5-CHLORO-3-KETO-I : 1-DIMETHYL-A4-TETRAHY DROUENZENE. 63 VI.-Action of RedtLciiig Ayeiits o"r, 5-CIiloi.o-3-keto- I ; I -rl i?, LCJ t h y l - ~ " - t c t 7 TI Ii y d 1.0 be I 1 y e 1 z e. By ARTHUR WILLIAM CrtossLm and NORA RENOUF, Salters' Research Fellow. SOME little time since (Trans., 1905, 8'7, 14S7), the authors described the action of sodium in moist ethereal solution on 5-chloro-3-keto-1 : 1- dimethyl-A4-tetrahydrobenzene (I) and showed that the main product , I. 11. of the reaction was 3-hydroxy-1 : 1-dimethylhexahydrobenzene (11), which may be described as the limit reduction product of the chloro- ketone, The main object of the present investigation was to find reducing agents, less powerful than sodium in moist ethereal solution, which would be discriminating in their action; so that it might be possible to prepare from cliloroketodimethyltetrahydrobenzene, first, a ketodimethylletrahydrobenzene, differing from the former only in that chlorine would be replaced by hydrogen, and secondly, the corre- sponding ketodimo t hylhexahydrobenzene.Chloroketodimethyltetrahydrobenzene is readily prepared from dimethyldihydroresorcin (see page 69) by the action of phosphorus trichloride, the yield being $5 per cent. of the theoretical amount ; and as other substituted dihydroresorcins give equally good yields of the corresponding chloroketones, i t is evident that if the above- mentioned reactions could be realised, they would furnish very ready methods for preparing substituted ketotetrahydro- and ketohexahydro- benzenes, which substances, especially the former, are not easy t o obtain by the present known methods. Complete success has attended the experiments, and further work is in progress with the object of proving that the reactions are general ones.Purticular attention is being paid to the ketones (111 and IV) derivable from trimethyldihydroresorcio, on account of the great similarity in the Y 111. groupings which they contain when compared with those of camphor. The action of sodium in moist ethereal solution on chloroketo- dirnethyltetrahydrobenzene has been further investigated, and it has been proved that the addition of a small quantity of alcohol to the ether has a beneficial effect, considerably increasing the yield of hydroxydimethylhexahydrobenzene and rendering it much easier64 CROSSLEY AND RENOUF: ACTION OF REDUCING AGENTS ON to remgve the chlorine completely from the chloroketone.Further, several interesting substances have been isolated from the resinous by-product (Trans., 1905, 87, 1494) which are dicyclic in composition. Similar resins have been encountered in all the reducing actions which have been*tried, and as the constitutions of the substances derived from them have close relationship with one mother, it will be most convenient to describe first the single nucleus compounds produced in these reactions and afterwmds the dicyclic derivatives. Single Nucleus Conzpou?cds. Having found that alcohol had such a decided influence in the reaction between chloroketodimethgltetrahydrobenzene and sodium in moist ethereal solution, it was thought advisable to try the action of sodium in absolute alcoholic solution. The reaction proceeded, however, in an unexpected direction, and though giving an interesting body from quite anot'her point of view, did not further t'he object of the present inquiry.It demonstrated the fact that the chlorine atom in chloroketodimethyltetrahydrobenzene is very reactive, a fact which greatly enhances the possibilities of t h e use of this and similar chloroketones for synthetical purposes. The reaction gives rise to small quantities of hydroxydimethyl- hexahydrobenzene (11), but principally 3-l~ydroxy-5-ethoxpl : 1 - climethyl~~exctZiyclrobenxene (VI). Evidently the sodium ethoxide formed A , iNaiOEt v.------- VI. in the first stages of the reduction reacts with the chlorine atom of the chloroketone t o give the substmce rapresented by formula V, which is then further reduced to the corresponding saturated compound.The constitution of the latter is proved by analysis and by the facts that a Zeisel determination shows it to contain an ethoxy-group, and that, when treated with acetyl- or benzoyl-chlorides, it yields acetyl- or benzoyl-derivatives respectively. The next reducing agent employed was zinc dust in aqueous alcoholic solution, which, as previously shown (rrans., 1905, 87, 1497; 1906, 89, 43), readily replaces halogen by hydrogen in siturated hydroaromatic substances, but in the present instance its action is too powerful, as it gives a mixtureA of the ketones represented by formulze V I I and VIII, containing approximately 30 par cent.of the latter. However, 3-keto-1 : 1 -dirneth~2-~~-letrahydro- beizxene (VII) may be obtained quite pure by replacing zinc dust by $11, VfII.5-CHLORO-3-KETO-P : 1-DIMETHYL-A4-TETRAHYDROBENZENE. 65 zinc filings, either in the cold or on heating, or by using the zinc copper couple. It is a colourless liquid boiling a t 88.5' at 32 mm., and its ketonic nature is proved by. the fact that it gives a semi- carbnzone and an oxime. When oxidised with potassium per- inanganate in the cold, it yields as-diniethylsuccinic acid and the lactone of a-hydroxy-PP-dimethylglutaric acid : These products mere also obtained by the oxidation of chloroketo- dimethyltetrahydrobenzene (Trans., 1903, 83, 11 9) together with /3P-dimethylglutaric acid.The production of the lactone of a-hydroxy- PP-dimethylglutaric acid proves quite definitely the constitution of the ketone, although its formation cannot be expIained on the lines suggested in the case of the chloroketone. From the present instance, i t would appear that the production of this lactone must be regarded as almost certain in any oxidation where the degradation of a ring is concerned and which might reasonably be expected to yield P/3- d imet h y lgl u t a r ic acid. 3-Keto-1 : 1-dirnethyZ~~exa7Lydrobeizxene (VIII) may be produced from chloroketodimethyltetrahydrobenzene by heating it with zinc dust in glacial or dilute acetic acid solution. It is a colourless liquid boiling about 10' lower (75.5' a t 25 mm.) than the corresponding unsaturated ketone.It forms an oxime and a semicarbazone, and when oxidised with potassium permanganate gives only PP-dimethyladipic acid (IX), a fact which proves its constitution beyond doubt. One slight difference is observable in the specimens of the ketone, according as to whether glacial or dilute acetic acid