ROTATION OF OPTICALLY ACTIVE COMPOUNDS. I. 167 XVTII.--The Influence o f Solvents on the Rotation of Optically Active Compounds. I. Influence of Wates., Methyl Alcohol, Ethyl Alcohol, n-Propyl Alcohol, and Glycerol on the Rotation of Ethyl Tartrate. By T. S. PATTERSON. ALTHOUGH a t the present time the results of a considerable number of investigations relative to the influence of solvents on the rotation of optically active compounds are available, they seem insufficient t o allow of the deduction of any satisfactory generalisations. A few theories have been suggested attributing the phenomena observed, for instance, to difference of solubility of the active substance in the various solvents used or to the formation of chemical combinations of active substance and solution, but no one of these can be said to have met with much success.The work already done appears to be neither sufficiently extensive nor sufficiently systematic. Only in the case of the application of the electrolytic dissociation theory to explain the simi- larity of the rotations of dilute aqueous solutions of different salts of the same optically active acid or base, do facts and theory correspond, and this, however satisfactory the correspondence may be, is merely a verification of the dissociation theory; it does not help us at all to understand the general effect of solvents on the rotation of optically active substances. For, supposing that the active ion could be obtained free in two different solvents, it would in all probability not have the same rotation in each.Different solvents would presumably have different effects on active ions just as they have on undissociatedcom- pounds, and it seems also impossible to explain the behaviour of solu- tions of optically active substances by assuming varying degrees of electrolytic dissociation in each. This property of causing electro- lytic dissociation to any considerable extent is only possessed by a few liquids ; in the rest, the physical molecular weight of the solute is the N Z168 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE same as the chemical molecular weight, or it is greater, The question whether the existence of the solute as associated, simple, or dissoci- ated molecules in solution merely depends on some one property of solvents which is developed to a different degree in each, does not yet seem to admit of answer.However that may be, the theory which asserts that variation of rotation and varying degree of association are causally connected with each other has received a certain amount of credence. I n the case of homogeneous compounds, P. F. Frankland (Trans., 1899, 75, 347) has been able to trace a very interesting connection between the abnormal (that is, B priori unexpected) rotations often shown by the lower mem- bers of homologous series of active compounds and their association factor as calculated by I. Traube's method, and several investiga- tions have been undertaken with the object of ascertaining whether a direct connection could be proved between the rotation and associa- tion factors of an optically active substance in solution.The evidence which has thus been collected is unsatisfactory ; i t perhaps does not abso- lutely disprove the possibility of the connection sought for, but i t leaves plenty of room for doubt as to its probability. The whole question of rotation in solution is one still requiring much careful investigation, and it is the object of this paper, as it will be OF succeeding ones, to add at least something to the data on the subject. As the possible connec- tion of association and rotation arises naturally in discussing the ex- periments detailed below, the short criticism of existing work which is necessary is introduced further on (p. 184). Ethyl tartrate has been chosen for investigation for the present, since it can be obtained pure without much difficulty, has a sufficiently high rotation to allow of fairly accurate measurement, and is miscible in all proportions with many organic solvents, so that it is possible t o draw, complete concentration curves, and, finally, has a simple con- stitution which is fairly well understood.The ethyl tartrate used in this investigation was prepared by boiling tartaric acid with ethyl alcohol (in the proportion of 1 molecule to 4) for some hours. The solution was then cooled and saturated with hydrogen chloride at a low temperature. After an interval of about 12 hours, the gas was removed as f a r as possible in a vacuum and further expelled, along with excess of alcohol, by heating on a water-bath at a low pressure. A weight of alcohol equal t o that previously used was then added and the solution again saturated with hydrogen chloride, which was driven off as before.The residual ethyl tartrate was carefully fractionated several times under about 15 mm. pressure until two successive fractions had identical rotations. The observed rotation of the product thus prepared was, from the figures given on p. 198, + 9.244' in II 100 mm. tube, and the specific rota-ROTATION OF OPTICALLY ACTIVE COMPOUNDS. I. 169 tion, therefore, +7.67". This value agrees well with that found by other investigators who have prepared the est-er in the same manner, as the following figures show : t. UD Perkin" (Trans., 1887, 51, 363) ........................ 20 9-65 Frankland and Patterson (Trans., 1898,73, 188) ... 20 9.31 Frankland and McCrae (Trans., 1898, 73, 310) ......20 9.323 Pictet (Jahresber., 1882, 356) ........................... 20' +9*236O Rodger and Brame (Trans., 1898, '73, 304) ............ 20.1 9.37 The optical behaviour of ethyl tartrate in water, methyl alcohol and ethyl alcohol has already been studied by Landolt (Annslen, 1877, 189, 311), who, however, in his experiments used an ester having a specific rotation of + 8.31' at Z O O and a density 1.1989. That these values do not agree with those of other investigators he assumes to be due to admixture with some inactive impurity, probably alcohol. The following investigation is somewhat more extensive than his, especially with regard to temperature, it being possible from the figures and curves given t o deduce the rotation of ethyl tartrate in each of the solvents dealt with, at any temperature within the limits of the experiments and at! any concentration whatever, with fair accuracy.The instrument used was a Laurent half-shadow polarimeter reading direct to one minute. I t is fitted with a large jacket which for use at higher temperatures is filled with hot water and allowed to cool down slowly, observations being made at intervals as the desired temperatures are reached. For readings at loo", steam is passed through the jacket until the rotation is constant. The liquids t o be examined are contained in tubes differing only slightly from the ordinary form. Fig. 1 shows a t A and B two different forms for the ends. I n A , a short piece of indiarubber tubing is drawn past the end of the tube so as to project a little beyond the disc, over mhich it folds t o some extent.It is then squeezed against the disc by the brass ring, which is screwed to a corresponding metal collar fitting against the glass flange a t the end of the tube. B shows the end of a tube of different form, which was made by Messrs. Schmidt and Haensch, Berlin. The end is flanged as before, but is then ground out so that the disc fits closely into position and flush with the face of the tube. A rubber washer, which has only a very small annular space to cover, is held in position by brass end * Perkin, in the course of his preparation, dissolved the ester in ether, and treated the solution with anhydrous potassiuni carbonate, which seems to have effected some slight purification beyond that accomplished by simple distillation.170 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE pieces.This is a very suitable form but much more troublesome to make than the other. The tubes are also fitted with a lateral tube, C , for filling and to allow of expansion of the liquid on heating. This is constricted a t the distal end, which is ground flat so as to be about 2 inch in dia- meter. When the tube has been filled so that no air bubble can be seen on looking through it, the open end of the side piece is closed with a small microscope cover-glass held in position by a rubber band. A small rubber balloon is then slipped over the side piece. This arrangement is found to prevent evaporation very satisfactorily, the density of the solution and the rotation, as will be observed from the figures given, being, in most cases, almost identical before and after the observations at higher temperatures had been made, even when the heating had been carried up to within ten or fifteen degrees of the boiling point of the solvent.FIG. 1. The rotations are recorded in the order in which they were observed, so that a comparison of the first and last figures will show to what extent evaporation of the solution has affected the experiments. The densities, which are all relative to that of water a t 4O, were determined by means of Ostwald pyknometers of about 8 C.C. capacity. The error of a determination is probably not more than about three or four units in the fourth place. Percentage composition throughout this paper is understood to mean grams of active substance per 100 grams of solution.The expression ‘‘ after experiment ” attached to some of the density data means that the determination was made with the solution taken from the polarimeter tube after the rotation had been observed. The observed rotations are generally given to three places of deci- mals, since two successive determinations, each the mean of ten settings of the instrument, agree usually to within a few units in the thirdROTATION OF OPTICALLY ACTIVE COMPOUNDS. I, 171 place. The specific rotation is, however, only given to two places of decimals, and in many cases the probable error of a figure for specific rotation may be fairly large, although seldom more than 0*1', even with dilute solutions. Nearly all the experimental data in this paper are represented by the curves shown in the different figures, and since these give a corn- prehensive idea of the results very much more quickly and clearly than the numbers themselves, the latter have been collected and placed at the end of the paper, where reference can easily be made to them when necessary.The rotation of ethyl tartrate is known to be very sensitive to change of temperature, but although observations have been recorded a t 100' and at 20' by various observers and between 12' and 20' by Perkin (Trans., 1887, 51, 368), the rotation does not appear to have been examined at temperatures between 20' and looo. The rotation of pure ethyl tartrate a t various temperatures was therefore first determined. The figures obtained will be found on p.198 and the curve obtained from them is shown in Figs. 2,3,4, 5, and 6 (the fiducial points being only introduced, however, in Fig. 2), in order that the rotation of the free substance and its solutions may be easily compared. EthyE Turtrute in Water. Solutions of percentage composition 1, 2.5, 4.999, 9.994, 24.954, 49.993, 74.99, were made up and examined a t various temperatures. The figures obtained will be found onpp. 199-201, and the results are plotted in Fig. 2 as curves which show clearly the behaviour of the rotations of aqueous solutions of ethyl tartrate with regard both to temperature and to concentration, the former being, however, the more obvious. It will be noticed that the curves show a slight amount of irregularity, which is due to the fact that the solutions were heated up first to a comparatively low temperature and the rotation observed, then allowed to cool, and another observation made in order to see whether the heating had produced any permanent effect on the rota- tion of the solution.The heating was then again carried to a higher temperature than before, another observation made, the solution allowed to cool again, and so on. The results show that ethyl tartrate is rather more stable in aqueous solution than might have been expected. A 50 per cent. solution may be hested up to 75', and kept a t that temperature for some time without showing, on cooling, any noticeable change in rotation, and this applies to other solutions as well. Only a very slight decrease is noticeable in the rotation of a 5 per cent.solution even after standing172 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE for fifteen days, although a 10 per cent. solution showed a greater change after standing a shorter time, It is apparent from the curves shown in Fig. 2 that the specific rotation of ethyl tartrate is profoundly modified by solution in water, as was indeed already known from the experiments of Landolt, and from the evidence adduced later on (p. 181) there seems no reason whatever to ascribe this to chemical action of the solvent on the ester. FIG. 2.