1884 PATTERSON AND KAYE : CLXXXII1.-Studies in Optical Superposition. Part II: By THOMAS STEWART PATTER~ON and JOHN KAYE. IN the first part of this investigation (Patterson and Taylor, Trans., 1905, 87, 33, 122), di-Z-menthyl d-tartrate and its diacetyl derivative were described and data were given for their rotations, as well as for that of menthol, both in the homogeneous state and in solution. The present communication deals with the corresponding I-menthyl deriv- atives of Ltartaric acid.STUDIES IN OPTICAL SUPERPOSITION. PART 11. 1885 I n the first part, however, the chief problem for the elucidation of which the investigation was undertaken was not discussed, because the data secured were insufficient to supply a definite answer. The data which we now publish, whilst having, as in the former case, an interest of their own, are still insufficient to decide the main question at issue, but in consequence of the appearance of a paper by Rosanoff (J.dmer. Chent. Xoc., 1906, 28, 525) we will state briefly here the ideas which we wish ultimately to test. The rotation of a compound built up of two active radicles, for example, d-amyl Z-lactate, may be regarded as the sum of two rotations, say ao and p", one contributed by each radicle. van't Hoff is responsible for the assumption, very reasonable at the time it was made, that the rotation of Z-amyl Z-lactate may be represented as equal to -a'+pO, a quantity obtained by merely reversing the sign of rotation of the amyl radicle. This assumption is known as the principle of optical superposition, and a considerable amount of experi- mental work has been undertaken by Guye and by Walden with the object of proving it.A consideration of this work convinced us, firstly, that the experi- ments of Guye and Walden have in reality no bearing on the matter in hand, and, secondly, that the validity of van't Hoff's assumption is distinctly open to doubt.. According to a reference in Meyer and Jacobson's Lehrbuch (vol. i, 2nd edition, p. l04), Urban (Arch. Pharrn., 1904, 242, 51) has arrived at a similar conclusion, but unfortunately we have not been able to consult the original memoir. Lowry in a recent paper (Trans., 1906, 89, 1039) also calls attention to a case in which van't Hoff's assump- tion is contradicted. We may examine briefly an example of Walden's results.He prepared, for instance, three amyl lactates, and determined their rotations with the following results : C.Iw [MID. I. +Amy1 1-lactate .................. - 6-38" - 10.21" 11. 2-Amy1 i-lactate.. ................ +2-64 +4*22 111. 2-Amy1 Z-lactate .................. - 3.93 - 6.29 Since the sum of the molecular rotations of I and I1 ( - 5-99') is nearly the same as the rotation of I11 ( - 6 * 2 9 O ) , andfrom a number of other similar series of figures, Walden concludes that "die lande- siibliche Auffassung von der algebraischen Superposition der optischen Eigenschaf ten verschiedener asymmetrischen Kohlenstoffatome in einer Molekul findet ihre Bestatigung " (Zeit. physikal. Chern., 1895, 17, 724). That is, the experiment quoted above is taken as proving that the rotation of the active lactyl radicle in I is the same as in 111, the * The paper has recently been republished in Zeit.phpsikad. Chent. 1906, 56, 566. 6 ~ 21886 PATTERSON AND KAYE: rotation of the active amyl radicle in I1 is the same as in 111, and that therefore the rotat.ion due to the lactyl radicle is the same whether it be combined with a laevo- or a dextro-amyl radicle. It may be said at once that the mistake made here lies in regarding i-amyl and i-lactyl as simple radicles. The substances I and I1 above are not homogeneous compounds.* When Z-lactic acid acts on i-amyl alcohol, two substances are formed-we may assume for the moment that they are formed in equal proportions-so that if small letters represent amyl radicles and capital letters represent lactyl radicles, we have Therefore, in supposing the rotation of I to be that of i-amyl Z-lactate, Walden makes the tacit assumption that Z-amyl and d-amyl have the same effect on the rotation of Z-lactgl, the very point which was under investigation.Further, of the four compounds composing I and 11, d - L forming half of I and Z-D forming half of I1 are enantiomorphs, so that their rotations would be equal and opposite, say, p” and - p”, whilst the other