198 Witlis.—Age and Area. 
been regarded as insoluble problems in geographical distribution. No 
theory based upon adaptation to conditions could hope to explain why the 
genera and species were distributed as they are—why, for example, they 
overlapped one another in such a casual way, some occupying large, some 
small, areas, and no two coinciding. It was almost impossible to conceive 
that conditions should overlap in this extraordinary way (cf. the map of the 
distribution of Ranunculus in New Zealand in 8, p. 343, and in 11 , p. 156), 
and perhaps even more impossible was it to understand why, as a rule, 
a genus retained its area unbroken, except by large barriers like expanses 
of sea. If there were much killing out of species by competition, one would 
expect a good many genera to show broken areas. 
So utterly hopeless did it seem to be ever to find explanation of the 
simplest facts of distribution upon the theory of Natural Selection, that for 
many years work upon general distribution had been practically abandoned, 
or confined to speculation with little fact to back it (cf. Hooker’s remark, 
11 , p. 104). Practically the only important work of the last twenty years 
has been that of Guppy, which has led him to conclusions diametrically 
opposed to the Darwinian theory. Even so simple a fact as that Coleus 
elongatus is confined to the summit of Ritigala in Ceylon, while C. barbatus 
occurs there and also ranges tropical Asia and Africa—a fact which can be 
matched from almost any large genus of animal or plant—is quite 
incapable of explanation by the theory of Natural Selection. 
If one adopts the very simple theory of Age and Area, which has been 
placed in a position of great strength by the fact that it can be, and has 
been, so successfully employed in making predictions, all this is made clear. 
If we start with a hypothesis A , and from it deduce that B must therefore 
occur, and then find, upon examining the facts, that B does occur, we get 
a very strong argument indeed in favour of A. When a hypothesis A has 
been successfully used in this way on many occasions, it can only be 
disproved by producing a new hypothesis X, which will explain what A 
explains, and allow at least the same predictions, or as many, to be made. 
Successful prediction goes a long way in proof of the correctness of the 
hypothesis with which we set out, when carried out many times, without 
failure, as has been the case with Age and Area. This method of proof, 
however, is as yet somewhat new and unfamiliar to most biologists, having 
been only rarely employed, and then mostly in connexion with work upon 
breeding on Mendelian lines. Age and Area has now been used to make 
more than a hundred predictions about geographical distribution and 
endemism. As every one of these has proved correct, within comparatively 
small limits, very strong evidence—not a priori , but in support of a rival 
hypothesis—is now required to displace it. 
Age and Area imagines that species spread, under the pull of the many 
and various factors acting upon them, at a rate which over long periods of 