-Speci$c rotation of aqueous solutions of ethyl tartrate. + 27" + 25 + 23 + 21 +19 2 +17 R 5 +15 + 13 3 u & + 11 + 9 I + 7 * + 5 10" 20" 30" 40" 50" 60" 70" 80" 96" 100" Temperature.The curve for a solution of p = 75 lies considerably above that for the free ester, the specific rotation of the solution a t all temperatures for which the curve holds is greater than that of the pure tartrate, and a t the same time, i t should be carefully noticed, the influence of tempera- ture upon the rotation becomes less marked ; the curve for the solution is flatter than that for the ester. Both these influences are more evident in a p = 50 solution ; the specific rotation is again increased,ROTATION OF OPTICALLY ACTIVE COMPOUNDS. I. 173 the observed rotation for this solution being actually greater than that caused by the same tube filled with pure ethyl tartrate, and the specific rotation of the solution instead of increasing with increase of tempera- ture, exhibits just the opposite behaviour, diminishing slightly.The specific rotation gradually increases with further dilution up to a com- position of p = 10, when the maximum influence of the solvent appears to be reached, since the value of [a]: is almost identical for solutions of p = 10, 5, 2.5, and 1. A 10 per cent. solution may, apparently, SO far as rotation is concerned, he regarded as one infinitely dilute, and this maximum influence which the water can exert on the ethyl tar- trate molecules dissolved in it, not only increases the rotation of the latter t o between three and four times its original value, but pro- duces such a condition in them that increase of temperature causes a diminution in their rotation t o nearly the same extent that it produces a rise in the rotation of the free ester, the latter effect being quite a s remarkable and interesting as the former.Landolt's experiments seem t o show that the rotation of a solution of ethyl tartrate in water is a linear function of the concentration. This, however, is not borne out by my results, as will be seen by reference to Fig. 8, where, amongst others, the concentration rotation curve for aqueous solutions is shown. The curve is not a straight line, even between p = 25 and p= 100, but is at first concave to the concen- tration axis and then convex to it, having a point of inflection at about p=50. Ethyl Tartrate in Methyl AZcohol. The figures obtained in the examination of solutions of ethyl tar- trate in methyl alcohol for which p = 5,10, 25.01, 56, and 75 will be found on pp.202-204. They are represented graphically in Fig. 3,which shows the variation in specific rotation with change of temperature. The general appearance of the curves is in all eases similar, but they possibly tend to approach one another somewhat a t higher tempera- tures, although only very slightly, and they preserve also practically the same form as that of the homogeneous ester a t any rate up to about 60'. On mixing ethyl tartrate, then, with methyl alcohol, the rotation of the dissolved molecules slowly increases, the influence of the alcohol seeming to reach its maximum when the dilution has been carried to about p = 10, since the curve for a p = 5 solution coincides almost exactly with that of one for which p = 10, though perhaps it is just a little higher." This variation of rotation with concentration can be seen from Fig.8. As the dilution increases the rotation also increases * The ciirve fororp=5 is not shown in the figure, since the fiducial points are apt t o be coufused with those for ap=lO solution.174 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE fo a maximum of about + 1 l o 2 5 O , the influence of methyl alcoh'ol being therefore much less than that of water, but in this case, what- ever change takes place in the molecule on solution, increase of tempera- fure has the same effect as on the free ester, the rotation increases, whilst in water it diminishes. Ethyl Tartmte in Ethyl Alcohol. The solutions examined in the cases of this solvent were of 5, 10.94, The experimental figures will be found 20, 40, and 60-01 per cent.FIG. 3.-Ethyl tartrate in, methyl alcohol. + 14" + 13 + 12 + 11 rz' -$ 2 3 +10 3 & + 9 u +8 + 7 + 6 I 10" 20" 30" 40" 50" 60" Temperature. on pp. 204-206, and are represented graphically in the accompanying diagram, Fig. 4. It is obvious from the curves obtained that the general behaviour of ethyl tartrate in ethyl alcohol is much the same as in methyl alcohol, although the latter has a markedly greater effect. The specific rotation of dilute solutions is slightly higher than that of the homogeneous substance, but as will be seen from the concentra- tion-rotation curve (Fig. S ) , the effect of increasing dilution in thisROTATION OF OPTICALLY ACTIVE COMPOUNDS. I. 175 and the previous case is different.Addition of methyl alcohol to ethyl tartrate causes at first a fairly rapid increase in rotation, which be- comes gradually less and less as the dilution increases. Addition of ethyl alcohol, however, causes little or no change in specific rotation until the percentage composition of the solution is about 60. Further addition of alcohol then causes increase of specik rotation, the rate of increase being greater the more dilute the solution becomes. The FIG. 4.-EthyE tartrate in. ethyl alcohol. + 15' I- 14 + 13 -t- 12 z4 s g & s 4-11 s i-10 @ + 9 + 8 +7 + 6 10" 20" 30" 40" 50" 60" 70 80" 90 l'emperattcre. curve obtained is therefore convex to the concentration axis, the maximum effect of the ethyl alcohol being only reached in infinitelr dilute solution.It may be noticed, too, from Fig, 4 that, although the effect of in- crease of temperature on the ethyl alcoholic solutions and on ethyl tartrate is much the same up to about 50°, it appears to be somewhat different at higher temperatures when the influence of increasing temperature has less effect on the pure ester than on its solutions.176 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE Ethyl Tartrate in n-Propyl Alcohol. The n-propyl alcohol which was used in this investigation, although bought as pure, proved to be laxorotatory, but only to so slight an extent that the influence of the impurity is probably completely eliminated by the introduction of a small correction in calculating the specific rotation from the data obtained (see p. 207).2.5, 5, 7.71, FIG. &-Ethyl tartrate in n-propyl alcohol. + 14' + 13 + 12 + 11 i .$ +10 E u u $ 9 & + 8 +7 +6 + 5 10 2G" 30" 40" 50" 60" 70" 80" 90" Temperature. 10, 17.5, 25, 37.51, 49.83, and '74.99 per cent. solutions were examined. The experimental numbers will be found on pp. 208-210, and are repre- sented graphica.lly by the curves in Fig. 5. The curves for four of these solutions are, however, omitted in the diagram, as their intro- duction only tended to cause confusion. The influence of the other solvents examined is t o increase the specific rotation of the dissolved ethyl tartrate, but this effect of the solvent becomes less in passing from water to methyl alcohol and ethylROTATION OF OPTICALLY ACTIVE COMPOUNDS. I. 177 alcohol, and in n-propyl alcohol becomes such that at low temperatures fhe rotation is depressed below that of the free ester.The specific rotation of a p = 75 solution is at 20" about, 1.l0 lower than that of pure ethyl tartrate, the addition of alcohol to give a, solution of p = 49.S3 diminishing the specific rotation still more. With further dilution, however, the specific rotation gradually rises again, the curve for p = 25 lying above those for the concentrations just mentioned, so that when p = 5, the specific rotation of the solution is almost as high FIG. 6.--Ei%yl tartrate in glycerol. 4- 14' + 13 + 11 + 10 + G + 5 Temperature. as that of the free ethyl tartrate. The specific rotation of a solution, however, never seems to quite equal that of the pure ester, and at 20" appears to be constant for any concentration less than p = 10, which is apparent from the curve marked "propyl alcohol" in Fig.8. No solution, therefore, of ethyl tartrate in propyl alcohol below 30' has a rotation as great as that of the pure ester; but it should be noticed that since the rate of variation of specific rotation with variation of temperature is distinctly greater in the solutions, and especially in the178 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE more dilute ones, than in the free ester, this does not hold at tempera- tures above about 36O. Another interesting fact to which attention may be drawn i s the existence, for any temperature within the limits of these experiments, of a concentration of minimum rotation. Only a few similar cases are known, and in this one the phenomenon is very distinctly marked.As will be seen from Fig. 8, the minimum value of the specific rotation (+ 6.4') occurs (at 20°) when p=57. For other tempera- tures, the concentration at which it occurs is different, being greater the higher the temperature. Ethyl TuTtruts in Glycerol. The curves in Fig 6 represent the results obtained in the examination of solutions containing 5 , 9.9, 23.45, 48-12, 69-93, and 89.98 per cent. of ethyl tartrate. The curves do not always pass exactly through the experimental points," the irregularity being due to the fact that these experiments were carried out before the others and with apparatus less suitable than was obtained later. It will be noticed in the first place that the specific rotation of ethyl tartrate may suffer considerable variation by mixture of the ester with glycerol, being either greater or less than, or equal to, that of the free ester, according t o the composition and temperature of the solution.The phenomena observed are somewhat