928 ABELL: THE CONDENSATION OF XCVII1.- The Condeizscction of Phenyl Ethyl Ketone and Benxaldeh yde. By ROBERT DUNCOMBE ABELL, B.Sc. (Wales), 1851 Exhibition Scholar of the University College of North Wales, Bangor. ACETOPHENONE condenses with benzaldehyde in three proportions : (i) One mol. of acetophenone with 1 mol. of benzaldehyde, t o form an unsaturated ketone, benzalacetophenone (Claisen and Clapardde, Ber., 1881, 14, 2464 ; Claisen Ber., 1887, 20, 655). c6H,*COoCH3 4- C,$&*CHO = H,O -I- C,H,*CO*CH:CH*C,H,. (ii) Two mols. of acetophenone with 1 mol. of benzaldehyde, to form benzaldiacetophenone or 2-phen yl-l : 3-dibenzoglpropane (Kostanecki and Rossbach, Bey., 1896, 29, 1492). C6H5*CO*CH, CGH5*CO* vH2 O:CH*C,H, = 7H*C,H5 + H,O. C,H5*CO*CH3 C6H5 CO CH, (iii) Three mols.of acetophenone with 2 mols. of benzaldehyde, t o form two stereoisorneric dibenzaltriacetophenones or 2 : 4-diphenpl- 1 : 3 : 5-tribenzoylpentanes (ibid.). C,H,*CO*CH, C,H; ClO*v H, 0: CH*C,H, yH*C,H5 O:C€1~C6H, yH* C6H5 C6H5*CO*CH, = C,K,*CO*yH + 2H20. C6H,*CO*CH, C6H,* CO CH, Ketones of the form R*CO*CH,*R, for example, phenyl ethyl ketone, C6H,*CO*CH2=CH3, do not undergo this third form of condensation. The property of condensation is intimately associated with the group -CO*CH2-, the hydrogen atoms of which are slightly electro- negative in character. The formation of the preceding compounds may be regarded as taking place in the following manner : (i) One rnol. of acetophenone condenses with 1 mol. of benzaldebyde t o form benzalacetophenone. C6H5*CO*CH, + C,H,*CHO - H,O = C,H,*CO*CH:CH*C,H, (ii) One mol.of benzalacetophenone then unites with 1 mol. of a cetophenone to form 2-phenyl-1 : 3-dibenzoylpropane. C,H,-CO*EH C6H,*CO*$!H2 + C6H,~CO*CH, C,H5*C0 *CH, CH*C,H5 = yH*C,HrPHENYL ETHYL KETONE AND BENZALDEHYDE. 929 (iii) One mol. of this 1 : 5-diketone again unites with 1 mol. of benzalacetophenone to form 2 : 4-diphenyl-1 : 3 : 5-tribenzoylpentane. prejent in the acetophenone taking part in these reactions, whereas in the third reaction this group is present in the 1 : 5-diketone. If now phenyl ethyl ketone, C6H,*CO*CH,*CH,, take the place of acetophenone, this substance also contains the group -CO*CH,-, and consequently reactions corresponding to i and ii are to be expected, but not the third reaction, because the 1 : 5-diketone, 2-phenyl-1 : 3-di- methyl-1 : 3-di benzoylpropane, yH*C,H,, which we should expect to be formed in the second reaction, would no longer contain this group.It was therefore expected that benzaldehyde and phenyl ethyl ketone mould condense in two proportions : (i) One mol. of phenyl ethyl ketone with 1 mol. of benzaldehyde to form the unsaturated ketone, benzalpropiophenone, or benzalmethyl- acetophenone : C6H,*CO* yH*CH, C,R5'CO* CH*CH, CGH,*CO*CH2*CH3 -!- C,H,*CHO = H,O + C,H,*CO*R*CH, CH*C,H; (ii) Two mols of phenyl ethyl ketone with 1 mol. of benzaldehyde to C,H,*CO* CH2* CH, C,H,*~O*$!H*CH, C,H,*CO*CH,*CH, C,H,*CO*CH*CH, The following experiments were undertaken with the view of in- Mixtures of phenyl ethyl ketone and form one or more stereoisomeric 1 : 5-diketones, for example : O:CH*C,H, = yH*c6H5 + H,O.vestigating these reactions. benzaldehyde werq treated with the following reagents : Sodium methoxide, 20 per cent. solution. Sodium ethoxide, 20 per cent. solution. Potassium hydroxide in aqueous alcoholic solution. Dry hydrogen chloride.930 