CSRROXYLIC AClDS DERIVED FlZOSI CTCLOBUTANE, ETC. 26635CCX LIX .-Curboxylic Acids Derived j+owb cycloButune,cyclopentane, cycloHexune, and cyclo [Reptune.By LEONARD JAMES GOLDSWORTIIY andWILLIAM HENRY PERKIN, jun.THE present investigation is one of a series which has been insti-tuted with the object of obtaining further evidence relating to thecomparative readiness of formation and stability of cyclic struc-tures containing varying numbers of carbon atoms. Judging by theyields produced in analogous reactions, experience is roughly inaccordance with Baeyer’s ‘‘ Spannungstheorie,” and seems t o indi-cate that, in the cyclopropase, cyclobutane, cyclopentane, andcyclohexanel series, derivatives of cyclopropane are produced withthe greatest difficulty, and t,hat, whilst derivatives of cyclobutaneand cyclohexane are much niore readily obtained, the tendency t oform cyclopentane derivatives is so pronounced that these areof ten produced in quantitative yields, and not infrequently duringreactions which might be expected t o lead t o the formation ofother ring complexes.The evidence on this point, however, is oftenconflicting, since it has frequently been observed that, althoughsome cyclic derivatives are obtained in very small yields, otherclerivatives of tho same ring seem to be produced under very similarconditions with great readiness. Thus the yield of ethyl (ydo-propane-l : 1-dicarboxylate (I) obtained when ethylene dihromitlereacts with t’lie sodium derivative of ethyl inaloiiate is very small,whereas ethyl cyclopropane-1 : 2-dicarboxylate (11) is readily pre2666 GOLDSWORTHY AND PEHKIN :pared in good yield when ethylene dibromide is replaced by ethyla@-dibrornopropionate in this interaction :Other similar cases have been observed, and one of the chiefdifficulties which arises is t o distinguish between the effect on theyield of tlie reactivity of the interacting substances on the onehand and of the readiness of formation of the closed ring on theother.It is clear that a useful generalisation cannot be formulateduntil a much larger number of cyclic carboxylic acids and otherderivatives have been prepared and investigated, and, in tlie presentcommunication, we describe some new carboxylic acids derivedfrom cycZo-butane, -pentane, -hexane, and -heptane.I n the firstplace, we have prepared the cis- and trans-modifications of cyclo-b utan e-l : 2 : 3-tricarboxylic acid,by causing ethyl ap-dibromopropionate t o react with the disodiuinderivative of ethyl ethanetetracarbosylate, when the reaction pro-ceeds t o the extenli. of about 50 per cent. in the required direction :The prod-uct, after hydrolysis and elimination of carbon dioxideby heating at 190°, yields cis-cyclobutane-1 : 2 : 3-t~iccrrbosy& acid(m. p. 141-143*), and this, when heated with hydrochloric acid at180°, is converted into the trans-acid, which melts at 168-170°.In the cyclopentane series the cis- and traiis-modifications of the1 : 2 : 4-tricarboxylic acid have $ready been prepared from thedisodiurn derivative of ethyl pentane-aay yeehexacarboxylate by theaction of iodine and subsequent liydrolysis and elimination ofcarbon dioxide :(Bottomley and Perkin, T., 1900, 77, 296).We have now suc-ceeded in obtaining t'he same acids much more conveniently and ina much better yield by the action of ethyl ap-dibromopropionate onthe disodium derivative of ethyl propane-aayy-tetracarboxylate(ethyl met,hylenedimalon ate) :CH,*$J(CO,Et),CH,Br*CHBr*CO,Et CH2.CH*C0,EtCARBOXYLIC ACIDS DEIi LVED FROM CYCCOBUTANE, ETC. 2667When the product of this interaction is hydrolysed, simultaneouselimination of one molecule of carbon dioxide takes place and acrystalline cyclopentanetetracarboxylic acid is formed. The crudeacid decomposes a t 180°, and yields a syrupy mass, from whichtrans-cyclopentane-1 : 2 : 4-tnricarboxylic acid is obtained by theaction of hydrochloric acid at 190°.It melts a t 127-130°, and isconverted, by heating with acetic anhydride and subsequent distil-lation, into the anhydro-cis-acid :a highly