DERIVATIVES OF DIBENZOYLME3ITYLENE. 1311 CXXXII1.-Derivatives of Dibenxoylrnesitylene. By WILLIAM HOBSON MILLS, Fellow of Jesus College, Cambridge, and THOMAS HILL EASTERFIELD. THE synthetical production of polycyclic hydrocarbons which contain more than four rings has hitherto been but rarely effected. For this reason, it seemed desirable to undertake a further study of the substance dibenzoylmsitylene, CH3 C6H5*CO(\,CO*C H prepared by Louise in 1885 (Ann. Chim. Phys., [vi], 6, 234) ; more particularly to work out a method for its preparation which would enable it to be obtained in quantity, and to study the acids which would be derived from it by the oxidation of two of its methyl groups to carboxyl. For just as o-benzoylbenzoic acid is condensible by phosphoric oxide (Behr and van Dorp, Ber., 1874, 7, 578) or concen- trated sulphuric acid (Liebermann, Ber., 1874, 7, 805 ; Perkin, Trans., 1891, 59,1012) to anthraquinone, so was i t to be expected t h a t the two dibenzoyluvitic acids.: CH,”GH, ’’ would, under the influence of dehydrating agents, be convertible into quinones of the formulae : respectively, from which the corresponding hydrocarbon might be prepared by reduction. The present paper deals mainly with the preparation of dibenzoyl- poesitylene and with the acids obtained from if by oxidation. An1312 MILLS AND EASTERFIELD : account of the action of dehydrating agents on the latter is reserved for a later communication. Louise (Eoc. cit.) prepared dibenzoylmesitylene by treating a mixture of mesitylene and benzoyl chloride with a very small quantity of aluminium chloride a t a high temperature. The high temperature, however, and its attendant disadvantages may be avoided, provided that the benzoyl chloride is allowed to act on the mesitylene in warm carbon disulphide solution in the presence OF a large excess of aluminium chloride, the procedure thus closely corresponding to that worked out by Victor Meyer (Ber., 1896,29, 846, 1413) for the produc- tion of the diacetyl derivatives of durene and mesitylene.I n this mmner, the preparation can be carried out on a comparatively large scale, and excellent yields may be obtained. Incidentally, the action of reducing agents on the ketone was studied. Zinc and potash acting on an alcoholic solution of the ketone reduce it to dihydroxydibenzylmesitylene, CH.9 a substance not yet obtained in a crystalline condition. When boiled with hydriodic acid (b. p. 127') and yellow phosphorus, dihydroxydibenzylmeaitylene is converted i n t i a beautifully crystalline compound melting at 89O. ' I t s composition and vapour density, taken together with its mode of formation, show conclusively that this substance must be dibernxylmesitylene : Louise (Ann. Chim. Phys., 1885, [vi], 6, 197), however, has described as dibenzylmesitylene a substance melting at 1 31°, which he obtained directly from mesitylene by the action of benzyl chloride and aluminium chloride. Since tribenzylmesitylene would have a composition so similar to that of dibenzylmesitylene that analysis alone could not distinguish between them, it seems not impossible that the substance obtained by Louise may be tribenzylmesitylene. It was, however, in the oxidation of the ketone that our interest was chiefly centred.After various preliminary experiments, it was found that dilute nitric acid was the only suitable oxidising agent.DERIVATIVES OF DIBENZOYLMESITYLENE. 1313 When heated with dilute nitric acid in sealed tubes at 140' for 14 hours, dibenzoylmesitylene is converted into a mixture of acids, consisting mainly of the two desired clibenzoyluvitic acids. CH3 C,H~*CO/\CO* C,H, UO,H()CO,H r. It was to be expected that would be effected by treating CO,H C,H, CO/'\CO*C,H, C0,H(/CH3 11. the separation of two such substances the mixture with methyl alcohol and hydrogen chloride, when the acid of formula I would yield a normal ester, whilst that of formula 11, in accordance with Victor Meyer's esterification rule, would give an acid ester only.As a matter of fact, however, both acids were by this treatment rapidly converted into their dimethyl esters. Further investigation showed that these acids were really the two