JAPACONITINE AND THE ALKALOIDS OY JAPANESE ACONITE. 45 V.-Contribution to our Knowledge o f the Aco~zite Alkaloids. Part XV. On Japaconitine and the Alkaloids of Japanese Aconite. By WYNDHABI R. DUNSTAN, F.R.S., and HAROLD &I. READ, Assistant Chemist in the Scientific Department of the Imperial Institute. THE examination of the physiologically active alkaloid which exists in Japanese aconite roots has already formed the subject of several communications to this and other societies ; but in view both of the conflicting statements as to its composition and relationship t o aconitine and of the more recent work which has been carried out by one of US on aconitine and pseudaconitine (the crystalline, toxic alkaloids of46 DUNSTAN AND READ : JAPACONITINE Aconitum NapeZZus and A .ferox respectively), it was thought desirable that the investigation should be extended to the alkaloids of Japanese aconite. Japanese aconite roots seem to have come into commerce about twenty years ago ; they are now imported regularly, and are regarded as more toxic than those of A. Napellus. An exhaustive report by Dr. A. Langgaard on (‘Japanese and Chinese Aconite Roots” was published in 1881 (Arch. Pharrn., 18, 161), and from this it appears that, although the native practitioner employs many varieties of aconite, that most frequently used and exported is (( Kuza-Uzu,” which has been identified by various authorities as A . Chinense, A . Fischevi, and A crystalline alkaloid was first obtained from Japanese aconite roots by Paul and Kiogzett (Pharm. J., 1877, [iii], 8, 173).The alkaloid was soluble in ether, insoluble in water, and formed uncrystal- lisable salts. From the results of a single combustion for carbon and hydrogen, and one nitrogen determination, the formula C,,H,,O,N was adopted. This, however, was not controlled by the analysis of the platinum salt, This alkaloid was said not to suffer hydrolysis into benzoic acid and a basic substance. I n 1879, Wright, Lug, and Menke extended their investigations of the alkaloids of other aconites to those of Japanese aconite. The general results of their work (Trans,, 1879, 35, 387) may be briefly stated as follows : 1. The roots imported from Japan were considerably richor in active crystalline alkaloids, as well as in non-crystalline bases, than A.Napellus. 2. Only one crystalline alkaloid, named japaconitine, was present. This melted at 184-186”, and the formula C66H,8021N, was proposed, the base being regarded as the sesqui-apo-derivative of a parent sub- stance having the formula C,,H,7012N. 3. Japaconitine formed readily crystallisable salts, especially with nitric, hydrochloric, and hydrobromic acids. The hydrobromide has the formula C66H880,1N,,2H&? 4- 5H20. 4. The alkaloid could be entirely extracted from the roots by means of alcohol alone. 5. When hydrolgsed, japaconitine furnished benzoic acid and a new base, jspaconine, which was amorphous, and formed amorphous salts. 6. When either the parent base or the hydrolytic base is benzoylated, a derivative was formed containing four benzoyl groups for every C,, originally present.4. Lycoctonum respectively. For the hypothetical parent base, the formulaAND THE ALKALOIDS OF JAPANESE ACONITE. 47 was proposed, whilst japaconitine, being the sesqui-apo-derivative, was given the formula I n 1885, K. F. Mandelin examined a specimen of japaconitine pre- pared by Merck, and stated that the alkaloid from Japanese roots agreed in chemical and physical behaviour with aconitine derived from A . Xape llus. Liibbe '(Chem. Centr., 1890, ii, 14s) arrived at the same conclusion, working with material which he extracted from '' kuza-uzu " roots im- ported direct from Japan. The alkaloid which he isolated crystal- lised in rhombic prisms melting a t 183-184'. It must be noted that the proof of the identity of his base with that obtained from A .Napellus by Wright rests on the results of a single combustion, Still more recently, Freund and Beck (Bey., 1894, 27, 723) have stated that " they have no doubt of the identity of the alkaloid from both sources," although they do not give any details of the work which led them to this conclusion. The roots from which we have extracted the alkaloid used for this investigation were purchased in commerce as Japanese aconite roots. Mr. E. M. Holmes, F.L.S., Curator of the Pharmaceutical Society's Museum, kindly examined a specimen of these, and informed us that they are undoubtedly ordinary Japanese roots, such as described by Langgaard (Zoc. cit., and Pharrn. J., 1881, [iii], 11, 1043), and known in Japan as " kuza-uzu." I. JAPACONITINE.Extraction of the Alkaloid. We have extracted the alkaloid by the following method. The finely ground root is percolated with a mixture of amyl alcohol and wood spirit, in the proportion of 1 part of the former to 5 of the latter. The wood spirit is distilled from the mixture under reduced pressure on a water-bath, the temperature of which is kept below 60'; the residual solution of the alkaloids in amyl alcohol is extracted with dilute (0.5 per cent.) sulphuric acid, and the latter thoroughly washed with ether to remove all traces of aruyl alcohol. The alkaloid is now fractionally precipitated by sodium carbonate or ammonia from the aqueous solution of the sulphate, and is extracted by shaking with ether. When the whole of the bases soluble in ether have been removed, the extraction is continued with chloroform.The aqueous mother liquors still contain a small quantity of alkaloid, The complete exhaustion takes a considerable time.48 DUNSTAN AND READ : JAPACONITINE a certain proportion of which may be removed by saturating the solu- tion with sodium sulphate and then shaking with chloroform, The ethereal solution is now washed with water to