values for the atomic weights of the elements.
2. The estimation of the atomic weights of the elements. On account of
practical difficulties, it is not always possible to fix the atomic weight of some
elements by vapour density determinations (Avogadro's rule), and by specific heat
determinations (Dulong and Petit's rule), and the atomic weights of these elements
were frequently assigned on somewhat uncertain grounds. According to C. L.
Winkler, indium has the equivalent weight 37 '8. The correct atomic weight must
be some multiple of this, and for no special reason, the atomic weight was once
taken to be 37'8x2=75'6. In that case, indium would fall between arsenic and
selenium where it would be quite mis-matched. Mendeleeff proposed to make
indium tervalent, like aluminium, so that the atomic weight became 37'8x3=113'4,
and the element fell in the table between cadmium and tin where it fits very well.
The subsequent determination of the specific heat of indium, 0'0577, corroborated
the change made by Mendeleeff in the atomic weight-from 75'6 to 113'4. Beryllium,
THE CLASSIFICATION OF THE ELEMENTS 263
uranium, and a number of the rare earths at one time did not fit very well into the
table, but MendeleefFs alteration of the supposed atomic weights to make these
elements fit the table were subsequently justified by vapour density determinations
of the volatile chlorides, or by specific heat determinations.
3. The prediction of the properties o! hitherto undiscovered elements.
When an empty space occurs in MendeleefFs table, it is assumed that an element
will one day be discovered which will fill that place ; and conversely, if a new element
were found to correspond with a place in the table already filled, it would be sus-
pected that the supposed element is not really elemental. In attempting to imitate
Mendeleeff, and predict the properties of missing elements in the table, attention
is paid to the composition and properties of the more important compounds
hydroxides, oxides, haloid salts, etc. so as to bring out (1) the family characters
of the group to which it belongs ; (2) the character of the series to which it belongs ;
(3) its position in the series and group so that a comparison can be made with the
properties of other known elements similarly situated in neighbouring groups or
series ; and (4) the relations of the particular group and series in which it occurs
with other groups and series. In order to avoid introducing new names when speak-
ing of unknown elements, represented by gaps in the table, Mendeleeff designated
them by prefixing a Sanscrit numeral eka (one), dwi (two), tri (three), etc. to
the names of the preceding analogous elements of the odd or even numbered series
of the same group. Thus, the unknown elements of group I will be called eka-
C8esium,and dwi-csesium. Were strontium unknown, it would be called eka-calcium.
In addition to the prediction of germanium, gallium, and scandium already dis-
cussed, Mendeleeff foretold the possible discovery of eka- and dwi-csesium ; of eka-
niobium En =146 ; of eka-tantalum Et=235 ; of dwi-tellurium Dt=212 ;
and of the analogies of manganese : eka-manganese Em=100 ; and tri-manganese
-Tm=190.
The case of the so-called inert gases is of more recent date. The discovery
of argon and helium could not have been predicted from Mendeleeffs periodic law,
but after these elements had been discovered, and accommodated in the periodic
table between the strongly acid halogen family and the strongly basic alkali metals,
the probable existence of other similar inert gases was indicated. When an ex-
haustive search was made, krypton, neon, and xenon were discovered with pro-
perties and atomic weights which could have been predicted from the arrangement
which was made for argon and helium in MendeleefE's table.
4. The correction of the values of atomic weights. If the atomic weight
of an element does not fit with the regular course of, say, the atomic volume curve,
Fig. 4, the atomic weight is probably in error. Thus, the atomic weights of
platinum, iridium, and osmium at that time were probably too high, and subsequent
determinations verified this inference. For example, the atomic weights of these
elements were :
Platinum. Iridium. Osmium.
In 1870 . . . 196-7 196-7 198*6
In 1919 195-2 193-1 190-9
6. Some Defects in the Periodic Law
The scientific value of thoroughly sound hypotheses is enhanced daily both by known
facts that they are continually assimilating, and new facts that they are con
revealing. J. WARD (1899).
There are some misfits in the MendeleefE's table as we have it to-day, owing
to the fact that at least three pairs of elements would be mis-matched if they were
simply classed according to their atomic weights : argon (39'88) and potass:
(39-10) ; cobalt (58'97) and nickel (58'68) ; and tellurium (127'5) and iodine (126*92).
