1066 POPE AND PEACREY: By WILLIAM JACIiSON POPE and STAXLEY JOIIN PEACHEY. IN the present communication, we describe the separation of com- pensated tetrahydroquinaldine into its optically active components by means of Keychler's clestrocitlllphorsulphoaic acid ; we have previously discussed the reasons for using this acid in preference to tartaric acid (Trans., lS9S, 73, 893). The new method adopted, which is a generally applicable one, is based upon the following considerations. The solubilities of the salts (d13 tlA and ZB d A ) of a dextrorotntory acid ( d A ) with a dextro- and a lwo-base (tZR and ZB) mould hardly be expected to differ considerably, because the solubility is partly a func- tion of tho cheinicnl nature of the salts. If, however, the salt, ZB dA, is the less soluble ant1 only sufficient of the active acid, dA, necessary to the formation of this salt is added, the balance of acid required to dissolve the base befug made'up by adding the requisite amount of an optically inactive acid, such as hydrochloric acid, which forms corn-RESOLUTION OF TETRAHYDROQUINALDINE. 1067 paratively soluble salts with the base, it would be expected that on crystallisation the greater part of the lavo-base would separate as the sparingly soluble salt, ZB dA, whilst the mother liquors would retain the dextro-base of which the hydrochloride, dB HC1, is very soluble.The disadvantage of the ordinary method of separating externally compensated bases by cry stallisation with excess of tartaric acid lies very largely in the fact that the solubilities of the two salts, d B dd and ZBdA, are not suEciently different to permit of the pure salts being easily isolated by fractional crystallisation ; this is illustrated by Ladenburg’s observntion (Ber., 1894, 27, 77) that the crude dextro- tetrshydroqiiiiialdine dextrobitartrate obtained from the inactive base must be recrystsllised many times before it is obtained pure, By applying our method to this base (using one molecular proportion each of dextro-a-bromoctcmphorsulphonic acid and of hydrochloric acid for the dissolution of two molecular proportions of externally compensated base) and crystsllising the first separation twice from alcohol to remove mechanically retained impurities, a pure sample of Itwotetrahydro- quinaldine dextro-a-bromocamphorsulplionate is obtained.Although the sepczration mas best effectod in this WHY, instructive results were obtained by modifying the method of applying the principle above explained. Thus, hot aqueous solutions of one molecular proportion of ammon- ium dextro-a-bromocamphorsulphonste and of two inolecular proportions of racemic tetrahydroquinalcline hydrochloride were mixed and the liquid allowed to cool ; a large separation of lzevotetrahydroquinaldine dextro-~-bromocamphorsulpl~~nate was thus obtained in a practically pure condition, this being the least soluble salt whichcould be formed. A further great advantage of making use of the formation of a sparingly soluble salt of a strong acid in the resolution of externally compensated bases is found in the fact that such sparingly soluble salts are not decomposed by feeble acids.Thus, a fairly serviceable method of obtaining lavotetrahy droquinaldine dextro-a-bromocamphor- sulphonate consists in adding an aqueous solution of one molecular proportion of ammonium dextro-a-bromocamphorsulphonate to an acetic acid solution of two molecular proportions of the externally compensated base; the solution soon affords a copious deposit of crystalsof lsvotetrahydroquinaldinedextro-a-bi~o~oc~mphors~il~~~onate. I. LEVOTETRAHYDROQUINALDINE, L~votets.n~~ldPoz2Linalclilze DeztTo-a-bro~)1oca~~~p~~o~8su~~~~o~~(~CloH13N,Clo H,,BrO*SO,H. The most convenient method of preparing laevotetrahydroquinaldine dextro-a-bromocamphorsulphonate consists in mixing a hot con-1068 POPE AND PEACHEY: centrated solution of one molecular proportion of crude externally compensated tetrahydroyuinaldine hydrochloride with a hot concen- trated solution of one molecular proportion of the base in one mole- cular proportion of a concentrated dextro-a-bromocamphorsulphonic acid solution prepared as we have previously described (Trans,, 1898, '75, 895).By the time the solution has cooled to the ordinary tem- perature, the greater part of the salt of the laevo-base has crystallised ; the separation is filtered off, washed with dilute alcohol, and recrystal- lised from boiling absolute alcohol. After two recrystallisations, the substance is obtained as a mass of colourless needles melting at 323-225O. It is sparingly soluble in cold, and moderately so in hot alcohol or water; it is more soluble in glacial acetic acid, but nearly insoluble in acetone or ethylic acetate.The crystals deposited as the hot alcoholic solution cools are long, flattened needles with oblicluely placed end faces ; the extinction in the large flat face is straight with the sides, and through this face the bisectrix of a large optic axial angle emerges nearly normally. Tho double refraction is negative in sign and the optic axial dispersion is marked, the optic axial angle for red light being greater than that for blue ; the optic axial plane is parallel to the long edges of the crystals. No goniometrical measurements could be obtained. After melting the salt on a microscope slide under a cover slip and subsequently cooling quickly, it remains amorphous, and signs of crystallisation only appear a t the edges of the preparation after several days preserva- tion at the ordinary temperature.By alternately cooling and heating the liquid film several times, crystallisation may be started, and once started proceeds rapidly at temperatures not much below the melting point ; at tlie ordinary temperature, however, the crystallisstion proceeds with extreme slowness even after it has been started at a higher temperature. The crystalline film thus obtained consists of a radiate mass of long needles which crack across the direction of growth as cooling proceeds ; these needles are crystallographically identical with the crystals deposited from the alcoholic solution, and mostly show the optically negative bisectrix of a large optic axial angle emerging nearly normally to the plate.The following analytical results were obtained : 0.1969 gave 0.3793 CO, ancl 0.1104 H,O. 0.9654 ,, 25.9 C.C. of dry nitrogen at lGOand 749 mm. N=3*14. 0.6144 ,, 0.3534 AgBr and 0.3172 BaSO,. 5 r = 17.55 ; S = 7.09. C,oH2,0,NBrS requires C = 52.40 ; H = 6.11 ; N = 3.06 ; Br = 17.47 ; S = 6.99 per cent. The specific rotatory power of the salt was determined in absolute C = 52.54 ; H = 6.23. 0.1986 ,, 0.3843 CO, ,, 0,111 1 H,O. C = 52.77 ; H = 6-22.RESOT,UTION OF TETRAHYDROQUINAT.DIN E. 1069 alcohol and in glacial acetic acid solutions. A solution in glacial acetic acid of 0.5005 grain in 35.3 C.C. gave aU + 1-54' a t 31.3' in a 200 mm. tube, whence [a I,, + 38,s" a i d [Ill,, + 177.6". A solution i n absolute alcohol of 0.2005 grani in 25.2 C.C.gave a,+0*66° at 31' iii a 200 mm. tube, whence [ + 41.5" slid [ A1 J1, + 1W. Chef ully purified lrevotetr~tliydro~~iiinsldine dextro-a-brornocamplior- sulphonate is distilled in n current of steam with a slight excess of soda ; the base is obtained :is a colourless oil saspencled in the aqueous distillate and is extracted \I itli ether. The ethereal solution is dried with potash and the solvent distilled off; the residual pale yellow oil is then distilled uncle? reduced pressnrc. It passes over as a colourless oil boiling at 15s' under 59 inni. pressure, and is sparingly soluble in water, but miscible with the ordinary organic solvents ; it has a faint basic oclour and gridunlly beconies yellow on exposure to the air.The following nn:dgticul results were obtaiiied : 0.1Si6 gave 0.5593 CO, and 0.151 1 H20. 