is employed in their preparation. In the latter case, the ketone gives a faint colour reaction with concentrated sulphuric acid, indicating the presence of traces of ketodimethyltetrahydrobenzsne, which is not, however, produced in sufficient amount to influence the analysis of the substance, and which can be completely and easily removed by treatment with a small quantity of dilute potassium permanganate solution in the cold. The choice between dilute and glacial acetic acid would be influenced according a d to whether the solid by-product (m.p. 148', see page 82) was required for investigation, as a larger proportion of the latter substance is obtained when dilute acetic acid is employed, although the yield of ketodimethylhexahydrobenzene is not so good as when glacial acetic acid is used. VOL. XCI. F66 CROSSLEY AND RENOUF: ACTION OF REDUCING AGENTS ON Other reducing agents investigated were zinc dust in strongly alkaline solution, also zinc dust and hydrogen chloride i n alcoholic solution. I n the former case, the hydrolytic action of the potassium hydroxide overshadows the reducing action of the zinc, with the result that the only product isolated was dimethyldiliydrorosorcin (X). I n the latter case, a mixture of ketodimethyltetrnhydrobenzene and keto- dilnethylhexahydrobenzene was obtained, together with dimethyldi- hydroresorcin a.nd it,s ethyl ether (XI) (Trans., 1899, 75, 775).c H2--CO>CH cMe2<CH 2' C("H) X. XI. Here again hydrolysis must first take place, giving rise to dimethyldihydroresorcin, which is then esterified by the alcoholic hydrogen chloride. Dicyclic Compounds. The compounds of this nature met with are all derivatives of a substance formed theoretically by the removal of one hydrogen atom from each of two hexahydrobenzene rings, with consequent production of dic!clic derivatives. It has been decided to refer to this substance as dicyclohexane, and to indicate the positions of the various substitut- ing groups by adopting the following scheme of numbering : ,UH,*CH,, /CH2* ; H2, CH2Q t i)CH*CHFg: 3t 4?)CH2. \C: H2 CH 2/ CH2*CH2/ When zinc dust acts on chloroketodimethyltetrahydrobenzene in aqueous solution, there is obtaiued a mixture of ketodimethyltetra- hydrobenzene and ketodimethylhexahydrobenzene (VII and VIII), in which the former largely predominates.If this mixture of ketones is again treated several times with zinc dust, it is possible tto isolate pure ketodimethylhexahydrobenzene and a semi-solid mass from which a crystalline product has been separated, melting at 14S0, and having the composition C,,H,,02, and which is believed to be 1 : 1'-dihydroxy- 5 : 5 : 5' : 5'-tetramethyZ-A2:2'-dicyclohexene (XII). Another possibility XIT. for a substance of this composition would be that represented by formula XIII, which could be produced from chloroketodimethyltetra- hydrobenzene * by the zinc coupling two molecules together and then partially reducing the unsaturated diketone so formed.* To make the formnlce of these dicyclic compounds strictly comparable with those of the siiigle nucleus derivatives, chloroketodimethyltetrahydrobenzeiie should be writ ten CH,/ CRIe,'CH,. 'CCI, instead of as in forinula I. \CO--.CHY5-CHLORO-3-KETO-1: 1-DIMETHYL-A*-TETRAHYDROBENZENE. 67 XII1. Such a presumption is, however, a t once rendered impossible when it is remembered that the substance is not formed directly from the chloroketone but from the halogen-free mixture of ketones, obtained by the initial action of zinc dust on chloroketodimethyltetrahydro- benzene. There can be therefore no doubt in this case that the substance C16H2602 owes its origin to pinacone formation, an exactly analogous case having been observed i n the reduction of ketohexa- hydrobenzene (Trans., 1904, 85, 141 5), where considerable amounts of 1 : 1-dihydroxydicpclohexane (XIV), (ketohexamethylene pinacone) were formed.XIV. Moreover, the behaviour of dihydroxytetramethyldicyclohexene is quite in accord with that of a substance having formula XII, for i t does not give a colour reaction with concentrated sulphuric acid, nor can it be acetylated or benzoylated under the conditions employed. Further, its unsaturated nature is proved by the fact that it readily absorbs bromine with elimination of hydrogen bromide and formation of what is believed to be a tribromoanhydride, but the exact con- stitution of this compound is a t present doubtful.Also, when treated with sodium in moist ethereal solution, it absorbs four atoms of hydrogen to give 1 : l'-dihyclroxy-5 : 5 : 5' : 5'-~etrc~methyZclicyclofiexane (XV). Here again this saturated pinacone does not give a colour reaction with concentrated sulphuric acid, but it is surprising to find that it is acetylated or benzoglated very readily. No explanation of this fact can be offered on the present occasion, but i t is hoped that, as the work proceeds, other similarly constituted pinacones will be encountered, when the matter will receive further attention. It may a t first sight appear strange that zinc dust in aqueous alcohol is a reagent sufficiently strong to reduce the double bond in ket odimethyltetrahydrobenzene, giving ketodimethylhexahydrobenzene, and also strong enough to form dihydroxytetramethyldicyclohexene from ketodimethyltetrahydrobenzene, yet is not strong enough t o F 268 CROXSLEY AND RENOUF: ACTION OF REDUCING AGENTS ON reduce the dicyclohexene compound to the corresponding saturated derivative.But such is undoubtedly the case, for on boiling an aqueous alcoholic solution of dihydroxytetramethyldicgclohexene with zinc dust for twelve hours, no trace of any rednctioa having taken place could be detected and only the unsaturated dicpclohexene compo und was recovered. Moreover, the reduction of ketodimethyltetrslzydrobenzene t.0 keto- dimethylhexahydrobenzene by zinc dust in aqueous alcohol proceeds extremely slowly, and from the experiments described on page 75 it can only be concluded that, under the influence of this particular reducing agent, t.he unsaturated pinacone is more easily formed than ketodimethylhesahydro benzene.I n the preparation of hydroxydimethylhexahydrobenzene consider- able difficulty was experienced a t one t,ime (see page 70) in obtaining the product free from halogen, and on examining the resinous by- product formed under these conditions, two solid substances were separated from it melting respectively a t 178' and 173-174". The former of these has the composition C,,H,202 argd is apparently 