characteristic derivative, which melts at 215-21 To, andfurnishes the cis-acid (m. p. 146-148") on hydrolysis.We next, attempted the synthesis of cycloheznn e-1 : 2 : 4-tricarb.oxylic acid (hexahydrntrimellitic acid),an acid which does not appear t o have been previously described,and we ultimately succeeded in the following manner. The disodiumderivative of ethyl butanetetracarbaxylate was caused t o react, withethyl aP-dibromopropionate, when deconiposition takes place in thefollowing manner :The ester thus produced yields, on hydrolysis and elimination ofcarbon dioxide, a mixture of stereoisomeric acids, from which, byheating with hydrochloric ;rcid a t 190°, trans-cycloheznne-1 : 2 : 4-tricarboxylic acid was isolated, melting at 220-222O.When thisacid was digested with acetic anhydride and the product distilled,most of i t decomposed, but a sinall quantity of a distillate wasobtained, which, on hydrolysis, yielded the cis-acid as a crystallinecrust melting a t 225O. It is remarkable that the cis- and trans-modifications of this acid should have almost identical meltingpoints, and be so very similar in other properties that i t was a tfirst thought t h a t they were identical.However, a mixture of equal parts of the two preparations wasfound t o soften a t 198--30Clo, and to he alniost completely nielteda t 208O, so that they cannot be identical, and there can be littledoubt that they are the cis- and trcrm-modifications of cyclohexane-1 : 3 : 4-tricarboxylic acid2668 GOLDSWORTHY AND PERKIN :Finally, we have succeeded in synthesising trans-cyclokcpta~ze-1 2 : 4-t~i~c(rbox?/lic acidone of the few derivatives of cycZohept,ane which have, so far, bee11obtained.For the purpose of this synthesis, the disodium deriv-ztive of ethyl pentane-aaee-tetracarboxylate was digested with ethylup-dibroiiiopropionate when a coinplicated reaction took place, butan ester, evidently ethyl cyclopentane-1 : 1 : 2 : 4 : 4-pentacarboxylate,was produced in small quantity according t o the equation:CRTa(CO2Et),*CH2*CH2*C€J2*CNa(CC),Er), + CH,BI *CIIHBi *CO.,EtWiien the product of this interaction was hydrolysed, the syrupyacid mixture heated a t 200° and then esterified, i t yielded a smallquantity of ethyl cycloheptane-1 : 2 : 4-tricarboxylate (b.p.212-215O/30 mm.).On hydrolysis a syrupy mixture of stereoisomeric modificationswas obtained, from which, by heating with hydrochloric acid a t190°, trans-cycloheptane-1 : 2 : 4-tricnrboxylic acid was isolated, melt-ing a t 198--200°.Apparently there is little tendency in the direction of the forma-tion of the cycloheptane ring, since the yield of the above acid wasso small t h a t we were unable to examine it a t all completely.Further experiments, which are in progress, will show whether thisring and also the cyclo-octane and still larger rings are alwaysproduced with more difficulty than the simpler rings containingthree, four, five, and six carbon atoms.EXPERIMENTAL.Preparutio 1% of E t T~yl ap-Dib ro mopropionut e , CH,Br *CHBr *CO,Et.Dwing t h e course of this and other investigations, large quanti-ties of ethyl up-dibromopropionate were required, and, as the pricecharged for this ester is prohibitive, we have made a series ofcomparative experiments on the best conditions for its preparation,and find that the following process works well.Ally1 alcohol *(180 grains or 210 c.c.) is mixed with a n equal volume of chloro-forni (or carbon disulphide), the solutioii cooled in ice and salt,Crude coniuiercinl ally1 nlcoliol contains a large rtnioutit of water, and, if this isiiseil as tlio shrting-point, i t is shaken with l)otas5inni carbonate as long as thisdirsol~red, tlic aqueous layer r u n off and the nlcohol clehytfrnte 1 with a furtherquantity of potassinm carbonatc.It is tlicii fractionated nit11 ail ( ficient colunii~ ;the portion dirtilling a t 95-98' was collected for iise in the above preparationCAltUOXYLIC iiClDS DERIVED FROM CYULOBUTANE, ETC. 2669and tl.en bromine (496 grams or 156 c.c.) is gradually added, carebeing taken that