FORMATION OF a- AND 6-ACROSE FROM GLYCOLLIC ALDEHYDE. 129 XII1.-Formation of a- and 6-Acrose from GZycoZZic A Zdehyde. By HENRY JACKSON, B.A., Fellow of Downing College, Cambridge. GLYCOLLIC aldehyde, first obtained in a dilute aqueous solution by Fischer and Landsteiner (Bey., 1892, 25, 2549), was shown by these authors to lose its power of reducing Fehling's solution in the cold after treatment with 1 per cent. aqueous caustic soda at 0" for 15 hours. If the condensation product, after acidification with acetic acid, was heated with phenylhydrazine acetate on the water-bath for 8 hours, a crude osazone separated out which, after purification with ether and crystallisation, first from hot water and finally from dry benzene, gave a pure osazone which melted at 166-168' and had a composition very similar to that required for tetrosazone.I n 1897, Fenton found (Trans., 71,375) that when glycollic aldehyde in the form of a syrup is heated at 100" under very diminished pres- sure, it undergoes condensation, and by fractional solution with absolute alcohol he obtained a sparingly soluble portion which, on treatment with phenylhydrazine acetate, gave a normal hexosazone.130 JACKSON : FORMATION OF a- AND P-ACROSE In a communication to the British Association this year, it wag shown by Fenton and the author that if a dilute aqueous solution of glycollic aldehyde, prepared from dihydroxymaleic acid ( Fenton, Trans., 1895, 67, 778) or glycol (Fenton and Jackson, Trans., 1899, 76, 2), is treated with dilute sohitions of sodium or calcium hydr- oxides at the ordinary temperature of the laboratory (about 15O), it quickly undergoes condensation, losing its power of reducing Fehling’s solution in the cold and of restoring tha colour to an alcoholic solu- tion of magenta, which had been decolorised by sulphur dioxide.After neutralising the product with acetic acid and warming on the water-bath for 3-4 hours with phenylhydrazine acetate, a beautiful, yellow osazone separated out on cooling, which, after crystallisation from boiling water,and afterwards from benzene and from ethyl acetate, melted sharply at 158O, and on analysis was found t o be a normal hexosazone. Its melting point and action towards solvents pointed to its identity with p-acrosazone which Fischer and Tafel obtained from the condensation product of ‘‘ glycerose ” (Bw., 1887, 20, 3384).Considering the readiness with which the condensation by alkalis took place at the ordinary temperature, it appeared to the author that it would be interesting to repeat the experiments conducted by Fischer and Landsteiner, and, with .Mr. Fenton’s .approval, this has been done. Action of Dilute Caustic Soda Solution ut Oofoq* 15 Hours. The details of the experiment were briefly as follows :-Pure glycollic aldehyde was diluted with distilled water until a solution containing 3 per cent. of the aldehyde was obtained, the strength being determined by Fehling’s solution, as the reducing power of tho pure aldehyde is known (Trans., 1899, 75, 579). To the solution, cooled to Oo, dilute caustic soda solution of known strength was added until the mixture contained 1 per cent.of the alkali. The combined solutions were then kept at 0’ for 15 hours, when i t was found t h a t the liquid, whioh was originally colourless, had become orange-yellow. It still had the power of reducing Fehling’s solution in the cold, although not so strongly as glycollic aldehyde, and also of restoring the colour to decolorised magenta solution. After neutralising the condensation product with acetic acid, phenyl- hydrazine acetate was added in excess, and the mixture allowed t o stand for 12 hours. The solution was then filtered and heated on the water-bath at 100’ for 4 hours; longer heating was not found to materially increase the yield, whilst the osazone was far more resinous. After cooling and standing, a bulky, dirty yellow osazone separated out, which, after removal by filtration and drying in the air, wasFROM ULYCOLLIC ALDEHYDE.131 rubbed with small quantities of dry benzene to remove the resinous matter, arid then warmed with 50 times its weight of dry benzene. As only about half was soluble, the solution was quickly filtered on the pump from the sparingly soluble portion. The benzene solution, on standing, deposited a flocculent mass of yellow needles, which were collected and examined. Part soluble in Benzene.-After being recrystallised from hot water and dried in the air, this fraction was crystallised twice from the least possible amount of boiling dry benzene. It was thus obtained as a mass of fine needles which melted sharply at 167", and on analysis gave the following numbers, proving it to be tetrosazone : 0.1230 gave 0.2893 CO, and 0,0672 H,O.C = 64% ; H= 6.07. 0.1452 ,, 0.3409 CO, ,, 0.0785 H20, C=64*03 ; H=6.01. 0.1710 ,, 27.5 C.C. nitrogen a t 19O and 758 mm. N= 18.50. C16H,,0,N, requires C = 64-43 ; H = 6-04 ; N = 18.9 per cent. These numbers agree closely with those obtained by Fischer aEd Landsteiner, and it would appear that tetrose is undoubtedly one of the condensation products of glycollic aldehyde. Part sparingly soluble in Benzene.-This portion was again boiled with small quantities of dry benzene t o remove all traces of tetr- osazone. The osazone was dried in the air, and finally at 100"; on analysis, it gave the following numbers : 0.1484 gave 0.3270 CO, and 0.0854 H,O.C =: 60.23 ; H = 6.37. 