Prbyest, Mewes} Periodic Law of the Chemical Elements. 03
properties. ‘The absence of an exa& regularity in the
variations of atomic weights arises from two comparisons.
It arises, first, from the fact that the differences in the
magnitudes of atomic weights of corresponding members
of different series and different groups are far from agreeing,
 even if we take into account the admissible errors of
determination. Thus, the difference between the atomic
weights of Na and Li, or, for the sake of brevity,
Na—~Li=16, as also K—Na=16, but Mg—~Be=14'6; on
she other hand, Ti-Si=20, V-P=20, Pt—Pd=gs,
Au—Ag=8g, Hg—Cd=§8, Pb—~Sn=8g, Bi~Sb=86. It
is difficult to admit that the gradual diminution of the
lifferences is only a matter of chance. We ought rather
to think that the relation between Pt and Pd is not altozether
 the same as that between Bi and 8b, and therefore
the differences cannot be the same. We might notice
that the properties of Bi and of Sb differ by a certain
aumber of individual charaéers, whose origin arises in
he slight individual separations of their atomic weights.
A second proof of the fact that periodic differences are
n reality irregular, is found in the precise determina-:iops
 of the atomic weights of the alkaline metals and the
aaloids, which were done by Stas,
If H = 1 and O = 15°96
Li = 7004 i Y
Na oo) difference  15°976
K = 39040} ” 1455
Cl = 35°368 .
Br = av) y hgh
1 =126'533} 3 40°793

THE CHEMICAL NEWS

a

V. On THE Use oF THE PERIODIC LAW FOR THE
CORRECTION OF ATOMIC WEIGHTS.
EveryBoDY knows the fate of Prout’s hypothesis. Careful
 observations having proved that the numbers of atomic
weights can contain fractions—Stas having shown that
lefinite fradtions cannot be admitted—we can no longer
:ntertain any doubts, in spite of Marignac’s brilliant opposition,
 but that Prout’s hypothesis overstepped the faés.
[ cannot even see how any reasonable argument can
support it. Let us admit that the matter of all the elements
 is completely homogeneous : there is no reason why
we should also admit, that un ponderatle particles, or #
atoms of an element, in changing into one atom of a
second element, will give the same n ponderable particles,
or that the atom of the second element will be # times as
heavy as the atom of the first. We can consider the law
of the invariability of weights as a special case of the law
of the invariability of force or of motion. The cause of
weight is evidently a particular kind of movement of
matter, and there is no reason for denying the possibility
of transforming these movements into chemical energy,
sr into some other form of movement. Two phenomena
which we can now notice in the elements, viz., the invariability
 and the indestrudibility of the atomic weights,
ate, up to the present, in intimate and even historical relation.
 If, then, a known element was to decompose
‘tself, or if it was to change into another one, these
shenomena might be accompanied by a diminution or an
.ncrease in weight. In this manner can be explained, up
to a certain point, the difference of the chemical energy in
different elements. In expressing this idea here I wish to
say only that there is some possibility of the opinions of
chemists on the composed nature of elements becoming
aniform without adopting Prout’s hypothesis.
From a pra@ical point of view, Prout’s hypothesis is
iefetive, in so far as it concerns directly, small numbers.
Dur usual determinations of atomic weights do not often
agree to a fiftieth part with the whole number, often even
to nearly five or six units, whilst in Prout’s hypothesis it
is only a question of tenths. Thus, as we have already
seen, the different determinations of the atomic weight of
titanium vary from 57 to 48.
But from a theoretical as well as from a practical point
of view, it is very important to have a principle which
will lead to the discovery of the great errors which are
sometimes made in the determination of atomic weights,
and to have an idea of greater numbers. If in these last
we discover a true regularity, it is certain that eventually
we shall discover such a regularity that will enable to
letermine theoretically, and with great precision, the
magnitude of atomic weights. The periodic law, even in
its present state, enables us to discover moderate errors in
the determinations of atomic weights, as we have seen in
the case of titanium. We are now going to give examples
of a more rigorous nature.
To corret atomic weights exactly it is necessary to
submit the individual properties of elements to an exa&
and comparative examination, for it is these properties
which cause the perturbations in the regalar variations of
she amount of the atomic weights. Already, in the preceding
 pages, I have noticed the evident existence of such

Wemust see that there is a stri& comparison in the
lifferences between the values of the atomic weights of
:lements that correspond in the system, but we also notice
dy the side of these comparisons individual separations.
Therefore the elements possess general properties, desendent
 periodically on their atomic weights (fcr example,
:he property of given determinate forms of oxidations,
which are modified spontaneously and gradually), and inlividual
 properties which are determined by the separaions
 mentioned above.
We do not really know anything of the just-mentioned
relation, except the periodic property ; and even this one
is not properly understood. It is therefore impossible to
determine exaly the amount of the separations, any more
than we can corre&t positively the magnitude of atomic
weights. We can only determine limits, certainly very
near together, between which the amount of the atomic
weight of a given element should be.
According to the periodic law, the atomic weight of
:ellurium should be greater than that of Sb= 22, and less
:han that of I=127; that is to say that the atomic weight
of tellurium ought to be about 125, because from every point
of view, atomic analogies assign it a place between Sb and
[. Considering that Ag—Cu=43, that Cd—Zn=47, that
S5b—-As=47, that I-Br=47, that Cs—Rb=48, that
Ba—Sr=50, we are led to presume that in the same
manner Te— Se= 47 (nearly), for Te is found between Sh
ind I, as Se is between As and Rr; and as Se, whose
atomic weight is 78, has been more completely studied,
and can be more easily purified than Te, the number found
for Se should be accepted before the atomic weight found
‘or Te; this is why this latter should be about 78447 =125,
This result does not agree with the numbers last given by
Berzélius (who determined the quantity of TeQ, obtained
by the oxidation of Te by means of nitric acid).

0 = 1596, 0 = 160
Te = 12813, Te = 12845
= 12797, = 12828
= 127°96, = 12828
Hauer (who estimated the bromine of K,TeBrs as AgBr)
>btained corresponding numbers; still, Te is generally
aken to be = 128. Further, as the first determinations,
nade by Berzélius (1812), gave the number 116, as the
ater experiments have given the number 129, the exa&i-