CHEMISTRY regarded as a constant or variable quantity—or perhaps it is better to say, in determining the maximum valency of an element. In substance this difficulty arises from the difficulty previously referred to of obtaining and applying a satisfactory standard of valency. A third, and perhaps, the greatest difficulty of all arises from the circumstance that valency is a dependent variable, the degree in which it is manifest depending on the conditions under which the element is placed. Apparently the only saturated molecules known to us at present, i.e., molecules destitute of all power of exercising an external influence, are those of gases molecules suc^ as helium an^ argon. Even hydrogen is to some extent unsaturated, to judge from its behaviour towards the metals platinum and palladium. On account of its properties generally, however, and of the extreme difficulty with which it is compelled to assume the liquid state, it may be said to follow helium closely; consequently it may be supposed that the hydrogen atom is a strictly univalent atom, and that when it is associated either with itself or with another radicle its affinity becomes all but neutralized. Tested by their power of combining either with hydrogen or methyl, the elements generally (in so far as these tests are applicable) may be resolved into the four groups of which the four hydrides HC1, H20, H3N, H4C are types. But, except in the last of these groups, the compounds produced are unsaturated. Hydrogen chloride and ammonia very readily liquefy in presence of water, in which they are intensely soluble; and water, considering its composition in comparison with that of hydrogen chloride and ammonia, and the fact that the analogous sulphur compound is gaseous and but slightly soluble in water, is altogether remarkable on account of its high boiling-point, its high capacity for heat, and its extraordinary power of dissolving oxygenated compounds generally and also chlorides, bromides and iodides. Apparently the only plausible—if not the only possible—interpretation of these peculiarities is that the three hydrides in question are all eminently unsaturated compounds. Assuming this to be the case, their attractive power may be regarded as exercised either (1) by the molecules as wholes; or (2) by their constituents independently, although not necessarily to equal extents; or (3) by one or other of the constituents. The study of carbon compounds appears to justify the belief that in very many cases the activities which they display proceed rather from centres than from the molecules .as wholes. Taking into account the estimate we are able to form on general grounds of the properties of hydrogen, on the one hand, and on the other of elements such 1 as c Affinity hl°rine> oxygen and nitrogen, it therefore seems probable that in the hydrides of these elements the attractive power is exercised entirely by the element which is associated with the hydrogen; that chlorine in hydrogen chloride, oxygen in water, and nitrogen in ammonia are each possessed of considerable residual affinity. It is to be noted that this residual affinity is by no means indiscriminately exercised. To put the point in conventional terms, whereas oxygen is prepared to behave as a dyad to all comers—to associate with them by means of its two affinities—fluorine alone being scorned, it is prepared to exhibit residual affinity only in presence of a select few; and most interesting of all, it seems to accord preferential treatment to itself and its near allies, the halogens. The behaviour of oxygen is all the more remarkable when it is contrasted with that of sulphur, the element with which it is most nearly allied; as tested by its power of combining with other substances, in comparison with Avater, sulphuretted hydrogen
717
is absolutely inert, and yet sulphur commonly ranks as an element of higher valency than oxygen. The peculiarity of oxygen referred to is specially manifest in the properties of ordinary water; the production of a substance so remarkable among hydrides, it may almost be said, can only be explained by the assumption that the residual affinity of the oxygen in the molecule of the dihydride OH2 compels molecule to unite with molecule. To Avhat extent complication proceeds we cannot yet say—the estimates vary; but of one thing we are certain, that placid as the fluid is in appearance, it must be pictured as at all times the seat of an extraordinary turmoil of dissociative changes. The view that oxygen is not to be regarded as capable of acting only as a dyad, but that it can also manifest a higher valency, has been definitely advocated in various quarters of late years—notably by Meldola, by Briihl in explanation of the properties of hydrogen peroxide, and by Collie (cf. Tra?is. Chem. Soc. 1899, p. 715); but in all these cases the oxygen has been pictured as acting as a tetrad, not merely as displaying residual affinity. The argument which leads us to regard the oxygen in water and many other compounds as possessed of residual affinity would lead us also to conclude that this is also manifested—although in very different degrees—by the chlorine in hydrogen chloride and by the nitrogen in ammonia. The question then arises whether the union of these two compounds, when ammonium chloride is formed, takes place in such a way that the elements become redistributed and separately centred around the nitrogen atom so that this acts as a pentad; or whether merely the residual affinity of each molecule conies into operation. This is the old historic question which was disStructure cussed by Kekule and his contemporaries in the of am’sixties, and decided by him in the latter sense, monium but which the general verdict of chemists has comdecided in favour of the doctrine of varying pounds. valency. In the interval numerous facts have come to light which bear on the problem although they do not solve it. Baker’s recent observation, already referred to, may be held to meet Kekule’s objection that ammonium chloride cannot be volatilized unchanged; but as the compound at once breaks down in presence of a mere trace of water, it is a question whether the objection does not retain its force. An argument of real weight in favour of nitrogen being regarded as a true pentad in the ammonium compounds has been furnished by the discovery of derivatives of the hypothetical amine oxide HgNO, corresponding Avith those which have long been known of the equally hypothetical phosphine oxide H3PO; their existence was first recognized by Wernick and Wolffenstein in 1898. Such compounds are formed on oxidizing tertiary amines—dimethylaniline, for example —by means of hydrogen peroxide, and by the action of methyl and ethyl iodides on hydroxylamine. They are highly basic substances, which combine directly with acids (cf. Bamberger, Dunstan and Goulding, Trans. Chem,. Soc. 1899, p. 792). On the assumption that the ammonium salts are molecular compounds, a quaternary ammonium derivative formed by uniting an amine Nabc with the iodide Id should differ from one produced from the amine Nabd and the iodide Ic. Experiments to test this assumption were made in 1875 by Meyer and Lecco, who on the one hand combined ethyldimethylamine with ethyl iodide, and on the other methyldiethylamine with methyl iodide; finding the products to have identical properties, they expressed the opinion that the ammonium derivatives were to be regarded as atomic compounds. Similar experiments made by Ladenburg at about the same date led to inconclusive