0.2977 ,, 34.6 C.C. of dry nitrogen a t 3loand 7GI mm. N=9.65. C = 81.31 ; I3 = 8.95. U,,H,,N requires C = 82-63 ; 13 = 8-84 ; N = 9.52 per cent, The base has a,-59*34' in n 100 r i m . tube a t 30'; whence [a],-5S.13° and [M]L,-S5*440 a t 20'. The relative density is 1.02365 at 14*5"/1". Further details as to the physical properties are given in a subsequent paper (this vol., p. 11 1 I), L c e t . o t e t , ~ * c c k ~ ( l ~ ~ o ~ u ~ ~ t ( ~ ~ ~ i ~ L e Hydvochloricle, C:,,,H,,N,IICl + H,O. A warm solution of piwe l ~ v o t e t ~ ~ l ~ y c l r o c ~ ~ u ~ i i ~ l ~ l i ~ ~ e in excess of concentrated hydrochloric acid yields on cooling a large separation of the hydrochloride as a white, crystalline pomdei- ; this is filtered on the pump? waslied with concentrated hydrochloric acid, in which the salt is moderately soluble, and recrystnllised from absolute alcohol.It melts a t 196*5--197*5'and is fairly soluble in water, less so in absolute alcohol, and sparingly soluble in acetone. The following analyses of the air-dried material show it to contain 1II,O which is lost a t 100" : Ob143S gave 0.1031 E20 and 0.3115 0,. 1,9828 lost 0.1765 H,O in 4 hours :xb 100'. C!,,H,,N,HC!l+ H,O requires C: = 59.5s ; H = 7.94 ; C1 r= 1'7.57 ; H,O = S.93 per cent. afterwards does not melt so sharply as before, C=59*55 ; €I= 7.94. 0.3162 ,) 03%S A.gUl. u1= 17.66. II,O=S*90. On prolonged heating at 1 OO', the material volatilises slowly, and VOL LXXV.4 c1070 POPE AND PEACHEY: The alcoholic solution, on spontaneous evaporation, deposits the hydrated salt. as large, transparent, lustrous tablets belonging to the orthorhombic system (Fig. 1). The form c{OOl) is dominant and gives very good results on measurement; the forms q{O11) and ~ { 1 0 1 ) are the next largest :tnd are well developed. The pyramid o(561) is always present and fairly large, but the dome d{501) is very rarely observed. There is a perfect cleavage parallel to ~ { l O l ) , and the optic axial iater- ference figure may be observed through a flake hacked parallel to b(010). The axis-b is the acute bisectrix, and the optic axial plane is af100) ; the optic axial angle is not large, and the double refraction is fairly strong and negative in sign.The optic axial dispersion is marked, the angle for red light being greater than that for blue. FIG. 1. FIG. 2, No evidence of the sphenoidal hernihedrisrn mas obtained. The form 0{561) is never completely developed, but usually only one half of its total number of faces is present; this, however, is apparently due to the crystal growing whilst resting upon the face c{OOT), which results in an imperfect development of those faces cutting tJhe negative end of the c-axis. Thus the crystals very usually present the appearance shown in Fig. 2, only the four faces (561), (561), (561), and (561) of the form o [56 1 being pisesent. We were unable to obtain any evidence of erantiomorphous hemihedrism by etching ; the action of water and alcohol upcn the faces of the pinacoid cjOOlf gives rise to etch- figures having the appearance shown in Figs.3 and 4 respectively, whilst upon a cleavage plate parallel to ~ ( 1 0 1 ) water produces etch- figures of the outline shown in Fig. 5. These figures appear t o be quite in accordance with holohedral symmetry. FIG. 3. tcM FIG, 4. up F I G . 5. Crystalline System.-Orthorhombic. a : b : c = 0 -8627 : 1 : 1 *4124. Forms observed.-c(OOl), p(O1 I}, r(101), ~'(501), and o(561).RESOLUTION OF TETRAHYDROQOINALDINE, The following angular measuremeiits were obtained : 1071 Anglc. cq =001 : 0 l i ‘14 = O l J : 011 c$J =001:011 cr = O O l : lo! 9’7’ = l O i : 101 C)’ =001: 101 C)” = 001 : 501 7.r’ = l 0 l : 501 w‘ = 101 : 501 ‘I/’ = 011 : 101 qo =011 :661 co =OOl: 661 co = O O l : 561 Number of observations. 37 14 12 41 16 24 70 11 14 6 18 15 7 Liini ts.54’19’- 55O11’ T O S - il 5 1.24 42 -12s 37 57 59-- 59 1 62 9 6:: 1s 121 4 ---l22 0 82 16 - 53 25 23 54 - 24 50 38 0 - 39 4 i d ii- 7’2 57 50 2 - 51 5 84 36 - 85 41 94 40 - 05 27 Mean observed. 5 i”42’ 70 40 165 15 5s 35 62 47 121 31 ti2 54 21 20 35 26 72 31 GO 35 85 4 94 59 Calculated. - iO”36, 126 18 6 2 50 121 25 83 2 24 27 35 23 i 2 28 50 40 85 9 94 51 - After cautionsly melting aiihydroiis I:-evotetrahydroqninaldine hydro- chloride under a cover slip 011 a microscope slide, it begins to crystallise quite readily a t about 60’, yielding a film which is macroscopically very transparent. As the temperature falls, however, the speed of growth decreases, and a t the ordinary temperature crystallisation proceeds so slowly that it cannot be followed microscopically, Further, the crystalline material obtained at temperatures not far removed from the melting point, consists of long needles or of large square plates showing straight extiiiction ; these are striated parallel to one side and the strice are the trace of the optic axial plane, n bisectrix of positive double refraction emerging normally to the large square face.The optic axes emerge outside the microscope field, and this material is of orthorhoinbic crystalline form. As the very transparent film cools, large patches of i t successively change with great rapidity, becoming very white and opaque; in a thick film, tho change is alinost explosive in character, sometimes even throwing the cover slip into the air.Notwithstanding this, it is very improbable that the change is due to polymorphism, as, although the material becomes so opaque, yet the optical properties can still be made out and are not appreciably altered; the rupture is apparently due to strain set up during cooling. At a little below the temperature at which the hydrochloride solidifies most readily, crystallisatioii still proceeds, although more slowly, the same modification being produced as at higher temperatures ; the crystallisation proceeds from ceutres just as before with production of radiate aggregates of long needles, showing well-defined extinction and optical properties. These aggre- gates, however, do not fly to pieces on cooling, but remain transparent, merely cracking across the longer dimensions of the needles. At the Q c 21072 POPE AND PEACHEY: In ailtieons solntion.ordinary temperature, the substance crystallises very sluggishly (about 1 mm. in 3 days), yielding a very opaque film showing aggregate polarisation ; this confused crystallisation is structurally identical with the broad needles produced so easily just below the melting point, because here and there in the mass transparent fragments may be observed having optical properties identical with those of the material produced a t the higher temperature, and also because no conversion of the one material into the other occurs on standing. Since crystallisa- tion only occurs readily at about GO', it is necessary, in order to obtain a crystalline film rapidly, to allow the hot molten film to cool until cryst.allieation starts from centres, and then to warm i t over the lamp until the whole has solidified.The rotatory power OF the hydrated salt, C!loI~$,IXCl + K,O, was determined in aqueous and absolute alcoholic solutions. 0,4932 gram, made up to 25.15 C.C. with water a t lS.9', gtve uD - 2.37' in a 200 mu. tube; 0.5451 gram, made ~p to 25.15 C.C. with absolute alcohol at 18*9', gave aD - 3.03' in a 200 mm. tube ; whence the following values : In alcoholic solution. C,,H,,N,HCl+ H,O ............... C,,H,,N,HCl ........................ C,,K,,N,HCI ........................ I I [a],, - 60.4" [ a ] , - 6 6 4 [MI, - 131.7 [a],) - 69'9" [a],, - 77.4 [MI,,- 140'8 Lcevotetral~~di.opui?zcclcliize Picrate, C,,T~,,N,C,H,(NO,),.OII. Laevotetrahydroquinaldine picrate is prepared by crystallising n mixture of the lxwo-base and picric acid in the requisite proportion from hot alcohol ; it forms dark yellow plates or needles melting at 148--150', but crystals suitable for goniometrical examination could not be obtained.It is sparingly soluble in water and moderately so in alcohol, acetone, benzene, or ethylic acetate. The following analytical results were obtained with material crystallised from absolute alcohol : 0-1693 gave 0.3145 CO, and 0.0670 H,O. C = 50.66 ; H = 4.39. 