similar to those due t o the action of n-propyl alcohol. The addition of about 10 per cent. of glycerol lowers the specific rotation of the ethyl tartrate at 20' by 1.5') another 20 per cent. of glycerol causes a further diminution, after which, however, a rapid increase in rotation takes place with increasing dilution, but the effect of glycerol is different in two respects from that of n-propyl alcohol. Firstly, the specific rotation of dilute solutions is much higher at low temperatures than that of the ester itself, and secondly, owing to the fact that the temperature coeffi- cient is less for dilute solutions than for the pure ester, the tendency is for the specific rotation of all solutions t o sink below that of the free active substance at higher temperatures.It is obvious that here again we meet with a very good example of solutions of minimum rotation, the phenomenon being perhaps more marked in this than in any case previously observed. A number of concentration-rotation curves for different; temperatures are shown in Fig 7, from which it will be seen that the occurrence of the minimum rotation is more pronounced at low temperatures than at higher ones. * This is particularly the case for a p = 5 solution at higher temperatures, and is more obvious in the concentration-rotation curves.The data for these solutions will be found on pp. 211-213.ROTATION OF OPTICALLY ACTIVE COMPOUNDS. x. 179 The difference between the maximum and minimum rotations at 10' is about 6*2O, whilst at 100' it is only 2*5O, so that with increasing temperature the concentration-rotation curve becomes flatter. The position of the minimum, too, seems to shift to a less concentration with rise of temperature. Thus at 10' the minimum specific rotation FIG, 7. --Ethyl tartrate in glycerol. Concentmtion-rotat$on curves. 10 20 30 40 50 60 70 80 90 100 Percentage composition of solution in gram of ethy,? tartrate per 100 grams of solution. is found at p = 70, a t 25' it lies at p = 65, at 50' a t p = 60, a t 75' about p = 55, and at 100' at p = 53.Discwss~oon of Results. Passing now to a general discussion of these results, since water, methyl alcohol, ethyl alcohol, and n-propyl alcohol form part of a homo- logous series, it is natural to seek, in the first place, for some effect on180 PATTERSON: THE INFLUENCE OF SOLVENTS ON TEIE the rotation of ethyl tartrate due to their influence and varying in a gradual manner from one solvent to another. It is only in dilute solution that the maximum influence of the solvent can be exerted, and FIG. 8.-Relationship of specdjic rotation and concentration, and of mo2ecular- solution-volume and Concentration at 20". 10 20 30 40 50 60 70 80 90 100 Concentration. in such-for which p = 10 or less-a gradual variation may be noticed : (1) 1% the value of the speciJc rotation, which decreases as the molecular weight of the solvent increases.ROTATION OF OPTICALLY ACTIVE COMPOUNDS.I. 181 (2j 17% the fawn of the concentration curves, which gradually changes from the concave for water and methyl alcohol to the slightly convex for ethyl alcohol-with the possibility of a concentration of minimum rotation-and the still more convex curve for n-propyl alcohol with a distinct minimum, and, if we include glycerol, to its still more marked minimum and convexity. This is apparent from the curves in Fig. 8. (3) In, the efect of increase of tenzperatuve upon corresponding solutions. In water, the coefficient is negative, in methyl alcohol it is positive, but perhaps a little less than for the pure ester, i n ethyl alcohol it is the same as for the pure ester and for m-propyl alcohol distinctly greater, as is apparent from an examination of Figs.2, 3, 4, and 5. These three regularities then in the behaviour of sollitions of ethyl tartrate in the above solvents may be distinguished ; it remains t o correlate, if possible, the variation in rotation with some other similarly variable physical property of the dissolved substance, the solvent or the solution. The results obtained may first be discussed with regard to existing ideas, the most important of which, as already stated, attributes varia- tion of rotation to varying degrees of association of the molecules of the active substance in the different solutions, The variations of the rotation of ethyl tartrate described above are so considerable that if they be really due to this cause one might expect to trace the con- nection experimentally, and it is noticeable that the rotations stand in the same order as the dissociating power of the solvents used.That the behaviour of aqueous solutions of ethyl tartrate is remark- able is obvious from the description of it on p. 172 ; it shows peculiari- ties which easily suggest the possibility of an exceptional character for such solutions, notably in the remarkably high specific rotation and in the influence of temperature change. It is possible that in aqueous solution ethyl tartrate suffers hydrolysis to ethyl hydrogen tartrate which also has a high rotation (for c=2*252 [aID =21.8 [Fayollat, Compt. rend., 1893, 11'7, 630]), but this is improbable on account of the fact that no permanent change seems to take place in the solutions on heating ; the rotation returns to its original value when the solution cools.The fact that the molecular rotation of a 10 per cent. aqueous solu- tion of ethyl tartrate is +53*82' whilst that of similar solutions of the neutral tartrates is about + 60' suggests the possibility of electro- lytic dissociation in the former as in the latter, and this, could it be proved to occur in ethyl tartrate solutions, might explain, not only their high rotation, but possibly also the anomalous influence of change of temperature as well. If the rotation of dilute aqueous solutions of ethyl tartrate and the neutral tartrates depends on the same thing, VOL. LXXIX. 0182 PATTERSON: THE INFLUEWE OF SOLVENTS ON THE Constant 18'6 0.2594 0.3304 0.5467 0.5940 0.8928 namely, the existence of the free tartaryl ion in solution, both ought to behave alike with regard to change of temperature, and this appears to afford a means of determining whether the two solutions are similarly constituted.I n order to apply this criterion, a p = 13.68 solution of Rochelle salt (equivalent to a p = 10 solution of ethyl tartrate) was made up and examined in the polarimeter at various temperatures. The experi- mental figures will be found on p. 214 and are represented by the curve in Fig 2 marked '(Rochelle salt," It will be seen that the specific rotation of the Rochelle salt, instead of diminishing, increases somewhat* with rise of temperature, the rate of increase, however, becoming less and less at higher temperatures, so that between 60' and 100" it is almost independent of temperature and it must be con- cluded that the conditions in the solutions of this salt and ethyl tartrate are not similar.There remains, however, the possibility of varying degrees of association of the ethyl tartrate in the different solvents used, so in order to ascertain if the complexity of the active substance varies sufficiently with the solvent to account for the rotations observed, the following molecular weight determinations were made. In water, the cryoscopic method was used with these results : 9.0578 11-2638 10 '29 06 10 '4460 9.7404 Calc. mol. wt. = 206. Weight of Weight of substance solvent taken, 1 used. Percentage composition of solution, Grams per 100 grams solution.2-78 2 -85 5 *05 5 -38 8 '40 I 0.265" 0.280 0.495 0,535 0.835 Mean ..... M. 201 194.8 199.5 197.2 204'2 199-3 The molecular weight determinations in methyl, ethyl, and a-propyl alcohol were made ebullioscopically. An apparatus similar (except that it was not graduated for volume measurement) to that recently described by H. N. McCoy (Amer. Chem. J., 1900, 23, 353) with a thermometer reading direct to twentieths of a degree was used. As a preliminary experiment, the molecular weight of thiocarbanilide was determined in methyl alcohol in which 231.7 instead of the calculated value, 228, was found. * This had already been shown by Hadrich.ROTATION OF OPTICALLY ACTIVE COMPOUNDS. I. 183 No. The following are the results : Weight of ethyl tar- trate.I I1 I11 IV v Weight of solvent. 1,1972 1.0654 1'1972 1'7844 i m 4 Percentage composition of solution. Grams sub- stauce per 100 grams solution. I I1 111 IV I I I 1,5225 33-58 ! 2'0650 32.72 ' 2.0650 1-5225 y; 1 Methyl alcohol. Constant 8-8 : I I1 111 IV V VI 1.3161 1.7504 1'6126 1'8143 1.5560 1.7504 33'24 24.37 25.18 25'02 17'16 28-544 22'750 19.808 22'060 17.064 16'060 3-48 4.17 4 *54 6 '66 9 '42 4.34 5'94 6.61 9 '57 4'41 7'14 7 -53 7 -60 8 -36 9 *83 At. 0.144" 0.175 0.168 0.300 0-385 Mean., . 0-245 0'360 0-365 0.555 Mean., , 0.401 0'630 0'660 0'605 0.650 0.755 Mean.. , Mol. wt. 22 0 -1 219'8 249.0 209.2 237.7 227.2 - 216.5 205.1 226.6 223'1 217'8 182.8 194 196.1 216 221 *6 229 -5 206.6 The distillates from the methyl alcohol determinations mere collected and examined in the polarimeter.No rotation could be detected, so that there was probably no volatilisation of ethyl tartrate. It would appear from the above results that, under the conditions of the experiments, ethyl tartrate exists in these solutions in simple molecules, but obviously any comparison of rotations must be made a t the same temperature; it would have little meaning t o compare the rotation of ethyl tartrate in water a t 0" with those in methyl, ethyl, and propyl alcohols at their respective boiling points, and there- fore the question arises whether we may assume that ethyl tartrate in these last three solvents exists in simple molecules at a temperature of, say, 20" as well as at the boiling points, This is, of course, diffi- cult to answer.I n some cases it does appear that the molecular weight of a given substance determined at the freezing point in a par- ficular solvent is greater than that obtained in the same solvent at 0 2184 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE I. Propyl dipropionyltartrate the boiling point, but if at all, is only slightly greater, and therefore the small amount of association which might be assumed to occur in the above cases a t the ordinary temperature scarcely seems to be a suffi- cient cause to which to attribute the great variation in rotation. The experimental evidence which has previously been collected on this subject is somewhat unsatisfactory. It is difficult to review it briefly, but a few figures may be given which seem to show the in- sufficiency of the hypothesis that variation of the rotation of an active substance in different solvents is due t o corresponding variation of association.Freundler (Ann. Chim. Phys., 1895, [vii], 4, 256) has found that, in a number of substances examined by him, when the molecular weight is normal, the rotation is the same or very similar in the free state and in solution, whilst in other cases where the molecular weight is not normal the rotation differs in solution from that of the free active compound. He gives, amongst others, the following examples : Ethylene bromide 346 ~ Active substance, 11. Propyl diacetyltartrate ... y , diphenacetyltartrate 111. Methyl tartrate .............. ,, ............... 