104 FRXNKLASD AND MACOREUOR : ETHEREAL SALTS XII1.-Elhereal Salts of Active aizd Inclctive Monoben- xoyl-, Bibenzoyl-, Dipkenacetyl- and Dipropionyl- glyceric acids. By PERCY FRANKLAND, Ph.D., B.Sc., F.R.S., and JOHN MACGREGOR, M.A. IN pursuing our investigations on the relationship between chemical constitution and optical activity, we have prepared a number of derivatives of active glyceric acid (dextrorotatory) in which the carboxylic hydrogen is replaced by positive radicles, whilst the hydroxylic hydrogen atoms (either one or both) are replaced by the acid radicles benzoyl and phenacetyl. It will be remembered that we hare already (Trans., 1893, 63, 511, 1410, 1419, and 1894, 65, 750) prepared a number, not only of the ethereal salts of active glyceric acid itself, but also of active diacetylglyceiic acid, the com- pounds under consideration in tho present communication being intended to throw further light on the influence exerted on the rota- tion by making these substitutions in the molecule. Methylic Dibeiuoy lglljcerate (Active).This was prepared by running active methylic glycemte from a dropping funnel into twice the calculated quantity of benzoylOF ACTIVE AND INAC'I'IVE GLYCERIC ACIDS. 105 chloride, contained in a flask heated at 120" and finally to 1gO" by means of an oil bath; a vigorous action takes place, hydrogen chloride being given off. The excess of benzoyl chloride was distilled of€ under reduced pressure, and the residue then frac- tionated under diminished pressure ; this was attended with some difficulty in consequence of the solidification of the distillate in the lateral tube of the distilling flask.A large part of the liquid passes over at 245-247". After some hours, the distillate crystallised in tufts of long, flat needles radiating from centres ; these crystals are easily soluble in chloroform, acetone, and benzene, soluble also, especially on warming, in methylated spirit, from which recrystal- lisat'ion can most advantageously be effected, the long, radiating needles having the appearance of thistle-down. The crystals melt at The crystalline substance was submitted to hydrolysis with 58-59'. alcoholic caustic potash with the following results. I. 1.2500 gram required 0.6449 gram KOH for hydrolysis. 11. 1.1032 ), ,, 0,5706 7 , 7 7 7 9 I. 11. 100 parts by weight of substance required KOH for hydrolysis .., . . . . . . . . . . . . . 51.59 51.73 100 parts by weight of methylic monobenzoylglycerate, CIIH,,O,, 100 parts by weight of methy!ic dibenzoylglycerate, C&1606, On combustion the following results were obtained. 0.2093 gave 0.5044 CO, and 0.0927 H,O. require 50.00 parts KOH. require 51.86 parts KOH. Calculated for methylic Calculated for methylic Found. dibenzoylglycerate. monobenzoylglycerate. Carbon.. , . 65-72 65.85 58.93 per cent. Thus, both hydrolysis and combustion clearly show that the sub- stance is pure methylic dibenzoyl and iiot monobenzoylglycerate. Owing to the high melting point of this substance, the rotation had to be determined at higher temperatures in a state of fusion, and the rotation for ordinary temperatures deduced by extrapolation.The density, compared with water at 4", was determined at 65' and '39" respectively. Hydrogen., 4.92 4*%3 5.35 ,, a 6 5 0 ~ 1.1836. d 99O/4" 1.1581. from which it appears that the diminution in density proceeds by 0*00075 for 1" rise i n temperature. VOL. LXlX. I106 FRANKLAND AND MACGREGOR : ETHEREAL SALTS The substance was submitted to polarimetric examination at the following temperatures i n a tube 44 mm. long, placed in an air chamber surrounded by a water jacket. Observed rotation. aD in 44 mm. tube. Temp. 80.5'. ..... +10*26' 77.0 ...... 10.51 74.0 ...... 10.67 71.5 ...... 10.79 67.0 ...... 11.10 63.5 ...... 11.30 59-5 ...... 11.57 Density compared. with water a t 4'. 1.1720 1.1 746 1.1769 1.1787 1.1821 1.1 847 1.1877 [a1 D.+ 19.89O 20.33 20.60 20.80 21-33 21.67 22-13 On plotting out the above figures on a diagram (see p. 122) in which the specific rotations are represented by ordinat,es and the temperatures of observa,tion as abscissse, it is found that the observed specific rotations lie almost exactly on a straight line, from which by extrapolation the value for the specific rotation at 15' can be obtained as [aID = + 26.89'. This value may therefore serve for purposes of comparison with active compounds which have had their rotation determined at that temperature. Methy lic Dibenzoylglycerate (Inactive). This was prepared in substantially the same way as the active compound above, The solutions in methylated spirit exhibited less tendency to crystallise than similar solutions of the active compound, hut ultimately very similar crystals, long needles 1 inch in length and radiating from centres, were obtained ; these melted at 44-46", or about 14Obelow those of the active compound.0.2483 gave 0.5978 COz and 0.1093 H,O. Methylic dibenzoylglycerate, Cl8HI6O6, requires C = 65.85 ; H = 4.88 per cent. The crystals are soluble in the same solvents as those of the active compound. The molecular weight of this compound, as determined by the cryoscopic method, has been ascertained in benzene, nitrobenzene, ethylene dibromide, and acetic acid solutions (see next paper, p. 123). In these solutions, there is no evidence of the existence of the double molecules corresponding to a racemate, the cryoscopic values for the molecular weight being essentially similar to those obtained for the active methylic dibenzoylglycerate.The following results were obtained on combustion. C = 65-66 ; H = 4.89.OF ACTIVE AND INACTIVE QLYCERIC ACIDS. 