ABELL: THE CONDENSATION OF Condensation of Phenyl Ethyl Ketone and Benxcddehyde by means of a 20 per cent. solution of Sodium Methozide. A mixture of 5 grams of benzaldehyde and 6.3 grams of phenyl ethyl ketone was treated with 15 C.C. of a 20 per cent. solution of sodium methoxide. After standing for 7 days, the gelatinous, yellow mass was dissolved by shaking with ether and dilute hydrochloric acid. The ethereal solution was washed with a dilute solution of sodium hydroxide and afterwards with water, and was then dried over fused sodium sulphate.After distilling off the ether, a clear oil remained which was distilled under 25 mm. pressure, until the thermometer rose rapidly to 200'. A fraction boiling at 95-120' consisting of un- changed benzaldehyde and phenyl ethyl ketone was collected separately. As soon as the thermometer registered 200°, the oil distilling over began t o solidify. The distillation was then stopped, the residue in the flask washed out with dry ether, and the ether evaporated off in n current of dry air. The oil which mas thus obtained crystallised almost completely on standing for a few hours in a vacuum over sul- phuric acid. (1) Examination of the CqistaZZine Portion of the Coiodensc6tion Pro- duct.-The crystalline mass was triturated with cold light petroleum which dissolved out an oil leaving behind a mass of fine crystals melt- ing a t 92-96'.This substance, upon repeated recrystallisation from large quantities of boiling light petroleum, to which a few drops of alcohol had been added, separated in needles melting at 98-99'. It is easily soluble in alcohol, benzene, chloroform, acetone, or ether, but only very sparingly so in boiling light petroleum. On analysis : 0.2 gave 0.581 CO, and 0.1345 H,O. C=79*22; H='i*47. 0.2012 ,, 0.5845 CO, ,, 0.1355 H,O. C=79*21 ; H=7*48. C,H,O requires C = 79.33 ; H = 7.44 per cent. Determinations of the molecular weight by the boiling point method 0.2992 dissolved in 29.16 chloroform gave a rise of temperature 0-5192 dissolved in 29.16 chloroform gave a rise of temperature gave the following results : 0.158'. Mol.wt. =238. 0,258'. Mol. wt. =252. Mean =245.-Cl6Hl8O2 has mol. wt. =242. The substance is not acted on by potassium permanganate or brom- ine and is therefore saturated. Further, it cannot contain a ketonic or an aldehydic group because it does not react with phenylhydrazine, hydroxylamine, &c. DiacetyZ Derivative, C,GHl,(O*CO*CH,),.-Two grams of the sub-PHENYL ETHYL KETONE AND BENZALDEHYDE. 931 stance were boiled with 2 grams of fused sodium acetate and 20 grams of acetic anhydride for 4 hours. The oil which first separated on treatment with water rapidly became crystalline and after recrys tal- lisation from alcohol, formed prisms melting at 123-124'.On analysis : 0.2005 gave 0.541. CO, and 0.184 H,O. C = 73.61 ; H = 6.86. 0.202 ,, 0.545 CO, ,, 0.1232 H,O. C = 73.57 ; H = 6.77. C,oH2,04 requires C = 73.62 ; H = 6.74 per cent. The substance, is therefore, the diacetyl derivative of C16HlS02, which is thus proved to contain two hydroxyl groups. Oxidation of Substance C,,H1,02 (m. p. 98-99').