128 PERKIN AND THORPE: XV.-aa-Dirrzethylbulc(Ize-ap~-t.51.icarboxylic Acid, y-Keto- PP-di172etJ2Y~entc~rnetJaYlene-a-cai’boxylic Acid, and the Synthesis o f Inactive a- Campholactone of Inuctive a- Campholy tic Acid curd of ,8- Cam- pholytic Acid (isoLuuronolic A c i d ) . By WILLIAM HENRY PERKIN, jun., and JOCELYR FIELD THORPE. DURING the course of a long series of researches on the constitution of camphor and its derivatives, attention has been repeatedly directed t o the advisability of synthesising the more important substances obtained by the degradation of camphor, in order, thus, to definitely establish their constitution. It has frequently been found that analytical methods have, alone, not been sufficient to prove the constitutions of such degradation products, and in the more difficult cases, as, for example, in the determination of the constitutions of carnphoronic and isocamphoronic acids, a definite proof was first obtained by synthetical means (Perkin and Thorpe, Trans,, 1897, 71, 1169 ; 1899,75, 897; Perkin, Trans., 1902, 81, 246).I n continuing our experiments in this direction, we have endeavoured t o find some means of synthesising, not only open chain acids, such as those mentioned above, but also closed chain compounds more closely allied to camphor and camphoric acid, We hope that it may eventu- ally be found possible to prepare these substances themselves, in quantities sufficient for definite identification, by synthetical means, but the problem is an exceedingly difficult one.* In the present paper, we describe a series of reactions which has led * This investigation, an abstract of which appeared in the P~oceedings, 1903, 19, 61, was completed and this paper written before the publication of G.Iiomppa’s complete synthesis of camphoric and dehydrocaniphoric acids (Bcr., 1903, 36, 4832).~~-DIMETHYLBUTAKE-~P&TRICARHOXY LIC ACID. 129 t o the synthesis of inactive a-campholytic acid, and indirectly t o t h a t of P-campholytic acid (isolauronolic acid). When ethyl cyanoacetate is digested in alcoholic solution with sodium ethoxide and ethyl bromoisobutyrate, sodium bromide separ- ates and ethyl cyanodimethylsuccinate is formed, thus : C0,Et yH*CN CO,Et*CHNa*UN + CO,Et*CMe,Br = CO,Et.CBle, + NaEr (compare Bone and Sprankling, Trans., 1899, 75, S54). The sodium compound of this cyano-ester interacts readily with ethyl P-iodopropionate, forming ethyl P-cycLizo-aa-clinzethylbutarae-aps- tricurbox ylat e, C0,E t Q (CN)Na CO,Et*CMe, + CH,T*C H,*CO,Et = C0,Et *~(CN)*CH,*CH,*CO,Et C0,Et CMe2 + NaI, but, unfortunately, during this reaction, a large amount of the P-iodo- propionic ester is decomposed with the elimination of hydrogen iodide and formation of ethyl acrylate, CH,:CH*CO,Et, and by no variation in conditions could this be avoided. This loss of valuable material would have made further progress impossible had i t not been discovered that the acrylate thus produced may be converted into ethyl cyanodi- methylbutanetricarboxylate by the following series of reactions. Ethyl acrylate condense., with the sodium compound of ethyl cyano- acetate, yielding the sodium compound of ethyl a-cyanoglutarate, thus : CO2Et*CHN&*CN + CH,:CH*CO,Et = CO,Et*CNa( CN) *CH,*CH,*CO,Et, a n ester which had already been prepared by L.Barthe (Compt. Tend., 1894, 118, 1268) from the sodium compound of ethyl cyanoacetate and ethyl P-bromopropionate. The sodium compound of ethyl a-cyano- glutarute, when digested in alcoholic solution with ethyl a-bromoiso- butyrate, gives a good yield of ethyl cyanodimethylbutanetricarb- ox ylate, C0,Et*CNa(CN)*CH,*CH2*C02Et + CO,Et*CMe,Br = CO,Et*Q( CN)*CH,*CH,*CO,Et + NaBr, C0,Et*CMe2 When boiled with hydrochloric acid, this cyano-ester is hydrolysed with the elimination of carbon dioxide and formation of aa-dimethyl- butune-afl6-tricar boxglic m i d , VOL, LXXXV.K130 PERKLN AND THORPE: a crystalline, readily soluble acid which melts a t 155-151". When this acid is distilled under reduced pressure, it loses water and is con- verted into the intern-a2 up-ccnlhydride of dimethglbutanetricarboxylic acid, CH*C H,*CH,* C0,H /A / \ W-O*CO*C'Me, which melts a t 9 8 O and dissolves in water, regenerating the tribltsic acid. If the dry sodium salt of the acid is heated with acetic anhydride at 140°, carbon dioxide is evolved and a new acid is pro- duced,* decomposition taking place according t o the equation : The y-keto-PP-dimetlzylpentamethylene-a-carboxylic acid thus obtained melts at 109-llOo and yields an oxime and a sparingly soluble semi- carbazone; i t is isomeric with, and very closely allied to, the 6-keto- /3/3-dinzsthyZpentamethylene-a-carboxylic acid, ,CH,-- QO CMe, GH,' CO,H*CH\ which Perkin and Thorpe (Trans., 1901, 79, 782) obtained by the reduction of dimethyiketodicyclopentanedicarboxylic acid, and which melts a t 103".The latter acid is not easily acted on by sodium amalgam, b u t if a large excess is employed, i t is ultimately reduced to the corresponding hydroxydimethylpentamethylenecarboxylic acid This acid must exist in trans- and &-modifications, should, like all cis-y-hydroxy-acids, readily yield a the latter of which lactone, and since the acid (m. p. 115;) which-was actually obtained showed no tendency to lose water with formation of a lactone, it is evidently the trans- modification. On the other hand, it is interesting t o note that the y-keto-acid (m.p. l l O o ) , unlike the 8-keto-acid (m. p. 103O), is not reduced at all even when the aqueous solution of its sodium salt is boiled with a large excess of sodium amalgam, and this difference in behaviour is doubtless due t o the proximity of the >CMe, complex to the keto- * If this curious reaction should prove to be a general one, it will afford a con- venient means of synthesising many importaiit closed chain keto-acids and, for this reason, experiments are being carried out by one of us with the object of ascertaining the exact conditions under which the change takes place.U~-DIMETHYLBUTANE-~B~-TRICARBOXY LIC ACID. 131 group in the latter case. Reduction of the y-keto-acid may, however, be effected by adding sodium to the boiling alcoholic solution