remove traces of alkali, and calcium chloride is added. The latter serves a double purpose in precipitating most of the colouring matter extracted by the ether, and at the same time drying the solution. Since the calcium chloride does not show any tendency to combine with the alkaloid, this method is very advantageous in avoiding a great deal of subsequent purification. The dried and filtered solution is caref iilly distilled, and, when sufficiently concentrated, the nearly pure jap- aconitine crystallises out in minute, colourless rosettes of prismatic needles.Prom the mother liquor, a further quantity of crystalline japaconitine may be obtained. The best method for working up the solution in chloroform is to wash it with water to remove traces of alkali, evaporate to a small volume, drive off the residual chloroform with a current of air, and then to add ether to the pale brown varnish which remains. The greater part of the varnish dissolves in the ether, and on pouring off and evaporating the latter solution, more crystal- line japaconitine is obtained. The extract left undissolved by the ether will be described later (see page 62).Proprt ies of Japaconitine. The crystalline base obtained in this manner melts at about 195-196'. For the final purification, it is converted into a salt, preferably the hydrobromide, by dissolving in the diluted acid and evaporating the neutral solution in a vacuous desiccator until crystals form. The crystallisation may be considerably hastened by well stirring the solution with a glass rod. We have found that, in evaporat- ing the solution, it is well to avoid entirely the aid of heat, since the saving of time ef3Fected is more than counterbalanced by the discolora- tion of the resulting salt. Having recrystallised the salt, either from water or from a mixture of alcohol and ether, until the melting point remains unchanged, the base may be regenerated in the usual manner.When thus purified, japaconitine crystallises from ether, alcohol, or chloroform in colourless, anhydrous rosetkes of needles, The base recrystallised from ether was analysed. 05518 lost 0.002 after 30 minutes a t 100'. Loss = 0.362 per cent 0.6267 ,, 0.0003 ,, 90 ,, looo, LOSS =0*047 ,, The base therefore does not combine with the solvent. Many attempts were made to obtain large crystals for crystallo- graphic measurement, but these, unfortunately, have been unsuccessful. This difficulty in obtaining large crystals of japaconitine contrastsAND THE ATJKATAOIDS OF JAPANESE ACONITE. 49 sharply with the ease with which, by the same methods, t'he well- defined, hexagonal prisms so characteristic of nconitine are obtained. The largest jnpaconitine crystals we have succeeded in preparing were obtained by the gradiial precipitation of a solution in dry chloroform with light petroleum.These were kindly examined for ns by Mr. W. J. Pope, who describes them as follows : '' The crystals of japaconitine are very small, transparent, colourless needles possessing a brilliant lustre; the sides of the needles are made up of two pairs of parallel prism faces. A number of measure- ments of the prism angle showed this to have the value 77" 35'- 77" 54' ; no measurements in which the end faces of the needles were involved could be made owing to the minute size of these end faces. " The extinction in the prism faces is straight, and on examination in convergent light the interference figure of a uniaxial substance, or of a biaxial substance with a very small optici axial angle, is seen ernerg- ing in the angle between the prism faces; the double refraction is negative in sign.On examining the crushed crystals microscopically, pieces may sometimes be found in which the optic axial interference figure occupies the centre of the field; some pieces show a uniaxial figure, whilst in others the interference figure is biaxial with a very small optic axial angle. ''On comparing the above description with that of aconitine given by Tutton (Trans., 1891, 59, 288) or of pseudaconitine by Pope (Trans., 1897, 71, 352), co points of similarity are traceable ; it must there- fore be concluded that japaconitine is crystallographically different from both aconitine and pseudaconitine." Japaconitine is very soluble in acetone, wet alcohol, and chloro- form, and, when amorphous, in dry alcohol and in ether.When crystalline, it slowly dissolves in boiling ether and alcohol. I t is almost insoluble in dry light petroleum, and is but slightly soluble in water. It is readily crystallised by the addition either of water to its solution in alcohol or in acetone, or of petroleum t o its solution in chloroform, or in alcohol and ether. The pure base melts at 204.5" (corr.) to a clear, pale yellow liquid, which, a few seconds later, effervesces rapidly with the evolution of acetic acid. XpeciJic Rotation of Japaconitine Lubbe (Zoc. cit.) has stated that a 3 per cent. alcoholic solution of japaconitine is inactive. We find, on the contrary, that both alcoholic and chlorotorm solutions are dextrorotatory, like those of aconitine (Dunstan and Ince, Trans., 1891, 59, 281).VOL. LXXVIJ. E50 DUNSTAN AND READ : JAPACONITINE ( I ) In chloroform. [a], = + 20.26' (mean). ,lo" = + 0.622' ; I = 2 dcm. ; c = 1.602 ; [all: + 19.41'. d 5 " = +0.6'; 1=2 dcm.; c=1*42; [a]: +21*12O. ,lSSO = + 17" 2'; I= 2 dcm. ; c=0*605 ; (2) In alcohol. [aID= +23*6O. The specific rotatory power of jspaconitine is therefore much greater than that of aconitine, which is only + 11Oin alcohol (Dunstan and Ince, Zoc. cit.), whilst in chloroform we find it to be + 14.61Oat 18'. + 2306~. Determination of Methoxyl Groups in Japaconitine. The number of methoxyl groups in the base was estimated in the 0.1806 gave 0.268 AgI. CH,O = 19.58. C,,H,707N(OCH,)4 requires CH,O = 19.1 6 per cent.usual manner, using a slight modification of Zeisel's method. 