COMPLEX HALOGEN ACIDS AND SALTS. 141
trations, such as by measurements of the freezing-point, of electromotive force,
etc. Thus not only can the extreme cases be distinguished — in which there
is merely a dissociation into a simple cation and a complex anion or else a
complete breaking down into the simple ions — but also it can be shown, in
the transition cases, to what degree the two kinds of dissociation prevail.
The extent to which the nature of the compound, and indeed the nature
of the elements forming the complex as well as of those outside of the com-
plex, favors the first or second form of dissociation is now understood in a
general way. The formation of complexes seems to be favored when the
anions of the simple compounds are the same. The stability of the com-
plex increases from the chlorine compounds through the other halogen to
the cyanogen compounds; it increases further, as far as metal-containing
anions are concerned, with the "nobleness" of the metal. The latter property
of metals is intimately connected with the discharge potential of their ions —
a quantity capable of exact measurement — (electroaffinity, cf. p. 67), hence
it would seem possible to predict, from data concerning discharge potentials,
the relative tendency of the metals to form complexes. Abegg and Bodlander
have in fact proposed a new system of chemical classification^1 based upon this
fundamental property of the atoms. Among the numerous investigations
which these ideas have instigated, mention should be made of the measure-
ments carried out by Bodlander of the so-called "stability-constants" which
give the proportion in which simple metal ions dissociate from complex metal-
containing anions.^2
The composition of the complex halogen compounds shows that one atom
of non-haloid nature occupies a peculiar position in the complex, while the
remaining atoms of the complex, which are either all of the same kind or
closely related, are most frequently either four or six in number. Start-
ing from these facts Werner has used with success the so-called "Coordina-
tion Theory,"^3 which he himself proposed, to explain the constitution of
these compounds. According to this theory the halogen atoms are situated
in space around the other atom of the complex, the "central atom"; when
they number four they may be supposed to be located at the corners of a
tetrahedron, and when they number six, at the corners of an octahedron which
surrounds the central atom (see p. 166). Werner has, in constructing his
theory, also introduced the idea of "secondary valences" which differ from
the "principal valences" by being weaker and incapable of binding electrons.
Thus aluminium, when its three principal valences are saturated with fluo-
rine, can bring into play its three secondary valences, and thereby bind the
fluorine atoms in three molecules of sodium fluoride:
[F ,.F_]_N£
F_^Afl..F— —Na
LF/ XF—J—Na(^1) Abegg and Bodlander: Die Electroaffinitat. Z. anorg. Chem. 20,453 (1899).
(^2) Bodlander: Ber. 36, 3933 (1903).
(^3) For a full discussion of this theory see A. Werner, Neuere Anschau-
ungen auf dem Gebiete der anorganischen Chemie, 4th Edition, 1919.