LEAN : TETRAHYDROFURFURAN-2 : 5-DICARBOXYLIC ACID. 103 XI.-Ethyl Dibromobutanetetracarboxylate and the Synthesis of Tetrahydrofurfuran-2 : 5-diccwboxylic Acid. By BEVAN LEAN, D,Sc., B.A. IN a paper communicated t o the Society a few years ago (Trans., 1894, 65, 995), the author described a number of homologues of butanetetracarboxylic acid and of adipic acid. It was shown that ethyl butanetetracarboxylate, when treated with sodium ethoxide and alkyl haloids, yields disubstitution derivatives of the general formula ( C02C2H,),RC*CH,* CH2* CR( C02C2H5)2. From these, symmetrical dialkyl butanetetracarboxylic acids and dialkyl ndipic acids were obtained, and it was shown that the latter always existed in two modifications, usually differing from one another markedly in melting point, solubility, and other physical properties.A t the time that this communication was made, it had also been found that bromine could readily replace the hydrogen atoms of the two CH groups in ethyl butanetetracarboxylate, and that the two bromine atoms of the resulting compound could, in turn, be replaced by hydroxyl groups on digestion with barium hydroxide. The in- vestigation of dihydroxybutanetetracarboxylic acid and of its deriva- tives proved, however, to be a matter of much greater di5culty and interest than was anticipated, and it now appears desirable to no longer postpone the communication of the results of this part of the inquiry. Ethyl butanetetrncarboxylate, like ethyl pentanetetracarboxylate (Perkin, Trans., 1891, 59, 827)) readily reacts, in chloroform solution, with bromine with the formation of ethyl as-dibroruobutanetetra- carboxylate and evolution of hydrogen bromide, thus : $!H,*CH(CO C H 5 2 ) + 2Br2 = CH2* I CBr(C02C2H6)2 + 2HBr, CH2* CH(C02C2H6)2 CH2* CBr(C0,C2H5)2 1 2104 LEAN : ETHYL DIBROMOBUTANETETRACARBOXYLATE ANb The beautiful, crystalline dibromo-compound thus obtained, when digested for some hours with a strong solution of barium hydroxide, yields an insoluble barium salt of dihydroxybutanetetrncarboxylic acid, crystallising apparently with one molecular proportion of water.When this salt is decomposed by sulphuric acid, dihydroxybutane- te t racarbox ylic acid, (CO,H),C(OH)*CH,* CH,* C(OH)(CO,H),, is obtained in solution, as is shown by the preparation and analysis of its silver salt, which has the composition C,H6010Ag,.On slowly concentrating the aqueous solution of dihgdroxybutane- tetracarboxylic acid over strong sulphuric acid, beautiful, long, prismatic needles are obtained of the corresponding 8-monolactone, (CO,H),F*CH,* CH,* ?(OH)*CO,H 0 co The silver salt prepared from this substance is silver dihydroxy- butanetetracarboxylate, C,H,O,,Ag,, thus confirming this view of its constitution. On heating an aqueous solution of this 8-lactone in a sealed tube at 150°, decomposition ensues with the elimination of carbon dioxide. I f the aqueous solution of the product is evaporated to a, small bulk on a water-bath and placed over sulphuric acid in a vacuum, small, star-shaped clusters of crystals appear after one or two days; if left longer over sulphuric acid, the whole eventually becomes solid, but the last traces of moisture disappear very slowly.The product had the composition C,H,O,, which is that of dihydroxyadipic acid, less 1 molecular proportion of water, C8Hlo0,0 = C6H805 + 2C0, + H,O. In this case, again, the simplest explanation would be that the substance is the Glactone of dihydroxyadipic acid, namely, CO,H*$!H*CH,*CH,* YH*OH 0 co Careful investigation, however, has shown tbat the silver salt prepared from i t has the composition C,H6O5Ag,, rendering it very improbable that the substance C,H,O, is a &lactone : for this t o be the case, it would be necessary to suppose the lactonic ring to be quite exception- ally stable, remaining intact when the substance is dissolved in water, and also to assume that the hydrogen atom of the hydroxyl group has an acidic character-a very unusual occurrence in an aliphatic compound.Further consideration shows, however, that the elimination of waterTETRAHYDHOFURFURAN-2 : 5-DICABBOXYLIC ACID. 105 actually take9 place between the two