308 HISTORY OF CHEMISTRY [LECT. xv.

Two different methods have been proposed in order to bring
these exceptions as far as possible into harmony with the
system, but neither of them wholly gets rid of the difficulty.

One party, under the leadership of Kekule,57 adheres to the
definition of valency given above, but admits that there is a
large class of substances to which it is not applicable. This
class is composed of the molecular compounds, the smallest
particles of which consist of aggregates of molecules held to-
gether by means of molecular forces. Examples of the class
are the compounds containing water of crystallisation (also
alcohol, benzene, etc., of crystallisation), the majority of
double salts, the ammonium salts, phosphorus pentachloride,
iodine trichloride, etc. No precise definition of them can
be given, but they are characterised generally by the facts that
they cannot pass, undecomposed, into the state of vapour
(although Thorpe found phosphorus pentafluoride to be an
exception to this rule58), and that they are easily formed
from and decomposed into their molecular constituents.

The adherents of constant valency are further obliged to
recognise the unsaturated compounds as exceptions. Even
although there are not a very great many of these com-
pounds, still their existence constitutes a serious objection
to the doctrine; and fruitless endeavours have been made
to weaken this objection on the ground of the tendency
exhibited by such substances to become saturated.60

The opponents of these views, whose first representatives
are Frankland and Couper,60 define the valency of an element
as its maximum saturating capacity, and under this definition
the unsaturated compounds cease to occupy an exceptional
position. In view of the fact that they further assume the
valency in the case of many elements to be considerably higher
than had previously been assumed—nitrogen and phosphorus,
for example, as quinquivalent, sulphur as sexivaient, iodine as

57 Lehrbuch der Chemie. i, 142, 443 ; Comptes Rendus. 58, 510,
68 Annalen. 182, 204* 89 Compare p. 269 and Horstmarm, Theoretische
Chemie. 295* 60 Compare pp. 231 and 254.