1X.-RESEARCHES ON THE RELATION OF THE MOLE-CULAR STRUCTURE OF CARBON COMPOUNDS TO THEIR ABSORPTION-SPECTRA. By W. N. HARTLEY F.R.S.E. &c. Professor of Chemistry Royal College of Science Dublin. Part VI. On the Comtitution of Pyriditie Picolhae Quinoliue ant1 Cyanuric A cia. WrmN an atom of carbon is united to an atom of nitrogen no absorp-tion-bands are seen in the ultra-violet spectra transmitted by such a combination. This conclusion may be drawn from the experiments of 31. Soret and of the late Dr. W. A. Miller who proved the photo-graphic transparency of hydrocyanic acid and the cyanides. I 11:tve considered i t necessary to make independent observations 0 3 1 account of the more delicate nature of the instrument and the photo-graphic process I employ and because it is necessary to take into consideration the quantity of substance in the various solutions examined.A strong solution of hydrocyanic acid was prepared by distilling potassic ferrocyanide in the usual way with dilute sulphuric acid ; this was redistilled and its strength determined with a volumetric silver solution. It contained 11.9 per cent. of HCN. The liquid was examined in the undiluted state in layers 30 mm. and 15 mm. in thickness and subsequently in various degrees of dilution. Hgdrocyanic acid is a remarkably diactinic substance and does not under any ordinsry circumstances exhibit absorption-bands. Hence we may draw the conclusion that-The simple union of cwbon to nitrogen does not cause selective absoiptioir o f the ultra-violet rays.It has been suggested and it is generally believed that pyridine and its homologues are compounds in which an atom of triad nitrogsn replaces an atom of carbon in a closed chain of six atoms. Koerner’s formula for quinoline represents it as naphthalene in which an atom of carbon in one of the benzene nuclei is replaced by triad nitrogen, and moreover the recent synthesis of quinoline by Baejer has cou-firmed this theory of its constitution (Bey. 12 1320). Without com-mitting ourselves to views as to the internal structure of the molecules, we may write the two substances thus :-VOL. XLI. 46 HARTLET RELATION VF TIIE JIOLECULAR STRUCTURE C c Py ridine. H Quinoline. As already shown if a molecule contains only two pairs of carbon-atoms doubly linked it will show no absorption-bands and it is coil-ceivable that if a carbon-atom be replaced by nitrogen as in thc formulae given above the nature of the absorption will not b ~ materially affected by this modification because tbe structure of tllt: nucleus of the compound is unaltered ; and although there is one atom of hydrogen less i n the molecule we know that the addition uf hydrogen-atoms provided they do not interfere with the structure of the nucleus has no effect on the absorption so that we may expect the witlidrawal of one atom to have as little influence.In fact we should expect to see a banded spectrum whatever element might bc introduced into the benzene nucleus if the atom were capable o€ vibrating in unison with the carbon-atom it has replaced.Likewise if we regard picoline arid the two pyridine-dicarboxylic acids as modifications of the pyridine nucleus drawn above an(! &anding in the same relation to it that toluene and the phthalic acids occupy with regard to benzene we should expect them also to exhibit absorption- bands. Quinine shows a remarkable banded spectrum which is probably due to the conjugation of four ppidine or two quinoline nuclei since Dobbie and Ramsay (Chem. SOC. J. 1878; 1879 p. 189) have obtained two pyridine-dicarboxylic acids and pyridine-tricarboxyliu acid by the oxidation of this alkaloid. Samples of the two formein substances were kindly given to me by Dr. Ramsay and also some di-pyridine and picoline. Particulars regarding the spectra of these substances are the following :-a-Pyrit7irLe-dic~rbux2/llc Acid.- The substance was dissolved in absolute alcohol and solutions containing m1G3 T;m m;T*, and m&m were examined. A strong band of' absorption tvas noticed, which was not extinguished even in the weakest of these solutions. The band is most strongly marked a t a dilution of 1 in 20,000 ; it lies between the lines 1 7 and 18 Cd wave-lengths 2743.4 and 2,5i4?, Diagram X. 6-Py7iditie-diccLrbo~i~Z~c Acid. - The solutions in absolute alcoho 8. PPRIDINE-DI-CARBOXYLTG ACID. Scale. 10,000 parts to half an inch. Emison & Sons. L i d . S’ Martins Lane.%? DIAQBAM X. a. PYRIDINE-I>I-CARBOXYLIC ACID. Numbers indicating the position of cadmium Lines of known wave length. Scale. 10,OOO parts to half an iuoh. Harrison & S.ons.Litll. S Martins Lane,W. DIAGRAM XI. P I C O L I N E . Numbers indicating the position of cadmium lines of known wave length. Ordinates = volumes of solution containing one volume of Picoline. Scale. 20,000 parts to half an inch OF CARUOS COJIPOUXDS TO THEIR ABSORPTIOS-SPECTRA. 4 7 examined contained i,&5 dG m+5G h6 & and of their weight of the substance. There was a considerable absorption, though not of great intensity in this last liquid the bands stretching from below 1 7 to 18 Cd. The band of transmitted rays did not extend as far as 22 Cd while in the previous case they went up to 23 Cd Diagram IS. PicoZine.-Solutims in absoliite alcohol containing y$m hTi, The solution containing shows a strong absorption-band lying beyond 1 7 Cd but a broad band of rays is faintly transmitted further on.Solutions containing m:35 show a strong absorption-band lying between 17 and 18 Cd and a band of transmitted rays which how-ever are faint between 18 and 25 Cd. The absorption-band is narrowed but not destroyed by dilution to ha and the transmitted rays are still faint. As this substance is a methyl-pyridine it affords us direct evidence of the fact that the substitution of N for carbon in the benzene nucleus does not interfere with the selective absorption, but only modifies it in some degree. The chief modification is seen in the greatly increased. intensity of the absorption. By comparing the Diagram XI with that representing the absorp-tion of methyl-benzene (Phil. Trans. 