2 10 
Willis .— Age and Area. 
and vice versa, both really work upon what are essentially the same 
principles, and that they produce results of the most extraordinary 
parallelism. 
The Distribution of Genera by Sizes in Definite Areas; 
a Difficulty for the Supporter of Relicdom. 
A commencement has been made with this subject in my book, upon 
p. 189, where it is shown that in South America, to take one instance, there 
exist, confined to this continent, 18 per cent, of the world’s monotypes, but 
only 16 per cent, of the ditypes, 13 per cent, of the genera with five species, 
11 per cent, of those with ten, and so on, the proportions steadily diminish¬ 
ing. To explain this by Natural Selection appears to me an impossibility, 
for the same phenomenon is shown by other continents and islands, and 
Natural Selection could not produce such a regular and mechanical result, 
nor would relics be likely to exist in a steadily diminishing proportion, with 
the greatest number at the point of death. 
But other problems arise out of these facts, which are perhaps even 
more difficult for the natural selectionist. How explain, for example, the 
fact that the actual numbers and proportions (from my Dictionary, 4th ed., 
as usual) of the genera confined to the three most isolated continents are: 
Africa. 
North A meric a. 
South Amer 
Total of genera 
1,740 
1,308 
U 739 
With r species 
835 48 % 
612 47 % 
887 
5 i% 
2 
254 14 
224 17 
263 
15 
3 >> 
136 7 
in 8 
125 
1 
5 > y 
97 5 
58 4 
CO 
•<r 
5 
10 „ 
5 6 3 
42 3 
59 
3 
The percentages are counted downwards; 48 % of the African genera are monotypes, 51 % of 
the South American. 
Not only do the proportions decrease regularly, but they are closely the 
same for all three. Neither of these facts can be explained by Natural 
Selection, though they are quite simple to Age and Area, if we imagine 
genera to be ‘ casually ’ formed at different spots. 
Or, to take another case, why, on the theory of Natural Selection, should 
Australia, which is a large island, have only 241 out of its 565 endemic 
genera monotypic, or 42 per cent., 1 while Java has 57 out of 59, or 96 per 
cent.? The smaller islands than Australia,taken as a whole, have 1,037 out 
of 1,582 endemic genera monotypic, or 65 per cent. The explanation of 
relicdom is somewhat hardly pressed to explain such cases as this, or the 
still more difficult case that while Java has 96 per cent, of its endemic 
genera monotypic, Socotra has 89 per cent., Japan 80 per cent, Ceylon 
76 per cent., Madagascar and New Zealand 70 per cent., New Caledonia 
57 per cent., and the Hawaiian Islands merely 31 per cent. (cf. figures on 
1 i. e. less than Africa, North America, or South America. 