0,2146 ,, 0.4003 CO, ,, 0.0835 I€,O. C=50*S7; H=4*34. Cl,H,,07N, requires C! = 51.06 ; I€ = 4.26 per cent, A solution of 0.5014 gram, made with 25.3 C.C. of absolute alcohol at 20°, gave aD - 1.31' in a 200 mm. tube, wllence [.ID - 33.0' and [RI],, - 124'.Since the molecular rotatory power of hvotetrahydro. quinaldine in absolute alcohol is [bf]L, - 94*lo, it would appear that the picrate for the most part is not dissociated in alcoholic solution,RESOLUTION OF TETRAHYDROQUINALDINE. 1073 13enzoylZevotetrccl~ycEropzcinccltEine, C,,H,,N* COD C,H,. On suspending purified ltevotetrahydroquinnldine in warm caustic soda solution and running in rather more than the calculated quantity of benzoic chloride with continual agitation, an opaque, yellowish oil separates which rapidly solrclifies to LZ hard, crystalline mass ; this is ground in a mortar and filtered, being wall washed with water an3 dilute hydrochloric acid. The substance is best purified by crystsllisa- tion from acetone ; if, as sometimes happens, LZ green colouring matter is procluced, this is best destroyed by crystallisation from ethylic acetate.After ultimately crystnllising from absolute alcohol, the benzoyl derivative is obtained in colourless crystals melting at 11 7 - 5 4 1 So. It is very soluble in benzene and moderately soluble in cold alcohol, less so iu cold acetone or ethylic acetate ; it is nearly insoluble in light petroleum or boiling water. This substance is, as would be expected, appreciably more soluble than the corresponding racemic compound. The following resnlts were obtained on analysis : 0.1996 gave 0,5934 CO, and 0.1830 H,O. 0.6108 ,, 30.0 C.C. of clry nitrogen a t 21' and 7.55 mm. N = 5.69. C,,H,,OM requires C = 51.27 ; H = G.77 ; N = 5-58 per9 cent. I n order to obtain evidence as to the racemic nature of externally compensated benzoyltetrahydroqninaldine, it was necesswy to compare the densities and crystalline forms of the active and inactive sub- stances.The densities welie determined by Retgers' method (Zeit. ylqsill.cd. C'hem., 1889,3,497), using a solution of barium mercuric iodide diluted with water ; the results seem slightly more accurate than those obtained by Retgers with isomorphous mixtures, using organic liquids. The following results were obtained with crystals of benzoyllaevo- tetrahydroquinddine deposited from acetone or ethylia acetate solutions : C = 81.0s ; H = 6.84. 0.2066 ,, 0.6133 CO, ,, 0.12'76 H20. C=Sl.O-i; kI=G*86. c21~?'=1*2113 ; 1.2113 ; 1.2116 ; meau=1*3116 ; the molecular volume of the crystslline material is thus 207.1 6 at 14*5"/4".The crystallographic properties of the benzoyl derivatives of lavo- and dextro-tetrahgdroquinaldine were fully studied and compared ; the crystals of the two substances being enantiomorphously related, as was to be expected, the following crystallographic description includes both componnds. The opticrtlly active benzoyltetrahydro~uinaldi~ies are so soluble in benzene that good crystals could not be obtained from this solvent j1074 POPE AND PEACHEY: cold solutions in acetone or ethylic acetate, however, deposit on spon- taneous evaporation well-developed crystals suitable for goniometrical examination. The crystals are usually elongated in the direction of Pro. 7. the c-axis, and then a1-e generally developed o d y at one end; fre- quently, however, crystals o€ the typical habit shown in Pigs. 6 and '7 are obtained, and the examination of these crystals shows them to be FIG.8. FIG. 9. + 71" hemimorphic, the b-axis being polar. The pinacoids a(100) and b(010) are predominant, and generally of about the sitme facial development ; neither of these forms gives very good reflections. The form pw(il1) is usually fairly large, whilst the pyramid p o f l l l ) is quite small ;RESOLUTION OF TETRAHYDROQUINALDINE. 1075 crystals exhibiting the number of faces required in the holohedral di- vision of the monosymmetric system are often deposited from ethylic acetate solution, but the hemimorphism is generally betrayed by the unequal sizes of the faces, as in Figs. 8 and 9.There is a very perfect cleavage parallel t o cc[lOOj, and the acute bisectris emerges through a cleavage fragment a t the edge of the field ; observation of the estinctioii in b(010) shows that the acute bisectris lies in the plttiie of symuuuetry at 17.5” to the fnce-normal to ~~(100). Both optic axes are visible in convergent light uudor a l/l?lth inch oil immersion objective ; thc optic axial clispeilsion is marked, the optic asial angle for 13ed being greater thiill that for blue light. The double refraction is positive in sign, and the optic axial plane is perpendicular to the ylme of symnictry. Crystalline systcm.-~louohyiiillletric : Ileniimorpliic. CC: b : C = 1.0377 : 1 : 0.4361. p = 8S3 15’. Forms present on l ~ e n z o y l l ~ v o t e t r a l ~ y d r o ~ ~ ~ ~ i ~ i a l ~ ~ i n e (Figs.6 ixnd 8). -c(;lOO1, b;01O~,+pofllIf and +,u@i) ; sometimes also -pof11I) Pornrs present on beiizoyldeuti.otetrrLllyd~o~i~ii~:~ldii~e (Figs. 7 and 9).--~[100~, b10101, - - p o { l l l ; , -pzu[llli ; sometimes also + p o ( l l l ) -,,Lo;iii;. aacl + p‘”; 1 1 1 ;. The following angultir measurements were o1)tniiied : Azlgle. Calcnlat ctl . After melting the substnnce on a microscope slide un(ler n cover slip, tho liquid film can usually be cooled to the ordinary temperataro without any solidification occurring ; soinetiiiies, however, crystallisn- tion sets in whilst the film is very hot and then proceeds rapidly at the high temperature, but stops entirely when the slide cools. Tho crystallisation, having oiico st,arted, can be caused to proceed rapidly until complete by cautiously heating the film at ft temperature1076 POPE AND PEACHEY: below its melting point.The crystalline film consists of large, well- defined individual flakes; the larger faces of some of these are nearly perpendiculai- to the acute bisectrix of a fairly small optic axial anglo showing positive double ilefraction, but those of others are nearly perpendicular to tho obtuse bisectrix, in which the double refraction is apparently negative in sign. The optic axial dispersion is marked, the optic axial angle for red being greater than that for blue light. At the ordinary temperature, the liquid film solidifies very slowly indeed to a mass of interlaced needles showing aggregate polarisation ; some of these needles can be seen to lie perpendicularly to the optically negative bisectris of a large optic axial angle.Both the films obtained at high and low temperatures are probably structurally iden tics1 and also identical with the crystals deposited from solution. The rotation const:ints of benzoyllrc.votetrcihydroquindcline are of considerable interest, nticl show tlint the introduction of the acidic group has converted the 1,uro~otatory base into a highly dexti-o- rotatory compound. This is tlie more reinsrkn ble since the piperidine bases, dextro a-pipecoline, destroisopipecoline, coniine, and destroiso- coniine, have the specific rotatory powers [ + 369', + 33*29O, + 13*79", and + S.19" respectively, whilst their benzoyl clerivatives are also dextrorotatory and have tho values [.ID + 35.3?, + 33*35O, + 37*7", and + 29.1" respectively ; the introduction of the benzoyl group into the pipecolines scarcely nl tei-s the specific rotatory power (Ladenburg, Bey., 1893, 26, 854).A case somewhat similar to that of lxvo- tetrahydrocluinnldine and its benzoyl derivative has been iiivestigated by Forster (Trans., ISDS, 73, 386), who fiiuls that clextrobornylamine having [Ill,, + 69.6' yields a benzoyl derivative having the molecular rotatory power in alcoholic solution of [ill], - 56*0°, the molecular rotatory power thus changing by 126*63 ; in the case now recorded, a change of nearly 1000" in molecular rotatory power attends the conversion of the base into its benzoyl derivative. It is further of interest to note that, whilst the change of rotatory power occurs in the same mnke on passing from either bornylamine or neobornyl- amine to its hydrochloride and then to its benzoyl derivative, the direction of the cliange in rotntory power alters in the case of tetra- hydroquinaldine, as is shown in the following table : Rase ......................................Hydrocliloride in nlcohol.. .......... 