half of I is the same as the other half of 11. If, then, the pure compound Z - L has a rotation equal to a’ for a given length of tube, it is obvious that the rotation of I for the Same length of tube will be equal to $ao+ ip”, whilst that of I1 will be equal to &ao - 8p9 their sum being, of course, a’, the rotation of the pure compound Z -L.What Walden has done in these cases, therefore, is merely to measure the rotation of Z- L in two different ways, and it is not surprising that the results agree fairly closely. The experiments have been carried out in such a manner as to eliminate in each case the influence they were intended to discover. There is, indeed, so far as we are aware, only one recorded instance in which this question can be directly tested, and, curiously enough, thedata are supplied by Walden himself, who, however, makes no reference to their bearing on this particular subject. I n a paper, “Ueber die optiache Drehung stereoisomerer Verbindungen ” (Zeit. p h y s i k d . Chem., 1896, 20, 377), he gives values for the rotations of the di-Z-umyl esters of racemic and i-tartaric acid, comparing them with the values for the rotations of the di-Z-amyl esters of maleic and fumaric acid, Now the Z-amyl ester of racemic acid will, of course, connist of two non-enantiomorphic substances, and if we assume definite rotation * This was clearlyseen by Frankland and Price in the analogous case of the amyl glycerates (Trans., 1897, 71, 267).These authors, however, were not thereby led to doubt the validity of van’t Hoff’s assumption. On the contrary, they accept it as the basis of some of their arguments.STUDIES IN OPTICAL SUPERPOSITION. PART 11. 1887 values for the different active carbon atoms in these compounds according to the idea of van't Hoff, Guye, and Walden, we shall have, if R = C,H,,*O*CO- Di-2-amyl i-tartaric ester.Di-Z-amyl racemic ester. / I. (deztro-acid? 11. (kevo-ac?d). 111. (2u" + 28"). (2u" - 28"). (2a"). Obviously the rotation of the racemic ester, since it is supposed to consist of equal parts of I and 11, should be - - 2ao+ 2PO - 2 + 2 2ao - 2PO = 2ao, that is, should be the same as the rotation of the i-tartaric ester. Walden's figures are : [MI?. Di-Z-amyl i-tartrate ..................... + 13.83" Di-2-amyl &-tartrate .................. + 9 77 Between these numbers there is a difference of 4.06°, almost half the rotation of the racemic ester, and much greater than the difference (0.3") i n the case of the amyl lactic esters already mentioned. These data of Guye and Walden are not, however, experimentally sound.I n all cases, as has been pointed out, the active compounds the rotations of which have been determined were mixtures, and herein lies a possible source of error. To take the last example quoted, Wben pure Z-amyl alcohol acts on racemic acid, it is probable that the d- and Z-acids do not esterify a t the same rate, and since in the preparation of an ester the esterification is seldom or never complete, it is by no means unlikely that after distiIIation the resultant ester does not contain equal proportions of the d- and Z-acid radicles. Further, the amyl alcohol used by Walde; was itself also a mixture, which renders the experiment still more complicated and unreliable. The esterification of the i-acid would, of course, only be affected by the latter cause. These considerations apply, however, to all the experiments with Z-amyl alcohol and &acids, and therefore it might be expected that the experimental error in each series would be much the same.This is the case in all except the experiments with i-tartaric and racemic acid. The difference in the rotations of the esters of these acids can hardly be ascribed entirely to experimental error, and that1888 PATTERSON AND KAYE: t h i s difference has been found goes, we think, some way towards dispioving van’t Hoff’s assumption. In Rosanoff’s paper, already referred to, the line of argument is exactly similar to ours,* and he arrives at the same conclusion, namely, that the principle ” of optical superposition is, at least, very doubtful. He also suggests that in the experiments of Walden and of Guye solvent influence may exercise a disturbing effect.That this may be so is possible, but we are inclined to think that any solvent action which may come into play is likely to be of much less consequence