1140 FERRISS AND TCTRNER: CXXIX-Studies in Ring Formation. Part IIL The Condensation of Aromatie Amines with a- and &Diketones and with 4 &-Diacetyldiphenyl. By CLARENCE VICTOR FERRISS and EUSTACE EBENE~ZER TURNER. The Comle?zmtion of Bemzidime a d Toli&na with BerzziJ a d CONDENSATIO: products off loenzidinel and tdidine with belnzil or glyolxal were deeoribed by Cain and Mioklelthwait (T, 1914 105, 1437) but the prelselnt a-utlhors after clarefully repeating their eixperiment8 are unable1 to confirm the formulz assigned by them. to the protducts obtained from benzidiim aiid these1 two diketoncs. Whilst the oolndsnsation product from b e n d and belnzidina has now been obtained in a statel of purity its cornplelte analysis and a determinatlion of its molleloular weight by the oryoscopic; method have not leid to1 any definite colncllusions as to1 its stmature.Con-delnsation seelms to1 take place1 in a more aomplelx mannelr than is assmeld by bhe above autholrs. Hydrollysis of the1 profduck with rninelral acid showed tha,t more mollelculee of benzil t<han of benzidinel elnter intlol the reladioin. Again although such a con-deineatioln might be elxpelcted to1 prolceeld quantlitativdy we1 have ;not sucweded in obtaining more than 70 pelr wnt. of the aaloulated G l y o ~ x d STUDIES IN RING FORMATION. PART III. 1141 yield nolr is an appreoia,brle quantitly o:f aaeltylb8enzidine formeid as might be supposed. Of t,hel prolducta ob,taineId from glyoxal t,hat formed with bemidinel is a highly insoilubls norn-crystlallina mass of unknown rno1e:cula)r comple,xity.The oo8ndensaf8ion olf glyolxad aad tolidine ga,vei a substqa4nce of tlhel e,mpiricad folrmula folund bly Cain and Miaklet8hwait b'ut its insolubility relnders a mollelcular-welight det8elrmination impossib3ei so tlha,t4 itl ma,y possibly be f ormeld by the colndelnsation not olf ojiiel molelmle oC glyolxad with one of t401idinel, b,ut by tlhe condelnsa.tion od two! molleiculelsi od each together. The Codenmtion of A mines and Dimnines with Ac1e'tylaCe:tome a,& Benzoy lac e t on,e. Condensation prolductsi o F acetyla.celt onel a,nd benzoylamt~olns with bemidinel and othelr b,a,sea ha,vve previoasly been dmcrib,ed by one of us (T. 1915 107 1495; 1917 1111 1). Further examples have now b'ee'n studieid.h o ' n g olthelr reeults it hae been found thab monolaicet?/liso~rrol;olyl~e,netoil~~nei is a liquid ah the ordinary teimpeira~turei and dz;acetyli.,o~olrylzdenetoli~~e a so;lid melting a.t 1 0 8 O (uncolrr.) whe'ress t'hhe c'olrresponding colmpolunds froem beazidine mellt res8pect8ivetly a.t 137O and 198O' (cornparre benzylide'ne-o-t.oluidine1 a liquid a'nd benzylidelnelaniline m. p. 48-49O). Combes (BuLl. Soc. c h k . 1888 [ii] 49 89) wha first studied tIhe a,ctioln of aminea on a&,ylacetone stla,teld thati the prolduclt of aondensa,tion of p4olluidinel a.nd a,cetyla,ce~t~oae melted ah 39-40°, a8pparelntlly reterring t*o the trimelthylquinoiline f olmed by its furthelr colndelnsa.tioln in the preeelnm of sulphuria soid. Pfitzinger ( J . pr. Chem. 1888 [ii] 38 40) givea t'he melting polint of the trimeithylquinolinel in question as 63-64'.The present aut,hoas ha,ve found t,ha,t pLttoluidinel and a,mkIylamt~one colndense undelr the influeaae of helast aJone t,ol give1 tsh el tkriinetchylquinoline the melting polintl o€ which is 64O a fa.ct which renders p-toluidine an ab~no~rma81 compound as regards its aation on amtylacet80ae. Ketones of the DZphenyl Series. Attempts to1 Condense 4 4f-Dicnce t y ldiph eny 1 with o-Ph eny 1 ertediamhe. Hitherto our knowledge of ketones derived from diphenyl has bwn r estriche d t o p hen y 1 a celt op hen one 4 -p hem y lbenzo phenoas, 4 4/-dibeazoyldiphenyl diphenylyl benzyl ketone and diphenylyl