107 Ethylic Dibe?azoylglycemte (Active). This was prepared in the same way as the corresponding methylic compound, by running the active ethylic: glycerate into an excess of Inenzoyl chloride heated t o 149-175'. On subsequent fractionation, the greater part distilled over under diminished pressure at 240-260°, and eventually the true boiling poiiit was found to lie between 254' and 258' under about, 10 mm. pressure, 0.2498 gave 0.6096 C02 and 0.1185 H,O. C = 66.56 ; H = 5.27. C,9H1806 requires C = 66.67 ; H = 5.26 per cent. During the severe weather of the last winter this liquid began t o crystallise, and by means of these crystals it was found possible to start its crys tallisation in strong alcoholic solution.A further quantity of this active ethylic dibenzoylglycerate was subsequently prepared with the object of obtaining it., if possible, without distillation. The excess of benzoyl chloride was removed by distillation under diminished pressura, and the residue in the flask was dissolved out with alcohol, but the alcoholic solution could not be brought t o crystallise even on sowing with some crystals of the previous preparation. The alcohol was, therefore, distilled off, and the residue, after being washed with a solution of sodium carbonate t o remove any benzoic acid that might have been formed from the benzoyl chloride, was dissolved in ether and washed with water.The ethereal solution was evaporated, and the residue placed in it vacuum desiccator, a crystal of the previous pi-eparation being added ; crystallisation could, however, not be induced. This undis- tilled product was also submitted to combust'ion and to polarimetric examination in benzene solution ; these determinations, however, showed that it was of inferior purity and inferior rotatory power to the preparation obtained by distillation. The product was therefore distilled under diminished pressure, and in the distillate crystallisa- tion was induced by sowing with a crystal from the prerious pre- paration. The crystals, which are needles radiating from centres, melt at 25'; they are much more soluble in alcohol than the cor- responding methyl compound. 0.2522 gave 0.6144 GO, and 0.1192 H,O.C = 66.44 ; H = 5.25. C19H,,0s requires C = 66.67 ; H = 5.26 per cent. The optical activity of this crystalline specimen was determined i n benzene solution, and found to be the same as that exhibited by the benzene solution of the first preparation, which had a t the time only been obtained in the liquid state. Thus, the liquid ethylic dibenzoyiglycerate obtained in the firsti preparation by distillation alone, and that obtained in the secoiid I 2108 FRANKLAND AND MACGREGOK. : ETHEREAL SALTS preparatioii by distillation and subsequent crystallisation were of equal purity, as determined both by combustion and by polarimetric observation. The density, compared with water a t 4", was determined a t 60" and 98.5" respectively. a i50/40 1.2010.a 60°/40 1.1596. d 9 8 . ~ / 4 0 1.1282. The substance was submitted to polarimetric examination at the following temperatures. Observed rotation. Density compared Temp. CZD i n 44 mm. tube. with water a t 4'. 1.1 D- 83.0'. ..... +10-02O 1.1407 + 19.95' 60-0 . . . . . . 11-49 1-1596 22.52 49.5 ...... 12.11 1.1693 23.53 22.0 ...... 13.65 1.1946 25.96 16.5 ...... 13.87 1.1996 26.28 On plotting the above figures on a diagram (see p. l22), it is found that the specific rotations for the temperatures from 83' to 49.5' lie almost exactly on a straight line, whilst the line joining the values for the lower temperatures exhibits a very slight but distinct droop. Thus, i f the value of the specific rotation for 16.5" be calculated from the change i n rotation between the higher temperatures of observation, the following values are obtained.Value of [a]D a t 16.5", calculated from specific rotations observed at + 27-39'. . . . . . . . . . 83-60.0" +27.06 . . . . . . . . . . 83-49.5 M hich values depart, therefore, but slight'ly from the actually observed r d u e for 16.3", which was found to be [a]D = +26*23". On further extrapolating the value of [alD for 15" fyom the values actually obtained for 22" and 16*5O, the result is Temperatures. +26.50 .......... 83-22.0 [ a ] D = +26*37" for 13". which is only 0.5" inferior to the specific rotation calculated for the inethylic compound at the same temperature. I n thus extrapola- ting, it is difficult to know which part of the line t o calculate from; thus, in the above we have taken as the basis of calcula- tion the variation in rotation between 22" and 16*,5O, a range of tem- perature which is wholly below the melting poiiit of the compound, namely, 25".A more rational mode of extrapolation would appear 10 be to calculate from ranges of temperature which bear the same relationship to the melting points of the two compounds. Thus, the inethylic compound melts a t 5g0, and the ethylic a t 25" ; we should,O F ACTIVE AND INACTIVE GLYCERIC ACIDS. 109 therefore, calculate in the case of the former from the variation in observed specific rotation hetween 80.5' and 59.5', whilst in the case of the ethylic salt from the variation between 49.5' and 22".Calcu- lating in this way, the values become [ U ] D for 15". Methylic diberizoylglycerate . . . . . . . + 26.89 E thy lic 9 , . .. .. .. +26-58 In consequence of the very slight difference between the specific rotations of methylic and ethylic dibenzoylglycerates at the tem- peratures referred t o above, we thought it desirable to ascertain their rotations at a much higher temperature, and for this purpose we passed a current of aniline vapour through the jacket chamber surronnding the internal metal tube in which the glass polarimeter tube is placed. In this way, a constant temperature of 183' was obtained in the in- ternal tube, but as at this temperature it would have been impossible to use the india-rubber rings, with which we ordinarily bring a gentle pressure to bear on the glass discs closing the ends of the polarimeter tube, we employed instead small spiral springs of copper wire, which