276 Willis . — The Floras of the Outlying Islands of 
the more genera in a family the older the family will be, and therefore the 
better represented. 
Table XII. 
Fam. repres. in N.Z. by 
In N.Z. 
In Stewart. 
In islands. 
Not there. 
1 genus 
36 
16 
15 
4i % 
21 
2 genera 
15 
7 
10 
66% 
5 
3 > > 
L5 
13 
10 
66% 
5 
4-5 33 
10 
10 
10 
100% 
— 
6-10 „ 
9 
8 
8 
90% 
1 
over 10 „ 
6 
6 
6 
100 % 
— 
(18) Similarly we 
may repeat the prediction 
(3) 
on the 
same page, and 
say that the same thing will be true with regard to the genera. This also 
proves to be the case, the proportions of genera that occur being 21, 35, 583 
55, 66 , 68 , and 80 per cent., a progressive table, save for the one slight drop 
from 58 to 55 in the middle. 
(19) We may go on to predict that the plants reaching 3 groups of 
islands, New Zealand, and Stewart, 2 groups, 1 group, Stewart only, and 
the main islands of New Zealand only, should show a progressive decrease 
in the average size of families and genera. 
Table XIII. 
Reaching Aver, size of fain. Aver, size of genus. 
3 islands 
88 
6 1 
2 33 
4 2 
13 
1 3, 
2 9 
7.8 
Stewart only 
18.5 
6-8 
N.Z. as a whole 
15 
4.2 
N.Z. only 
2.9 
i*5 
(20) All the predictions we have made with regard to the flowering 
plants should also hold in the case of the ferns, and a very cursory exami- 
nation shows that they do, so that it would be a work of supererogation to 
go into details ; but we may predict that as the ferns are on the whole an 
older group than the flowering plants, those of the flora of Stewart will be 
* on the whole better represented. Examination shows that no less than 
64 per cent, of the ferns of Stewart are represented on the islands, against 
only 40 per cent, of the species of flowering plants. 
There are other predictions that might be made with regard to 
families, genera, and species, but we must go on to other more general 
features, and may begin with the species endemic to the islands only, which 
are fairly numerous. 
(21) Upon the hypothesis of Natural Selection one might expect that 
some of these endemics at any rate would be relicts. If this were so, one 
might surely expect to find some of them occurring upon two of the groups 
of islands and not in New Zealand ; they might have spread to these 
1 Only 5 genera with 30 species, too small a number for a reliable result. 