3 : 3'-cZiketo-5 : 5 : 5' : 5'-tetmmethgl- A-I 'l'-dicyclohexene (XVI), formed XVI. by the direct coupling of 2 molecules of chloroketodimethyltetra- hydrobenzene by the sodium.It is highly coloured (yellow), gives a brick-red disemicarbazone, thus proving its diketonic nature, and is unsaturated as shown by its ready absorption of bromine. It could not be detected in the resin formed when alcohol was added t o the ether used in the reduction of chloroketodimethyltetrahydro- benzene, but much larger quantities of the second substance melting at 173-174" and also another compound melting sharply a t 2 1 2 O were isolated. The latter proved to be identical with 1 : 1'-dihydroxy- 5 : 5 : 5' : 5'-tetramethyldicycZohexane (see page 67). For a long time the substance melting a t 173-1 74" was thought to be homogeneous, as it gave on analysis numbers a.greeing with the formula C16H3002, nor mas its melting point altered by many recrystallisations, and moreover i t sublimed in needles which melted a t 171-172".Nevertheless i t was found t o be a mixture, for on acetylation it gave two diacetyl derivatives melting at 130" and 68", and on benzoylation, two dibenzoyl derivatives melting at 199" and 134". The former of the diacetyl and dibenzoyl compounds (m. p. 130' and 199") proved to be diacetyl- and dibenzoyl-1 : l'-dihydroxy-5 : 5 : 5' : 5'-tetramethyldi- cydohexane, which substance they yielded on hydrolysis,5- CHLORO-3-KETO- 1 : 1 -DIMETHY L-A4-TETRAHY DROBENZENE. 69 The above-mentioned derivatives melting a t 68’ and 134’ were separately hydrolysed with alcoholic potassium hydroxide, when they each gave a substance, C,,H,,O,, melting a t 183O, which is believed to be 3 ; 3’-dihydrozy-5 : 5 : 5’ ; 5’-tet~anwthyldkyclo?~xane (XVII).It XVII. is readily ncetylated and benzoylated, and, unlike the unsaturated 01’ saturated pinacones, gives a decided colour roaction with sulphuric acid. Its formation would be due to the further reduction of diketo- tetrsmethyldicyclohexene (XVI), although, unfortunately, sixflicient of the latter material could not be isolated to try the action of reducing agents upon it. The substance melting at 173-174e is therefore a mixture of 1 : 1’-dihydroxy-5 : 5 : 5’ : 5’-tetramethyldicyclohexane (XV) and 3 : 3’-di- hydroxy-5 : 5 : 5’ : 5’-tetramethyldicyclohexane (XVII), in the approxi- mate proportion of one part of the former to three parts of the latter, and it has been ascertained that a mixture of these substances in these proportions melts at 173-1’74’ and sublimes in needles meltinm 9 at 171-172’.It would appear, therefore, that in the reduction of chloroketo- dimethyltetrahydrobenzene with sodium in moist ethereal solution, dicyclic compounds are formed both by the process of pinncone forma- tion and by the coupling reaction of the sodium. Further, when the reduction is inefficient, there are formed 3 : 3’-diketo-5 : 5 : 5’ : 5’-tetra- rnethyl-Al’l’-dicycZohexene (XVI) and small amounts of 1 : 1’-di- hydroxy- and 3 : 3’-dihydroxy-5 : 5 : 5’ : 5’-tetrnmethyldicycZohexane ; but when the reduction is facilitated by addition of alcohol to the ether, the first of these three substances cannot be isolated, as it is further reduced to 3 : 3’-dihydroxy-5 : 5 : 5’ : 5’-tetramethyldicycZo- hexane.EX PERIMENTAL. Chloroketodimethyltetrahydrobenzene was prepared as already described (Trans., 1903, 83, 117), except that the heating with phosphorus trichloride was continued for three hours instead of two and a half hours. Moreover, after working up the product, the alkaline washings should be acidified with sulphuric acid, when a considerable amount of solid matter separates, consisting largely of unaltered dimethyldihydroresorcin, and if this is again treated with phosphorus trichloride, the yield of chloroketodirnethyltetrahydro- benzene may be increased t o 75 per cent. (instead of 66 per cent.) of the theoretical amountl. The boiling point of the ketone, after repebted distillation, is some-70 CROSSLEY AND RENOUF: ACTION OF REDUCING AGENTS ON what lower than previously stated (Zoc.cit.), namely 9 9 O at 20 mm,, and its great stability is a point worth calling attention to. On exposure to air and light, i t darkens in colour and deposits crystals of dimethyldihydroresorcin hydrochloride, but if protected from air and light i t may be kept for an almost indefinite period. A specimen which had been bottled in this way for three years was found t o be faintly yellow, and contained 1 gram of crystalline dirnethyldihydro- resorcin hydrochloride. After filtration, there remained 37 grams of a liquid, which were dissolved in ether, the ethereal solution washed with potassium hydroxide solution, then with water, dried over calcium chloride, and the ether evaporated. On distilling the residue in a vacuum, 36 grams passed over quite constantly with the above- mentioned boiling point of chloroketodimeth yltetrahydrobenzene. Reduction of CId oroketodimet~i~ltetrc6?~~dro be?izene, I.iti'th Sodium in moist Ethereal Solution. The main product obtained by this reaction is 3-hydroxy-1 : 1- dimethylhexahydrobenzene, which has already been described in det3ail (Trans., 1905, 87, 1494). Considerable difficulty mas ex- perienced at one time in obtaining this substance free from chlorine, until it mas redised that in the original preparation the ether employed had not been washed with water to remove alcohol. Further experiments have shown that, if the amounts of materials as already stated be used, with the addition of 20 C.C. of absolute alcohol to the ether, the product of the reaction is always free from halogen. Aforeover, the yield of hydroxydimethylhexahydrobenzene is thereby increased to 60 per cent.of the theoretical amount. During the preparation resinous matter is always formed, varying in amount from 25 to 30 grams from 100 grams of chloroketodimethyl- tetrahydrobenzene. The material obtained when difficulty was ex- perienced in producing hydroxydimethylhexahydrobenzene free from halogen was a thick, red jelly, which was dissolved in the smallest possible amount of light petroleum (b. p. SO-l0O0) and allowed to stand. A viscid, yellow solid gradually separated (2.3 grams from 75 grams of the jelly), which was repeatedly crystallised from methyl alcohol : 0.1078 gave 0.3076 CO, and 0.0871 H,O.3 : 3'-Diketo-5 : 5 : 5' : 5'-tetranaethyl-A1 '"-dicyclohexene, C = 77-82 ; H = 8.97. C,,H,,O, requires C = 78.04 ; H = 8.94 per cent. is readily soluble in the cold in chloroform or benzene, and crystallises5-CHLORO-3-KETO-1 : ~-DIMETHYL-A*-TETRAHYDROBENZENE. 