a-naphthyl ketone. Satisfsctolry proof olf the constitution of these keitlonee has only beeln giveln in the uasel of 4 4/-dibenzoyldiphenyl.This substance originally obtained by Wolf (Bey. 1881 14 2031 1142 FERRISS AND TURNER: by the adion of benzolyl chloride on diphenyl in the presence of aluminium chloride was pretpa.reld by Ullma,nn (Annulen' 1904, 332 79) from 4-iodobe8nzo~phelnolne~ and colpper brolnm and thus shown t,o be1 the 4 4/-delriva,tive. Froam thip it follows thatl Wolf's phelnylb,enzophenone is a 4-deriva.tive . Va.rious e'rrolrs harm been noltimld in the1 liteIra,tnrel t>racreable t,o the work of Adam (Anln. Chim. Phys. 1886 [vi] 15 224). This aut<holr as a relsult od a' more or less sjstematic study od the Friedel and Crafts readioln in t'he diphenyl serie,s concludeld t'hat in all suoh relactiom saw olnel 3-suh~stlitlutied diphelnyls were prolduced a aonolusion based on the1 fa,ct that these' compounds ga8vei on olxida8-tion a,n acid mellting at 160-161° which adam t,oolk t,o be diphen y l- S-ca,rb,oxylic acid .His ,ofne' e,xcelp tio,n diphetny lb elnzOc phenolne obt8a.ineld from oarb,onyl chlolride a,nd diphenyl proved to be identica,l witlh the prolduot obta,insd by Weilelr (Ber. 1873 6, 1181) from methylal and diphelnyl and gave on oxidation an acid (diphe'nyl-3-aarboxylic) me~lting a8t 21 7-21 8O. Adam's phenyla,cdolphenoae was state,d by Volrlandes (Bey., 1907 40 4535) to! be the 4-delrivamtJve~ since deriveid oompoands formeid a,nisotropio liquids a prolpesty which would not be p o e s e 8 d by 3-derivat'ives. The aation 09 a.ceityl cIiloiride1 on diphenyl in the prese'nce of aluminium chloride has now beleln sho,-wn t,o relsult in t,he formatioln olf 4-pheinylaceitophenoae (m.p. 121') and 4 4/-dia:cetyldiphenyl (m. p. 19O-19lo). The fomelr (idelntica,l with Adam's phelnyl-acetophenonel) gives oa oxidaDion diphenyl-4- and nolt. 8,s st.atod by Adam diphenyl-3-carbolxylio acid. The constitlutlio8n of these ketlonee has been set beyoad doubt b,y tlheir synthesis respectively, from 4-oyanob a.nd 4 4'-dicya.no~-diphe~nyl b,y melam of magne8sium methyl iodide. Since Adam's phenylam4tophe8nonel is the 4-de.riva8tive and since his supposed S-et,hyldiphe,nyl ga,vel t,his keltonec o,n partial oixidation, this elthyl deirivat'ive must be 4-ethyldiphenyl. It is evidentl, simila,rly tha.t his supposed 3 -methyldipheayl is actua'lly 4-methyl-diphenyl .A.gain the diphenylyl henzyl ketone ob,ta.ined by Papeke! (Ber., 188'8 2 I 1339) by t,hs ac,tion olf phe8iiylamtyl chlolrids on diphenyl in the presence of a,luminium chlolride has now belea sho'wn definitely to be the 4-derivaft.ive by it5 synthesis from 4-cyano-diphenyl a,nd magnesium benzyl chloride. I n additio'n 4 4'-d~~~he,i~yZdialcetyld~ph.e~~yZ which is not pro-duaed tot a,ny e'xtent in the Friede'l a,nd Craftls rela,ct,ion me8ntiioned above has b,e.eln prepamd from 4 4/-dicya.noldiphe1nyl. It is evideint from tb abwe tlha.t in all lrnown Friedel an STUDIES IN RIKQ FORMATION. PART 111. 1143 Crafts' reactions betIwee.n diphenyl and halogen compolunds 4-deriv-at'ives are obtlained a' fa,ot which justifiels Sohmidlin and Gascia-Baniis ( B e y .1912 45 3183) in their assumption als to t'he constit,ution of the diphmylyl a-naphthyl ketone prepamred bsy bhem. This inveatiigation into tihe ket,ones of t,he diphenyl seriels was made a,s t,he rwult of d,esiring t*ol determine the1 spa,tial reilation of the tlwo a4crehy1 groaps in 4 4'-dia'aeltyldiphe.nyl ; this. ke,tone might be expelcteld i f t,he Kanfler formula for derivat'ivea of diphenyl is valid t,o condense with 0,-phsnylenediamine to give1 a compound of the mnstitutlion (I). When the two substlances in question a m warmed together in acetic acid solution