the land in
question over land on the margin of cultivation.

A clear grasp of this relation helps us to perceive that an increase
in the prosperity of the community may sometimes influence rents in an
unexpected way. It all depends on the causes which have given rise to
the increased prosperity. An advance, for instance, in agricultural
science will facilitate a more abundant supply of foodstuffs; but it
will not necessarily increase the aggregate rents of agricultural
land. For if it takes the form, say, of the discovery of some new
artificial manure, it will very likely facilitate production on the
less fertile soils far more than it will on the more fertile soils
where artificial manures are not so necessary. It will thus tend to
diminish the differential advantages of working on the more fertile
farms, and their rents will accordingly fall, possibly by much more in
the aggregate than any increase in the rents of the farms near the
margin of cultivation. The point may, perhaps, be better understood if
we pass from agricultural to urban land, and ask what would be the
effect on site values of a great improvement in the facilities of
internal transport. Push the case to an extreme, and suppose passenger
transport to become so cheap and so quick that there ceases to be any
advantage in living in a town so as to be near your place of work.
Urban landlords would no longer be able to obtain the high rents they
now receive for the sites of houses in or near a town. For most people
would prefer to move out into the country where sites can be obtained
at little more than an agricultural rent. The country covers so large
an area relatively to the towns that the supply of rural sites would
be still very plentiful as compared with the demand. Their rents would
not, therefore, rise by very much, although the rents of the housing
sites in towns would fall heavily. Of course, there are other factors
to be taken into account before we could pronounce upon the effect on
aggregate rents. Central sites for shops might, for instance, fetch a
higher rental than before. The purpose of this discussion is not to
generalize but to show the danger of generalizing about rents in the
aggregate, or land as a whole.

Sec.4. _The Margin of Transference_. The last illustration may serve,
however, to remind us of an obvious fact which we must now take into
account. The same piece of land may be used for a variety of purposes.
It may have been