71 from light petroleum (b. p. S0-1OO0), ethyl acetate, or methyl alcohol in deep yellow needles melting at 178'. When a solution of bromine in chloroform was added to a solution of the ketone in the same solvent and the whole slightly warmed, the bromine disappeared and hydrogen bromide was evolved. On evaporating the solvent, a white solid remained, which crystallised from absolute alcohol and melted a t 165" with evolution of gas, but the amount of material was too small t o permit of the reaction being thoroughly investigated.The disernicarbaxoTze was prepared in the usual manner and separated from the alcoholic solution as a micro-crystalline, brick-red powder melting about 273' with decomposition. On account of its insolubility in the ordinary organic solvents, it was not found possible to purify it by crystallisation, and it mas therefore analysed after being well washed with absolute alcohol : Found, N = 23.38. C,,H,,O,N, requires N = 23.33 per cent. The light petroleum solution of the original red jelly gave a second small crop of solid matter, which was white, and consisted of 1 : l'-dihydroxy-5 : 5 : 5' : 5'-tetramethyZdicyclohexane, melting a t 212", (see page 77) and of 3 : 3'-dihydroxy-5 : 5 : 5' : 5'-tetramethyldicycZo- hexane melting at 183" (see page 72).When the reduction of chloroketodimethyltetrahydrobenzene by sodium in moist ethereal solution is carried out as described on page 70, the by-product sets to a hard, transparent, reddish-yellow resin. This was dissolved in benzene, light petroleum (b. p. 40-60') added, and the whole allowed to stand, when solid matter separated, which was filtered and well washed with light petroleum. By treatment in this manner, 80 grams of resin yielded 15 grams of a white solid, from which two substances of constant melting point ( A and B) were separated by means of ;t somewhat tedious process of fractional ex- traction with ethyl acetate and repeated crystallisation from the same solvent. The smaller fraction A , weighing 2.2 grams, crystallised from ethyl acetate in oblique, square plates melting a t 212', and was proved by analysis and the preparation from it of acetyl and benzoyl derivatives, to be identical with 1 : l'-dihydroxy-5 : 5 : 5' : 5'-tetrarnethyZdicyclohexccne, described on page 77.The larger fraction, B (5-1 grams), was analysed with the following results : 0,1275 gave 0.3529 CO, and 0.1353 H,O. This substance crystallised from ethyl acetate in radiating clusters C=75.48; H= 11.79. C,,H,,O, requires C = 75.59 ; H = 11.81 per cent.72 CROSSLEY AND RENOUF: ACTION OF REDUCING AGENTS ON of transparent, leaf-like aggregates melting a t 173-174O,* nor was this melting point altered by repeated crystallisation. On account of this fact, and also because i t sublimed unchanged (m. p. 171-172') in long, silky needles, it was for a long time thought that the body was homogeneous and consisted of 3 : 3'-dihydroxy-5 : 5 : 5' : 5'-tetramethyl- dicyclohexane; but the following experiments prove that this is not the case.Action of Benxoyl Chloride on the Substance, m. p. 173--174'.-One gram of the substance was heated on the water-bath for two hours with excess of benzoyl chloride, the whole shaken with sodium hydroxide solution, extracted with ether, the ethereal solution washed with water until free from alkali, dried over calcium chloride, and the ether evaporated. The solid residue, weighing 1.6 grams, was dissolved in a large amount of absolute alcohol, when, on cooling, 0.4 gram of solid separated (filtrate = A ) melting a t 186-192", and, after a further crystallisation, a t 199'.This substance was proved by the mixed melting point method and by analysis to be identical with the dibenzoyl derivative of 1 : 1'-dihydroxy-5 : 5 : 5' : 5'-tetramethyldicycZohexane de- scribed on page 77, which latter substance (m. p. 212') it gave on hydrolysis with alcoholic potassium hydroxide. The filtrate A was evaporated, when several fractions of crystals (in all 1.0 gram) were obtained having the same melting point. These were purified by recrystallisation from ethyl alcohol : 0.1174 gave 0.3335 CO, and 0.0892 H,O. C,,H,,O, requires C = 77.92 ; H = 8.22 per cent. The dibenxoyl derivative of 3 : 3'-dihydroxy-5 : 5 : 5' : 5'-tetramethyl- dicyclohexane is readily soluble in the cold in benzene or chloro- form, moderately soluble in acetone or ethyl acetate, and crystal- lises from absolute alcohol in radiating clusters of flattened needles melting at 133-134'.When hydrolysed by boiling with alcoholic potassium hydroxide and the solution poured into water, a white solid separated which was purified by recrystallisat ion from benzene : C = 77.48 ; H = 8.44. 0.1071 gave 0-2968 CO, and 0.1121 H,O. 3 : 3'-Dihydroxy-5 : 5 : 5' : 5'-tetramethyZdicyclo?~exane, C=75.57; H=11*63. Cl,H,o02 requires C = 77.59 j H = 11 *8 1 per cent. is readily soluble in the cold in absolute alcohol or acetone, less so in * I t has since been ascertained that on standing the crystals become opaque and the melting point less sharp. A specimen melting at 173-174", when first isolated, melted after six weeks a t the same temperature, but did not become clear until 180", and two months later partially melted at 173-174", but did not clarify until 195".5-CHLORO-3-KETO-1: I-DIMETHYL-A4-TETRAHYDROBENZENE.73 ethyl acetate, crystallises From chloroform or benzene in clusters of scaly needles melting a t 183', and sublimes unchanged in fern-like aggregates of flattened needles. It does not decolorise a chloroform solution of bromine, gives with concentrated sulphuric acid a salmon- pink colour turning to deep orange, and is converted by benzoyl chloride into the above-mentioned dibenzoyl derivative melting at 133-134", and no other substance. On mixing three parts of this product with one part of 1 : 1'-dihydroxy- 5 : 5 : 5' : 5'-tetramethyldicyclohexane, that is, in the proportion which formed the constant melting mixt8ure ( 173-1 74") of these two sub- stances, the melting point was 172--173O, but complete clarification did not take place until 179'.Moreover, this mixture sublimed in silky needles melting a t 171-172" and becoming quite clear a t 176'. Action of Acetyl Chloride oy6 the Substance m. p. 173--174'.