condensation occurs and a yelllow sollid is predpita*ted from the mixture on a,dding methyl or ethyl alcohol. As in the case of the condeasa-tion produot of benzil and bnzidine howevelr it has not been found pmibri to ascribe a definite consttitlutioln to the product.ombtained dt'hough the elvideince is prob'ably in favour of a corn-p o z d farmed by the condensatioln af two moleculee of keltone with one of diarmine. E x P E R I M E N T A L. GoYndensatim of B e n d with Bemidine. This condensation has been carried out as described by Cain and Micklethwait (loc. cit.) 1-84 grams of bnzidine 2.1 grams of benzil and 10 C.O. of glacial acietio acid being used. After twenty minutels' heating of this 1nihn-s on the boiling-water bath itl was cooleld and scratched when the canary-yedlow precipitate separated, further warming then causing further separation of precipitate. The yield was found to be increased by perioldicl cooling and scratch-ing.The precipitate was oodlscted washed with a little glaaial acettio acid and then with dilute amtio acid and dried in a vaouum over potassium hydroxide. The greenish-yellow solid so obtained (equivalent in quantity t o a 70 per cent. conversion to the suppoeed oompound) crystalliaed readily from benzelne but a lees green product was obtained using xylene as solvent the crystals being then washed with light petroleum [Found @=85.1; H=5.5. C=85-2; H=5*7. N-5-3 5.3 pelr cent. M.W. (by oryoscopic method in bromoform) =496] 1144 FERRISS AND TDRNER: The su6stamce is readily solluble in hob xylene benzene or pyridina and almost insoluble in add xylene benzene alcohol or glaaoid acetia add. An at'tempt ta obtain Cain a,nd Micklehhwait's ethyl alcohol additive m p a u n d was rna.de as folllows the co,ndensa.tion product was collected dissolved in hnzonei the solution evaporated to dry-ness after the additioa of a little aloolhol and the residue again t'reated with alcohol and the httelr eva,pora.teld.Evaporahion of the hnzeae solution obtlaineid as abwe withoutl a,dding alcoholl, gave a well-defineld product identioal wit.h that ob,t'ained in t'he presence od alclo~holl (Found C = 85.0 ; H = 5.5 ; N = 5.7 per cent.). The substmca meJteld ah 239-240° a mixtare with tIha produd from xylelne (above) melt'ing a t the sa,me tempelrattare. Simila'rly an a,ttempt was made to prepare Cain and Miokle thwait'e melt,hyl alcohol sddit'ive compound f o4owing t'helir deacribeid method m closely as possible.The product could not be melted undelr methyl adcoholl even 0 1 1 1 prolonged bloiling a,nd when dried melted at 231-234O (Found C=80*0; H = 4 * 2 ; N =5.8 The produot was a>gain trea;teid with melthyl alcohol (far the fourth time,) and the whale empoiratad (Found C = 80.4 ; H = 3.9 per cent.; m. p. 231-234O). It was then crystallised from xylene (Found C = 8'1.3 ; H.= 4.9 per cent.; m. p. 233-236O) a selaond clrystallisa.tTioln frolm xylene giving a product mdt,ing at 234-237O (Found C=83.7; H=5.0 per cent.). The preselnt autlhors a,ttribate the law melting paint etc. of tho first s f thew products to the fa,& that the green impuritiee were still presentl and the subselquent re'sults to the rmova.1 of tthme impurities. I n olrder if pmsib'le t'ol throlw mom light o a the aoadensa,t.ion just deeoribed the melting points od mixtures of b e n d and h z i d i n e have been deltRlrmine8d.The benzil used was purified by recrystldlisation frolm alcohoil and meltsd a4t 94.4O. The benzidine was purified by recrystlallisation successively f r m watelr and dilute aloohol and then several timels frolm alaohol. Pure belnzidine olbttaineld in this wa,y mdts a8t 127.4O. All mixtures oif bnzil a,nd benzidine melted below the1 lafter t,emperature the ourve of mixed metlting polints b,eing part of a,n invelrteld pasaboh the axis of which ran vertically froim the1 point. represent8ing two molecules of belnzil t,ol one olf bennidine. No1 e8videtnw folr t,he formahion of a compound was forthaoming. per ceint.) STUDIES IN RING FORMATION.PART 111. 