dibenzoyluvitic acids. The interesting fact was thus established that the benzoyl group as an ortho-substituent forms an exception to Victor Illeyer's esterification rule." None the less, this treatment led to a method of separating the two acids, for whilst the methyl ester of one is readily soluble in methyl alcohol, that of the other is so sparingly soluble that it separates out almost completely from the esterificntion mixti ire. By saponifying the two esters thus obtained, there resulted the corresponding dibenzoyluvitic acids, that from the sparingly soluble ester melting at 262O, that from the other a t 213'.There was, however, still no clue as to which of the two possible formuls belonged to which acid. To gain some evidence on this point, their rates of esterification were compared, but since the acid with the sparingly soluble ester was found to be the more rapidly esterifiable, no definite conclusion could be drawn, i t being impossible to be certain that the removal of the ester from the sphere of action through its insolubility did not accelerate its rate of formation relatively to that of the ester of the other acid-the two acids could not be esterified under comparable conditions.Another method suggested itself which promised to yield more definite results. This was to prepare the two dibenzoylmesitylenic acids which would result from the oxidation of one only of the three methyl groups of dibenzoylmesitylene : * Some time after this had been observed, an account of the similar behaviour of 3 : 6-dichlorobenzoylbenzoic acid was published by Graebe (Ber., 1900, 33, 2026).1314 MILLS AND EASTERFIELD : CH3 CO,H c6H5*cd\co C6H, c,H~~co/'\co- C,H, CO,HI/CH, CH,j)CH, 111, IV. Then, whereas the unsymmetrical acid (111) on f urthei. oxidation should yield a mixture of the two dibenzoyluvitic acids, the sym- metrical acid (IT) should yield only one of these, namely, that of the formula I1 : CO,H CO,H()CH, Y and thus would be fixed the constitution of all four acids in question.Although dibenzoylmesitylene is not attacked by dilute nitric acid when boiled with it under ordinary conditions, it is readily oxidised when the boiling point of the diluted acid is raised to 125' by the addition of large quantities of potassium and sodium nitrates, and it then yields a mixture of all the five acids which can result from it by the successive conversion of its three methyl groups into carboxyl. Of this mixture, one of the required dibenzoylmesitylenic acids (of melt- ing point 174') forms by far the largest constituent. The other acids are present in relatively small quantity, and the isolation of the second dibenzoylmesitylenic acid was a matter of some difficulty ; it was finally effected by crystallising the mixture of acids from ethyl acetate and from alcohol. The two dibenzoylmesitylenic acids thus obtained were then oxidised in alkaline solution with potassium permanganate, with the result that that of melting point 1'74' yielded a mixture of both dibenzoyluvitic acids, whilst that of meiting point 2 2 2 O under the same conditions yielded only the dibenzoyluvitic acid melting at 213'.The constitution of the five acids resulting from the oxidation of dibenzoylmesitylene was thus settled. The dibeqzoylmesitylenic acid of melting point 1 7 4 O , which, on oxidation, gives the two dibenzoyluvitic acids, must be the unsym- metrical acid of formula 111 : C,H,* CO/\CO*C,H, CH3 Y' 111. ' x C,H,*CO/\CO-C,H~ CO,H?)CH, CH CO,H c6H,*CO/\CO*C,H, c H 'C*IO/\CO*C,H, CO,H(/CO,H 6 0 , H u C H 3 I.11.DERIVATIVES OF DIBENZOY LMESITY LENE. 1315 The dibenzoylmesitybnic acid of melting point 222O, which, on oxidation, gives only one dibenzoyluvitic acid, must be the symmetrical acid of formula IV : C0,H CO,H C,K,.CO/\CO*C~N, \/ CO,H/ 'CH, IV. I1 It therefore follows that dibenzoyluvitic acid melting a t 213' is the unsymmetrical acid (formula II), and that melting at 262' is the symmetrical compound (formula I). EXPERIMENTAL. Preparation of Bi benxoylmesitytene, One hundred grams of aluminium chloride are rapidly pounded and covered with 250 C.C. of dry carbon disulphide. Eighty-five grams of benzoyl chloride are then added, and while the