264 INORGANIC AND THEORETICAL CHEMISTRY
G. Kruss and F. W. Schmidt (1889) 1 attributed the difficulty with cobalt and nickel to
the presence of a hitherto undiscovered element in nickel which they named gnomium.
This explanation, however, had to be discarded. It did not survive the ordeal
remorselessly applied to conjectures of this kind. No gnomium has yet been
found. Again, the case of iodine and tellurium has been studied with relentless
vigour stimulated largely by D. I. Mendeleeff's prediction : " The atomic weight
of tellurium must be between 123 and 126, and cannot be 128." Iodine most
certainly belongs to the same group as the other halogens, and tellurium undoubtedly
belongs to the selenium group ; these elements are accordingly placed among
their own family relations in spite of the fact that if their atomic weights were
alone considered tellurium would be ranked with the halogens, and iodine with
selenium. B. Brauner (1889) suggested that ordinary tellurium is a complex
containing a- and jS-tellurium ; and it was inferred that true tellurium say a-Te
has an atomic weight 125, and that the other form of this element has a higher
atomic weight, and will find a place in the periodic system in the valency below
tellurium. D. I. Mendeleeff cafled this undiscovered element dwi-tellurium, Dt,
and he sketched some of its physical and chemical properties ; but tellurium, said
G. Wyrouboff, has been tortured in every conceivable manner : it has been melted,
sublimed, oxidized, hydrogenized, phenylated, dissolved, crystallized, fractioned,
and precipitated ; yet nothing but failure has followed all attempts to get an
atomic weight lower than iodine or to fraction the element into two others. Nothing
has developed which would warrant a belief in Mendeleeff's " must." Hence, in
spite of the fact that " the laws of nature admit of no exception," faith in the law
has led chemists to allocate these discordant elements according to their chemical
properties and not according to their atomic weights. To put the matter bluntly,
the procedure runs : It is necessary either to reject the periodic law or to reject
the number 127'5 for tellurium ; the periodic law cannot be rejected because it is
the very embodiment of truth, nay, truth itself ; ergo, in spite of all evidence to the
contrary, the number 127'5 must be wrong. Bode's law of astronomy successfully
predicted the asteroids and allocated their proper place in the solar system ; but
the subsequent discovery of Neptune did not agree with Bode's law. The law was
accordingly abandoned and it is now regarded as a curiosity. Mendeleeff's law
may have to go the same way. B. Brauner's assumption that tellurium is a mixture
of true tellurium with a higher homologue, may be a good working hypothesis for
stimulating experiments on this element, but to maintain the thesis against all
evidence to the contrary ' ' may afford an easier way out of the difficulty than by
working steadily at the cause of the discrepancy, but it affords at best a
feeble and undignified cover for one's retreat." This method must be dubbed
unscientific, but the circumstantial evidence justifies the procedure in the ex-
pectation that a consistent system will ultimately grow from the truth and error
engrafted into the " law." It is not very probable that the principle underlying
the periodic law will be abandoned because it is founded on a vast assemblage
of facts of different kinds ; and because it seems to be plastic enough to fulfil
subsequent requirements. The central problem in inorganic chemistry, said
W Ramsay (1904), is to answer the question : Why this incomplete concordance ?
Allocation of hydrogen. The location of hydrogen in the table, as already
indicated, is difficult, because hydrogen seems to be without companions. It is
univalent, and thus falls either with the alkali metals (D. I. Mendeleeff, 1869 ;
G. Martin, 1901) or with the halogens (0. Mason, 1896 ; W. Crookes, 1898 ; W.
Ramsay, 1901). Although D. I. Mendeleeff 2 rather inclined to the belief that
hydrogen occupies an " isolated independent position," he said that in virtue of
" its salt-like oxide H 2 0, and the salts H, it must stand in the first group ; " that
" the nearest analogue to hydrogen is sodium which also stands in an odd series of
the first group ; " and that " the more remote analogues of hydrogen are copper,
silver, and gold." The attempts to displace hydrogen from its position at the head *
of the alkali metal group, and to place it with the halogens have not been very
THE CLASSIFICATION OF THE ELEMENTS 265
successful, but in either case many of the arguments appear rather strained and far-
fetched ; they run pro et con. somewhat as follows :
(1) Unlike hydrogen, the monad alkali metals appear to bo monatomic, but hydrogen
too is probably monatomic at a high enough temperature.