1145 Condenmtiom of Benzdhe with Glyojxd. On effeating the coadensa,tion as described by Cain and Mickls thwait a dark broiwn amorphous precipit'ate was obtlained. On drying it gave a hard mass a a examina,tioa of whiclh gave no relsults worth reloolrding. Codematiron. of Totlidine with GlyoxaL. This mndelnsattlion gave the resdts published by Cain and Mickle-thwait (Folund C = 82.2 ; N = 5.9 ; N = 12.2. Calc. for CI6Hl4N2 : C=82*0; H=6.0; N=12.0 per cent.). It was not found possible to oarry out a deitelrmina8tioln of rnoledar weight so that the aolnstitation of the substanw camoh be deduced. Codematilorn of A cetylacetome with p-Toltuidine. A mixture of p-tohidinel and aoetylamtone in moleimlar propor-tions was gently boileld for a,n hour.On cooling the whole set solid a mall polrtion after being left in contact with porous porcelain melting atl 64O and a t the same temperature after crystallisation frolm light peltroleurn. Almost the theoretiodly potssiblei yield of 2 4 6-trimethylquinodine was thus obtained (Folund N=8.2. Calc. N=8-2 per mntl.). M o s f L o o c e t y l i s o ~ o ~ ~ l i d e ~ e t o l l ~ ~ ~ ~ a d its p-Nitrolb enzylidene Deri;vative N0,*c6R,*CH:NgC6H,Me~*C6H3Me*N:CMe*CH2Aa. A boiiling solutioln od 2.1 grams of tlo;lidine in 15 C.U. of xylene was slowly treated with 1 gram od amtylambne dissolved in 5 C.C. olf xyleinel heating being continued folr a further hour. On dis-tilling off the xylene a dark viscid liquid was obtained which could not be aanwd t o orystallise.In order tlo confirm its con-stitution it was convelrtsd into the pnitrobenzylidene derivative by boiling it in alcohollic solutioin with a slight excess od p-nitro-benzaldehyds. The brick-reld preoipitate obtained in this way was clrystallised frolm a mixture of pyridine and alcolhol (Found : N = 10.5. The high pelrmlnt.age of nitroigeln foand was due to the preeenw of a m a l l quantity od di(pnitrobelnzy1idenel)-tolidine which when wolrking with. barelly sufficient material cannoit readily be relmoved. N - p - N i t r o i b e n z y l ~ ~ e u n e m o i n o a c e t y l i s o l ( p r o ~ y l ~ ~ n ~ t ~ l ~ ~ e melts a t 245-24707 is readily soluble in hot. pyridine and sparingly so in cold adcolhol or acetione. C,H,,03N3 relquires N = 9.8 per cent.) 1146 FERRISS AND 'TURNER: Dia ce t y 1 isolpro~pylideme t o~lidke ( C6H3Mel* N CMel* CH Ac) .A sodution olf tolidine (8 grams) in 15 grams od aceltylaoeltone was heated a t the boliling point folr three hours tihe excms of diketone distJlled off and the1 yellow crystalline! residue obt aiiied on cooling crystalliseld from a mixture of xylene and light p&roletlm from which i t sepa,rateld in yebw needles meltling a t 108O (unoorr.) (Found N = 7.7. C,HB0,N2 requires N = 7.4 per cent;.). 2 :4 8 21 41 81-Hexamethyl-6 61-d~pinolyE. The prelwding oolmpound was heated with ten times its weight of concentrated sulphuric asid a t 125O for tlwo holurs the cooled solutioln poureld int'ol water the1 result ing sollution relndelreld alkaline, and the white preloipitatel collledeld and orystadlised from :rylene (Found N = 8.4.The base crystallises from xylene in minute cubes melting at 252*5" and is almolst insoluble in eithelr hot ocr colld alcolhholl carbon tetir achlolrid el or toluene . The platzhichlosdde was obtained in the usual ma,nner and folrms brownish-yellow needlee (Found Pt = 26.2. C,H2,N,,~I,PtCl, requirels Pt = 26.14 per cent'.). C24H24N2 require8 N = 8.2 pefr centl.). Molno b enzoylisolp-o