-C--O -c:o IAccepting the common view of the triazo-group, a n d comparing itwith the diazo-complex as occurring in diazomethane and diazoaceticester,N--N<I I N CH,<#CO,Et*CH<N, Nit will be recognised that two-thirds of the nit,rogen in the first-namedradicle stands in that relation to the remainder which is borne by thenitrogen in alphatic diazo-compounds to the carbon with which i t iscombined.Now there exists a considerable body of evidence to showt h a t the nitrogen in diazomethane and diazoacetic ester is capable oftaking part in additive action, the products parting easily withthe element in question.For instance, Buchner (Annulan, 1893, 273,214; compare also Buchner and Papendieck, Zoc. cit., 232 and 246;Buchner and Witter, Zoc. cit., 239) found that when ethyl diazo-acetate acts on the esters of unsaturated acids, pyrazolinedicarboxylicesters are produced, which lose nitrogen when heated, yielding tri-methyleued i ca rboxy 1 i c esters :...................C*2Me*RH + W>CH*C02Me = CO,Me* 7”-i-N: ......... s;T iCH, N CH,--CH;C02Me ;von Pechmann, again (Bey., 1894, 27, lS90), obtained methylpyrazoline-4 : 5-dicarboxylate by the interaction of diazomethane andmekhyl fumarate,CO,Me*?H---C I H 2..+ QH”N= co2Me* CH:-N: N j’CO,Me*EHCO,Me*CH N= ......................the pyrazolinecarboxy lic esters being here represented in their pseudo-form t o indicate more clearly the loss of nitrogen.Another ring-system which readily parts with the element arises from diazomethaneand thiocarbimides, von Pechmann and Nold (Bey., 1896, 29, 2588)having shown that phenylaminothiodiazole,behaves in this manner when fused, whilst the recent conversionof aldehydes into methyl ketones by the action of diazomethanFORSTER AND FIERZ : THE THIAZO-GROUP. PART 1. 75(Schlotterbeck, Ber., 1907, 40, 479) has been explained with someplausibility by assuming the intermediate formation of furodiazoles,Hence the possibility of mutual satisfaction of affinity between thecarbonyl and triazo-groups appears by no means remote, and isindependent of the alternative representations,-N:NiN and -CH:NiN,which might be used for the azoimide complex and aliphatic diazo-radicle respectively.Another argument in favour of this explanation may be drawn fromthe behnviour of triazoformic, as compared with that of triazoacetic,ester. When ethyl triazoformate is treated with aqueous potash, thesubstance is promptly hydrolysed, nitrogen being eliminated exclusivelyin the form of hydrazoic acid :N,*CO,*C,H, + H20 = HN, + CO, + C,H,*OH ;here there is no hydrogen attached to the carbon atom with which thetriazo-group is combined, and consequently the tendency towards ring-formation as indicated above cannot find expression, whereas triazo-acetic acid so far resembles triazocamphor as to yield nitrogen whenboiled with excess of potash, unaccompanied by hydrazoic acid,N,*CH,*CO, K + N, + NH :CH*CO,H,although the negative hydroxyl group certainly exerts a powerfulretarding influence on the change.Whilst, however, this hypothesisof potential ring-formation appears to us a reasonable one, difficultyarises in connexion with triazoacetoxime and the oxime of camphoryl-azoimide. It has beon shown that the latter substance yields hydrazoicacid instead of nitrogen with alcoholic potash (Trans., 1907, 91, 874),but we now find that triazoacetoxime furnishes both hydrazoic acid andnitrogen with the utmost readiness, and a comparison of the formulae :does not suggest an explanation af this apparent discrepancy.Tri-azoacetoxime is a very labile substance, however, and cannot bedistilled, even under 2 mm. pressure, without undergoing decomposi-tion; moreover, we believe it has a tendency to undergo transforma-tion into the nitroso-modification, and consequently it may be supposedthat the liberation of nitrogen is due to intermediate association of thetriazo- and nitroso-groups76 FORSTER AND FIERZ: THE TRIAZO-GROUP. PART I.opportunity for which does not present itself in the case of thecamphoryl derivative.The properties of triazoacetic ester naturally invite comparisonwith those of the corresponding