2 9 >t Propyl , , ............... Y9 I .Calc.I Benzene Nitrobenzene Acetic acid Benzene Ethylene bromide 9 , 318 470 470 178 234 234 Founc 342 277 378 377 411 306 326 - lolution. +5.4" + 1.2 + 14.6 + 27 -2 - 8.8 + 20.1 - 0.6 Free. + 5 3 " +13'4 + 20.9 + 20-9 +2'14 + 12-44 + 1244 From these figures, Freundler has deduced two '' laws," against which, however, it is not difficult t o bring objections. It should be noticed, for instance, that propyl diphenacetyltartrate is dissociated t o exactly the same extent in nitrobenzene and acetic acid, but whilst in the former solvent the rotation is depressed, in the latter it is raised. Again, propyl tartrate is associated in benzene and in ethyl- ene bromide, the association being accompanied in the former solvent by an increase in rotation of about 60 per cent., and in the latter by a decrease in rotation of over 100 per cent.on that of the free sub- stance. Surely it may be expected that if these variations in mole- cular weight and rotation are connected with each other, the effects of association or dissociation should be, at any rate, consistent. It should be noticed that of the three possibilities, (1) simultaneous normality or abnormality of molecular weight and rotation ; (2) ab- normal molecular weight with normal rotation ; (3) normal molecularROTATION OF OPTICALLY ACTIVE COMPOUNDS. I. 185 weight with abnormal rotation ; the last is much the most important. Of these, (1) may indicate, but does not necessarily prove, a causal con- nection between variation of rotation and variation of association ; (2) does not necessarily disprove such a connection ; but (3) if it can be shown that any active substance has normal molecular weight in several different solvents and at the same time very different values for the specific rotation in these solvents, then unless some very good reason can be given to account for this unexpected behaviour, the association hypothesis must be considered disproved. The conditions numbered (3) seem to be met with in the case of ethyl tartrate detailed above, and the following figures taken from a paper by Frankland and Picknrd (Trans,, 1896, 69, 131) appear to furnish another example.Methyl dibenxoylglycerate [ a]1,5" + 26*899 M = 328. ~ ~~~~~ Molecular weight. I Percentage com- position of solution. Solvent : Nitrobenzene. 3 '9 5.3 6.7 7.8 Solvent : Acetic acid.2.0 3.4 5 '1 5.6 7 *7 16'2 327.1 317.8 315.8 322.5 327.9 304'9 306.4 327-6 324-9 323.7 Rotation. Percentage com- position of solution. 5.5 17.4 4.7 13-6 20.62 21'72 33.27 32-61 Thus the molecular weights are in both cases undoubtedly normal, Finally, a case cited by Walden (Zed. physikal. Chem., 1895, 17, whilst there is a very considerable difference in the rotations. 705) may be mentioned. Ethyl mandelate [ a ] , - 123 *lo. I n acetone. I n carbon diszclphide. Molecular weight normal Molecular weight normal [UID - goo, [a]D -180'.186 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE If, therefore, variation of rotation does really depend on variation of association, it will be necessary to attribute a greater influence t o an undetectable degree of association than even the constitution of the active substance itself can have. Of the chemists who seem inclined to adopt this hypothesis, Pope and Peachey (Trans,, 1899, 75, 1111) have perhaps pro- nounced themselves most strongly in its favour, but owing to the fact that they give no direct experimental evidence whatever in support of their views, judgment a s to the correctneEs of their con- clusions must be suspended.I n the paper in question, these authors seek t o base:a method of discriminating betweenracemicandnon-racemic liquids, which consists in ascertaining whether the value of the rotation of one of the forms of an optically active substance changes when dissolved in an inactive mixture of both forms. If the externally compensated substance be racemic, then, according t o these authors, the molecular condition of the active formwill alterwhen dissolved in it, this being the case ‘‘ since an optically active substance necessarily * has different rotation constants according as it is associated t o different degrees,” and therefore, although the evidence is ‘( rather meagre,” ‘‘ we must expect to find that the specific rotatory power of substances having high association factors in the pure liquid state varies considerably with change of solvent and of concentration, whilst those substances having in the pure liquid state association factors approximating to unity would in solution have specific rotatory powers but slightly dependent on the solvent and the concentration.” The authors assume that the association factor of E-tetrahydroquin- aldine is about 1.5-which may or may not be the case-and then show (p.1116) that when dissolved in different media this substance gives various values for [a]= lying between -45.9O and - 97.6O ; ‘( the specific rotatory power of the base in piperidine solution is less than one-half of what it is in carbon tetrachloride solution. These large variations in specific rotatory power with change of solvent can only be attributed to differences in the degree of association of the base in the various solutions.” Considering that molecular weight determinations could have been carried out in at least seven of the nine solvents (excluding acetic acid) used, it does not seem necessary to resort to the indirect (( corro- borative evidence” which the authors addme, namely, that the specific rotation of I-tetrahydroquinaldine is the.same in the free state and when dissolved in its lower homologue, tetrahydroquinoline, this being the case (p. 1117) ‘‘ because the association factor remains almost unchanged.” * In this and the other passages quoted, the italics do not occur in the original.ROTATION OF OPTICALLY ACTIVE COMPOUNDS. I. 187 I-Tetrahydroquinaldine was then dissolved in the externally com- pensated base, If, in the latter, ic the two antipodes are not quite mutually indifferent, the association factor would change on admixture and lzevotetrahydroquinaldine could not have the same specific rotatory power when dissolved in the externally compensated base as solvent as when solvent-free,” The rotation in these two conditions, it appears, is identical, and this proves ‘( in the most conclusive manner possible,” that the externally compensated base is not racemic.On the other hand, ‘ I the determinations of the densities and refraction constants of lzevo- and externally compensated tetrahydroquinaldine indicate with great probability that the association factor is the same in both.” Is not this evidence quite as conclusive as, or even more conclu- sive than, that derived from the rotation data, involving as it does fewer purely arbitrary assumptions? It will be seen a t once that Pope and Peachey’s statements do not rest on any solid foundation, and their paper has been referred to here because the results of the experiments detailed in the present communication seem to be absolutely a t variance with the fundamental assumptions of the authors.If, as they assume, substances which from their nature should have high association factors exhibit very different rotatory powers as the solvent is changed, then it follows, as they admit (p. 1112), that ‘‘ those which should be nearly monomolecular [must] vary but slightly in specific rotatory power in like circumstances.” Pope and Peachey depend for evidence as to the association of tetrahydroquinaldine on some data given by Traube regarding aniline, pyridine, quinoline, and piperidine. Now according to Traube’s method of calculation, ethyl tartrate is a unimolecular substance. Its molecular volume at 15’ (170.1) agrees closely with the cal- culated value (171.3 [Frankland, Trans., 1899, 75, 349]), and therefore the value of its specific rotation in different solvents should be very similar.As a matter of fact, however, the specific rotation of a 5 per cent. aqueous solution (+ 26’) is just three and a half times as great as that of a 5 per cent. n-propyl alcohol solution (+ 7.4”), whilst there is no reason to suppose that any marked difference in degree of association exists in the two solvents. Although, therefore, it may be still premature to deny the connec- tion between association and rotation, that hypothesis can scarcely be considered strong enough to discourage an attempt t o trace the phenomena of rotation in solution to some other cause, to some physical property of solvents which, apriori, might be expected to exercise a marked influence on any substance dissolved in them.188 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE The Relationship between the Rotation of Active Xubstances in XoZution and the Intemccl Pressure of the Solvent.Are we acquainted with such a property of solvents as this? The (1). There must be some reasonable probability of its connection (2). It must be capable of approximate or relative measurement. (3). I n order to account for the very considerable changes which occur in the rotation, it must vary in different solvents between wide limits. These requirements seem to be met by that property of liquids known as the (( internal pressure,” which often assumes enormous proportions and which varies very greatly in different liquids, It was in the hope OF connecting this pressure with rotation that the present investigation was commenced, but the idea appears to have been originally suggested by Tammann, and is attributed to him by Siertsema (compare Abstr., 1900, 78, ii, 329).The latter author has determined the influence of external pressure on the rotation of solutions of sucrose, and if external pressure is capable of influencing rotation, the fact is an encouragement to the investigation of the effect of internal pressure in this direction. The values given by the various authorities for the internal pressure of any given liquid are often very different. According to Tammann, its value in water a t Oo is about 22,000 atmospheres, whilst in ether at the boiling point it is nearly 2,500 atmospheres; according to Ostwald, the values are about half the above, the figures of other investigators being again different, but the relation amongst themselves of the figures of one authority for a number of liquids is generally much the same as that of the figures given by another.I n order to see how this pressure* would act, we can suppose a molecule of ethyl tartrate taken from amongst a large number of similar molecules and placed amongst a large number of water mole- cules, The pressure on the molecule changes then from the value which it has in ethyl tartrate to that which i t has in water. The first effect which we are accustomed to associate with change of pressure is change of volume. The volume of the ethyl tartrate molecule will change, and although, according to Tammann (Zeit.physikal. Chem., 1896, 21, 529) this change of volume is the sum of several changes, we may assume as a first approximation that in dilute * Although the word pressure is used throughout for the sake of clearness, it is without any intention of instituting a too strict analogy between this property and ordinary hydrostatic pressure. few necessary conditions which i t must fulfil are these : with rotation,ROTATION OF OPTICALLY ACTIVE COMPOUNDS. I. 189 solution the change in volume is suffered entirely by