than the possibility of selective esterification, mentioned above, which Rosanoff has overlooked. Rosanoff, in his paper (p. 529), in discussing a case of supposed optical superposition, says, “ On the other hand, cases like Landolt’s, if general instead of exceptional, would lead, not to the principle of optical superposition, but to the theorem that the rotabory power of an active radicle is independent of the chemical composition of the rest of the molecule,” and to this passage he adds a footnote in the words: ‘‘ Patterson and Taylor (Trans., 1905, 87, 33) seem t o think that this is really what is meant by optical superposition.” It is difficult to understand how Rosanoff could have fallen into this error, since in the second paragraph of the paper to which he refers there occurs the passage, “ When in a simple active molecule, such as that of lactic acid, the replaceable hydrogen atoms are substituted by radicles like methyl and ethyl or scetyl and benzoyl, the change in rotation which occurs with each substitution is probably due, not merely to the addition of a new group, but also to a modification, a slight molecular rearrangement of the active radicle itself.That is, tbe lacfyl radicle, supposing it could be detached from a molecule of methyl lactate without suffering any other change, would show, when examined polarimetrically, a rotation differing from that of a lactyl radiclo separated, in the same manner, from a molecule of some other lactate.” The idea suggested here is surely the antithesis of the conception that the rotatory power of an active radicle is independent of the chemical composition of the rest of the molecule ! Rosanoff‘s paper contains no new experimental work, but some is promised, and, therefore, to prevent any possible overlapping we may state that the next part of this investigation will deal with the menthyl esters of i-tartaric acid, and we hope then to obtain data * The foregoing, except for the references to more recent work, was written, in a more extended form, more than three years ago, and was intended to serve as an introduction to Part I of this investigation.For the reason given on p. 1885, however, it was omitted from that paper. One other point we may refer to.STUDIES IN OPTICAL SWPERPOSITION. PART 11. 1889 which shall supply a satisfactory solution to the problem of opticaI superposition. EX PER I M E NTA L. The sodium ammonium I-tartrate used in this investigation was That the salt was pure is shown by the following figures : 2.5 grams made up to 50 C.C. with water gave in a 2-dcm. tube the prepared by us from racemic acid. rotations : t. a,. t. OD. 16.3" -4.650" I 20'1" - 4.670" from which cty - 4,669. from the dextro-acid, gave at the same concentration the figures : A recrystallised specimen of sodium ammonium d-tartrate, prepared t . a,. 13-9" + 4-62' 16 '9 4-634 t.%I. 17'2' + 4-640" 22 *5 4.669 from which ar +4.655. Ri-1-menthyl-1-tartrate. -For the preparation of this substance it is not necessary to isolate I-tartaric acid from the sodium ammonium salt. The process used was as follows : some sodium ammonium tartrate was thoroughly dehydrated by heating in the steam bath for three to four hours. Fifty grams of the salt were placed in a round-bottomed flask with 158 grams of menthol, and dry hydrogen chloride was passed into the mixture, first in the cold and then at 110-130' for about twelve hours. The liquid mixture in the Bask was then separated from the chlorides of sodium and ammonium formed, by pouring off while hot into a distilling flask. Most of the menthol was then distilled off under diminished pressure and steam blown through the viscid residue to remove the last traces.The ester in the flask was extracted with ether in which it did not seem very soluble, heating being necessary. A viscid brown solution was thus obtained, and on adding a small quantity of sodium carbonate solution (a few drops) a copious white precipitate separated. This, which proved to be sodium Z-menthyl- Z-tartrate, was filtered off and examined later. The ethereal extract was washed with sodium carbonate solution, then with water, and dried over anhydrous sodium sulphate. After removing the ether an attempt was made to distil the menthyl tartrate, but without success, as decomposition occurred. Attempts were also made to crystallise the ester after it had been purified by boiling, in ethyl alcoholic solution, with animal charcoal and then precipitating with water, but these were for a long time unsuccessful.Finally, we found that if the ester was dissolved in pyridirie and water added, a crystalline substance1890 PATTERSON AND KAYE: Eepsrated out. This compound, which contains pyridine, was re- crystallised twice from light petroleum, when it melted at 69--70°.