A very small amount of a new alkaloid, eschscholtzidine, present in E. californica, was s11on.n to be eschscholtzine in lvhich a methylenedioxy group is replaced by two methoxyls.The present authors have recorded the isolatioll of a number of alltaloids from Eschscholtzia californica Cham.(1). The fraction which yielded eschscholtzine has now yielded a nev-base, escl~scholtzidine, which so far has not been crystallized either free or in the form of salts. I t was ultimately purified by means of column chromatography, and its purity checked by the thin-layer technique (R, 0.51 on silica gel with cyclohexanechloroform-diethylamine (7: 2: 1 ) ) .T h a t this allcaloid was related t o eschscholtzine (2) was suspected because of its chemical behavior and its source. The nuclear magnetic resonance spectrum confirlned this close relation, and in fact indicated the structure shown by I.("Tz-oh~e OhIe Its nuclear nlagiletic resonance spectrum was superposable on that of eschscholtzine, except that 1101~ t~v o methoxyls (T 6.18 (31-1) and 6.27 (3EI)) and a methylenedioxy group (T 4.23 (21-1)) were evident. I-Ience it appeared t o have structure I. This was readily confirmed by first delnethylenating the alltaloid t o the corresponding dihydroxy base followed by methylation with diazomethane. The product was identical with a specimen of argenlonine lcindly supplied by Dr. T. 0. Soine. EXPERIMENTALThe final mother liquors, from \vhich no further alkaloids could be crystallized either a s free bases or as salts (hydrochloride, nitrate, and hydrobromide) (from 270 kg of E. californica), were combined, dissolved in dilute hydrochloric acid, and filtered. The solutior~ was then extracted with several portions of chloroform. The alltaloids were regenerated from the fraction that was extracted with chloroform, dissolved in ether, and dried with potassium hydroxide.The solution was then passed through a long colun~n of alumina (250 g, p1-I 8.15, activity 0.98) and eluted with dry cther. The first fraction colored on exposure to air and was discarded. The large middle fraction was again passed through a second column, and its middle fraction proved to be chromatographically pure; yield 3 g. For analysis a portion was distilled a t 180' and 1 nlm.

Eschscholtzine, a new alkaloid from Eschscholtzia californica, has been shown to be the bismethylenedioxy analogue of 1V-methylpavine (argemonine). The mass spectrometric fragmentation of allcaloids of the pavine type is discussed.Two of us have reported the isolation of a new alkaloid, eschscholtzine, from Eschscholtzia californica Cham.(1) and a mass spectrum* of it showed the molecular formula to be C~~H I~O~N .In the meantime, a fuller mass spectrometric study of eschscholtzine a t Liverpool allowed the correct structure t o be derived which had independently been determined by nuclear magnetic resonance (n.m.r.) and chemical degradation a t Guelph. The mass spectrum of eschscholtzine shows a very small parent peak a t m/e 323 when the hot inlet system was used; however, by direct inlet of the sample, the parent peak increased to about one-third the size of the base peak. In both spectra (hot inlet or direct inlet) only one intense fragment peak appeared a t m/e 188. This behavior is similar to that of the 1-benzylisoquinoline alkaloids (2) but on this basis the only reasonable structure for the fragment ion is (I) which corresponds to m/e 190 and is thus unacceptable.However, the spectra of pavine (IV, R = H) and N-methylpavine (IV, R = Me) parallel that of eschscholtzine in showing a single intense fragment ion a t m/e 190 and one a t m/e 204. These ions are formulated as I1 (R = H) and I1 (R = Me) and their formation can be rationalized as shown below. The nature of the ion from eschscholtzine can now be understood as the aromatic ion (111) and its forination strongly indicated the methylenedioxy analogue (V) of N-methylpavine (IV, R = Me) (argemonine) as the structure of eschscholtzine.The striking similarity of the n.m.r. spectra of argemonine (IV, R = hlIe) (3, 4) and eschscholtzine (V) supported this indication. Both n.m.r. spectra exhibit impressive symmetry and they are largely superiinposable except for the slight shifts described in Table I. The aromatic and N-methyl protons call for no special comiment but the protons of the two methylenedioxy groups appear as a quartet of lines of almost equal intensity: 4.17, 4.21, 4.22, and 4.26 T . The two methylenedioxy groups are thus allnost equivalent but the two protons of the 0-CH2-0 system are not identical as a result of their being in different environments with respect to the aromatic rings. The eschscholtzine molecule is non-planar and may resemble a partly opened hinge. A similar non-equivalence of the methylenedioxy protons has been encountered in the aporphine alkaloids, diceiltrine and bulbocapnine methyl ether (5).

Die N‐Pyruvylderivate von Glycin, Alanin, Valin und Leucin werden auf folgendem einfachen Weg erhalten: Kupplung der Brenztraubensäure mit den Aminosäure‐benzylestern mit Hilfe von Phosphoroxychlorid in Tetrahydrofuran/ Pyridin und anschließende Abspaltung der Benzylreste mit katalytisch erregtem Wasserstoff. Zur Erkennung der Substanzen im Papierchromatogramm und ‐pherogramm dient Nitroprussid‐Na, dann NH3, wobei tiefblaue Flecken sichtbar werden.