diazo-compound, prepared by Curtius,and the contrast between the two substances brings out once more thestability of the triazo-group as compared with the diazo-complex.Boiling water, iodine, mercuric oxide, and ammmiacal silver oxideleave the ester unchanged, and sodium does not dissolve in the coldsubstance; it is, moreover, colourless, and the odour is very faint.Furthermore, triazoacetic ester may be hydrolysed to the acid, whichis a well-defined substance, ,and does not decompose below iOOo,whereas attempts to liberate diazoacetic acid by passing carbon dioxidethrough the hydrolysed ester result in liberation of nitrogen, whilstconcentrated aqueous alkalis induce simultaneous hydrolysis and poly-merisation, leading to bisdiazoacetic acid,We are engaged in studying the triazo-derivatives of other typicalmembers of the aliphatic series, including alcohols, aldehydes, ketones,and esters ; such compounds have been prepared from methyl ethylketone, malonic ester, and acetoacetic ester, and we hope to describethese in a subsequent communication.E x P E R I hi E N T A L.T&xoacetic Acid, N,* CH,*CO,H.Fifty grams of triazoacetic ester were shaken with a, 20 per cent.solution of potassium hydroxide containing 21 grams, this being aslight deficit from one molecular proportion ; the temperature roseand the oil disappeared slowly, but there was no liberation of gas.After being twice extracted with ether, the neutral solution wastreated with the calculated amount of sulphuric acid and extractedfifteen times with ether, which, when dried with ignited sodiumsulphate, left a very pale yellow, oily liqlrid on evaporation.This washeated in boiling water during one hour under 2 mm. pressure, whenthe acid was found to be sufficiently anhydrous to solidify in meltingice, but there was no distillation at this temperature, and it was notconsidered safe to heat the substance more strongly.Triazoacetic acid crystallises in hygroscopic, glassy plates, andmelts a t about 1 6 O . I t has a very faint odour, suggesting that ofbutyric acid without the pungent effect; the substance is a strongacid, and feels greasy when rubbed between the fingers. On a hotplate, the acid detonates with a moderate explosion and takes fire,but when it is heated in a capillary tube a violent detonation occurs.It does not reduce ammoniacal silver oxide, even on boiling thesolutionFORSTER AND FIERZ: THE TRTAZO-GROUP.PART I. 77The analysis of trinzoacetic acid and ester has presented unusualobstacles, and more than twenty attempts have been made t o obtainsatisfactory results. The principal difficulty lies in the fact thatwhen combustion is conducted under ordinary conditions, methaneappears among the products, whilst adopting the device which hasbeen suggested to meet this drawback, namely, substitution of leadchromate for copper oxide and mixing the substance with cuprouschloride (Dunstan and Carr, Proc., 1896, 12, 48, and Haas, Trans.,1906, 89, 570), low results were obtained consequent on the pro-duction of methylamine. In the case of the acid, combustion for per-centage of carbon and hydrogen was finally carried out in a tubecontaining platinised asbestos, an attempt being made 60 maintainthroughout the operation a large excess of oxygen, which was usedinstead of air ; for the purpose of estimating nitrogen, advantage wastaken of the fact that when potassium triazoacetate is heated withexcess of potash the triazo-group undergoes disruption, and accordinglythe weighed substance was mixed with 20 per cent.aqueous potashbefore being placed in the combustion tube :0.3306 gave 0.2913 CO, and 0.0892 H,O.0.0947 ,, 34.25 C.C. of nitrogen at 18.5' and 747 mm. N = 41.64.C,H,O,N, requires C = 23.76 ; H = 2-97 ; N = 41.58 per cent.I n view of the violence with which aromatic azoimides lose nitrogenwhen treated with concentrated sulphuric acid, tho behaviour of theagent towards triazoacetic acid is remarkable; when mixed with con-centrated sulphuric acid on a watch-glass, no change takes placeimmediately, and only on vigorous stirring with a glass rod does gasappear, very slowly at first, but quickly increasing in briskness.Estimation of molecular weight was conducted in benzene and inphenol.I n the former solvent, 221 units represents the average ofthree experiments, whilst in phenol the mean of four amounted to 99units, the solution being brown ; the formula C,K,O,N, requires101 units.Salts of Triazoacetic Acid.