answered the purpose admirably.The following resulhs were obtained- Methy lic? diben zoy lgly cerate. Ethylic dibenzojlglycerate. aD = +4*12'. Z 44 mm. a, = +4*02'. Z 44 mm. d 183'/4" 1,0951. d 183'14' 1.0599. c- -v----d at 183'. The specific rotations of these compounds at this high temperature may also be arrived at, by extrapolation from the observations made at lower temperatures, arid these calculated values are in both cases most remarkably concordant with the above experimental results ; thus, the extrapolated values are for 18.3' Methylic dibenzoylglycerate . . . . [a]D = +9*01' Ethylic 9 ) . . . . ,, = + 8 * i 8 In order to ascertain whether exposure to this high temperature had permanently altered the rotatory power of these compounds, the rotations were again taken in the same specimens at one of the lower temperatures at which they had been previously determined, with the result that precisely the original figures were obtained.We have thus shown that the rotation of these two compounds is exceedingly sensitive to temperature, [a]= at 15' being i n each case three times as great as it is at 183". (See also diagram, p. 122).110 FRANKLAND AND MACGREQOR : ETHEREAL SALTS P~opylic Dibenzoylglycerute ( A c t i v e ) . This was prepared on the same lines as the corresponding methylic and ethylic compounds. The active propylic glycerate was run into twice the calculated quantity of benzoyl chloride a t 140°, and the mixture then raised to 180'.After removing the excess of benzoyl chloride by distillation, mach difficulty was experienced in distilling the residue a t a low pressure, in consequence of violent bumping, nor was this tendency to bump removed by dissolving the et4hereal salt i n ether and then washing with a solution of sodium carbonate. Ultimately, after repeatNed fractionation, the main distiliate passed over between 267' and 269". This product, on combustion, yielded the following results. I. 0.2436 gave 0.5992 CO, and 0.1221 H,O. C = 67.09; H = 5.57. 11. 0.2341 ,, 0.5769 ,, ,, 0.1191 ,, C = 67.21; H = 5.65. C2,H2,06 requires C = 67-42; H = 5.62 per cent. The density, compared with water at 4", was determined a t 15", 25O, 60', and 98*5', d 15'14" 1 1807 ; d 60*0°/4" 1.1399 ; d 25"ld' 1.1727 ; d 98*5"/4" 1.1079.from which it appears that the diminution in density proceeds by 0*00080 for 1" rise in temperature between 15' and 25', by 0.00094 for 1' between 25" and 60', and by 0.00083 €or 1' between 60" and 98.5'. The substance was submitted to polarimetric examination at the following temperatures in a tube 44 mm. long, placed in an air chamber surrounded by a water jacket. Observed rotation UD in 44 nim. tube. Temp. 19.5' . . . . . +10.73' 340 .. . . . 10.10 38.0 . .. . . 9.96 48.0 . .. . . 9-53 56-0 .. . . . 9.10 68.2 . .. .. 8-47 78.0 , .. .. 8.02 87.0 ..... 7.54 Density compared with water a t 4'. 1.1771 1,1642 1.1605 1.1511 1.1436 1.1331 1.1250 1.1175 cab + 20.71O 19.71 19-50 18.81 18-08 16.99 16.20 15.34 These figures have been plotted on a diagram (see p.122), from which it appears a,gain that the line joining the specific rotations taken a t the higher temperatures is almost straight, but that it drops slightly with the lower temperatures. In extrapo1at)ing the value of [aID for 15', we have no melting point to take into consideration, as we have only had to deal withOF ACTIVE AND INACTIVE GLPCERIC ACIDS. 111 this substance in the liquid state ; we may therefore reasonably cal- culate from the variation in specific rotation observed between 38O and 19*5", Erom which results the value [ a ] , = +21*00 for 15". The rotation of the propylic is thus markedly inferior to that of either the methylic or ethylic compounds. Methylic Diphenylacetylg lycerate (Active).The method of preparation was similar to that employed in the case of the other ethereal salts described above. The active methylic glycerste was run into a large excess of the phenylacetyl chloride ; the action appeared to commence at BOO, becoming more vigorous at 105' ; the temperature was finally raised to 160'. After the excess of phenylacetyl chloride had been removed by distillation under diminished pressure, the residue was washed with a warm solution of sodium carbonate, extracted with ether, separated, and the ethereal solution washed with water. The ether was then removed by distil- lation, and the residual liquid fractionated at low pressure, the main distillate being obbined at 265-273'. This distillate was agairz washed with sodium carbonate solution, extracted with ether, shaken up with animal charcoal, as it was slightly coloured, and then dried in a vacuum desiccator.A combustion made with this product showed it to be impure methylic diphenylacetylglycerate. It was further purified by repeated distillation in 8 vacuum, hhe boiling point being 266-270" under 1 7 mni. pressure. On combustion, t8he following results were obtained. 0.2385 gave 0,5877 CO, and 0.1231 H,O. The density was determined at the following temperatures C = 67.20; H = 5.73. C,oHzoOs requires C = 67-41 ; H = 5.62 per cent. d 142"/4O 1.1975 ; cl 41"/4' 1.1737 ; d 8Oo/4O 1.1404. from which it appears that the diminution in density proceeds by 0.00089 f o r 1' rise in temperature between 14.2' and 41°, and by 0.00085 for 1' between 41' and SO3.Polarimetric determinations were made at the following tempera- tures. Observed rotation, Density compared Temp. U D in 92.35 mm. tube. with water a t 4O. La] D. 14.5'- . . . . -17.76' 1.1972 - lCi-06O 34.0 -35.0". . . . -16.62 1.1794 - 15-25 45.5 4 6 . 