282 



NA TURE 



[September 2, 1909 



Taking into account this principle, it is impossible to 

 arrange a long series of elements such as the rare earths 

 continuously in order of atomic weight, as they would be 

 brought into everv family in the table by such a proctdun- : 

 the difficultv iias'been got over by Brauner, who has pro- 

 posed to arrange a large number of the rare earths in a 

 single vertical series under barium. Biltz has made a 

 similar proposal. 



The principle had been advocated by me previously m 

 an article written for the " Encyclopa;dia Britannica." ' 



In the arrangement 1 have proposed, it is not only 

 assumed that there may be as many as sixteen vertical series 

 of elements of which the elements from hydrogen to oxygen 

 are initial terms, some series being at present unrepre- 

 sented, it is also suggested that groups of elements occur 

 in perhaps four of these series, numbers 4, 8, 12 and 16, the 

 largest being that of the so-called rare earths in series 8. 



The principle which is assumed to be in operation is that 

 which is so clearly manifest in the case of hydrocarbons : 

 successive vertical series of elements correspond to succes- 

 sive isologous series of homologous hydrocarbons. In the 

 case of the hydrocarbons, the passage from one isologous 

 series to another often takes place from a term several 

 places removed from the origin of the series— for example, 

 from benzene, C^H^, which may be regarded as primarily a 

 derivative of hexane to naphthalene, C,,Hj, which is not an 

 immediate derivative of benzene but of butylbenzene. It is 

 conceivable that at the genesis of the elements a process was 

 at work corresponding to that by which a hydrocarbon 

 such as naphthalene is derived from benzene, and by which 

 the former then serves in turn as the point of departure 

 for more complex hydrocarbons of other series. There is 

 no reason, from this' point of view, why progression should 

 not take place along a particular line and that terms should 

 exist in a series through which this line passes but below it 



for example, that antimony and iodine may bear a direct 



linear relationship, but that tellurium, insteadof being the 

 element in the progression series in the oxygen group, is a 

 homologue of greater weight. The same view may be taken 

 of selenium. In this way, it would be possible to maintain 

 selenium and tellurium in the oxygen-sulphur series, from 

 which they cannot well be separated, whilst retaining 

 Mendelejeff's conception of a genetic relationship along the 

 series. The only departure involved is in assuming that 

 instead of forming a single linear series ascending regu- 

 larly in spiral progression — a series which can, as it were, 

 be strung on a single spirally wound cord — the elements 

 closely simulate a series of homologous isologous hydro- 

 carbons. From this point of view, it is easy also to under- 

 stand that some vertical series are unrepresented. 



In discussing the chief attributes of the elements none is 

 so difficult to deal with as that of valency, using the term 

 in the broadest possible sense, not merely as indicative of 

 the number of units of affinity but as including the, at 

 present, all but incomprehensible problems of residual 

 affinity and elementary character. I discussed the subject 

 somewhat fuUv in my former address, dwelling especially 

 on the properties of negative elements and their power of 

 acting as linking agents ; this view has met with ample 

 confirmation in the interval, and will, I believe, be found to 

 be of wide application in the future. I have already referred 

 to the manner in which it is exemplified by silica. 



The greatest advance in the discussion of the problems 

 of valency in recent years is that made by Barlow and 

 Pope, as their method of treatment is one which applies to 

 solid substances — the correlation of structure with crystal- 

 line form which it effects promises to be of far-reaching 

 importance. 



Apart from hydrogen, carbon is the one element of 

 certain character, always acting as a tetrad — its affinities 

 may be only incompletely satisfied but they are always 

 exercised, it may be supposed, even in ethenoid and similar 

 compounds ; carbon monoxide apparently is the only excep- 

 tion to this rule, its relative inactivity being one of the 

 most puzzling enigmas of our science, especially as the oxide 

 becomes one of the most active of known substances when 

 only two atoms of hydrogen are added to it. Most other 

 elements (non-metallic) seem to vary in valency, the valency 

 beyond a certain minimum being dependent on the nature 

 1 Cf. Roy. Soc. Proc, 1902, vol. l.xx., pp. 36-94. 

 XO. 2079, VOL. 81] 



of the association. Of late years, attention has been directed 

 in particular to the quadrivalency of oxygen in many of its 

 compounds. 



