Experime&s on the Syathesis of Rotehow ctc. IWt XII. 279 57. Experimnts on the Synthesis of Rotenone and its Derivatives. Part XII. The 2 ;2-Dimethyl-A3-chromen Residue of Toxicarol. By WALTER BRIDGE,REGINALD G. HEYES,and ALEXANDER ROBERTSON. IN Part V (J. 1935 681) the structure of apo- and of dehydroapo-toxicarol was clearly demonstrated and the view was expressed that toxicarol and certain of its derivatives contain a 2 :2-dimethyl-A3-chromen nucleus involving the phloroglucinol residue analogous to that obtaining in deguelin and tephrosin (Part 11,J. 1932 1380). This sug- gestion was based mainly on the production of acetone from toxicarol and deguelin by hydrolytic fission in conjunction with the fact that compared with toxicarol coumarones and coumarans e.g.tubanol and isotubanol are relatively stable to hot aqueous sodium hydroxide and before directing attention to the remaining experimental details required for the complete structure of toxicarol it was considered essential that clear evidence of the presence of the chromen residue should be obtained In the first instance the oxidation of 0-acetyldihydrodehydrotoxicarolwas studied because it appeared reasonable to expect that if toxicarol contained the 2 :2-dimethyl-A3-chromen system it would probably give rise to an acid analogous to tephrosindicarboxylic acid (Part 11 loc. cit.). Efforts in this direction however have been entirely unsuccessful; e.g. the oxidation of the acetate of dehydrotoxicarol in acetone with aqueous potassium permanganate gave in addition to 2-hydroxy-4 :5-dimethoxybenzoic acid and rissic acid a neutral product (an acetate) m.p. 149” which did not appear to be suitable for further step-wise degradation. Conse-quently we turned our attention to the hydrolytic fission of dehydrotoxicarol and of dihydro-toxicarolic acid having as our objective the isolation of phenols analogous to tubanol and dihydrot ubanol. Degradation of dehydrotoxicarol with warm aqueous sodium hydroxide under a variety of conditions failed to give the expected product but scission of dihydrotoxicarolic acid with hot 50% reagent yielded a crystalline dihydric phenol CllHI4O3 which behaved as a satu-rated compound in the presence of hydrogen and a palladium-charcoal catalyst formed a diacetate insoluble in cold aqueous sodium hydroxide did not contain a carbonyl group and appeared to have one ethereal oxygen atom.On the basis of the 2 2-dimethyl-A3-chromen hypothesis and the fact that toxicarol has been shown to contain a phloroglucinol residue this substance was considered likely to be 5 7-dihydroxy-2 2-dimethyZchroman (11); the properties of the phenol appeared to exclude a structure of the type (111)and from the argu- ments presented in Part V (loc.cit.) the dihydrofuran structure (IV) was considered improb- able. That the compound has formula (11)is conclusively established by the fact that it was found to be identical with authentic 5 7-dihydroxy-2 2-dimethylchroman obtained in excellent yield by the application of Clemmensen’s method of reduction to 5 7-dihydroxy-2 2-dimethylchromanone(VI R = H).(1.1 (11.) (111.) Although the condensation of phloroglucinol and pp-dimethylacryl chloride with the aid of aluminium chloride gives a good yield of the chromanone (VI R = H) the reaction does not afford conclusive evidence regarding the structure of the product and since the standard chromanone synthesis (Robinson and co-workers J. 1926 945) is inapplicable 280 Bridge Heyes and Robertsoiz ExPeYiments on the the structure of the compound has been deduced indirectly. It behaves as a saturated substance towards hydrogen in the presence of a palladium catalyst and forms a 2 :4-ainitrophenylhydrazoneand a dimethyl ether (VI R = Me) which is insoluble in cold aqueous sodium hydroxide and does not give a ferric reaction properties sufficing to distinguish the chromanone from the isomeric compounds (V) (VII) and (VIII) but not from the p-coumayanone (XII,R =H) all of which could conceivably have been formed by the Friedel- Crafts reaction.Although under certain conditions the chromanone is accompanied by small amounts of a by-product no trace of an isomeric compound has been observed. The by-product which has riot yet been obtained analytically pure appears to have the formula *p5,as it gives a mono-2 :4-dinitrophen~yZhydrazoneagreeing with this composition and is in all probability of the type (IX). 