pyldene 6 enzidine, NH,*C,H,*C,H,*N:CMe,*CH,Bz. A sdutioln of 3'24 grams of beazoyla8mtonei in 15 C.C. of xylene was added gra,dually to a bo3ling solut'ion of 3.68 grams' od benzidine in 15 0.0. olf xylelne the heading continueld f o r a further hour the sollventq remotveld by distillation and the yelllolw residue orystallised from a mixture of pyridine and alaohol when it sepa<rated in slender pale yellolw neleldles me'lting at! 179O (Foand : N = 8.7.C,,H,,ON requires N = 8.5 per cent.). A ce t y Zisopropylideme b e m o yliso F o p y lideneb eitzidine, CH,Ac* CMe:N* C6H,*C6H,*N CXZe8* CH~BZ. A boiling solutioln of the1 preceding compolund (0.761 gram) in 10 C.C. otf xylene was sloiwly treiakted with a sollution of 0.232 gram of amtylamltoae in 5 C.O. of xylene and the reacrtioln allowed to complete itself during a further sholrt boiling. On removing the xylene by distillattion a yelloivrr regidue relmainsd which was arystlallised frotm EL mixture of pyridine and alcohol (Found: N=7.0. C,,H,O,N requires N=6.8 per wt.) STUDIES I N RING FORMATION.PART 111. 1147 The compound fosms yellow le'aflelte melting a t 234-236O and is sparingly soiluble in alcohol relaclily so in cold chlolrolform and dissollves in hot pyridine. Mocnolb e n z o y l i s o ~ r o ~ ~ l i d e n e t olidine, NI-I,.C,E43Mel*C,H3Mel*N:CMe*CH~~z. This substaiioel was olbtained in a manne'r similar tol tthat described folr the1 corresponding benzidine compolund and crystal-lises from a mixture of pyxidinel and alco4holl in pale yellow needles meltling at 170O. It acquires a grelen colour in nio$istl air owing to partial oxidation (Found N = 8.0. C,4H2,0N2 requiree N = 7.9 pelr cent.). Com2ensatio.n of A cetyl Chlol.i.cFe with Diphenyt. (a,) Pmpraltion o f 4-Phenylacetophenome.-A~~tyl chloride (8 grams) was mixed with 100 O.C.of carbon disulphide and al hrge excess of diphenyl (35 grams) and 20 grams of anhydrous aluminium chloride were slowly added. The reaction was carried to1 mmpleltioa by helating and the coded mixture decolmpmeld with ice and dilute hydrochlorio acid. The carbon disulphide was removeld in a current of steam and the rwidue cooled filtered, and dried. The whitel solid sol obtained was extractejd with boil-ing light petroleum until a sample of the residue oln drying mellted a t 115-120O. The residue oonsisting of almostl pure phenylacetoc pheiione was crystallised from aloohol and melteld a t 120-12lo (uncorr.). It possessed the properties asaribejd to it by Adam (Zoc. cit.) (Found C=85-2; H=6*2. Calo. C=85-7; H=6.1 per cent.). (13) Preparaition of Phenylaicetophenone and 4 4I-DiucetyZ-dipkenyl .-Anhydroas aluminium chloride1 (26.7 grams) was covelred with a mixture of 100 C.C.of carbon disulphide and 15-4 grams of diphenyl. Awtyl chloride (15.7 grams) was then slowly added and the evolution of hydrolgen chloride allowed to become complete by subsequent heating in warm water. The carbon disulphide was evaporated the mixture decomposed as bef orel and the resulting white! solid collelcteld and drield. Elaborate1 fractional crystallisation from alcohol gavel pheiiylacetophenmel (molderately soluble) and 4 4/-diucetyZdiphenyI (sparingly soluble). The latter was purifield by crystallisation successively from dilute aeetio aoid, alcohol and oarbon tetrachloride ; itl forms almost colourless leiaflets melting at 190-1910 (uncolrr.) (Found C=80*5; H=5*9.C,6Hl,0 requires C = 80.7 ; H = 5.9 per wnt.) 1148 FERRISS AND TURNER: P r e p r a t i m of 4-Cyanodiph en yl, The preparation of this substanae from diazotised 4-amine diphenyl and pot aasium cmproayanide proved unsatisf a,ctory a large quantity of a polymeride being formed. It is intelreating t o note thatl Kaiselr (Anmlem 1890 257 100) obtained