0.1852 ,, 0.279 AgI. CH,O= 19.87. Japaconitine therefore, like aconitine, contains four methoxyl groups. Combustions of Japaconitine, The pure base was burnt in a current of oxygen, using cupric oxide 0.2857 gave 0.6534 CO, and 0.2005 H20. and a silver spiral. C = 62.37 ; H = 7.79. 0.2814 ,, 0.6410 CO, ,, 0.1790 H,O. C = 62.12 ; H=7*06. 0.1888 ,, 0.4338 CO, ,, 0.1198 H,O. C=62.66; H=7.05. 0.2049 ,, 0,4740 CO, ,, 0.1322 H20. C = 63-09 ; H= 7.16. Mean, C = 62.56 ; H = 7-26 per cent. Two combustions for nitrogen by the absolute method, using lead ehromate, furnished the following data : 0.3035 gave 8.892 mg. moist nitrogen. N= 2-92. 0.3105 ,, 7.864 mg. moist nitrogen. N= 2.53.C34H490,,N requires C = 63.06 ; H = 7-57 ; N = 2.16 per cent. We therefore adopt this formula provisionally as best representing the composition of japaconitine. This formula differs only very slightly from that proposed for aconitine by Freund and Beck (Ber., 1894,27, Salts of Japaconitine. 433). Japaconitine, like aconitine, furnishes a number of well crystallised salts. Japaconitine hydvochlmide is readily obtained by dissolving the pure base in dilute hydrochloric acid, the solution being left faintly acidAND THE ALKALOIDS OF JAPANESE ACONITE. 5 1 to litmus. The salt crystnllisee out in plates melting at 149-150'. When crystallised by dissolving in alcohol and precipitating with ether, lustrous rosettes of hexagonal plates, melting at 1 49-150°, are obtained.The salt crystallised from aqueous alcohol and ether contains three molecular proportions of water. 0.7800 lost 0.0640 after 1; hours at 105-110°. 0.3230 gave 0.0604 AgC1. Cl=4.62. C,,H,,O,,N,HCl+ 3H20 requires C1= 4.81 ; H20 = 7*3 per cent. It has already been shown (Dunstan and Ince, Zoc. cit., Dunstan and Cam, Trans., 1897, 71,350) that although aconitine and pseudaconitine are dextrorotatory, their salts are lsvorotatory. Japaconitine also exhibits this peculiarity. The tri-hydrated salt dissolved in water was used for the following determination : Loss =8*2. = - 1.466' ; I = 2 dcm. ; c = 3.076 ; [ u ~ ~ " = - 2 3 . 8 O . A determination of the specific rotation of aconitine hydrochloride was made a t the same temperature for comparison : u1V= - 1.433'; I = 2 dcm.; c = 2.315 ; [cty:= - 30.9'. Japaconitir~e hydrohomide crystallises readily from water, or from alcohol and ether, The crystals deposited from the latter solvents are very similar to those of the hydrochloride, being rosettes of hexagonal plates. The salt crystallised from aqueous alcohol and ether contains four molecular proportions of water. 1.2860 lost 0.1044 after 2 hours at 105-110'. 0.3060 gave 0.0726 AgBr. C,,H,,O,,N,HBr + 4H20 requires Br = 10.00 ; H,O = 9.00 per cent. Japaconitine hydriodide crystallises from water in minute rosettes melting at 195-197'. When recrystallised from alcohol and ether, the salt melts at 207*5-208*5°. Japaconitine nitrate crystallises from water in minute rosettes of needles melting a t 173-177'. When recrystallised from alcohol and ether, the salt contains one molecular proportion of water and melts a t 194', effervescing sharply a t 199'.The water of crystallisation is not lost until the salt is heated to 115--120'. When dried, the salt melts at 172-173'. Loss =8*11. 0,3320 ,, 0.0278 ,, ,, 100-110'. LOSS ~ 8 . 3 7 . Br = 10.09. After being dried at looo, 0.5315 lost 0.01 a t 115-120' for 18 hours. The hydrated salt gave the following numbers on analysis : Loss = 1.9 per cent, E 252 DUNSTAN AND READ : JAPACONITINE 0.1762 gave 0.3620 CO, and 0.1 145 H,O. C = 56.03 ; H = 7-29. C,,H,SOllN,HNO, + H,O requires C = 56.04 ; H = 7.14 ; H,O = 2.47 per cent. Japacondtine thiocyanate is prepared by adding an aqueous solution of ammonium thiocyanate to an aqueous solution of japaconitine hydrochloride, The crystalline salt, melting at 1 20°, is immediately precipitated if the alkaloidal solution is strong.When dissolved in alcohol and precipitated with water, the salt crystallises in lustrous needles. These melt at 190-1 92'. Aconitine thiocyanate, crystal- lised from water, melts at 145'; crystsllised from alcohol and ether, it melts at 193-195'. Japaconitine aurichlovide is obtained as a bulky, amorphous, canary- yellow precipitate by the addition of an aqueous solution of auric chloride to an aqueous solution of japaconitine hydrochloride slightly acidified with hydrochloric acid. The precipitate is quickly filtered, washed with water until the washings are neutral, and then dried in a vacuum. The amorphous aurichloride is readily soluble in alcohol, chloroform, or acetone, slightly soluble in water.A few minutes after the amorphous salt is dissolved in a little alcohol, the greater part of it separates in minute, golden-yellow, opaque needles which melt at 231' (uncorr.). The same salt may be prepared by precipitating (1) methyl or ethyl alcohol, or acetone solutions with ether or with water, (2) chloroform solutions with ether or with ether and petroleum. In each case, crystalline japaconitine aurichloride, melting at 231', is obtained. On two occasions we have obtained yellow crystals of an auri- chloride melting at a much lower temperature. First, by dissolving the amorphous salt in dry chloroform and allowing the solvent to evaporate spontaneously, yellow prisms separated. These melted in- distinctly at 152-154'.They were recrystallised by dissolving in chloroform and precipitating with ether, when similar prisms melting at 153' were obtained. On attempting to recrystallise these from alcohol, it was found that they Separated from that solvent almost immediately after solution, in the opaque, small, canary-yellow rosettes which we find t o be so characteristic of japaconitine auri- chloride. On the second occasion, a crystalline, golden-yellow aurichloride, melting indistinctly at 154-160°, was obtained after the addition of dry, light petroleum to an alcoholic solution of the amorphous salt. These crystals dissolved in chloroform, and on the addition of ether were precipitated in opaque rosettes melting at 231'. They were therefore japaconitine aurichloride.One of us has already shown (Dunstan and Jowett, Trans., 1893, These rosettes melted at 231'.AND THE ALKALOIDS OF JAPANESE ACONITE. 