1879 PI. 25) a very close and striking resemblance is noticed but with the following differences :-the narrow band of transmitted rays which is a little more refrangible than 1 7 Cd> and appears to he characteristic of benzene hydrocarbons, is absent from the picoline spectrum and the broad and similar band of faintly transmitted rays which appears first in solutions containing 1 part of toluene in 2,000 of alcohol does not become visible until a dilution of 1 in 15,000 in the case of picoline.And whereas a dilhtion of 3,000 times extinguishes a distinct absorption-band in the case of toluene the similar band of picoline is continued to a dilution of 1 part of the substance in 100,000 of liquid. QmhoZine.-The pure substance was obtained by fractionating a specimen purchased from Messrs.Burgoyne and Burbidges ; the portion distilling between 233" and 240" was regarded as the purest. Its identity was established and its purity determined in the following manner. Hydrochloric acid gas was passed into a portion of tlic quinoline until it solidified on cooling into a hard crystalline mass. This hydrochloride was dissolved in water and a solution of platinic chloride was added. The platinoquinoline hydrochloride precipitated was washed by decantation separated by filtration and again washed until free from platinic chloride then dried a t first a t a tem-perature not exceeding 80",and finally heated for a short time to 100". A weighed portion of the compound was then ignited and tlzc residual platinum determined. ,,:,, ~ z o o o - ,,:,, and iB$om of the substance were examined.E 48 HARTLEY RELATION OF TEE MOLECULAR STRUCTURE ETC. Weight of substance. Weight of platinum. Per cent. Pt. gram. gram. I 0.7516 0.2192 29.1 G 11 0.8008 0.2326 29-04 The theoretical quantity of platinum in the salt is 29.44 per cent. Photographs were taken of solutions of the base in alcohol. The solutions contained -I- -I_- -1- -I- ~ ___ 1000 5000 10000) 1 5 0 0 0 7 2 0 0 0 0 9 S O A O 0 40A00, 3h and One narrow band of absorption is seen just a little beyond 12 Cd in solutions Containing lo&u8. Two narrow bands and a broad one are seen in solutions containing =fm and traces of absorption con-tinue until the dilution has reached It is interesting to compare the quinoline Diagram XII with the naphthalene Diagram 111 (Chem.SOC. J. 1881). Two of the four absorption-bands have bzcome fused and the remaining ones are much lowered in refrangibility; a t the same time the intensity of the absorption is decreased. Thus three strong bands are seen in solutions containing 1 part of substance in 100,000 of a solution of naphthalene, but with a dilution of 1 in 50,000 they have all disappeared in the case of quinoline. From the ease with which cyanic acid becomes converted into cyanuric acid evolving a considerable amount of heat during thc change and on account of the necessity for the absorption of a large amount of heat to effect its decomposition into the simpler molecule, it appears highly probable that the carbon and nitrogen-atoms con-stitute a closed chaio forming a compound which may graphically be represented thus :-0 I€ their weight of the substance.I I1 \ x / HOC COH Such an atomic grouping would probabIy be the cause of absorption-bands the hydroxyls in the molecule however being inactive in this respect as I have shown by previous observations (see No. 11 April, 1881 this Journal). I n order to gain some knowledge of the constitution of cyanuric acid a beautifully crystallised specimen was dissolved and the ultra-violet spectrum was submitted to its action. A solution of 1 part, by weight in 200 of water gave a strong absorption-band stretchin,g from wave-length 2687 to about 2630 an interval lying midway hetween 17 and 18 Cd. The band of rays beyond this which was b u t faintly transmitted does not extend t o 22 Cd.A solution con 30,000 .tu,oao 50,auo I a 0,mo Absorption bands traceable in solutions contailring ha of the substance. Ordinates =proportion of alcohol to one part by weight of Quinoliiia Thickness of layel- of liquid = 15 mm. DIAGRAN XIII. C Y A N U R I C A O I D . Numbers indicating the position of cadmium lines of known wave length C+ p43. c+ - 9 7011 ?Z 17 I6 0. 50 100 I 50 2 00 Oidinates = The propoiition of liquid contJainmg one part by weight of tlie sut,staiice. Scale. 100 p&ts to Iialf an inch. 1 in 2,000 transmits all rays. Harrison b Ssns.Lith. S'Martins L3ne.VJ. ATKINSON AND TOSHIDA ON PEPPERPIIINT CAPIIPHOR ETC. 49 taining 1 part of substance in 2,000 of water is quite diactinic. The presence of the absorption-band proves I think that there is a union between the 6 atoms of nitrogen and carbon similar to that suggested by the graphic formula given aboye because otherwise the carbons and nitrogens would stand to each other in the same relation that obtains in hydrocyanic acid which is a perfectly diactinic substance.Jud,aing from the evidence afforded by the study of carbon compouiids, the account of which has been given in my former papers it is not a t all likely that the carbon-atoms in cyanuric acid can be only singly linked ; a t the same time the union of the atoms cannot be of so intimate a nature as that occurring in benzene because its actinic absorption is so much less powerful. It appears therefore that this substance possesses a nucleus with a compactness of structure intermediate between that of benzene hexchloride and that of benzene. Now as the atoms of the nucleus in the first body are singly linked if we accept the justifiable conclusion that those in cyanuric acid are alternately doubly linked it would appear probable that in benzene each carbon-atom is united with other three carbons since this is the only manner in which a more compact atomic linking than that expressed in the formulz of benzene hexchloride and cyanuric acid can be made consistent with the perfect symmetry of the compound or in other words with the perfect equality of function of the six carbon-atoms