-One gram of the substance was heated with an excess of acetyl chloride for two hours, and the solvent evaporated, when the viscous residue so obtained rapidly solidified on rubbing with a few drops of alcohol. It was spread on a porous plate and crystallised from absolute alcohol, when 0.2 gram of needle-shaped crystals separated which melted a t 130°, and proved to be identical with the diacetyl derivative of 1 : l'-dihydroxy-5 : 5 : 5' : 5'-tetram6thyldicycZoh~xnne described on page 77.On addition of water to the alcoholic mother liquor, further amounts of needle-shaped crystals separated, which were purified by crystallisahion from dilute alcohol : 0.1124 gave 0,2916 GO, and 0.1060 H,O. C,oH3,0, requires C = 71.01 ; H = 10.06 per cent. The dicccetyl derivative of 3 : 3'-dihydroxy-5 : 5 : 5' : 5'-tetramethyl- dicyclohexane is extremely soluble in the ordinary organic media, and crystallises, as above stated, in clusters of slender needles melting at 68". It is readily hydrolysed by alcoholic potassium hydroxide, yielding 3 : 3'-dibydroxy-5 : 5 : 5' : 5'-tetramethyldicycZohexane melting at 183". C = 70.75 ; H= 10.47. 11. TVith Sodium in Absolute Alcoholic Solution. Twenty grams of cbloroketodimethyltetrahydrobenzene were dis- solved in 400 c .~ . of absolute alcohol in a flask attached to a reverse condenser, and 32 grams of sodium, cut in thin slices, gradually added. At the end of the reaction the whole was poured into a large volume of water, extracted three times with ether, the ethereal solution washed with water, dried over calcium chloride, the ether evaporated, and the residue distilled under diminished pressure, when the following fractions were obtained at 35 mm. : 95-135' = 4.5 grams, 135-150' = 10.7 grams, resinous residue = 1.5 grams.74 CROSSLEY AND RENOUF: ACTION OF REDUCING AGENTS ON The fraction 95-1 35' gave with sulphuric acid the marked colour reaction characteristic of 3-hydroxy-1 : 1 -dimethylhexahydrobenzene (Trans., 1905, 87,,1495), and no doubt consisted of a mixture of this substance with 3-J~ydroxy-5-ethoxy-1 : 1-dimetJ~yZJ~exahgdrobenxene (see below).The latter compound will, however, form the starting point of another investigation, and a detailed description of its properties and reactions will be reserved for a future communication. Thefi.action 135-150" was redistilled, when S grams of a colourless liquid passed over quite constantly at 135' a t 25 mm. : 0.1499 gave 0.3835 CO, and 0.1559 H,O. 3-Hy d roxy- 5 - e t J~oxy - I : 1 -dime t JL y 1 Jhexah y clr o be nxen e, C= 69.77; H = 11.55. C,,H,,O, requires C = 69-76 ; H = 11 -62 per cent. is a colourless, oily liquid boiling a t 135" at 25 mm., and possessing a faint celery-like odour. It does not solidify when cooled in a mixture of ice and hydrochloric acid, does not decolorise a soliition of bromine in chloroform, and with concentrated sulphuric acid gives only a very faint orange-pink colour.That it contains an ethoxy- group was proved by a Zeisel determination, carried out according to the directions given by Sir W. H. Perkin (Trans., 1903, 83, 1367) : 0.3445 gave 0,4397 AgI. C,H,,O*OC,H, requires -OC,H, = 26.16 per cent. The result is somewhat low, but, as pointed out by Perkin, ethoxy- determinations, as a rule, give results from 1 to 2 per cent. below the calculated. The acetyl derivative, prepared by the action of acetyl chloride, is a colourless, refractive, oily liquid boiling a t 129" a t 22 mm., and possessing a sweet, slightly camphoraceous odour : -OC,H, = 24.44. Found, C = 67.33 ; H = 10.41.C,,H,,O, requires C = 67.29 ; H = 10.28 per cent. The benxoyl derivative, obtained in the usual manner, is a faintly coloured, highly refractive liquid boiling a t 236' a t 50 mm., and having an odour somewhat resembling that of ethyl benzoate : Found, C = 73.75 ; H = S.85. CI7H,,O, requires C = 73-91 ; H = 8.69 per cent. III. FitJi Zinc Dust in Aqueous Alcoholic Solution. Two quantities of 29 grams each of chloroketodimethyltetrahydro- benzene were separately dissolved in 108 C.C. of 90 per cent. alcohol, 40 grams of zinc dust mixed with an equal volume of sand added, and the5-CHLORO-3-KETO-1 : 1 -DIMETHYL-A4-TETRAHYDROBENZENE. 7 5 whole heated on the water-bath for four to five hours. The major portion of the alcohol was then distilled off, the residue poured into water, extracted six times with ether, the ethereal solution washed with water, dried over calcium chloride, and the ether evaporated.The residue did not boil constantly (95-102" at 78 mm.), but the major portion passed over between 100-102° a t 7s mm., and was analysed with the following result : 0.1403 gave 0.3952 CO, and 0.1303 H,O. C = 76.82 ; H= 10.31. C,H,,O requires C = 77.42 ; H = 9&67 per cent. C,H,,O ,, C - 76.19 ; H = 11.11 ,, This colourless liquid, which was free from halogen, possessed a pungent, camphoraceous odour, and gave the marked colour reaction with concentrated sulphuric acid characteristic of 3-keto-1 : 1- dimethyl- A4-tetrahydrobenzene (see page 78), but the above analysis proved it to be a mixture of this substance with 3-keto-1 : l-dimethylhexahydro- benzene * (see page 81), and this is further proved by the fact that on oxidation as-dimethylsuccinic acid, the lactone of a-hydroxy-pfl- dimethylglutaric acid, and Pfl-dimethyladipic acid were obtained, and by the following experiments.Nineteen grams of the mixed ketones were dissolved in 52 C.C. of 90 per cent. alcohol, 19 grams of zinc dust and an equal volume of sand added, and the whole heated on the water-bath for ten hours. The major portion of the alcohol was distilled off, the residue poured into water and distilled in steam (residue of distillation=A). The distillate was extracted six times with ether, and the residue obtained on evaporation of the ether again heated with zinc dust in aqueous alcoholic solut,ion ; this process was repeated three times, when it mas found that the residue of the steam distillation did not furnish any further solid matter.The liquid volatile with steam (6 grams) still gave the colour reaction of the unsaturated ketone, and in order to remove these last traces it was treated with dilute potassium permanganate solution in the cold. Only a very small amount of the oxidising agent was used up and the recovered liquid, which boiled quite constantly a t 80" a t 36 mm., no longer gave a colour with sulphuric acid, and consisted of pure 3-keto-1 : I-dimethylhexahydrobenzeae (see page 81) : Found, C = 75.87 ; C,H,,O requires C = 76.19 ; H = 11.11 per cent. The semicarbazone crystallised from alcohol in radiating clusters of scaly needles melting a t 195O with decomposition and evolution of gas.