the dissolved substance. Traube, in his recent work on molecular solution volume, assumes that it is suffered by the solvent, but his grounds for doing so are not completely convincing. It seems somewhat unwarrantable t o suppose that when one molecule of ethyl tartrate is dissolved in, say, one hundred molecules of water, the latter should be altered and not the former.A t any rate, a volume change does take place on solution, which we seem at liberty to attribute to the ethyl tartrate, and which we may regard in the meantime as a measure of the change of internal pressure. If now the molecule of ethyl tartrate were quite regular, this change of pressure would probably produce no corresponding change in rotation. It is, however, assumed to be asymmetric, and conse- quently when the volume alters so also will the shape. But it is the shape, or something corresponding to the shape, of the molecule that conditions the value of the rotation, and therefore with altera- tion of volume a corresponding alteration of rotation may be expected.A mechanical conception of the process is not difficult to form, but the simplest illustration (suggested by Dr. Shroud) is afforded by a figure (say a cube) cut out of a substance whose elasticities are dif- ferent along the three axes. Such a figure, subjected to hydrostatic pressure, would alter, not only in volume, but in shape as well. This change in asymmetry of an active molecule will bear some proportion t o the change in rotstion, and it should also bear a relationship to the change in volume, and we may therefore expect to find a connection between the rotation of a substance dissolved in various media and its volume in the same media. The data for calculating the change in volume of ethyl tartrate when dissolved in several solvents are given by the density determinations, so without troubling in the meantime about its cause, we may turn t o a comparison of the change in volume with that in specific rotation.The volume of a gram-molecule of a compound in solution may be calculated from the following formula which has recently been used by Traube : M + S s M.S. V. (molecular-solution-volume) = - - - d 6 ' X= weight of solvent associated with 1 gram-molecule of d = density of solution. 8 = density of solvent. Where N = molecular weight of dissolved substance. dissolved substance. If the solution is one of percentage composition p , then M grams of sub-190 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE stance are associated with M(loo - p ) grams solvent, that is, P a= w o o -P) P and, substituting this value of S in the above equation, we find M.S.K=--( M 100 7-6 100--p 1.P The figures which have been calculated by means of this formula will be found on pp. 214-215, and in Fig. 8 these values of molecular- solution-volume at 20° are plotted relatively to concentration, and below the corresponding concentration-rotation curves for the same temperature. It is evident, in the first place, that the values obtained for molecular-solution-volume in dilute solutions are somew hat uncer- tain, this being due to the difficulty of carrying out the density deter- minations with sufficient accuracy, the effect of a slight error being great in dilute solutions, as is apparent from an examination of the formula used in the calculation, It is therefore rather difficult t o determine how the most probable curve should be drawn in each case from the experimental data.The curves for water and methyl alcohol are not very satisfactory for p < 10, whilst that for ethyl alcohol is the least satisfactory of all; i t has been drawn, however, as nearly as possible between the values for p = 5 and p = 10.94. The ethyl tartrate molecule evidently undergoes a very considerable change in volume on solution in a large quantity of water. At infinite dilution, the molecular-solution-volume seems to be about 157.5 c,c. at 20°, that of the free ester being 170.9 C.C. In methyl alcohol, the change in volume is also considerable but not so great as in the case of water, the value at infinite dilution being about 159.3 C.C.In ethyl alcohol, the volume is 164 C.C. whilst in It-propyl alcohol it is 167.7 C.C. If now the corresponding concentration-rotation curves are examined, it will be noticed that the values of the rotations at infinite dilution stand in the inverse order, and although the rotations do not seem to be quantitatively related to the values of the molecular- solution-volume, there may be a qualitative relationship. There can be little doubt that the order of the values of molecular- solution-volume at isfinite dilution in the above four cases is correct, although the values themselves are a little uncertain, but it is more difficult to say whether the rotation of ethyl tartrate in glycerol can be similarly explained by the value of its molecular-solution-volume in that solvent.As has already been remarked, the accurate deter- mination of molecular-solution-volume becomes more and more difticult as the dilution of the solution increases, and it may be that tbe curveROTATION OF OPTICALLY ACTIVE COMPOUNDS. I. 191 11.S.V. (infinite dilution). Solvent. drawn in Fig. 8 for glycerol is not correct and that the molecular-solution- volume a t infinite dilution is greater than that of ethyl tartrate in ethyl alcohol. Nevertheless, the curve obtained from the experi-- mental figures is so regular as to be some guarantee of its accuracy, and assuming it to be correct the molecular-solution-volume of ethyl. tartrate in glycerol at infinite dilution has a value between those found in ethyl alcohol and methyl alcohol, namely, 163.3, and, in agree- ment with this, the rotation of glycerol is greater than in ethyl alcohol and less than in methyl alcohol.At infinite dilution, therefore, the order of the rotations and molecular-solution-volumes correspond inversely, a small volume being associated with a high rotation, as is. apparent from the following table : [alD (infinite dilntion). ........................... 157.7 .............. 159.3 ........................ 163.3 ................ 164 Water Methyl alcohol Glycerol Ethyl alcohol.. n-Propyl alcohol ............ 1 167'5 I 26-15' 11.50 10.57 9'13 7 '40 Change in aD due to solution. 18.49" 3-84 2'91 1'47 - 0.26 The parallelism of these figures is, as a first approximation and in a qualitative sense, fairly satisfactory, but if this relationship is not merely accidental, that is, if variation in molecular-solution-volume does really determine variation in specific rotation, then me may expect to find a connection, not only at infinite dilution, but under ail circum- stances.That is to' say, the curves for rotation should correspond throughout with the true curve for molecular-solution-volume. Now it will be seen in Fig, 8 that the molecular-solution-volume curves for ethyl tartrate in water, methyl alcohol, ethyl alcohol, and n-propyl alcohol are all, in fact, of much the same form, they show a gradation of a similar order to that found in the rotation curves, and this connection between rotation and volume becomes much more striking when glycerol is also taken into account, because its behaviour differs markedly from that of the other solvents ; it presents something of the character of an exception, and if an exceptional variation in rotation is accompanied by exceptional variation in molecular-solution- volume, the suggested correlation of these two phenomena becomes more probable. The concentration-rotation curve for this last solvent is, as has already been mentioned, a remarkable one.At 20°, the rotation a t infinite dilution is + 10~6~. As the concentration increases, the rotation diminishes much more rapidly than in the three other alcohols, and corresponding with this the volume increases much more192 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE rapidly between p = 0 and p = 20, and in the meantime this qualitative relationship is sufficient, It can scarcely be expected that the molecu- lar-solution-volume curves should cut one another at exactly the same concentrations as those at which the corresponding rotation curves intersect.It might be possible to trace this close connection if the trwe curves for molecular-solution-volume were known ; the curves drawn are, however, only approximations, and in all probability not very satis- factory ones. All the contraction on solution has been assumed to take place in the ethyl tartrate. Probably even in a p = 5 solution the error thus committed is considerable, and in a p = 20 solution it must certainly be great. I n the diagrams, however, this volume change is assumed to take place, for all concentrations, in the ethyl tartrate only, which is certainly incorrect.I n reality, the total volume change consists of at least two changes, one in the solute and one in the solvent, but, what is of the greatest importance here, it is not possible to separate it into these two or more simple changes. The curves on the lower part of Fig. 8 are there- fore not correct, although they can probably still give some indication as to the actual behaviour of the substances examined. This (the merely approximate nature of the curves) explains why solutions having the same specific rotation need not necessarily show the same molecular-solution-volume for the dissolved ethyl tartrate. The volume of the tartrate may really be the same in two different solu- tions whilst the volume change in the solvents is not the same.For instance, a p = 25 solution in glycerol has the same rotation ( + 7.6") as a p = 55 solution in ethyl alcohol, although the corresponding volumes are not the same, being for the former 169 C.C. and for the latter 16'7.5 C.C. It is evident that in this case a greater volume change is likely to have taken place in the ethyl alcohol of the latter solution than in the glycerol of the former, both of these changes, however, from the method of calculation, being ascribed to the ethyl tartrate alone. Such a connection as this between molecuIar-solution-volume and rotation appears to render possible a rational explanation of that very interesting phenomenon, the occurrence of a minimum rotation of ethyl tartrate dissolved in glycerol or a-propyl alcohol. For if rotation is really dependent on molecular-solution-volume and in glycerol solution at 20° the minimum rotation occurs when p = 65, then it follows that the molecular-solution-volume for the same temperature should be a maximum at that concentration.Now the molecular- solution-volume curve for glycerol rises rapidly with increasing con- crntration up to about p = 25, after which the increase is much more gradual ; but at about p = 25 it is probable that the glycerol also suffers considerable change in volume and if this be contraction it willROTATION OF OPTICALLY ACTIVE COMPOUSDS. I. 193 counteract the effect of expansion in the ethyl tartrate. That is t o say, the ethyl tartrate ma-yreally continue to expand with increasing con- centration, the state of affairs not being represented by the full line in the figure but rather by the broken one, until a t about p = 65 a maximum volume of about 173 C.C.is reached, the volume then again diminishing rapidly to 171 C.C. when p=lOO. This assumes the possibility of a. value greater than normal for ethyl