NOTES AND OBSERVATIONS.
545
tracts it as well as B, and that B also attracts D. Then, in the mixture we have two
other forces, one connecting A with C, and the other connecting B with D. We may
express these forces by the symbols AXB, CXD, AXC, BXD.
The two first may be called maintaining forces, because they tend to keep things in their
present condition. The forces AXC and BXD may be called divellent, because they
are conceived to have a tendency to separate A from B, and C from D. This supposition
is founded on the authority of the general fact of separation of two ingredients by pre
senting a third substance which has affinity to one of them.
From this manner of conceiving the subject, it is inferred that if the sum of the forces
AXC and BXD, exceed that of the forces AXB and CXD, we ſhall have a double ex
change, and two new compounds. This may happen, therefore, although the attraction
of A for B ſhould exceéd its attraction for C, provided it do not exceed it so much as
the attraction of D for B exceeds its attraction for C.
Since it is a general fact in chemistry, that substances act most powerfully in their
simple state, it may appear strange that C, when already united with D, ſhould overcome
the stronger attraction of A for B, which it cannot do alone. This ſhews us, by the way,
that the prevalence in chemical action depends rather on the manner and concomitant cir
cumstances of the action, than on the measurable intensity or magnitude of a particular
force. Dr. Cullen used to represent this chemical fact by the diagram in the margin,
where the numbers placed between the substances express the supposed attractive forces
exerted between the substances. This diagram suggests the notion of bodies attached to
the ends of two rods or levers A D and B C, moveable round their intersection E. Were
this the case, it is certain that the attraction of A for C, and of B for D, tend to separate
A from B, and C from D ; and what is afserted above will happen, —the levers will close
between A C and B D, and A will apply itself to C, and B to D. Dr. Black first em
ployed this diagram,—but he gave it up, because it suggested a notion not chemical, but
mechanical. Levers can have no place here. It suggests also an erroneous notion. The
levers produce the effect, only in consequence of a connection which they establish be
tween A and D, and between B and C. Now, in by far the greatest number of cases in
which a double exchange is observed, we know of no such connection. He used to ex
press the cases more in the stile of chemical phenomena, by faying that, in order to have a
double exchange, the partiality of D for B must exceed that of A for B. But still this is
merely a figurative expression of an unknown cause.
It is not easy to conceive any mode of operation which will clearly produce the ob
served effect. It confirms, I think, the considerations mentioned in note 21. p. 276. and
it receives some illustration from the magnetical phenomena mentioned there.
When we say that a double exchange happens when AXCBXD exceeds
AXB+CXD, we seem to say something like instruction. But, in truth, it is only an
inference of this greater partiality from the observed effect. We have generally no other
32
VoL. I.