hydroxyl groups, and that the substance is tetrahydrofurfuran-2 : 5-dicarboxylic acid, formed in the following way : C02H*C y 3 2 - p 2 y*C02H = C0,H.C y"s'yH2 C*CO,H + H20. \/ 0 AH OH It was found that the tetrahydrofurfurandicarboxylic acid (C,H,05) obtained by heating an aqueous solution of dihydroxybutanetetra- oarboxylic acid begins t o melt about 65", but on raising the tempera- ture the fusion proceeds only gradually, and is not complete until about 120'.The absence of a definite melting point suggested the possi- bility of the substance being a mixture of stereoisomeric acids, and experiments were instituted to separate them if possible by fractional crystallisation from water. Using 32 grams of material, there was little difficulty in separating 6 grams of a tetrahydrofurfurandicarboxylic acid which melted at 123-125'. This was shown to be dibasic by the preparation and analysis of its silver salt, which was found to have the composition C,H,O,Ag,. It is to be noted that it crystallised fairly readily from a little water in star-shaped clusters, and that it was not necessary to dry the crystals over strong sulphuric acid for analysis, exposure in the air being sufficient. It appeared to have no tendency to combine with a molecular proportion of water, as would be the case with a &lactone, for an aqueous solution, after being boiled, gave again the silver salt mentioned above, and not a salt of the formula c6H,06Ag2.The mother liquor from which this acid had been separated was care- fully examined, as is described later (p. 11 3), and from it were isolated 8 grams of small, white crystals, which when dried in the air melted at 56.5-62'. It was at first expected that this substance was a second tetrahydrof urf urandicarboxylic acid ; analysis, however, proved i 1; to have the composition C,H,O,+H,O or C6H1006. The silver salt prepared from it had the composition C6H,Ag20,, a result which led to the conclusion that the substance C,H,,06 was not dihydroxyadipic acid, but the isomeric monohydrate of tetrahydrofurfurandicarboxylic acid.The only solvent from which it was found at all feasible to crystallise this substance was strong hydrochloric acid ; it then melted at 63-64', and analysis proved it to be unchanged in composition. When placed a Few hours over strong sulphuric acid,it was found that the monohydrate, C6H1006, could no longer be fused below 70°, and after 8 days the product melted a t 93-95' and had the composition C,H,O,. Its silver salt was found to have the composition C,H,o,Ag,, ooufirming the conclusion that the substawe was a tetrahydrofurf urm-106 LEAN : ETHYL DIBROMOBUTANETETRACARBOXYLATE AND dicarboxylic acid, and not the isomeric &monolactone of dihydroxy- adipic acid. When this acid (m. p.93-94O) was exposed in the air or dissolved in a little water and the solution evaporated over solid potassium hydroxide, the product melted a t 57-62’, the monohydrate having been reformed. I n the previous paragraphs it has been shown that if an aqueous solution of dihydroxybutanetetracarboxylic acid is heated in a sealed tube at 150°, the product must be regarded as a mixture of isomeric tetrahydrofurfurandicarboxylic acids, the one melting at 123--125O, the other at about 93-94’. The isomerism of these must be geo- metric ; they are, in fact, cis- and trans-forms, recalling the isomeric hexahydroterephthalic acids, and may be represented thus : So far as is known, this is the first case of geometrical isomerism which has been established in the furfuran series.Few instances are recorded of the formation of a furfuran derivative by the removal of the elements of water from an open chain hydroxylic compound. Fischer (Bey., 1891, 24, 2140) has shown that water can be split off from the tetrahydroxyadipic acids, with the formation of 8 furfurandicarboxylic acid (dehydromucic acid), thus : QH(OH)-YH*OH FH-QH YH*OH QH’OH - 3H20 = CO2H*C C*CO,H. C0,H C0,H \/ 0 This action, however, is not strictly analogous t o the case discussed in the present paper, because water was eliminated only when the hydroxyadipic acids were heated with strong acids. Another instance which is more