726
CHEMISTRY
carbon, for example, is in harmony, he considers, with the circumstance that the greatest distortion must be involved in its formation, as if deflected into parallelism each valency will be drawn out of its position through 109° 28'. The values in other cases are— CH2 ch2 / / ch2 ch2—ch2 h2c ch2 h2c ch9 / I I I III h2c—ch2 ch2—ch2 h2c—ch2 h2c ch2 / ch2 24° 44' 9° 34' 0° 44' - 5° 16' The general behaviour of the several types of hydrocarbons is certainly in accordance with this conception, and it is a remarkable fact that when benzene, apparently the most stable of all cycloids, is reduced by heating with hydrogen iodide, it is not, as was long supposed to be the case, converted into hexamethylene, but into methylpentamethylene (cf. MarkownikofF, Liebig’s Annalen, 1898, 302,1); and by no means a few other cases of the conversion of six-carbon rings into five-carbon rings have been recorded in recent years. Similar considerations will apply to rings containing other polyad elements besides carbon, but it is impossible to discuss these cases at present on account of the inadequacy of the evidence as to the “ valency ” of such elements. As an illustration it may be pointed out, however, that in the case of the two known types of lactones— the half-acid, half-alcoholic ethers formed from hydroxyacids by internal condensation—the y-lactones, which contain four carbon atoms and one oxygen atom in the ring, are more readily formed and more stable (less readily hydrolysed) than the S-lactones, which contain one oxygen and five carbon atoms in the ring. That the number of atoms which can be associated in a ring by single affinities is limited there can be no doubt, but there is not yet sufficient evidence to show where the limit must be placed. Baeyer has suggested that his hypothesis may also be applied to explain the instability of acetylene and its derivatives, and the still greater instability of the polyacetylene compounds. The objections to all such “ explanations ” is their purely “ mechanical ” character, and in using them it is necessary to bear this in mind. More especially must it be remembered that what is needed is not so much an explanation of instability as of the greater activity of certain types of compound. This dynamical side of the problem has hitherto been practically entirely left out of account, it being impossible to deal with it with the means at present at our command. In the case of closed chain hydrocarbons in which the carbon atoms are not fully saturated with hydrogen as they are in the poly-methylenes, conditions prevail Kekule’s which have the effect of introducing limitations, 80 apparently there are but few forms possible. Of these the benzene ring, composed of six atoms of carbon, is the most stable, this ring, or collocations of such rings, being produced particularly in changes occurring at very high temperatures—as in the formation of coal-tar for example. The question of the intimate structure of benzene is therefore of special interest. It has been a subject of constant discussion since Kekule in 1865 enunciated the conception that the six carbon atoms are directly united together in a closed chain, each of them being also united with a single hydrogen atom, so that the cycloid consists of six CH groups, and is therefore a symmetrical structure. This has been proved to be the case in the most absolute manner possible. Each carbon atom in the ring is, ex hypothesi, associated with neighbour-
formula.
ing carbon atoms in such a manner that at least two of its affinities are engaged—one on either side—and a third affinity of each atom is in combination with hydrogen. To account for the disposal of the fourth affinity Kekule suggested as most probable that the carbon atoms were united alternately by one and two affinities, and to the present day this hypothesis is regarded with favour. Two great objections have been urged against it, however. The first is that di-derivatives in which the radicles are associated with contiguous carbon atoms—so-called orthocompounds—should differ according as these carbon atoms are associated by a single or by two affinities. No such difference has been observed. This objection has always been felt to be of the greatest weight. To dispose of it, Kekule suggested that the distribution of the affinities within the ring must be supposed to vary, so that at one moment the configuration will be that represented by the first, and at the next that represented by the second of the following figures:— A /C HC CB HC CB I II HC CH c/ H H On this assumption the compound AB is a mixture in equal proportions of substances which cannot be separately dealt with. The explanation has never been regarded as satisfactory, no independent evidence having been obtained which would make it necessary to admit that such an oscillation of the affinities can and does take place. On the contrary, the results of modern inquiry appear uniformly to favour the view that when such changes occur they are not the result of mere internal oscillation, but are occasioned by catalytic influences, and take place only because the molecule becomes included with other bodies in a system. An argument of far greater weight against the ethenoid formula of benzene is furnished by the behaviour of benzenoid compounds generally in comparison with that of ethenoid compounds. Ethylene, and practically all of its derivatives, are eminently unsaturated compounds, combining—under appropriate conditions—with bromine, the halhydrides, etc., with the greatest readiness. Benzene, therefore, if it contain these ethenoid linkages, should be even more active than ethylene. But such is not the case, the behaviour of benzenoid compounds generally being rather that of saturated than of unsaturated compounds, as in most cases they afford substitution derivatives. This argument becomes convincing when dihydro- and tetrahydro-benzene are taken into account. These behave as ethenoid compounds, being prone to change and combining as readily as the corresponding open chain derivatives do with various other substances. It cannot, therefore, be argued that the association of several ethenoid linkages in a ring has the effect of reducing their activity. The diagonal formula advocated by Claus, in which the affinities are represented as united in pairs across the ring, although not open to the criticisms which apply to Kekule s formula, has met with but little sup- other port; there would seem, in fact, to be an instinc- formula.-. tive objection to the assumption that affinities can act across one another. This difficulty is got over in the “ centric ” formula proposed by the writer (Phil. Mag. ; Trans. Chem. Soc. 1887, 264), and shortly afterwards by v. Baeyer, in which the fourth affinity of each carbon atom is assumed to be only directed towards the centre of the ring—no one of the carbon atoms being regarded as