of the acid, when an a1 mos t quantitative y ie Id of trons-y-h ydrox y-PP-dime th y Z- psntccmethylene-a-carboxylic acid (m.p. 10 lo), is obtained. The most remarkable property of this acid is its great stability" since, although it is a y-hydroxy-acid, it does not yield a lactone on boiling with dilute sulphuric acid and, even wheln distilled, it passes over for the most part unchanged. This unusual stability, even of a trans-modification of a y-hydroxy- acid, led us to suspect that some intramolecular change might have taken place during the energetic reduction with sodium and alcohol, but tbat this is not the case is proved by the behaviour of the hydroxy-acid on treatment with chromic acid, when it is very readily oxidised and converted into the y-keto-acid from which it had been obtained by reduction. When t~ans-y-hydroxydimethylpentamethylenecarboxylic acid is heated with hydrobromic acid, i t yields y-b~omodim~thylpenta?~e~~yl~n~- carboxylic acid, and this viscid, syrupy substance, when boiIed with sodium carbonate, loses hydrogen bromide with the formation of dimethylcyclopentenecurhxylic acid (b.p. 2 36') ; * The reason for the remarkable stability of this hydroxy-acid is difficult to understand, but it niay be mentioned that other closed chain y-hydroxy-acids are known-for example, the y-hydroxyhexahydro-p-tolnic acid : /Ye CHzI CH,\/GH, ICH*OH, CH*CO,H prepared by Tiemann and Semmler (Ber. , 1895, 28, little tendency to yield lactones (compare Perkin 1375). 2143)--which also exhibit very and Tates, Trans., 1901, 79, On the other hand, trans-hydroxyhexahydro-xylic acid, OH'C'CH, c H/'WH* I 1 CH2\/CH'CH3 HC*CO,H ' in which the hydroxy-group is in the 6-position with respect to the carboxyl radicle, readily yields a lactone on distillation (Perkin and Lees, Trans., 1901, 79, 344), and this and other similar observations seem to point to the possibility of lactone formation taking place more readily in the 6- thaii in the y-position in ring compounds. E 2132 PERKIN AND THORPE: Action of Magnesium Methyl Iodide on Ethyl Ketodimethylpcnta- methylenecarboxylate.Formation of a-Campholactone und Synthesis of a-Cmnpholytic Acid and P-Campholytic Acid (is0 Lauronolic Acid). As stated on p. 128, the primary object of this research was the synthesis of closed chain compounds closely allied to camphor and this was accomplished i n the following way.E thy1 ketodimethylpent amet hylenecarboxylate is readily acted on by magnesium methyl iodide with the formation of several substances, among which is a new lactone produced according to the following scheme : CH,*FO CH;YMe(OMgI) CH,.yMe-O I 5% -+ I yille, -+- I y ~ e , 1 c1 H,*CH*CO,Et UH,* CH*CO,H U H,*CH--UO On account of its close relationship to campholactone, from which, indeed, it only differs in the reversal of the position of the lacfone group, this new lactone has been named inactive a-cumpho- lactone.* Inactive a-campholactone is an oil which smells strongly of pepper- mint; and distils at 155-157O under 50 mm. pressure ; when treated with hydrobromic acid, it yields inactive y- bromotrimetiiylpentamethylene- cavboxylic acid, which melts at 108" and, when digested with sodium carbonate, loses hydrogen bromide with the formation of the corresponding unsaturated acid : ++ I n the preliniinary notice of this research in the Proceedings (Zoc.cit.), we named the new lactone " isocampholactone " in order to ciiiphasise its great similarity to campholactone. We snbsequently found, however, that this name had already been given t o a lactone of unknown constitution which Noyes (Ber., 1895, 28, 553) had prepared by the action of nitrous acid on aniinolmronic acid. T'his latter lactone, which melts a t 23" and has quite different properties to our synthetical lactone, was not obtained in sufficient quantity for analysis, but i t s method of forination makes it probable that it is isoxieric (possibly stereoisomeric) with campholactone.Owing t o the nanie " isocampholactone " having been previously used by Noyes, we have altered that of our lactone to " a-campholactone."ACID. 133 CH,* ?Me-0 CH2-yMoBr I ?Me, I --+ I CAfe, -+ CH,*CH-CO CH,~H.CO,H The formula of this unsaturated acid is that of inactive a-cccmpholytic acid, an acid which was first obtained by one of us (Perkin, Trans., 1903, 83, 853) by the reduction of a-camphylic acid, C,H,,*CO,H, with sodium amalgam. I n order to prove that the synthetical acid is inuctive a-campholytic acid, it was digested with dilute eulphuric acid when crystals were formed which melted a t 132' and were easily identified as P-campholytic acid (isolauronolic m i d ) , the well- known remarkable isomeric change indicated by the f ormuh, CH-yMe CH,*yMe, OH,* CH* C0,H CH,*C*CO,H a-Campholytic acid.isolauronolic acid. having taken place. That the isolauronolic acid thus synthesised is iden- tical with the acid obtained from camphoric acid was shown by mixing equal quantities of the two preparations when the mixture melted at 132*, the melting point of the constituents. Furthermore, the synthetical acid, on oxidation with permanganate, yielded isolauronic acid. There can thus be no doubt that the above series of reactions consti- tutes a synthesis of both inactive a-campholytic acid and of isolauronolic acid (P-campholytic acid) and therefore the constitution of these important acids may now be taken as definitely established. If the formulE of a-campholactone and the derivatives obtained from i t are carefully examined, it will be a t once seen how very closely these synthetical substances are related t o camphoric acid.Thus if y-bromo t rime t h ylpen tame thy lenecar box ylic acid be taken as an example, it is only necessary t o replace the bromine atom in that substance by the carboxyl group in order to accomplish a synthesis of camphoric acid : CH,*yMeBr CH,*$!Me*CO,H CH,*CH*CO,H CH,* C H*CO,H Many experiments were made, under the most varied conditions, in the hope of being able to bring about this change, but with little success. Thus the bromo-acid and its ester mere treated with silver cyanide, with solid potassium cyanide, alone and in the presence of