Sept. 2, 1886] 
Mills, is converted into metastyrol, aldehyd into paraldehyd, the 
cyanates into cyanurates, and turpentine into metaterebenthene. 
At the critical points above referred to heat is liberated in 
especial abundance, and the bodies thus formed are known as 
polymers. If we could gradually cool down substances through 
a vast range of temperature, we should then probably discover a 
much greater number of such critical points, or points of 
multiple proportion, than we have been able to discover experl- 
mentally. 
The heat given out in the act of polymerisation naturally 
reverses to some extent the polymerisation itself, and so causes a 
partial return to the previous condition of things. This forward 
and backward movement, several times repeated, constitutes 
“periodicity.” Dr. Mills regards variable stars as instances, 
now in evidence, of the genesis of elementary bodies. c 
From a study of the classification of the elements, Dr. Mills is 
of opinion that the only known polymers of the primitive matter 
are arsenic, antimony, and perhaps erbium and osmium ; whilst 
zirconium, ruthenium, samarium, and platinum approximate to 
the positions of other polymers. Hence, from this genetic view, 
these elements may be described as products of successive poly- 
merisations. 
I must now call attention to a method of illustrating the 
periodic law, proposed by my friend Prof. Emerson Reynolds, 
of the University of Dublin, which will here assist us. Prof. 
Reynolds points out that in each period the general properties of 
the elements vary from one to another with approximate 
regularity until we reach the seventh member, which is in more 
or less striking contrast with the first element of the same period, 
as well as with the first of the next. Thus chlorine, the 
_ seventh member of Mendeléeff’s third period, contrasts sharply 
wr | 
both with sodium, the first member of the same series, and with 
potassium, the first member of the next series, whilst, on the 
other hand, sodium and potassium are closely analogous. The 
six elements whose atomic weights intervene between sodium 
and potassium vary in properties, step by step, until chlorine, 
the contrast to sodium, is reached. But from chlorine to 
potassium, the analogue of sodium, there is a change in pro- 
perties fev saléum. Further, such alternations of gradual and 
abrupt transitions are observed as the atomic weights increase. 
If we thus recognise a contrast in properties—more or less 
decided—between the first and the last members of each series, 
we can scarcely help admitting the existence of a point of mean 
variation within each system. In general, the fourth element of 
each series possesses the properties we might expect a transition- 
element to exhibit. If we examine a particular period—for in- 
stance, that one whose meso-element is silicon, we note :—/*¢rst, 
that the three elements of lower atomic weight than silicon, viz. 
sodium, magnesium, and aluminium, are distinctly electro-positive 
in character, while those of higher atomic weight, viz. phos- 
phorus, sulphur, and chlorine, are as distinctly edectro-megative. 
Throughout the best-known periods this remarkable subdivision 
is observable, although, as might be anticipated, the differences 
become less strongly marked as the atomic weights increase. 
Secondly, that the members above and below the meso-element 
fall into pairs of elements, which, while exhibiting certain 
analogies, are generally in more or less direct chemical contrast. 
Thus, in the silicon period we have—- 
Si iv 
+ Al” pr 
+ Mg” Se 
+Na’ Cl— 
This divison also happens, in many cases, to coincide with 
some characteristic valence of the contrasted elements. It is 
noteworthy, however, that the members on the electro-negative 
side exhibit the most marked tendency to vibration in atom-fixing 
power, so that valence alone is an untrustworthy guide to the 
probable position of an element in a period. 
Thus for the purpose of graphic trans!ation Prof. Reynolds 
considers that the fourth member of a period—silicon, for example 
—may be placed at the apex of a symmetrical curve, which shall 
represent, for that particular period, the direction in which the 
properties of the series of elements vary with rising atomic weights. 
In the drawing before you (Fig. 1) I have modified Prof. 
Reynolds’s diagram in one or two points. I have turned it the 
reverse way, as it is more convenient to’ start from the top and 
proceed downwards. I have represented the pendulous swing 
as gradually declining in amplitude, according to a mathe- 
NATURE 427 
matical law, and I have introduced another half-swing of the 
pendulum between cerium and lead, which not only renders the 
oscillations more symmetrical, but brings gold, mercury, thal- 
lium, lead, and bismuth on the side where they are in complete 
harmony with members of foregoing groups, instead of being out 
of harmony with them. ‘This modification has another ad- 
vantage, inasmuch as it leaves many gaps to be hereafter filled 
in with new elements just when the development of research is 
beginning to demand room for such expansion. 