* We then found that if the menthyl tartrate was dissolved in light petroleum and a crystal of the above substance added, menthyl tartrate crystallised in fine needles melting at 42'. It was recrystallised from light petroleum, when i t melted a t the same temperature as before, and on analysis 0.1900 gave 0.4719 CO, and 0.1711 H20. C2,H,,0, requires H = 9.86 ; C = 67.61 per cent.We made two attempts to determine the rotation of this ester, one with a specimen prepared before we had obtained any in the crystalline condition, and therefore purified by precipitation with water from ethyl alcoholic solution, and another with the crystallised substance. Since in both cases we observed the same behaviour we need quote only from the latter. The substance was melted and poured into a 30 mm. tube. It did not crystallise on cooling, so the first rotation was taken at t : 16*3', when uD - 23.86'. On raising the tempemture the rotation increised, and at 134.9' had the value uD - 24.12'. On the following morning, however, the rotation had not returned to the original value, but a t 18*4', uD - 24-31', and a day or two later a t 130°, uD - 24.51'.On cooling again to 9*5', aD had become -24.69', and when heated to 98.2', uD -24.83'. The last observation made at 9.3' gave uD - 24*77', and at 99.3' uD had the value - 24-84'. These observations extended over about a fortnight ; they show no constancy, but to what this is due we cannot explain. The density of the menthyl tartrate was also determined twice. The results agreed closely, as is shown by the following figures: Seriee I was carried out with an oil purified by precipitation, Series I1 with the crystallised substance. C= 6'7.73 ; H = 10.00. Series I. Series 11. t. t. Density. - 71.5" 1*0028 79 0'9968 94" - 0.9866 112 0'9725 8eries I. Series 11. t 1. Density. - 136" 0'9536 - 0'9541 - 154 0'9396 137" Obviously, specific and molecular rotations deduced from the above values are of little importance, but the following numbers for the extremes at low and at high temperatures may be given :- to.a: (30 mm.). Density. [a15 w1;. - 23.86" 1 -0450 - 76.11" - 324 '2" { l;:: 24.77 1 -0505 78.60 334-8 24-13 0.9547 84'22 358'8 { I;:; 24'84 0'9816 84.36 859.4 * This substance will be more fully investigated later.STUDIES IN OPTICAL SUPERPOSITION. PART XI. 1891 Sodium 1-Menthyl 1-Tartrate.-As already mentioned (p. 1889), a sub- stance was precipitated in rather a curious manner from the ethereal extract of crude menthyl tartrate on the addition of a drop or two of sodium carbonate solution. This compound was not very soluble in hot water. On cooling, the solution became turbid, and, on standing, clusters of needle-shaped crystals separated and the solution became quite clear.The substance was recrystallised from water. It did not melt when heated to 200O. On igniting with sulphuric acid : 1.0091 gave 0.208 Na2S04. Na = 6.68. 0.7013 ,, 0.147 Na2S04. Na = 6.79. 0.3480, dried at loo', lost 0.0191 H20. The rotation of the compound was determined in aqueous solution H20=5*48. C,,H2,0,Na,H20 requires Na = 7.01 ; H,O = 5.48 per cent. with the following result : P. to. a: (170 mm.). Density. [a]:. [MI:* 0.3623 48.6" - 0.466" 0.9896 - 76-46" - 250'8" Menthyl diacetyl-1-tartrate was prepared by boiling menthyl tartrate with excess of acetyl chloride for several ho;xrs. The residue, after the acetyl chloride had been distilled off, was washed with water and sodium carbonate Bolution, when it became solid.The compound, when dry, mas dissolved in hot methyl alcohol and boiled under a reflux condenser with animal charcoal, the solution filtered and allowed to cool. The crystals which separated were recrystallised from aqueous methyl alcohol, when they melted at 102.5'. On analysis : 0.1947 gave 0.4705 CO, and 0.1558 H,O. C = 65.90 ; H = S.89. 