53 63, 994) that aconitine aurichloride may exist in three distinct iso- meric forms, the nature of the solvent and of the precipitant being apparently the factors which determine the production of these modi- fications. Although we have been unable to find the exact conditions for the formation of the salt of japaconitine melting at 154-156", it is evident from these results that japaconitine aurichloride exists in a t least two isomeric forms. The stable modification melting a t 231", being that most generally produced, we propose to name japaconitine a-aurichloride, and the unstable variety, melting indistinctly at 154-1 6O0,japaconitine P-ccurichZo&te.The melting points of aconitine a-, /3-, y-aurichlorides are 135.5O, 151-152", and 176" respectively. I n this connection, i t is of interest to note that japaconitine and aconitine, when mixed, may, by a tedious process, be separated by repeatedly fractionating the aurichlorides of the mixed bases by crystallisation from alcohol and from a mixture of alcohol and ether. Japaconitine a-aurichloride, melting a t 231°, crystallises out in the first fractions, and aconitine P-aurichloride, melting at 151-152" in the later fractions. Japaconitine a-aurichloride forms neither a hydrate nor an alcoholate. The crystalline salt, after being dried on a porous tile, lost no weight, either when allowed to remain in a vacuous desiccator for several days, or when heated in the water oven.A determination of the gold and chlorine gave the following results : 0.2408 gave 0.0474 AU and required 9.8 C.C. of N / l O AgNO, (of which 983 C.C. =3*55 C1) for complete precipitation of the chlorine. Whence AU = 19.68 ; C1= 14.69 per cent. The salt was burnt for carbon and hydrogen, but the determina- tion of the latter was of no value. 0,2014 gave 0.3010 CO,. C= 40.76 per cent. per cent, C3PH49011N,HAuC14 requires AU = 19.95 ; C1= 14.38 ; C = 41 *33 Physiological Action of Japaconitine. The physiological action of japaconitine is being investigated by Professor Cash. It may be etated here that it is an intensely toxic alkaloid which, in general, acts like aconitine, the resemblance between the two alkaloids being very close indeed.It produces the characteristic tingling of the lips, tongue, and skin, with long-continued tactile and thermic perception at the seat of applicatioii. It slows and then accelerates the heart, producing a sequence of ventricular upon auricular action. The death occasioned by it appears, however, to be primarily respiratory. Whilst the lethal59 DUNSTAN AND READ : JAPACONITINE dose of japaconitine towards frogs (Ranu esculenta and R. tempomria) and mammais is not identical with that of aconitine, the difference in toxicity is not great, moreover the variation in the effect produced by an equal dose is one rather of degree than of kind. Action of Methyl Iodide on Japaconitine. MethyGapaconitine. When japaconitino is heated in a closed tube to a temperature of 110-112' with a slight excess of methyl iodide, a crystalline methiodide is formed.This methiodide crystallises from the mother liquor with extreme readiness in large rosettes of colourless needles which melt 0.2240 dissolved in alcohol and precipitated with silver nitrate, st 224-225'. furnished 0.0690 AgI. I = 16.64. C,4H4,011N,CH3T requires I = 16.09 per cent. On adding dilute potash to an aqueous solution of japaconitine meth- iodide, a flocculent precipitate of rnethygapaconitike, C3,H,,0,,N*CH3, is thrown down. This base may be readily extracted by chloroform or ether, and crystallisqs from the latter in minute rosettes of colourless needles, which melt a t 206'. A combustion of the pure base furnished the following data : 0*1142 gave 0.2670 CO, and 0.0790 H,O.Methyljapaconitine ccuyichloride crystallises from a mixture of alcohol and ether in minute rosettes which melt at 223-225O. The readiness with which japaconitine furnishes a methyl derivative is another point of difference between it and aconitine (Dunstan and Jowett, Proc., 1894, 10, 96). C = 63.76 ; H = 7.69. C35H51011N requires C = 6354 ; H = 7.71 per cent. Action of Acetyl Chloride on Japaconitine. By allowing acetyl chloride to act on japaconitine for some hours a triacetyl derivative melting at 166' is formed. I n addition to tbis, a small quantity of a derivative very soluble in aIcoho1, and melting at 184-186', is produced, but the amount obtained was too small for further examination. Priucety$'apaconitine, C34H46011N( COO CH,),, crys tallises in colourless rosettes melting at about 166'.It is not very soluble in alcohol, and may thus be easily separated from the derivative melting at 184-186'. It is soluble in ether and in chloroform, but insoluble in water. Tricccetyljccpaconitine. 0.1043 gave 0.2357 CO, and 0,0697 H,O. 0-2475 C = 61-63 ; H = 7.41. ,, on hydrolysis with potash, 29.5 per cent. acetic acid. CaH55014N, requires C = 62.09 ; H = 7-1 1 ; C2H,0, = 31.04 per cent.AND THE ALKALOIDS OF JAPANESE ACONITE 55 11. JAPBENZACONINE. Hydrolysis of Japaconitine. When japaconitine is hydrolysed, it furnishes acetic acid, and a new crystalline base, which from its analogy to benzaconine, the hydrolytic base from aconitine, we purpose naming jupbenxaconine.This hydro- lysis may be effected in neutral, in alkaline, or in slightly acid solution. The last is the most convenient, since in the presence of alkalis, the new base, japbenzaconine, readily suffers further hydrolysis into benzoic acid, and another base, japaconine (vide ilzfra). Although any of the salts of japaconitine may be used for the hydrolysis, we have found the sulphate to give the most satisfactory results. 