5 . . . . -16.03 1-1694 - 14-84 49.5 -50.5 . . . . -15.88 1-1660 - 14.74 61.0 -62.0 . . . . -15.58 1.1563 - 14.40 69.0 -70.0 . . . , -15.08 1.1495 - 14.20 77.5 - . . . . -14.88 1.1427 - 14.10112 FRANKLAND AND MACGREGOR : ETHEREAL SALTS Thus the introduction of the two phenylacetyl groups has an entirely different effect from the introduction of the two benzoyl groups, for whilst ruet>hylic dibenzoylglycerate is strongly dextro- rotatory, the diphenylacetylglycerate is levorotatory, in fact even more so than methylic diacetylglycernte. This profound difference in the rotatory effect produced by tlie beiizoyl and phenylacetyl- groups respectively will be again referred to later (see p. 119).Although the diphenylacetylglycerat e resembles the diace tyl- glycerate so closely in its rotation, there is one point in which it strikingly differs, and that is in the sensitiveness of t h e rotation to temperature; thus, from the above table it will be sem that tho lsvorotation becomes diminished with rise of temperature, whilst in the case of the diacctylglycerates, as we Iiave already shown (Trans., I 894, 65, 765), the laevorotation increases with increase of tempera- ture, the same being true of the glycerates themselves. The clibenzoyl- glycerates, again, have their dextzorotation diniinished with increase of temperature.These remarkable differences in beliaviour with regard to tempera- ture obviously indicaf e that the dissymmetry of the molecule which leads to the laevorotation of the glycerates and diacet~ylglycerates becomes exaggerated with rise of temperature, whilst the dissym- metry of the molecule which occasions the laevorotation of the diph en y lacetyl gl y cerat e diminishes with increase of temperature, arid similarly the dissymmetry which brings about the dcxtrorotatiou of the dibenzoylglycerates becomes moderated by rise of temperature. Methylic ~Ionotenzoylglycerate (Active). The active rnethylic gljcerate was mixed with the calculated quan- tity of benzoyl chloride in the cold.There was no visible action until the mixture was heated to 83", and it became vigorous a t 100' ; the oil-bath was finally raised to 180°, and maintained there for 20 minutes. The product, after being washed with sodium carbonate solution, and with water, was put into a vacuum desiccator, but crystallisation did riot ensue. Numerous unsuccessful attempts mere made to obtain crystals by using the most varied solvents, in conjunc- tion with a freezing mixture ; benzene, aldehyde, chloroform, toluene, aniline, ether, acetone, methglated spirit, propylic alcohol, carbon tetrachloride, and light petroleum were all emplo_red, to 110 purpose. In i i g h t yetrolenrn, the liquid was only very slightly soluble. The liquid was stibmitted to conrbust>ion, with the following results.0.2368 gave 0.5120 CO, and 0.1072 H20. Methylic monobenzoylglycerate, CllH1206, requires C = 58.93 ; H = 5-36 per cent. C = 58.97; H = 5.03.OF ACTIVE AND IXACTIT7E GLYCEBIC ACIDS. 1.13 The liqnid, which analysis had thus shown to be pure metbylic monobenzoylglycerate, was examined with the polayimeter. a, = + 5.95"; t = 13' ; d 1 3 ' / 4 O = 1.2655 ; 1 = 50 mm. It was then distilled under a reduced pressure of 10 mm., when This distillate, on polarimetric examination, gave the following about one-half of it passed over between 180' and 240'. results. U D = +7*13'; I = 50 mm.; t = 13'. Thus this distillate possessed a considerably higher dextrorotation than the original liquid from which it was obtained, clearly pointing, therefore, to the fact that the original liquid was a mixture of the two possible isomeric methylic monobenzoylglycerates, and from our subsequent experiences with the ethylic compouud, there can be little doubt tlhat the more volatile part is the a-modification with the higher dextrorotation, whilst the /3-compound with the higher boiling point, and which remained for the most part in the distilling flask, must have a lower dextrorotation, or even a 1Evorotatioa.Tbe quantity of material a t our disposal was not sufficient to enable u s to obtain either of these isomers iii a pure state, so that for the present we have had to be content with having indicated this probable relationship between their rotations, namely, that the nzethy2ic a-monobenzoylglycerate is dextrorotntory, but much less s9 t h a n metltylic dibenzoylglyctrate, whilst methy lic f3-monobenz~ylglycerate i s certainly less dextromtatory thaw the a- compound, and possibly even hvorotatory. (For another possible explanation of these phenomena see p.115.) dlet h y 1 ic $Ion obenzoy lglyce rate (51 active). This was prepared in the same way as the active compound, only using inactive methylic gljcerate. On drying the product, the ethereal solntion of which had been washed with sodium carbonate solution and water in the desiccator, it became pasty and turbid, i n consequence of the appemacce of small crystals, the quantity of which iucreased after prolonged standing ; a large proportion of t h e material, however, still remained as a viscid liquid. The ci-ystals were found to be very so!uble in alcohol, chloroform, and acetone, but recrystallieation from hot benzene prored most serviceable for their purification ; in this way the cr,ystals are obtained in small, nodular groups, exhibiting a radiated structure.The melting point of the purified crystals was 92.5-93-5'. 