-The silver salt, C,H,O,N,Ag, obtainedas a curdy precipitate on adding silver nitrate to a neutral solution ofpotassium triazoacetate, may be crystallised from boiling water inpresence of a few drops of dilute nitric acid, separating in colourless,lustrous needles. An attemptl t o estimate the silver by cautiousevaporation with nitric acid having led to a slight detonation, aweighed quantity of the substance was reduced with ammoniumsulphide, the silver sulphide thereby precipitated being convertedinto silver.For the purpose of estimating nitrogen, the saltwas mixed with 20 per cent. potash before being placed i n thetube :C = 24-03 ; H = 2.9978 FORSTER AND FlERZ: THE TKlAZO-GROUP. PART 1.0.2661 gave 0.1385 Ag. Ag= 52.05.0.1923 ,, 34.4 C.C. of nitrogen a t 19O and '745 mm. N = 20.50.C,H,O,N,Ag requires Ag = 51.92 ; N = 20.20 per cent.When heated on an iron plate, the substance detonates mildly andburns brightly.An attempt to prepare the copper salt led to a curious result.Having noticed that copper sulphate develops a deep green colorationwhen mixed with potassium triazoacetate, copper oxide was dissolvedin an aqueous solution of the free acid.It was noticed, however, thatthis solution steadily liberates gas when heated on the water-bath,and, although the liquid may be concentrated at 40°, an att.empt toobtain crystals of the copper salt by leaving the liquid in a desiccatorfailed, because at a high concentration gas was evolved, even at theordinary temperature. A dark green, hygroscopic powder finallyremained, and, on warming an aqueous solution of this product withdilute potash, it remain6d momentarily clear, but suddenly precipi-tated cuprous oxide and liberated ammonia,The potassium salt is freely soluble in water, and is precipitatedby concentrated potash.A neutral, moderately dilute solution issurprisingly stable, and may be boiled without evolving gas, but onadding to the hot liquid some 40 per cent. potash, torrents of gas areliberated, followed, after a momentary pause, by another rush of gas,consisting of nitrogen and ammonia, and the effervescence is continuedin this characteristic fashion by further addition of alkali. We havemade several attempts to isolate the products of this change, whichshould include glyoxylic acid, but hitherto we have been able to recog-nise onlyoxalic acid, which probably arises from glyoxylic acid by theaction of potash.When the potassium salt has been treated withexcess of alkali in the manner indicated above, the liquid depositscrystals on cooling, but the composition of this product appears tovary considerably according to the conditions of the experiment. Onone occasion, a small quantity of a substance was obtained which leftno residue on evaporation with concentrated sulphuric acid, andappeared to be an ammonium salt; it reduced ammoniacal silveroxide and Fehling's solutions immediately without warming, but thisproperty disappeared on boiling with potash, and the reaction is beingstudied therefore more fully.A neutral solution of patassium triazoacetate gives a lustrous,crystalline precipitate with lead nitrate, and a deep red colorationwith ferric chlorideFORSTER AND FIERZ: THE TRIAZO-GROUP.PART I. 79Ethyl Z'riccxoacetccte, N,*CH,*CO,*C,H,.Two hundred grams of ethyl chloroacetate and 100 grams of alcoholwere heated under reflux during three hours with 120 grams ofsodium azide, and sufficient water to maintain the salt in solution; acurrent of steam was then passed through the liquid, from which theester was quickly removed. On diluting the distillahe with water,adding a considerable quantity of crystallised sodium acetate, andallowing the heavy oil to separate, 190 grams of the substance wereobtained, and this mas shaken twice with water, dried with