nearly parallel seems t o exist in the case of Tiemann’s ‘isosaccharic acid’ (Bey., 1884, 17, 247 ; 1886, 19, 1257). This was at first regarded as a tetrahydroxyadipic acid, C,H,,OS, but more recently (Ber., 1894, 27, 11 8) was shown t o have the composition C,H&.Tiemann gave the name ‘ nor-isosaccharic acid’ to the tetrahydroxyadipic acid fromwhich ‘ isosaccharic acid ’ might be considered to be derived, and pointed out that on the analogy of saccharic and mucic acids, the substance C,H,07 might be regarded as the lactone of this acid but for the facts that it proved to be dibasic, andTETRAHYDROFURFURAN-2 : b-DICARBOXYLIC ACID. 107 that when heated alone, or in a stream of dry hydrogen chloride, it was changed, without charring, into pyromucic acid, He therefore con- cluded that ‘ isosaccharic acid ’ was really 3 : 4-dihydroxytetrahydro- f urf wan-2 : 5-dicarboxylic acid, OH*HC---CH* OH Of the various compounds prepared from this substance, Tiemann found that several were derivatives of what he continued to term isosaccharic acid,’ but that others contained, in addition, the elements of one molecular proportion of water.The latter, he pointed out, might be regarded as f urfuran derivatives crystallised with lH20, but, as other dihydroxyfurfurans were not known, and as a tetrahydroxy- adipic acid, which lost water when its aqueous solution was evaporated, might be expected to undergo hydration under the influence of chemical agents, he continued to speak of them as formed from ‘ nor-isosaccharic acid.’ In confirmation of this view, he described a tetracetyl derivative of ‘nor-isosaccharic acid’ (Bey., 1894,27, 128)’ but it is not stated how this was isolated, nor are any analyses quoted.It may be noted that some of the derivatives of ‘ nor-isosaccharic acid ’ were found to lose readily one molecular proportion of water, giving rise to corresponding derivatives of ‘ isosaccharic acid,’ and that the latter readily recombined with water in molecular proportion. A number of salts of ‘ nor-isosaccharic acid ’ were described, but almost all of these contained water of crystallisation, and it is not clear how Tiemann satisfied himself that they were not derived from ‘ isosaccharic acid.’ From a review of these facts, it appears t o the author of the present paper that the compounds described by Tiemann as derivatives of 6 isosaccharic acid,’ or of ‘ nor-isosacuharic acid,’ are in reality de- rivatives of 3 : 4-dihydroxytetrahydrofurfuran-2 : 5-dicarboxglic acid, and that the names ‘ isosaccharic acid ’ and ‘ nor-isosaccbaric acid ’ should be abandoned.EXPERIMENTAL. QH2* CBr(CO,C,H,), Eth,yl Dib.r.omobutultetetracarbox~lccte, CH,* CBr(CO 2 2 5 2 80 grams of ethyl butsnetetracarboxylate, free from ethyl butane- tricarhoxylate (compare Lean and Lees, Trans., 1897, ’73, 1062), were dissolved in 230 grams of chloroform in a flask connected with a reflux condenser, and 73 grams of dry bromine run in drop by drop, the flask108 LEAN : EI'HYL DIBROMOBUTANETETRACARBOXYLATE AND being kept cool during the operation. The liquid remained colourless until nearly the whole of the bromine had been added; the flask was then warmed a t 60' for 5 hours on the water-bath, and finally for 1 hour at 70'.The product, which mas still coloured by bromine, was shaken with a small quantity of a solution of sodium hydrogen sulphite, and the residual heavy, yellowish oil washed with water and dehydrated by cal- cium chloride. After the chloroform had been removed as far as possible by evaporation on the water-bath, the residue, on standing, solidified almost completely, forming a beautiful, crystalline mass (about 108 grams). This was broken up and separated as completely as possible from oil by washing with a little ethyl alcohol with the aid of a pump. The white, crystalline product (87 grams) was further freed from traces of oil by dissolving i t in the least possible quantity of hot ethyl alcohol, when the solution deposited, on standing, about 76 grams of colourless, monoclinic prisms.If crystallised again from alcohol, almost exactly the same weight (75 grams) was recovered. On analysis : 0,1517 gave 0,2127 CO, and 0.0680 H,O. 0,1639 ,, 0.1283 AgBr. Br= 32.01. 