mixture is gently boiled under a reflux apparatus, 17 grams of mesitylene are slowly dropped in from a dropping funnel.The liquid a t once assumes a brownish-red colo u r . After boiling for 14 hours (by which time the aluminium chloride has almost completely dissolved), the mixture is allowed to cool and then decomposed by pouring on to lfr kilos. of pounded ice. After treating with concentrated hydrochloric acid t o dissolve the precipitated alumina, the carbon disulphide, containing the ketone and the excess of benzoyl chloride in solution, is separated, hydrochloric acid added, and the mixture heated in a current of steam-the heating must be continued for a t least half-anbhour after the carbon disulphide has been driven over to ensure complete decomposition of the benzoyl chloride, After cooling somewhat, the crude ketone sets t o a solid brown mass a t the bottom of the flask, from which the warm dilute acid, containing much benzoic acid in sol~ution, can be poured away.Dilut,e caustic soda is then added and the mixture again heated with steam just long enough t o ensure thorough agitation of the melted ketone with the hot soda. After cooling, the crude ketone is dissolved in hot alcohol, from which i t crystallises in somewhat pinkish plates. When recrystallised from alcohol, it is sufficiently pure for use in further operations. The yield is about '70 per cent. of the theoretical, but for the success of the preparation it is absolutely essential t h a t all the material be of the very best quality. To obtain the ketone perfectly pure, it must be distilled in a vacuum.1316 MILLS AND EASTERFIFXD : It boils at 275-285' under 19 mm.pressure. The resinous mass thus obtained is crystallised from light petroleum and then from alcohol, and forms beautiful, white plates which melt at 117'. On analysis : 0.1748 gave 0.5372 GO, and 0.0960 H,O. Dibenzoylmesitylene also crystallises well from carbon disulphide or In ether, benzene, or chloroform, i t is extremely C = 83.8 ; H = 6-1. C,,H,,O, requires C = 84.1 ; H = 6.1 per cent. glacial acetic acid, soluble. To Louise's account of its properties, there is nothing to add. Oxidation of Dibenxoylmesitylene. The ketone is boiled in quantities of 20 grams with 40 C.C. of nitric acid, sp. gr. 1.5, diluted with 160 C.C. of water to which are added 240 grams each OF potassium and sodium nitrates. The boiling is continued for 6 hours, the temperature assumed by the mixture being 1 2 5 O .When cold, the mixture is extracted with ether, the ethereal solution washed free from nitric acid by careful treatment with dilute caustic soda, and finally with water ; the acids formed are then removed from unaltered ketone and nitro-derivatives by shaking with dilute sodium carbonate solution. The resulting solution is just neutralised with hydrochloric acid, the carbon dioxide expelled by gently warming under reduced pressure, and excess of calcium chloride solution then added. A considerable quantity of a yellowish-white, gummy, calcium salt is then precipitated, which, on vigorous stirring, ad heres completely to the sides of the vessel, so that the clear mother liquor can readily be poured away.The latter contains the soluble calcium salts of the two dibenzoyluvitic acids together with some dibenzoyltrimesic acid. It is best worked up together with the oxidation product of as-dibenzoyl- mesitylenic acid. The mixture of-dibenzoylmesitylenic acids is isolated from the gummy calcium salt by shaking with hydrochloric acid and ether. The residuo from the ether is fractionally crystallised from ethyl acetate, In this way, from 320 grams of dibenzoylmesit,ylene, 5 grams of an acid melting at 219' were obtained. This is symmetrical dibenaoyl- mesitylenic acid. The acid contained in the mother liquors was obtained in crystalline form by evaporating off the ethyl acetate and allowing the residues to crystallise from glacial acetic acid. Unsymmct&a I dibenzoylmesitylenic acid separates in a finely crystalline condition, but crystallisation takes pIace remarkably slowly, separation not being complete until after the lapse of several days..DERIVATIVES OF DIBENZOY LMESITYLENE. 