tartrate in solution, which is, however, surely as possible as one less than the normal. So far it would seem that the assumption of a relationship between molecular-solution-volume and rotation is at least worthy of considera- tion, but there are considerable difficulties to be overcome before the connection can be regarded as proved. One of these is met with in the fact that in n-propyl alcohol, although the molecular-solution- volume at infinite dilution is only 167.1, the rotation is lower than that of the pure ester by 0.26' instead of being higher.The dis- crepancy is not very great, and an explanation can scarcely be expected until more data have been obtained. Another difficulty occurs when the influence of temperature change upon the rotation of these solutions is considered. Except in one case-solution in water-increase of temperature causes increase of rotation. But increase of temperature also causes increase of mole- cular-solution-volume and therefore ought to be attended by decrease of rotation. We have here a direct contradiction, but the following consideration will show that i t is not inexplicable. Let us take the case of free ethyl tartrate, whose rotation, as is well known, increases rapidly with rise of temperature.Imagine one particular molecule, A, in the liquid kept a t a definite temperature, T, whilst all the others are heated to a higher temperature. The pressure on the molecule A will decrease, its volume will increase, and its rotation should also decrease ; that is to say, the molecule becomes less asymmetric. Now let the molecule A be also heated to the higher temperature. The effect will be expansion of the molecule A against a certain pressure-certain forces-resulting in another increase of volume. I n this second case, however, the proximate cause of change of volume is not the same as before-the effort comes from within the molecule, the change is not due to variation in the properties of surrounding molecules-and now the expansion may take place in such a way that the molecule becomes more asymmetric again, and since we know in general that a slight change in the temperature of a liquid or solid will produce a much greater alteration of volume than an enormous chauge of pressure can bring about, the second of the operations just mentioned will probably have a greater effect than the first on the rota- tionof the molecule A, so that the net result is an increased rotation.194 PA4TTERSON: THE INFLUENCE OF SOLVENTS ON THE The same volume might be arrived at either by heating or by diminution of pressure alone, but the shape, that is, the asymmetry, of the molecule would not be similar in each case.The asymmetry of the molecule then depends on temperature and pressure-or some- thing analogous to pressure-but it is only constant for definite values of both variables; the asymmetry is not so simply conditioned as the volume.If this is admitted, then the simultaneous increase of rotation and volume presents no difficulty either in the case of solution in the alcohols or in water, although in the latter case the rotation in dilute solutions (anything less than p = 55) decreases with increase of temperature. I n dilute aqueous solution, where the internal pressure is great, the asymmetry of the ethyl tartrate molecule has become such that the effect of increasing temperature is to produce a less .asymmetric molecule. Under a low pressure (solutions in the alcohols) effort from within the molecule produces greater asymmetry ; under a high pressure,* it produces a less asymmetry, and consequently between these extremes there may be a pressure under which the molecule of ethyl tartrate has an asymmetry practically unaltered by heating, increase of temperature causing expansion of the molecule certainly, the shape, however, remaining always the same.This particular case appears to be found in an aqueous solution for which p=55. The rotation of such a solution is practically insensitive to Semperature. I n this connection, it should also be noticed that the ethyl tartrate molecule appears to be most sensitive t o temperature in those solvents in which its molecular-solution-volume is greatest, which fact is in agreement with the above considerations. In dilute glycerol solutions, the sensitiveness of ethyl tartrate is probably slightly less than that of the free ester, whilst in methyl and ethyl alcohols it is practically the same as that of the free ester.I n m-propyl alcohol, however, it is rather greater. This corresponds with what has been suggested above; the greatest sensitiveness is shown in that solvent in which molecular-solution-volume is high, that is, in n-propyl alcohol. It is obvious, however, that here again we meet with difficulties, for the molecular-solution-volume of the ethyl tartrate in an insensitive aqueous solution (that is, of p=55) is about 163 c.c., which is higher than the volume in infinitely dilute methyl alcoholic solution (159.3 c.c.), and therefore the rotation of dilute solutions in the latter solvent ought also to decrease with increasing temperature, which, of course, is not the case.We must remember, however, that it is not possible to tell what the true molecular-solution-volume of ethyl tartrate in 55 * Always in this particular case of ethyl tartrate, of course. With some other molecule of different asymmetry, the phmomena might be reversed.ROTATION OF OPTICALLY ACTIVE COMPOUNDS. I. 195 per cent aqueous solution reaIIg is, and consequently must be content in the meantime with the indications already pointed out. This relationship of rotation and molecular-solution-volume has been first discussed because the figures necessary for the calculation of both these quantities may be directly obtained from the data collected in this investigation. As already remarked, however, the volume change accompanying solution is probably more complicated than has been assumed, and it is worth while to try to trace the cause of variation in rotation still further back, namely, to that property of liquids which has been supposed to be the cause of variation in voluiiie, the internal pressure.It may be that the variations in volume due to solution are not directly proportional to variation of internal pressure whilst the variations of rotation are, and that therefore there may be a closer and more obvious connection in the latter case than in the former. Barmwater (Zeit. physikal. Chem., 1899, 28, 124) has calculated this quantity for a number of substances, and Traube has suggested a method of calculation based on his work on molecular-solution-volume, whilst others have been proposed by van der Waals, Stefan and Tammann.The choice of a particular set of figures would involve a critical discussion of the various methods of calculation, and this we may avoid by considering, instead of the pressure, that which Briihl (Zeit. physikal. Chem., 1899, 30, 43) calls the medial energy or heat of disgregation of a liquid, because it is from this quantity that Stefan and Tammann both derive-although by slightly different reasoning- their figures for internal pressure. The heat of disgregation is calculated from the formula liquid, E = mechanical where M= mol. wt. of where D = heat of disgregation, IZ = heat of vaporisation, p = vapour pressure, P=volume of 1 gram of vapour, Pl=volume of 1 gram of equivalent, which, since we may set RF p ( P - V1)= - M ’ substance, reduces to The heat of disgregation therefore represents the amount of energy necessary to overcome the internal forces of a liquid, and to separate its particles from each other at any particular temperature and pressure.196 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE 575 '9 272.1 2234 By calculation from the above formula, the following figures are obtained.303.8 48*7 Solvent (t = 0"). 274.5 q7.G 26'4 Water .......................... Methyl alcohol.. .............. Ethyl alcohol .................. +24' (~'10) 9.9 14.1 (p=lO) 12.3 @ = l o ) o.5 11'8 ( p = 5 ) D. 1 A. Water ........................... Methyl alcohol ................ Ethyl alcohol.. ................ n-Propyl alcohol.. ............ 523'1 248'6 201 -6 175'2 t-26.5" ( p = l O ) 9'5 ($?=lo) 6'2 (p=5) A.17.0 3 -3 It will be noticed that here the differences in rotation and heat of disgregation are nearly proportional to each other. By making the calculations for a higher temperature, n-propyl alcohol may also be included in the table. Its heat of vaporisation has only been determined at the boiling point, and has been found to be 166' (cal.). By assuming that it varies in the same manner with temperature as those OF methyl and ethyl alcohols, the value a t 60' can be approximately obtained. From the number thus deduced, the heat of disgregation in the following table has been calculated : I D. Solvent (t= 60'). I These figures are also in fairly close agreement with each other ; the rotation in the different solvents appears to decrease in much the same proportion as the heat of disgregation.Such an agreement may -of course, be merely accidental, and the examination of several other series of solvents will be required to determine the point, but the figures are certainly striking. Plotted on a system of coordinates, the one property is seen to vary almost linearly with the other. It is not necessary, however, to enter into any further discussion regarding this relationship, for nearly everything that could have been said here has been said for molecular-solution-volume, and applies almost equally to both. It would appear, then, that molecular-solution-volume, heat of dis- gregation and rotation have some connection with each other, and in the paper already mentioned Briihl shows that the heat of disgregation and the dielectric constant of a substance are also related phenomenaROTATION OF OPTICALLY ACTIVE COMPOUNDS.I. 197 Association factor of solvent. the latter again varying in an analogous manner with dissociating power, as Nernst has pointed out, and thus we are led back again to a problem which was discussed earlier in this paper-the relationship of dissociation and rotation, As has already been shown, no definite connection can be deduced from the various researches which have been carried out on the subject, but it is nevertheless possible that although the dissociating power of a liquid-a term usually applied to a solvent with regard only to its behaviour towards electrolytes- does not cause actual dissociation in substances other than electrolytes dissolved in that liquid, it may nevertheless modify them.Thus a substance such as ethyl tartrate dissolved in two different liquids may exist in simple molecules in both, but still the force which we call dissociating power is acting t o a different extent in each case, and although unable to cause any decomposition of the molecule into ions, may yet exert some other influence on it which will be evident as change of rotation, for instance. As a summary of the relationships discussed in this paper, the following table may be added. The heat of disgregation is given for 60°, whilst the other figures are for lower temperatures. The association factors of the solvents have been introduced, as it is of interest to compare their values with the other figures given.M. S. V. of ethyl Traube, 15". Ramsay and Shields. tartrate. Infinite dilution, 20". Water , . . . . . . . . . . . Methyl alcohol. Glycerol . .. . , . . . Ethyl alcohol .. n-Propylalcohol f 26 '2" 1 '644" (20") 11.5 2*32* (20") 10 -6 9-1 1*65* (20O) 7'4 2'25 § (46'3" 2-3-f 1.798 1-90? 