68 Origin of Elementary Substances. J
—- Cn : ome mT ——p—ms Fe ie bea
during recent years, there is no question that the alkaline the typical molecules from which the elements were
metals, lithium, sodium, potassium, rubidium, and :volved, and is a valuable aid in the classification of
czesium, belong to the group which I have classified under slementary species, this property in the present state of
Hn. Chemists are also agreed that silver, notwithstand- knowledge is not in many cases sufficient, of iiself, to
ing the great divergence of some of its chara@eristics from indicate the group to which an element belongs. This
those of the alkaline metals, also belongs to the same will be seen from the recognised bivalency of copper and
group. Now some of the physical and chemical proper. mercury, and by the doubtful quantivalence of silver, and
ties of copper and mercury are more nearly allied to those by analogy ef sodium, all of which belong to the series
of silver than to metals of other groups, and recent in. Hn. That tetratcmic lead =208, is a member of the group
vestigations have shown that silver may, like copper, be Han, is shown by the isomorphism of its oxide, carbonate,
regarded as bivalent, since many of its compounds can be and sulphate, with the oxides, carbonates, and sulphates
represented by formule exactly analogous to those oi of barium, strontium, and calcium, besides which there is
cuprous compounds with which they are isomorphous.* no other place vacant in the system of elements where
The position of Hg, Ag, and Cu, as alternate members of one with the atomic weight and physical properties of
the series Hn, indicate their relationship with sodium, and lead would fit. .
are thereby brought into still closer conne&ion with Li, Were it not for the analogous physical properties, and
K, Rb, and Cs, That a relationship exists between the numerical relations subsisting among the elements
sodium and silver, by the isomorphism of their anhydrous grouped as forms of H3n, their classification from the
sulphates and in other ways, has already been pointed property of quantivalence alone would have hardly been
out by Odling. The greater specific gravity of sodium, possible, There can, however, be little doubt that
while possessing a lower atomic weight than potassium, aluminium, yttrium, erbium, and thorium are rightly
its passivity in the liquid state to the action of chlorine, :lassified together, and that indium and thallium are true
and its inferior volatility and oxidability to K, confirm analogues of each other. As considerable interest
the relationship of Na to the heavy metals of the series. attaches to this group at the present time, on account of
From what cause elements possessing physical proper. the recent additions which have been made to it by the
ties so widely different should be associated alternately aid of spectral analysis, I here show the atomicities of its
in regular order in the same series can only, in the present members in a separate table, calculated on the same
state of knowledge, be a subje of speculation; but, if principles as those in tables II. and III.
the views which I have enunciated on the formation of
the types Hn—Hyn be correc, it may be conceived that
after the transition of the cosmical vapours from the spiral
to the annular form, the gaseous material of each pair of
members might rotate in concentric zones, separate from
each other by an interval of space. It may be further
conceived that the rotating zones of elementary matte
were of sufficient thickness to cause a difference of density
between their upper and lower regions. That the zones
were in a highly ele@rical condition, and that their mutual
influence on each other, through the annular space between
 them, would induce opposite ele@rical conditions
in their external and internal regions; all the inner and
denser regions of the zones being in a negative, and the
puter or rarer regions in a positive ele@rical condition.
Each zone would then be in a condition to form an electro.
positive and an electro-negative element, which, on a
Bubsequent condensation, would separate and form two
zones of elements having dissimilar properties alter.
nating with the other members of the same series.
Just as silver and sodium are the connecting links be.
tween Hg and Cu, and the alkaline metals Li, K, Rb, and
Cs, 80 do cadmium and magnesium connec lead and
zinc with the alkaline earth metals glucinum, calcium,
strontium, and barium, which I have classified as forms
of Han. The classification of glucinum, with the alkaline
 rarth metals has only recently been made, but
chemists are not yet agreed upon the atomic weight of
this element, as it has been fixed at Gl=74 (Awdejew) and
Gl=g4 (Reynolds). It may, however be suspected from
the anomalously high specific gravity assigned to glucinum
(2:10) as compared with that of magnesium (sp. gr. 1°74),
and with their homologues of position, Li (sp. gr.o's5g) and
Na (sp. gr.o'97), that this element has not been isolated in
a state of puiity.} By assigning to glucinum the atomic
weight Gl=8, it enters as a multiple into all the members
of the series H2x, and may be regarded as the produ& of
the first, second, or third powers of Hz.
While the property of quantivalence would appear to
be correlated with the number of hydrogen particles in
2nd Suppl. 1088.
t * Watts, Dic. Chem.,” Suppl,, 1030.
$ Since this paper was written, MM. Nilson and Petterson have
tommunicated to the French Academy the results of their researches
on the physical properties of glucinum, and have found for the meta
a density equal to 1°64, which, although still too high, the theoretica:
density Ling about 13, is less than that of magnesium, and, con.
sequently, stands in the same order of density as lithium anc
sodium, —Comptes Rendus, Apzil 18t, 1877, p. 825.

[VV

It will be observed that there are three elements missing
‘n this group, the atomic weights of which can be prelicted
 in like manner with those of the missing elements
‘n the preceding groups. The table also affords the
means of correcting and determining the atomicities of
the series which, from their rarity, have not been
sufficiently investigated. It will be further observed that,
besides the similar numerical relations of the members of
this group with those shown in tables II. and III, the
atomic weights are all multiples of 3, and are classified
accordingly as forms of H3n.
The spectral rea&ions of this series of elements are
remarkable from the oxides of carbon and erbium giving
a spe@rum of lines at low temperatures, and by the
simplicity of the speé&ral lines of indium and thallium
in the more refrangible parts of the spe@rum. The atomic
weights of C, Al, Tl, and Th are identical with those
generally received, and afford presumptive evidence that
the atomic weights of the intermediate members are
equally correct. It will, however, be observed that the
atomic weights of yttrium and indium are double the
accepted numbers (Y=61'7, In=75'6); but in regard to
the latter element it has not yet been definitely agreed
which multiple of 376, the original determination, shall
be the classical one, as the atomicity has been fixed by
different chemists at 75'6, 113, and 150, the number
assigned to it in the table. The relations which the
double atomic weights of In and Yt have to each other,
and with their homologues of position, Cs, Ba, Ag, and
Cd, in tables II. and III, render it highly probable that
the atomic weights of Y and In in the table are corre.
For similar reasons it is probable that the atomic weight
of erbium will be found to be 177. It is only very
recently that any investigations of the atomic weight of
this rare element have been made, from the difficulty