0 0 0 (VII.) 0 (VIII.) (IX.) (X.1 On general grounds it is almost certain that the formation of the chromanone (VI R = Hjwould proceed by way of the unsaturated ketone (V) a mechanism which is strictly analogous to that obtaining in the synthesis of flavanones by the condensation of cinnamoyl chlorides with rn-dihydroxyphenols.Nevertheless cyclisation of the ketone (V) could obviously have taken the alternative route with the formation of the coumaranone (XII, R = H) and hence it was essential that clear proof of the chromanone structure of the pro- duct should be obtained. For this purpose evidence of the presence of the group COCH in the dimethyl ether of the chromanone would have been conclusive but our efforts to demon- strate this by standard methods were unsuccessful. Condensation of the dimethyl ether with veratraldehyde in alcoholic sodium hydroxide led to the formation of a compound n1.p.152.5-l53" which from appearance (yellow colour) and crimson-red coloration with concentrated sulphuric acid appeared to be a benzylidece derivative of the chromanone but which unexpectedly gave a dark red coloration with alcoholic ferric chloride. Analysis indicated the presence of a molecule of water more than that required for 5 :7-dimethoxy-2 :2-dimethyl-3-veratrylidenechromanoneand hence it would seem that this compound is the styryl derivative (X). This is supported by the fact that when a solution of the chro- manone dimethyl ether in 5% alcoholic potassium hydroxide is kept for 24 hours at room temperature the product obtained on acidification has the properties of an o-hydroxyphenyl ketone. On the other hand attempts to form a veratrylidene derivative by the standard hydrogen chloride-acetic acid method led to the isolatioii of an almost colourless amorphous product which could not be purified and may have contained a bis-derivative of the alde- hyde.Similarly condensation of the dimethyl ether with 2-hydroxy-4-methoxybenzalde-hyde in ethyl acetate by means of hydrogen chloride in the usual manner yielded a red chromenopyrylium-like product which has not yet been obtained pure. Efforts to prepare an isonitroso-derivative of the dimethyl ether by the usual method were equally unsuccessful ; a compound m. p. i24-125" was invariably obtained in small yield which however did not contain nitrogen and was accompanied by a considerable amount of impurities giving a ferric reaction. Since these attempts to obtain decisive evidence for the chromanone structure proposed for the Friedel-Crafts product (VI R = H) and its dimethyl ether (VI,R =Me) were nnpro- Sylzthesis of Rotenone md its Derivatives.Part XII. 281 ductive the synthesis of the isomeric p-counzaranone (XII R = H) was next investigated. In contrast to the formation of the chromanone (VI R =; H) the condensation of phloro-0 MeOuy$HeCHMe2 HO/\OH Br RO,q/\C €-I*CHMe j)CO-L~*CBMe,OH \/-OOR OMe COR (XI.) (XII.) (XIII.) glucinol with a-bromoisovaleryl chloride in nitrobenzene by the Friedel-Crafts method furnished along with traces oi a bromo-compound a small yield of a product which ap-peared to be (XII R = H). This compound behaves as a saturated substance towards hydrogen in the presence of a palladium or platinum catalyst but unlike the isomeride (VI R = S),does not appear to form a 2 4-dinitrophenylhydrazonereadily and on reduc- tion by Clemmensen’s method gives in addition to much amorphous material only traces of a phenolic compound.By the methyl iodide-potassium carbonate method an alkali- insoluble methylation product was obtained which could not be satisfactorily purified from more highly methylated by-products (probably trimethyl ether of enolic form of coumaran- one) but appeared to consist mainly of the dimethyl ether (XII R = Me) since it gave rise to a good yield of a 2 4-dipzit~opheizylhyd~azone, subsequently found to be identical with the 2 4-dinitrophenylhydrazone of authentic 4 6-dimethoxy-2-isopropyl-3-coumaranone (XII R = Me).The latter compound was prepared by an unambiguous method and hence the p-coumaranone structure of (XII R = H) is established. Further since comparison of the 2 4-dinitrophenylhydrazones of