9 , ,, water .............. Benzoyl derivative i n alcohol ...... I - 55'4" - 140's - 121.7 f 514 *7 Eoriivliuniiie, [MI,,. 4 69'6" $43'0 - 56 +O - Neobornyl- amine, [ h l ID. - 47'9O - 73.s - 114'8 -RESOLUTION OF TETRAHYDROQUINALDINE. 1077 The variation of the rotation constants of non-electrolytes with change of solvent has up to the present been but little stndied, its, with the exception of the work of Freundlor (dm.CJiim Z'Jqs., 1895, [ vii], 4, SSG), few results of theoretical importance have been derived from snch determinations. As, however, me sliow in a subsequent paper (this vol., p. 11 11) that vnlunble inforinntion concerning the strate of molecular aggregation of optically fictive substances is derivable from the variations in rotation constants referred to, a series of determinations of the specific rotatory powers of benzoyllzvotetrahydroqninsldine in vai*ious solvents hns been made. The results are stated in the following table, in which w denotes the weight of substance contained in v C.C. of solution a t teinpernture t, and c is the concentration i n grams per 100 C.C. of solution : Sol vent. lhnzcnc ............ ,, ............ ,, ............(!hl or0 f o m ......... A ce t one, .............. ICthyIic ncclate ... Acetic acid ........ Ethylic nlcoliol ... 11.. 0.2502 0 a . 5 0 0 7 1 .ooos 0.5010 0 . 4 99 3 0.5002 0 '5009 0'%000 2'. t. + 4.96' +9.9S + 19.93 4- 12.09 + 12.55 f 12'99 -?-1::.oo -t 14-41 [ J13 I,. + 623 '0" + 629.6 + 762.2 f 794'6 + 514.7 + 919.8 +911'1 + 630.8 IIydmlgsis qf 13eitxo!/llcevotP,i.l.rclr~cl~*oqzLii~~~~~l~~.--Siiice we proposed to prepare p r e clel;trotetraliydrocluin,zlcliiie by hydrolysing its benzoyl derivative, mid since nicemisation frequently accompanies chemical change, it mas desirable to ascertain whether benzoyl- hvotetrahydroquinaldine yields only the parent base on hydrolysis. The powdered benzoyl derivative WLS hydrolysecl by boiling for some days with concentrated Bydrocldoric acid ; after rendering alkaline with soda, and extracting with ether, the ethereal solution was washed with water m d evaporated to (zrynehs, hydrochloric acid being added towards the end.Tho crystidline hydrochloride was then ground up with acelone in an agate mortar, separated by filtration, and spread on a porous plate; it was colourless, and 0,5009 gram, dissolved in water and made np to 25.8 C.C. at 23*0", gave a,, - 3.63' in a SO0 mm. tube, whence [a], - 66.2'. This being the specific rotatory power of the Izevohydrochloride, i t is obvious that no rsceinisatioa attends the hydrolysis,1078 POPE AND PEACHEY: TI. EQUILIBRIA BETWEEN OPTICAL ISOMERIDES AND THEIR SOLUTIONS. EI?'CCCtZ.'O?LCd Crptcclliscition of C~zccle Dext~~otetrccl~yd~oquincckli~ze Dextrocnniplioi.su~~~onat z.Preliminary experiment,s showed that, on crystnllising externally compensated tet,rahydroquinaldiiie with destro-a-bromocamphor- sulphonic acid, the first separat,ion of 'crystals yielded a hvorotntory base on t,reatinent wit.11 alktdi ; similarly, it w:ts fonnd that, on c i y - tallising the iiiactive base with Reycliler's clext~rocam~~liorsiil~~ionic acid, tlie first separation of crystals gave a destrorotntory base when rendered alkaliue wit.11 socla and distilled in steam. We coiisequently expected that, after sepiw;zting the iiinjoi- 1)iLrt of tlic lzvo-isomeride as bromocnmpliorPulplion:tte, pire ~:lestroi;etri~liydr~~~~i~i~l~innldine woulcl be readily o1)t;tinnble from the mother liq~iors by isolating the base and converting it into dextroc~inpliorsulplio~~nte ; by recryst~dlising t.liis salt, we expected easily to obtain a pure salt of clestrot,et.ruliydro(~iiinxldine, and to dispense with the rather tedious process originally em- ployed by Laclenbuiig (fie?.., 1S94, 2'7, 77) f o r prepwing the destro- rotatory base. Singnlizrly eiioiigli, however, this plan was a failure, but the results obtained ;we of siiflicient interest to iiierit description.The crude destro-base was separated froiii tlie NO ther liquors and dissolvecl in a liot ethylic acetate solution of a molecular proportion of Reychler's dext roc.anipliorsul~~1ionic acid. On cooling, a copious separittion of t.he campliorsu1phona.t~ occurred, and this was re- peatedly crysttdlisecl from hot ethylic acetate, tlie s d t separating in colourless, aciculnr aggregates.The fresh crystals melt a t 65 -SOo, but after preliminary drying in a vacuum the iiielting point is 125-1 37" ; t.he lower inelting point is due ti.pp:irently to mechanically retaiiiecl solvent, because the crystals do not lose nplrecinbly in weighh at 100". The salt wts reci*ystnllisecl four t,irnes from ethylic acetate, the melting points, after hying, being (1) 1 2 1-1 3 3 O , (3) 125-1 2i0, (3) 135-13'io, and (4) l25-137", so t,li:Lt tlie material liad been recrystallised twice after att,aining the coustant melt,ing point 125-lSTo, and might rensonnbly be expected to be n pure salt. After treatment with soclii an51 distillation in steam, t,he base was obtained as a colourless oil boiling nt 156' under 54 nim.pressure. It gAre the rotatory power al,+ 39*4-i0 in a 100 niin. tnbe, and since the pure 1:evo-base gives aD - 59*3i0 in a 100 mu. tube, still contains about one-sixth of its weight of lt-cvo-base. The base ivas therefore again coiivei*ted into dest~oceii~pliorsulphon~ ate and crystdlised from boiling acetone ; the salt is very sparingly soluble in hot dry acetone, but dissolves readily if the solvent contains eeveral per cent. of water. The first separation melted at 125-12$",RESOLUTION OF TETRARYDROQUINACDINE. 1079 and after drying a t ZOO', a solution of 1 gram made up to 25 C.C. with water gave aD + 3.60' in a 200 mm. tube, whence [.ID + 45.0' ; the mother liquor gave a separation melting at 125-127', and after drying at 100° an aqueous solution of 0,4374 ,a~-nm of this, made up t o 25 C.C.with water, gave u,+1.53 in a 300 mm. tube, whence a,, + 44.3'. The main separation having [a],, + 45.0" was once more crystallised from acetone and again melted at 145---1373 ; an aqueous solution of 0,4374 gram of this in 35 C.C. gitve al, + 1.57' in a 200 mm. tnbe, whence [.ID + 44.9". These three fractions Iinving identical melt- ing points and practically the same specific rotatory power, the base wits again separated as before from the fritction having [air, + 44.9". After distillation under reduced pressure, the base had the rotatory power aD + 53.04" in a 100 mni. tube, a value considerably below that obt'ained for the pure Ix-vo-isomericle, namely, a, -- 59-24'. These results illustrate well the extreme difficulty which may be met with in isolating a pure salt of the type cZ13 clil, from a mixture of the types cZB clrf and Ill cld.Had we not possessed previous knowledge of the specific rotatory power of clestrotetrrtEiydroquinxIdine, we should, on the basis of the above results, have been justified in supposing the rotatory power of the destro-base to be uD + 53' in a 100 mm. tube. It is therefore necessary t o emphnsise the fact that in sepnrating a mixture of the type dB CEB, ZB tEA, no criterion of purity is necessarily afforded by the identity of melting point or rotatory power of several consecutive fri1ctions. Fructioncd C'rjystcdliscctioii of Crude I)extrotelrcc7c?lclropuisiccl~ille Hgdr och lwr it le . We have shown that l,.evotetrahyclroqiiinnldiiie hydrochloride has the coniposition C1,,H1,,N,HC1 + 8,0, whilst the mcemic isomeride is anhydrous ; it might therefore lie expected that by crystalliaing the hydrochloride