02136 ,, 0.5164 CO, ,, 0.1711 H,O. C=65.93; H=8.90. C,,H,,O, requires C = 65 a88 ; H = 9.02 per cent. The molecular weight of the compound was determined cryoscopically in beuzene solution (K=60), the following data being obtained: Theoretical M. W. = 510. Grams or substance per Weight of Weight of 100 grams of substance. solvent. solvent. A. M. W. 0.1903 gram 16-65 grams 1'14 0~11" 519.5 0.2863 ,, 16-65 ,, 1.72 0.17 505.7 0.4466 ,, 16.65 ,, 2-68 0 -28 478'9 1-0580 ,, 17'15 ,, 6.17 0.67 460'4 I n this case, therefore, as was also found for the corresponding derivative of d-tartaric acid (Patterson and Taylor, Trans., 1905, 87, 40), the molecular weight diminishes with increasing con- centration, but the substance seems to be unimolecular i n very dilute solution.1892 PATTERSON AND KAYE: The rotation of the substance in the homogeneous condition was then determined with the following result.The ester remained super- cooled for a long time, and therefore the observations could be extended over a wide range of temperature. The numbers are recorded in the order in which they were obtained. Rotation of Di-l-rnmtlbyl Diacet yl-l-tmtrate. to. 122%" 129.0 146-5 1045 15.0 49 -3 103.0 150.0 16.0 uz (30 rum.).- 22.135" 22.080 22.020 22'195 22.237 22,204 22.166 22.042 22'242 Densities determined : to ......... 98" d . . .......... 0'9894 Density. 0.9697 - 0.9646 0'9505 0'9842 1-0558 1.0283 0,9853 0.9477 1'0550 blk". * 76.10" 76.31 77.23 75.18 70.21 71-98 75-00 77'54 70'28 122" 142" 0'9701 0.9541 [MI:. - 388.1" 389'2 393.9 383.4 358'1 367.2 382.5 395-4 358.4 There is no sign here that any permanent alteration of rotation had occurred. The rotations of Z-menthyl Z-tartrate, and Z-menthyl diacetyl l-tartrate were then determined in ethyl alcohol, benzene, and nitrobenzene with the following results : 1- MenthyZ LTartmte. Solvent : Ethyl Alcohol. 2 3 9 to. Density. a: (170 mm.). I. 2.41729 14.25" 0'8009 - 2.490" 27.6 07893 2.491 38'9 0.7779 2,493 20'0 * - - 11.7.05393 18.0" 0.8055 - 7 '29" 24 -6 0.8000 7-30 48'0 0-7800 7 -30 20.0 * - - Deneities determined : to. d. 1". I. 19.56" 0'79615 11. 18.25" 23-22 0*79305 22-95 29 '8 0 78741 38 -6 Interpolated. [a]:. 75-64' 76.77 77.90 76.06 * 7548" 76.10 78.05 75.60 [ M I 5 - 321 '7" 327-0 331.8 324-0 * 324'2 332.5 322.2 * - 321.5" d. 0-8051 6 0*80136 0.78766STUDIES IN OPTICAL SUPERPOSITION. PART 11. 1893 I-Menthyl 1- Tartrate (continued). Solvent : Benzene. P. to. Density. I. 2.73118 18.3" 0'8826 31.0 0.8690 38.0 0.861 5 20.0 I 11. 5-39447 15.1" 0.8890 24.5 0'8790 35.9 0.8670 20.0 * - Densities determined : t". d. I. 18-5" 0.88196 21 *5 0 *87 9 10 27 -0 0.87342 as (170 mm.). [a]:. [MI$ - 3'001" - 73.22" - 311.9" 3'000 74.35 316.5 3-000 75-00 319.5 - 73-33 * 312.4 - 6.083" - 74.57" - 317'7" 6-071 75.30 320'8 6,065 76.34 325 *2 - 74.94 * 319.2 ' to.d. 11. 17.75" 0.88591 21 -60 0 -88202 30.15 0.87303 Solvent : Nitrobenzene. 1. P. to. Density. a: (170 mm.). [a]:. [MI:. 1.98411 13-8" 1.2047 - 3.610" - 88'84" - 378.5" 20 -2 1 *1983 3.600 41 -0 1.1775 3.528 51 5 1.1667 3'483 13.6 1 '2049 3.619 20'0 * - - 25 *3 1.1857 9.610 31 -3 1-1797 9-529 42.0 1'1692 9,482 20.0 * - - 11. 5.34773 13-5" 1 *1975 - 9 -727" Densities determined : to. a. to. I. 17-85" 1 *20083 11. 17.50" 20.80 1.19783 29.75 29 *30 1 *la933 43'50 l-MentA yl Diacet yl-l- tartrate. Solvent : Ethyl Alcohol. I. 11. P* to. Density. a: (1 70 mm. ). 89-06 379.4 a8-87 378.5 88'43 376.7 89.08 379.5 89-02 * 379.2 * 89-15 379.7 88-86 378.5 89.19 379.9 89.32 * 380.5 * 89-37' - 380.70 d.1.19337 1.18149 1.1679 3.9479 20.2" 32.6 42.0 20.0 * 5.80053 16.7" 24 -8 38 -0 20'0 * 0.7985 - 3-a7i0 - 72.210 o w 9 3.874 73-18 0.7795 3.874 73.97 0.8053 -5.675" -71.52" 0.7953 5.675 72.29 0.7854 5,680 73-34 * Interpolated. - - 72-20 * - 71.77 * - - 368'3" 373.3 375.2 368'2 * 368.7 374.0 366.0 * - 364'7"1894 PATTERSON AND KAYE: l-Menthyl DiacetyZ-1-tartrate (continued). Densities determined : to. d. to. d. I. 21.0" 0.79799 11. 19'1" 0.80333 25.5 0'79407 24 -0 0.79895 39.4 0-78200 37.2 0'78784 Solvent : Benzene. P. to. I. 2.06209 2lmO" 31-5 15.0 20.0 11. 5.21901 15.3" 27 -5 20.0 i t Density. a$ (170 mm.). [u]:. 0.8795 - 1.885" - 60-97" 0.8685 1.895 62-08 0.8860 1'895 60.85 - 61-17 * I 0.8905 - 4.822" - 61~01" 0.8775 4-827 62'01 - - 61'37 * [MI:. - 310'9" 316'6 310.3 312.0 * 316.2 313.0 * - 311.1" Dmitiss determined : to. d .to. d. I. 17.75" 0 -883 1 9 11. 