anhydrous hydrogen cyanide, and with aqueous and alcoholic solutions of potassium cyanide, but although small quantities of a crystalline sub- 1 p e 2 I p e ? I -+ I ?Me2 I I ?Me,134 PEREIN AND THORPE: stance having properties identical with those of i-camphoric acid were obtained in two cases (see p.146), the amount was too small for analysis and definite identification. The cause of the failure in these experiments is the great readiness with which the bromo-acid loses hydrogen bromide with the formation of a-campholytic acid. Again a-campholactone was heated with potassium cyanide and with anhydrous formic acid, but in both cases, a-campholytic acid was the product of the reaction. Lastly, a synthesis, similar to that of benzoic acid from bromobenzene by the action of magnesium powder and carbon dioxide, was attempted with the ester of bromotrimethylpentametbylene- carboxylic acid, but again without success.It has not been thought worth while to describe all these various experiments in detail in this paper, the above short account being sufficient to indicate the reactions which were attempted with the object of synthesising carnphoric acid from a-campholactone. Condensation oj Ethy I C yanodimetluJsuccinnte with Eth y I P-Iodopro- pionate. Formation of Ethyl Cyanodi~methylbutanetricarboxylate, CH,*CH,*C02Et co,Et* )<CMe2 CO,E t I n carrying out this condensation, it was found that even slight differences in temperature and other conditions had a very marked influence on the yield of the condensation product, and the best results were usually obtained by working in the following way.Sodium (23 grams) is dissolved in absolute alcohol (350 c.c.) and, after cooling thoroughly, mixed with ethyl cyanodimethylsuccinate (227 grams) and then ethyl /I-iodopropionate (228 grams) added in small quantities, care being taken, by cooling with water, that the temperature never rises above 25O. After 1 2 hours, the product is heated on the water- bath for 2 hours, water is then added, and the oily ester, which smells strongly of ethyl acrylat-e, is twice extracted with ether. The ethereal solution is thoroughly washed with water, dried over calcium chloride, and the ether very slowly distilled off; the residual oil is then distilled a t first under the ordinary pressure until the temperature rises to 130° in order that the ethyl acrylate which comes over below this may be collected for subsequent use (see p.136). The distillation flask is then attached to the vacuum apparatus and the distillation continued under 20 mm. pressure when, after several fractionatiuns, ethyE cycLnodimethylbzttanetricarboxyl~te is readily obtained pure as a colourless oil boiling a t about 210' (20 mm.). 0.2976 gave 11 -3 C.C. of nitrogen at 23' and 760 mm. N = 4.3. C,6H,@6N requires N = 4.3 per cent.aa-DIMETHYLBUT ANE-aps-TRfCnRBOXYLIC ACID. 1 35 The yield of this ester varies in a remarkable way in different experiments carried out, apparently, under precisely the same condi- tions. In one instance, an almost theoretical yield was obtained, but in other cases hardly a trace was produced, the average yield being about 25 per cent.of the theoretical. The remainder consists of ethyl acrylate and unchanged ethyl cyanodimethylsuccinate and sometimes a considerable quantity of an oil having a high boiling point is produced. I)irnethyZbutunetricarboxylic Acid, C0,H*CH<CMe,.002H CH,*CH,*CO,H The cyan ogen group in e t h y 1 cyanodi me t h y 1 bu tanetricar boxy lat e is hydrolysed only with great difficulty, even on long boiling with hydro- chloric or dilute sulphuric acid the process is far from complete, and it was only by employing concentrated sulphuric acid in the first instance that a satisfactory result was ultimately obtained. The cyano-ester (in quantities of 50 grams) is dissolved in an equal volume of concen- trated sulphuric acid, and, when the solution is cold, water is added until oily drops just comnience to separate, The whole is then boiled on the sand-bath, the alcohol produced being allowed to escape through the air-condenser, and small quantities of water are added from time to time in order to prevent charring, After 8 hours, the solution is diluted with an equal volume of water, saturated with ammonium sul- phate, and extracted several times with ether; the solvent is then evaporated and the residue dissolved in a small quantity of water and mixed with an equal volume of concentrated hydrochloric acid.After two days, the crystalline crust which has separated is collected and purified by recrystallisation from hydrochloric acid, from which it separates as a sandy powder.0.1444 gave 0.2615 CO, and 0.0860 H,O. C = 49.4 ; H = 6.6. 0.1295 ,, 0.2359 CO, ,, 0.0750 H,O. C=49*5 ; H=6*4. C,H,,O, requires C = 49.5 ; H = 6-4 per cent. Dimet~~llbzctccnetricurboxylic ucid melts a t 155-157" and is readily soluble in water, but sparingly so in concentrated hydrochloric acid. It is rather sparingly soluble in dry ether, and crystallises from this solvent in small, glistening prisms or leaflets. On titration with decinormal caustic soda, 0.201 8 gram neutralised 0.1 108 gram of NaOH, whereas this quantity of a tribasic acid, C,H,,O,, should neutralise 0.1110 gram of NaOH. The hydrochloric acid mother liquors from the purification of the acid contain a large quantity of very impure material ; this is extracted with ether, and, after evaporating off the solvent, the residue is esteri- fied by boiling with alcohol and sulphuric acid.The ester is extracted136 PERKIN AND THORPE: in the usual way and fractionated under reduced pressure, when pure ethyl dimetlLyEbutunets.icurboxylute, CMe,( C0,Et)-CH( CO,E t)*CH,*CH,* CO,E t, is obtained as a colourless oil boiling at about 195" (40 mm.). 0.1722 gave 0.3761 CO, and 0.1319 H,O. From this ester, the pure tribasic acid is rekdily obtained by hydro- C = 59.5 ; H = 8.5. C1,H,,06 requires C = 59.6 ; H = 8.6 per cent. lyfiis with concentrated hydrochloric acid. The Internal Anhydride of Di~netl~ylbutccn/Let~ic~i~box~lic Acid, C H*CH,*CH,*CO,H CO*O*CO*CMe, When dimethylbutanetricarboxylic acid is heated under reduced pressure (45 mm.), i t melts and then, at once, gives off water, and the anhydro-acid distils at about 255" as a viscid, colourless syrup which soon begins t o crystallise in stellate groups and gradually becomes solid.By dissolving in hot benzene and allowing to cool slowly, the substance is obtained in well-defined, four-sided plates, but if the solution is rapidly stirred, it separates as a sandy, crystalline pre- cipitate. 