* The action of zinc dust on chloroketodimetbyltetrahydrobeiizeiie iii the cold gives precisely these same results, €€ = 11 -14.76 CROSSLEY AND RENOUF: ACTION OF REDUCING AGENTS ON The residue of the steam distillation ( A ) was extracted with ether, the ethereal solution dried over calcium chloride, and the ether evaporated, when 5.7 grams of a viscid, semi-solid mass remained. This and similar material (in all 9.3 grams) from the residues of the other steam distillations were triturated with light petroIeum (b. p. 40-6O0) and, after filtering, yielded 3.9 grams of a clean, white solid, which after crystallisation from ethyl acetate gave 2.5 grams melting sharply at 148' : * 0,1092 gave 0.3077 CO, and 0.1048 H20.C,,H2,0, requires C = 76.SO ; H = 10.40 per cent. 1 : l'-Dihydrox~-5 : 5 : 5' : 5'-tetramethyl-A2 21-di~y~lohe~en~, C = 76.84 ; H = 10.66. is fairly readily soluble in the cold in chloroform, benzene, or acetone, less readily in alcohol or ethyl acetate, and crystallises from the latter solvent in stout, transparent, rhombohedral plates melting at 148O. It sublimes unchanged in microscopic, rhombohedral plates, and does not give a colour reaction with sulphuric acid, nor was it found possible to prepare acetyl or benzoyl derivatives under ordinary con- ditions. Action of Bromine.-One gram of the substance was dissolved in 15 C.C. of chloroform and a solution of bromine in the same solvent gradually added, when, without the solution becoming sppreciably warm, the bromine was rapidly absorbed and torrents of hydrogen bromide evolved.The chloroform was evaporated and the residue (2 grams) purified by crystallisation from benzene : 0.1360 gave 0.2026 CO, and 0.0560 H20. 0.1102 ,, 0.1316 AgRr. Br=50*81. The tribromoanhgdride (1) of dihg drox y tet ramethyldicyclohexene is very slightly soluble, even on boiling in alcohol, acetone, or ethyl acetate, but is sufficiently soluble to be crystallised from either benzene or chloroform, when it separates in small, glistening, flattened needles, which on heating in a capillary tube begin to darken a t 245' and melt with complete decomposition a t 250'. It only dissolves slowly when heated with fuming nitric acid, and, on cooling, faintly yellow, scaly needles separate, melting at 2 12' with decomposition, but no more definite information can be given a t the present time owing to lack of material.* There is a second product of lower inelting point present in this and also in the solid isolated from the action of zinc dust and acetic acid on chloroketodiinethyl- tetrahydrobenzene (see p. 83), but up t o the present time it has not been found possible to draw any very definite conclusions as t o the constitution of this sub- stance, though it certainly possesses the forniula C16H2602. C = 40.63 ; H = 4.57. C,,H,,OBr, requires C = 40.93 ; H = 4.47 ; Br = 5 1 *1'7 per cent.5-CHLORO-3-KETO-1 : I-DIMETHYL-A4-TETRAHYDROBENZENE. 77 Action of Sodium in moist Ethereal Xo1ution.-Two grams of dihydroxytetramethyldicyclohexene were dissolved in a mixture of 20 C.C.of alcohol and 40 C.C. of ether, the solution floated on 30 C.C. of water, and 12 grams of sodium, cut in very thin slices, gradually added. As the reaction proceeded, a white solid separated which dissolved on the addition of further small amounts of alcohol and ether. The major portion of the solvents was then evaporated, the residue poured into water, the solid which separated, filtered, washed with water, spread on plate (2.3 grams), and crystallised from ethyl acetate : 0.1077 gave 0.2985 CO, and 0.1154 H20. 1 : l’-Dihydroxy-5 : 5 ; 5‘ ; 5’-tetramet7~?/ldicyclo?~exane, C=75*58; H= 11.90. C,,H,,O, requires C = 75.59 ; H = 11.81 per cent. is but slightly soluble in chloroform or benzene, not readily so in acetone, alcohol, or ethyl acetate, and crystallises from the latter solvent in transparent, oblique, square plates melting at 212’.It sublimes unchanged in fern-like aggregates, and does not give a colour reaction with concentrated sulphuric acid. Although so readily formed by the action of sodium in moist ethereal solution on dihydroxytetra- methyldicyclohexme, no trace of it could be found on heating the iatter substance for twelve hours with zinc dust in aqueous alcoholic solution. The diacetyl derivative, prepared by the direct action OF acetyl chloride, is readily soluble in the cold in benzene, chloroform, acetone or ethyl acetate, and crystallises from alcohol in sheaves of needles or on slow crystallisation in well-formed rectangular prisms melting a t 130° : Found, C = 70.99 ; H = 10.31.C20H8404 requires C = 71.00 ; H = 10.06 per cent. The dibenxoyl derivative, prepared as described on page 72, is readily soluble in the cold in benzene or chloroform, moderately soluble on heating in aIcohol or acetone, and crystallises from ethyl acetate in transparent, four-sided plates melting a t 199’ ; Found, C = 77.82 ; H = 8-26. C,oH,,O, requires C = 77.92 ; H = 8.22 per cent. Both the diacetyl and dibenzoyl derivatives regenerate dihydroxy- tetramethyldicyclohexane (m. p. 2 12’) when hydrolysed with alcoholic potassium hydroxide.78 CROSSLEY AND RENOUF: ACTION OF REDUCING AGENTS ON IT. TrTith Zinc Filings in Aqueous AZcoT~oZic Solution. Twenty grams of chloroketodimethyltetrahydrobenzene were dis- solved in 108 c .~ . of 90 per cent. alcohol, 30 grams of zinc filings added, and the whole heated on the water-bath, during which time zinc chloride separated. The heating was continued for five hours, when the major portion of the alcohol was distilled off, the residue poured into water, the solution extracted six times with ether, the ethereal solution washed once with water, dried over calcium chloride, and the ether evaporated. The residue was again treated with zinc filings and 90 per cent. alcohol, and the whole process repeated six times, when the resulting liquid was found t o be practically free from halogen. On distilling under diminished pressure, the major portion (free from halogen) passed over quite Constantly a t 89.5' a t 30 mm., leaving a small residue which contained chlorine. This fraction was again distilled and analysed : 0.1227 gave 0.3471 CO, and 0.107s H,O.C8H,,0 requires C = 77.42 ; H = 9.67 per cent. As the yield of the ketone prepared in this manner is not very satisfactory, the reaction mas tried in the cold, when the amount formed is considerably increased, as is also the case when using the zinc-copper coup]& For this purpose 10 grams of chloroketodimethyl- tetrahydrobenzene were dissolved in 15 C.C. of 90per cent. alcohol, 10 grams of zinc-copper couple added, and the whole allowed to stand a t the ordinary temperature, when zinc chloride rapidly separated. After forty-eight hours i t was filtered (filtrate = A ) , the residue mashed with alcohol, the alcoholic solution evaporated, added to A , which was again treated with the zinc-copper couple and this process repeated four times, when the addition of fresh zinc-copper couple did not pro- duce any further separation of zinc chloride.The whole mas then poured into water and worked up as described above. On distilling the residue under diminished pressure, it passed over for the most part quite constantly, leaving but a small residue which contained chlorine, and after a second distillation it boiled a t 83.5' a t 25 mm. : C = 77-21 ; H = 10.01. C = 77.15 ; H = 9.76. 