1-67? 1.667 Heat of disgrega- tion of solvent, 60". 523.1 248.6 201 '6 175-2 157'7 159.3 163.3 164 167'5 Dielectric constant of solvent. From this table, and from what has already been said, i t appears that a relationship, satisfactory in a qualitative sense, can be * Proc. Roy. Soc., 1894, 56, 180. t These numbers are not quoted by Traube?, but are calculated according to his 2 Nernst, Theor. Chemie, 3rd edition, p. 305. 11 Thwing, Zeit. physikal. Chem., 1894, 141, 293.l T Traube, Ueber den Raum der Atome, Ahrsns Sammlung, pp. 32 and 41. VOL. LXXIX. P direotions. Trans., 1893, 63, 1102.198 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE established between variation of rotation in solution and variation of molecular-solution-volume-at least so far as this series of solvents is concerned-and when these phenomena of rotation are traced further back to what may with reason be regarded as the cause of variation of molecular-solution-volume, namely, differences of internal pressure, or what is probably the same thing, of heat of disgregation, very similar regularities are observed, which seems to show that the original assumptions, dependence of volume on internal pressure and rotation on both, are justified. Finally, it may be pointed out that if the idea developed here be correct, greater account must be taken, when considering the rotation of homogeneous active substances, of their own internal forces ; the molecular rotation is not that of a free molecule of the compound.The molecular rotation of a homogeneous liquid is the rotation of the molecule subjected to the internal forces of that liquid. Rotation of ethyl tartrate. Temperature. 10 -8" 37 -6 33.7 29.9 20.1 89 '4 84.4 77 '1 67'2 55 -1 46.1 25 -1 16 11.3 100 an (100 mm.). + 8.047" 11'354 10'842 10-392 9'244 15.129 14'725 14510 14.110 13.600 12-792 12.067 9'900 8.719 8.089 Density. 2.2144 1.1913 1.1952 1.2051 1,1230 1.1349 1'1399 1 *1472 1.1576 1.1697 1.1789 1'2000 1 '2094 1,2140 I -1873 Densities determined : Temperature. 16.8" 37'2" 46.8" 58'3" 68.1" Density ......1.2087 1.1878 1.1783 1'1665 1'1566 + 6-63" 9 -56 9 -10 8 -70 7.67 13.47 12.97 12.73 12-30 11.75 10-94 10.24 8 -25 7-21 6.66 76'2" 99.4" 1 *1484 1.1248 Ethyl Twtrate in Watev. The distilled water used in these experiments was well boiled before use.ROTATION OF OPTICALLY ACTIVE COMPOUNDS. I. 199 Temperature. a, (400 mm.). Density. r a I$ + 26.06" 25-01 25.67 26.30 20" 50.8 27-2 22 -3 1.043 0.995 1 -026 1 -052 1 '0006 0'9945 0.9990 1*0001 Densities determined ; 26'2" 0.9993 Temperature ............ Density .................. 15.8" 1.0017 p = 2-5. Temperature. aD (400 mm.). Density. [alf. 14.7" 50.3 27'3 14.6 + 2,612" 2'450 2.572 2'593 1.0053 0.9940 1'0023 1'0053 -1- 25 *98" 24-65 25.66 25-80 55" Densities determined : Temperature......... 14.6" 26%" 41 '6" Density ............... 1 *0053 1.0024 0.9971 It 0*9912* * Although the water used in making up this solution had been boiled, it was found difficult to carry out the density determinations a t these higher tempera- tures owing to the separation of air-bubbles. These two figures are probably therefore too low. p = 4.999. Temperature. a, (400 mm.). Density. 15.3" 30 -8 16.6 (after standing 14 days) + 5 '304" 5'199 5.289 1 *0110 1,0077 1'0108 + 26 -23" 25.80 26-16 Densities determined : After standing 16 days. 36.1" 14.4" 1.0066 1'0113 Temperature ......... 17 '6" Density ............... 1*0106 P 2200 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE p = 9.994. Temperature. 21'3" 14'8 29.4 18'6 20 -7 (after standing 5 days) a, (400 inm.).+ 10.598" 10-701 10.415 10-650 10.490 Density. 1'0211 1.0231 1'0184 1.0220 1 '021 3 + 25.95" 26.17 2557 26.05 25.69 Densities determined : After standing 5 davs. Temperature ...... 19'2" 13" Density.. . ... . . . . . . 1.0220 1 -0237 27.5" 34-4" 14 *lo 18 - 6 O 1.0192 1.0164 1'0235 1.0220 p = 24.954. Temperature. 15" 32 '5 25 20.6 15 44.9 15 a,, (249 '6 mm.). Densities determined : + 15.725" 14.867 15.314 15'505 15.747 14.249 15-714 Density. 1.0597 1.0511 1,0549 1.0571 1.0597 1-0457 1.0597 4- 23.83" 22'71 23-31 23.55 23-85 21 -88 23-81 After heating After heating to 33'3". to 44'9". Temperature . . .. 16" 23" 33.3" 15'8" 15.2" Density ......... 1,0594 1'0562 1'0509 1.0595 1'0598ROTATION OF OPTICALLY ACTIVE COMPOUNDS. I. 201 a, (100 mm.). p = 49,993.Density. [ 41;. Temperature. 14.7" 34-2 26 -8 19-9 15.1 46 15 Twelve hours later : 15'3 52.6 14.5 66 -2 56 -1 15.9 17 aD (100 mm.). + 9'759" 9.379 9.529 9,670 9.742 9.197 9'727 9.739 9'112 9.750 8-894 9 '065 9'704 9 '689 Density. 1'1193 1.1052 1.1102 1.1153 1.1190 1'0973 1.1190 1.1188 1.0927 1.1194 1.0835 1 -0903 1'1184 1'1175 r41:. + 17-44" 16.97 17.17 17'34 17.41 16-76 17'39 17.41 16.68 17-42 16'42 16'63 17-35 17.34 Densities determined : After experiment. Temperature.. ..... 15.7' 22'3' 353" 63.8" 70.3" 16.8" Density ............ 1.1186 1.1137 1'1046 1.0807 1 *0752 1.1180 p = 74.99. Temperature. 16.2' 49 -1 45*0 30.6 19.5 16 67'2 53'4 18 Densities determined 1 + 10'069" 11'100 11*000 10'587 10'210 10.049 11'497 11'194 10.144 Temperature.. ....... 18 *3" Density .............. 1 *1690 1.1707 1.1408 1'1446 1,1579 1.1679 1.1709 1 -1248 1-1369 1.1691 + 11-47" 12.97 12-82 12.19 11 '66 11.44 13.63 13.13 11'57 After experiment, 58" 18.2' 1.1323 1.1691 Ethyl Turtrate in Methyl AZcohol.The methyl alcohol used was Kahlbaum's best quality and was redistilled from some sodium which had been carefully freed from petroleum.202 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE Temperature. The density of the methyl alcohol was determined at various tem- peratures, the following numbers being obtained : Temperature.. 16' 2 9' 3s-4' 48' Density . . , . . , 0.7953 0-7830 0.7741 0-7646 From these figures, by extrapolation, we find that the density at 0' is 0-8105, whilst, according to Dittmar and Fawsit, the density of pure methyl alcohol a t 0' is 0-81015. uD (400 mm.).p - 5 . 14'8" 13'2 49 '2 40 '2 32.5 24-3 12.2 + 1,845" 1 *773 2.121 2'081 1.993 I 1'913 I 1.750 Density. 0.8120 0.8137 0.7796 0.7860 0'7949 0.8029 0,8147 Densities determined : Temperature.. . . . , 20.2" 33.8" 46.7" Density .... ..... ... 0.8068 0'7939 0.7815 +11*36" 10.92 13.60 13.24 12-54 11.91 10'74 After experiment. 18.3" 0'8068 p=10. Temperature. 18'9" 12.8 16.7 53 46.7 42'6 35.5 27 16 '2 13 a, (400 mm.). + 3'744" 3.545 3.715 4.368 4.320 4.238 4.138 3.946 3-695 3.615 Density. 0'8240 0.8300 0'8288 0.7905 0 *79 70 0.8010 0.8080 0,8162 0.8268 0'8306 + 11'36" 10.68 11.21 13-81 13-55 13'23 12.81 12.09 11-17 10'88 After experiment. Temperature ... ... 22" 30.6" 39.6" 45'8" 14.2' Density .. . . , . . .. 0.8210 0 ,8128 0.8040 0'7980 0.8286ROTATION OF OPTICALLY ACTIVE COMPOUNDS.I. 203 p = 25.01. ~~ Temperature. 18.3" 20 13.2 48'1 45 -6 42.8 39-4 33'4 25 19'2 18.9 aD (249'6 mm.). + 6.044" 6.110 5.810 7.068 6.997 6'913 6'813 6.627 6'328 6'074 6'065 Densities deteerrnined : Temperature ... . . . 14 *lo Density. 0.8757 0.8741 0.8807 0.8468 0 '849 2 0.8518 0.8553 0'8610 0'8693 0.8750 0-8751 27'3" 36" Dens'ty ......... 0,8799 0.8671 0.8585 [a15 -t- 11 '06" 11.20 10-57 13'37 13'20 13 '00 12.76 12-33 11.66 11.12 11-10 43.2" o-a515 p = 50000. Temperature. 13" 16 53.8 46 -9 43.2 38.2 24.9 17 13'2 14'7 34.8 a, (100 mm.). + 4.781" 4.891 6.088 5'968 5.863 5-711 5-616 5.281 4.966 4.820 4.908 Density. 0.9762 0.9743 0.9366 0.9435 0'9471 0.9521 0.9555 0.9655 0.9733 0.9770 0-9755 Temperature ...,.. 19.8" 34.4" 44.6" Density ......... 0.9707 0.9561 0-9460 +9*80" 10.04 13-00 12-65 12.00 11.76 10-94 10.21 9-87 10.06 12-38 51" 0,9395204 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE p = 75. Temperature. 14'9" 12.5 51 '1 45.6 40'8 33.7 24-7 17.1 a, (100 mm.). + 7'012' 6-839 9.380 9.014 8-792 8.374 7.767 7'207 Dmsities determined : ~ Density. 1.0879 1 -0901 1 -0504 1.0564 1.0610 1.0685 1'0778 1.0853 [ a 15 Temperature ...... 18'3" 32'8" 39.7' 53.2" Density . . , . . . ... . . , 1.0842 1.0692 1'0626 1 -0488 + 8-59" 8-24 11-91 11'38 11 -05 10.43 9.61 8 -85 After experim.ent. 17 '6" 1 *0853 Ethyl Tartrate in Ethyl Alcohol. The ethyl alcohol used was carefully distilled over sodium. Its density was determined with the following results : Temperature 17.6' 30.4O 4 1 0 6 ~ 58 *2O Density ......0.7932 0.7822 0.7723 0.7575 This gives, by extrapolation, the number 0-8090 at Oo, whilst Mendeldef found 0.80625 (Landolt-Bornstein). p = 5a0013. ~~ Temperature. 18.8' 17 11 13'7 15-9 51 -8 42.9 37'2 31 -1 21.7 30.3 23 uD (400 mm.). + 1,419" 1 '370 1'262 1'304 1 '350 1'840 1.729 1'660 1,565 1 *417 1.551 1.459 Density, 0.8072 0.8090 0'8143 0'8119 0.8099 0-7779 0.7858 0.7908 0.7961 0-8046 0-7970 0.8035 + 8-79' 8.47 7-75 8-03 8 '33 11.81 11'01 10-49 9.83 8.80 9-73 9.08 After experiment. Temperature .... , ,. , ... 20-5" 30.6" 38" 54'4" 16" Density ........... ,... 0'8056 0'7969 0'7900 0.7754 0.8097ROTATION OF OPTICALLY ACTIVE COMPOUNDS. I. 205 p = 10.94. Temperature. 18.6" 7 '1 59.1 52 '8 41 23 '6 6.6 a, (400 mm.). Demsitier determined : + 3.023" 2.553 4-220 4'067 3.772 3'243 2-542 Density.0-8251 0'8353 0-7893 0.7950 0.8052 0.8208 0'8359 I: 4:. + 8 '37' 6-98 12-22 11-69 10.71 9.03 6 '95 After experiment. Temperature .. ... ...... 17.4" 38.4" 55.8" 65.8' 20'2" Density ...... ......... ... 0.8263 0.8079 0'7922 0.7826 0'8240 p = 20°003. Temperature. 16.2" 8.7 64.4 59.7 54 45-6 37.9 23.6 9-5 (249-6 mm.). + 3.348" 2.965 5.078 4.945 4'803 4.551 4.296 3.756 3,076 Density. 0.8569 0-8639 0.8119 0'8161 0'8215 0.8293 0.8367 0.8499 0.8632 + 7-82" 6-87 12'53 12.14 11 *71 10.99 10-28 8.85 7-14 Densities determined : After experiment. Temperature 131' 18.9" 33O 35*2" 39" 46.4" 69" 15%" Density .... 0.8595 0.8544 0.8418 0,8397 0'8362 0.8294 0.8074 0.8577206 PATTERSON: THE INFLUEX'CE OF SOLVENTS ON THE p = 40.002.Temperature. 19-7" 10 *2 60.3 55 51 -2 44 '9 40 -1 36.7 25 12.4 I Density. aD (249 '6 mm. ). + 7 *260" 6'263 10.335 10'058 9'831 9-368 9'063 8.788 7.785 6 *528 0.9244 0.9334 0.8857 0.8907 0'8944 0.9004 0.9050 0'9083 0.9194 0'9312 + 797" 6 -72 11'69 11'31 11.01 10.42 10-03 9'69 8 '48 7.02 After experiment. Temperature ........... 16.7" 33.3" 43'8" 62'8" 19" Density ............... 0.9272 0.9114 0.9017 0.8833 0.9258 p = 60.01. Temperature. 1 aD (100 mm. ). 1 1 l- 21.3" 11.1 14 56.7 48.2 39 34 22 + 4.667" 3'987 4.140 6.480 6'122 5'662 5'389 4.732 Density. 1.0040 1.0141 1.0113 0'9690 0.9774 0.9865 0,9914 1'0030 + 7-75" 6.55 6-82 11-24 10-44 9-56 9-06 7 '86 Densities determined : Temperature ............ 17.5" 28 '9" 47.8' 59 *lo Density ................1.0079 0.9969 0'9780 0'9668 Ethyl Tartrate in n-Propyl Alcohol. The n-propyl alcohol used was of Kahlbaum's best quality and was Its density was determined at various temperatures with the follow- Temperature 20° 2 3 ~ 4 ~ 3 2 O 40° 62*8O 69*6* Density.. .... 