0.1776 gave 0.3503 CO, and 0.0959 H,O. C,Hl,O, requires C = 54.0 ; H = 6.0 per cent. The anhydride molts at 98" and is sparingly soluble i n cold water ; it dissolves, however, readily on boiling with water, and i f a n equal volume of hydrochloric acid is added t o the concentrated solution, cryst,als separate which melt at 155-157" aud consist of the pure tribasic acid.C=53*8; H=6*0. Formation of Ethyl ~yanod~rnethy~butanet,.icccrbox~Zc~~~ from Ethyl Acrylccte, Ethyl Cyanoucetacte, and Btli y l Brornoisobutyrate. The reasons which led to this method being carefully worked out are given in the introduction. Sodium (23 grams) was dissolved i n 300 grams of alcohol, mixed with ethyl cyancacetate (1 13 grams), and, after well cooling, ethyl acrylate (100 grams) added in several portions. The condensation takes place very rapidly and with evolu- tion of a considerable amount OF he&, and in a short time the whole of the sodium compound of the ethyl cyanoacetate will have passed into solution. The mixture is heated on the water-bath for about two hours, and until a drop of the liquid, on dilution with water, no~~-DIMETAYLBUTANE-~/~&TRICARBOXYLIC ACID.137 longer deposits an oil; it is then mixed with ethyl bromoisobutyrate (195 grams) and heated in soda-water bottles in a boiling salt-bath for eight hours. On diluting with water, extracting the oily product with ether, and fractionating under reduced pressure, pure ethyl cyano- dimethylbutanetricarboxylate is obtained, the yield being 75 -80 per cent. of the theoretical. 0.2646 gave 10.2 C.C. of nitrogen a t 20' and 765 mm. That this cyano-ester has the same constitution as that obtained by the process described on p. 134 was proved by hydrolysis with sulphuric acid, when a good yield of dimethylbutanetricarboxylic acid was obtained which melted a t 155-157O. N = 4.4. C,6H2s0,N requires N = 4.3 per cent. 0.1516 gave 0,2759 C'O, and 0.0886 H,O. Much of the acid required for this investigation was prepared by C = 49.6 ; H = 6.5.CoH,,06 requires C = 49.5 ; H = 6.4 per cent. the foregoing process. Purtiul Hpdmlysis of Ethyl C'ywnodimetl~~Zb~ctcclzetricccrbox?/kt~. During the course of this investigation, a number of experiments were instituted with the object of eliminating the carboxyl radicle adjacent to the cyanogen group in ethyl cyanodimethylbutanetricarb- oxylate, and the results obtained may be briefly described as follows. When an alcoholic solution of 12 grams of pure caustic potash is mixed with 35 grams of the cyano-ester and the whole allowed to remain overnight, a crystalline potassium salt separates in quantity, This was collected at the pump, washed with alcohol, in which it is sparingly soluble, and dried a t 100' ; it then weighed 10 grams.0.3484 gave 0.1781 K,SO,. K = 22.9. C,,H,,O,NK, requires K = 22.4 per cent. The constitution of this salt is probably represented by the formula It was very soluble in water and gave no precipitate on acidifying with hydrochloric acid ; the whole was therefore repeatedly extracted with ether, the ethereal solution dried over calcium chloride and evaporated, when a colourless oil was obtained, which, over sulphuric acid in a vacuum desiccator, became very viscid but did not crystal- lise. The analysis shows that this substance is ethyl dihydroge?z cyanodimet?~ylbzctanetrical.boxplate, corresponding with the above potass- ium salt.138 PERKIN AND THORPE: 0.1974 gave 8.3 C.C. of nitrogen a t 17" and 759 mm.C,,HI7O,N requires N = 5-2 per cent. When this dibasic acid is heated at 150°, i t decomposes with evolu- tion of carbon dioxide, and, when the evolution of gas has ceased, the residue distils almost completely at 345-250' (50 mm.) as a viscid oil which consists of ethyl ?&ydi*oge?a cyaPzodimethyZbzctanedicn~doxyZccte, N=4*9. 0 2138 gave 11.5 C.C. of nitrogen a t 20' and 759 mm. N=6*2. C,,H170,N requires N = 6.2 per cent. This acid is formed when the sodium salt of dimethylbutsnetri- carboxylic acid is heated with acetic anhydride. The sodium salt is prepared by dissoIving the pure acid (109 grams) in water, adding 80 grams of anhydrous sodium carbonate, and evaporating the alkaline liquid t o dryness. I n order t o obtain t h e salt as finely divided as possible, it is powdered and passed through a fine wire sieve, being then again dried at 100' and transferred t o a flask, into the neck of which is fitted a tube bent twice a t right angles.After adding 150 grams of freshly distilled acetic anhydride, the whole is slowly heated in a n oil-bath when it will be noticed that carbon dioxide begins to come off below loo', as can be readily seen by causing the tube attached t o the flask to dip under baryta water. The temperature is kept at 135-140' for about 6 hours, then 50 grams of acetic anhydride are added and the heating continued until the evolution of carbon dioxide has practically ceased, which is the case after about three hours. As soon as t h e product is cold, i t is esterified by adding an excess of a solution of one volume of sul- phuric acid in three volumes of alcohol and heating for 6 hours on the water-bath in an open flask in order t h a t as much ethyl acetate as possible may be removed by evaporation, The product, which smells stiongly of ethyl acetate, is diluted with water and extracted several times with ether ; the ethereal solution is then thoroughly washed with water and dilute sodium carbonate solution" and evaporated.On fractionating this oil under 100 mm. pressure, about 50 grams distil a t 160-180" leaving a viscid, nearly black residue? in the retort * The dark sodium carbonate extract, on acidifying and extracting with ether, yields a brown oil from which considerable quantities of ketodimethylpentamethyl- enecarboxylic acid may be obtained by treatment with semicarbazide or by esterifi- cation and distillation under reduced pressure and subsequent hydrolysis. t The residues from several preparations were mixed and fractionated underW-DIMETHY LBUTANE-C@-TRICARBOXY LIC ACID.139 and, on repeatedly fractionating the distillate, an oil (35 grams) is obtained which boils at 170-172' (100 mm.) and consists of pure ethyl ~tod~methy~pe~ztccmeth~~enecccrboxy~nte. 0.1765 gave 0.4196 CO, and 0.1397 H,O. C = 64.8 ; H = 8.8. 