0.1183 gave 0.3349 CO, and 0.1066 H,O. C,H,,O requires C = 77.42 ; H = 9.67 per cent. As large quantities of this ketone are being prepared for another investigation, a description of the best method of obtaining it is re- served for a future communication.C H *CO 3-Keto-1 : 1-dimetl~yLA4-tetrahydrobenxene, CMe,<CH,2.CH>CH, is a colourless, highly refractive liquid, boiling a t 88.5' at 32 mm., and possessing an odour of almonds, which soon becomes disguised by a5-CHLORO-3-KETO-1 : 1-DIMETIIY L-A4-TETRAHYDROBENZENE. 79 pungent smell of camphor. When treated with a n equal bulk of con- centrated sulphuric acid it gives a blood-red colour, turning to plum- red, then gradually to violet and finally disappears, A solution of bromine in chloroform is immediately decolorised on addition of a chloroform solution of the ketone, and an attempt was made to determine the bromine absorption value, but without success, for although there is an apparent end reaction when 1 molecule of bromine has been added, yet a t t h i s point hydrogen bromide is evolved, and the amount increases rapidly with further addition of bromine.The semicurbaxone prepared by adding the ketone to a concentrated alcoholic solution of seniicarbazide acetate, crystallises from methyl alcohol in nacreous scales melting a t 195’ to a clear yellow liquid which slowly evolves gas. The preparation requires to be carried through as rapidly as possible on account of the ease with which tho semicarbazone decomposes, especially when in solution : Found, N = 23.39. Action of Nydroxylamine. -Two grams of hydroxylamine hydro- chloride were dissolved in the smallest amount of water, alcohol and 2 grams of the ketone added, the solution neutralised with sodium hydroxide and allowed to stand twelve hours.It was then poured into a saturated solution of brine, the whole extracted with ether, the ethereal solution washed with water, carefully dried over calcium chloride, and the ether distilled off. As the slightly coloured residue showed no sign of solidification on standing or on cooling, and as, unlike the oxime of the corresponding saturated ketone (see page Sl), it could not be distilled even under a low pressure without complete decomFosition, the nitrogen was determined in a specimen of the liquid prepared as above described : C9H,,0N, requires N = 23.20 per cent. Found, N = 11.15. Although the nitrogen found is not very closely in accord with the calculated amount, it is sufficiently near to show that the substance is a simple oxime and not a hydroxy lamino-oxime, which would require 16.27 per cent.of nitrogen. Moreover, .when the crude oxime was treated with benzoyl chloride, although decomposition took place resulting mainly in the formation of a dark coloured liquid, a small quantity of a solid benzoyl deriv- ative was isolated, crystallising from absolute alcohol in nacreous, scaly needles melting a t 171-172” and containing 6.10 per cent. of nitrogen, whereas the calculated amount for the benzoyl derivative of the simple oxime is 5.76 per cent. Oxidation of the Ketone.-Five grams of the ketone were suspended in 125 C.C. of water, and a cold saturated solution of potassium C,H,,ON requires N= 10.07 per cent.80 CROSSEEY AND RENOUF: ACTION OF l2EDUCING AGENTS ON permanganate added until no longer decolorised.As oxidation took place very rapidly, the solution was cooled by the addition of small quantities of ice and then worked up in the usual way, when 5.2 grams of a white solid were obtained. This was dissolved in water and the solution saturated with hydrogen chloride, when, on standing, 1 *8 grams of needle-shaped crystals separated melting a t 139-1 40°, nor was this melting point lowered on mixing with pure ns-dimethyl- succinic acid. The identity of this substance was further proved by converting a portion into the anilic acid, which crystallised from methyl alcohol in nacreous, scaly needles melting at 187' with evolution of gas. The mother liquor from the as-dimethylsuccinic acid was evaporated to dryness, i n d the solid residue (2.5 grams) heated with excess of acetyl chloride for two hours.On evaporation of the solvent, the residue solidified ra.pidly when stirred, and after frequent crystallisa- tion from benzene was obtained in stellar aggregates of transparent needles melting at 110-1 1 lo : 0.1100 gave 0,2142 CO, and 0,0654 H,O. 0.2019 required 12.85 C.C. N/lO NaOH. On heating the solution used for this titration, a further 11.5 C.C. of 8/10 NaOH were required for neutralisation. This substance is evidently, therefore, the lactone of a-hydroxy-PP-dimethylglutaric acid (compare Perkin and Thorpe, Trans., 1899, 75, 56). C = 53.10 ; I€ = 6-60. C7H,,0, requires C = 53.16 ; H = 6.33 per cent. Calculated, 12.78 C.C. C7H,,04 molecular weight calculated 158. Found, 157. V. Vith Zinc Dust in Acetic Acid Solution.Two quantities of 20 grams of chloroketodimethyltetrahydrobenzene were separately dissolved in 80 grams of glacial acetic acid contained in a flask attached to a reverse condenser and heated on a sand-bath. Thirty-three grams of zinc dust were gradually added, a t first in very small amounts as the reaction is a vigorous one. The operation takes about twenty hours, at the end of which time the whole was neutral- ised with sodium hydroxide and distilled in steam (residue = A), the distillate extracted six times with ether, the ethereal solution washed with potassium hydroxide, then with water, dried over calcium chloride, and the halogen free residue obtained on evaporation of the ether distilled under low pressure and analysed : 0.1080 gave 0.3017 CO, and 0.1090 H,O.The same substance may be obtained by using diIute instead of For this purpose, 20 grams of chloroketodimethyl- C= 76.18 ; H= 11-21. C,H,,O requires C = 76.19 ; H= 11.11 per cent. glacial acet'ic acid.5-CHLORO-3-KETO-1 : I-DIMETHYL-A*-TETRAHYDROBENZENE. 