0.8039 0*8012 0.7942 0.78'75 0.7682 0.7622 carefully distilled over clean sodium before use. ing results :ROTATION OF OPTICALLY ACTIVE COMPOUNDS. I. 207 Landolt-Bornstein's tables give d Oo/Oo 0.8205 hence d 0°/4' 0*8204, whilst by extrapolation from the above figures we find d 0°/4' 0%210, so that, judged by the density, the alcohol used seemed almost pure. However, after three solutions had been examined some doubt arose as to whether a portion of one of them had not been returned by mistake to the bottle containing the pure propyl alcohol instead of to that for residues, and to determine this, some of the former liquid was examined polarimetrically and found t o have a slight laevorotation.This might be due to presence of ethyl tartrate (although in that case a positive rotation was to be expected), so the propyl alcohol was redistilled and, on examination, the distillate and the residue left in the flask were found to be laevorotatory to almost exactly the same extent, which although proving that no mistake had been made in the first instance, also showed the propyl alcohol to be somewhat impure. This rotation of the alcohol being only very slight and several experi- ments having already been carried out, it seemed unnecessary to repeat them, since the quantity of impurity present probably did not influence the effect of the propyl alcohol on the rotation of the ethyl tartrate, except by superposition. The rotation of the propyl alcohol was therefore carefully determined : uD - 0.067O at 18.8' in a 400 mm.tube. uD -0994' at 68.5' ,, 99 ;;_Its rotation is thus very small, but becomes of some importance in the case of dilute solutions, and consequently the results of the experiments performed have been approximately corrected for the rotation of the propyl alcohol, the length of the tube and the com- position of the solution examined being taken into account. p = 2.5004. Temperature. (400 mm. ) Corr. I Obs. I I I Densities determined : Temperature . , . . . . . . . 16 *6O Densi ty... ... .... . . ..... , 0.8146 I Density. I [a]:. (400 mm.). I I +0-554" 0-8152 + 6.79" 0.690 1 0.8045 1 8-58 0.574 0 '81 20 7.07 20.9" 32.5" 0.8111 0.8017208 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE Temperature. p= 4.9996. 0b:2i400 Density. 0.8279 0.7970 0.8023 0'8102 0'8191 0.8254 0.8279 0.8826 17'6" 20 68.8 62'3 52'6 45'7 28 '3 18'4 - f 1 -072" 1 *125 1.867 1 *775 1.662 1.572 1 -308 1.108 Corr. + 0.064" 0.066 0.091 0'088 0-082 0.079 0.069 0-065 True aD (400 mm.). + 1-136O 1'191 1.958 1.863 1 '744 1'651 1 *377 1.173 Density. 0'8201 0.8174 0'7763 0.7821 0.7907 0.7970 0'8112 0'8197 + 6'93" 7'29 12.61 11.91 11.03 10-36 8'49 7 *16 Densities determined : After experiment. Temperature 16.7" Density ...... 0'8210 33.2" 44.9" 58" 80.2" 18.8" 0.8075 0,7977 0.7863 0.7661 0.8193 p = 7-71 3.Temperature. 18.7" 55-1 48'7 39'5 28'9 21'3 18'5 13 Obs. a, (400 mm.). + 1 -766" 2'751 2.616 2-396 2.082 1'858 1 *76l 1.570 Corr. + 0'062" 0.089 0'084 0-077 0.070 0'064 0'062 0.058 True aD. (400 mm.). + 1 *828" 2.840 2.700 2'473 2.152 1.922 1.823 1.628 ra1;. +7.16" 11.55 10'90 9-89 8 '52 7.55 7-13 6 -34 Densities determined : After experiment. Temperature 23.5' Density ...... 0.8236 34.2" 45-8" 58'4" i9.6" 20'8" 0.8148 0'8050 0.7941 0.7763 0.8268ROTATION OF OPTICALLY ACTIVE COMPOUNDS, I. 209 Density. p = 17*507. I: a 1;. ~~ True a, (249-6 mm.). + 2.082" 2'939 2-705 2.395 I I f 0-011" I- 2.093" 0.9259 + 6-03" 0'014 2.953 0-9080 8 -67 0-013 2-718 0'9122 7 -94 0-012 2-407 0-9194 6 '98 15 '9" 29'7 19.7 + 2'354" + 0'035" 3'010 0.040 2.555 0 037 + 2'389" 3.050 2'592 Densities delermirted ; Temperature............ 18 '9" Density.. ................ 0'8570 0.8596 0.8480 0.8563 28'2" 0.8492 + 6 '36" 8 -23 6 -93 p = 25. Temperature. 18'9" 68'2 63.3 57'9 51.9 42 -3 33'1 18'8 Obs. a, (249'6 mm.). + 30626~ 6.125 5.945 5.756 5,486 5.076 4.496 3-608 ~ Corr. + 0.033" 0.053 G.051 0 -049 0.047 0 *043 0.039 0'033 True an. 249'6 mm.). + 3,659" 6.178 5'996 5.805 5-533 5-119 4.535 3'641 Density. 0.8799 0 '83 58 0.8402 0.8453 0.8509 0.8596 0.8678 0.8800 Dewit& determined : Temperature ...... 17.7" 31 -6" 52.1" 70 -6" Density ............ 0.8810 0.8691 0.8506 0'8334 [a];. + 6 '67" 11.84 11.44 11.01 10'42 9 '54 8.37 6.63 After experiment. 19" 0,8802 p = 37-51. Temperature. 15.9" 36 31 23 '2 Ob8' an (loo I Corr. 1 True uD (loo I Density.1 [a]:. mm.). mm.). Densities determined : Temperature ............ 17'8" 35-6" Density ................. 0 '9242 0.9082210 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE p = 49.834. Temperature. 19" 71 -2 63.6 58 '2 52.4 41-2 28 19 Obs. aD (100 mm.). + 3.046" 5 '305 5 -021 4.830 4-613 4,150 3.525 3-031 Densities determined : Temperature ......... Density ............... - ~~ Corr. + 0'009" 0,016 0-01 5 0-014 0,014 0.012 0.011 0.009 True aD (100 mm.). -I 3.055" 5 *321 5-036 4-844 4.627 4.162 3.536 3.040 Density. 0.9687 0.9237 0.9262 0.9514 0.9370 0'9479 0.9601 0.9687 [ 42. 4- 6 '33" 11-56 10.91 10-43 9.91 8.81 7.39 6 '30 After experiment. 19.8" 31.20 43.8" 59-40 80.20 i s e 0.9678 0.9571 0.9453 0'9303 0'9099 0.9696 p = 74.99. No correction has been made in this case for the rotation of the propyl alcohol.Temperature. 20' 79 73.6 69.9 60'5 33-3 47.1 17.7 a, (100 mm.). + 5'276" 9'065 8.928 8'791 8-356 6'495 7.605 5-133 Density. 1,0756 1.0169 1 '0221 1 -0259 1 -0350 1.0622 1,0487 1 *on30 c 1:. + 6 '54" 11.89 11 '65 11.43 10.77 8.15 9.67 6 '35 Dernsitk determined : After After experiment. * experiment. * Temperature ...... 21" 38'3" 50.6" 71.2" 18" 20.8" Density ............ 1 *0747 1.0576 1'0448 1 '0248 1.0833 1.0811 In this case, the density after experiment differs more than is usual from the original density.ROTATION OF OPTICALLY ACTIVE COMPOUNDS. I. 2'11 Temperature. Ethyl Tartrate in GZycep*oZ. It boiled between 177O and 1 7 8 O under 20 mm. pressure, the temperature of the bath being 225-235'. Its density was determined at different temperatures with the following results : The glycerol used was carefully redistilled in a vacuum.Temperature 13.2 30' 5 4O 7 5 . 5 O 99*5O Density .,.... 1,2651 1.2552 1.2397 1.2256 1.2097 u,, (200.mm). p = 4-985. Temperature. 1 U, (249'6 mm). 98'2" 17 77.5 47.6 35.7 26'8 + 2'12" 1-52 1-95 1'87 1.77 1 -66 Densitiee determined: Temperature . , , . . . . . , Density. .. .. . .. . . . . . . . ... 17 *lo 1.2620 Density. 1'2080 1.2617 1 *2198 1.2425 1.2500 1.2555 + 14.13" 9'68 12.88 12.10 11 '38 10'63 40" 57" 99" 1.2475 1 '2366 1 *2076 p = 9.906. 99" 17 72.6 85.8 57 52-7 17.2 12.1 7a + 3-11 2 '26 3 -04 3'00 2-94 2.86 2-80 2.25 2.14 Density. 1'2050 1'2600 1'2193 1 '2230 1.2277 1'2340 1.2365 1 *2601 1.2632 I Q 1:. + 13.01" 9-07 12.59 12-28 12.09 11-70 11 '43 9-01 a -55 Temperature.. .. .. 17.3 37' 57" 68.2" 99.5" Density , , . . . . . . , 1-2601 1.2474 1 *2338 1 '2263 1 '2044212 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE p = 23455. Temperature. 15.4" 100 75 68'8 64'9 55 28 '3 25 '8 13 10.5 8 aD (200 mm.). -I- 4.21" 6 '94 6-60 6 *44 6'35 5-99 5.00 4'87 4'18 4.03 3 *89 Density. 1 '2098 1'1944 1'2125 I '21 73 1.2200 1.2271 1'2460 1.2480 1.2566 1 -2580 1.2597 t.1:. + 7-42' 12'39 11-61 11.28 11-10 10-41 8.55 8 -32 7.09 6'83 6.58 Densities determined : Temperature ...... 8.5" 21.4" 45 '2" 60' 100" Density ... . . . ,. . . .. 1.2600 1 *2512 1'2344 1 '2238 1 -1944 p = 48.125. Temperature. 100" 79 70.5 65.8 51 #2 46.5 41.8 39 -1 24 6 *5 aD (200 mm.). Densities determined : +12.97" 12-14 11 -57 11.31 10.49 9 *60 9-12 7 *24 4 '92 a -89 Density.1'1749 1.1920 1'1990 1 '2030 1'2145 1'2187 1 '2222 1 '2243 1'2368 1.2607 rai:. + 11 '47 10.58 10.03 9 -77 8 -98 8 -18 7-76 7 -54 6.08 4 *09 Temperature . . . . . . I 0" 36.3" 55" 7 0" 100" Density . . . . . . . . . . . . 1'2480 1'2269 1'2116 1'1993 1 -1749ROTATION OF OPTICALLY ACTIVE COMPOUNDS. I. 213 Temperatcre. p = 69.93. C Z ~ (200 mm.). 98.5" 16 78'4 70.7 60-8 48.5 40.6 33 '9 19 Densities determined : + 18.82" 7.91 17-48 16.62 15-37 14.17 13-04 11-87 8 -78 Density. 1-1592 1'2310 1.1763 1.1832 1'1920 1.2028 1'2097 1-2153 1'2287 Temperature ............ 19" 45" 59.5" 80" Density . . . , . . . , , .,. , . , ., , 1.2289 1.2059 1.1932 1'1752 + 11-61" 4.59 10.63 10-03 9 *22 8'42 7-71 6.98 5.11 97" 100" 1'1592 1'1575 p = 89.98. Temperature. 98.5" 15.7 a3 77.8 70'2 66 '1 62.5 49'5 37.9 32.2 13 a, (200 mm.). + 25 *64" 11-72 24'49 24-03 23.01 22 *41 21.83 20.17 17-77 ' 16'73 11-25 ~ Density. 1'1392 1-2190 1.1540 1'1589 1'1665 1.1702 1,1736 1,1865 1.1975 1.2030 1.2212 + 12.50" 5 -34 11-79 11 -52 10.95 10.64 10'34 9 -44 8-25 7 *73 5'12 78" 100 Densities determined : Temperature ..... 8" 17" 35" 53" 72" Density ............ 1.2271 1.2178 1'2004 1.1828 1.1643 1.1582 l.1377 VOL. LXXIX. Q214 PATTERSON: THE INFLUENCE OF SOLVENTS ON THE Ezperirnent with Rochelle Salt, C,H4O,NaK,4H,O. p = 13.686. Temperature. a, (400 mm.). 15.6" 58.9 34 14.8 99'2 14 20" ? S 9 9 9 ) i b , +12-737" 12.834 12.905 12.772 12.482 12.760 5 10 25 50 75 10 Density. 1.0711 1'0521 1.0640 1.0715 1.0273 1-0719 + 21 '72" 22'28 22-15 21.77 22-34 21 '74 Densities determined : Temperature. 15.6" 30.8' 55" 70.1" 14.7" 98.8" Density ...... 1*0710 1'0654 1'0541 1.0457 1,0714 1.0269 Molacular-sol~tion-volume of Ethyl Tartrate in various ,Solvents. M=206. Molecular volume of ethyl tartrate at 20" = 206/1'2053 = 170.91 = +7*67. I t. Water : 20" I 9 9 9 a ? 9 ) ib 2 '5 5 10 24.954 49 '993 74'99 10 ~~~~~ d. 1 -0041 1'0100 1.0216 1.0574 1.1153 1.1673 1.0245 0.8070 0'8229 0.8741 0,9703 1'0824 0'8327 6. M. S. V. 158 '25 159.06 159.44 160.05 163.05 166.50 156.15 159.43 160'68 161'83 164.31 166'97 159'24 ~~ [a]:. + 25-82'" 26-10 26-00 23.60 17.33 11.70 26'30 11 -50 11-48 11.20 10.50 9-12 10'60 * This value is probably rather low.ROTATION OF OPTICALLY ACTIVE COMPOUNDS. I. 215 Molecular-solution-volume of E8hyl Tartrate, &c. (continued). 20" 9 9 9 , 2 1 ib I p. t . 5 10.94 20 40 60.01 10.94 Ethylialcohol : rt-Propyl alcohol : 20" 9 , 9 ) 11 3 , I , 1 , 1 , ib Glycerol : 20" 9 , 1 , 9 , 1 ) ib 15 2 -5 5 7 :713 10 17'507 25 37'51 49.834 74'99 5 4 -98 9.906 23-45 48.125 69-93 89.98 9'906 9'906 d. 0.8061 0 '8240 0.8532 0,9240 1,0054 0.8328 0 '8 11 3" 0-8183 0 '826 1 * 0'8339 0.8561 0.8790 0'7222 0.9677 1-0757 0.8267 1'2600 1 *2581 1-2521 1.2399 1.2277 1.2149 1.2649 1.2614 6. 0,7912 9 , 1 1 1 1 o - ~ b o o 0.8043 0*%55t 0.8043 1 J 1 ) $ 9 9 , 0*&27 1 '2608 1 7 9 1 9 1 9 1.6365 1 '2638 M. S. v. 163.57 165.89 165'93 166730 167.90 164.74 167.30 168.10 168-27 168-51 167.80 169-01 168.76 169.56 169.95 167-27 165.44 166'99 168.84 169.17 169-84 170-24 164.91 166.33 8.82 8 -57 8 -30 7-90 7 -55 7 '60 7 *30 7 *30 7 -38 6.97 7.00 6 '73 6-59 6 *41 6'54 5 -80 9 -97 9.30 7 *80 5-62 5 -19 6 -05 8.28 8 '82 * These are not experimental values. They have been obtained by interpolation from a density-concentration curve constructed from the other figures. This is rendered necessary by the fact that in so dilute a solution 3s one of p=2*5 a very slight error in density makes a very large error in M.S.V. The experimental values are, forp=2.5, d=0*8118, andp=7*713, d=0*8265. j- This experiment was done much later than the others, and when the alcohol used had probably absorbed some moisture. This research is at present being extended in order that the effect of a considerable number of other solvents, not only upon the rotation of ethyl tartrate, but upon those of other active substances as well, Q 221 6 MELLOR: ON THE UNION OF may be ascertained, and in order t o determine whether the ideas sug- gested here can be further developed so as to explain existing difficultiers and discover new regularities. YORXSHIRE COLLEGE, LEEDS.