0,1584 ,, 0.3771 CO, ,, 0.1243 H,O. C=64.9; Hs8.7. C,,H,,O, requires C = 66.2 ; H = 8.7 per cent. This ester is hydrolysad by boiling with an excess of methyl-alcoholic potash ; the solution is diluted with water, neutmlised, and evaporated until free from alcohol.After acidifying with hydrochloric acid and extracting six times with pure ether, the ethereal solution is dried over calcium chloride and evaporated, when a pale yellow oil is obtained which, on cooling, rapidly solidifies. The mass is left in contact with porous porcelain t o remove a trace of oily impurity, and then dissolved in a small amount of water, from which the pure acid separates slowly in groups of colourless needles. 0,1485 gave 0,3360 GO, and 0,1043 H,O. C = 61.6 ; H= 7.8. 0.1634 ,, 0.3686 CO, ,? 0.1144 H,O. C = 61.5 ; H = 7.8. C,H,,O, requires C = 61.5 ; H = 7.7 per cent. Ketodirnethylpe~tamethyle~zecccr~oxylic cicid melts at 109-1 10' ; it dissolves readily in benzene, acetone, ether, alcohol, and chloroform, but is sparingly soluble in light petroleum.It is exceedingly soluble in hot water but much less so in the cold, and it may be obtained in long, colourless needles if the hot solution is allowed t o cool very slowly. The pure acid also crystallises readily from ether in the form of hard, glistening pribms. A slightly alkalke solution of the ammonium salt cf the acid gives no precipitate with barium or calcium chloride, even on boiling ; with copper sulphate, no precipitate is formed in the cold, but, on boiling, a pale blue, apparently crystalline, copper salt separates. The oxirne, C,H,,(C:N*OH)*CO,H, was prepared by dissolving the acid in a considerable excess of caustic potash solution, adding twice the calculated quantity of hydroxylamine hydrochloride, and allowing the whole to remain for two days.On acidifying with hydrochloric acid, a crystalline precipitate separated, which was extracted with ether, the ethereal solution was carefully dried with calcium chloride and evaporated t o a small bulk, when the oxime soon began t o separ- ate in crystalline crusts. 30 min. pressure, when about one-third distilled a t 150-190". This portion, on hydrolysis with hydrochloric acid, yielded a coiisiderable quantity of crude di- me thylbutanetricarboxy 1 ic acid, which was purified by recrys tallisati 017, first from hydrochloric acid and then from ether.140 PERKIN AND THOKPE: 0.203 gave 15.3 C.C. nitrogen at 19' and 737 mm. C,H,,O,N requires N = 8.2 per cent. KetozimedinaetJ~yZpentameth~ZenecarboxyZic acid melts at about 1 9 5 O , but it turns brown and commences to decompose below this tempera- ture.It is somewhat sparingly soluble even in boiling water, and separates, on cooling, as a sandy powder. I t dissolves readily in sodium carbonate, and, when heated in a test-tube, i t decomposes with a slight explosion, and yields an oily distillate which crystallises on cooling. The senaicccrbazoize, CG~11(C:N*NH.CO*NH2)*C0,H, separates at once a s a sandy powder when a concentrated aqueous solution of the acid is mixed with semicarbazide hydrochloride and sodium acetate, It melts at about 217' and is very sparingly soIubIe in water. N = 8.3. 0.1544 gave 26.6 C.C. nit'rogen at 19' and 751 mm. N = 19.5. C9HljO3N3 requires N = 19.7 per cent. Bedzcction of Ketodimet~?/~entametJ~~~enecc6~~o~yZ~c Acid. Formation of trans- y- IJycls.ox~/dinzeth y11,3entumetJLyleneca./.box~lic Acid, 7 H, YH O H CH, Chle, \/ CH *C 0,H It is stated in the introduction that ketodimethylpentamethylene- carboxylic acid is not reduced by sodium amalgam, and this was shown by the fact t h a t when 30 grams of the acid, dissolved in diIute aqueous caustic soda, was boiIed with 2500 grams of 4 per cent.sodium amalgam, and the solution subsequently acidified and extracted,with ether, almost the whole of t h e keto-acid was recovered unchanged. When, however, the boiling alcoholic' solution of the acid is treated with sodium, reduc- tion is readily effected. The pure acid (30 grams) was dissolved in about 500 C.C. of absolute alcohol, heated t o boiling in a reflux apparatus, and treated, as rapidly as possible, with 70 grams of sodium, boiling alcohol being added from time to time as the reaction slackened. The product was dissolved in much water, evaporated on t h e water-bath until free from alcohol, acidified with an excess of hydrochloric acid, and heated on the water-bath for half a n hour in the expectation that the hydroxy-acid, if present, would thus be con- verted into its lactone.As, however, no neutral oil separated, t h e whole was saturated with ammonium sulphate, and extracted 10 times with ether. The ethereal solution was dried over calcium chloride and evaporated, when 28 grams of a viscid oil remained and solidified on cooling. That no unchanged keto-acid was present i n this product was shown by treating a small portion with semicarbazide and sodium~~x-DIME'~HYLBUTANE-~@~-TRICA~~BOXY LIC ACID.141 acetate, when no trace of the very sparingly soluble semicarbazone (p. 140) was formed. Considerable dificulty was experienced in re- crystallising the crude reduced acid, but this was ultimately accom- plished by dissolving it in dry ether, adding an equal volume of benzene, and allowing the solution to evaporate slowly at the ordinary temperature. The crystalline crusts thus obtained were drained on porous porcelain, and dissolved in a small quantity of hot water, when, on slowly cooling, the pure acid separated in well-defined, glistening prisms. 0.1690 gave 0,3762 CO, and 0.1326 H,O. C=60-7 ; H=8.7. 0.1353 ,, 0'3012 CO, >, 0*1075 H,O. G 6 0 . 