a moldelratdy good yield od diphenyl-5-carbolxylic acid by the hydrollysis of the toltal crude p r d u d . The method dacTibeld by Rassow (Annalen, 1894 282 143) for the preparatJoln of the nitrils was finally adopted; it is however tedious and gives a large quantity od the polymeride. Prepwatiom of 4-Phenylacetophenone (4-AcetyUiphenyt). A solution of 0.3 gram olf 4-cyalno~diphenyl in 100 0.0.of benzene was added t a a Grignard relagelnt preipared frolm 1-5 gram od melthyl iodide 0.24 gram of magnesium and 40 C.C. of ether. The ether was then rmoveld by distillation aad the reaiduel heiated t a boiling for four hours them cooled deoomposed with water and acid and the benzene layer selparateld dried and evapolrated. The residue was extraatad witlh boiling alcohol in the presenm of wood char-coal the extracts filtereld waporateld and the residue t(reated as bedolre. In this way a small quantity oif EL white solid was obtained crystlallising f roan alcoho'l in colourl~s needles melting at 120-121° and producing no depression of the mellting point of the phenylaoetophenone prepared as described abolve. The la\tter is thered ore 4-phenylatceltophenone.Oxidation undelr the exact aotnditions described by Adam (Zoc. &t .) gave a mixtlure of unohanged ketone and diphenyl-4-earboxylia a,cid melting a t 220O. The latter was eixtraoted from the mixture by means of soldim hydroxidel and the extract filhred and acidified eltc. The law melting point (160-161O) found by Adam was evidently due to1 the preaenm od unohanged ketonel. Prfipwaitiorc of 4-Diphenylyl Benzyl Retom. (a) This substance was obtaineld by condensing diphenyl (1 moll.) with phenylamtyl ohloride (2 mods.) in carbon disulphide solution by means 04 aluminium chloride the resatioln mixture being tretated in the usual manner. The blaak ta'rry solid olbtainetd in tlhis way was edsacted relpeatedly with boliling alooiholl the extracts preoipitlated witlh waster and t,ho resulting yelbw soilid freed from phelnylamtia said by extra&ion with hot aqueous a m o a i a .The product was finally crystdised repelatredly fro STUDIES IN RING FOR*MATION. PART IU. 1149 adcoholl and then mellteld a,t 150". Its prolperties agreed with tlhose desoribetd by Fapcke (Zoc. cit.). ( b ) ,4 solution olf 0.9 gram of 4-cyanodiphenyl in 100 0.0. of benzene was added to1 aI Grignard relagent prepared from 2 grams of belnzyl chloride 0.36 gram of magnesium and 50 C.U. of ether. The1 latter solvent was removed by distillation the relsidue heated to1 boliling folr six hours and theln decomposeld in the usual manner. Extractioln of the1 solid finally obtaineld with light peltroleurn and benzelne gavel a white solid melting after crystlallisation from alcohol a t 150° and producing no depression of the1 melting point of the product from (a).The latter is thelrelfore 4-diphenylyl benzyl ketone. Prepav-altiom of 4 4/-Dicymodiphenyt. This substance has hithelrta only been prepared from tlhe mrre-sponding disulphonia aoid by fusion with potassium oyanide. The following method was found to1 be more convenient$. A solutqion of 17.2 grams of benzidine in 42.5 O.C. of concren-tlrated hydrolchloiric acid and 150 C.C. 09 watelr was diazoltised with a concentratled solution of sodium nitrite and the solutioln added gradually t a a warm solution prepareld by mixing 50 grams of cuprio sulphatei pentahydrats (in 200 U.U. of wa8br) with 55 grams of potassium cyanide (in 100 C.G.of water) benzene being added to prevelntl the accumulation of froth. The relaction was completed by helating to a holt water-bath temperature( the brown precipitake collleated washed with boiling water and drield and then extracted repeatedly with much boiling alcohol; the filtered extracts were predpitated with water and the combined precipitakes dried and orystallised from pyridine. The pure nitrile obtained in this way melt8 a t 2 3 5 O (uncorr.) as stated by Doebner (Bev. 1876 9 272). 