0.2431 gave 0.5213 CO, and.0.1170 H,O. C = 58.48 ; H = 5.35. Methylic monobenzoylglycerate, C,1H,20,, requires C = 58.93 ; H = 5-36 per cent.114 FRANKLAND AND MACGREGOR : ETHEREAL SALTS The product was again recrystallised from benzene, after which the melting point was found to be 92-5-93', and 0x1 combustion the fol- lowing results were obtained. Methylic monobenzoylglycerate, CllHI2O5, requires C = 58.93 ; H = 5.36 per cent. The substance was again recrystallised, and another combustion 0.1197 gave 0*3570 CO, and 0.0594 H,O.CllHlZO5 requires C = 58.93 ; €3 = 5.36 per cent. As already pointed out in the case of the active compound, there should be two mcthylic monobenzo~lglycerates, but whether the crystals thus obtained are the z- or the p-compound we have not yet directly determined ; from the fact, however, that the crystals appeared in the originally liquid product, and that a large portion of the laotter permanently remained in the liquid state, i t wonld appear probable that both the a- and the ,%compounds had been formed in the reaction, and that the one was crystallisable, and the other not. It may fur- ther be suggested that, by analogy, the crystallisable substance is probably the p-, and the unorptallisable the a-, compound; thus a-chloropropionic acid is ti liquid boiling a t 186', whilst /3-chloro propionic acid is a solid, the melting point of which is variously given as 35.5-41' by Krestovnikoff (J.Russ. Chenz. SOC., 11, 248): and 58" by Richter (Zeit. f. Chew,., 1868, 451) ; again a-chlorobutyric acid is a viscid liquid (Markovnikoff, Annulen, 1870, 153, 241), whilst P-chlorobutyric acid is a crystalline body, melting at 98 -- 99" (Mark- ovnikofi, Zeit. f. Chem., 1868, 621). Similarly a-bromopropionic acid is a liquid boiling a t 205.5", and solidifying a t -17" (Keknl6, Anmden, 1864, 130, IS), whilst p-bromopropionic acid is a solid, melting a t 61.5' (Richter, 2 e i t . f . Chem., 1868, 449). 0.2112 gave 0.4530 COz and 0.1039 H,O. C = 53-50 ; H = 5.47.made. C = 58.55 ; H = 5.51. Ethylic Il.loizob~tzoylgllJce,.ate (Actice). This was prepared in the same way as the two methylic mono- benzoylg1ycerates described above. After washing the crude product with sodium carbonate solution and water, it wa3 placed in the vacuum desiccator, in which, after six Reeks, it began t o crystallise, but the greater part remained liquid. The crystals were pressed between filter paper, and dissolved in hot, light petroleum, from which, on cooling, radiating needles were obtained, melting at 62". Still finer crystals were obtained from petroleum spirit (b. p. 80-117'). Thus in this case again it would appear that both the a- and the /3-compounds are formed in the reaction, and that one of these isON ACTIVE AND INACTIVE GLPCERIC ACIDS.113 crystallisable, and the other not, and for the reasons given above, the crystallisable one is in all probability the P-compound. The above crystals were submitted to combustion, with the follow- ing results. 0.2503 gave 0,5527 CO, and 0.1332 H20. C = 60.22 ; H = 5.91. 0.2293 ,, 0.5057 ,, ,, 0.1220 ,, C = 630.15; H = 5.91. ClzHlaOs requires C = 60.50 ; H = 5.88 per cent. The crystalline ethylic monobenzoylglycerate was examined polari- metrically in a stmate of fusion, at the following temperatures. Observed rotation, aD. Temp. 2 = Mmm. Density, t0/4'. [.ID- 67.0". ..... -4.98 1.1547 - 9.80" 78.5 ...... -4.88 1.1438 - 9-70 88.3 ...... -4.88 2.1344 -9.77 Thus the solid ethylic monobenzoylglycerat,e, which is presumably the P-compound, has a strong lzevorotation, which, moreover, remains practically unaltered by change of temperature, and thus difters markedly from all the disubstituted ethereal salts of glyceric acid which we have examined.The liquid portion of the crude ethereal salt, from which the above solid ethylic P-monobenzoylglycerate had crystallised, was examined in the polarimeter, and was found to possess a dextrorotation. This is, therefore, as in the case of the methylic monohenzoylglycerate, con- sistent witb the supposition that there were two et hylic monobenzoyl- glycerates formed, of which the solid one (doubtless the /+-compound) was obtained in a pure state, and was found to be levorotatory, whilst the liquid one (doubtless the a-compound) is dextrorotatory. We would, however, point out thaii the facts also admit of another explanation, namely, that whilst the solid ethylic monobenzoyl- glycerate possesses the IEvorotation given above, the dextrorotation of the liquid frnm which it separated may be due to ethylic dibenzoyl- glycerate.A similar explanation would also fit the facts in the case of the methylic compound. We cannot positirely decide between these two alternative hypotheses until we have prepared these com- pounds on a larger scale. Particularly interesting with regard t o the solid ethylic moLo- benzoylglycerate are the circumstances (1) that its specific rotation is almost identical with t h a t of ethylic: glycerate itself (for a fuller discussion of this point see p. lZO), and (2) that the specific rotation is extremely insensitive to temperature.This insensitiveness is evident from the figures given above, but we have submitted it to a still more severe test by determining the specific rotation" at 136-13'7" * The ethylic monobenzoylglycerate used in this experiment was not quite pure.11 6 FRANKLAND AND NACGIREGOR : ETHEREAL SALTS (using the vapour of xylene in the jacket-tube of the polarimeter), with the following result. ED = - 4-22' ; t = 136.5' ; Z = 44 mm ; d 136*5'/4' = 1.0886 ; As pointed out in the footnote below, the substance used for this experiment was not quite pure, its rotation at 71" being slightly below what it should be ; on increasing the specific rotation at 136.53 in the proportion 9.07 : 9.75 :: 8-81 = 947, we obtain the corrected [ a ] D = - 9-47' at 136.5', or almost exactly the same figure as for the lower temperatures at which the observa- tions were previously made.The specific rotation of this ethylic rnonobenzoylglycerate is thus almost wholly independent of tempera- ture, our experiments showing that it suffers only the most trivial diminution in value between 67' and 136.5'. The relationship between ethylic glycerate and ethylic monobenzoyl- glycerate is thus a very remarkable one, for whilst the molecular dissymmetry of the latter remains pracbically constant at all tem- peratures, the molecular dissymmetry of the former (leading to lavo- rotation) increases with the temperature, and at 15' its molecular dissymmetry (as measured by specific rotation) happens to be