the temperature remains below 5 O . The chloro-form is then removed by distillation from the water-bath underdiminished pressure, and the residual crude dibromopropyl alcoholneed not be fractionated, but can be directly oxidised t o up-dibromo-propionic acid.The crude dibromo-alcohol (150 grams) is mixedwith ordinary concentrated nitric acid (210 grams or 150 c.c.) andfuming nitric acid (90 grans or 60 C.C. of D 1.5) in a capaciousflask provided with a ground-in condenser and placed in water, andthe water is then very gradually heated. As soon as the initialviolent reaction has subsided, the water is raised t o the boilingpoint and maintained a t that for seven hours; the product is thenleft overnight in the ice-chest t o cool, when the dibroinopropionicacid usually crystallises, but, if not, the heating is continued andthe mass seeded. The acid is collected on a Bucliner funnel withoutfilter paper, drained on porous porcelain, and a further quantityniay be obtained by concentrating the mother liquors, so that thetotal yield is about 75 per cent.of that theoretically possible. I norder t o obtain the ethyl ester, the acid is dissolved in one anda-half tiixes its own weight of a 10 per cent. sorution of sulphuricacid in alcohol and heated on the water-bath for six hours. Wateris then added, the heavy ester extracted with ether, the etherealsolution washed first with dilute sodium carbonate, then withwater, dried, and the ether distilled off. The residual ethyl ap-di-bromopropionate distils almost completely a t 140-150°/ 100 mm.,and this was the material used in the following experiments.The cis- and trans-cycloRutnne-1 : 2 : 3-tricarboxylic Acids,CH(COP)>CH. CO,H.CH~<C€€(CO,H)I n order t o obtain these acids, the first, step was the synthesis ofethyl cyclobutane-1 : 1 : 2 : 2 : 3-pentacarboxylate (p.2666), and thiswas accomplished in the following way : Ethyl ethanetetracarboxylate (31 grams) was mixed with a little alcohol,* and then witha solution of sodium (4.6 grams) in alcohol (100 c.c.). and, afterkeeping for a few minutes until the whole of the ester had passedinto solution, ethyl up-dibrornopropionate (27 grams) was graduallyadded. I n a short time the mixture became warm, copious pre-cipitation of sodium broniide took place, and the process was coni-1)lstecl by lieating in a soda-water bottle in boiling water for fourhonrs. The product was mixed with water, extracted with ether,t h e ethereal solution washed well, dried, and the ether distilled off.* The alroliol u-etl i i i all these espcrinieiits was cdrefully dehydrated by distillationfirst over lirne a i d then over calciiiiil2670 GOLDSWORTHY AND PERKIN :The crude ester which appears to distil a t about 230°/20 mm.neednot be purified before conversion into cyclobutanetricarboxylic acid,but i3 a t once hydrolysed by boiling with 50 per cent. excess of25 per cent. methyl-alcoholic potassium hydroxide for four hours.Water is then added, the product evaporated until quite freefrom methyl alcohol, the solution is then acidified with excess ofconcentrated hydrochloric acid, evaporated to dryness, and theresidm extracted with ether in a Soxhlet apparatus.After distil-ling off tho ether, crude cyclobutanepentacarboxylic acid remains asa slightly brown, viscid syrup, and this is heated in an oil-bath a t190° for one hour. when the evolution of carbon dioxide will haveceased. The acid which remains does not readily crystallise, and istherefore purified by conversion into the ester which is obtained byboiling the acid with ten times its weight of 10 per cent. alcoholicsulphuric acid in a reflux apparatus for seven hours. The productis diluted with water, extracted with ether, the ethereal solutionwashed well with sodium carbonate, dried, and evaporated, and theresidue fractionated under 40 mm. pressure. A small quantity ofoil passes over below 1900, and probably contains ethyl succinate;then almost the whole of the remainder distils a t 190-205O, and onrefractionatioa, pure ethyl ciscyclobutane-1 : 2 : 