304 



NATURE 



{yan, 26, 1888 



case on record, however, of any haloid compound other than an 

 alkylic compound behaving in this manner ; the chlorides, 

 bromides, and iodides of every element except carbon are almost 

 at once converted into hydroxides by such treatment, and a 

 nitrogen iodide would surely be acted on. The behaviour of 

 the iodine is much more nearly that of iodine in methyl iodide, 

 and, it may be said, exactly that of the iodine in iodobenzene ; 

 indeed it would seem that in the alkyl-ammonium haloid com- 

 pounds the halogen is always less easily displaced by the action 

 of alkalies than it is in the parent haloid alkylic compound. 



The remarkable resemblance of the tetra-substituted am- 

 monium hydroxides to potassium hydroxide has led to their 

 being regarded as in every respect analogous to this latter, 

 and would appear to preclude the idea that they are molecular 

 compounds of an alcohol with an ammonium derivative. But 

 attentive consideration of their properties will suffice, I think, 

 to show that the apparent discreiJancies are not only explic- 

 able, but that they actually support the molecular compound 

 hypothesis. Thus it might be said to be improbable that tetra- 

 methylammonium hydroxide should behave as a powerful base, 

 and have the same heat of neutralization as potassium hydroxide, 

 if methyl-alcohol were one of its proximate constituents ; but it 

 is to be remembered that the salt which results from the action 

 of an acid on methyl-alcohol is liable to suffer reconversion into 

 the alcohol by the action of the water produced in the inter- 

 change ; also that in many cases the methyl salt is insoluble in 

 water, or nearly so. The heat developed on neutralizing methyl - 

 alcohol therefore falls far short in amount of that which would 

 be evolved if the interchange were complete, and if the product 

 were capable of interacting with water, and perhaps also with 

 itself in the way that apparently is possible in the case of metallic 

 salts. In the case of the tetramethylammonium hydroxide, the 

 action of acids is total as the change is irreversible, or almost 

 so, under the conditions which obtain during the formation of 

 the salt, just as in the case of the conversion of potassium 

 hydroxide into a salt ; moreover, the product is easily soluble, 

 even when acids like muriatic are used. Why the methyl- 

 alcohol, or other methyl derivative, retained in the ammonium 

 compound behaves so differently as compared with the unasso- 

 ciated methyl derivative, is a question which, for the present, 

 we must be content to put aside unanswered. I am also of 

 opinion that in discussing their constitution no particular weight 

 can be attached to the mode in which the tetralkylic ammonium 

 hydroxides undergo decomposition when heated, as the products 

 in some cases are an amine and an alcohol, but in others an 

 olefine and water, instead of an alcohol ; in the ease of the 

 phosphonium salts the diversity is still greater (Chem. Soc. 

 Proceedings, 1886, p. 164). That amines may act as "de- 

 hydrating" agents in the manner required if the molecular 

 compound hypothesis be adopted, appears by no means im- 

 probable. 



What is here stated of the tetramethyl compounds is true 

 of tetralkylic ammonium haloid compounds generally, in the 

 sense that they are all less readily acted on by alkalies than are 

 the parent alkylic haloid compounds ; but just as these latter 

 are more readily attacked by alkalies and other agents the more 

 complex the alkyl, so are the tetralkyl ammonium compounds; 

 in no case, however, do they manifest a reactivity at all com- 

 parable with that of simple metallic or non-metallic haloid 

 compounds—always excepting those of carbon. 



The argument used above would apply equally to the phos- 

 phonium and sulphine compounds ; indeed with greater force. 



In many other respects the behaviour of nitrogen in aminic 

 compounds is altogether peculiar and irreconcilable with the 

 assumption of pentadicity. Thus it is commonly pointed out 

 that the basic properties of aniline, for example, become lessened 

 and ultimately almost annulled by the introduction of chlorine 

 or bromine into the phenyl radicle ; and that acetamide, 

 C2H3O . NHg, and other similar compounds formed by the 

 introduction of acid radicles into ammonia are all but destitute 

 of basic properties ; the power to form ammonium compounds, 

 therefore, is not a simple function of the nitrogen atom, but is 

 largely dependent on the nature of the radicles associated with 

 the nitrogen atom. Other illustrations are afforded by the 

 hydrazines. Thus phenyl-hydrazine, CgHs.NH.NHj, al- 

 though it contains two atoms of (triad) nitrogen, forms with 

 hydrogen chloride the compound CgHj. N.2H3 . HCl, which 

 crystallizes unchanged from fuming muriatic acid, in which, 

 moreover, it is almost insoluble. Ethyl-hydrazine, however, 

 forms a dichlorhydride, CgHg . NgHg . 2HCI, but on evapor- 

 ating the aqueous solution of this salt a monochlorhydride is 



obtained ; and unsymmetric diethyl-hydrazine, (C2Hg)2N . NHg, 

 is a monobase like phenyl hydrazine. 