0.41 22 gram of japaconitine was neutralised with dilute sulphuric acid and the solu- tion heated in a sealed tube for 9 hours at 115-130'. The slightly discoloured solution was then neutralised with soda. A definite amount of decinormal sulphuric acid was now added and the benzoic acid, a small quantity of which is invariably produced, removed by shaking with benzene.The aqueous portion and the washings from the benzene were titrated with soda (1020 C.C. = 4.9 H2S04) of which 6.6 C.C. were required, equivalent to 9-41 per cent. acetic acid. The equation C34H49011N + H,O = C2H,02 + C,,H,701,N requires 9-27 per cent. The identity of the acid was proved by an analysis of its silver salt. 0.1485 gave 0.0957 Ag. C2H,02Ag requires Ag = 64.66 per cent. The identity of the benzoic acid, extracted by benzene, was deter- mined by its melting point, which was 121', as well as by other of its physical characters (vide infra). No methyl alcohol or other product could be detected. The hydrolysis, therefore, takes place according to the equation given above. Wright and Luff do not record the production of acetic acid during the hydrolysis of japaconitine, nor did they observe the formation of japbenzaconine.Their account of the hydrolysis is that the japaconitine is decomposed, forming japaconine and benxoic acid. Ag = 64.44. Properties of Japbenxaconine. The new base is precipitated from the aqueous solution of its salts by the addition of alkali, and may be partially removed by shaking with ether. Its complete removal necessitates a vigorous shaking with chloroform. By the evaporation of its ethereal solution, jap- benzaconinc is obtained as a, colourless varnish. The latter readily56 DUNSTAN AND READ : JAPACONITINE dissolves in dilute hydrochloric acid, the solution, after partial evapo- ration and stirring, yielding a crystalline salt. Although no difficulty attends the crystallisation of the salts of japbenzaconine, many at- tempts were made to crystallise the base before it was accomplished.The following method finally proved successful. A slight excess of dilute ammonia was added to an aqueous solution of the pure hydrochloride, the white flocks of the precipitated base dissolved by shaking with ether, and the ethereal solution washed and dried over calcium chloride. The dried solution was filtered into a stoppered bottle, and sufficient light petroleum added to produce a faint opalescence. A considerable quantity of amorphous japbenzaconine separated, and after standing 2 or 3 days the clear mother liquor was poured off into another bottle. After several days, minute rosettes appeared. These melted at 176"; further small quantities of crystals having the same melting point were obtained by the careful addition to the mother liquor of small quantities of light petroleum. When purified by recrystallisation from a mixture of ether and petroleum, the base melted a t 183'.It could also be crystallised by dissolving in alcohol and precipitating with water. The crystals thus obtained melted at about 180'. Jspbenzaconine, therefore, diff ers very markedly from benzaconine in the fact that it may, when pure, be fairly readily crystalliaed. The only well-defined crystals we have obtained were rhombohedra1 plates. When dissolved in dilute acids, the solution is distinctly bitter, and quite free from the tingling sensation so characteristic of japaconitine and of aconitine. The base, crgstallised from ether, is anhydrous.0.1967 lost 00003 after 1 hour at 103", that is, 1.53 per cent. Spec;fic rotation of japbenxmonine. I n alcohol. C Z ~ * ' ~ = + 1.633'; 1=2 dcm. ; c=2.033; 1aJ1g5"= f40.16". The specific rotation of benzaconine is + 4.48 (Dunstan and Harrison, Japbenzaconine is therefore nearly ten times Trans., 1893, 63, 443). more optically active than benzaconine. Combustions of japbenzaconine : 0.1922 gave 0.4510 CO, and 0.1505 H,O. C= 63.99 ; H= 8.7. 0.1857 9, 0.4365 CO, ,, 0.1252 H,O. C=64*10; H=7*49. C,,H&,,N requires C = 63.47 ; H = 7.76. Salts of Japbenxaconine. The salts of japbenzaconine crystallise with extreme readiness. Japbenxaconine hydrocklo&ie crystttllises f 1'0111 water in Minute When recrystallised from alcohol and rosettes melting at 244-2459AND THE ALKALOIDS OF JAPANESE, ACONITE.57 ether, the melting point is raised t o 253'. alcohol and ether, contains one moleciilar proportion of water. The salt, crystallised from 0.2234 hydrated salt lost 0.0044 H,O after 1; hours at 100'. 0.2009 hydrated salt lost 0.0057 H,O after heating hour at 110'. 0.1950 hydrated salt gave 0.0416 AgC1. Cl=5*27. H,O = 1.97. H,O = 2.33. 0.2066 ,, ,, 0,0446 AgC1. C1= 5.34. C,,H,70,0N,HC1 + H,O requires H,O = 2.73 ; 01 = 5.38 per cent, Specific rotation of the hydrated hydrochloride in water. d2"" = - 0.8'; .J= 2 dcm. ; c = 2.028 ; [ a]:30 = - 19-73' The specific rotation of benzaconine hydrochloride, dissolved in water, was taken for comparison. ~~'23" = 0.7C5O ; I = 2 dcm. ; c = 1.492 ; [ a]:30 = - 25.1 3'.Combustions of the anhydrous hydrochloride. 