of the destro-base containing a little of the lava-isomeride, dextrotetrahydroquinaldine hydi+ochloricle could easily be obtained in a pure state.Here ngsin we met with failure, and again the failure WAS more interesting than success would have been. A quantity of crude dextrotetrahydroquinaldine obtained in the manner described above from the inother liquors remaining after the preparation of l ~ v o t e t ~ n l i y d r o ~ ~ ~ i i i n l d i i i e destro-(z-bronrocaiiiphol.sul- phonate was converted in to hydrochloride, which wi~s then frnctionally crystallised from absolute alcoliol. The first fractions consisted almost entirely of the rnceiuic hydrochloride, which would be expected to be less soluble than its optically active com1ionent.s ; they had [ + 6' or + 7", and were only iuechanically contixminated with t'he dextro- hydrochloride by reason of adherent mother liquor.These fractions1080 POPE AND I'EACHEY: were recrystallised from absolute alcohol, and readily yielded pure racemic hydrochloride ; the mother liquors in which the dextro- hydrochloride had concentrated mere then added to the main mother liquor, and the fractionation continued. After a series of deposits had been obtained of this low rotatory power ( [ ~ ] , + 6 ~ to 7"), the specific rotatory power suddenly rose considerably ; thus a separation having [ .IL, + 6.28" was followed by one having [.ID + 41.90", and the succeeding fractions had specific rotatory powers varying from [aID + 55' t o + 58" in 2 per cent.aqueous solutions. Further, the repeated recrystallisation of these fractions failed on all occasions, with one exception, to yield a product having the specific rotatory power of dextrotetr~l~ydroquinaldine hydrochloride, namely, [a],, + 62.8' in aqueous solution, the specific rotatory power of the doposits varying from [u],+55" to +58". This behaviour showed that fractional crystallisation of a mixture of racemic and dextrorotatory tetral~ydroqnin~ldine hydrochlorides from alcohol could not be used as a practic:il method for obtaining the pure dextro-salt ; the racemic hydrochloride can be separated from the mixture until only a few per cent.of it remains with the deutro-salt, and further crystallisation fails to sensibly reduce this proportion. The explanation of this behaviour is to be found in Roozebooni's lucid discussion of the isothermnls for systems consisting of a solvent, a racemic compound and its optically active componeuts (Zeit. pkysiknl. Chem., 1899, 28, 494). In Fig. 10 the ordinates denote FIG. 10. the quantity of dextro-compound, and the absciss;le the quantity of Jsvo-compound in the saturated solution; the curve Zbcd is an iso- thermal representing the composition of saturated soltitions of the dextro- and lwo-compound and their mixtures at some particular temyerature. The branches Zb and dc of the curve represent solutiopsRESOLUTION OF TETRAHYDIiOQUINALDINE. 1081 in contact with levo- or dextro-material respectively, in contact, therefore, with one solid phase ; the branch bc represents solutions in equilibrium with the raceruic compound, again, therefore, with one solid phase.The points b and c, however, represent solutious which are in equilibrium with two solid phases, namely, the racemic and one optically active component. Suppose we start with a solution contain- ing both dextro- and lrevo-material in the proportion corresponding t o the point, Jl on tlie curve ; on crystallisation, the solution mill deposit raceniic compound, and the composition of the saturated mother liquor will change in the direction of the point c ; after the composition at c is reached, the subsequent fractions mill be of identical composition and specific rotation, b d h dextro- and raceiuic salt being deposited in a fixed proportion, and the solution retaining the composition at c until the whole of the solvent has evapor;tted.These considerations hold if the temperature of crystallisation remains constant, and in our attempts to isolate the dextro-salt from the mixture, crystallisation always oc- curred at the ordinixry temperature ; in spite of slight temperature fluctuations, me mere in general unable to change the composition of the mixture to any serviceable extent from that at the point c. This is ap- parently the first recorded instancein which equilibrium in contact withan optically active and racemic pliase at the point c has iiiterposed practical difficulty, and me hoped to overcome the difficulty by sowing dextro- hydrochloride into supersaturated solutions obtained by lowering the temperature or by cooling saturated solutions in contact with the dextro- salt so that the solution becomes a labile one of composition lying upon the branch ac.Only on one occasion, however, did we succeed in getting a separat ion of pure clextro te t rah y droquinaldine liy drochloride. This is somewhat noteworthy, because Bipping and Pope observed just the converse beliaviour with mixtures of potassium sodium dextro- and laevo-tartrates (Trans., 1899, 77, 45), and pointed out the ease with which ‘( a racemic compound might, under certain conditions, be resolved into its optically active components by simple crystallisation at tem- peratures at which the racemic compound is more stable than the mixture of the two optically active salts.” Attempts were made to change the composition of the solutions of the mixed hydrochlorides from that represented by the point c by alternate crystallisations from two solvents, on the supposition that the com- position a t this point mould differ for various solvents; our efforts in this direction were not successful, the only useful solvent besides alcohol being a mixture of alcollol and acetone.1082 POPE AND PEACHEY: Il'mctioiza Z CrystallisatioPz of Jiixtwes of Deetyo- and Bcicentic Benxoyt- t e t mlyd r o pzc i.r mldine.I n spite of our failure to isolate pure dextrotetrahydroquinsldine hydrochloride in quantity from its mixture with the racemic compound and the consequent expectation of a similar behaviour with the benzoyl derivatives, we mere able easily to obtain beiizoyldestrotetrsliydroquin- nldine in quantit8p by fractionally crybtallising a mixture of this sub- stance with its raccmic isorneride.Since we have previously shown that no raceinisation occurs on liyclrolysing bemoyllrevotetrahydroquin- aldine, we could readily obtain pure destrotetraliydroquinaldine in quantity from its beuzoyl derivative. The conversion of the mixture of clextro-base with a much smaller proportion of its hvo-isoruericle into benzoyl derivative is effected as previously described (p. 1073) ; the benzoyl derivative is then system- atically fractionated from its hot acetone solution by recrystallising the early separations, which consist largely of the raceinic compound, and adding the opticidly active mother liquors to the main solution. After the greater part of the raceniic material present has been eliminated, the subseqnent deposits consist of nearly pure benzoyldextrotetrahydro- quinaldine,mhilst occasionally a separution of low specific rotatory power occurs ; by continuing the fractionation and determining the specific rotatory power of each deposit, practically all of the original benzoyl derivative may be obtained in two lots, the one inactive, and the ocher baving [ u ] ~ - 2 .1 7 . 4 O in a 3 per cent. benzene solution a t 16'. Since our specimen of pure beuzoyl1~votetrahydroclrrin:Lldiiie had [ uID + 247.3' at 15" in benzene solution, the materid of [I air, - 247.4" is obviously a pure snbstiince.I n this case, therefore, the existence of the equilibrium point c (Fig. lo) causes no practical difficulty, owing, doubtless, to a greater facility of supersaturation. 