0,2385 ,, 0.1775 AgBr. Br = 31-67. C!= 38.33 ; H=4*98. C,6H,,08Br, requires C = 38.09 ; H = 4.76 ; Br = 31-74 per cent. E thy1 di bromobutanetetracarboxylate crystallises i n well-formed, mono- clinic crystals from a warm solution in light petroleum, or in methyl or ethyl alcohol, melts at 83O, and is readily soluble in ether, benzene, toluene, or glacial acetic acid. VH2* C(OH)(CO,H), Dih?/drox~bzltanetetracurbox?llic Acid, cH,. C(OH-(CO,H), - When ethyl dibromobutanetetracarboxylnte is hydrolysed in a glass flask, a considerable quantity of si1ic.t and alkali is introduced into solution, from which it is very difficult to free the product, no solvent besides water having been found from which dihydroxybutanetetra- carboxylic acid can be crystallised ; the hydrolysis therefore was carried oiit in a silver flask of 750 O.C.capacity. 80 grams of ethyl dibromobutanetetrncnrboxylate (1 mol.) along with 157 grams of barium hydroxide (2 mols.) freshly crystnllised in a platinum basin,and about 300c.c. of water were introduced into thesilver flask, to the neck of which a small reflux condenser was attached. The mixture was then boiled on a sand-bath for 6 hours. A further quantity of barium hydroxide (157 grams) was then added, and the heating continued for another 6 hours. The product, whiIe still hot, was filtered with the aid of a pump, and the insoluble, white, crystallineTETRAHYDROFURFURAN-2 : 5-DICARBOXYLIC ACID.109 barium dihydroxybutanetetracarboxylate was washed many times with hot water, then suspended in water, and decomposed with the necessary amount of sulphuric acid. After filtration from barium sulphate, the solution was always found to contain a small quantity of bromine ; t o remove this, silver hydroxide was added, and, lastly, the excess of silver was removed by sulphuretted hydrogen. The filtered solution was then evaporated t o about 150 C.C. on the water bath. Silver Salt.-A portion of the solution was neutxalised with ammonium hydroxide solution, poiired into a large excess of silver nitrate solution, and the mixture well shaken ; the white, amorphous precipitahe was collected on a filter, well washed with water and dried on a porous plate, and finilly over sulphuric acid.0.2910 gave 0.1820 AgBr. Ag= 62.54. C,H,0,,Ag4 requires Ag = 62-23 per cent. Bcwium SaZt.--Some of the barium salt formed in preparing the acid was washed many times with boiling water, and afterwards dried on n porous tile and by exposure on a watch glass in the air for some days. 0.6248 gave 0.5218 Bn80,. Ba = 49.1 2. C8H,0,,,Ba, + H,O requires Ba = 49.50 per cent. This salt apparently ~rystitllise~ with lH,O, but the water cannot be estimated by desiccation a t 1 loo, as further decomposition ensues. A strong solution of dihydroxgbutanetetracarboxylic acid has an extremely acrid taste, whilst a dilute solution has a taste very similar t o that of alum.The acid readily liberates carbon dioxide from a carbonate. When the aqueous solution of the acid was concentrated over sulphuric acid in a vacuum, a deposit of barium sulphate formed before crystallisation began. On further evaporation, a little more barium sulphate separated out, and shortly afterwards beautiful, long, prismatic needles of the monolactone of dihydroxybutanetetracarb- oxylic acid began to separ:tte from the slightly yellow, gelatinous mass. &Monolactone OJ Di~~ydroxl/butanetet~acnrbox?/lic Acid3 (CO,H),Q*C H,* CH,. Y(OI-I)CO,H. 0-- co The gelatinous product, in which crystals had begun t o form, was stirred up and after exposure for some days in a vacuum over sulphuric acid became solid and dry. The hard, white, porcelain-like mass was powdered and again exposed over sulphuric acid.The yield from 80 grams of ethyl dibromobutanetetrrtcarboxylate was usually 30-35 grams. The substance began to soften at 145O, and at 1 5 6 O it was110 LEAN : ETHYL DIBROMOBUTAXETETHACARBOXYLATE AND completely fused and frothed up the capillary tube. A t a higher temperature, it charred rapidly, On igniting 1.0966 grams of the substance, 0.0056 gram or 0.5 per cent, of ash was left, which proved to be mainly barium sulphate. In the following analyses, a correction of 0.5 per cent. was made upon the weight of substance taken. No solvent, besides water, was found from which the acid could be crystallised. 