the p-coumaranone (XII R = Me) and of the chro- manone dimethyl ether (VI R = Me) showed that the two compounds were isomeric and not identical therefore the chromanone structure assigned to the latter ether and the parent dihydroxy-compound (VI R = H) is confirmed. The analogous formation of 7-hydroxy- 2 2-dimethylchromanone and of 6-hydroxy-2-isopropyl-3-coumaranone from resorcinol and the appropriate acid chlorides has also been established (compare Arima and Okamoto J. Chem. SOC.Japan 1929 58,344),thus affording collateral evidence for the structures herein assigned to the phloroglucinol derivatives.These results will be described in a subsequent communication. In view of the established formation of the chromanone (VI R = H) when pp-dimethyl- acryl chloride is employed in the Friedel-Crafts reaction the p-coumaranone (XII R = H) must arise from the intermediate a-bromo-ketone (XI) by direct loss of hydrogen bromide without the formation of an unsaturated ketone (V). This cyclisation may well occur during the process of steam-distillation necessarily employed in working up the reaction mixture. The authentic coumaranone (XII R = Me) was prepared according to a standard pro- cedure (Stephen and co-workers J. 1936 896) Interaction of the sodium derivative of phloroglucinol dimethyl ether and ethyl a-bromoisovalerate gave rise to a satisfactory yield of the ester (XIII R = OEt) readily hydrolysed to the acid (XIII R = OH).Cyclisation of the acid chloride (XIII R = C1) in benzene with aluminium chloride furnished a product from which the coumaranone (XII R = Me) was obtained as a viscous oil conveniently characterised by the formation of a 2 4-dinitrophenylhydrazone. Although the chromanone structure of 5 7-dihydroxy-2 2-dimethylchromanone (VI R = H) normally implies that the reduction product is the chroman (11) the ease of opening of the chromanone ring system (VI R = Me) under certain conditions (see also experimental section) led us to consider the possibility that on application of the Clemmensen reaction to (VI R = H) reduction might be preceded by ring scission to give (V) resulting in the formation of (111) which could then undergo cyclisation to give the dihydrofuran (IV).The evidence at our disposal however appears to exclude this mechanism ; e.g. by analogy with unsubstituted chromanones and flavanones a ring system of the type (VI) would not be expected to open readily in the presence of acidic reagents ; reduction of the chromanone appears to proceed normally giving an excellent yield of (11) [compare p-coumaranone (XII R = H)]. In any case even if the reduction of the chromanone (VI R = H) did take place by way of the intermediate (111) there is substantial evidence that compounds of the latter type cyclise to chroman and not coumaran systems e.g. formation of p-dihydro-rotenone and its relationship to deguelin and tephrosin (Part 11,J.1932 1384; conversion of coumarins into chromanols and chromens Heilbron and Hill J. 1927 2005). Attempts to obtain independent evidence on this question by oxidation of the chroman (IT.)have so far been unsuccessful. CH 0 The isolation of the chroman (11)together with the structures established for apo-and dehydroapo-toxicarol in Part V (Zoc. cit.) affords substantial proof that dihydrotoxicarolic acid is represented by formula (I)and that dehydrodihydrotoxicarol has formula (XIV) or (XV). Further since the ethylenic linkage of toxicarol is known to be present in the C, residue lost in the formation of apotoxicarol it is now clear that toxicarol contains a 2 2-dimethyl-A3-chromen residue and hence the alternative structures suggested for toxicarol dehydrotoxicarol and for dihydrotoxicarol in Part V are substantiated.Formation of Acetone fyom 2 :2-Dimethyl-A3-chromens.