(2) If placed at the head of the halogen table, hydrogen is in close contiguity with the
other gaseous elements, but the extreme mobility and lightness of the hydrogen molecules
may be a powerful factor in determining its gaseity ; after all gaseity is a mere accident
of temperature.
(3) Hydrogen is electropositive like the alkali metals, but it is not now considered to
be a metal ; hydrogen does not exhibit the metallic properties characteristic of the family
of alkali metals, and towards lithium it behaves like nitrogen, oxygen, and the halogens
in forming a hydride. This argument is of little weight when no objection is raised to the
allocation of nitrogen and bismuth ; or of carbon and lead in the one family group.
(4) The difference between two consecutive elements usually ranges between 15 and 20,
and this agrees better with superposing hydrogen above fluorine than above lithium (7) ;
as G. Martin (1901) has pointed out this argument simply depends on the arbitrary selection
of subtraction as a criterion ; if division be selected, quite a different conclusion is obtained.
Thus, progressing upwards from potassium, the ratio K : Na = l'7 ; Na : Li = 3- 3 ; and,
following the same rule, Li : H = 6'9, which is near to the observed value.
(5) If hydrogen be placed above lithium, six gaps for undiscovered elements are crowded
in between hydrogen and helium, or helium must come in an unnatural intermediate
position, say, above carbon or nitrogen. In view of the gaps in the old periodic tables
which were subsequently filled, there is, however, no particular objection to the assumption
that these undiscovered elements have a real existence even if they have not yet been
discovered.
(6) The mutual replacement of hydrogen and the metals which has led to the acids
being regarded as salts of hydrogen, establishes a clear analogy between hydrogen
and the alkali metals ; against this it must be remembered that there is an equally striking
analogy between hydrogen and the halogens, for these elements can mutually replace one
another in many organic compounds with no more effect on the general properties of the
resulting compounds than is produced by the substitution of one halogen with another.
This argument loses much weight if it be remembered that the behaviour of a compound
is determined by its constitution rather than by the chemical nature of the atoms them-
selves, and that " the most diverse radicles may displace other radicles in a compound
and perform a similar function to that of the displaced radicles without materially affecting
the fundamental characteristics of the body into which they have entered."
(7) If the behaviour of the halogens towards oxygen be selected as a criterion, the
diminishing stability of the oxygen compounds with diminishing atomic weight culminates
in fluorine. No stable oxides are known, hydrogen oxide is a very stable compound,
totally unlike the halogen oxides. The halogens in their known oxides have a maximum
valency of seven, while the maximum valency of hydrogen is one.
(8) 'Then again, there is a great contrast between the stable hydrogen compounds with
the halogens, and the instability of the hydrides of the alkali metals. In a rough sort of
way the former property suggests dissimilarity ; the latter, similarity. Hence also
B. Brauner (1901) asks : How can such a positive element as hydrogen stand at the head
of such negative elements as the halogens ? The elements at the head of a sub-group are
always more negative and less positive than the lower members of that sub-group.
Accordingly, it will be evident that the position of hydrogen has not been definitely
settled, and that hydrogen appears to be a rogue element quite out of place in the general
scheme. Some suppose that hydrogen is a member of an extinct or yet undiscovered
series of independent elements, but whether hydrogen is the alpha or the onu-ijn.
is indeterminate because it would be eligible for a place either in group I or group
VII according to the properties selected for comparison. The supposed first
member of the series is called " proto-fluorine " ; so also the elements " proto-
beryllium " and " proto-boron " have been invented, the former with an atomic
weight 1'33, and the latter, 2. All this, however, is mere speculation.
Allocation of the rare earths. This also presents some difficulties. Most
of the rare earths can be distributed about the table according to their atomic
weights, or they can be relegated to a class by themselves. B. Brauner (1902),
who has made a special study of the rare earths, considers that they should all be
grouped together with cerium between barium and tantalum so that " Ce, 140'2J
in the table stands for : Ce, 140'25 ; Pr, 140'6 ; Nd, 144'3 ; Sa, 150'4 ; Eu, 152 ;
Gd, 157-3 ; Tb, 159'2 ; Dy, 162'5 ; Er, 167'7 ; Tm, 168-5 ; Yb, 172'0 . . .
has been called the asteroid theory of the rare earths. The properties of the
266
INORGANIC AND THEORETICAL CHEMISTRY
rare earths, however, are not well enough known to give us much confidence in
the various proposals which have been made for dealing with them ; and con-
sequently , Mendeleeff considered that the installation of these elements should
be deferred ; a similar remark applies to the radioactive elements. Here F. Soddy
and A. Fleck (1913) 3 assume :
All members occupying the same place in the periodic system are chemically identical
with one another, and are not separable from one another by chemical process, although
their atomic weights may vary over several units.