17.85" 0.88763 21.25 0.87941 20.87 0:8a450 30.50 0.86930 30.10 0.87483 Solvent : Nitrobenzene. P. to. I. 2.59034 16.8" 28 -8 58 -1 20.0 .!+ 11. 5'35741 15 -3' 29.8 44.0 20-0 * Density. af (170 mm.). [u];. 1.2016 - 3'710" - 70.09" 1.1898 3.698 70.56 1.1610 3'666 71.68 - - 70.06 * 1.1985 - 7.570" - 69.36" 1'1842 7,568 70.18 1'1704 7.546 70.80 - c 69-54 * m1:* - 357'5" 359-8 365.5 357.3 * - 353 -7 357.9 361 -0 354.7 * Dernsities determirzecl: to. d. to. d. I. 16.66" 1.20187 11. 17.75" 1 '19594 22.80 1.19579 21 -42 1.19231 32.80 1.18594 30.65 1.1833 48.4 1-1661 * Interpolated. The table below is a synopsis of these data. I n it are given the rotation values (by interpolation) of the rarioiis active compounds examined in 5 per cent.solutions in ethyl alcohol, benzene, and nitro- benzene. We think it better to give these values rather than numbers for infinitely dilute solution, since, perhaps, for purposes of comparisonSTUDIES IN OPTICAL SUPERPOSITION. PART IT. 1895 it is preferable to deal with concentrations which can be practically realised. Z-Menthyl d-tartrate. I-Menthyl I-tartrate. Condition. w1y. CMI-20". A. Homogeneous ..................... - 284.0" - 325 -0" ( 1 ) 41 '0" In ethyl alcohol (5 per cent.) ... 305 '1 323-0 17-9 , , benzene J Y 293.6 318-4 24'8 , , nitrobenzene , , ... 246.0 380.4 134'4 I- Men t h y 1 diace tyl- &Men t h yl diace tgl- Homogeneous .....................-256 4" - 359.6" 103.2" In ethyl alcohol (5 per cent.) ... 268-2 367-0 98'8 , , benzene ,, * * a 285.0 312-9 27 '9 ,, nitrobenzene ), ... 238.0 355.1 117.1 d-tartrate. E-tartrate. The chief point of interest which we find in these data is that the differences between the rotations of the corresponding d- and I-com- pounds, under these different conditions or in solution, are not constant. Thus, for example, solution in benzene increases the negative rotation of I-menthyl diacetyl-&tartrate by 2 8 * 6 O , whilst it diminishes that of I-menthyl diacetyl-I-tartrate by 46*7", so that whereas the difference in the rotations of the homogeneous compounds as given in the &column is 103*2O, in benzene solution it is orily 27.9'. I n ethyl alcohol solution the rotations of both acetyl tartrates are increased, but that of the dextro- to a greater extent than that of the laevo-compound. I n nitrobenzene the opposite is the case: both rotations are diminished, that of the dextro-compound to the greater extent.The behaviour in each of these three cases is different, and the fourth possibility is found in the behaviour of I-menthyl d- and I-tartrates in nitrobenzene, the rotation of the former being diminished by 3 8 O and that of the latter increased by about 5 5 O . It is not difficult to understand this if we assume that the solvent exerts its influence separately on each of the three active groups of which the molecule is composed, and that our measurements represent the sum of these effects. Thus the influence of a solvent on the menthyl part of I-menthyl diacetjl-d-tartrate will probably be the same, or much the same, as its action on the menthyl radicle of I-menthyl diacetyl-I-tartrate, whilst its influences on the tartary 1 parts of the molecules will probably be equal, or almost equal, but in opposite senses.If, then, the solvent influence is represented by a+P in one case, it will be represented by a - p (or nearly) in the other, and it will depend on the relative magnitudes of a and 6 for different solvents which of the four possible casos exemplified above shall occur. It is also possible to trace other approximately additive phenomena somewhat similar to those discussed in Part I. (pp. 38, 39) when we consider the change of rotation of these compounds with variation of1896 PATTERSON AND KAYE: temperature.at 100" and 0" are the following : The numbers for menthyl diacetyl d- and Z-tartrates Z-Menthyl I - Menth yl diacetyl-&tartrate. diacetyl-Z-tartrate. - -- to. [MI;. A. [MI;. A. - 381 '8" - 27 .a0 354.0 39 - 2 O 100" - 227 '6" 0 266'8 The rotation of the d-tartrate is diminished (that is, becomes less negative) by 39.2' on heating from 0' to 100'. This corresponds with the behaviour of ethyl diacetyl-d-tartrate the positive rotation of which increases on heating (Patterson and McCrae, Trans., 1900, 77, 1098). The rotation of the 2-menthyl diacetyl-&tartrate, on the other hand, increases by 27.8' (that is, becomes more negative). Since