111. DEXTROTETRAIIYD~~OQUINALDINE. Ir:en~o~~tlezt~otetrccl~~dl.oq'ZLiiZCilCZi.r2e. The general properties of benzoyldextrotetrahydroquinaldine need no separate description ; its crystalline form has already been discussed (p. 1074), and analyticd results proving its composition have been obtained. The deterniinatioiis of the density of beazoyldextrotetrshydroquin- aldme were niacle in the same way as were those of the antipodes and with the following results : ,,,14%. - - 1.2111 ; 1,2115 ; 1.2115. Alean= 1.2114, The iiiolecular volume at 15*6°/40 is thus 207.20 ; the mean density andRESOLUTION OF TETRAHYDROQUINALDINE. 1083 molecular volume differ by only 0.03 per cent, from the corresponding values obtained for benzoyllfevotetrahydi-oquinaldiiie.Benzoyldextrotetrahydroquinaldine is highly l,.evorotatory, and i t s rotation constants are given in the following table : Hoiixeiie ............... ,) ............... ) , ............... ) ) ............... C!llloroforlll.. .......... Bceto11u ............... E t.liylic acc- tate ...... Acetic acid ............ Etliylic a h a ( )I101 ...... w. - 1. 1 so 19 I 3 1 i I i 1 i I i l i I T 17 - a I,. A co1ilpikriso11 of these iiiimbers with those olhinecl with the antipodes shows tliat tlie two sets dif'fer nritmlmeticnlly but slightly, the difference being at8tributable to tlie different working temperature.In both cases, the solutioiis in glacial acetic acid Show far the highest rotation constiin t s. L)e;~~trotetrcch!/~~rozu~~~~~ Ztliiie. For the preparation of this base, finclr-powdered ~)enzoylclestrotetr~- hydrociuinaldiiie is heated in n yeilux al)pai-;ttus with excess of con- centrated hydrochloric acid ; after several ditys' heating, a copious separation of benzoic acid occurs 011 cooling, and ilie hydrolysis is complete. Slight excess of socln is :dcled and the product distilled in a current of steam ; the base is estr:icted with etlier niid purified RS before described ; the purity of the product \vils ascertained by it de- terminrttiou of its density and specific rotatory power. The density was found to be tl'.';: 1.01926, nud the rotatory power mas a,+5!3.2l0 in n 100 turn.tube at 20°, mlieuce [u],+58*01) at 20". Ladenburg gives the rotatory power, aI, + 6S.35" in a 100 im~. tnbe at IG", for this base, and calculates the specific rot;~tory power as [ a ID + 55*09O ; the density number, d';'' 1.043, which lie uses is somewhat too high (Ber., 1S91, 27, 76). I) ext yo t e t r*uliplclm pi n cc Z tliiz e II&o ck lor ide, C, 11 N, I4 C1+ 1% 0. The hydrochloride of dextrotetrr~liSdroquiiialdine was prepared i n the same may as t h a t of the lrevo-isoiiieride. It crystallises i n large, orthorhombic monohydrated tablets which melt at 196.6-197*5°, and are crystallographically indistinguishable from the crystals of the laevo-isomeride (Figs, 1 and 2, p. 1070).1054 POPE AND PEACHEY: The specific rotatory power of the salt dried a t 100' was determined in aqueous solution; 0.4806 gram of the anhydrous salt, made up to 25.3 C.C.with water a t 2 1 * 4 O , gave uD + 2'51' in a 300 mm. tuWe. The specific rotatory power is thus [ a ] [ , + 66*1', which compares perfectly well with the value [ alD - 66.4' obtained for laevotetrahydroquinaldine hydrochloride (p. 10.72). LV. ~IOLECULAR ROTATORY POWERS OF SALTS OP OPTICALLY ACTIVE BASES WITH OPTICALLY ACTIVE ACIDS. A simple law should connect the inolecular rotatory powers of optically active salts of the type d13 dA and 112 d A , in fairly dilute aqueous solutions, provided that the base aid acid, 1; and A , are so strong that the salts are pr:ictically wholly dissociated. The algebraic cliff erence between the molecular rotatory powers of the two salts should be equal to twice the molecular rotatory powers of the hydrochloride or similar salt of the active base, whilst the algebraic sum should be equal to twice the niolecular rotatory power of a metallic salt of the active acid.Since the truth of this consequence of the electrolytic dissociation hypothesis has previously only been tested by Walden upon Ealts of feeble bases, like morphine (Zeit. plqsSaZ. Cltem., 1894, 15, 206), it seemed desirable to prepnre salts of the strong bases, dextro- and Ievo-tetrahydroquinaldine, with a powerful optically active acid, when the principle enunciated above should be found to hold even more rigidly than in the case referred to. For this purpose, the salts of dextro- and Izevo-tetrahydroquinaldine with Reychler's destro- csmphorsulphonic acid mere prepared ; the former salt.wag also required in orcler to compare its properties with those of the impure compound obtained in the fractioaal crystnlliastion of the mixture. Dextrotetmh~lcl~oqu~~~c~Zi~~~~e L)extrocc~n~~~~o~s.zc~~l~onicte, Cl,~I~,,N,C,,H,,O~SO,H. Dextrotetrahydroquinaldine dextrocninphorsulphonate is obtained by crystallising a benzene solution of the component base and acid in long, colourless, flat prisms melting a t 128-129"; the prism zone is com- posed of six faces, whilst ruther obliquely placed end faces complete the crystal. An optic axis is observed to emerge through the large face, and the extinction in this face is practically parallel with the longer sides ; the crystals are very possibly anorthic, but good crystals could not be obtained for measurement.The salt is soluble in the ordinary organic solvents, including carbon bisulphide and ethylene dibromide, The following analytical results mere obtained with material dried a t 100' :ltESOLUTlON OF TETlihHYDROQUINAT~DINE. 1085 0.1246 gave 0.2887 CO, and 0.0879 H,O. 0.1317 ,, 0.3054 CO, ,, 0.0923 H,O. C = 63.24 ; H = 7.79. 0,4162 ,, 0.2699 B~xYO,. S=8*93. C = 63.18 ; H = 7.84. C,,H,,O,NS requires C = 63-32 ; H = 7.65 ; S = 8.64 per cent. The specific rotation was determined in aqueous solution : 0.507'7 grain, made up to 25 C.C. with water a t 19*1', gave a, + 1%5O in a 200-mm. tube ; whence [air, + 45.7" and [MI, + 173.3O. I n discussing the fractional crystallisation of a mixture of the dextro- caniphor~;ulphon,ztes of destro- and I~vo-tetr,zhydroquinaldine, it wts pointed out (p.1079) that we were unable to obtain tlie destrocainphor- sulphocate of a higher specific rotatory power than [ aID + 45.0' in aqueous solution, arid that this salt yielded a base having the rotatory power u~,+ 53' in place of + 59' ; since it is iiow shown that dextro- tetrit1iydroclnin:ldine clestrocani~ho~sulplioi~ate has [ u ID + 45*7', i t is evident that the inaterial previously obtained was, as we stated, still coiitnmiiiated with salt of the liero-base. Lcevo~etrccr7L2/~I~oqui,~a~cEis~e next,.ocnniiil~ol.szc~)?~o~ante, U , , , I r I , , N , C l o ~ ~ , , O . s ~ ~ ~ ~ . LE vot et rsh y clro~uinnldiiie d ex trocalllphorsulphon~ t e is prepmed by crystallisiug the requisite quantities of the biise and acid together from benzene solution ; it sepwates in colotirless, flattened needles melting a t 137-138".The needles are scluilre ended with truncated corners ; tlie extinction in the large face is sti*;light with the sides, aiid a, bisec trix emerges perpendicu1:wly through the face. The following analytical results mere obtained with material dried at 100" ; 0.2022 gzve 0.4680 CO, and 0.1416 H20. c! = 63.12 ; H = 7.78. 0.1209 ,, O.BS02 GO, ,, O.OS82 H,O. C = 63.21 ; 11 = 7.83. 0.4615 ,, 0.2S3S BdW,. 8=8*85. C,,H,,O,NS requires C = 63.33 ; I€ = 7.65 ; S = 8.64 per cent. 0.5073 gram, made up to 35 C.C. with mster at 19*1", gave aI, - 0.74" in a 200 mm. tube, whence [a]D - 18.3' and [ 1x3, - 69.5". Amntonium I)e~t,.ocai,LI~~oi.su~)~o~L(ite, C,,€I,,O*SO,NH,.The ammonium salt of des trocamphorsulphonic acid has been described by Reychler (Bull. Soc. C'him., 189S, [iii], 19, 120); we obtained it in long, colourless needles which, after drying at loo", gave analytical results agreeing with those required for the above formula. The following rotation determinations were made : 0.2503 gram, made up t o 25 C.C. with water a t 16", gave aD+0.i2* in a 200 mm. tube ; whence [ a ] , + 21*Oo. VOL. LXXV. 4 D1086 POPE AND PEACHEY: 0*5000 gram, made iip to 25 C.C. mith water at 18', gave uD+ 0.83' in a 200 mm. tube ; whence [.ID + 20.7". 