-When heated with a mixture of potassium dichromate and sulphuric acid, this sub- stance is oxidised to benzoic and carbonic acids. One gram of the compound was next dissolved in 40 grams of glacial acetic acid and oxidised at the boiling point with 0.6 gram of chromic acid, added in small quantities at a time. The green solution so ob- tained was precipitated with water, and the precipitate extracted with ether.The ethereal solution, after being mashed with sodium hydroxide and then with water, was dried over fused sodium sulphate and the ether evaporated off in a current of dry air. An oil remained which readily crystallised, and after recrystallisation from alcohol was ob- tained in needles melting at 82*5-S4'. On analysis : 0.1717 gave 0.5072 CO, and 0.0922 H,O. C=80*54; H=5.96. 0*2005 ,, 0.5935 CO, ,, 0.108 H,O. C=S0*59 ; H=5*98. C16H1402 requires C = 80.67 ; H = 5.88 per cent. The oxidation product (0.5 gram) being in all probability a ketone o r an aldehyde, was dissolved in alcohol, and mixed with a 50 per cent. solution of hydrazine hydrate (0.3 gram). After standing for 13 hours, the solution was precipitated with water, and the precipitate, after recrystallisation from alcohol, was obtained in glistening plates melting at 222-223'.It was dried a t 120' and analysed : 0,2022 gave 21.7 C.C. nitrogen at 16' and 740 mm. C16H14N2 requires N = 11 -96 per cent. The compound C,,H,,O, (m. p. 825--84O), and the compound CI6Hl4N2 (m. p. 222-2523'), are therefore identical with the methyl- dibenzoylmethane and the 3 : 5-diphenyl-4-methylpyrazole obtained by Beyme (Biss. Leipxig, 1900). From this it follows that the methyl- dibenzoylmethane is obtained from the corresponding 1 : 3-disecondary alcohol, namely, 1 *3-diphenyl-2-methyltrimethylene glycol, for example : N = 12.12. C,H,*FH*OH C6H,*$?0 FH*CH, + 2 0 = YH*CH, + 2H,O. C6H,*CH*OH Cp,H,*CO932 ABELL: THE CONDENSATION OF The methyldibenzoylmethane being a 1 : 3-diketone, reacts with hydrazine to form 3 : 5-diphenyl-4-methylpyrazole, for example : C,H,*YO NH2 CGH5.E- NH FH*CH3 + 1 = 7*CH3 I + 2H,O.C6H5*C0 NH2 CGH,*C=-IIJ The formation of 1 : 3-diphenyl-Z-methyltrimethylene glycol is remarkable, but may be explained on the assumption that the un- saturated benzalpropiophenone is first formed by the condensation of 1 mol. of phenyl ethyl ketone with 1 mol. of benzaldehyde, thus : C6H5*CO*CH,*CH3 + C,H,*CHO = ~6H,*cOwfl*CH3 + H20. CH*C6H5 This unsaturated ketone might then take up a mol. of water to form a saturated ketone-alcohol, the ketonic group being af termards reduced by the sodium methoxide to the corresponding secondary alcoholic group. It is scarcely possible that the reduction of the ketonic group takes place before the elements of water have been taken up by the benzal- propioyhenone to form a saturated compound, because the double bond would in all probability be first attacked by the nascent hydrogen and one or more of the following compounds would be formed : C,H,*CO*CH( CH,) -CH,*C,H,.C6H,*CH(OH)*C(CH3):CH*C6H5. C,H ,*CH(OH)* CH( CH,)*CH, C6H,. none of which has been found in the crystalline part of the product of the original condensation, but may be present in the solution of the oily part in light petroleum. (2) Examination of the Oily Portion of the Condensation Pt*oduct.