7 ; H=8.S. C,HI,03 requires C = 60.8 ; H = 8.S per cent. trans-y-HydroxydimethyZpent~~cmethyZenecarboxyZic acid melts a t 1OO-10lo and is readily soluble in water, alcohol, and ether.That it is a monobasic acid was shown by its behaviour with decinormal caustic soda, when 0.1977 gram of the acid required for neutralisition 0.0807 gram of NaOH, whereas this amount of a monobasic acid, C8Hl4O3, should neutralise 0.0506 gram of NaOH. This y-hydroxy-acid is remarkably stable, and shows no tendency to become converted into a lactone when boiled with dilute hydro- chloric or sulphuric acid, and, even on distillation, it does not yield a lactone, as is proved by the following experiment. Three grams of the pure acid were distilled under 40 mm. pressure, when the whole passed over a t 205--210° as a viscid, colourless oil resembling glycerol, and this, on cooling, solidified to a brittle, trans- parent, glassy mass.This product was completely soluble in cold dilute sodium carbonate solution, showing that it was not a lactone, and, without further purification, i t yielded numbers which differ little from those required by the hydroxy-acid. 0,1649 gave 0.3718 CO, and 0.1348 H,O. C = 61.4 ; H = 9.0. C,H,,03 requires C = 60.S ; H = 8.S per cent. Oxidation by Means of Chg*omic Acid.-Owing to the curious properties of this hydroxy-acid, it was thought necessary to investigate its behaviour on oxidation in order to prove that its constitution is that represented by the formula indicated on p. 140. The pure acid (2 grams) was dissolved in water, mixed with potassium dichromate (1-4 grams) and a small amount of sulphuric acid, and gradually heated to boiling when oxidittion readily took place.The product was extracted with ether, the ethereal solution dried over calcium chloride and evaporated to a small bulk, when crystalline crusts (1.5 grams) slowly separated, which melted a t logo, and consisted of pure ketodi- meth y Ipentameth y lenecarboxylic acid.142 PERKIN AND TRORPE: 0.1558 gave 0,3513 GO, and 0.1096 H,O. C=61.5; H=7*8, C,H,,O, requires C = 61.5 ; H = 7.7 per cent. y-Bromodimethylpentccmethylenecu~~oxyZ~c Acid, CH2*YHBr CH,*CH*CO,H I ?Me, ' CHZYH UH, *CH.CO,H cmd Dimethylcyclo~entenecarboxylic Acid, 1 ?Me, . Finely powdered y-hydroxydimethylpentamethylenecarboxylic acid dissolves readily in fuming hydrobromic acid (saturated a t OO), but the bromo-acid does not separate after a time as is the case with the corresponding trimethyl acid (see p.145). The solution was there- fore heated in a sealed tube at 1 O O O f o r about fifteen minutes, when, on cooling, an oily layer had separated on the surface of the hydrobromic acid. The whole was diluted with water, extracted with ether, the ethereal solution thoroughIy washed with water, dried over calcium chloride, and evaporat<ed, when an almost colourless oil remained which, after leaving for some hours over sulphuric acid in a vacuum desic- cator, was analysed with the following result : 0.3373 gave 0.2346 AgBr. Br =; 35.9. CsH,,O,Br requires Br = 36.2 per cent. y-Bromodimethyl~entameti~yZe~-Lecarboxylic acid, when prepared in this way, shows no signs of crystallising, even when left for a long time in a freezing mixture, and, on exposure to the air, it becomes darker and finally almost black.A quantity of the freshly prepared bromo-acid was dissolved in excess of sodium carbonate and boiled for several minutes ; the solution was then acidified and extracted with ether. After drying over calcium chloride and evaporating, an oil was obtained which distilled for the most part a t 160-170" (50 mm.), but there was a considerable quantity of a viscid oil left in the flask, which probably consisted of regenerated y-hydroxy-acid. The distil- late was fractionated under the ordinary pressure, when almost the whole passed over a t 236" (760 rum.) as a n oil which was colourless when hot but became deep sage-green on cooling, a behaviour often observed in connection with substances belonging to the camphor and terpene series.Owing to its method of formation, there can be no doubt that the oily acid thus obtained is dimetl~ylcycloperntenecarboxylic acid. That it is an unsaturated acid is shown by the fact that its solution in sodium carbonate instantly decolorises permanganate.U~-DIMETKYLBUTAN E-C@&TRICBRBOXY LIC ACID. 143 0.2128 gave 0.5315 CO, and 0.1638 H,O. C=68*2 ; H=8*5. 0.1730 ,, 0.4336 CO, ,, 0.1335 H,O. C=68.3; Hz8.6. C8Hl,0, requires C = 68.6 ; H = 8.6 per cent. 17126 Action of Magnesium Methyl Iodide on Ethyl Ketodimethyl- pentamethyZen,ecarboxylccte. Fornaation"of Inactive a-Campholactone. CH,*FMe-O I YVe, 1 . UH,*CH--CO After a long series of comparative experiments, t h e following pro- cess was ultimately adopted for the preparation of a-campholactone.Finely-divided magnesium filings (7.2 grams) are placed in a, 2 litre flask connected with a long water condenser aud covered with ether (300 c c.) which had been very carefully freed from water, and finally distilled over phosphoric oxide. Methyl iodide (45 grams) is then added in two portions, the vigorous reaction being kept under control by immersing the flask in ice water occasionally. Ethyl ketodimethylpentamethylenecarboxylate (37 grams) is dis- solved in pure ether (300 c.c.) in a large flask, connected with a reflux condenser, and immersed in powdered ice and water; the ethereal solution of magnesium methyl iodide is then cautiously poured down the condenser tube in several small quantities, the flask being well agitated and allowed t o cool thoroughly after each addition.As each quantity of the magnesium methyl iodide comes into contact with the ethereal solution of the keto-ester, a yellowish-white turbidity is produced, and ultimately the rnagnssium compound separates on the sides of the flask as a sticky mass. After half an hour, the pro- duct is decompossd by the careful addition of dilute hydrochloric acid, the yellow ethereal solution is separated, washed with dilute hydro- chloric acid and evaporated, and the residne, which is a mixture of the lactone, unchanged keto-ester, and a neutral substance (which has not yet been investigated), is treated as follows. The deep yellow oilis dissolved in methyl