1.0008 grams, made up to 25 C.C. with water at 18', gave aD+ 1.65' in n 200 mm. tube ; whence [.ID + 20.6'. Since we are dealing all through with solutions of about 3 per cent., me may take for compsrison with other salts the values [a I,, + 80 7' and [ 111: JD + 51.7'.The following table gives the specific and molecular rotatory powers of the various salts of dextrocamphorsulphonic acid with which me are concerned : d-C,,Hl SN, h C ' , ,,HI 5 0 *SO:3TI .................. Z-CIO H,:jN, d-("lo If l,O*SOylI ................. I-C,nI-TI ,N, HC1,. .................................. NH,, d-C?,oH,,O*SOsH ........................ $ 4 5 7 ' - 18.8 - 66 '.I + 2 O ' i -t 173.3" -I- 5 1 . i -69.5 7 1 - 121 *7 The algebraic cliff erence between the molecular rotatory powers of salts (1) and (2) is 343.8' and the half of this, namely, 121*4', should be equal to the molecular rotatory powers of active tetrahydroquin- aldine hydrochloride, namely, 12 1.7'.Further, the algebraic sum of moleculas rotatory powers (1) and (2) is 103.5, and the half of this, namely, 51*3', should be equal to the molecular rotatory power of ammonium dextrocamphorsulphonate, namely, 5 1 *7". The agreement even in these comparatively concentrated solutions is very close, V. EXTERNALLY COMPENSATED TETRAHYDROQUINALDINE. For the preparation of the tetrnhydroquinnldi~ie used in the present work, quinaldine was reduced with tin nud hydrochloric acid essen- tially as described by Walter (fie,.., 1S92, 25, 1261); when the reduction was complete, slight excess of soda mas added and the liquid subjected to prolonged clistillntiou in a current of steam. The tetrah ydroqninnldi ne was oxtr;tc t ed from the dis t illiit e with ether, the ethereal solution dried over potash and fractionally distilled ; the fraction boiling a t 240-250' wils taken as containing a11 the hydro- quinaldine and is sufficiently pure for immediate resolution into its optically active components as described above.Racemic Tet,.cc~~cll.oi.rLincclcl.iize IIyclrocl~loricEe, C1,,H,,N,HCl. A quanbity of the crude base boiling nt 240-250' was dissolved in large excess of hydrochloric acid aucl the solution allowed ta crystal-RESOLUTION OF TETRAHYDROQUINALDINE. 1087 lise ; the deposited hydrochloride was crystallised a number of times from absolute alcohol and then melted a t 196-197-5'. This salt was prepared by Fischer and Steche (Anncdeiz, 1887, 242, 358) and de- scribed as not very soluble in water, but easily soluble in alcohol ; it is, however, less soluble in alcohol than in water.Our analytical results show that the salt has the composition C,,H,,N,HCI, as deposited from aqueous or alcoholic solutions ; it is consequently a true racemic compound, because its optically active components under similar conditions ci-ystallise with water. I t separates on spontaneous evaporation of its pure aqueous 01' alcoholic solution either in small, colourless, transparent crystals of rhombohedra1 habit, or in lon& thin needles, both melting at 196-197*5°. These crystals are very poorly developed and the faces slicw considerable striation ; on microscopic examination, they are usually found t o be composite, a lnmellated twin structure somewhat resembling that characteristic of labradorite being observed. The goniometricnl examination of these crystals gave no result other than that the crystals are geometrically pseudorhom- bohedral in accordance with the lamellation detected microscopically.Large crystals better suited for measurement are obtained by spon- taneous evaporation of a solution of the salt in hydrochloric acid ; these consist of large, colourless, transparent tablets of Door facial development. The form c f 001) is domiii:mt Fro. 11. L (Fig. 11) and the prism p{ 110) is also well developed ; the pinacoicl CI [ 100) is usually smaller, and the faces of the dome r[lOll, are only observed as narrow replacements. There is nn extremely perfect cleavage parallel to c(OO1) and, when hacked in other directions, the cryst.& merely shear parallel to the cleavage. A bisectrix of positive double refraction emerges nearly normally to c(001), but the orientation or" the optic axial plane could not be ascertained.Crystalline system.-Monosymmetric, C.C b : ~ = 0 * 9 3 4 : 1 : 0.035. p= 71" 46'. Forms observed.-cc(100}, c(OOl}, p{110}, and ~ { l o l ) . 4 D 21088 POPE AND PEACHEY: The following angular measurements mere obtained : Angle. a11 = 100 : _I10 271, = 110 : lL0 pp =110 : 110 CM: = 1 0 0 : ~ 0 1 CT =001 : 101 fir =lo0 : 101 C]? =001 ' 110 C]? =ooi:rio p r =110 : 101 Number of observations. 31 18 15 42 31 3s 44 14 8 Limits. 40'48'- 42" 7' 96 4- 97 52 82 26- 84 1 70 4 i - 72 12 53 43- 55 26 53 6 - 55 13 75 19 - 77 14 102 54-101 7 63 0- 64 56 Mean. 41'35' 96 53 83 19 il 32 54 42 6 4 5 76 25 103 29 63 49 Cnlcnla tcd. - 96"50' 83 10 71 46 54 9 - - 103 32 63 55 After melting the racemic hydrochloride, it solidifies whilst hot less readily than the lavo-isomeride, but when cold solidifies with distinctly greater rapidity; it is quite easy to obtain a liquid film of the racemic substance at the ordinary temperature, but after 48 hours most of the liquid has crystnllised.On allowing the liquid film to cool considerably below the melting point and then again warnling, the whole may soon be made t o crystnllise ; crystallisation proceeds from centres and results in the production of broad, individual flakes, which are usually striated in n direction parallel to the extinction. The strk are parallel to the trace of the optic plane in the fragment, and a bisectrix emerges nearly norinally to the filce; the optic axial axes are outside the field, and the double refraction of the bisectris, probably the obtuse one, is of positive sign.As cooling proceeds, the plates crack considerably along lines perpendicular t o the direction of growth. Optically, this material is very similar to :'he orthorhombio plates of the hvo-hydrochloride, but morphologically it seems very different, the s p a r e plates never being observed ; the inactive material produced at the high temperature does not fall to pieces during cooling in the violent manner affected by the optically active substance. On rapidly cooling the molten film to the ordinary temperature, it remains liquid, but after a day or so a fringe of the modification described above, which polnrises brilliantly, forms round the edge of the preparation and subsequently a macroscopically opaque growth makes its appearance in the body of the film; this growth consists of minute, square-ended needles showing oblique estinction and, being of low double refraction, yolarises far less brillimtly than the first- described modification ; the needles exhibit an oblique optic axial emergence and are interlaced in a highly confused manner.After several days, this modification occupies nearly the whole of the film and encroaches upon the brilliantly polnrising fringe, the latterR ESO L U T I 0 N 0 1 T ETR AH Y D R 0 Q U I N A I, b I N E . 1089 Ftpparently becoming converted into the former. The externally compensated hydrochloride is therefore almost certainly dirnorphous.Race?& ~etr.al&ydl.oquinaZ~irre Picrute, C:,,H13N,C,H,(N0,),*OH, Racemic tetrahydroqninaldino picrate is obtained by crystallising the externally compensated base with the requisite quantity of picric acid from absolute alcohol ; it is more sparingly soluble in organic solvents and in water than the picrate of the hvo-base. The crystals deposited from the alcoholic solution melt a t 153-154’ and do not give good results on iiieasuremeiit ; they are apparently anorthic and consist of a predominant form n(100}, a smaller one, c[OOl}, and still smnller ones, p t l l 0 ; and p’(T10). The approximate angles are : QC = 100 : 001 = 74” 5.i‘ ((2) = 100 : 110 = 32’ 13’ cp = O O l : 110=8l 20 pp’= 110 :’110= 79 0 cp’ = 001 : 110 =; 7s 6 C$ =Too : i i o = 6s 47 There is a perfect cle2vage parallel to c(OO1) and the acute bisectrix emerges through ~ ( 0 0 1 ) ; the optic axial angle is large, the double refractim is negative in sign, and the optic axial dispersion is so marked that no defiuite extinction is observed for white light in c(OO1).The estinction in cc(100) is nearly straight with the edge ac, and the optic axial plane is nearly parallel to ~(100). The following analytical results were obtained with material crystallised from alcohol : 0.2055 gave 0.3523 CO, and 0.0829 H,O. C = 50.74 ; H = 4.48. 