81 tetrahydrobenzene were dissolved in 48 grams of glacial acetic acid and 40 C.C. of water and treated with 33 grams of zinc dust. Towards the end of the reaction it was necessary to add small quantities of glacial acetic acid. The product was worked up as described above, when the steam distillate (residue of distillation = B ) yielded a liquid boiling a t 77.5" at 27 mm : 0.1316 gave 0.3660 CO, and 0.1319 H,O. C,HI,O requires C = 76.1 9 ; H = 11 -1 1 per cent. The only difference between this specimen of the ketone and that prepared by the action of glacial acetic acid was that the former gave a faint colour reaction indicative of the presence of the corresponding unsaturated ketone (see page 75).The amount present must have been extremely small, as it did not affect the above quoted analysis, and after treating this specimen of the ketone in the cold with dilute potassium permanganate it no longer gave the colour reaction. C- 75.86 ; H = 11.13. clear, colourless, refractive liquid boiling a t 75.50 at 25 mm. and possessing a strong camphoraceous odour. The yield when using glacial acetic acid is 50-55 per cent. of the theoretical amount, and with dilute acetic acid somewhat less. The ketone does not give a colour reaction with sulphuric acid. When dissolved in chloroform and a solution of bromine in the same solvent added, no bromine is absorbed for some little time, then the colour suddenly disappears, hydrogen bromide is evolved, and on adding more bromine it is rapidly absorbed and torrents of hydrogen bromide are given off.This seems, perhaps, an unusual behaoiour for a saturated compound, but it has been ascertained that ketohexahydrobenzene behaves in exact'ly the same way towards bromine, and the phenomenon is probably connected with the conversion of the bodies into derivatives of the aromatic series, a point which is receiving a t tention. The semicurbaxone, prepared in the usual manner, crystallises from absolute alcohol in radiating clusters of glistening, flattened needles melting at 195" with evolution of gas and much greater decomposition than the semicarbazone of the corresponding unsaturated ketone (see page 79) : Found, N = 22.72.C,HI7ON, requires N = 22.95 per cent. The oxime was prepared from 5 grams of the ketone exactly as described on page 79. It distilled quite constantly a t 132' a t 37 mm. as a clear, colourless, syrupy liquid with a sickly odour somewhat reminiscent of both camphor and celery : Found, N = 9*S3. C8H,,0N requires N = 9.93 per cent. As it did not solidify on cooling or on standing for some consider- VOL. XCI. G82 5-CHLORO-3-KETO-1 1-DIMEI'HPL-A4-TET€idHY DROBENZENE. able time, it was converted into the benaoyl derivative, which is readily soluble in the cold in the usual organic solvents, but crystallises from dilute alcohol in nacreous scales melting a t 69' : Found, N = 6-05, C,,H,,O,N requires Tu' = 5.7 1 per cent.This substance was then hydrolysed by warming with potassium hydroxide (1 : a), when suddenly the solution became quite clew, and on passing carbon dioxide through i t the oxime was prc- cipitated. I t was extracted with ether, Bc., when, on standing some time in a vacuum, it solidified completely. The solid melted a t 43-44', was extremely soluble in the cold in the usual organic media, and was proved by analysis to consist of the pure oxime. Oxidation of the Keto9zs.-Ten grams of the ketone were suspended in 250 C.C. of water and a saturated solution of potassium perman- ganate gradually added. The oxidising agent was only used up extremely slowly in the cold, and therefore the whole was heated on the water-bath, when the reaction required about forty hours for com- pletion. The product was treated in the usual way, yielding 5 grams of solid, which were dissolved in water and the solution saturated with hydrogen chloride, but even on long standing no crystals were deposited (compare Trans., 1906, 89, 1552). The solution mas evaporated, care being taken to get rid of' all the hydrogen chloride, and the residue crystallised from a mixture of chloroform and light petroleum, when 2.8 grams of radiating clusters of rhombic plates separated, melting a t 85-86", nor was this melting point lowered on mixing with pure PP-dimethyladipic acid : 0.1193 gave 0.2411 CO, and 0.08'79 H20. C,H1,O, requires C = 55.17 ; H = S.04 per cent. On evaporating the mother liquor from the dimethyladipic acid, there remained a solid residue, which was dissolved in water and the solution saturated with hydrogen chloride, but on standing only a few minute crystals separated. These were filtered and hydrogen chloride allowed to escape from the solution, when compact crystals formed, consisting of pure PP-dimethyladipic acid, thus proving that only minute traces of the isomeric aa-dimethyladipic acid could have been produced during the oxidation. The residue from the steam distillation A (see page 80) was acidified with acetic acid, extracted with ether, the ethereal solution washed with potassium hydroxide solution, then with water, dried over calcium chloride, and the ether evaporated, when a residue of 9.2 grams remained which partially solidified. This was triturated with light petroleum (40-60°), yielding 3 grams of solid, from which on crystallisation from ethyl acetate there were obtained 2 grams of pure 1 : 1'-dihydroxy-5 : 5 : 5' : 5'-tetramethyl-A2:2'-dicycZohexene (see C=55*12; H=S*18.THE VISCOSITY OF LIQUln MIXTURES. 83 page 76). The residue from the steam distillation B (see page 80) was treated in a similar manner, except that the trituration with light petroleum was unnecessary, when it yielded 1.4 grams of 1 : 1'-di- hydroxy-5 : 5 : 5' : 5'-tetramethyl- A2'2-dicyclohexene. Both the above solid residues from A and B contained rz large pro- portion of the low melting solid referred to in the footnote on page 76. The authors take this opportunity of expressing their thanks to the Research Fund Committee of the Chemical Society for a grant which has, in part, defrayed the expenses of this investigation. RESEARCH LABORATORY, PHARMACEUTICAL SOCIETY, 17, BLOOMSBURY SQUARE, W.C.