Hence it may well be argued that we have no reason to 

 assume that nitrogen is pentad in the ammonium compounds, 

 or phosphorus pentad in the phosphoniucn compounds, or 

 sulphur tetrad ^ in the sulphine compounds ; but that these are 

 all to be reckoned as molecular compounds. 



What then is the valency of the elements in question ? and 

 what is a molecular compound ? 



In answer to the first of these questions, the proposition 

 may be advanced that gasefiable hydrogen compounds are 

 the only compounds available for the direct determination of 

 valency, and that the valency of an element — the number of 

 unit charges necessarily associated with its atom — is given by 

 the number of hydrogen atoms combined with the single atom 

 of the element in its gasefiable hydride.* In cases where such 

 hydrides are unknown, the determination of valency is very 

 difficult ; it can be but provisionally effected, and only by most 

 carefully weighing all the evidence relating to the constitution of 

 the compounds available for discussion. 



But if it be granted, for example, that nitrogen is a triad, 

 and that iodine is a monad, how are we to explain the fact that 

 the methyl compounds of these two elements unite to form 

 so well characterized a molecular compound as tetramethyl- 

 ammonium iodide ? how are such molecular compounds con- 

 stituted ? My own view has long been that the nitrogen and 

 iodine in such a case are both possessed of a certain amount of 

 residual affinity ; and I would define a molecular compound as 

 one formed by the coalescence of two or more molecules, un- 

 attended by redistribution of the constituent radicles, and in 

 which the integrant molecules are united by residual affinities. 

 In other words, the unit charge must be capable in certain cases 

 of directly promoting the association, not merely of two, but of 

 at least three, atoms. ' To put this hypothesis in terms which 

 cannot be misunderstood, let unit valency or charge be repre- 

 sented by a unit line, and further be it supposed that the charge 

 penetrates the atom, then the atom with its unit charge may be 

 represented thus : — 



f 



e 



f 



i.e. the unit charge may be held to consist of three portions, 

 the buried portion «, and the free portions /-f-/'. The facts, 

 as they present themselves to me, also appear to necessitate 

 the assumption that, in the case of different elements, the charge 

 penetrates the atom — and in the case of some polyad atoms, 

 different directions in the atom — with varying degrees of freedom.^ 

 The union of two atoms may then be pictured as an overlapping 

 of the unit lines. If the atoms are freely penetrated by their 

 charges, each atom may tend to move out to the end of the line, 

 leaving either no portion, or but a very small portion, free ; a 

 conception of this order would appear to apply in the case of 

 hydrogen, and may be represented thus : — 



H. 



H, 



But if the atom be not easily penetrated by its charge, it will 

 not move out to the end of its line, and the resulting com- 

 pound molecule will possess more or less " residual affinity ; " 

 this conception would appear to apply to the non-metals 

 generally, and to some of the metals ; it may be illustrated 

 thus : — 



e 



^ 



v_y 



e 



JZh. 



KJ 



II. 



CI. 



HCl. 



I have thought it permissible to state my views in this form 

 merely in order to advance the study of molecular compounds 



' Probably one of the strongest arguments in favour of the conclusion tha* 

 sulphur is divalent may be based on its inactivity in the cbsed-chain com- 

 pound thiophen, which does not unite wUh methyl iodide, nor does the 

 sulphur in it or its homologues permit of oxidation in the manner that is 

 characteristic of the element m thioethers. 



^ If this be granted, it follows that the maximum number of charges 

 which an ato.n can carry is four ; in ottier words, t.iat the posiibla maximum 

 valency is attained in the case of carbon. 



3 This is practically but a modification of Helmholtz's statement that 

 " the phenomena are the same as if equivalents of positive and negative 

 electricity were attracted by different atoms, and perhaps also by the 

 different v.-ilues of affinity belonging to the same atom, with different force. 