I do not, however, wish to infer that the gaps in Mendeléeff’s 
table, and in this graphic representation of it, necessarily mean 
that there are elements actually existing to fill up the gaps ; 
these gaps may only mean that at the birth of the elements 
there was an easy potentiality of the formation of an element 
which would fit into the place. 
Following the curve from hydrogen downwards we find that 
the elements forming Mendeléeff’s eighth group are to be found 
near three of the ten nodal points. These bodies are “ inter- 
periodic,” both because their atomic weights exclude them from 
thesmall periods into which the other elements fall, and because 
their chemical relations with certain members of the adjacent 
periods show that they are probably interperiodic in the sense of 
being transitional. 
This eighth group is divided into the three triplets—iron, nickel, 
and cobalt; rhodium, ruthenium, and palladium; iridium, 
osmium, and platinum, The members of each triplet have often 
been regarded as modifications of one single form of matter. 
Notice how accurately the series of like bodies fits into this 
scheme. Beginning at the top, run the eye down analogous 
positions in each oscillation, taking either the electro-positive 
or electro-negative swings :— 
No Be ee Nae Micy  A Geis, 1Cl (Cc 
Ve Ca Ie Gu “Zn Ga Ge wAsieSe Br oTi 
Nb Sr Rb Ag €d In Sn Sb yTe LT Zr 
— Ba Cs — —_ 
Ta Au Hg Tl Pb Bi 
Notice, also, how orderly the metals discovered by spectrum 
analysis fit in their places—gallium, indium, and thallium. 
The symmetry of nearly all this series proclaims at once that 
we are working in the right direction. We can also learn much 
from the anomalies here visible. Look at the places marked with 
a circle: didymium, samarium, holmium, erbium, ytterbium, 
and thulium. Didymium cannot follow in order after the triad 
nitrogen, vanadium, columbium ; nor erbium follow phosphorus, 
arsenic, antimony; nor thulium follow chlorine, bromine, 
iodine; nor ytterbium follow potassium, rubidium, caesium. 
The inference to be drawn is that these bodies are out of place, 
owing to their atomic weights not having been correctly deter- 
mined —an inference which is strengthened by the knowledge 
that the elementary character of some of these bodies is more 
than doubtful, whilst the chemical attributes of most of them are 
unknown. 
The more I study the arrangement of this zigzag curve the more 
I am convinced that he who grasps the key will be permitted to 
unlock some of the deepest mysteries of creation. Let us imagine 
ifit is possible to get a glimpse of a few of the secrets here hidden. 
Let us picture the very beginnings of time, before geological ages, 
before the earth was thrown off from the central nucleus of 
molten fluid, before even the sun himself had consolidated from 
the original Aro¢y/e.' Let us still imagine that at this primal 
stage all was in an ultra-gaseous state, at a temperature incon- 
ceivably hotter? than anything now existing in the visible 
universe ; so high, indeed, that the chem‘cal atoms could not 
yet have been formed, being still far above their dissociation- 
point. In so far as frotyle is capable of radiating or reflecting 
light, this vast sea of incandescent mist, to an astronomer in a 
distant star, might have appeared as a nebula, showing in the 
spectroscope a few isolated lines, forecasts of hydrogen, carbon, 
and nitrogen spectra. 
1 We require a word, analogous to protoplasm, to express the idea of the 
original primal matter existing before the evolution of the chemical ele- 
ments. ‘The wordI hive ventured to use for this purpose is compounded of 
po (earlier than)and ®» (the stuff of which things are made). ‘The word is 
scarcely a new coinage, for 620 years ago Roger Bacon wrote in his “‘De 
Arte Chymiz’’:—‘' The elements are made out of #7, and every element 
is converted into the nature of another element.” 
2 T am constrained to use words expressive of high temperature; but I 
confess I am unable clearly to associate with Aroty/e the idea of hot or cold. 
Teniperature, radiation, and free cooling seem to require the periodic 
motions that take place in the chemical atoms; and the introduction of 
centres of periodic motion into srotye would constitute its being so far 
changed into chemical atoms. 
