when it comes out, makes a room warmer, although it may very well
happen, if a fire is dying at the same time, that the room grows
colder in point of fact. For in our general statement we assume
implicitly that "other things" such as fires, are unchanged. But
assumptions of this kind are legitimate only when there is no reason
to suppose that the cause, the effects of which are being studied,
will itself produce a change in the "other things." If (as I have
often been told; I really do not know if it is true) the rays of the
sun help to put a fire out, the statement made above would be the
better for some qualification.

Now we can only say that an increase in demand raises price if we
assume the conditions of supply (as represented by the supply curve)
to remain unchanged. But in practice, an increase in demand may cause
a change in the _conditions_ of supply. An increase, for instance, in
the demand for a commodity may give rise to a revolution in the
methods of production, to the introduction of labor-saving machinery
and so forth, which will eventually result in the commodity being
produced more cheaply. It will certainly take a considerable time
before reactions of this kind can exert an appreciable influence; and
we can, therefore, feel reasonably sure that over a short period an
increase in demand will raise the price. But we cannot be sure what
the ultimate effect will be. A similar alteration in the condition of
demand is less likely to result from an increase or decrease in
supply; but it may conceivably occur. We must, therefore, be careful
to qualify any general propositions which we lay down in this
connection, by explicit reference to a short period of time. We can
add the following to our body of laws:--

LAW IV. An increase in demand, or a decrease in supply will tend to