CH.xviii] THE HOLLOW CURVE OF DISTRIBUTION 197 



facts, and of the striking way in which it has been found, in 

 regard, for instance, to the flora of New Zealand, that predictions 

 as to distribution may be made upon a basis of age only, and yet 

 be reasonably accurate, it would seem probable that age is by 

 far the principal factor in determining geographical distribution. 



This of course simply means, as has already been explained, 

 that the resultant effect of the many factors that are operativ^e 

 upon any individual species (and still more upon any group of 

 ten allied forms) is so uniform, when long periods of time are 

 dealt with, that dispersal goes very largely with age. The great 

 difference between this and the older view is that we can no 

 longer look upon the dispersal of species as having reached its 

 limits. Before the rise of the theory of natural selection, as has 

 been pointed out on p. 3, the effects of age were recognised, but 

 in the last sixty years they have been more and more lost to 

 view. The figures that have been given above, however, show 

 that in reality they are perhaps the principal features that are 

 apparent in distribution. 



But Size and Space also enters into the question, and if we 

 consider this principle also to be vahd, as indeed seems sho\vn 

 by the many cases of its application that have been given above, 

 then we shall expect that as Age, Size, and Space (or Area) go 

 together, the phenomena exhibited by Size will be more or less 

 like those exhibited by Space, inasmuch as Age is the only active 

 factor of the three. Actual examination soon shows tliat this is 

 the case, and that genera of one coimtry, endemic or not, ar- 

 ranged by sizes, form a hollow curve like those formed by species 

 in order of area; they begin with many monotypes, and a good 

 many ditypes, and taper off into a more or less long tail of larger 

 genera. This of course means that the hollow curve enters not 

 only into geographical distribution, but also into evolution, for 

 nothing but evolution could produce the size of a genus. 



The hollow curve shows in the distribution into sizes of the 

 endemic genera of all islands (p. 17G), of the endemic genera of 

 individual islands (p. 175), of those of Brazil (p. 176), and other 

 countries. It shows again in the composition, by sizes of genera, 

 of the floras of Great Britain, Ceylon, New Zealand, India, etc., 

 and shows in the division of these into portions of the country, 

 single families of reasonable size, and so on; it shows again in 

 the composition of the lists of genera with one, two, three, or 

 more species. Once more it shows in the composition of the 

 lists of genera confined to larger areas of the world, such as single 