these changes are in opposite directions it seems reasonable to suppose that they are chiefly due to those parts of the molecules which are of opposite configurations, namely, the acetyl-tartaryl radicles, and that tho numerical difference between them, 11*4', represents twice the change of rotation of two menthyl radicles attached to an acetyl- tartarpl radicle on heating from 0' to 100'.Thus, of the total change in rotation (39.2O) of the dextro-compound, +33*5' are due to the acetyl-d-tartaryl part of the molecule, and +2*85' to each of the menthyl radicles, all these changes being in the same direction, whilst in the laevo-compound the total changeof rotation ( - 27.8) is made up of - 33.5, due to the acetyl-Z-tartaryl group, and + 5.7', due to the two menthyl groups. The independent effect of each active group in the molecule seems, therefore, to be, at least roughly, traceable, although, of course, the above argtiment ignores the possibility that the change in rotation of, say, a menthyl radicle might be different according to whether it were combined with a d- or Z-acetyl-tartaryl group.It is likely, however, that such an influence would be slight, which is borne out to some extent by the fact that the change deduced above for a menthyl radicle attached to an acetyl-tartaryl group (2.85') is in fair agree- ment with that actually observed in Z-menthyl acetate (2.3') which was discussed in Part I. (Trans., 1905, 87, 38). We arrive at a somewhat similar conclusion from an examination of the variation of density of these compounds with change of tempera- ture which the following data exhibit.STUDIES IN OPTICAL SUPERPOSITION.PART 11. 1897 Z-Menth yl d- tartrate. I-Men thyl 2-tartrate. 1 *0040 i::::; 0.0617 0'9430 - to. bensity. A. Density. A. 0*0610 70" 150 I-Menthyl I-Men thyl diacetyl-d-tartrate. diacetyl-1-tartrate. 0.0641 1*0118 Q.0640 0 '9477 70" 1'0202 - 150 0-9562 It will be seen from these numbers that the change of density due to rise of temperature from '70° to 150' has almost the same value for the two simple tartrates, namely, 0.0610 and 0,0617, and that the same holds for the acetyl derivatives for which the changes are 0.0641 and 0.0640, but the change for either of the simple tartrates is quite different from that for an acetyl derivative, compare for example 0.0610 and 0.0641. From this we may conclude that the I-menthyl group expands to the same extent on heating, whether it be combined with a d- or a Etartaryl radicle, and that the latter groups also expand to the same extent when combined with an I-menthyl radicle.All these changes must of course be in the same sense, and differ in this respect from the corresponding rotation changes. So far we have dealt with variation of density; we may now con- sider the actual densities of the substances examined. Taking values at looo we have 2-Menthyl d-tartrate 0.9922 I-Menthyl &tartrate 0'9811 o'olll I-Menthyl diaoetyl-1-tartrate 0.9877 0'0085 Between the densities of the simple esters there is a difference of 0.0111, and between those of the acetyl derivatives a difference of 0.0086. On the other hand, the difference in density between I-menthyl d-tar- trate and I-menthyl diacetyl-d-tartrate is only 0-0040, whilst the difference in density between the corresponding laevo-compounds is 0.0066.Thus there is a greater difermce in denaity between the two simple tartrates than exists between one of these and its acetyl derivative. The difference in density caused by mere spatial change of the groups composing the molecules is distinctly greater in both cases than that due t o a considerable difference of chemical composition. This difference between two compounds which differ only spatially is perhaps shown more clearly in their molecular volumes. Substance. M.V.1OO'. A* Substance. Density. A. Substance. Density. A. Z-Menthyl diacetyl-d-tartrate 0.9962 4.9 C.C. 4'4 0.c I-Menthyl d-tartrate ... .. . . . . . . . . . . 429'3 C.C. I-Menthyl 2-tartrate ... , , . . . , . . . . . , 434'2 , , E-Menthyl diacetyl-d-tartrate ... ... 512.0 C.C. 1-Menthyl diacetyl-2-tartrate ...... 