- The solution of the oil in light petroleum was dried by means of fused sodium sulphate, the solvent distilled off, and the residual oil distilled under reduced pressure. The fraction boiling at 2 10 -2 1 3 O under 23 mm.pressure formed a clear, yellowish oil which furnished the following results upon analysis : 0.2175 gave 0-6742 CO, and 0.1287 H20. These figures do not agree with those required by benzalpropio- phenone, C1,H,,O (C = 86.48 ; H = 6.31 per cent.). The oil, however, has an unsaturated constitution, since it readily absorbs bromine. It is therefore probably benzalpropiophenone con- taining a small quantity of the 1 : 3-disecondary alcohol which boils at about the =me temperature (the pressure being the same in the two cases), or the impurity may consist of small quantities of other pro- C = 84-54 ; H= 6-57, 0.2 ,, 0.621 CO, ,, 0.1197 H20. C-84.68; H=6.65.PHENYL ETHYL KETONE AND BENSALDEHYDE. 933 ducts of the reaction which cannot be separated from the benzalpropio- phenone by distillation.On comparison of the figures with those required by the formuh of the various products which might result from the reaction, it will be seen that they agree most closely with the figures for benzalpropio- phenone. The figures are here given for comparison : C,H,*CO*C(CH,):CH*C,H, requires C = 86.48 ; H = 6.31. c~H,=CH(OH)=C(CR,):CH* C,H, C',H5.C0.CH.(CH,).CH2.C6H~ } 9 9 = 85*73 ; = 7'14* C6H,*CH(OH)*CH(CH3)*CH,*C,H5 ,, C = 84.96 ; H = 7.96. Found (mean) C = 84.61 ; H = 6.61. This conclusion is confirmed by the reaction of benzalpropiophenone with phenyl ethyl ketone described on p. 936. Condensation of P h e n y l EthJ Ketone and BenxuldelLydle 6y means of a 20 pel. cent. solution of Sodium Ethoxide.A mixture of 5 grams of benzaldehyde and 6.2 grams of phenyl ethyl ketone was treated with 15 C.C. of a 20 per cent. solution of sodium ethoxide. After standing for 7 days, the hard, pink-coloured product was shaken up with ether and dilute hydrochloric acid, when a small quantity of a crystalline compound melting at 160-1 63' remained undissolved. This substance, after recry stallisation from alcohol, melted at 162-163'. The ethereal solution was treated in the same way as that obtained from the condensation by means of sodium methoxide and contained : i. Unchanged benzaldehyde and pheayl ethyl ketone. ii. 1 : 3-Diphenyl- 2-methyltrimethylene glycol, C,,H,,O, (m. p. 98-99'), crystallising in needles. iii. A yellowish oil boiling at 810-213O under 23 mm. pressure, This was analysed, with the following results : 0,206 gave 0.6435 CO, and 0.1227 H,O.C = 85.19 ; H = 6.62. 0.206 ,, 0.6415 CO, ,, 0,1242 H,O. C 4 4 . 9 2 ; H= 6.69. C,,H,,O requires C = 86.48 ; I€ = 6.31 per cent. This oil readily absorbs bromine and is consequently slightly impure For further proof of iv. A compound crystallising in small, six-sided crystals melting a t 0.201 benzalpropiophenone as in the preceding case. this conclusion, see p. 935. 163-163', which furnished the following results on analysis : gave 0.621 CO, and 0.1 245 H,O. C = 84.25 ; H = 6.88. 0.2005 ,? 0.619 CO, ,, 0.1235 H,O. C=84.19; H=6*84. C,,H,40, requires C = 84.27 ; H = 6.74 per cent. VOL. LXXIX. 323934 ABELL: THE CONDENSATION OF This substance is obviously 2-phenyl-1 : 3-dimethyl-1 : 3-dibenzoyl- propane, formed by the condensation of 2 mols.of phenyl ethyl ketone and 1 mol. of benzaldehyde with separation of 1 mol. of water (p. 929). To test the coi*rectness of this view, a mixture of 5 grams of benz- aldehyde (1 mol.) and 12.6 grams of phenyl ethyl ketone (2 mols.) was treated with 15 C.C. of a 20 per cent. solution of sodium ethoxide, with the result that the compound C,,H,,O, (m. p. 163-163') was obtained in slightly greater quantity. The preparation of pure benzalpropiophenone was next attempted, potassium