3-tricarboxylate isobtained as a colourless oil boiling a t 195--197O/40 mm.:0'2246 gavo 0'4714 CO, and 0.1540 H20. C=57.2; H=7.6.C,,H,,,O, requires C = 57.3 ; H = 7.3 per cent.I n the preparation of the tricarboxylic acid, the ester distillinga t 190-205°/40 mm. was hydrolysed by boiling with excess of25 per cent,. methyl-alcoholic potassium hydroxide for four hours.Water was then added, the solution evaporated until quite freefrom methyl alcohol, mixed with excess of hydrochloric acid, eva-poreted, and the dry residue extracted with ether in a Soxhletapparatus. The ethereal solution deposited, after boiling off theether, a Eolicl acid, which was left in contact with porous porcelain,and then rwrystallised from concentrated hydrochloric acid :0,1410 gave 0'2332 CO, and 0.0592 H,O.C=45.1; H=4%.C,H,O, requires C = 44.7 ; H = 4.3 per cent.On titration, 0.1226 required 0-0772 NaOH for neutralisation,whereas this amount of a tribasic acid, C,H,O,, should neutralise0'0782 NaOH.cis-cycloButane-1 : 2 : 3-t~icarboxylic acid melts a t 141-143O, andis readily soluble in water or alcohol, but rather sparingly so incold concentrated hydrochloric acid.I n order to obtain the tmns-modification, the cis-acid (2 grams)was heated with concentrated hydrochloric acid (15 c.c.) in a sealedtube at 180° for two hours. After diluting with water and filterinciiImoxl’r,Ic ACIDS DERIVED FROM CYCLOBUTANE, ETC. 267 1from a trace of carbonaceous matter, the solution was concentrated,when, on standing over solid potassium hydroxide, a hard, glassy,crystalline mass, consisting of flat, glistening plates, graduallyseparated :0.1031 gave 0.1699 CO, and 0.0411 H,O.C=44*9; H=4*4.C,H,O, requires C= 44.7 ; H =4.2 per cent.On titrntion, 0.3500 neutralised 0.2236 NaOH, whereas thisamount of a tribasic acid, C7H,0,, should neutralise 0‘2234 NaOH.trans-cycloButrrne-1 : 2 : 3-tricarboxylic ncid melts a t 168-l7Oo,and is readily soluble in watler, but sparingly so in concentratedhydrochloric acid. The solution of the acid in excess of ammoniagives, when hoiled with barium chloride, a very sparingly solublebarium salt, and the cis-acid behaves in a similar manner. Bothacids are oxidised with difficulty, even when their alkaline solutionsare boiled with permanganate.The Trianilide of the trans-L4ccid.--In order to prepare this deriv-ative, the pure trans-acid was heated with thionyl chloride in asealed tube a t loOD for an hour; the product was then evaporated,and the residual syrup dissolved in benzene.Aniline was addedin excess, and, as soon as the vigorous action had subsided, thebenzene was evaporated, the mass treated with dilute hydrocliloricacid, and the precipitate collected and drained on porous porcelain.The trianilide of trans-cyclobutane-1 : 2 : 3-tricarboxylic acid meltsa t 252O, and separates from alcohol, in which i t is sparingly soluble,as a voluminous, almost gelatinous, mass of needles. For aaalysjsi t was recrystallised by dissolving in acetone and ;1 ddiiig benzene ;tlie acetone was then distilled off, and tlie solution set aside, whenthe substance separated as a crystalline crust of needles :0.2449 gave 21.5 C.C.N, a t 1 7 O and 763 inm.C2jH230XN3 requires N = 10.2 per cent.In order t o make sure that transformation of the trans-acid intothe cis-modification had not taken place during the heating withthionyl chloride, soiiie G f the syrupy product of this action wasdecomposed by boiliq with water and the acid recrystallised, w1iei-Li t melted a t 168-170O.hT==10*2.The cis- and trans-cycloPeutane-1 : 2 : 4-tricnrborylic ,-1 cids,The etliyl cyclopentane-1 : 1 : 2 : 4 : 4-pentacarhoxylate employed inthis synthesis of these acids was obtained nnder the followinqcmiidi tions.Ethyl inethylenedimalonate, (CO,Et),CH*CH,*CH(CO,Et), (33.2672 GOLDSWORTHY AND PERKIN :crams), dissolved in alcohol (50 c.c.) was added t o a cold solution ofsodium (4.6 grams) in alcohol (100 c.c.), and, after a few minutes,the disodium derivative was inixed with ethyl ab-dibromopropionate(26 grams), when gradual