516'4 ,,1898 PATTERSON AND KAYE: The volume change due to re-arrangement is greater in the smaller than in the larger molecule. We have also calculated, from our density determinations, valued for the solution volumes of these two compounds in the various solvents used. We should mention, however, that owing to scarcity of material the data were obtained (except in the case of Z-menthyl diacetyl- I-tartrate in nitrobenzene) with a pyknometer having a capacity of only 7 C.C. The values for M.S.V. are therefore probably not so accurate as those formerly given for the corresponding derivatives of d-tartaric acid.The values, a t both concentrations examined, for the diacetyl-Z-tartrate in ethyl alcohol and benzene agree fairly closely and therefore confirm each other. The values for the same substance in nitrobenzene are the most reliable of all, having been obtained with a larger pyknometer. Moleczctar Solution VoZumes. 1-Menthyl Ltartrate, [ M F = 325O (?). M.V.20" = 426/1*042 = 408.9 C.C. Ethyl alcohol (6 20"/4"=0'79045) ... 2'41729 0.79577 389*8* C.C. - 324.0" 7'05393 0.80393 410.8 ), 322.2 Benzene (6 20"/4"=0'87784) ......... 2.73118 0.88053 431.0 C.C. -312'4" 5'39447 0.88364 426.2 ,, 319'2 Nitrobenzene (6 20"/4"-1'20319) ... 1.98411 1.19864 380.6" C.C. - 379.2" 5.34773 1.19095 422'1 ,, 3805 Solvent. P. d. M.8.V.200. * These values are doubtful.Z-Menthyl diacetyl-Etartrate, [M]r = - 359.6'. M.V.20"= 510/1*O518 = 484.9 C.C. Pa d. M.S.V.20. Ethyl alcohol ...... 3.9479 0.79887 472.9 C.C. 5230053 0.80252 477.9 ,, Benzene ... ... .. . ... .. , 2'06209 0.88076 487.5 C.C. 5'21901 0'88542 485.7 )) Nitrobenzene ...... 2.59034 1.19856 487'1 C.C. 5 *35741 1.19371 486.7 ,, [MITO", - 368 '2' 366'0 - 312*0' 313.0 - 357.3" 354'7 These numbers do not exhibit any simple relationship between solu- tion volume and rotation, but from the table below it will be observed that in all cases except one the volume of the Z-tartrate in solution is greater than that of the d-tartrate, which is, as has already been remarked, also the case for the compounds in the homogeneous state.STUDIES IN OPl'lCAL ST;PERPOSITIOS.PART 11. 1899 1- Jienthyl I- tar tratc. Z- 3it.n thy1 tl- tar tra te. -' 7<7 7- Solve11t. $I. Ji.S.V."O". 21. i r . s . v ~ . A. Ethyl alcohol ...... 7.1 410.8 C.C. 7.9 398'3 C.C. 4-12.5 C.C. Eenzene ............ 5 . 4 426'2 ,, 7.4 409.5 ,} + l S * 7 ,, Nitrobenzene ...... 5.3 422.1 , , 6'5 408.2 ,, +13*9 ,, I-Menthyl I-Menthy 1 diacetyl-I-tartrate. diacetyl-&tartrate. Ethyl alcohol ..... 5.8 477.9 C.C. 7 - 3 480.0 C.C. -2.1 C . C . Eenzeiic ............ 5.2 485.7 7 7 7.8 482 5 ,, +3'2 5 , Nitrobenzene ..... 5.4 48ti.7 ,, 6 ' 3 483'0 ), 4-3'7 ,) That the molecular volumes or molecular solution-volumes of partial enantiomorphs, like those we have been dealing with, should be different is by no means remarkable. It is merely in agreement with the very well-known fact t h a t the solubility of compounds such, for instance, as cinchonine-d- and I-tartrate or mannose, glucose and galactose is different, and it is further in agreement with the fact that such substances differ also in melting point. The melting points of the substances examined in this investigation, for example, are as follows :- 84.5" * + Trdus., 1905, 87, 124-126. I-llenthyl d-tartrate ...... 74--75" I-Menthyl diacetyl-&tartrate ... { I-Meathyl I-tartrate ...... 42 I-Menthyl cliacetyl-Z-tartrate . . , 102 5 Prom the fact that when an active group AE combines respectively with two other enantiomorphic groups, B d and Bl, the changes in the reacting groups, which accompany combination, are not the same in the two cases, since the melting points, solubility and density of the resulting substances are different, and, in spite of the fact that, on the other hand, the variation of volume and, 60 far a3 we can judge, the variation of rotation of the group Al, with change of temperature seem to be independent of the configuration of the other group with which it is combined, it is scarcely to be expected that the rotation of such partially enantiomorphic substances should be a purely additive property. As has been said, however, there exist no unimpeachable data to settle the question, but we hope with the next part of this investigation t o produce some evidence of a decisive character. -"- The substance is climorplious. It gives us pleasure, in conclusion, to acknowledge our indebtedness to the Research Fund Committees of the Chemical Society and of t h e Royal Society investigation. THE UNIVERSITY, G LASGOW. for grants which defrayed the expenses of this VOL. LXXXIX. G I