The quadrivalency of sulphur in substances such as tri- 

 methylsulphonium iodide, .Mc-,SI, having been proved to 

 demonstration by the production of optically active com- 

 pounds of this type (Pope and Peachey), it can no longer 

 be supposed that in such cases we are dealing with com- 

 pounds in which the negative constituents of the parent 

 molecules are conjoined, e.g. Mel : SMe,. And yet we 

 must contemplate the existence of such compounds as pos- 

 sible — in the case of nitrogen, for example, as ammonia 

 must be supposed to form the compound NH, : OH, in 

 preference to the hydroxide NH^.OH, the latter being only 

 a very minor constituent, the former the major component 

 of the aqueous solution of the gas ; hydrogen chloride, on 

 the other hand, appears only to afford one product with 

 ammonia, viz. NH,.C1. The existence of such differences 

 affords clear proof in the case of the non-metallic elements 

 other than carbon that valency is not merely a variable but 

 also a reciprocal or dependent function. 



There is no reason to suppose that hydrogen ever acts 

 otherwise than as a simple monad ; and the behaviour of 

 the alkalies and alkaline earths in salts would seem to 

 justify the conclusion that they have no tendency to vary 

 in valency, were it not for the existence of well-defined non- 

 volatile hydrides of these metals which arc clearly sub- 

 stances of some degree of molecular complexity. Such com- 

 pounds are illustrations of the difficulties which surround 

 the subject. It has long been clear that the exhibition of 

 the higher valency by an element is a process of a different 

 order from that manifest when it exerts only its lower 

 proper valency measured in terms of positive radicles such 

 as H or C„fio„+i radicles. What that difference is we 

 are not able at present to decide — carbon (together with 

 silicon) differs from almost all other elements, especially 

 in combining with hydrogen and analogous radicles to the 

 extent of its maximum valency. 



The proposition I made in 1888 (Fhil Mag., Series V., 

 25, 21) that the valency lines should, in some cases, be 

 represented as passing through the atom, so that each is 

 capable of acting in two directions, is the only consistent 

 mode of e.xpressing varying valency which has been 

 devised, the only one, moreover, by which attention is 

 directed to the great difference. 



In many cases probably there has been a tendency to 

 exaggerate the valency value — in the case of chlorine, for 

 example, in assuming that it functions as a heptad in the 

 perchlorates. In this and many other instances, it suffices 

 to assume that the chlorine and oxygen atoms are united 

 in a closed ring, the chlorine functioning as a triad. Some 

 such explanation will doubtless be given of the structure 

 of the metallic ammonias and similar compounds. The 

 co-ordination values introduced by Werner serve only to 

 establish certain empirical relationships and are useful for 

 the purposes of classification. The perhaps more rational 

 plan of dealing with such compounds suggested by Abegg 

 has a similar value. 



It is to the advantage of the hypothesis formulated by 

 Barlow and Pope that the elements are represented as of 

 constant valency in so far as their relative volume spheres 

 of influence are concerned — the compound in which the 

 higher valency is manifest being derived from that of 

 lower valency by the opening out of the close packed arrange- 

 ment and the insertion of certain new elements ; but the 

 fact that in such cases the volume is altered not in one 

 direction alone in the crystalline structure but propor- 

 tionately in all directions would seem to show that the 

 volume sphere of atomic influence does actually change ; 

 the change is one, however, which affects all the atoms in 

 the complex proportionately. 



At present, unfortunately, our methods of treating the 

 problems of valency are such that we cannot in any way 

 give expression to the energy side of the phenomena. 



Of late there has been talk of electrons in this connection, 

 but what is said is little more than superficial paraphrase, 

 in the advanced scientific slang of the day, of the ideas 

 which have long been current. When, following Odiing,, 

 we represent valency by dashes written after the elementary 

 symbol, we give clear expression by means of a simple 

 convention to certain ideas that are well understood by all 