0.1548 gave 0.2168 CO, and 0-0505 H,O. C = 38.24 ; H = 3.65, 0.1616 ,, 0.2304 00, ,, 0.0523 H,O.(2-38.86 ; H=3.61. C,H,,Olo requires C = 36-06 ; H = 3-78 per cent. C,H, 0, ,, C=38.68; H-3.25 ,, These analyses showed that the substance consisted probably of the &monolactone of dihydroxybutanetetracarboxylic acid : i t was proved to be a lactone, and not a furfuran derivative, by analysis of the silver salt. The silver salt was prepared, and proved t o be tetrabasic. 0.3410 gave 0.3697 AgBr. Ag = 62.27. 0.245 ,, 0.2662 AgBr. Ag= 62.41. The lactone of dihydroxybutanetetracarboxylic acid crystallises in long needles on slow evaporation of its aqueous solution. It is readily soluble in water, ether, methyl, or ethyl alcohol, but in- soluble in benzene, toluene, or light petroleum. On exposure in the air, it absorbs moisture, but only very slowly. C,HGOloAg, requires Ag = 62.23 per cent.Tetrahydrofurfuran-2 : 5-dicarboxylic Acid. On heating an aqueous solution of the lactone of dihydroxybutane- tetracarboxylic acid in a sealed glass tube, the acid decomposed with the elimination of carbon dioxide. About 2 grams of the lactone of dihydroxybutanetetracarboxylic acid, dissolved in 20 C.C. of water, were heated at 150' in a sealed tube for 6 hours. A very considerable pressure was developed within the tube, so much so that until a special quality of Jena glass tubing was obtained, almost every tube was shattered, and much valuable material lost. On opening the t4ube, a violent escape of carbon dioxide took place, and the decomposition was found to be complete. Fifty-six grams of the lactone were successfully decomposed, and the contents of the tubes were mixed and filtered from a little sediment. A portion was examined as follows : 120 C.C.were evaporated to a small bulk on a water-bath and placed over sulphuric acid in a vacuum; in the course of the next night, a small sediment separated out, which proved to be barium sulphate (compare p. 109) ; th' IS wasTETRAHYDROFURFURAN-2 : 5-DICAHBOXYLIC ACID. 111 removed by filtration and the slow evaporation continued. 'In the course of another day, small, star-shaped forms resembling snow- crystals began to appear, and after four more days almost the whole became solid. The cake was then broken up and placed for a fort- night over sulphuric acid in a vacuum ; the last traces of moisture seemed to disappear very slowly. On ignition, a small amount of mineral matter was left ; 0.5327 gram gave 0.0068 gram or 1-3 per cent, of ash, This was mainly barium sulphate and silica, introduced unavoidably in previous operations ; as no way was found of removing this inorganic material, the necessary correction was made in the following analyses, which proved that the product is not dihydroxyadipic acid.0.1253 gave 0.2056 CO, and 0.0598 H,O. C = 44.73 ; H=5*33. 0.1758 ,, 0.2902 CO, ,, 0,0823 H20. C=44.99 ; H=5*23. CGHl006 requires C = 40.42 ; H = 5.66 per cent. C,H,O, ,, C=44*95 ; H=5*03 5, The siZuei* scdt was prepared in the usual way from a neutral solu- tion of the ammonium salt; it was dried over sulphuric acid and anal y sed. 0.1871 silver salt gave, on ignition, 0.1078 Ag. Ag = 57.63. 0.1692 ,, 9 , ,, 0.0976 Ag.Ag = 57.68. C,H,0,Ag2 requires Ag = 55-09 per cent. cGH,05Ag, ,, Ag=57.73 ,, These analyses, showing the silver salt to be dibasic but with the composition C,H60,Ag,, make it very improbable that the acid C,H,05 is the &lactone of dihydroxyadipic acid ; it must, in fact, be regarded as a tetrahydrofurf urandicarboxyiic acid. This conclusion is supported by the fact that the same silver salt was obtained, without the previous isolation of the acid C,H805, directly from an aqueous solution of dihydroxybutanetetracarboxylic acid, heated at 170' in a sealed tube for 6 hours, and then evaporated on a water- bath with addition of water t o ensure the removal of carbon dioxide. 