-The formation of acetone by hydrolytic fission appears to be a general property of certain natural 2 2-dimethyl-A3-chromens deguelin toxicarol xanthoxyletin and xanthyletin (Part V Zoc. cit. ; J. 1936 627 1828) and in these cases this type of scission takes place much more readily than in the case of analogous compounds substituted in the 4-position ; thus we have observed that the iormation of acetone could not be detected when 7-hydroxy-2 :2 4-trimethyl-A3-chromen was boiled with 50% aqueous sodium hydroxide for 3 hours. Under the experimental conditions employed for the hydrolysis of the natural chromens (Zoc.cit.) it is extremely unlikely that the formation of acetone arises by fission across the double bond i.e. between the 3-and the 4-C atom of the chromen system as a first step. This route would doubtless lead to the production of a-hydroxyisobutyric acid as an intermediate stage and although this compound is stated to give acetone on fusion with solid sodium hydroxide it is stable under conditions considerably more drastic than those necessary for the fission of the A3-chromen systems of the aforementioned natural compounds. In view of the occurrence of members of the rotenone series-rotenone deguelin te- phrosin and toxicarol-in the same or closely related plant species and of the fact that these compounds contain the C unit E>C-C-C embodied either in a 2 2-dimethyl-A3-chromen (type XVII) or in an isomeric coumaran system (type XVIII) it is reasonable to assume that both structures originate from a common type of intermediate e.g.(XVI). On the further assumption that the attachment of the C residue is one of the end processes in the phytocheinical production of these compounds we may express the formation of the intermediate type (XVI) thus CH2 0 0 Me,CO Me*CHO+-+ Me,C( OH) *CH,*CHO [or CH,:CMe*CH,-CHO] /\//-+ CH -+ CH CH (XVI.) (XVII.) (XVIII.) The substance (XVI) could then undergo cyclisation in two ways forming either the chro- men (XVII) or the dihydrocoumarone (XVIII). Reversal of this scheme offers a feasible mechanism for the production of acetone by hydrolytic fission of 2 :2-dimethyl-A3-chromens unsubstituted in the 3- and 4-positions analogous to that suggested by Heilbron and Hill (loc.cit.) for the decomposition of 2 2-diphenyl-4-met hyl-A3-chromens the final stage of Synthesis of Rotenow and its Deuivatives. Part XII. 283 which represents the reversal of an aldol condensation. This mechanism as indicated by these authors finds a close analogy in the fission of certain coumarins discovered by Baker (J.,1925 12'7 2349); e.g. 7-hydroxy-3-phenyl-4-methylcoumarin gives mainly resorcinol and benzyl methyl ketone which must arise by way of a-phenylacetoacetic acid. EXPERIMENTAL. Oxidation of DehydrotoxicarolA cetate with Potassium Permangunate.-Treatment of dehydro-toxicarol (2.7 g.) with acetic anhydride (10 c.c.) and pyridine (5 c.c.) on the steam-bath for 2 hours gave the acetate which formed pale straw-coloured needles or rods m.p. 231-232" identical with a specimen obtained by Clark's method (1. Amer. Chem. SOG.,1931 53 2264). A solution of potassium pcrmanganate (1.87 g.) in water (50 c.c.) was added to a well-stirred mixture of the acetate (1.7 8.) in acetone (750 c.c.) in the course of 1Q hours. Next day the manganese dioxide was recovered by filtration the greater part of the acetone evaporated in a vacuum the residual liquor treated with excess of aqueous sodium carbonate and the insolublc product (A) collected. Acidification of the aqueous filtrate precipitated 2-hydroxy-4 5-dimeth-oxybenzoic acid which had m.p 213-214' after repeated crystallisation from aqueous acetone and was identical with a synthetical specimen (J. 1931 2432). The precipitate of manganese dioxide was extracted with boiling water (20 C.C. x 5) and the combined extracts acidified with hydrochloric acid (Congo-red) and concentrated in a vacuum. After the removal of a small quantity of 2-hydroxy-4 :5-dimethoxybenzoic acid the residual liquor on extraction with ether gave a small amount of rissic acid which formed colourless prisms m. p. 256" from methyl alcohol and was identified by comparison with an authentic specimen (J. 1932 1380). The aforementioned solid (A) was found to consist of a mixture of unchanged dehydrotoxi- carol acetate and a neutral compound (B),from which the latter was extracted by means of warm methyl alcohol.Evaporation of the extracts left (R) as colourless plates contaminated with a small amount of dehydrotoxicarol acetate which was removed by means of warm benzene (30-35'). Repeated crystallisation of the residue from dilute acetone and then dilute methyl alcohol gave (B)in colourless diamond-shaped plates ni. p. 149" after sintering at 143" (Found in air-dried material :C 59.4 59.8 59-9;H 5.5 5.4 5.5%. Found in a specimen dried in a vacuum at 110' :C 60.6 60.7 60.8; H 5.3 5.2 5.7; OMe 14.4; CH,*CO 10.1%). This compound which had a negative ferric reaction was insoluble in aqueous sodium carbonate but slowly dissolved in 8% aqueous sodium hydroxide. Acidification of the alkaline solution gave an amorphous solid having a green ferric reaction.Hydrolytic Fission of Dihydrotoxicarolic Acid.