hydroxide or hydrogen chloride gas being used as the condensing agent. Condensation, of Pheql EtAyZ Ketone and Benddehyde hy means of Potassium Hydroxide. A mixture of 12.6 grams of phenyl ethyl ketone and 10 grams of benzaldehyde was dissolved in 100 C.C.of alcohol, and 3-4 C.C. of a concentrated aqueous solution of potassium hydroxide were added ; the whole was diluted with water, but not in such quantity as to precipitate the ingredients, and the mixture allowed to stand for 7 days. The aqueous alcoholic solution was precipitated with water and extracted with ether. The ethereal solution, dried by means of fused sodium sulphate, was evaporated, and the remaining oil distilled under 22 mm. pressure and collected in the following fractions : i. Benzaldehyde, b. p. 85-88' (185-190' under atmospheric pressure). ii. Phenyl ethyl ketone, b. p. 103-!08' (210-215O atmospheric pressure). iii. A yellowish unsaturated oil, b. p. 200-220'. The last portion was redistilled a number of times, and the portion boiling at 210-213' under 23 mm.pressure analysed : 0.2175 gave 0.685 GO, and 0,1227 H,O. Benzalpropiophenone can therefore be prepared pure by this method, A small quantity of a yellowish oil separated. C = 86-35 ; H = 6-27, C,,H,,O requires C = 86.48 ; H = 6.31 per cent. but the yield (10 per cent. of the theoretical) is small, Condensation of Phertyl Ethyl IrTtone and Benxaldeiyde by means of lfydrogen Chloride. When n mixture of inoleculnr quantities of phony1 ethyl ketone and benzaldehyde is treated with dry hydrogen chloride until 1 mol. has been absorbed, a dark oil is obtained which readily splits off hydrogen chloride, and from which no definite chlorine compound could bePHENYL ETHYL KETONE AND BENZALDEHYDE. 935 isolated. This oil after being kept for 12 hours in a closed vessel was gently heated on a sand-bath in a current of air until it no longer gave off hydrogen chloride, and was then distilled under reduced pressure.The fraction boiling at 234-235' under 25 mm. pressure, which still contained chlorine, was analysed with the following results : 0.2345 gave 0.69S5 CO, and 0.1275 H,O. 0.232 ,, 0.7267 CO, ,, 0.13 H,O. C=85*43; H=6.22. 0,4632 ,, 0.0239 AgCI. C1=1*27. 0.4947 ,, 0.02714 AgCI. Cl= 1.35. C = S 4 . S ; H= 6-31. CI6HL50Cl requires C = 74.29 ; H = 5.80 ; C1= 13-73 per cent. C,,H,,O ,, C = 86.48 ; H= 6.31 per cent. From these figures it is evident that about nine-tenths of the chlorine has been eliminated by heating, one-tenth still remaining after a number of distillations.It is therefore probable that in the reaction two isomeric compounds are formed, and that one of these is much less stable than the other. The remaining chlorine can be removed as hydrochloric acid by boiling the substance with an aqueous 50 per cent. solution of potassium hydroxide. Benzazpropiophenone so obtained is straw-yellow and distils without decomposition a t 210-213' under 23 mm. pressure ; the yield amounts to '70-80 per cent. of t,he theoretical. It was analysed with the following results : 0,2122 gave 0.6732 CO, and 0.122 H,O. C=86.50; H=6*38. 