rise of temperature, followed by a vigor-ous action, set in, and sodium bromide separated. The actionwas completed by heating for three hours in a soda-water bottle,the product was diluted with water, extracted with ether, and afterwashing well, drying, and evaporating off the ether, the residual oilwas distilled, when almost the whole quantity passed over at226--240°/15 mm., and, on redistillation, the boiling point of theethyl cyclopentanepentacarboxylate was observed t o be about234--236O/15 min.The ester was hydrolysed in the' usual mannerby boiling with excess of 25 per cent. methyl-alcoholic potassiumhydroxide for four hours, the product was diluted with water,evaporated until free from methyl alcohol, acidified, again evapor-ated, and the mass extracted with ether in a Soxhlet apparatus.When the ethereal solution was evaporated, a syrupy massremained, which becams partly solid while still on the water-bath.This was stirred with concentrated hydrochloric acid, and the sandypowder which separated was collected, washed with hydrochloricacid, left in contact with porous porcelain in a vacuum desiccator,and then analysed:0.1670 gave 0.2768 CO, and 0.0644 H,O.C =43*7 ; H=4*3.C,H,,O, requires C = 43.9 ; H = 4.0 per cent.It is obvious that' during the hydrolysis under the above condi-tions one molecule of carbon dioxide had been eliminated, and thatthe above acid is cyclopentaite-l : 2 : 2 : 4-tetracarhoxylic acid,Ethyl cycloPentane-1 : 2 : 4-tricarboxylnte, C5Hi(C02H)3. - Inorder t o prepare this ester t'he crude' product of the hydrolysis ofethyl cyclopentaiiepentacarboxylate was heated a t 190° for half-an-hour, when a glassy mass remained, which did not crystallise. Thiswas converted into the ester by heating with alcohol and sulphuricacid in the usual manner (p. 2670), and the pure substance distilleda t 205-210°/40 mm.:0.1938 gave 0.4186 CO, and 0.1360 H,O.C,PH2206 requires C = 58.7 ; I$= 7.8 per cent.Tlie corresponding trimethyl ester, C,Hi(CO,Me),, liati beell pre-viously obtained (T., 1900, 77, 303), and distilled at 164-166*/12 mm.Tlie pure triethyl ester was hydrolysecl with niethyl-alcoholic potassium hydroxide in the usual manner, and, afterremoval of the metthyl alcohol, acidifying and evaporating t oC=58*9; H=7*8CAHBOSYLIC ACIDS DERlVED FROM CYCLOBUTANE, E'I'C. 26'73dryness, the mays was extracted with ether in a Soxlilet apparatus.Ou evaporation, the ethereal solution deposited a syrup, which,when rubbed with coiicentrated hydrochloric acid, crystallised, an(1the solid mass, after washing with hydrochloric acid and dryiiig 011porous porcelain, melted indefinitely a t 115-l25Oy and was appa-rently a mixture of the cis- and trans-modifications of cyclopentane-1 : 2 : 4-tricarboxylic acid.In order to demonst'rate this, half of thesubstance was heated with hydrochloric acid in a sealed tube at180° for four hours. The product was diluted with water, filteredfrom a small arnount of carbonaceous matter, and evaporated todryness; it was then dissolved in water, digested with animalcharcoal, and again evaporated. When the syrupy residue wasrubbed with concentrated hydrochloric acid, it crystallised withdifficulty (compare T., 1900, 77, 304), and the solid mass, aftercontact with porous porcelain and recrystallisation from hydro-chloric acid, melted a t 129-130°, and consisted of tmns-cyclo-pentane-1 : 2 : 4-tricarboxplic acid :0.1264 gave 0.2224 CO, and 0.0568 H,O.C = 417.9 ; H = 5.0.C,H,,O, requires C = 47.5 ; H = 4.9 per cent.On titration 0.1528 neutralised 0.0880 NaOH, whereas thisamount of a tribasic acid, C8Hl,,06, should neutralise 0.0907 NaOH.cis-cycloPentane-1 : 2 : 4-tricarboxylic Acid. - This modificationwas obtained by heating the crude mixture of acids melting a t115-125O (see above) with acetic anhydride for an hour, and thendistilling the product, when anhydro-cis-cyclopentanetricarboxylicacid solidified in the neck of the retort. After recrystallisation froma mixture of acetone and chloroform, this substance melted at215-217O, as stated