-A mixture of the acid (Clark J. Amer. Chem. SOC.,1932 54 2546) (5 g.) and 50% aqueous potassium hydroxide (20 c.c.) was heated (oil-bath) in nitrogen to 200-205" in the course of 14 minutes and then maintained at this temperature for I0 minutes. A solution of the cooled reaction mixture in water (150 c.c.)was acidified with hydro- chloric acid (Congo-red) warmed to about 70" mixed with charcoal and filtered (wash charcoal with 70 C.C. of water). After the addition of excess of sodium bicarbonate the filtrate was thoroughly extracted with ether (40 C.C. x lo) and the combined extracts washed with brine dried and evaporated leaving 5 7-dihydroxy-2 2-dimethylchroman as a light brown oil which gradually solidified.Crystallisation from benzene and then several times from water (charcoal) gave the compound in colourless hexagonal prisms (1 g.),m. p. 162-163" readily soluble in ethyl or methyl alcohol acetone and ethyl acetate (Found C 68.1 ; H 7.4. CllH1403 requires C 68-0;H 7.2%). In alcohol the ferric reaction is negative and in aqueous solution faint blue Acetylated with acetic anhydride (1-5 c.c.)and pyridine (0-5c.c.) on the steam-bath for 12 hour the chroman (0-5 g.) gave rise to the diacetate which separated from a little methyl alcohol in colourless prisms m. p. 86" (Found C 64.7 ; H 6-7. C15H&5 requires C 64.7 ; H 6.5%). This derivative is soluble in the usual organic solvents and insoluble in cold aqueous sodium hydroxide. PP-Dimethylacrylic A cid.-Attempts to prepare this compound according to Dutt's method (J.Indian Chem.SOG.,1924 1 297) were unsuccessful. A mixture of a-bromoisovaleric acid (" Organic Syntheses,'' XI p. 20) (40 g.) and quinoline (80 g.) was heated to 140-150" and when the vigorous reaction had ceased then maintained at 165-170" for 20 minutes. After the addition of excess of dilute hydrochloric acid to the cooled reaction mixture PP-dimethylacrylic acid was isolated with ether and purified by dis-tillation in a vacuum; yield 11.5 g. Crystallised from light petroleum (b. p. 40-60") it had m. p. 67-68'. When applied to a-bromoisovaleric acid the method used by Perkin (J. 1896,69 1470) for the preparation of acrylic acid from ethyl or-bromoisovalerate gave poor yields of an inferior product.Bridge Heyes and Robertson Exfieriments on the PP-Dimethylacry1 chloride (11g.) b. p. 45-48"/12 mm. was prepared by gently warming the acid (12 g.) with thionyl chloride and purified by distillation in a vacuum. 5 7-Dzhyd~oxy-2 2-dimethylchromanone (VI R = H).-Aluminium chloride (15.2 g.) was dissolved in a mixture of nitrobenzene (200 c.c.) and phloroglucinol (16-2 8.) (agitate). Pp-Di-methylacryl chloride (12.5 g.) was then gradually introduced and the reaction mixture kept at room temperature for 4 days. After the addition of ice (150 g.) and dilute hydrochloric acid (100 c.c.) the product mixed with nitrobenzene was isolated with ether the nitrobenzene was removed by means of steam and the hot aqueous liquor decanted from brown viscous resin.On being kept the cooled aqueous solution slowly deposited the chromanone ; a further quantity was obtained by extraction of the ,resinous residue with chloroform-light petroleum. Yield of crude material 14.5 g. Recrystallised from chloroform or dilute alcohol the compound formed colourless elongated prisms m. p. 198" readily soluble. in alcohol acetone or ethyl acetate and sparingly soluble in benzene (Found C 63.4; H 5.7. CllHl2O4 requires 63-5; H 5.8%). With alcoholic ferric chloride it gave a purple coloration. Excess of 2 4-dinitrophenylhydrazine hydrochloride in dilute hydrochloric acid was added to a solution of the chromanone (0.7 g.) in alcohol (20 c.c.) and the mixture warmed for 10 minutes on the water-bath and then kept for 16 hours at room temperature.The resulting 2 4-di.l.titro~henylhyd~azone formed blood-red needles m. p. 277-278". from alcohol (Found N 14.3. c~,H,,o,N requires N 14.4%). In the preparation of the chromanone it was found that the proportion of aluminium chloride employed had considerable effect upon the tractability of the product obtained and the foregoing procedure appeared to give an optimum result. In another experiment where phloroglucinol (1-25 mols.) acid chloride (1mol.) and aluminium chloride (2.5 mols.) were allowed to interact in nitrobenzene for 7 days the aqueous liquor left on removal of nitrobenzene gave only a small amount of the chromanone and the resinous product which did not yield crystalline material with solvents was dissolved in ether and the solution extracted with aqueous sodium carbonate.In the course of the extraction a small amount of a solid sodium derivative crystallised. Decom-position of this material in a little methyl alcohol with acetic acid gave a compound (IX) which separated from light petroleum (b. p. 80-100°) in colourless elongated prisms m. p. 134" (the melt becoming clear at 138") giving a red coloration with alcoholic ferric chloride (Found in material dried at 90" C 66.9 67.0 ; H 6.3 6-3; M 245 253. Calc. for CI6Hl8O5 C 66.2 ; H 6.2% ; M 290). The substance is soluble in alcohol acetone chloroform benzene or ethyl acetate and on treatment with 2 4-dinitrophenylhydrazinehydrochloride (method employed in the case of chromanone) gave rise to a 2 4-dinitro~henylhydrazone,which formed short red prisms m.p. 267-268' (Found C 56.3; H 4.9; N li.5. C,,H,,O,N requires C 56.2; H 4.7; N 11.9%). Acidification of the combined aqueous sodium carbonate extracts gave a further small amount of chromanone m. p. 198" after purification. 5 7-Dimethoxy-2 :2-dimethylchromanone (VI R = Me) .-Methylation of 5 7-dihydroxy-2 2-dimethylchromanone (2 g.) was effected with methyl iodide (4 c.c.) and potassium carbonate (7 g.) in boiling acetone (60 c.c.) in the course of about 7 hours ; after 4 hours more iodide (2 c.c.) and more carbonate (3 g.) were added. When a test portion of the reaction mixture did not give a ferric reaction the potassium salts were removed by filtration (wash with excess acetone) the solvent evaporated and an ethereal solution of the residue washed with dilute aqueous sodium hydroxide dried and evaporated.Crystallisation of the residual product from light petroleum (b. p. 60-80') gave the dimethyl ether in colourless plates (1.7 g.) m. p. 104.5-105" soluble in bmzene or alcohol and having a negative ferric reaction [Found C 66.1 ; H 6.9; OMe 26.7. CllHl,O,(OMe) requires C 66.1 ;H 6.8; OMe 26-3%]. When a solution of the chromanone (0.4 g.) in alcohol (13 c.c.) was boiled with 2 4-dinitro-phenylhydrazine hydrochloride (from 0.4 g. of the hydrazine) for 2 minutes the 2 4-dinitro-phenylhydrazone separated almost immediately ; it formed blood-red needles (0.5 g.) m. p. 240" from ethyl acetate (Found :N 13.5. C,,H,,0,N4 requires N 13.5%). The product obtained by treating the chromanone with semicarbazide acetate in the usual manner separated from alcohol in colourless elongated prisms m.p. 245" (Found :N 13.7. Calc. for C14H,,04N3 N 14.3%. Calc. for Cl4H1,O4N3,H,O N 14.0%). As this compound possesses the unexpected property of giving a reddish-brown coloration with alcoholic ferric chloride it is probably the semicarbazone of a hydrated derivative of the chromanone i.e. of p 2-dihydroxy-4 6-dimethoxyisovalerophenone(the semicarbazoiies of o-hydroxypropiophenone and of 3-chloro-2-hydroxyacetophenonegive dark brownish-violet ferric reactions). On treatment with hot alcoholic 2 :4-dinitrophenylhydrazinehydrochloride however this derivative was quan- Syfithesis of Rotenone and its Derivatives.Part XII. titatively converted into the 2 :4-dinitrophenylhydrazone,m. p. 236-237" which on being mixed with an authentic specimen m. p. 240° melted at 238-239". 5 :7-Dihydroxy-2 2-dimethyZckronzan (11).-A solution of 5 7-dihydroxy-2 :2-dimethyl-chromanone (2 g.) in a mixture of alcohol (20 c.c.) acetic acid (10 c.c.) water (46 c.c.) and con- centrated hydrochloric acid (34 c.c.) containing amalgamated zinc dust (Robinson and Shah J . 