0.1779 gave 0.1214 CO,, 0,0286 H20, and 0.1019 Ag. C= 18.60 ; H = 1-80 ; Ag = 57.28. C,H,O,Ag, requires C = 19.24 ; H = 1.62 ; Ag = 57.73 per cent.As stated in the introduction, the acid C,H& began to melt a t about 65', but on raising the temperature the fusion proceeded only graduallyand was not complete until about 120'. This absence of a definite melting point could not be attributed to impurity, or to the material being a mixture of entirely different substances in view of the analytical results adduced above, and pointed rather to the112 LEAN : ETHYL DIBROMOBUTANETETRACARBOXYLATE AND possibility of the substance consisting of a mixture of stereoisomeric acids. After preliminary experiments had shown that fracbional crystal- lisation from water, although tedious, promised to effect a separation, the method was carefully applied to 32 grams of the tetrahydro- furfurandicarboxylic acid, Tetrahydrofurfuran-2 : 5dicccrboxyZic Acid, m.p. 123-1 25'. Thirty-two grams of the acid C,H,O, were dissolved in water, and the aqueous solution evaporated slowly to a small bulk on the water-bath, filtered from a little sediment of barium sulphate, and placed over sulphuric acid in a vacuum, when star-shaped clusters of crystals soon began to appear. As soon as at least half of the product had crystal- lised, the crystals were separated as completely as possible from the brown, syrup-like mother liquor by the aid of a pump and dried over sulphuric acid. They melted between 90' and 122'. On repeating the fractional crystallisation, 6 grams of beautiful, white crystals were ultimately obtained which melted at 123-125' ; after solidification, they melted again at the same temperature.The crystals had the same melting point, whether dried a t 100' in a steam oven or by exposure in the air. On analysis, the substance proved to have a composition corre- sponding to that of tetrahydrof urf urandicarboxylic acid. 0.1372 gave 0,2264 CO, and 0.0638 H,O. C = 44.98 ; H = 5.12. C,H,O, requires C = 44.95 ; H = 5.03 per cent. The acid has a very acrid taste. It is extremely soluble in cold water, methyl or ethyl alcohol, acetone, or glacial acetic acid, but is not readily dissolved by ether, and is practically insoluble in chloroform, benzene, or light petroleum. It dissolves in boiling toluene, and is rapidly precipitated, on cooling, in arborescent masses of minute crys- tals.It can be crystallised from a concentrated aqueous solution and from a verystrong solution of hydrochloric acid. The acid is charred extremely readily if cautiously heated in a dry test-tube. A neutral solution of the ammonium salt does not readily yield a precipitate on the addition of solutions ol metallic salts, except in the case of silver and mercurous salts. The silver salt of the acid was prepared and analysed both by igni- tion and by combustion. It was found impossible to estimate the carbon accurately, on account of the great tendency of the salt t o explode, even at a very moderate temperature,TETRAHYDROFUBFURAN-2 : 5bDICARBOXYLIC ACID 113 0,2984 gave 0.1704 Ag. 0.2248 ,, 0.1374 CO,, 0.0382 H,O, and 0.1291 Ag. C = 16.66; C6H60,Ag, requires c = 19.24 ; H = 1 *62 ; Ag =L: 57.73 per cent.The basicity of the acid was determined by titration with a solution of pure barium hydroxide, of which 1 C.C. contained 0.007607 gram Ba(OH),. 0.5160 gram acid required 58-9 C.C. or 0.4480 gram Ba(OH), for neutralisation, whether phenolphthalein or litmus was used as an indicator. Calculated for C6H805, 0.4452 gram, and for G,H,,O,, 0.4002 gram of Ba(OH), would be required to form a dibasic salt with this amount of acid, a difference equivalent t o 6 C.C. of solution. An attempt was made t o open the ring by boiling half a gram of the acid, dissolved in a little water for 6 hours, in a small Geissler flask. The silver salt was then prepared and analysed, proving that no change had occurred; a result which made it still less probable that the substance could be a &lactone.Ag= 57.41. C,H60,Ag2 requires Ag = 57.73 per cent. Ag = 57.10. H = 1.90 ; Ag = 57.40. 0.2292 gave 0.1316 Ag. C6H80&g2 ,, Ag=55.09 ,, ~eti~ahydrofurfurcn-2 : 5-dicarboxylic Acid, m. p . 