1317 as-Dibenxoylmsitylsnic Acid, C,H,-CO CO-C,H, . COP, GL3 After two or three recrystallisations from glacial acetic acid and then from alcohol, the pure acid was obtained as a white, crystalline powder melting at 174-1 75". On analysis : 0.1242 gave 0.3515 CO, and 0.0556 H,O. C24H1804 requires C: = 77-09 ; H = 5.03 per cent. The acid dissolves with great readiness in ether, ethyl acetate, acetone, chloroform, or carbon disulphide ; it is less soluble in alcohol, glacial acetic acid, or benzene, whilst in water or light petroleum it is almost insoluble.The sodium salt of the acid crystallises in silky needles from a strong solution of the acid in sodium carbonate ; it can be purified by recrystallisation from a small quantity of water and is anhydrous : C=77.12; H=4*98. 0.2690 gave 0.0516 Na,SO,. The calcium salt is obtained, as already stated, as a transparent, gummy mass by precipitating a solution of the sodium salt with calcium chloride. It is soluble in alcohol and, to some extent, in ether. The barium, zinc, lead, copper, silver, and ferric salts are precipitated on treating a solu- tion of the ammonium salt with solutions of the salts of these metals. Methyl as-Dibenxoylmesitylnate.-A solution of the acid in methyl alcohol is saturated with hydrogen chloride and then boiled for 2 hours in a slow current of the gas.The ester is separated by dilution with water and extraction with ether in the usual way. The residue from the dried ethereal solution, on crystallisstion from alcohol, deposits the ester in radially striated masses. After repeated crystallisation from alcohol, it melts a t 125-126". Na = 6.20. C,,HI70,Na requires N = 6.05 per cent. The magnesium salt resembles the calcium salt. On analysis : 0.1192 gave 0.3377 CO, and 0,0560 H,O. The ester is readily soluble in the common organic solvents with C = 77.26 ; H = 5.22. C,4€T,o0, requires C = 77.42 ; H = 5.37 per cent. the exception of light petroleum and cold methyl or ethyl alcohol. VOL. LXXXI. . 4 T1318 MILLS AND EASTERFIELD : The substance obtained by fractionally crystallising the acid obtained from the gummy calcium salt from ethyl acetate was further repeatedly recrptallised from alcohol and then from glacial acetic acid.The product melted apparently constantly at 220°, but in a series of con- cordant analyses the carbon was found almost 1 per cent. too low. Bince the amount of substance was becoming too small for complete purification in this way, i t was dissolved in a solution of sodium car- bonate and allowed to crystallise. The sparingly soluble sodium salt separated out as a beautiful mass of somewhat thick needles. It was recrystallised from a solution of sodium carbonate, in which it seemed to be distinctly less soluble than in water. The acid, separated from the recrystallised salt and crystallised from methyl alcohol, melted at 221--222', and now gave, on analysis, results in satisfactory accord with those required for the formula C23H180.4 : 0.1464 gave 0.4127 CO, and 0.0673 H,O.C23H,804 requires C E 77.1 ; H = 5.0 per cent. The acid crystallises best from ethyl or methyl alcohol. It can also be recrystallised from ethyl acetate or glacial acetic acid and is fairly soluble in ether. A solution of the sodium salt gives sparingly soluble, amorphous precipitates with the soluble salts of silver, copper, lead, and iron. The magnesium, calcium, barium, and zinc salts separate slowly, in crystalline form, on adding a solution of a salt of the respective metal to a solution of the sodium salt of the acid, C=76.9 ; H=5-1. The Relative Rates of Esterificcction of the Two Di6enxoyZmesityZenic Acids.As it seemed of some interest to gain an idea as to the amount of obstruction offered by the benzoyl groups to esterification, an attempt was made to compare the rates of esterification of the two dibenzoyl- mesitylenic acids under comparable conditions. Unfortunately, the experiment was marred by the sparing solubility of the symmetrical acid in methyl alcohol, a portion having crystallised out during the esterification, but the result, nevertheless, served to show that diortho-substitution by two benzoyl groups produces aDERIVATIVES OF DIBENZOYLMESITYLENE. 