0.209 ,, 0.6625 CO, ,, 0.119 H,O. C = S6.45 ; H = 6.32. C,,H,,O requires C = 86.48 ; H = 6.31 per cent. On analysis : C = 84.49 ; H = 6.53. The phenylhydrazone, crgstallised from alcohol in yellow needles softening at 115' and melting at 127-128'.0*2012 gave 0.6235 CO, and 0.1 IS5 H,O. 0,2005 ,, 16 C.C. nitrogen a t 20' and 752 mm. N = 9.03, C,,HZoN2 requires C = 84.61 ; H = 6.41 ; N = S.97 per cent. Benzalpropiophenone dibromide, C6H,*CO*CBr(CH,)*CHBr*c6H~. -Ten grams of benzalpropiophenone mere dissolved in SO grams of chloroform, and 7.3 grams of bromine in 20-30 grams of chloroform gradually added t o the solution, , The bromine was rapidly decolorised with a slight development of heat, and only a small quantity of hydrogen bromide was evolved, The product was a t once poured into an evaporating dish, the chloroform evaporated in a current of dry air, and the thick viscid oil placed over sulphuric acid and potassium hydroxide in a vacuum until the weight was constant.The green viscous oil thus obtained did not crystallise even after long standing, or when cooled with a mixture of solid carbonic acid and ether. On analysis : 3 s 2936 ABELL: THE CONDENSATION OF 0.232 gave 0.438 CO, and 0.0797 H,O. C = 50.31 ; H = 3.81. 0.337 ,, 0.3205 AgBr. Br=41-'75. 0.2792 ,, 0.273 AgBr. Br=41-60. C1,H,,OBr, requires C = 50.26 ; H = 3-66 ; Br = 41.89 per cent. I n the various preparations of the dibromide, the amount of bromine 0.2515 ,, 0.4642 CO, ,, 0,0865 H,O. C=50'34; H=3.82. present was found to vary between 38 and 43 per cent. Condensation of Benxalpopiophenone with Phenyl Ethyl Ketone. It seemed probable that for the satisfactory preparation of Z-phenyl- 1 : 3-dimethyl-1 : 3-dibenzoylpropane, the previous formation of the unsaturated benzalpropiophenone is necessary.The following experi- ments were therefore made. A mixture of 4.4 grams of benzalpropiophenone and 2-7 grams of phenyl ethyl ketone was therefore treated with a solution of 0-5 gram of sodium in the least possible quantity of absolute alcohol. The mixture rapidly became pink in colour, with development of heat; crystals soon began to form, and in 48 hours the whole had solidified to a compact, crystalline mass. This was triturated with dilute alcohol, filtered, and washed with alcohol, when a nearly white, crystal- line mass melting at 15S-16Oo was obtained, which after recrystal- lisation from alcohol melted at 162-163'. It was almost insoluble in hot light petroleum, and only sparingly soluble in boiling alcohol and in ether. The yield of 2-phenyl-1 : 3-dimethyl-1 : 3-dibenzoylpropane was almost quantitative. Erom the alcoholic mother liquors, a small quantity of a compound much more soluble in alcohol was obtained.This crystallised in needles and melted a t 121-122'. On analysis : 0*3005 gave 0,6165 GO, and 0.1225 H,O. C = 83.85 ; H = 6.78. 0.2015 ,? 0.6202 CO, ,, 0.123 H,O. C=S3.96 ; H=6*78. C,,H,,O, requires C = 84.27 ; H = 6-74 per cent. As the analyses show, this compound is isomeric with 2-phenyl-1 : 3- dimethyl-1 : 3-dibenzoylpropane (m. p. 162-1 63O). The yield amounts to 1 gram from reactions yielding upwards of 60 grams of the isomeride. The mode of formation of these compounds, more especially of that of the higher melting point, by the above method indicates not only that they are isomeric 1 : 5-diketones, but is also a further proof of the unsaturated constitution of benzalpropiophenone.Similar reactions of unsaturated compounds with substances which contain electro-negative hydrogen atoms have been carried out by Michael (J. p. Chem., 1887, [ii], 35, 349), who studied, amongst otherPBENYL ETHYL KETONE AND BENZALDEHYDE. 