93-95', and its Hydrate. The brown, syrup-like mother liquor obtained in the course of the isolation of the acid of higher melting point was placed over sulphuric acid in a vacuum, but, even after standing some days, crystallisation did not begin until a fragment of the acid melting at 123-125' was added, and the solution vigorously stirred with a glass rod, when a pasty, crys- talline mass resulted. This was at once filtered by means of a pump, and gave about 6 C.C. of a reddish-brown mother liquor and 11 grams of small, nearly white, sand-like crystals, which, after being spread on a porous plate, were found to melt gradually between 50° and 65'.The crystals were dissolved in a little cold water, and the solution, after filtration from a slight sediment, allowed to evaporate slowly over sulphuric acid in a vacuum until it became syrupy, without, however, any separation of crystals occurring, but on adding a small fragment and stirring the syrup suddenly became almost solid with sensible evolution of heat. The product was dried in the air on a porous tile, giving 8 grams of small, white crystals, which were found to melt at 56.5-62"aiid when dried over solid caustic potmh in a vacuum,114 LEAN : ETHYL DISROMOBUTANETETRACARkOXYLA'l'E A N D showed no change in the melting point.A small portion was more completely dried in a steam oven at 100'; after melting, it did not solidify on cooling until scratched with a glass rod, when it at once became solid and brittle; it then melted at 59-62O. A portion mas also placed in a vacuum over sulphuric acid, and the next day was found to be caked together, crisp, and nodular ; it then began to melt a t 63O, but was not completely fused even at 70". This raising of the melting point was subsequently found t o be due t o the partial dehydration of the hydrate of tetrahydrofurfuran- dicarboxylic acid (m. p. 59-62O), the anhydrous acid having a higher melting point (see p. 115). I n the account of the isolation of these substances, mention is made of 5 C.C. of a reddish-brown mother liquor which might conceivably con- tain a compound of still lower melting point, although contaminated by impurities accumulated in the numerous processes of the prepara- tion.Careful examination, however, failed to reveal the presence of any other substance, Hydrate of Tetrahydrq%rfuraficrficrccn-2 ; 5-dicarboxylic Acid, C6H80, -+ H,O. -The small, wbite crystals (8 grams), which when dried in the air melted at 56-5-62', proved t o be a hydrate of tetrahydrofurfuran- dicarboxylic acid. Prior to analysis, the crystals were dried over solid caustic potash in a vacuum f o r 3 days; they gave a slight amount of ash after ignition (0.3 per cent.), and the necessary cor- rection WAS applied t o the analytical results : 0.1593 gave 0.2360 CO, and 0.0812 H20.C = 40.37 ; H= 5.70. 0.1188 ,, 0*1777 GO, ,, 0.0633 H,O. C =r 40.76 ; H = 5.95. C6HI0O6 requires C = 40.42 ; H = 5-66 per cent. C6H,0, ,, C144.95; H = ~ 5 * 0 3 ,, This conclusion was confirmed by determining the bccsicity of the acid with a solution of barium hydroxide containing 0.007607 gram Ba(OH), per C.C. 0.4340 gmm acid required 55.2 c.c., or 0.4801 gram Brt(OK), for neutralisation whether phenolphthalein or litmus was used as an indicator, Calculated for CGHZO06, 0-4176 gram, and for C6H805, 0.4645 gram of (BaOH), would be required to form a dibasic salt with this amount of acid. The formula C,HlOO6 represents both dihydroxyadipic acid and also monohydrated tetrahydrofurfurandicarboxylic acid. The for- mation and analysis of the silver salt showed, however, that the substance was the latter.The silver salt was prepared and dried a t looo 1TETRAHYDROFURFURAN-2 : 5-DICARBOXYLIC ACID. 115 0'1735 gave 0*1288 CO,, 0.0292 H,O, and U*OO91 Ag. C = 19.30 ; H = 1 *87 ; Ag = 57.14. C,H,O,Ag, requires C = 19624 ; H = 1.62 ; Ag = 57.73 per cent, From another preparation, made 3 years previously, a silver salt was prepared which gave Ag = 57.53 per cent. The hydrate of t etrahydrofurfurandicarboxylic acid is extremely soluble in cold water, methyl or ethyl alcohol, acetone, or glacial acetic acid, but is not at all readily soluble in boiling ether, and dissolves only sparingly in boiling benzene, toluene, or light