1319 retardation of the rate of esterification, which, although quite distinctly marked, is yet very small in comparison with that produced by such groups as CH,, C1, or NO,. The experiment was carried out a t the ordinary temperature, with sulphuric acid as catalyser in place of the much less convenient hydro- chloric acid.0.5 gram of each acid was dissolved in 20 C.C. of methyl alcohol, to each 4 C.C. of concentrated sulphuric acid were added slowly with cooling, and the mixtures were then allowed t o stand side by side for 16 hours. It was found that the amount of symmetrical acid which had crys- tallised out was 0.1 gram. This was filtered off and the esters were separated by dilution with water and extraction with ether in the usual way and weighed after removal of the ether. The unesterified acids were also similarly isolated, and weighed as a check on the result. The yields were : s-Acid. as-Acid. Ester ..................... 0.15 0.24 Acid ........................0.25 0.28 Acid cryst. during expt. 0.1 - H. Goldschmidt has shown (Ber., 1895, 28, 3218) that for esterifi- cation in thia way in the presence of a large excess of alcohol, the formula for the unimolecular reaction, k = tlog---, l a a - x is applicable. This gives for the esterification velocity of the 8-acid referred to that of the as-acid as 1 : 0.76 or 1 to 058, according as it is assumed that the quantity of acid which crystallised out was absent from or present in the reaction mixture during the whole period of reaction. The actual number will thus probably lie in the neighbour- hood of 0.65. Oxiahtion of s-Dibenxoylmesitytenic Acid. One gram of the acid was dissolved in 20 C.C. of hot dilute sodium carbonate and added to a molecular proportion of potassium perman- ganate (0.88 gram) dissolved in 10 C.C.of boiling water. The mixture was kept hot in the water-bath and became decolorised in about half- an-hour , The manganese dioxide was filtered off, the filtrate concentrated to 6-8 c.c., and allowed to stand overnight in order to allow the unoxidised acid to separate in the form of its sparingly soluble sodium salt. The small quantity of this which crgstallised out was filtered off and the filtrate was acidified and extracted with ether. The 4 T 21320 MILLS AND EASTERFIELD : residue from the ethereal solution was dissolved in 15 C.C. of methyl alcohol and treated with hydrogen chloride on the water-bath. No separation of the sparingly soluble ester of s-dibenzoyluvitic acid took place.The ester formed was isolated and crystallised several times from methyl alcohol. I n this way could be isolated a less soluble fraction melting sharply at 187'. This is the methyl estei- of dibenxoyltrimesic acid. The more soluble fractions were saponified by treating for 2 hours in the cold with methyl alcoholic potash. The mixture was diluted, the alcohol boiled away, and the solution acidified with dilute sulphuric acid and extracted with ether. The ethereal solution was concentrated to a small volume and the acid allowed to crystallise. It already melted at 21 lo. It was recrystallised from glacial acetic acid and found to be identical with the as-dibenzoyluvitic acid obtained by oxidising as-di- benzoylmesitylenic acid-in particular the melting point of a mixture cf the two preparations was the same as that of either separately.Hence the oxidation of the dibenzoylmesitylenic acid melting at 222' by a molecular proportion of potassium permanganate results in the production of only one dibenzoyluvitic acid (the as-acid) together with a small quantity of dibenzoyltrimesic acid. Oxidation of as-Dibenxoylmesit ylenic Acid. A boiling solution of 10 grams of the acid in 100 C.C. of dilute sodium carbonate solution was poured into a boiling solution of 8.8 grams of potassium permanganate in 100 C.C. of water and the mix- ture kept hot in the water-bath until decolorisation had taken place. This required about half-an-hour. After removal of the manganese dioxide, the filtrate was neutralised, concentrated somewhat, and mixed with a solution of calcium chloride to precipitate the unoxidised acid in the form of its gummy calcium salt.The filtrate was acidified, extracted with ether, and the