937 reactions, the behaviour of ethyl cinnamate with the sodium deriva- tives of ethyl malonate and ethyl acetoacetate. I n the above case also, the union of the unsaturated ketone benzal- propiophenone with phenyl ethyl ketone follows in the sense of Michael's " positive-negative " rule (J. yr. Chern., 18SS, [ii], 37, 522), according to which '' in the union of a sodium compound with an un- saturated compound, the sodium atom, replacing a hydrogen atom of the acid methylene (= CH,) group, links itself to the carbon atom (of the unsaturated compound) which has the strongest negative character," in this case to the carbon atom attached to benxoyl on the one side and to methyl on the other in the radicle, C,H,*CO*S.CH,.Consequently, the reaction may be represented by the following equation : C,H,* C0.E : CH, U,H,.CO*yNa*CH, CH*C,H, - - yH*C,H,. + CGH5-CO*CHNa*CH, CGH,*CO*CH*CH, The sodium compound so formed is then probably decomposed by the alcohol with regeneration of sodium ethoxide. I n carrying out this reaction, it is necessary to use an amount of sodium in agreement with the above equation in order to obtain the best results, and under these conditions the yield of 1 : 5-diketone is almost quantitative.Again, in this reaction it was found that methyl alcohol cannot take the place of ethyl alcohol. This explains how it is that the 1 :5-diketone is formed when phenyl ethyl ketone and benzaldehyde are condensed by means of sodium ethoxide, and not when sodium methoxide is used. As this reaction is practically quantitative, i t can be used t o ascer- tain whether the oils boiling a t 210-213' under 23 mm. pressure, obtained by the condensation of phenyl ethyl ketone and benzaldehyde by means of sodium methoxide and sodium ethoxide respectively, con- sist of benzalpropiophenone, as previously concluded. 2.2 grams of each of these oils were mixed with 1.3 grams of phenyl ethyl ketone and treated with 0.23 gram of sodium dissolved in abso- lute alcohol. The yield of 1 : 5-diketone was 3.3 grams, instead of 3.5 grams, that is 94 per cent.of the calculated amount in both cases, thus furnishing conclusive evidence that these oils are almost pure benzalpropiophenone. I n order to investigate further the structure of the isomerides, C25H240,, the reaction with ammonia and with hydroxylamine was studied in the hope of obtaining one and the same pyridine derivative by means of the following reaction : CH,*y H* CO*CGH, C H,* T-=q'C,H, CH,*CH* CO*C,H, C H 3 * C E C .C,H5 CGH5.7 tI + NH, = 2H,O + C,H,*FH TH938 CONDENSATION OF PHENYL ETHYL KETONE. Action of Awmonia o n P~eiz?lk~imet?~?/ZdibePl.no~~~oz)ccne, m. p. 162-1 63'. -One gram of the compound (m. p. 162-163') and 5 C.C.of a saturated alcoholic solution of ammonia were heated in a sealed tube to 200' for 8 hours. Upon opening the tube, in which there was scarcely any pressure, a n oil remained. The contents of the tube were dissolved in alcohol and allowed t o evaporate in a vacuum over sulphuric acid. A small quan- tity of a pink compound melting a t 155-156O was obtained, along with a large quantity of resinous matter. After numerous recrystal- lisations from alcohol, the former was obtained in colourless needles melting a t 155-156', and was analysed, with the following results: 0.2149 gave 8 C.C. nitrogen a t 15' and 759 mm. N=4*35, C,SH,3N requires N=4.16 per cent. C,,H,,N ,, N=4.18 ,, From the nitrogen found, this compound may be triphenyldimethyl- dihydropyridine, C,,H,,N, or