petrol- eum. It is soluble in concentrated hydrochloric acid, but if the solution is stirred and left over solid caustic potash the hydrate can be induced to crystallisc in beautiful, white plates.Concentrated hydrochloric acid is therefore the only solvent from which the hydrate can be crystallised without very great loss. The crystallisation was carried out as follows : 3 grams of the hydrate were dissolved in a little water, the solution filtered from a very small amount of sediment, and hydrogen chloride passed into it until the solution was saturated. After concentration in a vacuum, a white, crystalline precipitate was induced to form by adding a crystal and stirring vigorously. The product was collected, washed with a little strong hydrochloric acid, spread on a porous tile, and dried over solid caustic potash in a vacuum. I n this way, 1.3 grams were obtained, which on analysis proved t o be the substance C,H,,O, un- changed in composition.I t melted at 63-64', and this may probably be taken as the correct melting point rather than 56-5-62', the melting point before crystallisation from hydrochloric acid. An aqueous solution of the hydrate is intensely acrid. A neutral solution of its ammonium salt gives a white precipitate with silver or mercurous nitrate,' but no precipitate with barium nitrate, calcium chloride, lead acetate, or mercuric chloride. Tetrahyd~ofurfu~an-2 ; 5-dicarboxylic acid, m. p . 93-95°.--It has already been stated that, after the hydrate of tetrahydrofurfurandi- carboxylic acid had been placed for 1 day over sulphuric acid in a vacuum, the crystals were found to be caked together and could no longer be melted below 70'.A portion of the hydrate, which had been recrystallised from hydrochloric acid and melted at 63--64O, was exposed for 8 days over strong sulphuric acid in a vacuum. It then melted at 93-95'. On analysis : 0.1513 gave 0.2486 CO, and 0*0700 H,O. C = 44.79 ; H = 5-17. C,H,O, requires C = 44-95 ; H = 5.03 per cent. This result showed that the riiising OF the melting point mas due to the loss of a mcleculw proportion of water, and was confirmed by116 LEAN : TETRAHYDROFURFURAN-2 : 5-DICARBOXYLIC ACID. determining the actual loss in weight which occurred when the hydrate was exposed over sulphuric acid. 1.1737 gmm of the hydrate lost 0.0638 gram or 5-4 per cent. in 1 day; at the end of 10 days, the weight remained constant, and 0.1235 gram had been lost, or 10.52 per cent. The calculated loss is 10.10 per cent. A determination of the 6asicity of the substance (m. p. 93-95') with a solution of barium hydroxide containing 0.00917 gram barium hydroxide per C.C. confirmed the conclusion that it had the composition 0.4342 gram required 50.39 C.C. or 0.4622 gram Ba(OH), for neutralisation, whether phenolphthalein or litmus was used as the indicator. Calculated for C,H,,O,, 0-4176 gram, and for C6H805, 0.4645 gram of Ba(OH), would be required to form a dibasic salt with this amount of acid, 'GHSo5* The silver sult was prepared and analysed : 0,1752 gave 0.1227 CO,, 0.0295 H,O, and 0.1007 Ag. c6H,05Ag2 requires c = 19-24 ; I€ = 1.62 ; Ag = 57.73 per cent. It follows therefore that the substance C,HSo5 cannot be the &lactone of dihydroxyadipic acid. When a portion of the tetrahydro- furfurandicarboxylic acid melting a t 93-94' was dissolved in a little water and the solution allowed to evaporate over solid potassium hydr- oxide until dry, the product was found to melt at 57-62'. Further, on exposing some of the acid in the open air for 7 days, the melting point was lowered until it became 57-62'. There was not sufficient material for analysis. These results show that this tetrahydro- furfurandicarboxylic acid can be readily converted into its hydrate, C= 19.08; H = 1.88 j Ag = 57.46. A portion of the expense incurred in this investigation was defrayed by a grant awarded by the Government Grant Committee of the Royal Society, for which the author desires to express his thanks. The author's thanks are also due to Mr. F. H. Lees for his very care- f u l and assiduous assistance in the later stages of the work. THE OWENS COLLEGE, MANCHESTER ; AND ACKWORTH SCHOOL