ethereal solution dried by shaking for a short time with fused calcium chloride and then evaporated. The residue, dissolved in methyl alcohol, was treated with hydrogen chloride on the water-bath for two m three hours after saturation had taken place, The sparingly soluble ester of 8-dibenzyluvitic acid soon began to separate from this hot solution. After standing overnight, the crystalline ester was filtered off and washed with methyl alcohol. The more soluble ester of as-dibenzoyl- uvitic acid was isolated from the mother liquor by dilution with water, extraction with ether, and treatment with sodium carbonate in the usual way.DERIVATIVES OF DIBENZOYLMESITYLENE.1321 In one experiment, 10 grams of as-dibenzoylmesitylenic acid yielded 4.85 grams of sparingly soluble ester, 3-35 grams of soluble ester, and 1 -43 grams of unesterified acid. Methyl s-dibenxoyluuifate, (C6H5*CO),C,H(CH3)(C02*CH3)2.-The sparingly soluble substance obtained by esterification of the above- mentioned oxidation product consists mainly of this ester, but it con- tains also some of the ester of dibenzoyltrimesic acid, which is formed, to a limited extent, by the further oxidation of the two dibenzoyluvitic acids. It may be obtained free from the latter by recrystallisation successively from methyl alcohol, chloroform, and methyl alcohol again, and then forms small, white needles melting at 252'.It is very sparingly soluble in alcohol or ether (100 C.C. of alcohol dissolves 0.03 gram at the ordinary temperature), but dissolves much more readily in chloroform, and with the greatest ease in the other common organic solvents, excepting light petroleum. On analysis : 0.1259 gave 0.3328 CO, and 0.0547 H20. s-DibenxoyEuvitic Acid.-Finely divided methyl dibenzoyluvitate is suspended in hot alcohol and treated with an excess of a hot mixture of three volumes of alcohol with one of concentrated aqueous potash. The presence of water is necessary to prevent the precipitation of the potassium salt of the acid, which is sparingly soluble in alcohol. The ester rapidly goes into solution, and after standing for some time, saponification is complete.The acid is isolated by neutralising the product with dilute sulphuric acid, boiling off the alcohol, acidifying the residue, and taking up in ether. The residue from the ether is crystallised a few times from glacial acetic acid. The acid is thus obtained in prismatic crystals which melt at 262' and contain 2 mols. of water of crystallisation : C = 72.1 ; H = 4.8. C,,H,006 requires C = 72.1 ; H = 4-6 per cent. 0.3008, at 140°, lost 0.0251 H,O. The dried acid gave the following results on combustion : 001303 gave 0.3391 CO, and 0.0482 H20. C,,H,,O, requires C = 71 *1 ; H = 4.1 per cent. The acid dissolves very readily in alcohol, ether, ethyl acetate, or acetone, but is somewhat sparingly soluble in glacial acetic acid, and very sparingly eo in chloroform or benzene.The zinc and magnesium salts are soluble, and the calcium and barium salts sparingly so in water. Those of copper, iron, lead, and silver are formed as amorphous precipitates by treating the ammonium salt of the acid with a soluble salt of the appropriate metal. H,O=9*1. C,3H,60,,2H,0 requires H20 = 9.3 per cent. C = 70.9 ; H= 4.1.1322 MILLS AND EASTERFIELD : CO,H The readily soluble ester obtained by esterifying the product of the oxidation of as-di benzoylniesitylenic acid is saponified by treating its solution in methyl alcohol with an excess of methyl alcoholic potash. After 6 hours, the mixture is diluted with water and heated on the water-bath to drive off the methyl alcohol ; it is then acidified, and the acid extracted with a considerable volume of ether and purified as described under the oxidation of 8-dibenzoylmesitylenic acid, It is obtained from its solu- tion in glacial acetic acid as a fine, crystalline powder, from alcohol as fine, matted, silky needles, and melts a t 213' : 0.1098 gave 0.2855 CO, and 0.0412 H,O.It is almost insoluble in chloroform or benzene, very sparingly soluble in ether, and sparingly so in glacial acetic acid. It is very soluble in acetone. I n alcohol, it dissolves readily, but crystallises well from the somewhat