alcohol, and digested for 10 minutes with a solu- tion of caustic potash (20 grams) in methyl alcohol, by which means the lactone and keto-ester are hydrolysed; water is %hen added, and the neutral substance removed by extraction with ether.The aqueous solution is evaporated on the water-bath until free from ether and methyl alcohol; it is then diluted with water and acidified, when a viscid oil separates, which, on heating on the water-bath for 15 minutes, becomes limpid, owing to the conversion of the hydroxy-acid into the corresponding lactone. The whole is extracted with ether, the ethereal solution repeatedly washed with dilute aqueous sodium carbonate, dried over calcium chloride, and evaporated, when an oil144 PERKIN AND THOKPE: (6-8 grams) is obtained which, after repeated fractionation, distils constantly at 155-157' (50 mm.), and consists of pure inactive a-cam- pholactone.0.1649 gave 0.1227 (70, and 0.1392 H,O. C = 69.9 ; H = 9-3. 0.1485 ,, 0.3819 CO, ,, 0.1238 H,O. C = 70.0 ; H= 9.2. C,H,,O, requires C = 70.1 ; H = 9.1 per cent. Inactive a-campholactone is a pale yellow oil which has a most pungent odour of peppermint; it is readily volatile in steam, and distils under the ordinary pressure almost without decomposition. It is insoluble in sodium carbonate solution, but dissolves readily and completely in warm dilute aqueous caustic soda, and the well-cooled solution, on acidifying, remains clear for a considerable time, indicating that the hydroxy-acid corresponding with the lactone is present ; this acid is slowly decomposed a t the ordinary temperature, and more rapidly on warming, with a separation of the lactone.A special experi- ment was made in the hope of being able to obtain the hydroxy- acid in a crystalline condition, in order that a direct comparison might be made with the bydroxydihydrocampholytic acid described by Noyes* (Amer. CAem. J , 1894, 16, 307, 502 ; 1896, 18, 685), and which he considers has the same constitution, namely, CH,* FMe-OH CH,*CH*CO,H I p e , The pure lactone (2 grams) was dissolved in hot baryta water, the solution cooled with ice, 'acidified, and extracted with ether. On removing the ether by a current of dry air, a viscid oil remained, which was obviously the hydroxy-acid, since it dissolved readily and completely in dilute sodium carbonate solution. 0,1496 gave 0.3454 CO, and 0.1252 H,O.C,H,,O, requires C = 62.8 ; H = 9.3 per cent. Tnuctive y-cis-r7~ydro.xyts.imet?~ylpentametJ~~lenecur6~xylic acid, when lef ti over sulphuric acid in a vacuum desiccator, rapidly loses water with the formation of a-campholactone and, when boiled with hydrochloric acid, the change is complete in a few minutes. Hydroxydihydro- campholytic acid, on the contrary, is perfectly stable and, indeed, does not show any tendency t o yield a lactone even when boiled with acids; if then i t has the constitution represented above, i t must obviously be the tmns-modification. C = 64.0 ; H = 9.3. * Conipare Meyer and Jacobson (Lek~b~rch der 01.9. Chcnt., Vol. 11, 1021).~zcz-DIMETHYLBUTANE-CZ@~-TRICARBOXYLIC ACID.145 Inactive y-Bromotl.i~aethyZpentamethylenecccrboxylic Acid, CH,*QMeBr CH,*CH* CO,H and its Ethp? Ester. 1 ?Me2 When a-campholactone is mixed with three times its volume of fuming hydrobromic acid (saturated at 0.; it dissolves, but almost immediately the liquid clouds and a viscid oil separates on the surface of the acid. This oily layer soon darkens in colour and crystallises, and in preparing the pure bromo-acid i t is best to stir vigorously with a glass rod, so that solidification may be complete before decomposition sets in. The solid is thoroughly washed with water to remove hydrobromic acid and spread on porous porcelain t o separate the oily impurity. The residue, which should be almost colourless, is made into a paste with a small quantity of formic acid (sp.gr. 1.22) and again transferred t o porous porcelain, and after the operation has been repeated, a colourless, crystalline mass is obtained which softens at 105O and melts at about 108' with decom- position. 0.2203 gave 0.1'735 AgBr. Br= 33.5. C,H,,O,Br requires Br = 34.0 per cent. Inactive y-bromotrimethylpentamethylenecarboxylic acid readily decomposes with loss of hydrogen bromide and formation of inuctive a-campholytic acid (see p. 147) and therefore no attempts were made to purify it by recrystallisation. As explained in the introduction, i t is exactly similar in its chemical properties t o the hydrobromide of a-campholytic acid (cistrans-campholytic acid), which, according t o Noyes (Bey., 1898, 28, 551), melts at 98-100O and decomposes in moist air, evolving hydrogen bromide.The ethyl ester, C8H1,Br*C02Et.-In preparing this ester, the pure lactone (4 grams) was mixed with phosphorous pentabromide (15 grams), when a vigorous action took place and much hydrogen bromide was evolved. After 10 minutes, the whole was heated on the water-bath for 5 minutes and then poured, in a thin stream, into a large excess of absolute alcohol. As soon as the reaction Elad subsided, the solution mas allowed to cool, diluted with water, and the heavy oil extracted with ether, the ethereal solution was washed with water and dilute sodium carbonate solution, dried over calcium chloride, evaporated, and the &residual oil rapidly fractionated under 70 mm. pressure. A small quantity of oil passed over bdow 165' and then almost VOL.LXXXV, L146 PERKIN AND THORPE: the whole of the remainder distilled at 165-170' leaving only a small, black residue in the flask. 0.1854 gave 0,1377 AgBr. Br = 31.6. C,,H,,O,Br requires Br = 305 per cent. This oil was made the starting-point in a long series of experiments instituted i n the hope of synthesising camphoric acid (see p. 133) but without success. The above analysis and general properties of the oil leave, however, scarcely any doubt that it is inactive etlql y-bromotrimethylpentanzeth ylenecurboxylate. Action of Potussiuna Cycmide on Bromot~inaetl~?llZ,elztan~et~yl~ne- carboxylic Acid.