0,1894 gave 0.4175 CO, and 0.1240 H,O. C = 60.11 ; H = 7.27 0.1925 ,, 0.4242 CO, ,, 0.1281 H,O. C=60*09 ; H=7*39. C,,H,70,0N,HC1 requires C = 59.95 ; H = 7.49 per cent. Jnpbenxccconisie Iqdrobrornide crystallises from water in minute prisms, and- from alcohol and ether in rosettes melting indistinctly Japbenzaconine aurichloride is precipitated in amorphous, yellow flocks on mixing aqueous solutions of auric chloride and of japbenz- aconine hydrochloride. The dried amorphous salt crystallises immedi- ately upon t h e addition of dry alcohol, the crystals melting at about 2 12O. The salt may be readily crystallised from alcohol or chloroform hy precipitation with ether. The melting point varies with the nature of the solvent, the crystals from alcohol melting at 219', and those from chloroform and ether at 288'.By the addition of water to a n alcoholic solution, the salt is precipitated in oily drops. The crystal- line aurichloride is anhydrous. Analysis of the salt furnished the following results. 0.3320 gave 0.0675 Au and 0.189 AgCl. :tbout 205-217'. Au= 20.33 ; C1= 14.08. C,,H,70,0N,HAuCI, requires A u = 20.84 ; C1= 15.02 per cent. The hydrochloride of the regenerated base crystallised readily from water, melted at 246', and was identical with japbenzaconine hydro- chloride. A urichlo?.-jcc3,beiLxa con ine. -1V he 11 a I I a1 coholic so1 u ti on of j ap- benzaconine aurichloride, t o which a few drops of light petroleum58 DUNSTAN AND READ : JAPACONITIXE have been added, is allowed to stand for about a week, colourless, well- defined, lustrous octahedra are obtained.These octahedra, after re- crystallisation from alcohol and ether, melt a t 178" with the separation of a considerable quantity of gold. The mother liquor from which these crystals separated furnished a small quantity of colourless rosettes of needles melting at about 230" and containing gold. The quantity obtained was unfortunately too small for further examination. The octahedra, melting at 1 7 8 O , crystallise without any attached solvent. An analysis furnished the following data : 0.2332 gave 0.053 Au and 0.078 AgC1. C,2H,,0,0N=AuC12 requires Au = 22.6 ; C1= 8-23 per cent. The hydrochloride of the regenerated base crystallised from alcohol and ether in beautiful, lustrous rosettes melting a t 252-253", and identical with j ap benzaconine hydrochloride. Au=22.72 ; Cl=&27 III.JAPACONINE. Hycli*oZysis of Jccpbenxaconine. Japbenzaconine is slowly hydrolysed by strong alkalis or by cold dilute sulphuric acid, more rapidly on heating. The complete hydrolysis is, however, by no means easy to effect, since the continued heating with alkalis leads to the resinification of the hydrolytic products. This is obviated to a certain extent by carrying out the hydrolysis either in alcoholic solution, using an alcoholic solution of sodium hydroxide, or in a sealed tube. The former method was the one which we foucd most satisfactory, and gave fairly concordant results when the percentage of benzoic acid was determined.2.9015 grams of japbenzaconine were dissolved in alcohol, warmed, and a 30 per cent. alcoholic solution of sodium hydroxide added. The mixture was allowed to stand for 24 hours. Dilution with water pro- duced no precipitation of japbenznconine, showing that the hydrolysis was complete. The solution was neutralised with sulphuric acid, the alcohol removed by evaporation on a water-bath, and the aqueous solu- tion, after acidifying with sulphuric acid, extracted with benzene. The benzene solution was extracted with dilute soda, and the latter acidified with sulphuric acid and extracted with ether. The dried ethereal solu- tion furnished, after careful evaporation, 0,589 gram of benzoic acid (m. p. 121*5O) =19*98 per cent. The following equation therefore represents the hydrolysis, the calculated quantity of benzoic acid being 20.16 per cent.C52H47010N + HgO = C,H,* CO2H + C'25H,,O,N. The acid solution from which the benzoic acid had been removedAND THE ALKALOIDS OF JAPANESE ACONITE. 59 was now rendered alkaline, and extracted, first with ether and then with chloroform. The ether removed nothing from the solution, and the chloroform only a trace of alkaloid which was readily extracted again by shaking with water. The residual solution was therefore neutralised with sulphuric acid, and evaporated to dryness. The residue, most of which was sodium sulphate, was extracted with alcohol, the alcoholic solution made alkaline with sodium hydroxide, evaporated to dryness, and the brown alkaloidal residue dissolved in chloroform ; the chloroform solution gave, after evaporation, a very hygroscopic, dark brown varnish.This varnish, which was strongly alkaline to litmus and readily reduced Fehling's solution, was purified by neutralising with aqueous sulphuric acid and removing the greater part of the colouring matter by boiling the solution with charcoal. The pure base, japaconine, was obtained by precipitating the solution of the sulphate with barium hydroxide, filtering, and evaporating the aqueous solution. It was finally purified by the fractional precipitation of a solution in chloroform with ether. Most of the colouring matter is precipitated in the first fractions, the later fractions being almost colourless japaconine. Pwperties and Composition of Japc6conine. Thus purified, japaconine is a colourless, exceedingly hygroscopic, powder which we have so far been unable to crystallise.It is readily soluble in water, alcohol, chloroform and acetone, but almost in- soluble in ether and petroleum. It melts indefinitely between 9 7 O and 100'. Its specific rotation in water is + 1 0 . 8 8 O . algO = +0*4"; Z=2 dcm.; c=1.837; [ u ] F = +10*88O. Combustions of the amorphous japaconine. 0.2010 gave 0.4480 CO, and 0.1460 H,O. C = 60.78 ; H = 8.07. 0.1855 ,, 0.4058 CO, ,, 0,1193 H,O. C = 59.65 ; H= 7.14. 