by Bottomley and Perkin (Zoc.cit., p. 305).The solution of the anhydro-acid in water deposited, on concentra-tion, colourless crystals of cis-cy clopentane-1 : 2 : 4-tricarboxylic acid,melting at 146-148O.Y'he cis- aad trans-cycldexane-1 : 2 : 4-tm'carboxylic Acids(I~exahydrotrimellitic Acids),The ethyl butanetetracarboxylate,CH( C02E t)2*C15&= CH2=CH (C0,E t),,required for these experiments was prepared by the processdescribed by Perkin (T., 1894, 65, 578). This ester (34-6 grams),dissolved in an equal weight of alcohol, was mixed with a solutionof sodium (4.6 grams) in alcohol (100 c.c.) and then with ethylaS-dibromopropionate (26 grams), when, on setting aside, th2674 GOf,DSWOJtTHY AND PERKIN :mixture gradually becaaie quite liot. After heating for four 1iou.r~in a soda-water bottle in boiling water, water was added, the crudeester extracted with ether, and, since i t decomposed on distillation,it was a t once hydrolysed by boiling with excess of 25 per cent.methyl-alcoholic potassium hydroxide in the usual manner.Water was then added, the methyl alcohol carefully removed onth3 water-bath, the product acidified with excess of hydrochloricacid, evaporatled t o dryness, and extracted in a Soxhlet apparatuswith ether.On evaporation, the ethereal extxact deposited a syrup whichpartly solidified, and this was heated a t 190° for half-an-hour, thesyrupy residue being then esterified by boiling with alcoholicsulphuric acid and the ester fractionated (compare p.2670). Ethylcyclohezane-1 : 2 : 4-tricarboxylate was thus obtained as a colourlessoil, which distilled a t about 207O/30 mm.:0.1898 gave! 0.4142 CO, and 0.1374 H,O.C=59*5; H=8*0.This ester was hydrolysed by boiling with excess of 25 per cent.methyl-alcoholic potassium hydroxide for four hours, and, afterthe methyl alcohol had been completely removed, hydrochloric acidwas added, the whole evaporated to dryness, and extracted withether in a Soxhlet apparatus.The syrup which remained on distilling off the ether, only partlysolidified when it was rubbed with hydrochloric acid, and the solidhad an indefinite melting point, obviously consisting of two or morestereoisomeric modifications. I n order t o isolate one definite modi-fication (trans-), the whole was heated with concentrated hydro-chloric acid in a sealed tnbe a t 190° f o r four hours, the productwas dilute'd with water, filtered from a small amount of carbon-aceous matter, 6 ecolorised wfth animal charcoal, and evaporatedto dryness, during which the acid commenced to separate in colour-less crystals.Aftler remaining in the ice-chest for two days thecrystals were collected, dissolved in ether, filtered from a trace ofinorganic matter, and the ether removed, when a solid remainedwhich was crystallised from hydrochloric acid :0.1786 gave 0.3304 CO, and 0.0948 H20. C=50.4; H=5*9.C,H,,O, requires C=50.0; H=5-6 per cent.On titration 0.1526 required for neutralisation 0*0840 NaOH,whereas %his amount of a tribasic acid, C9H1206, neutralises 0.0847NaOH.tsans-cycloHexuae-1 : 2 : 4-tricurboxylic acid is rather sparinglysoluble in cold, but readily so in hot water, and has a markedtendency t o form supersaturated solutions, which only graduallyC,,H,,O, requires C = 60.0 ; H = 8.0 per centCARROXYLIC ACIDS DERIVED FROM CYCLORUTBNE, ETC.2675deposit, the acid as a hard, opaque, crystalline crust. It melts a tabout 220-222O t o a colourless, viscid syrup, but the exact' point, isdifficult to observe.cis-cycloHexane-1 : 2 : 4-tricarboxylic Acid.