-It was stated in the introduction that, in experiment- ing on the action of cyanides on this bromo-acid, an acid identical in properties with i-camphoric acid was obtained in very small quantities in two cases.I n one instance, the experiment was carried out as follows : freshly prepared bromotrimethylpentamethylenecarboxylic acid (10 grams) was dissolved in alcohol, mixed with a concentrated aqueous solution of pure potassium cyanide (5 grams) and 5 C.C. of anhydrous hydrogen cyanide, the mixture being allowed to remain for -14 days. The dark brown product was heated on t h e water-bath for one hour, evaporated, and the residue mixed with concentrated hydro- chloric acid and, after remaining overnight, heated for three hours on the water-bath. Water was then added and the whole distilled in steam until the distillate was free from P-campholytic acid (see p. 147). The residue was digested with animal charcoal, filtered, evaporated t o a small bulk, and repeatedly extracted with ether, when a yellow oil was obtained which became semi-solid when left for some weeks over sulphuric acid.This substance was drained on porous porcelain until all the oil had been absorbed and the residue recryatallised from a little water. As the crystalline product, which could not have weighed more than a few milligrams, melted indefinitely between 189' and 190', it was heated in a test-tube with a small quantity of acetic anhydride when, on cooling, crystals separated which were drained on porous porcelain and washed with a little acetic anhydride. These crystals closely resembled the anhydride of i-camphoric acid, since they melted at 215-217O and gave on hydrolysis an acid, melting at 204', whereas i-camphoric acid melts at 200--203° and yields an anhydride melting a t 2 2 1 O .Unfortunately, the amount a t our disposal was quite in- sufficient t o enable us to establish the identity of the substance by analysis.* * The method of formation of this small quantity of acid and its properties leave scarcely room for doubt that i t was i-camphoric acid, and experiments were in progress which it was hoped would have yielded enough material for definiteaa-DIMETHYLBUTANE-a,8&TRICARBOXYLlc ACID. 147 CHrVMe UH,*CH*CO,H Synthesis of Inuctive a-Campholytic Acid, I YRle, , and of C H,* ?Me, CH-C*CO,H gMe P-Campholytic Acid (isoLazcronolic Acid), 1 When y-bromotrimethylpentamethylenecarboxylic acid is mixed with excess of sodium carbonate solution, i t rapidly dissolves, and if the liquid is heated just to boiling, cooled, and acidified with hydro- chloric acid, a viscid, oily layer of inccctive a-campholytic acid separates.This mas extracted with ether, the ethereal solution washed until quite free from hydrochloric acid, dried over calcium chloride, evaporated, and the residual oil distilled under reduced pressure, when the whole passed over at 162-164" (45 mm.) as a colourless, viscid oil. 0.1441 gave 0,3696 CO, and 0,1185 H,O. C,H,,O, requires C = 70.1 ; H = 9.1 per cent. Methyl a-CamphoZytate.-This ester was prepared by leaving the acid in contact with methyl alcohol and sulphuric acid for a short timeand then precipitating with water. After extracting with ether, washing with sodium carbonate, drying over calcium chloride, and evaporating, an oil was obtained which distilled constantly at 200".C= 69.9 ; H = 9.1. 0.1475 gave 0.3882 CO, and 0.1287 H,@. C,H,,*CO,Me requires C = 71.4 ; H = 9.5 per cent. On hydrolysis with caustic potash and acidifying, this ester yielded liquid a-campholytic acid, showing that no isomeric change had taken place during esterification (see below). That the acid obtained in the way described above was inactive a-campholytic acid and identical with the acid first obtained by the reduction of a-camphylic acid with sodium amalgam (Trans., 1903, 83, 853) was proved by digesting a small quantity with 25 per cent. sulphuric acid for a few minutes, when, on cooling, a mass of crystals of /3-campholytic acid (isolauronolic acid) separated. These mere collected, washed with water, and recrystallised from dilute acetic acid, when leaflets were obtained which melted a t 1 3 0 O . 0.16S1 gave 0.4318 CO, and 0.1394 H,O. C = 71.7 ; H= 9.7. C = 70.0 ; H=9*1. C,H,,O, requires C = 70.1 ; H= 9.1 per cent. identification. In the meantime, Komppa (Zoc. cit. ) has published his brilliaiit synthesis of camphoris acid, which, once for all, establishes the correctness of Rredt's formula, and it is therefore quite unnecessary to investigate our much less satis- factory process any further. L 3148 NEEDHAM AND PERKIN : 0-NITROBENZOYLACETIC ACID. The identity of this acid with the P-campholytic acid, obtained by the action of aluminium chloride on camphoric anhydride, was proved, firstly, by mixing equal quantities of the two acids, when no alteration in the melting point could be detected, and, secondly, by converting the synthetical acid into isolauronic acid by oxidation. Two grams of the synthetical acid were dissolved in aqueous sodium carbonate, cooled with powdered ice, and then 2 per cent. potassium permanganate solution run in, drop by drop, until the pink colour just remained. The slight excess was destroyed by the addition of sodium sulphite, the whole was then heated to boiling and filtered, and the filtrate evaporated t o a small bulk. The yellow liquid gave, on acidifying, il mass of pale yellow crystals which, after crystallising from water, melted a t 133' and consisted of pure isolauronic acid. 0-1149 gave 0.2720 CO, and 0.0753 H,O. C,H,,O, requires C = 64.3 ; H = 7.1 per cent. Inactive a-campholytic acid has been repeatedly obtained during the course of the experiments which have been carried out with y-bromo- trimethylpentamethylenecarboxylic acid and with a-campholactone. Thus, for example, i t is formed almost quantitatively when the bromo-acid is digested with potassium cyanide in alcoholic solution and when a-campholactone is heated with dry potassium cyanide a t 1 1 oo. C = 64.5 ; H = 7.3. The authors wish t o express their thanks to Mr. D. T. Jones for his valuable assistance in carrying out these experiments, and to state t h a t much of the expense incurred during this investigation has been met by repeated grants from the Government Grant Fund of the Royal Society. OWENS COLLEGE, MANCHESTER.