calciumchloride, and distilled under 2 mm. pressure, when it boiled to the lastfew drops a t 44-46'.The difficulties presented by the analysis of this ester have beenmentioned above, and aftor eight attempts, the indicated results ofwhich vary between 19.8 and 34.3 per cent.of nitrogen, we are stillunable t o record a satisfactory estimation of this element; by theuse of platinised asbestos, however, fairly concordant determinationsof carbon and hydrogen have been obtained, although these have beenusually too high, owing to the di6culty of avoiding the formation ofnitrous fumes, even in presence of silver gauze :0.1571 gave 0*2100 CO, and 0.0790 H,O.We are indebted to Dr. Joshua for an independent estimation of0.1202 gave 0 1621) CO,.It happens, unfortunately, that the gasometric estimation of nitrogeneliminated by concentrated sulphuric acid, hot potash, or stannouschloride cannot be used to supplement the above analytical dataregarding the ester, because the action in each case takes an abnormalcourse, and the percentage of gas evolved by these agents agrees moreclosely with half the azidic nitrogen (16.3 per cent.) than with two-thirds (21.7 per cent.). So regular is this discrepancy from theexpected result that we became suspicious, before the analyticaldifficulties mere surmounted, regarding the identity of the ester, as itseemed possible that the triazo-group had conferred on acetic ester thecapacity t o form an alcoholate, because i t happens by chance thattwo-thirds of the nitrogen required by the formula,N,*CH,*C(OEt),*OHamounts to 16.0 per cent., or, roughly, half the nitrogen content ofthe simple ester.Accordingly, a specimen of triazoacetic ester waaprepared by heating 50 grams of ethyl chloroacetate under reflux withsodium azide dissolved in water, but the product was found to corre-C = 36.30 ; H = 5.62.C,H70,N, requires C = 37.21 ; H = 5.43 per cent.carbon by oxidation with a mixture of sulphuric and chromic acids :C=36*76 per cent80 FORSTER AND FIERZ: THE TRIAZO-GROUP. PART I.spond in every respect with the ester prepared in presence of alcohol ;the action proceeds in the same way, but is much more sluggish thanwhen alcohol is used. Tbe abnormal behaviour indicated above stillawaits explanation, therefore, and we hope to furnish this later.Triazoacetic ester is a limpid, colourless oil, having a faint, sweetodour, more suggestive of chloroform than of ethyl acetate, theresemblance to the latter becoming more marked in steam, with whichthe substance is readily volatile; when inhaled for some seconds, thevapour produces a throbbing sensation in the head, and slight palpita-tion of the heart.On mixing with concentrated sulphuric acid, thereis no effervescence at first, but, on stirring vigorously, gas is liberatedslowly, the disengagement becoming ultimately quite brisk ; withstannous chloride dissolved in concentrated hydrochloric acid, nitrogenis evolved immediately. When the ester is shaken and warmed with10 per cent. caustic potash, i t is quickly hydrolysed, forming a clearsolution of potassium triazoacetate, but concentrated alkali (40 percent.) appears to leave the substance unchanged, unless alcohol is addedor the temperature raised. Freshly-cnt sodium does not dissolve in itunless the liquid is heated, when vigorous action takes place; thesubstance is indifferent towards mercuric oxide and a solution ofiodine in potassium iodide.The specific gravity of triazoacetic ester is 1.127 compared withwater at 20'.An estimation of molecular weight in benzene gave119, 125, and 127 units, the formula C4H70,N, requiring 129.Yriaxoacetccnzide, N,*CH,*CO*NH,.When shaken with aqueous ammonia, the ester dissolved, and, onevaporating on the water-bath, a pale red liquid remained whichsolidified on cooling ; this was drained on earthenware, and recrystal-lised twice from hot benzene :0-141 I gave 68.4 C.C.of nitrogen a t 20' and 758 mm.C2H40N, requires N = 56.00 per cent.Triazoacetamide forms tough, lustrous, colourless needles, frequentlyexceeding an inch in length, and melting at 58'; it is readily solublein water and in alcohol, but