1934 1497) (80 g.) was kept at room temperature for 2 days treated with 12% hydrochloric acid (20 c.c.) heated on the steam-bath for 1hour and then refluxed for 6 hours more acid (20 c.c.) was added. The hot liquor was decanted cooled and extracted several times with ether and the combined extracts were well washed with aqueous sodium bicarbonate dried and evaporated leaving the chroman which separated from benzene and then water in colourless prisms (1.2 g.) m.p. 162-163" identical in every way with the natural product (Found c 67.9; H 7.3%). The diacetate formed colourless prisms m. p. and mixed m. p. 86" from a small volume of methyl alcohol. 4 6-Dihydroxy-2-isopro~y~-3-co.unzanone (XII R = H).-a-Bromoisovaleryl chloride (18 g.) was slowly added to a cooled solution of phloroglucinol (14 g.) and aluminium chloride (15 g.) in nitrobenzene (200 c.c.) and the mixture kept for 4 days; more aluminium chloride (6 g.) was added after 2 days. The reaction mixture was very slowly heated to 60" maintainetl at this temperature for 1hour cooled and treated with excess of ice and hydrochloric acid.TIE mixture was extracted several times with ether the combined extracts evaporated and the nitro- benzene removed from the residue by a current of steam. On being kept the residual aqueous liquor slowly deposited the coumurunone (4.5g.) which formed colourless rhombic prisms m. p. 196" from dilute methyl alcohol readily soluble in alcohol or ethyl acetate and sparingly soluble in benzene (Found C 63.3; H 5.8. C,,H,,O requires C 63.5; H 5.8%). Mixed with 5 7-dihydroxy-2 2-dimethylchromanone this compound melted at 170-184". With alcoholic ferric chloride it gave a purple coloration. After the separation of the crude coumaranone the aqueous liquor deposited traces of a pro-duct containing bromine m. p. 110-115° after being once recrystallised from water.4 6-Dimethoxy-2-iso~ropyZ(~)cou.unaranone (XII R = Me).-(A) Ethyl a-bromoisovalerate (20 g.) was gradually added to a solution of phloroglucinol dimethyl ether (20 g.) in alcohol (80 c.c.) containing sodium ethoxide (from 3 g. of sodium) and next day the mixture was gently refluxed for 6 hours. After the removal of the greater part of the alcohol by distillation in a vacuum the residue was diluted with water the solution extracted several times with ether and the combined ethereal extracts washed with excess of 1% aqueous sodium hydroxide to remove unchanged phloroglucinol dimethyl ether and evaporated. The residual mixed esters were hydrolysed with excess of 12% alcoholic sodium hydroxide at room temperature for 16 hours and on isolation in the usual manner the resulting 3 5-dimethoxy-a-phenoxyisovalericacid (XIII R = OH) (7g.) (fraction b.p. 160-170°/0~1 mm.) was separated from pg-dimethyl- acrylic acid (6 g.) by distillation in a high vacuum and then crystallised from aqueous alcohol forming colourless prisms m. p. 92". Prepared by means of phosphorus pentachloride (4.4 g.) the acid chloride (from 5 g. of acid) was dissolved in purified benzene (50 c.c.) and then treated with aluminium chloride (2.9 g. added in the course of 4 hour). 3 Hours later excess of ice and dilute hydrochloric acid was introduced and the product and benzene isolated with ether. After having been washed with aqueous sodium bicarbonate to remove unchanged acid the ether-benzene solution was evap- orated and the residue distilled in a high vacuum giving a main fraction (2 g.).Redistillation of this material gave a clear very viscous liquid b. p. 140-150°/0.1 mm. which had the charac- teristic coumaranone odour and gave rise to a 2 4-dinitrophenyZhydrazone,iorming deep red needles m. p. 185",from alcohol (Found N 13.2. C,,H,,07N0 requires N 13.5%). (B) 4 6-Dihydroxy-2-isopropylcoumaranone(1 g.) was methylated with methyl iodide (6 c.c. added in two portions) and excess of potassium carbonate in boiling acetone for 6 hours ; ;I test sample did not give a ferric reaction. On isolation and removal of phenolic material by nieans of dilute sodium hydroxide the product was obtained as a viscous oil which partly solidified and on treatment with hot alcoholic 2 4-dinitrophenylhydrazinehydrochloride gave rise to the coumaranone-2 4-dinitrophenylhydrazone,m. p. 185" identical in every way with a specimen obtained by method (A) (Found N 13.4y0). The authors are indebted to the Government Grants Committee of the Royal Society for a grant. UNIVBRSITY LIVERPOOL. [Received December 18th 1936.1 OF