718
CHEMISTRY
results. Lessen pointed out at the time that Meyer’s results did not necessarily justify the interpretation he gave to them, inasmuch as the differences likely to arise in such a case might be so slight as to escape detection, particularly as crystallographic identity—on which Meyer and Lecco had placed special reliance—was, as a matter of fact, to be expected. At a later date, owing to the introduction of stereo-chemical considerations, the problem assumed a different aspect, and it became clear that the existence of isomeric asymmetric optically-active ammonium compounds was to be expected. The existence of such compounds was rendered probable, but not proved, by observations made by Le Bel in 1891, and has lately been established beyond question by Pope and Peachey {Trans. Chem. Soc. 1899, p. 1127), who have succeeded in resolving benzylphenylallylmethylammonium iodide into two powerfully optically-active isomerides. Although the problem is not solved by this important discovery, as it is conceivable that a molecular compound might also exist in isomeric optically-active forms, it is brought nearer solution, as a possible method of determining the structure of ammonium derivatives is at last placed in our hands. Yet it is evident that there may, and indeed will, be exceptional difficulty in arriving at a conclusion, and that the utmost caution must be used in interpreting even negative results, as the possibility must be foreseen of interchanges of position readily taking place within the molecule. Le Bel has laid particular emphasis on this, but from a point of view which is open to question. In the case of compounds of non - metallic elements generally with halogens, the halogen is invariably displaced with great readiness by the use of alkalies. It is therefore highly remarkable that the compound N(CHg)4I, for example, should not be in the least affected by even the strongest caustic alkali, and that the halogen in such compounds behaves much as does that in alkylic haloids. This would seem to be a strong argument in favour of the view that the halogen is not simply associated with the nitrogen, but retained in combination with the hydrocarbon radicle. An even stronger case may be based on the similar behaviour of the phosphonium haloids, seeing that the compounds of phosphorus with halogens are so readily attacked even by water. It may be, however, that it is an effective answer to this argument to say that such behaviour affords proof of the error of the current view which asserts that when an alkali acts on ammonium haloids the halogen is attacked by the metal; in reality hydrogen is attacked by hydroxyl, the halogen only indirectly “falling a victim” to the metal, so that in the absence of hydrogen there can be little or no action. The case is one which well illustrates the difficulty of interpreting facts. A class of facts which more than any other favours the view that the ammonium derivatives are molecular compounds may now be referred to. The basic qualities of ammonia, it is well known, are invariably diminished by the introduction of other radicles in place of hydrogen, and the extraordinary variations in basicity manifest amongst its derivatives cannot fail to arrest attention. Aniline, for example, is a very weak base in comparison with ammonia. But the chloranilines are still weaker, trichloraniline being so feeble that although soluble in acids it is precipitated on adding water. A more remarkable case is that of phenylhydrazine, C6H5.lSriI.lSriI2, which yields only monacid salts; and hydrazine itself, H2N.]SrH2, although capable of forming diacid salts, has a marked" tendency to combine with only a single equivalent of acid. A large number of compounds are now known in which several nitrogen atoms are united in a ring, frequently in
conjunction with carbon. The nitrogen atoms seem to be united within the ring, on the one side by one and on the other by two affinities, so that the three affinities are entirely engaged within the ring. It is remarkable that the nitrogen in such compounds is in many cases entirely destitute of basic properties, and in fact inactive to an extraordinary degree. The same may be said of the nitrogen in compounds such as azobenzene, as the few compounds these form with acids are decomposed by water. But the most striking example of inactivity is afforded by the remarkable compound diazoimide, N^.NII —which may or may not have a ring structure; not only is this compound destitute of basic properties, but it actually behaves as a weak acid. Therefore it cannot be asserted that nitrogen is a “basic” element. It is basic, as a rule, only in the aminic condition ; the “ ammonium function ” is a strictly dependent function and highly discriminative. The determination of the valency of the halogen elements is of primary importance, as we are practically forced to rely upon them as standards in many cases. They commonly rank as well-defined monads, Valency but as it is clear that iodine may exercise triadic °^Jiggens functions, the remaining halogens can scarcely be denied the rank of potential triads. The question to be settled is whether any effects they may produce beyond the functions of a true monad are to be ascribed to residual affinity, or whether they are actual triads. At present the evidence is inferential, as there are no valid methods available of determining the structure of the compounds coming under consideration — for example, phosphorus pentachloride and the double chlorides generally. In all such instances we have to deal with at least two unknowns. Chemists have long been in the habit of dividing double chlorides into two classes—one comprising mere “double salts,” of which very little proper notice was taken ; the other compounds of a somewhat higher order of stability, such as theplatinichloridesandaurichlorides, which have been regarded by many as atomic compounds, and in conjunction with compounds such as 0s04 have been quoted as proof that platinum may function as an octad. Obviously, in the present state of our knowledge, the only fact of which they afford proof is that chlorides may combine. The supposition that the constituents in double chlorides generally may be held together through the agency of the chlorine atoms is one which cannot be dismissed from consideration ; we do not hesitate to explain the formation of double cyanides, which are analogous compounds, by such an assumption, and, moreover, the existence of compounds such as the chlorates and perchlorates, and the extraordinary inertness of a part of the halogen in some of the cobaltamine compounds, almost compel belief in the existence of stable rings containing halogen atoms in which the halogen acts as a linking element. The application of this view to fluorine is of particular importance. There can be no question that this element is possessed of residual affinity in a far higher degree than any of the allied elements. The vapour of hydrogen fluoride near to its boiling-point (19°) consists of complex molecules (Mallet), which, as Thorpe and Hambly have shown, are only completely resolved into simple molecules at a temperature near to the boiling-point of water {Trans. Chem. Soc. 1889, p. 163). Double fluorides of remarkable stability are formed with exceptional ease. The properties of some of these are worthy of special consideration— notably hydrogen silicon fluoride, a compound analogous in composition to hydrochloroplatinic acid. In consequence of its relation to carbon the valency of silicon may be regarded as established. If silicon be a tetrad, we are, however, almost compelled to assume that in hydrofluosilicic