equal to the uniform molecular dissymmetry of the ethylic monobenzoyl- glycerate molecule.Thus we have formerly (Trans., 1894, 65, 769) shown that the specific rotation (lavo) of ethylic glycerat'e increases by 0.035' for 1' rise in temperature, so that assuming this increase to proceed uniformly, at 136.5', the specific rotation of ethylic glycerate would be 13-19', and therefore greatly in excess of that of ethylic monobenzoylglycerate, The exact temperature, in fact, at which coincidence between their specific rotations should take place is 33.8'. These relations strikingly indicate t h e importance of taking into account temperature in conrection with the comparison of the specific rotations of different active substances. Methylic dip yopion ylgzycerate (Active). The method of preparation was the usual one of running the plycerate into excess of the acid chloride, but a difficulty arose in consequence of the propionyl chloride containiug phospliorus com- having suffered slight decomposition through a distillation to which it had been submitted, as was shown by its rotation a t 71" being aD = -4.59", 1 = 44 mm., the 5gure previous13 found with the pure substance.OF ACTIVE AND INACTIVE GLYCERTC ACIDS.117 pounds, which it nppears to be practically impossible to remove, owing to the proximity of their boiling points (propionyl chloride b.p. 80', PCI, b.p. 76", POCI, b.p. 107"). On examining the crude product, i t was found to be slightly dextrorotatory and to fume in con- tact with air ; on refractionating, the lower distillate, which fumed, showed a high dextrorotation, whilst the higher fraction had a laevo- iotation.As it was found impossible to entirely remove the fuming liquid by fractionation, the laevorotatory compound was washed successively with a solution of sodium carbonate and with water, then dried in a vacuum desiccator, and subsequently distilled ; the distillate thus obtained did not fume, and gave the following polarimetric result. CCD = -11.5"; I = 99.2 mm. at 148". This liquid was again distilled, but the amouut of distillate obtained was so small that the rotation could not be ascertained with t.he requisite accuracy, but it was approximately aD = -11.6-11*9" ( I = 99.2 mm.). The dextrorotation of the fuming liquid referred t o above was doubtless due to the presence of chlorine derivatives of glyceric acid formed by the action of the phosphorous compounds in the propionyl chloride used.A second preparation again yielded a crude product which fumed in the air, and this instead of being distilled was washed with a solution of sodium carbonate as described above. By repeated fractionation of t h i s washed product, a liquid was obtained having a rotation aD = -12.2" (I = 99.2 mm.) a t 14.2'. this was again fractionated, and the distillate then gave substantially the same rotation as before, thus a, = -6.23" (I = 50 mm.) a t 1.5*2=. The rotation was found to be considerably influenced by tempera,- ture, and, as in the case of the diacetylglycerates, the rotation increases with rise of temperature. The specific rotation m'ty be calculated from the above figures, and the density, which was found to be d 15'/4" 1.1349, thus This rotation is slightly inferior to that of methylic diacetyl- glycerate, which we have previously shown (Trans., 1893,63, 1430) to be [a], = -12.04".It is to be anticipated, therefore, that the di- butyr~lglycerates will exhibit a further diminution in levorotatioii and so on for the series of fatty acid radicles. Although the fact that the rotation obtained was essentially the same for the two118 FRANKLAND AND MACQREGOR : ETHEREAL SALTS separately prepared specimens of methylic dipropionylglycerate points to the rotation being correct, still we give the figure with some reserve in consequence of the phosphorus compounds present in tho propionyl chloride employed having conceivably interfered with the optical purity of the product. On combustion the following resnlts were obtained.0,2912 gave 0.5478 GOz, and 0.1818 H,O ; C = 51.30 ; Hz6.94. Again for hydrolysis with alcoholic potash. 11. 0.9426 ,, ,, 0.6798 ,, ,, = 72.22 ,, Calculated for methylic dipropionylglycerate = 72.41 ,, CIoH,,O6 reqnires C = 51.72; H = 6-90 per cent. I. 1.0506 gram required 0.7566 gram KOH = 72.01 per cent. Xunzrnary and Conclusions. (1.) Whilst the ethereal glycerates are laworotatory, and the diacetylglycerates more l~vorotatory still, the dibenzoylglycerates are powerfully dextro-rotatory, thus (1) H\ /CH2*OH (31, C Met.hylic glycerate [a]= = -4 'SO". (59) COOCH,/ \OH (17) (1) H\ /CHC*OH (31) C Propylic glycerate. i87) COOC,H/ '\OH (17) [alD =: - 12-94".(1) H\ /CH2*0C2H30 ($3) C (73) COOC2H,/ \OC,H,O (59) Ethylic diacetylglgcerate. [aJD = -16 -31'. (1) H, ,CHyOC;H,O (135) C (59) COOCH,/ \OC;H,O (121) Methvlic dibenzoylglycerate. [a]; = + 26 '89" (at 15"). [MID = + 88 5?0". ( 1 ) H\C/CH2*OH (31) ($3) COOC&.' 'OH (17) E thylic glycerat e [alD = -9*1S0. (1) H\ /CH2*0C2H30 (73) C (59) COOCH~/ \OC,H,O (59) Meth ylic diacetylglycerate, [.ID = -12'04*. C ( 8 ; ) COOC3H7/ \OC2H30 (59) (1) H\ /CH2*OC2H@ (73) Prop-ylic diacetyIglycerate. [a]= = -19 %7". (1) H, ,CH,.OC7H,O (135) C (73) COOC,H,/ \OC:H,O (121) Ethylic dibenzoylglycerate. [aID = + 26 '58" (at 15'). [MID = + 90 *go'. (1) CH2*OC7H,O (135) (87) COOC,H/ \OC;H,O (121) Propylic dibenzoylglycerate. [a]D = + 21 *ooo (at 15'). [MID = -k '74 *76". (2.) We have already shown (Trans., '1893, 63, 1415, and 1894,OF ACTlVE AND INACTIVE CILYCERIC ACIDS.119 65,754) that' in the series of the glycerates and the diacetylglycerates respectively, the rotations rise with the positive radicle, attaining in each series a maximum at the isobutyl compound. In the series of the dibenzoylglycerates, on the other hand, the rotations diminish from the methylic salts onwards, the rotations of the methylic