is most conveniently crystallised from hotbenzene, in which it is moderately soluble, whilst boiling petroleumalso dissolves it, but less freely. When thrown on a hot plate, i tdetonates feebly and takes fire.The substance is unusually resistant towards concentrated sulphuricacid, with which it must be warmed to about 50' before gas is liberated;cold 60 per cent. aqueous potash, however, attacks it immediately,torrents of nitrogen and ammonia being liberated, whilst more dilutesolutions of alkali disrupt the triazo-group very slowly.An aqueousN = 56.30FORSTER AND FIERZ: THE TRIAZO-GROUP. PART I. 81solution of the amide dissolves yellow mercuric oxide on boiling, butthe resulting compound, unlike mercury acetamide, does not lose themetal when treated with hydroxylamine hydrochloride, a precipitatebeing formed only on adding alkali, when effervescence takes place(compare Trans., 1898, 73, 785).Ethyl Triaxofoymccte, N,*CO,*C,H,.Methyl azoimidocarbonate was prepared by Curtius and Heidenreich( J . pr. Cheni., 1895, [ii], 52, 454) from ammonium azoimide andmethyl chlorocarbonate, and, in order to compare the condition of thetriazo-group in this type of compound with the behaviour of the samecomplex occurring in the acetic series, we have examined the corre-sponding ethyl ester, which was prepared by agitating 50 grams of ethylchlorocarbonate with an aqueous solution of sodium azide containing35 grams until the pungent odoiir of the original material was nolonger perceptible ; the heavy oil was then removed, dried with sodiumsulphate, and distilled under 2 mm.pressure, when i t boiled steadilyat 2 5 O .The colourless, limpid ester has sp. gr. 1.118 compared with water atIS0, and boils a t 114' under 769 mm. pressure, but is liable to explode ;the odour is more powerful than that of triazoacetic ester, and thedisagreeable effects of inhaling the vapour are much more marked.I t may be mixed with concentrated sulphuric acid or 40 per cent. potashwithout evolving gas, but the alkali hydrolyses it completely to hydrazoicacid, alcohol, and carbonic acid; it is therefore impossible to producetriazoformic acid, or even the salts, because a deficit of alkali merelyleaves the corresponding amount of ester unchanged.Alcoholicammonium sulphide reduces ethyl triazoformate to urethane, nitrogenbeing set free.Triaxoacetone (Acetonylazoimide), N,*CH,*CO*CH,.The monochloroacetone required for the production of triazoacetonewas prepared by Fritsch's method (Annalen, 1894, 279, 313), whichwe have found to yield excellent results.One hundred grams of chloroacetone (b. p. 119-12lo) mere shakenwith a concentrated aqueous solution of sodium azide containing 80grams, to which a few drops of glacial acetic acid had been added;after twenty-four hours, the pungent odour of chloroacetone being nolonger perceptible, the oil was extracted with ether, dried with sodiumsulphate, and distilled under 2 mm.pressure. The product, weighing90 grams, was shaken with freshly-ignited sodium sulphate, and againdistilled under 2 mm. pressure, the major portion boiling at 54*, andhaving sp. gr. 1.123 compared with water a t 18' :VOL. XCIII. 82 FORSTER AND FIERZ: THE TRIAZO-C4ROUP. PART I.0-2398 gave 0.3204 CO, and 0.1 132 H,O.0.1685C,H,ON, requires C=36.36 ; H=5*05 ; N=42*42 per cent.Freshly-distilled triazoacetone is a colourl6ss, highly refractiveliquid having a faint odour, but after a few days, even when preservedin a well-stoppered bottle and protected from light, the odour ofcarbylamine is noticeable in the specimen, which has become yellow.The substance exhibits no tendency to solidify, remaining quite limpidin a freezing mixture j it is sparingly soluble in water, and very readilyvolatile in steam.When dropped on a hot plate, triazoacetoneexplodes, and burns with a brilliant flame ; concentrated sulphuricacid decomposes the substance immediately, liberating nitrogen.Action of Alkali. -When triazoacetone is treated with concentratedaqueous potash (40 per cent.), nitrogen is liberated with almostexplosive violence, and the liquid becomes red; eyen with a 1 percent. solution