4 4t-Diacetykdiphenyl. A sohtion of 2 grams of 4:4/-dicyanodiphenyl in 150 0.0. of benzene was atdded to a Grignard reagent prepared from 6 grams of melthyl iodide 1 gram of magnesium and 50 C.U. of ether. The et'her was remowd by disti!!ation and tho rcsiduei healtd to boil-ing f o r six hours thein treated with water and exms of dilute sulphuria acid and the mixture again heated for an hour it being thought possible that otwing to the extreme difficulty of hydro-lysing the dinitrile (see Doebnsr Annalen 1874 172 116) the inhrme1diat.a magnesium additivs compolund f omed in the present reautio~n would mly slowly be deaomposed by cold aoid (see Forste 1150 STUDIES I N RING FORMATION.PART 111. and Judd T. 1905 87 368). The benzene layer was finally separabd and dried. When elvalpoNrated slightly it deposited the diacetyldiphenyl in olusters of nezdles or leaflet,s melting a t 190°, and producing no depreBsion of the mellting polinti oif the diamtyl-diphenyl prelpareld from diphenyl atnd aceltyl chloz5del (above). The lakter is therefore the 4 4/-delrivative.4 4I-Diphenyldiacet y1dGphenyl. A solution of 2 grams of 4:4/-dicyanodiphen;yl in 200 0.0. of benzene was added tlo a Grignard reagent prepareid from 7 grams of belnzyl ohlolride 1 gram of magnelsium and 60 0.0. of ether, when a reddish-brown precipitate wa8 formed. The ether was removed by distillation the rmidue belated to boiling f o r :our holurs decomposed wit'h water and dilute1 acid again helated and fina,lly filtereld and the two laeyers weirel separated. The1 residue from the filtration was combinetd with the residue obtained by eva,porating the benzene solution drield and mystallised from pyridine when it formed white lelaflets melting a t 2 2 4 O (unclorr.) (Found cO=86.7; H=5*6. C,,H,,O requires C=86*2; El-5.6 per cent.).Both of the ket+onea just dworibsd reaot normally with phenyl-hydraxine. C d m s a t i o m of 4 ; 41-Dkcetyldiphenyl with o l - P h e n ? / t e ~ e ~ ~ ~ ~ i n e . When the ketone and the diamine are gently hefated togeither for a few minutes a reld colour is developeld owing to! the folrm-ation of a substame which is prelcripitated by aloohol and is insoluble in tlhe latiar. Condelnsatioa in slightly diluteid glacial aaeltio aaid (to preclude the formation so far as possible) of benz-iminazole) gave WE- en equim olemlar quantities of the reacting substances were tqakein an orange-yeillow proiductl obtained by p r e aipita,tlion of the amt-ic acid sollution with methyl aloohol. On dissolving i t in pyridine and precipitating with light petroleam it was obtained in a miorocrystalline condition (m. p. [indefinih] about 1 5 0 O ) (Found C = 84.7 ; H = 5.7 ; N = 4-2. C,,B,O,N, [2 moils. of ketone condensing with 1 mol. of base with elimin-ation of 2 no3s. of water] requires C=83*2; H=5.8; N=5*1; C2,HI8Np [l mod. of ketone 1 mol. of base elimination of 2 moils. of warbr] require C=85*2; H=5.8; N=9*0 per cent.). The small quantity of t<he product rendered a tholrolugh investi-gakion kpoiwible but' itl is tbought tihat the evidence indicates the formation of a compound of the iminexolQ type by the condensa NIERENSTEM THE CONSTITUTION OF CATECHIN. PART 11. 1151 tlim of two moilmula of the diketone with one of ol-phenylene-di amine. We desire to thank Mr. R. G. Hook for some help with olne od tlhe experimelnts and the Research Fund Committee of the Chemical Sooielty far a grant which has partly defrayed the mt of this investigation. THE UNIVERSITY CHEMICAL LABORATORIES, SYDNEY N.S.W. [Received August 30th 1920.