0*1120 ,, 15.06 C.C. nitrogen a t 14" and 749 mm. N=15*80. C,,Hz20,N4 requires C = 60.33 ; H = 6.15 ; N = 15.64 per cent. It was evident from these numbers that the substance was either a hexosazone or a mixture of hexosazones. It was therefore warmed with 10 times its weight of ethyl acetate, when only a portion dis- solved ; the insoluble part was removed by filtration, and on standing a mass of yellow needles crystallised out from the solution; this was recrystallised from hot water and finally from ethyl acetate. The osazone melted sharply at 158". On analysis, the following numbers were obtained : 0°1380 gave 0.3042 CO, and 0.0746 H,O. C = 60.12 ; H = 6-02 per cent. The osazone dissolves sparingly in ether or benzene, but easily in ethyl acetate, and is soluble in hot water.The analytical results, melting point, and action towards solvents point to its identity with P-acrosasone. The portion sparingly soluble in ethyl acetate was left as a greenish- yellow powder. This was washed with small quantities of hot absoIufe 0,1660 ,, 0.3655 CO, ,, 0.0893 H,O. C=59*91 ; H=5*98.132 JACKSON: FORMATION OF a- AND P-ACROSE alcohol, and then crystallised from 95 per cent. alcohol. The osazone which separated out was next crystallised from 98 per cent, alcohol, and finally from hot absolute alcohol, in which i t was sparingly soluble. It was thus obtained aR a mass of yellowish needles, which were dried in the air and melted at 208-2109 On analysis, it was found to be a normal hexosazone : 0.1120 gave 0.2814 CO,.The melting point and sparing solubility towards solvents would point to the substance being a-acrosazone, obtained by Fischer and Passmore from '' formose " (Ber., 1889, 22, 359), and by Fischer and Tafel from the condensation product of '' glycerose " (Rev., 1887, 20, 3384). It would therefore seem that the condensation of glycollic aldehyde by dilute soda at Oo, if continued for a short time, results i n the formation of tetrose and of a- and p-acrose. The power which the product has of reducing Fehling's solution in the cold is probably due to tetrose, as i t has been shown that the dilute aqueous solution of tetrose obtained by the oxidation of erythritol is able to effect the reduction in the cold (Fenton and Jackson, Trans., 1899, 75, l ) , whereas the two hexoses only bring this about on warming.C = 60.42 per cent. Action of Dilute Caustic Soda Solution at 0" fop* 2 Days. In the next experiment, the condensation was allowed to go on for 2 days; the solution was then found to reduce Fehling's solution in the cold, but only to a F-ery slight extent. It was neutralised with acetic acid, heated with phenylhydrazine acetate a8 before and the crude osazone rubbed with dry benzene. On heating with a large quantity of benzene, only a very small portion dissolved, the major portion being very sparingly soluble in this solvent. From t h e benzene solution, a small quantity of a bright yellow, crystalline osazone was obtained which, after drying, melted at 157O, and gave on analysis the following numbers, indicating that it was a mixture of tetrosazone with a hexosazone : 0.1635 gave 0.3725 GO, and 0.09 H,O.C = 62.12 ; H = 6.11 per cent. By fractionation with dry ether, i t was separated into two portions, tetrosazone being readily soluble whilst pure p-acrosazone is only very sparingly soluble in this solvent. The portion soluble in ether melted at 166q and on analysis gave numbers corresponding t o those required for tetrosazone : 0.0930 gave 0.2175 CO,. C=63.86 per cent.FROM GLYCOLLIC ALDEHYDE. 133 whilst the sparingly soluble portion melted at 157' and on analysis gave numbers corresponding to those required for a hexosazone. 0.1103 gave 0.2233 CO,. The major portion of the condensation product, namely, that spar- ingly soluble in benzene, mas found on analysis to be a hexosazone, and was separated in the manner previously described into p-acrosazone melting at 1 5 8 O , and a-acrosazone melting at 208'. C = 60.7 per cent.Action of Dilute Caustic Soda Xolution at 0' for 6 Days. I n the final experiments, the condensation was allowed to proceed until the solution did not reduce Fehling's solution after standing for half an hour in the cold. This was found to be case after the dilute aqueous solutions of glycollic aldehyde and caustic soda (1 per cent.) had remained at Oo for 6 days, The solution was neutralised with acetic acid and warmed with phenylhydrazine acetate as before. On boiling with benzene, only a small quantity dissolved, and this crystallised out as a mass of yellow needles which melted at 1 5 8 O , and on analysis was found to consist of P-acrosazone mixed with a trace of tetrosazone. 0.1185 gave 0,2614 CO,. The major portion was, as before, separated into a- and p-acros- azones. It would therefore seem that the tetrose formed by the con- densation of glycollic: aldehyde is unstable in the presence of dilute alkalis, and this view is emphasised by the fact that in the condensa- tion at the ordinary temperature no tetrose could be found. C=6Oo60 per cent. Gclycollic aldehyde is the last member of the series of aldehydes from which synthetical hexoses can be obtained ; the formation of a- and p-acrose by its condensation is of much interest, as Fischer and his pupils have previously shown that the same sugars are formed by the condensation of formaldehyde and of glycerose. UNIVERSITY CHEMICAL LABORATOEY, CAMBRIDGE. VOL. LXXVXI