concentrated solution. The lead, copper, and silver salts of the acid are sparingly soluble and the calcium, barium, zinc, and magnesium salts are soluble in water. Saponification takes place readily in the cold. C = 70.9 ; H = 4.2. C2,H1606 requires C = 71.1 ; H = 4.1 per cent.Dibenxoyltrinzesic Acid, ( C6H, CO),C,H (CO,H),. Dibenzoylmesitylenic acid is dissolved in sodium carbonate and heated with a slight excess of potassium permanganate for 2-3 hours until decolorisation has taken place. The manganese dioxide is filtered off and the filtrate acidified with dilute sulphuric acid. The precipitate, which forms at once, is filtered off, and dibenzoyltrimesic acid then separates out slowly in the course of several hours. It is obtained pure by redissolving in alkali, repeating the above process, and then crystallising from glacial acetic acid. It crystallises from water in rosettes of needles. The crystallised acid contains 1; mols. of water of crystallisation : The pure acid melts at 249-250°. 0*381'7 air-dried acid lost, at 170', 0.0222 H20.The anhydrous acid is a hygroscopic powder. On analysis : 0,1412 gave 0.3410 CO, and 0.047 H20. H,O = 6.2. C,,H,,O,,l~H,O requires H,O = 6.4 per cent. C = 65.9 ; H = 3.7. C,,H,,O, requires C = 66.0 ; H = 3.35 per cent.DEElVATIVES OF DIBENZOYLMESITYLENE. 1323 The acid is very sparingly soluble in chloroform or benzene and somewhat sparingly so in cold glacial acetic acid. Alcohol or ether dissolves it readily. Boiling water takes it up in fair quantity, but deposits the greater part again on cooling. The calcium salt may be obtained as a slowly crystallising syrup by boiling the acid with calcium carbonate until neutral, filtering, and concentrating, A solution of the calcium salt in water remains clear when treated with solutions of barium or magnesium salts, but gives precipitates with those of silver, copper, or lead.Twenty grams of dibenzoylmesitylene are dissolved in 700 C.C. of alcohol and boiled for 4-5 hours with 200 grams of zinc dust, with the occasional addition of a few C.C. of concentrated aqueous potash. The zinc dust is filtered off, the alcoholic solution concen- trated, and the reduced product then precipitated by the addition of water and taken up in ether. The ethereal solution is dried and evaporated and the residue distilled a t a pressure of less than 50 mm. The distillate goes over mainly between 320' and 330'; it is a resin- ous, amber-coloured mass, which has not yet been obtained crystalline. It is very soluble in all organic solvents with the exception of light petroleum. Bi benxy hesitylene, ( CH,),C6H( CH,*C,H,),. Ten grams of dihydroxydibenzylmesitylene were heated with 200 C.C. of hydriodic acid (b. p. 127') and 2 grams of yellow phos- phorus for 3-4 hours in a reflux apparatus. The mixture was diluted and extracted with ether; the ethereal solution, washed free from hydriodic acid and dried with potash, was evaporated and the residue distilled under 20-30 mm. pressure, The hydrocarbon went over almost completely at 280' as a clear, faintly greenish- yellow, viscous liquid, which cry stallised on standing. After repeated recrystallisation from alcohol, it was obtained in the form of beautiful, highly refractive crystals melting at 89". On analysis: 0.1005 gave 0.3382 CO, and 0.0710 H,O. The vapour density was determined by Victor Meyer's air-displace- ment method at the temperature of boiling sulphur. In an atmo- sphere of nitrogen (in which the experiments were carried out), the substance is perfectly stable a t this temperature : C = 91.8 ; H = 7.8. C23H,, requires C = 92.0 ; H = 8-0 per cent.1324 COHEN AND DAKIN: 0~1000 gave 7.45 C.C. at 13' and 756 mm. Density= 159. 0.1050 ,, 7.6 ,, 15 ,, 757 ,, Density = 165. C,,H,, requires density = 150. Dibenzylmesitylene is soluble in all the common organic solvent^, It is best crystallised from alcohol, in including light petroleum, which it is only sparingly soluble in the cold. P a r t of the expense incurred in this work was defrayed by a grant made by the Research Fund Committee of the Chemical Society, for which the authors desire to express their indebtedness. UNIVERSITY CHEMICAL LABORATORY, CAMBRIDGE.