PT. II, CH. XXII] GEOGRAPHICAL DISTRIBUTION 229 



made— from the theory of natural selection, which are to-day 

 strenuously supported, and the belief in which seems to me the 

 chief preventive to further progress in the study of distribution, 

 are perhaps the following: 



(1) That distribution of species about the world has in general 

 been rapid. 



(2) That the present distribution of species and genera about 

 the world represents the maximum possible to those species and 

 genera, and that distribution is consequently a closed chapter. 



(3) That species and genera now existing occupy, as a rule, 

 just those places to which they are suited. 



(4) That species and genera occupying small areas are as a 

 general rule species and genera that are dying out (relics). 

 Natural selection could not produce them upon areas so small 

 as are occupied by a great many. It also demands that there 

 shall be a good many moribund forms; and therefore these 

 localised forms are assumed to be dying out. 



(5) That on the whole, in the same way, small genera (with 

 few species) are to be regarded as relics, and as in process of 

 dying out. 



As regards the first two of these, we have seen in Chapters ii 

 to V that there is no reason to suppose that as a general rule 

 dispersal in nature is anything but extraordinarily slow, the 

 ground being usually fully occupied by societies or associations 

 of plants, into Avhich entry will be difficult or even impossible. 

 This is confirmed by ordinary oljservation, for if one remember 

 the position of various clumps of plants from one's childliood. 

 one soon realises that if man have made no alterations in the 

 neighbourhood they will be found in the same places, without 

 having extended their area except in very rare instances. Dis- 

 persal may be rapid if there be (which is very rarely the case) 

 virgin soil available, or if man or other cause have made some 

 great alteration in conditions, but usually it will be a matter of 

 the most extreme slowness. The figures for areal distribution 

 that have been given above, showing that the "hollow curve" 

 is apparently' a universal rule, not only for totals, but for indi- 

 vidual families and genera, show clearly that dispersal follows a 

 largely "mechanical" course, and that if a species now occupy 

 a small area, it is in most cases because it has not had time to 

 occupy a larger one. If the areas occuj^ied had been determined 

 by natural selection, it is inconceivable that they should have 

 been thus graduated in sizes from many small to few large, with 