-In order t o obtainthis acid, the pure tram-acid (2 grams) was mixed with aceticanhydride (5 c.c.) in a test-tube, and gentlly boiled by means of asulphuric-acid bath for one hour; the temperature was then raised,and the excess of acetic anhydride distilled off.The test-tube wasdrawn out, and the whole heated as rapidly as pcssible over afree flame; when there was much decomposition and a voluminouscarbonaceous mass remained.The! small quantity of brown distillate, which partly solidifiedon keeping, was boiled with much water, filtered from a 1it;tls tar,decolorised with animal charcoal, and then evaporated to 'a smallbulk and mixed with an equal volume of concentrated hydrochloricacid. On remaining in the ice-chest a crystalline crust graduallyformed, and this was coilected and crystallised from water, inwhich it is rather sparingly soluble in the cold, and from whichthe cis-acid separated as a crust of colourless, warty masses, whichhad a quite different appearance from the crystals of the trans-acid :0.1432 gave 0.2630 CO, and 0.0752 H,O.C=50*2; H=5*8.C,HI20, requires C=50*0; H=5*6 per cent.On titration 0.1405 required 0.0764 NaOH f o r neutralisation,whereas this amount of a tribasic acid, C9Hl2O6, neutralises 0*0780NaOR.ciscycloHexnne-1 : 2 : 4-tricarboxylic acid softens a t 218O, meltsa t 225O, and, when it is mixed with an equal amount of thetrans-modification, the mixture softens very much a t 198-200°,and is almost completely melted a t 2 0 8 O (compare p. 2667).trans-cycloHeptaiae-1 : 2 : 4-tricarboxylic A cid,C0,H,CH<Cfl,--C'H2--C CH2*CH(C 02H) !1 2>CH*C02H.The investigation of this interesting acid has been rendered diffi-cult owing to the small yield which is produced by the followingprocess, and it has not been found possible to obtain bett'er resultsby varying these conditions.Ethyl cyczoheptane-1 : 1 : 2 : 4 : 4-penta-carboxylate (p. 2668) was first prepared by mixing 36 grams ofethyl pentanetetracarboxylate,CH(CO,Et),*CH,*CH,* CH,* CH( CO,Et),(Perkin, T., 1887, 41, 240), first with a solution of sodium ethoxidemade by dissolving sodium (4.6 grams) in alcohol (100 c.c.), an26'46 CARBOXYLIC A C I ~ S DERIVED FROM CYCLOHUTANE, ETC.t h e n with ethyl ap-t3il)l.ornopropioiiate (26 grams), and, after t l i pi iiitial somewhat vigorous action had subsided, the whole was heatedin a soda-water bottle in boiling water for four hours. Tlie productwas isolated by addiiig water arid extracting with ether, and wasliydrolysed in the usual irianner (p.2670), the crude, dark brown,syrupy polybasic acid being heated a t 200° f o r an hour, and theresidue esterified by boiling with alcohol and sulphuric acid. Ondistillation a considerable amount of ethyl pimelate,CO,Et*[CH,],*CO,Et,passed over a t about 190-195°/100 mm., and then crude ethylcycloheptanetricarboxylate distilled a t 200-230°/ 30 mm., leavinga considerable dark-coloured residue in the distilling flask. Thecr clde tricarboxylic ester was redistilled, and the fraction distillingatl 212--215°/30 mm. analysed, but the yield obtained was only7 grams, or 8 per cent. of that theoretically possible:0.1960 gave 0.4390 CO, and 0.1504 H@. C ' ' = G l - l ; H=8*5.Cl6HZ6O, requires C = 61.1 ; H = 8.3 per cent.This ester was hydrolysed by boiling with methyl-alcoholicpotassium hydroxide in the usual manner, and, after removal of thealcohol, excess of hydrochloric acid was added, the whole evaporatedto dryness, and extracted with ether in a Soxhlet apparatus. Theresidue from the ether was a syrup, which could not be induced tocrystallise, and was doubtless a mixture of stereoisomerides. Itwas heated with hydrochloric acid in a sealed tube a t 190° forfour hours, the product diluted with water, filtered from a smallamount of carbonaceous matter, and the hydrochloric acid removedby evaporation. The residue was dissolved in water, boiled withanimal charcoal, and again evaporated, when a viscid, colourlesssyrup was obtained, which, on rubbing, almost completely solidified.After contact with porous porcelain, the acid separated from hydro-chloric acid in hard, nodular masses:0*1100 gave 0.2106 CO, and 0.0623 H,O. C =52.2; H= 6.1.C,,H,,O, requires C= 52.2 ; H = 6.1 per cent.On titration 0.0488 required 0.0250 NaOH for neutralisation,whereas this amount of a tribasic acid, C10H1406, should neutralise0.0254 NaOH.trans-cycloHeptane-1 : 2 : 4-tricarboxylic acid melts a t 198-200°,and is readily soluble in water. Unfortunately the amount ofmaterial available was not sufficient for the preparation of thecis-acid.THE UNIVERSITY MUSEUM,OXFORD