Sept. 2, 1886] 
In this way it is conceivable that the succesion of events which 
gave us such groups as platinum, osmium, and iridium—palla- 
dium, ruthenium, and rhodium—iron, nickel, and cobalt, if the 
operation of genesis had been greatly more prolonged, would 
have resulted in the birth of only one element of these groups. 
It is also probable that, by a much more rapid rate of cooling, 
elements would originate even more closely related than are 
nickel and cobalt, and thus we should have formed the nearly 
allied elements of the cerium, yttrium, and similar groups; in 
fact, the minerals of the class of samarskite and gadolinite may 
be regarded as the cosmical lumber-room where the elements in 
a state of arrested development—the unconnecte 
of inorganic Darwinism—are finally aggregated. 
missing links 
I have said that the original frotyZe contained within itself the 
potentiality of all possible atomic weights. It may well be 
questioned whether there is an absolute uniformity in the mass of 
every ultimate atom of the same chemical element. Probably 
our atomic weights merely represent a mean value around 
which the actual atomic weights of the atoms vary within certain 
narrow limits. 
Each well-defined element represents a platform of stability 
In the first accreting 
connected by ladders of unstable bodies. 
NATURE 
429 
together of the primitive stuff the smallest atoms would form, 
then these would join together to form larger groups, the gulf 
across from one stage to another would be gradually bridged over, 
and the stable element appropriate to that stage would absorb, 
as it were, the unstable rungs of the ladder which led up to it. 
I conceive, therefore, that when we say the atomic weight of, for 
instance, calcium is 40, we really express the fact that, while the 
majority of calcium atoms have an actual atomic weight of 40, 
there are not a few which are represented by 39 or 41, a less 
number by 38 or 42, and so on. We are here reminded of 
Newton’s ‘‘ old worn particles.” 
Is it not possible, or even feasible, that these heavier and lighter 
atoms may have been in some cases subsequently sorted out bya 
process resembling chemical fractionation? This sorting out 
may have taken place in part while atomic matter was condens- 
ing from the primal state of intense ignition, but also it may have 
been partly effected in geological ages by successive solutions and 
re-precipitations of the various earths. 
This may seem an audacious speculation, but I do not think it 
is beyond the power of chemistry to test its feasibility. An in- 
vestigation on which I have been occupied for several years has 
yielded results which to me appear apposite to the question, and 
I therefore beg permission here to allude briefly to some of the 
2S 
Fig. 3. ¥aGaddinia) 
| 
F: 
results, reserving details for a subsequent communication to the 
Section. 
My work has been with the earths present in samarskite and 
gadolinite, separating them by systematic fractionation. Chemical 
fractionation, on which I hope to say more on another occasion, 
is very similar to the formation of a spectrum with a wide slit 
and a succession of shallow prisms. The centre portion remains 
unchanged for a long time, and the only approach to purity at 
first is at the two ends, while a considerable series of operations 
is needed to produce an appreciable change in the centre. The 
groups of didymium and yttrium earths are those which have 
chiefly occupied my attention. On comparing these rare earths 
we are at once struck with the close mutual similarity, verging 
almost into identity, of the members of the same group. 
The phosphorescent spectra of these earths when their an- 
hydrous sulphates are submitted to the induction discharge zz 
vacuo are extremely complicated, and change in their details in 
a puzzling manner. For many years I have been persistently 
groping on in almost hopeless endeavour to get a clue to the 
meaning I felt convinced was locked up in these systems of 
bands and lines. It was impossible to divest myself of the 
conviction that I was looking at a series of autograph inscriptions 
from the molecular world, evidently of intense interest, but 
2-4. 
written in a strange and baffling tongue. All attempts to 
decipher the mysterious signs were, however, for a long time, 
fruitless. I required a Rosetta stone. 
Down to a date comparatively recent nothing was more firmly 
fixed in my mind than the noticn that yttria was the oxide of a 
simple body, and that its phosphorescent spectrum gave a definite 
system of coloured bands, such as you see in the drawing before 
you (Fig. 2). Broadly speaking, there is a deep red band, a red 
band, a very luminous citron-coloured band, a pair of greenish- 
blue bands, and a blue band. It is true these bands varied 
slightly in relative intensities and in sharpness with almost every 
sample of yttria I examined ; but the general character of the 
spectrum remained unchanged, and I had got into the way of 
looking upon this spectrum as characteristic of yttria: all the 
bands being visible when the earth was present in quantity, 
whilst only the strongest band of all—the citron band—was 
visible when traces, such as millionths, were present. But that 
the whole system of bands spelled yttria and nothing but yttria I 
was firmly convinced. 
During the later fractionations of the yttria earths, and the 
continued observations of their spectra, certain suspicions which 
had troubled me for some time assumed consistent form. The 
bands which hitherto I had thought belonged to yttria began to 