0.1634 ,, 0.3645 GO, ,, 0.1108 H,O. C = 60.17; H= 7.53. C,,H,,O,N requires C = 59.88 ; H = 8-58 per cent. The salts of japaconine crystallise with great difficulty. None crystallise from water, and those which we have crystallised from alcohol and ether are so hygroscopic that we have, unfortunately, been unable t o examine them completely.Japaconine hydrobromide crystallises from alcohol and ether in rosettes of colourless, triangular plates melting sharply at 22 lo. Jccpaconine hydrochloride crystallises froin alcoholic solution after the addition of ether. Japaconine oxalute crystallises in colourless, dumb-bell shaped rosettes.60 DUNSTAN AND READ$: JAPACONITlNE A strong aqueous solution of the hydrochloride gives no precipitate with auric chloride or platinic chloride, arid the aqueous solution of the aurichloride is quickly reduced. Fehling's solution is also reduced when boiled with a solution of japaconine. In its general behaviour, japaconine resembles aconine ; its salts, however, crystallise much less readily than those of aconine.IV. PYROJAPACONITINE. Efeect of Heat on Japaconitine. It has already been shown (Dunstan and Carr, Trans., 1894, 65, 1'76 ; 189'7,71, 350) that aconitine and pseudaconitine decompose when heated at their melting points, furnishing acetic acid and new bases which were named pyraconitine and pyropseudaconitine respectively. We find that japaconitine behaves in a similar manner. When heated at about 200-210' for 10 minutes, the base gradually darkens, and finally melts, with the evolution of acetic acid. The amount of the latter, estimated by titration with decinormal soda solution, varies from 8.4 t o 8.9 per cent. The amount required for a loss of one molecule of acid from one molecule of japaconitine (C,,H,90,,N - C,H,O,) = 9-27 per cent.I n addition to the loss of the acetic acid, there is almost invariably a further loss varying from 3 to 4-5 per cent., the nature of which has not been determined. The dark brown varnish which remained after the evolution of the acetic acid was readily purified by dissolving in dilute sulphuric acid, exactly neutralising the solution with soda, and boiling with charcoal for a few minutes. Ammonia was added to the colourless solution thus obtained, and the bulky, flocculent precipitate of the pyro-base extracted by shaking several times with ether and finally with chloroform. The ethereal solution was dried with calcium chloride and evaporated to dryness, a pile yellow varnish being thus obtained. This varnish was dissolved in dilute bydrobrornic acid, and the solution evaporated t o a small bulk on a water-bath.After vigorous stirring, minute crystals of the salt separated, which melted a t 240° (pyraconitine hydrobromide melts at 282'). The base was regenerated and extracted with ether, and the ethereal solution, after partial evaporation, dried with calcium chloride, filtered into a stoppered bottle, and light petroleum added until a n opalescence formed. Colourless needles of the base crystallised after standing a few days. These, after crystallisation, melted at 167-- 168'. The alkaloid, like pyraconitine, is lzvorotatory. 2 0 " = - 3.33' ; Z = 2 dcm. ; c = 2.529 ; [u]z = - 65.89.AND THE ALKALOIDS OF ,JAPANESE ACONITE. 61 On combustion, tohe crystalline base furnished the following results : 0.1081 gave 0.2582 CO, and 0.0685 H,O.C=G5*14; K= 7.04. 0.1079 ,, 0.2572 (20, ,, 0.0652 H,O. C=64*92 ; H=6.71. C32H4,OgN requires C = 65.41 ; H = 7.66 per cent. A methoxyl determination by Zeisel's method shows that pyro- 0.1950 furnished 0.2810 AgI. C3,H,,0gN requires, for 3 inethoxyl groups, CH,O = 15.8 per cent. japacotdine, like jspaconitine, contains four methoxyl groups. CH,O = 19.00. 9 9 9 9 4 ?, ,, CH30=21-l ,, Pyrojapaconitine furnishes well crystallised salts. P?pojapaconitine hydrobromide sepasates from water in minute, colourless crystals melting a t 241". When this salt is recrystallised by dissolving in alcohol and precipitating with ether, it melts at 208-2209", but, after solidification, it remelts at about 237-238'. The melting point of this (208') variety is unchanged by repeated recrys- tallisation from dry alcohol and ether, but when recrystallised from water the salt melts at 241'.The salt, crystallised from its aqueous solution, contains two mole- cular proportions of water, one only of which is lost a t 100". 0.7140 lost 0.0293 H,O after 2 hours at 100'. H,O=4.10. 0.4415 ,, 0.0180 H20 ,, 70 minutes a t 100-130°. H,O = 4.8. The speci6c rotation of pyrojapaconitine hydrobromide in water is - 102.5°. - - 5.616'; I = 2 dcm. ; c = 2.7388 ; [ CC]~D~.= - 102.5". On analysis, the salt, dried a t looo, furnished the following data : 0-1815 gave 0.3706 C02 and 0*1030 H,O. 0.1910 , I 0.3910 CO, ,, 0.1110 H,O. C=55*83 ; H = 6.45. 0.4235 ,, 0.1160 AgBr. Br = 11.65. C = 55 68 ; H = 6-30. C3,H,,OSN,HBr + H,O requiresC = 55.97; H = 6.99; Br = 11.66 per cent.Unsuccessful attempts were made to prepare an anhydrous hydro- bromide by a direct method. Pyrojapaconitine aurichloride is precipitated in bulky, yellow flocks on the addition of aqueous auric chloride to an aqueous solution of pyrojapaconitine hydrochloride. It crystallises very easily from its solution in alcohol or chloroform after either spontaneous evaporation or the addition of ether. The melting point of the salt varies with the solvents from which it is crystallised, t h a t from chloroform melt- ing at about 160-161', and that from alcohol and ether at about 188-1 89".62 DUNSTAN AND READ : JAPACONITINE The following results wore obtsinetl by the analysis of the salt (m. p. 360-161") crystallised from chloroform : 0.3389 gave 0.0697 Au and 0*2025 AgC1.Au = 20.56 ; C1= 14.78. The salt (m. p. 188-189') crystallised from alcohol and ether gave 0.1742 gave 0.2656 CO, and 0.0740 H,O. C = 40.01 ; H = 4.72. 