The rare earths do not fall all in the same group in this sense because several of the
members fit well enough into the table, thus, ytterbium Yb, 172 fits into group
III, series 10, etc. The so-called isotopic elements will be discussed later.
If the properties of the elements are dependent on their atomic weights
the existence of two elements with different properties and approximately
the same atomic weights should be impossible. Hence the difficulty with
elements like cobalt and nickel ; ruthenium and rhodium, etc. The peculiarities
of these elements would never have been suspected from the periodic law. It might
also be added that some experiments with the radioactive elements have led to the
inference that " different elements not necessarily of identical atomic weight, do
occupy the same place in the table, and that when this occurs, these elements possess
an identical chemical nature." The evidence as to the identity of chemical pro-
perties is not very strong when it is remembered how very few chemical tests have
been made owing to the small amount of available material. Not very long ago
praseodymium and neodymium were considered to have identical chemical
properties.
Twin elements. R. Lorenz 4 has shown that certain elements have atomic
weights which approach each other in pairs, and which differ from each other by
at most l - 4 units ; and he applies the term twin elements to pairs of elements
whose atomic weights approach one another very closely within 1'4 of one another.
For example :
Boron-carbon .
S odium - magnesium
Aluminium-silicon
Phosphorus-sulphur
Potassium-calcium
Vanadium- chromium
Manganese-iron
Diff.
1-009 Nickel-cobalt .
1-322 Selenium-bromine
1-320 Palladium-silver
1-033 Tin-antimony
0-864 Iodine-tellurium
0*940 Tantalum- tungsten
0-910 Lead-bismuth
Diff.
0-660
0-893
1-238
1-190
0-736
1-20C
1-099
The elements usually show many similarities in their chemical behaviour, and their
separation presents some difficulties. Most of the twin-elements usually follow
one another in immediate succession, so that the atomic weight of a member of one
pair differs from that of the corresponding member of the next pair by approximately
4 or a multiple of 4, e.g. Na and Al, 4'022 ; Mg and Si, 4'02 ; Al and P, 3'95 ;
Si and S, 3'663 ; etc. Lorenz shows that elements which do not form twin pairs
may follow this rule if they be regarded as representing twin pairs with other
unknown elements the exceptions are H, Be, N, Zn, Ga, Rb, Y 3 Zr, Nb, In, Cs, Ba,
Ir, Au, and some rare earths elements.
Some elements are allocated places in the table according to their atomic
weights in opposition to their chemical properties. For instance, copper,
silver, and gold fall into one group with the alkali metals. The tervalency of gold
appears to be unconformable with the valency of its companions although in its
present position the series PtCl 4 , AuCl 3 , HgCl 2 , and T1C1 is suggestive. Beryllium
is peculiarly placed from this point of view. Thallium is very like lead, but its
sulphate and some other salts are quite different from lead salts. At least three
pairs of elements have been placed according to their properties irrespective of their
atomic weights, as indicated by the misfits mentioned in the preceding section.
Again, the so-called type-elements, Li, Be, B, C, N, 0, F, which stand at the heads
THE CLASSIFICATION OF THE ELEMENTS 267
of the family groups the vertical columns of Mendeleeff's table usually have
properties quite at variance with the other members of the family. In 1870,
Mendeleeff attributed this to their low atomic weight, for he said :
The elements of the first two series have the least atomic weights, and in consequence
of this very circumstance, although they bear the general properties of the group, they
still show many peculiar and independent properties.
The difficulty still remains, for these elements have not yet been altogether reconciled
to the groups to which they should be closely analogous. The test of any given
classification of the elements arises when the arguments why a given element should
be included rather in one class than in another are reviewed. For instance, in spite
of the unique properties of fluorine or of lithium, could the former be included in
any group other than the halogens, or lithium in any group other than the alkali
metals ? The answer is in the negative.
