to discuss the classification of the electro-primary species with a far greater degree of certainty than heretofore. One fact is clear
that a periodic arrangement was never more justified: formerly this involved placing them in the order of the magnitude of their
atomic weights and a sub-grouping under families; there was no
means of determining whether or no unassigned numbers were or
were not those of missing elements.
We are now on surer ground, as we may substitute atomic number the integer indicative of place in the evolutionary series for atomic weight : it were better perhaps to speak of this as the species number. In addition, we have to recognize the exis- tence, within some species at least, of sub-species or varieties dif- fering in atomic mass but in other respects, as a rule, so similar as to be indistinguishable except by special methods. These are the so-called isotopes. In the " isotopic elements," apparently, we are dealing, with substances which are closely related in electronic structure, corresponding to the terms in a series of homologous organic compounds.
No precise distinction can be drawn between the terms " chem- ical " and " physical " the chemist has availed himself so fully of physical methods that he has made them his own and has difficulty in giving any precise meaning to the expression " chem- ical property ": nevertheless, it has a clear connotation in his mind. The initial and second terms of the great series of paraffin hydrocarbons, methane (CH 4 ) and ethane (C 2 H 6 ), are, chem- ically speaking, identical; indeed this is true of the entire group, putting structure aside: the differences are mainly physical in mass, molecular magnitude, density, boiling point, etc.
In the accompanying table the electro-primaries are classified " periodically," in accordance with their " affinities." It is a striking fact that when arranged in the order of their " species numbers " they fall into eight great families: but progression is not along a continuous spiral. When the 25th place is reached, there is a precipitate fall through 26, 27 and 28; the series is then continued on the descending spiral until a similar precipitate fall takes place at 43 through 44, 45 and 46; again the progression is orderly until at 56 there is an astounding drop to 70; after a short interval, at 75 there is another fall similar to the first and second; during the remaining interval, progress is uniformly on the spiral. The species numbers, in some cases at least, serve but to indicate pockets in which homologues may be stored.
The view which was coming into favour in 1021 involves but an extension of the often discussed century-old hypothesis of Prout, that the elements are all multiples of hydrogen. It has long been held that the ratio of hydrogen to oxygen is 1-008:16, not i :i6; not only is this confirmed by Aston's observations but his measurements seem to justify the conclusion that integral values are to be assigned to all the electro-primaries other than hydrogen (cf. Aston, Jour. Chem. Soc., 1921). It is also a striking fact that no evidence of the existence of variants has been obtained in respect of species the determined atomic weight of which is not an integral value within the probable limits of error. Thus no evidence of the existence of variants is forth- coming in the case of helium, carbon, nitrogen, oxygen, fluorine, sodium, phosphorus and sulphur; but lithium, boron, neon, mag- nesium, silicon and chlorine are each to be regarded as a mixture of one or more varieties of a single species. The absolute depart- ure, however, from a whole number is no greater for lithium (6-94) than for sulphur (32-06), though in the latter case the difference is a much smaller proportion of the whole: sulphur is evidently a material to be further studied from this point of view.
The extent to which the accepted " atomic weight " differs from a whole number is no indication apparently of the values of which it is an integration. Thus lithium (6-94) appears to con- sist mainly of a constituent of mass 7 with only a small proportion of one of mass 6; but bromide (79-92) is a mixture in nearly equal proportions of variants of mass 79 and 81; still more remarkable is the composition of mercury (200-6), which appears to consist of 6 variants differing in mass from 197 to 204: krypton and xenon are of like complexity. Apparently the weights of separate species do not overlap. That phosphorus should be without variants and of mass 31 is remarkable, in view of the difference
of 16 units between many of the superposed terms in the first and second lines of the table it would have been less surprising had it proved to be a mixture of units of mass 30 and 32.
The table has other noteworthy features. The members of each of the three metallic triads, on the short precipice faces at the right of the table, might be placed in line and the arrange- ment would have the advantage of bringing out the homology between their corresponding successive terms Fe, Ru, Os; Co, Rh, Ir; Ni, Pd, Pt. The arrangement has the advantage of being parallel with that which must be adopted in the case of the great group of rare-earth primaries these cannot be entered across the table if this is, in any way, to be a picture of the homologies. manifest among the primaries. Maybe when this rare-earth group is fully studied, rhythmic variations such as are apparent in the three groups of triads will also be made obvious. It is noteworthy that, of the five presumed missing links in the record, two occur above the platinum triads in positions similar to that manganese has at the head of the iron triad; the third absentee may presumably also belong to the same great family and may well be a radio-active halogen. The fourth gap is in the rare- earth series; the fifth is in the radio-active region.
The classification of Cu, Ag, Au along with the alkali metals, still more that of the iron and platinum triads along with the halogens, may well seem peculiar; but in comparison with the as- tounding difference in properties between radium (a metal) and the emanation from it (a very volatile inert gas) the differences are not surprising. We may well be dealing with collaterals.
The arrangement is comparable with that of hydrocarbons in great families under the empirical symbols CnHja-K, C n H 2n , CnHfc, 2, or in a series and in that formed by benzene, naphtha- lene and anthracene. In the C n H 2n series, for example, the unsaturated ethenoid, hexene, is included together with the sat- urated hexamethylene (hexahydrobenzene). Benzene itself is usually non-valent but may act as a dyad, tetrad or hexad. Hither- to such variations appeared to be striking when observed among the " elements "; now that these are proved to be structurally complex, we may look forward to the explanation of their func- tional and physical peculiarities as consequences of structure, pre- cisely as in the case of organic compounds: why some are metals, others non-metals; why some metals are good and others bad conductors of electricity; why some species are coloured, others colourless; why some species, notwithstanding their great mass, are so remarkably volatile; for example, mercury and, most strange of all, the " emanation " from radium as in organic compounds variation in volatility corresponds very nearly with that of mass. Light may also be thrown upon the special prop- erties of compounds such as carbonic oxide and nitric oxide.
As the interest attaching to the correlation of function with inter- nal structure is now so great, attention may be called to a few special cases in which the structure may be supposed to alter. _
A phenomenon which has attracted great attention, in carbon compounds, is that of metameric (isodynamic) change, one of the earliest and most interesting cases observed being that of ethylic acetoacetate, which, according to circumstances, functions in two distinct ways, in correspondence with either the one or the other of the two formulae CHa.CO.COOEt or CH S : C(OH).COOEt. It has often been supposed that these are but two reciprocal forms and that the molecule is subject to constant, spontaneous, oscillatory change; the evidence is convincing, however, that like all other cases of chemical change, the alteration is the outcome of a more complex, reversible process conditioned by a conducting impurity. Thus CH s .CO.COOEt+HX1=;CN3.CX(OH).COOEt ^ CH 2 : C(OH).C- OOEt+HX. The two forms have been isolated and their stability shown to be a question of purity.
That alterations take place in internal electronic structure in simple compounds and even in " elementary " materials is already clear. Water affords one of the most striking examples.'in the sudden large increase in volume which it undergoes on conversion into ice. Assuming that the molecules are close packed in crystals, the increase cannot well be supposed to be due to their arrangement in any " open " form in ice. Can it be supposed that the electronic systems " expand " in some way? What is most striking is the suddenness with which the change takes place, at a definite temperature or nearly so for there is evidence that ice is present in water above the freezing point and also that ice contains water.
Evidence of " internal " structural change in elementary materials is to be found perhaps in the peculiar manner in which their heat-