0.2107 ?, 0.3933 CO, ,, O*OSSS H,O. C=50*01 ; H=4*35. Cl,H,,07N, requires C: = 51.06 ; H =t 4-26 per cent. &xte~*nd~y Coqie?tsnted Tet,*nkyclror/u.i?anZdine, Cl,HI ,N. Carefully purified rscemic tetrahydroquinnldine hydrochloride is distilled in a current of steam with addition of a slight excess of soda ; the distillate is extracted with purified ether, the ethereal solution dried with potash, and the ether distilled off.The residual nearly colourless oil is then distilled under reduced pressure and practically all distils a t 196’ under 207.5 mm. pressure. The base has the same density as its Izvo-component, and its physical properties are dealt with fully in a subsequent paper, in which it is shown that this base is merely a mixture of the two optically active constituents, Rme?nic Beur;oy I tetycchyclroqu imcldine, Cl,Hl JY CO C, H,. This substance has been prepared from tetrahydroquinaldine by Walter (Bey., 1892, 26, 1 293), using the Schotten-Baumann reaction.It separates on spontaneous evaporation of its cold ethylic acetate1090 POPE AND PEACHEY: solutions in magnificent, lustrous, monosymmetric prisms (Fig. 12) melting a t 117-118°, and showing the form bf010) dominant ; the forms ~ ( 1 1 0 ) and q{Oll> are the next largest forms present, whilst the pinacoid a{ loo} is always small and frequently absent,. The pinacoid c(OO1) is well developed and gives good results on measurement ; the form ~(101) is generally well represented, whilst the form r’[lOl> is FIG. 12. FIG. 13. poorly developed and rarely observed. The crystals deposited from solutions in benzene (Fig. 13) differ greatly in habit from those obtained from ethylic acetate, and lend t homselves better t o crystallo- graphic measurement than do the latter ; they exhibit the form p[llO> predominant, and also shorn /(TO1 ] and o ( l l 1 ) well developed, whilst the pinacoids a[lOO) and b(010j itre comparatively small and frequently Frtr.14. absent, The facial development of the form ~ [ l O l ) is less than that of r ’ ( f O 1 t . Crystals are frequently deposited from benzene solution of the liemiliedrd habit depicted in Fig. 14 ; these consist of parts of the forms cfOOl), o(T11), and p [ l l O > . No indications of pyroelectrical properties could be obtained to show that this habit is due to hemi- hedrism, so that, it is probably due merely to abnormal conditions of growth.flESOLUT1ON OF TETRAHYDROQUINALDINE. 1091 Crystalline system .-Monosymmetric.CG : h : c = 0.6765 : 1 : 0.6675. /3 = S1° 4'. Forms observed.--(c~100~, b[010{, c[OOl), p [ l l O ! , q(Oll], ~ [ l O l ] , The following angular nieasnrements were obtained : + ( i o i ) nncl o p i i ; . Angle. up = l o o :110 ?y = 010 : l i 0 21p =I10 : 110 cq =001: 011 b y = 010 : 01-1 'I? =011:031 qr =011 : 101 pq =110 : 011 p,' =IlO: 101- pr'=llO : 101 pq =110 : y o qr'= 01 1 : 101 C/' = 001 : LO1 cr' =001 : 101 o / ~ ' = l O o : 101 or/ = l o o : 011 $7" =Oll : 111 ( I . ~ = l o o : 111 co =001 : 11 1 (31 =ill :Ti0 tp =go1 : 110 T'O =LO1 : 111 no =111:!11 (C/' = l(l0 : 101 bo =010 :111 ~11nll)cr of obserratioiis. I,il ni ts. 33'42' 50 17 67 29 3.3 24 5li 37 66 51 50 19 i 8 35 51 3 67 33 65 37 56 49 40 41 40 12 4s 56 49 57 s.:! :j t.; 48 '24 51 59 5-1 10 43 14 s2 33 27 14 5:: 5s 62 56 Chlculntcd.The extinctions in the faces b(010) and p(110) are practically straight with the side np. The obtuse bisectris emerges nearly normdly throngli cc{lOOj and the obtmse bisectris is chservecl through a section hacked nearly perpendicular to the zone [ O O l ] ; the optic axial angle is large and the double refraction is negative in sign. The optic axial dispersion is slight and tlie angle for blue light is greater than that for red ; tlie plane of symmetry is the optic mid plane. After melting, the substance solidifies fairly rendily a t high temper- atures, and the whole film niny be caused to crystallise by alternate heating and cooling; if, however, the film be rapidly cooled, it may be obtained at the ordinary temperature as n liquid film which crystallises extremely slnggishly.The crystalline film obtained at high temperatures is composed of long needles radiating from centres j1092 RESOLUTION OF TEI'RAHY DROQUI NA LDIN E. these extinguish nearly straight with the direction !of growth and, on cooling, crack extensively in different directions. Most of the crystalline individuals lie nearly perpendicular to the positive obtuse bisectrix of the optic axial angle and the optic axial plane is parallel t o the direction of growth. Sometimes the acute bisectrix emerges nearly normally to the surface of the crystal fragment; the optic axial angle is large and the optic axial dispersion slight. The double refraction is negative in sign. The optical properties of this modifica- tion agree closely with tliose of the crystals deposited from solution, and tlie two are structurally identical, so that externally compensated benzoyltetrahydroquinaldine melts and solidifies as a rncemic compound. The molten film nt ordinary temperatures crystallises very slowly, yielding a mass of needles which shows aggregate polarisation. The density of the crystals was determined as in the case of those of the optically active constituents with the following results : dli?= 1,3373 ; 1.2380 ; 1.33'73. iVIean- 1.3375. The moleculnr volume at 14.5"/'4" is thus 303.8. Since tlie optically active benzoylt~trnhydroqninal~lines have the molecular volume 20'7.2 a t tho same temperature in the crystalline state, it is obvious that the crystals of the racemic substance conform to Liebisch's rule (Ansmlen, 1895,286, 140), their density being greater than that of their optically active components. I n the formation of crystalline raceinic benzoyl- tetrahydroquinaldine from its crystalline optically activo components at 14*5O, a contraction of abont 2.124 per cent. in volume occurs. VI. IXELTISCI POINTS OF ~PTIC'hLTAS ACTIYE AND EXTERNALLY COMPENSATED ISOMERIDES. The great interest which has been imported into the pestion of the melting points of optically active and externally compensated isomer- ides by the recent work of Roozeboom (Zeit. p l q d n l . Chem., 1899, 28, 505) rendered desirable the investigation of tlie melting points of the hydrochlorides and benzoyl derivatives of the tetrahyboquinaldines. Careful determinations of the melting points of highly purified racemic and optically active tetrahydroquinaldine hydrochlorides gave the values 196--107*5° and 19695-1 97*5O respectively ;.the melting point of the racemic material is not quite so sharp as that of the dextro- or hvo-component. An intimate mixture of about equal weights of racemic and 1;levo-material melted at 180-1 84". Similarly, active benzoyltetrahydroquinaldine melted at 117.5-1 1 8 * 5 O and the racemic material at 117-llSo, the racemic again melting less sharply than the active isomerides. A mixture of about equal parts of each melted at 108-110".ItESO1,UTION OF TE’l’RAHYDROPARATO1,UQUrNALDINE. 1093 Although the active and iiinctive isomerides melt at practically the same temperature, the externally compensated substance melts a s a racemic compound, and the melting point curves for t h e mixtures mould seem t o belong to Roozeboom’s Type 2 ; attention does not seem t o have been drawn previously to examples of this type. Our thanks are due to the Research Fund Committee of the Chemical Society for grants enabling the purchase of inaterials used i n the foregoing work.