of the alkali, gas evolution is quite brisk, and severalconcordant estimations of the nitrogen evolved during the changeindicated that this amounted to roughly 4 per cent.less than two-thirds :C = 36.44 ; H = 5.24.N=42*45. ,, 61.1 C.C. of nitrogen at 1 9 O and 761 mm.091550 gave 33.4 C.C. of nitrogen at 22' and 761 mm.C,H,ON, requires 2/3N = 28-3 per cent.By decomposing 5 grams of t,he triazo-ketone at one time, steamingthe product while alkaline, then adding dilute sulphuric acid and dis-tilling again, we were able to show that the deficit indicated above isdue to simultaneous production of hydrazoic acid, which mas easilyrecognised in the acid distillate. Excepting ammonia, however, theother products of the changes involved are not easily identified. Thefact that nitrogen, ammonia, and hydrazoic acid are eliminated, indi-cates that the following decomposition occurs :N = 24-3.I.N,*CH,*CO*CH, = N, + NH:CH*CO*CH,.11. NH:CH*CO°CH, + H20 = NH, + O:CH*CO*CH,.111. N,*CH,*CO*CH, + H20 = HN, + HO*CK,*CO*CH,.Accordingly, it should be possible to recogniso both pyroracemicaldehyde and acetol in the product, and, although we have failed toisolate these compounds, probably owing to the further action ofalkali, the presence sf reducing materials is indicated by vigorousaction on Fehling's solution and ammoniscal silver oxide.The 8emicarbazone.-On mixing 6 grams of triazoacetoue with6.5 grams of semicarbazide hydrochloride and 6 grams of sodiumacetate in water, the semicarbazone was precipitated immediately.After being recrystallised twice from absolute alcohol, it melted a t152' without evolving gas FORSTER AND FIERZ: THE TRIAZO-GROUP.PART I. 830.1258 gave 58.7 C.C. of nitrogen a t 22' and 761 mm.C,H,ON, requires N = 53.84 per cent.The semicarbazone crystallises from water or alcohol in long,brilliant needles ; it effervesces with aqueous potash and, moreslowly, with concentrated sulphuric acid.This derivative has been prepared from several specimens of triazo-acetone, as it forms a convenient substance by which to identify theketobe, and on one occasion the latter remained in association withexcess of semicarbazide during several hours. The product in thiscase was quite distinct from the semicarbazone described above, beingvery sparingly soluble in common media, and precipitated by alkalifrom solution in acids.A specimen recrystallised from glacial aceticacid was therefore analysed, but the combustion presented considerabledifficulty, as the substance has no definite melting point, and decom-poses a t a high temperature, leaving carbon :0.1476 gave 0.1733 CO, and 0.0763 H,O. C = 32.02 ; H = 5-78.0.1300 ,, 50.3 C.C. of nitrogen a t 17' and 758 mm. N = 45.43.C,H,,O,N, requires C= 32.25 ; H = 5.38 ; N = 45.16 per cent.We believe therefore that this compound is the his-semicarbazoneN = 54.06.of methylglyoxal,CH,*C( :N*NH*CO*NH,)*CH:N*NH*CO*NH,,because me have found that, under certain conditions, hydroxylamine iscapable of transforming acetonylazoimide into methylglyoxime,CH,*C(:NOH)*CH:NOH ;the latter change indicates a disposition to undergo oxidation on thepart of the terminal carbon atom, recalling the behaviour of fructosetowards phenylhydrazine, but, so far as we know, the alteration inquestion has not previously been effected by semicarbazide.Triaxoncetozime, Ns*CH,*C(:NOH)*CH,.Five grams of triazoacetone were warmed to 50' with a solution ofhydroxylamine containing 4 grams of the hydrochloride in 80 C.C.ofwater, neutralised with 3.2 grams of anhydrous sodium carbonate ; theoil dissolved, and the solution suddenly became turbid. After twohours' agitation, the product was extracted with redistilled, purifiedether, and dried with sodium sulphate, the solvent being removed byexpoeing the liquid to a pressure of 2 mm. during four hours :0.0786 gave 33% C.C.of nitrogen at 18' and 745 mm.This experiment was made subsequently to an attempt to distil50 grams of triazoacetoxlme under 2 mm. pressure. On this occasion,about 25 grams boiled a t 84O, whilst the residue in the flask graduallyN=49.01,C,H60N, requires N = 49.1 2 per cent.a 84 