744
CHEMISTRY
which they tend to take up in the unrestrained carbon atom, and it may be supposed that the extent to which the affinities of the paired carbon atoms are deflected varies according to the nature of their associates. But the deflection would vary only within narrow limits ; consequently, the limits within which the ethenoid “ constant ” may be expected to vary are narrow. In the case of benzenoid compounds a much wider variation may be possible. In discussing their chemical properties, it has already been pointed out that the centric system must be regarded as a plastic group within which considerable variation of the affinity relationships may take place. Physical properties tell the same tale. To give but one or two examples. No property affords a more sensitive test of changes in structure than the refractive power. If the refractive power of the Cg group in benzene be calculated by deducting from the molecular refractive power the value of six hydrogen atoms—hydrogen being supposed to have a constant value in all compounds—and the difference be divided by 6, the value for the carbon atom is very nearly 6. If a similar calculation be made for the methylated homologues of benzene, assigning to CH3 the value that it has in other series, the value for the carbon atom (in the C6 group) is found to increase from 6 in benzene to 6*15 in trimethylbenzene (mesitylene). If a similar calculation be made for aniline, assuming that NIP2 has the value assigned to it in paraffinoid amines (7'7 = 5-l + 2.1'3), a still higher value is obtained for carbon, viz. 6‘3. Whether these variations are to be ascribed solely to variations in the centric system, or whether the associated I'adicles also vary, it is at present impossible to decide ; the question has already been raised in discussing the occurrence of chemical change in the case of benzenoid compounds, and it has been pointed out that there is some reason to suppose that even the CH3 group has special properties in benzenoid hydrocarbons, but the argument is by no means conclusive. With regard to the alteration in physical properties consequent on the introduction of other elements into hydrocarbons, oxygen exercises a greater effect Special perhaps than any other, and one which is espe^of oxygen. cially noticeable in the case of properties which are mainly “ intermolecular ” effects. Thus the introduction of hydroxyl into the paraffins has the effect of converting very slightly viscous compounds of low boiling-point into relatively highly viscous compounds boiling at temperatures far above those at which the parent hydrocarbons are vaporized. There can be little if any doubt that the peculiar physical properties of oxygen compounds are to be correlated with their exceptional chemical activity—that the residual affinity which is possessed by oxygen in most of its compounds in so high a degree, and which conditions the preliminary associations determinative of chemical change, is also operative in causing the fundamental molecules to cling together and form complexes to an extent not met with in compounds of other elements, so that the special peculiarities of oxygen compounds may be ascribed to the oxygen. Whether the properties of the paraffinoid radicle are in any way modified by association with oxygen or other elements is a difficult question to decide. In discussing it geometrical considerations must needs be taken into account. In the normal paraffins and in the corresponding primary alcohols, the carbon atoms are arranged in a branchless chain, but in the secondary alcohols the hydrocarbon radicles have two, and in the tertiary alcohols three branches. The primary alcohols may, therefore, be pictured as simple filaments progressing, it may be, with a wave-like motion. As the hydrocarbon radicle is composed of saturated elements, it may be regarded as practically indifferent to all external attractions, and the
activity of the compound may be located at the oxygenated extremity or head, the hydrocarbon radicle being supposed to oscillate about a mean position relatively to the OH group. In the secondary and tertiary alcohols, in which the filaments are branched, the hydrocarbon radicles are brought into closer proximity with the OH group, and they will also tend to oscillate on one side of the mean position taken up by the primary radicle, unless the constituent groups are of such dimensions as to balance one another. Such differences in structure might alone condition a materially different behaviour of the molecules towards each other, which would be apparent in the viscosity, density, and boiling-point relationships. Optical properties, on the other hand, which are mainly conditioned by intermolecular structure, would be but slightly affected; and chemical properties should not be materially altered. As a matter of fact, considerable differences are noticed on contrasting the chemical behaviour of the lower with that of the higher terms of the series in the case not only of the alcohols, but also of the ketones and acids derived from the paraffins. As there is no reason to suppose that the hydrocarbon radicle undergoes any modification, it may be assumed that the oxygen is the inconstant element; but as only the lowest terms differ to any marked extent, it would seem that when a certain point is reached variation can no longer take place. To draw a rough parallel, the oxygen appears to behave much as if it were connected with the hydrocarbon radicle by a spring which cannot be extended beyond a certain limit; in the lower terms of the series the spring is not fully extended, but the limit is soon reached. Differences such as are observed between phenol, for example, and the paraffinoid alcohol, hexylic alcohol, are of interest from this point of view. Although of lower molecular weight, phenol is a crystalline solid at temperatures at which the alcohol is liquid, and as a liquid it is more viscid than the alcohol. Phenol also manifests weak acid properties, which may be regarded as evidence that the oxygen has some power of attracting bases. It has been shown also that phenol exists in many solutions as a double molecule. On all grounds, therefore, it seems probable that the oxygen in phenol has more residual affinity than that in hexylic alcohol. The great increase in viscosity which attends the introduction of several hydroxyl groups is also corroborative of the argument here used. The extreme viscosity of polyhydric alcohols, such as glycerol, is possibly a consequence of the arrangement of the three oxygen atoms in one plane in such a way as to constitute an “ oxygen surface.” Exact determinations of the viscosity relationships of isomeric asymmetric polyhydric alcohols of known configuration would be of great value as testing a point of this kind. Variations in the physical effects produced by oxygen are particularly apparent in compounds in which oxygen is associated with carbon by two affinities, much in the way that carbon is associated with carbon in ethenoids. Speaking generally, oxygen apparently behaves as though its two affinities took the form of a looped filament, the two ends of which represent its ordinary affinities, and the loop the dependent or residual affinity. The variations it exhibits may then be pictured as “variations in the size of the loop, no secondary point of attachment remaining when it is reduced to a mere knot. A factor of the highest importance, which in many cases conditions striking physical peculiarities, remains to be referred to, viz., that of co-operative action. If the ^ heat of formation of chlorinated paraffinoid com- ‘o‘£°’tjve pounds be calculated from the heats of combus- effect.” tion, it appears, as J. Thomsen has pointed out,. that on an average about 13,500 gram. deg. C. units of heat are developed in the fixation of a single atomic proportion ol