0.1694 ,, 0.2502 CO, ,, 0.0696 H20. C=40*28 ; H=4*56. C32H4,0gN,HAuCl, + H,O requires C = 40.63; H = 5.07; AU = 20.84 ; C1= 15.02 per cent. C32H,509N,HAuC'1, requires C = 41 -43 ; H = 4-96 ; AU = 21 -25 ; C1= 15-31 per cent. the following data : No aurichlor derivative could be obtained. V. PYBOJAPACONINE. Hydrolysis of Pyrojapaconitine. Pyrojapaconitine is hydrolysed with extreme readiness by alkalis, The products of the hydrolysis being complete in less than an hour. the reaction are benzoic acid and a new base, pyrojapaconine, C,,H,,O,N + H20 = C6H5* C02H + C2,H4,0,N.Pyrojapaconine may be extracted from its aqueous alkaline solution by shaking either with ether or chloroform, preferably the latter. Light petroleum precipitates the substance from the ethereal solution in colourless, amorphous flakes melting at about Z 23 -1 2 8 O . The base is very deliquescent. Its specific rotation is - 73.96'. a2'= -4.3'; Z=2 dcm.; ~ = 2 . 9 1 0 6 j [a]r=73.96°. Combustions of pyrojapaconine furnished the following results : 0.1495 gave 0.3385 CO, and 0.1060 H,O. C! = 61-75 ; H = 7 87. 0.1499 ,, O933'70 CO, ,, 0.1035 H,O. C=61*31; H=7*67. C,,H,,O,N requires C = 62.1 1 ; H = 8.48 per cent. All attempts to crystallise either the base or its salts were unsuccess- The aurichloride is soluble in warm water, but separates in a It is soluble in alcohol, and the deep yellow ful. vitreous mass on cooling.solution becomes colourless after long standing. TI. THE AMORPHOUS ALKALOIDS OF JAPANESE ACONITE. In working up the crystalline japaconitine used for the preceding Work, a small amount of an amorphous base was accumulated. This base furnished a hydrochloride, which crystallised readily fromAND THE AJ,RAJ,OIDS OF JAPANESE ACONLTE. 63 water in the minute crystals so characteristic of japbenzaconine hydrochloride, and melted at 241-242'. By rccrystallisation from alcohol and ether, the melting point was raised to 24GD. This is the melting point of japbenzaconine hydrochloride. I n order t o completely confirm the identity of this substance, a small quantity was converted into the aurichlor derivative by adding light petroleum to an alcoholic solution of the aurichloride.Colourless octahedra of aurichlorjccpbenxnconine me1 ting at 179' separated. The amorphous base which accompanies japaconitine io Japanese aconite roots is therefore japbenzaconine. The amount is exceedingly small (about 1 gram from l l g kilos.) as compared with the quantity of benzaconine contained in the roots of Aconitum Napellus. The results of this investigation, which has occupied us for several years, seem to leave no room for doubt that the crystalline alkaloid (japaconitine) of Japanese aconite, though closely resembling aconitine, the crystalline alkaloid of A. Napellus, is chemically a distinct individual. The chief differences which we have established between them and their derivatives are summarised in the folIowing tables : 1.-Japaconitine and Aconitifie.Substance. Ja paconitine ..................... Aconitine ........................ Japaconitine hydrochloride,. . Aconitine hydrochloride ...... Japaconitine hydrobromide.. . Aconitine hydrobroniide ...... Japaconitine hydriodide ...... Aconitine hydriodide ......... Japnconitine aurichloride ... Aconitine aurichloride ......... Triacetyljapaconitine ......... Triace tylaconitine ............ Melting point. 204'2" 196-1 97' 149-150" 149" 172-173' 163" 207-208" 226" (a) 135'5' ( B ) 152" ( Y ) 176" 166-167 207" Specific rotation. + 20.25" in chloroforrr + 14 -6" in chloroform - 23 '8" in water - 30.9" in water - 30 .So in water Remarks. Prismatic needles Hexagonal prisms64 JAPACONITINE AND THE ALKACOIIM OF JAPANESE ACONITE. I1 .--Japbenxaconine crnd Renxnconine. Melting point. Substance. Specific rotation. Remarks. Japbenzaconine ................. Benzacoiiiiie .................... f 40*16" i i i alvohol + 4 '48" in alcohol ltcctnngular plates Uncrystrtllised 182-1 83" 125" JapSenzaconine hydrochlorid Renzaconine hydrochloride . . - 19'73" in water - 28.7" in water 205-217" 282" Japbenzaconine hydrobromidl Benzaconine hydrobromide.. Japbenzaconine aurichloride. Benzaconine aurichloride.. ... Aurichlor-japbenzaconine ..... Aurichlor-benzaconine ....... 228" 125-135" 178" 204" 0 c tahedra 3ectanplar prisms Japaconine and Aconine. Melting point. Substance. Specific rotation. Remarks. Japaconine ....................... Aconine ........................... 97-100" 132" -t 10.88" in water f23" in water Amorphous, very hygroscopic Amorphous, very hygroscopic Japaconine hydrobromide . . , 221" Uncry stallised Japaconine hydrochloride . . I Aconine hydrochloride.. ....... 175'5" - 7-71" in waterSEPARATION OF ISOMERIDES CONTAINED IN XYLIDINE. 65 Pyrojapaconitine and Pyraconitine. Substance. Py rqjapaconitine .............. Pyraconitine .................... Pyrojapaconitine hgdro- chloride Pyraconitine hydrochloride.. Pyrojapnconitine hydro- bromide Pyraconitine hydrobromide.. Pyrojapaconitine aurichlorid Pyraconitine aurichloride.. ... Pyrojapaconine ................ Pyraconine ....................... Melting point. i53--16a0 167'5" 175-176" 248 9 5 " 161" Specific rotation. - 85-89" in alcohol Inactive - 102'5" - 46'8" in water - 73'96" in water - 90.99' in water Remarks. Crystalline Uncrystallised Uncrystallised Uncrystallised The conclusions to be drawn from our results as to the composition and properties of japaconitine are not in agreement with those of Wright and his fellow-workers. Neither can we confirm the opinions which have been advanced by Blandelin, Liibbe, and by Freund and Beck, all of whom regard japaconitine as identical with aconitine. For the present, we reserve a discussion of the exact nature of the relationship between japaconitine and aconitine, as we are not yet in a position t o decide finally which formula to select for either alkaloid as most correctly expressing its composition and the composition of its derivatives, SCIENTIFIC DEPARTMENT, IMPEEIAL INSTITUTE, S . W.