As regards many pieces of land, there is no doubt as to the purposes for which they can most profitably be used. This piece will command a much higher rent as a shop site than in any other capacity; for that piece house-building is the obvious employment; for another, agriculture. But in quite a number of instances there is considerable uncertainty. It is not clear whether upon this site it will be better to erect a house or a shop, or if the latter, what kind of a shop. It is not clear whether it will pay to use that farm land for a building scheme; and, within the domain of agriculture, which of course comprises an immense variety of really different industries, it is often a very moot point indeed whether a certain field should be left under grass, or brought under the plow. Cases of this sort are not phantoms of the imagination; they emerge on every side as concrete problems with which some one or other is dealing every day, and it is these cases which constitute the marginal land for the purposes of a particular occupation. The marginal sites for shops are the sites for which it is only just worth while to pay rents sufficient to entice them away from houses. And the rent for a site in Bond Street, or elsewhere, which is so much more suitable for shop purposes that no alternative use would be worth considering, will exceed the rent paid for one of these marginal sites by, roughly speaking, the extra advantage it possesses for shop purposes. Or will fall short of it, it may be well to add, to the extent of its comparative disadvantage. For there may be many such marginal sites, some of which will fetch low rents, and others very high rents indeed; the same site being often of great potential utility for a large variety of occupations. Between any two occupations there will thus usually be a margin of transference, which we must conceive not as a point, but as an irregular line, upon or near to which there will be many pieces of land, differing greatly in the rents which they fetch. These variations of rent will correspond to the differences between the advantages or derived utilities which the sites possess for both the occupations in question. The position of such margins of transference will of course alter as industrial conditions change, and, when they alter, the rents of sites which are not near any margin of transference will be affected also. Thus an increased demand for the products of any particular industry will make it profitable for that industry to offer higher rents, and thus draw land away from other occupations. This will have the effect of raising, though possibly to a very slight extent, the rents of sites which still remain in other uses; for there will be fewer of them available; and their derived utilities will consequently be increased.