and ethylic salts being almost identical, but considerably greater than the normal propylic. As the dextrorotation is conditioned by the benzoyl groups, whilst the positive radicles alone condition a negative rotation of the molecule, we should anticipate that by increasing the magnitude of the positive radicle the tendency towards a negative rotation would be increased, or in other words that ihe positive rotation would be diminished, and this is actually found to be the case.It would, however, have been anticipated thatl this diminution in posi- tive rotation should have proceeded more regularly, and that the positive rotation of ethylic dibenzoylglycerate should have been coil- siderably inferior to that of the methylic compound. If this line of argument be correct it is further to be anticipated that this diminution in the positive rotation of the dibenzoylglyce- rates will continue until the isobutylic compound is reached, beyond which the positive rotation should again increase. This point we have not yet had time t o determine. (3.) Whilst the presence of the two benzoyl groups thus condi- tions a positive rotation, the result is entirely otherwise if two phenacetyl groups be introduced iastead.Thus we have found methylic diphenacetylglycernte to be laevorotatory (1) H\ /CH,*OCJ3;O (149) C (59) COOCH,/ \OC,H,O (135) Methylic diphenacetylglycerate. [uJD = -16*06" (at 14'5'). In fact this compound has almost exactly the same rotation as ethylic diacetylglycerate, aiid a little higher rotation than that of methylic diacetylglycerate. This is a most striking illustration of what we have before called attention to (Trans., 1893, 63, 535), that the rotation is more power- fully influenced by the qualitative character than by the mere mass of the groups attached to the asymmetric carbon atom. (4.) We have also prepared some of the monobenzoylglycerate~.Of each of these, there should, obviously, always be two isomeric: modifications, according as the benzoyl gronp enters the glyceric acid molecule in the a- or in the P-positim. These compounds were obtained by mixing the calculated quantities of benzoyl chloride and [MI* = -57 ' 1 ' 7 O .120 FRANRLAND AND MBCGREGOR : ETHEREAL SALTS ethereal salt of glyceric acid and t,hen heating themixture as long as hydrogen chloride was evolved ; both isotners appear in every case to be formed. I n the case of methylic monobenzoylglycera~~e (active), both isomers being liquid, we have not yet been able to separate them perfectly from each other. We have, however, shown that the lower boiling isomer, which is in all probability the a-compound, has a greater dextrorotation than the higher boiling and, presumably, ,&compound ; i n fact the latter may even be l&vorotatory.A solid methylic monobenzoylglycerate (inactive) was obtained which melted at 92-5-93'. In the case of ethjlic monobenzoyl- glycerate (active) the higher boiling, and, therefore, presumably p-compound is solid (m. p. 6 2 O ) , and has been purified by repeated crystallisation. It is lsvorotatorg, and has almost exactly the same rotation as ethylic glycerate itself. Methylic a-monobenzoylglycerate. Dextrorotatory liquid. Met hylic /3- monoben zoy lgly cerat e. Less dextrorotatory liquid than the a-compound. (1) H\C/CH2*oH (31) (1) H\ /CH2*OCjH,O (135) C ( J 3 ) COOC,H,/ \OC;H@ (121) (73) COOC~H,,' \OH (I7) Ethylic a-monobenzoylglycerate. Ethylic 8-monobenzoylglycerate.Dextrorotatory liquid. [a]D = -9.80" at 67", but insensitive to temperature, hence probably the same a t 15'. (Melting point SZ".) [MID = - 23.32". The almost exact equality between the specific rotations of ethylic glycerate (;a], = - 9 . 1 8 O ) and of ethylic P-monobenzoylglycerate ([aIn = -9.80') is extremely remarkable, for it would surely be anticipated that the displacement of a single atom of hydrogen by such au enormous group as the benzoyl radicle should produce a pro- found change in the dissymmetry of the molecule as measured by' specific rotation. This replacement of hydrogen by benzoyl, wit'hout much effect on the rotation, is only realised in the case of the CH,*OH-group of ethylic glycerate, for, by replacing tlie H in ths OH-group by C,H,O, the product is dextrorotatory (see ethylic a-monobenzoylglycemte) , and, therefore, entirely different, even in, the sign of its rotation, from the original ethylic glycerate. (5.) The remarkable phenomenon of replacement without marked change in rotation is also exhibited in the case of the methylic diphenyl- acetylglycerate referred to in (3), for, as there pointed out, the rota- tion of this compound ([RID = -16.06' a t 14.5O) differs but slightly from that of methylie diacetylglycerate ( [ a ] , = -12.04'), althougliOF ACTIVE AKD INACTIVE GLYCERIC ACIDS.121 the constitutJona1 change is of such a profound character as the dis- placement of two hydrogen atoms by two C6H15-groups. (6.) Methylic dipropionylglycerate ([a]D = -10*97"), again, ex- hibited only a very slight difference in rotation from methylic diacetyl- glycerate ([a]= = -12*04"), although the constitutional change is considerable, consisting as it does in the replacement of two atoms of hydrogen by two CH3-groups.