triphenyldimethylpyridine, C,,H,,N, Considering the large amount of resin formed, it is possible that the same changes have gone on as were observed by Michael (Ber., 1885, 18, 2021) in the formation of ay-lutidine-P-carboxylic ester from aceto- acetic ester, acetaldehyde, and aldehydeammonia.A similar case was observed by Knoevenagel and Weissgerber (Ber., 1893, 26, 436) in the formation of pentaphenylpyridine from benzamarone and ammonia. There is, therefore, not merely a separation of water to form a hydro- pyridine derivative, but possibly a simultaneous oxidation of the hydro- pyridine ring with formation of 2 : 4 : 6-triphenyl-3 : 5-dimethylpyridine. This assumption is justified by the fact that the same compound is obtained by the action of hydroxylamine hydrochloride.Action of Hydvox?/Zamine Hydrochloride on P~enyldimethykdienxoyl- propane (m.p. 162-1 63°).-Twogramsof the substance (m. p. 162-163"), 1 gram of hydroxylamine hydrochloride, and 40 C.C. of 90 per cent. alcohol were heated in a sealed tube at 120-130' for 6 hours. The contents of the tube were precipitated with water and extracted with ether. The ethereal solution left, on evaporation, a pink, crystalline mass, which, aher many recrystallisations from alcohol, separated in colourless needles melting at 155-156'. The yield was 70 per cent. of the theoretical. On analysis : 0.201'7 gave 0,6622 CO, and 0.11s H,O. 0.223 ,, 8.6 C.C. nitrogen at 13' and 744 mm. N = 4.52. C=89*52; H=6.49, Ot2012 ,, 0.66 CO, ,, 0.116 H,O. C=89*44; H=6*40. C&H2,N requires C = 89.02 ; H = 6*S2 ; N = 4-16 per cent.C,,H,,N ,, C = 89-55 ; H = 6.2'7 ; N =4*1S ,, From the analyses, it is obvious that the compound has the formula C,,H,,N, hence hydroxylamine reacts smoothly with formation of 2 : 4 : 6-triphenyl-3 : 5-dimethylpyridine, thus :SODEAU: THE DECOMPOSITION OF CHLORATES. PART 1V. 93'3 CH,* $!H* CO* C6H5 CH,*Y=Y*C,H, CH,* CH*CO*C,H, CH,*C-C*CC,H, Action, of H y d r o q l a n h ~ Hydrochtode on Pheizylclinteth~ Zdibenxoyl- propane (m. p. 121-122°).-The isomeride (m. p. 121-122") was not treated with ammonia on account; of the very limited quantity obtain- able, 0.4 gram of the substance, 0.5 gram of hydroxylamine hydrochloride, and 10 C.C. of 90 per cent, alcohol were heated in a sealed tube a t 120-130' for 4 hours, and the product precipitated with water and extracted with ether, Upon evaporation, the ethereal solution de- posited a residue of pink crystals, which, af ter many recrystallisations, were obtained as colourless needles melting a t 155-156'. On analysis: 0.18 gave 6.8 C.C. nitrogen at 15' and 752 mm. C,5H,,N requires N = 4.18 per cent. This compound is identical in appearance and properties with that obtained from the isomeride melting a t 162-163' by means of am- monia and of hydroxylamine. The two substances have, therefore, the same structure, and are 1 : 5-diketones of the formula CGH,*yH 4- OH*NH, = 3IJ,O + C,H,*E 8 N= 4.37. CH,* TH*CO*C,H, C],H,*F]H 3 CH,*CH-CO*C6H, It is hoped that the substances mentioned in this paper mill form but their isomerism is to be explained upon stereochemical grounds. the subject of a future communication. The author wishes to express his indebtedness to Prof. J. Wislicenus for suggesting the above investigation, and for the privilege of being nllomed to pursue it in his laboratory. EI:STES CIIEMISCHES LAEOI~ATORIUM, UNIVEKSITAT, LEIPZIG.