FORSTER AND FIERZ: THE TRIAZO-GROUP. PART I.became dark brown, finally exploding with coiisiderable violence ; thedistillate was colourless, and had a faint odour of prussic acid, whichalone indicates decomposition, since the undistilled oxime is odourless,and, moreover, when analysed, furnished an amount of nitrogen 2 percent. below that required by the empirical ormula C,H,ON,. It isnoteworthy that bromoacetoxime also undergoes explosive decom-position when distilled.Triazoacetoxime is colourless, and does not solidify in the freezingmixture ; its behaviour towards concentrated sulphuric acid resemblesthat of triazoacetic acid and ester, liberation of nitrogen occurringonly after some delay, and on vigorous agitation.When treated with40 per cent. potash, nitrogen is liberated in considerable quantities, buthydrazoic acid is also produccd; if, however, more dilute alkali(20 per cent.) is employed without heating, all the nitrogen is eliminatedin the form of hydrazoic acid, along with a substance having theproperties of the oxime of acetol :N,*CH,*C(NOH)*CH, + H,O = HN, + HO*CH,*C(:NOH)*CH,,Before the unstable character of triazoacctoxime was appreciated, anattempt was made to prepare this compound by the action ofhydroxylamine sulphate instead of the free base.Twenty grams oftriazoacetone were suspended in 300 C.C. of water, and agitated with32 grams of hydroxylamine sulphate in 150 C.C. of water; the oildisappeared gradually but, the odour of hydrazoic acid becomingperceptible, the product was allowed to remain ten days in order tocomplete the change. The liquid was then extracted three times withether, and on removing the solvent, after drying with sodium sulphate,9 grams of a lustrous, crystalline solid separated from the oily residue ;on recrystallising this product from hot benzene, of which 200 C.C.were required by 1 gram, it was found t o be methylglyoxirne, andwas obtained in minute needles, melting at 157" without evolvinggas :0.1189 gave 35.4 C.C.of nitrogen at 18' and 756 mm.C,H,O,N, requires N = 27.45 per cent.The production of methylglyoxime and the bis-semicctrbazone ofmethylglyoxal by the action of hydroxylamine and semi-carbazide respectively on triazoacetone, is an interesting case of thattype of oxidation which leads to the formation of osazones. A similarobservation has been made in connexion with chloroacetoxime byHantzsch and Wild (Annalen, 1896, 289, 285) and Scholl andMatthaiopouIos (Ber., 1896,29, 1550), methylglyoxime being obtainedwhen excess of hydroxylamine acts on the substituted ketoxime. It isprobable, also, that the same change occurs when hydroxylamine actson hydroxyacetone, because Kling (Ann. Chirn. Phys., 1905, [viii], 5,N = 27.74VELOClTY OF FORMATION OF ACETOXIME. 85482), who performed the experiment, recorda the production of acompound melting at 153O, and an analysis which furnished 26.6 percent. of nitrogen ; Kling suggests the constitutional formula,CH,*C( :NOH)*CH,*N€€*OH,but, as there is no evidence for this view, it seems to us more probablethat he had in hand an imperfectly purified specimen of methyl-glyoxime.Z’he p-Yohenesulphonic Berivutive.-The oxime being a n oil, it witsconsidered desirable to characterise the substance further by preparinga crystalline derivative. Triazoacetoxime dissolved in pyridine wasaccordingly treated with p-toluenesulphonic chloride ; the oil whichseparated on dilution with water quickly solidified, and was re-crystallised from boiling light petroleum, in which the substancedissolves sparingly, 1 gram requiring about 100 c.c., from whichit crystallises in large, lustrous, striated plates, melting a t 7 3 O :0.1090 gave 20.2 C.C. of nitrogen at 23’ and 757 mm. N = 21.29.C,,H,,O,N,S requires N = 20.90 per cent.The substance, although snow-white when freshly crystallised,rapidly deteriorates, becoming brownin a few days. Thrown on ahot plate, i t detonates, and burns with a, brilliant flame ; it ismoderately soluble in cold methyl or ethyl alcohol, disso1viDg freely inchloroform, benzene, or ethyl acetate.ROYAL COLLEQE OF SCIENCE, LONDON.SOUTH KENSINGTON, S. W