(7.) It is worthy of remark, that in all these cases in which snbsti- tution is attended with comparatively little change in rotatory power, the substihtion takes place at a point which is comparatively remote from the asymmetric carbon atom ; whilst in those cases in which the substitution is so near the asymmetric carbon atom that it is only separated froin it by a single atom of oxygen, the change in rotatory power is very considerable. Thus, turning for illustrations of this principle t o the experimental material which we have ourselves furnished in this and previous communications, we find i3) (9) Paasiige from glycerates to dibenzoylglycerates causes great change in specific Passage from glycerates to diacetylglycerates causes pest change in specific Passage from glycerates to diphenylacetylglycerates cauees great change in Passage from lactates to benzoyllactates causes great change in specific rotation.,, glycerate to a-monobenzoylglycerate causes great change in Passage from methylic glycerate to ethylic glycerate causes great change in Passage from ethylic glycerate to propylic glycerate causes smaller change in Passage from propylic glycerate to isobutylic glycerate causes still smaller Passage from propylic glycerate to normal butylic glycerate causes practically Passage from normal butylic glycerete to heptylic glycerate causes very little Passage from heptylic glycerate to octylic glpcerate causes very little change Passage from metliylic dibenzoylglycerste to ethylic dibenzoylglycerate causes Passage from ethylic glycerate to ethylic B-nionabenzoylglycerate causes prac- Passage from methylic diacetj-lglycerate to methylic dip%enylacetjlglyceratc rotation.rotation. specific rotation. >> ,, acetyllactates >, 1 ) ,, specific rotation. specific rotation. specific rotation. change in specific rotation. no change in specific rotation. change in specific rotation. in specific mtation. practically no change in specfic rotation. tically no change in specific rotation. causes very little change in specific I &tion. An exception to the above rule is afforded by the specific rotations of the three dibenzoylglycerates which we have prepared, and in which we found that passage from the methylic to the ethylic com- VOL.LXIX. K122 ETEEREAL SALTS OF ACTIVE, ETC., GLPCERIC ACIDS. pound was attended with less effect on the rotaiion than passage from ethylic to propylic, although the latter substitution is more remote from the asymmetric carbon atom. Again, the same phenomenon is exhibited by the extensive mate- rial which has been furnished by Witlden (2eit.physikaZ. Chem., 1895, 17, 264) in connection with the derivatives of malic acid. Thus, bearing in mind the arrangement of the groups around the assym- metric carbon atom, SOOH p z , H C-OH -9 COOK the following results will be seen to show that replacement of the carboxylic hydrogen atoms produces ti considerable effect on the rotation ; but a much greater effect is obtained by replacing the hydroxylic hydrogen ; whilst after this latter hydrogen has been re- placed, all further substitutions carried out in the substituting group itself are almost entirely without influence on the rotation.Dimethylic malate.. . . . . . . . . Diethylic ,, . . .. .. . . .. Dipropylic ,) . . . . . . .. . . Diisopropglic ,, .... . .. .. niisobutylic , , . . . . . . . . . . Diamylic , , . . , . . . . . . . Dicaprylic , , . . . . . . . . . . C 4 D . - 6.85" - 10 -18 - 11 *62 - 10 '41 (about) - 11 *14 - 9-92 - 6-92 (about) Dimethylic acetylmalate. . . . . ,, propionylmalate . ), butyrylmalate . . . ,, isobutyrylmalate .,, isovalerylmalate . ,) chloracetylmalate ,, bromacetylmalate Dietliylic acetylmalate . . . . . , ,, propionylmslate . . . :, butyrylmalate.. . . . ,, isobutyi*ylmalate. . . ,, isovalerylmalate . . ., ,, bromecetylmalste. . , , brompropionylma- late.. . . . . . . . . . . , , brom butyry lmalate ,, bromisobutyrylmu- late.. . . . . . , . . . . -22.92 - 22 9 4 - 22 -22.36 - 22 -39 - 23 '30 - 22 '40 - 22 5 2 - 22 -20 -22.28 - 21 a99 - 22 *07 - 22 '48 -22.48 - 28 9 6 -22-57 Dipropplic acetylmalate . . . . ,) chloracetylmalate ,, butyrylmalate. . . . ,, isovalerylmalate.. ,, bromacetylmalate. Diisobutylic acetylmalate . . ,, butyrylmalate.. ,, isovalerylmalate ,, bromacetylma- late - - . . . . , , ialD. - 22 '85O - 23 -52 -22 .a - 21 '68 - 22 '24 -21 '88 - 21 '68 - 19 -91 - 20 '38 Maldiamide . . . . . . . . . . . . . . Malic acid . . . . . . . . . . . . . . . . Maldianilide . . . . . . . . . . . . . . Maldi-o-toluide . . . . . . . . . . . . Maldi-ptoluide.. . . . . . . . . . . Mal-B-nttghtbyl . . . . . . . . . . - 38 '0" - 5 * 75" - 60.0" - 66 -5" - 70 '0" - 51 '5" Chlorosuccinic acid . . . . . . . . 9 9 anhydride.. . Y, chloride . . . . Dime thylic chlorosuccinate. . Diethy lie .. Dipropy lic >, .. Diisobutylic ,, .. 9 , 'Diamy lic Y J .. -I. 52 '0' + 31 '0" (HboUt) + 29 -53 -I- 41 '42 + 27.50 t-25.63 + 21 '57 (about) + 21 -56 The rotation of these compounds was taken in solution.+ 27 -I 18 + 9 0 II 6 e J 1 -9 - 18 INFLUENCE OF TEMPERATURE ON THE SPECIFIC ROTATION OF COMPOUNDS DESCRIBED. --T-ROTATION OF ACTIVE COMPOUNDS I N ORGANIC SOLVENTS. 123 Dimethylic bromosuccinate.. + 51 -18 Dipropglic bromosuccinate.. + 38 '05 Dietbylic 11 . . -I- 40.96 1 Diisobutylic ,, .. +23*56 Thus in the above series such a change as that from dimethy'lic acetylmalate to dimethylic chloracetylmalstte is essentially similar to the change from methylic diacetylglycerate to methylie diphenyl- acetylglycerate; and in both cases the effect on rotation is remark- ably small. In malic acid, the hydroxyl group being directly attached to the asymmetric carbon atom, corresponds to the a-hydroxyl group in glyceric acid, and replacement OP this hydrogen in both case:: pro- duces a profound change in the rotation. Still rnore pyofound is the change in the rotation if the whole of the OH-group is substituted ; as, for example, in the passage from dimethylic malate to dimethylic chlorosuccinate. Again, i t appears that the replacement of the hydroxylic hydrogen by a hydrocarbon radicle produces far more effect on tfhe rotation than the substitution of the same atom of hydrogen by an acid radicle containing the same number of carbon atoms ; thus a com- parison map be made between 7 H2.C 0 OCH, Hg;:ECH3 H q O * C2H3G SH2* CO 0 CH, H - c OH bOOCH, COOCH, Q COOCH, [.ID = - 6 ' 8 5 O [uID = f60.9B0* [ u ] D = -22'92" ( Walden). (Purdie and W~lliamson) (Walden). Trans., 1895, 979.