Some elements which appear to be chemically similar are separated
in the table. For example, copper and mercury ; silver and thallium ; barium
and lead ; etc. The position of these elements in the table gives no hint of these
characteristics. Still, it might be argued that these elements exhibit many essential
differences. Thus, the physical properties of the cupric and mercuric chlorides
and sulphates show great contrasts. The stability of cuprous and mercurous
chlorides is also very different. Lead and barium dioxides appear to have a different
constitution. The unstable thallium sesquioxide, T1 2 3 , corresponds with the other
more stable sesquioxides in the group, but there are many important points
of resemblance between thallium and the alkali metals, and between silver and
lead. The extension of the periodic law to include compounds as well as elements
is not always satisfactory. Many examples will appear when the different family
groups are reviewed ; and J. Locke (1898) 5 inquires : Why should the relations
between magnesium and zinc be emphasized and the closer relation between
magnesium and manganese be ignored as if the explanation were not conditioned
by an equally important law of nature ? The worship of the periodic law as a fetish
may stimulate the pursuit of remote analogies in one direction, and close the door
to the search for closer analogies in other directions.
Multivalent elements. According to Mendeleeff, when an element like copper
forms two series of compounds in one of which it has the same valency as its neigh-
bour in a horizontal row, the compounds of the neighbouring elements are similar.
This is confirmed by the close resemblance between the bivalent compounds of
copper and zinc ; but, on the other hand, the close proximity of scandium to
titanium does not seem to confer on tervalent titanium compounds any of the
characteristic properties of scandium. Hence, Mendeleeff's classification does not
make clear the fact that heterologous elements i.e. elements belonging to different
groups in the periodic table may give analogous compounds when in the same
form of combination below their highest valency e.g. silver and thallous chlorides.
The compounds of ferric iron resemble those of tervalent aluminium and chromium,
while those of ferrous iron are like those of bivalent zinc and magnesium. Again,
the cuprous and cupric compounds have little more in common than has hydrogen
sulphide and sulphuric acid each pair of compounds has the same element copper
in the one case, sulphur in the other. Still further, the compounds of bivalent
vanadium resemble those of magnesium ; tervalent vanadium, those of aluminium ;
quadrivalent vanadium, those of silicon ; and quinquivalent vanadium, those
phosphorus. Hence, J. Locke (1898) and G. A. Barbieri (1914) recommended
the compounds of an element with different valencies be regarded as belonging
so many different elements. Ferric and ferrous iron are just as distinct primary
forms of matter as electricity and heat are forms of energy, and the one can be con
verted into the other, or into metallic iron. When a ferric salt is reduce
ferrous salt, or into metallic iron, the form of matter analogous to tervalent :
has ceased to exist.
268 INORGANIC AND THEORETICAL CHEMISTRY
The higher oxides. G. Wyrouboff represents the actual state of our knowledge
of the higher oxides of the elements by a chart constructed like the ideal curve,
Fig. 3. The selection of the characteristic oxides by Mendeleeff (Fig. 3) is quite
arbitrary, there appears to be no guiding principle. Sometimes it is the lower
oxide which is selected e.g. BaO in place of Ba0 2 ; sometimes a higher e.g.
Mn 2 7 in place of Mn0 2 , MnO. etc. Sometimes it is the more stable oxide e.g.
BaO and not Ba0 2 ; and sometimes it is the less stable one e.g. Cu 2 0, not CuO.
The curve of the actual oxides will doubtless be modified by future researches,
but it is far less regular and has more the character of a zig-zag line. No
longer can the higher oxides ranged along the same horizontal line be said to
have any relation with their chemical analogies ; the best established of which
may disappear ; and harmonious order is replaced by la plus absolue anarchie.
Against these views it has been urged that Mendeleeff purposed selecting the
highest salt-forming oxide in his table, and that he did not regard such oxides
as K 2 2 , Ba0 2 as salt-forming oxides. 6 If he distinctly specified the
salt-forming oxides, it is urged that the generalization cannot be impugned by the
consideration of another class of oxides altogether. Mendeleeff claims that the
true peroxides, Ba0 2 , Cr 2 7 , Ti0 2 , H 2 2 , cannot form corresponding salts, whereas
Pb0 2 and Bi 2 5 are distinctly salt-forming oxides in that the one corresponds
with the plumbates, and the other with the bismuthates. However, the existence
of the persulphates, pertungstates, and permolybdates does not harmonize with
| Links exchange | RSS feed 