Several lots of Eschscholbia californica have been investigated and the following alkaloids have been identified, in addition to a new allcaloid which has been named eschscholtzine: protopine, allocryptopine, cryptopine, chelerythrine, and lauroscholtzine, which is a new name for IV-methyllauroteta~line. Some plant lots did not yield cryptopine and not sufficient chelerythrine for positive identification.The presence of alkaloids, and particularly of protopine and of allocryptopine, in Eschsckoltzia calijornica Cham. has long been known (1, 2). Other alltaloids have been suspected and (or) reported and the author has recorded the presence of a new alkaloid, eschscholtzine (3).h~Iore plant inaterial has become available and a more complete examination is now reported. A slilall quantity of plant material (4 kg) grown locally yielded protopine, allocryptopine, cryptopine, chelerythrine, eschscholtzine, and a small amount of a phenolic base melting at 254' but not further characterized. The main lot (270 kg) did not yield this phenolic base nor cryptopine, and chelerythrine was found in only trace amounts and then evidently contaminated with sanguinarine. Escl~scholtzine has now been obtained crystalline and is C19M1704N. The phenolic fraction yielded N-methyllaurotetanine for which the trivial name lauroscholtzine is now proposed. EXPERIMENTALThe smaller lot of material (4 kg) consisted largely of plants grown fro111 seed of the wild plant. T h e large lot (270 kg) was grown from seed kindly supplied by Bodgers Seeds Limited, El Monte, California, and consisted of a mixture of cultivated varieties. The plants were grown on a plot of the Ontario Agricultural College, Guelph, Ontario, and special thanks are due to Dr. R. J. Hilton who made the necessary arrangements.The dried and gound plant material was extracted with methanol and the extract was largely freed of solvent. The residue was then acidified with dilute hydrochloric acid (Congo test paper) and the remaining solvent was expelled in a current of steam. While still hot paraffin was added to the mixture. After settling and cooling the lower aqueous layer was withdrawn, reheated with paraffin, and after cooling once more was filtered with the aid of Filtercel. The water insoluble portions were washed with a fresh portion of acidified water and this extract was added to a subsequent methanol extract. By this countercu>rent procedure it was possible to extract all of the alkaloid and do it with about 1 I of final solution per 2 kg of plant material.The total all~aloid mixture was recovered from the aqueous solution by chloroform extraction after making alkaline with ammo~lia. Much non-alltaloid material was obtained in the chloroform extract and this was largely eliminated by extracting the residue from the extract with dilute hydrochloric acid. The water soluble fraction thus obtained was extracted with chloroform (4) and yielded the following fractions:BC, bases whose hydrochlorides are extracted from aqueous solution with chloroform-eschscholtzine; ...

Previous work by many authors has led to the assumption that retamine might be (+)-12-hydroxysparteine. A partial synthesis of the enantiomorph of this compound has been effected by dehydration of (+)-13-hydroxylupanine and hydroboration of the product. The dehydration product consisted of two conlponents that were separated by thin-layer chro~natography and identified by the characteristics of their nuclear magnetic resonance (11.m.r.) spectra a s A1?J3-and A'3Jq-dehydrolupanine. Hydroboration of the A'?J3-isomer gave rise t o (-)-12-hydroxysparteine having, in thin-layer chromatography, the same Rf value a s natural retamine and the same optical rotation numerically, although of opposite sign. T h e synthetic base had the same infrared and n.m.r. spectra a s the alkaloid and the two had superimposable Debye-Scherrer patterns. Evidence is given showing the hydroxyl t o be equatorial.The lupin allraloid retainine, Clj~IZ60hT2, n1.p. 168", [a]D +46.2", was first isolated by Battandier and Malosse (I). I t was suggested by White (2) that it was probably a hydroxysparteine, and by Ribas, Sanchez, and Primo (3) that it was likely 6-hydroxysparteine (I). Later Ribas and Fraga (4) recognized that the properties of retainine could not be accominodated by a 6-hydroxy structure and, since the base could not be oxidized to a ketone, they suggested that it was either a 7-or a 9-hydroxy-(+)-sparteine. l'lore recently, in the light of new evidence, Ribas and his co-worliers (5) abandoned the 7-or 9-hydroxy structure in favor of 8-hydroxy-(+)-sparteine, noturithstanding the fact that retamine did not produce a lietone on oxidation. Finally, Bohlinann and his co-workers (6, 7) synthesized both epimers of each 7-hydroxy and 9-hydroxy-(+)-sparteine as well a s the t\vo epiiners of 8-hydroxy-(+)-sparteine and established that none of these was identical with retainine.From the results of Ribas et al. (5) the hydroxyl in retamine can be present only in ring A or ring D. Since, however, anhydroretainine obtained by the dehydration of the base a t high temperature gives rise on catalytic hydrogenation t o a mixture of sparteine and a-isosparteine, it must be concluded that the hydroxyl is located in ring D. Furthermore, retamine is not a carbinolainine and, therefore, the hydroxyl cannot occupy positions 11 or 15. Also both epimers of 13-hydroxysparteine are linown (8) and neither is identical with the alkaloid. Consequently, Bohlinann et al. (7) have suggested that retainine was either 12-hydroxysparteine I1 or 14-hydroxysparteine. I t was assumed that the dehydration of 12-hydroxysparteine would give rise to A1l,l2-dehydrosparteine, and since this product on catalytic hydrogenation should form a-isosparteine exclusively (9, 10) whereas the dehydration product of retamine, depending on the temperature a t which i t was