339 
its Applicability to the Ferns , &c. 
The forty-one endemics found in all the islands are thus explained 
as due to the continual growth of the number by new arrivals from lower in 
the scale. But when the islands separated, there would still be many that 
had not yet covered the whole area, and as the endemics are younger, these 
would be more numerous than in the case of the wides, in proportion at any 
rate. They would increase in the usual way from those occupying large 
down to those occupying small areas. Once the islands were well separated, 
they would necessarily be confined, in the great majority of cases, each to 
its own island, its two islands, or whatever number it then occupied. The 
actual facts of the table above bear out these suppositions completely. 
If, as there seems just reason to suppose, endemic species may them- 
selves give rise to other endemics, then the numbers now found on one 
island are probably much in excess of those that actually occurred at the 
moment of separation. The sudden rise from 113 on two islands to 273 on 
one goes to support this supposition. 
The facts of the Hawaiian flora thus fit in easily with my hypothesis 
of £ age and area ’, but without considerable use of supplementary hypo- 
theses, based merely on supposition, cannot be made to fit any other. 
III. The Distribution of Callitris in Australia. 
So far I have not attempted to include within the operation of my 
‘ age and area ’ hypothesis anything but the Angiosperms, but a very little 
investigation shows that it is in reality more general. Happening to have 
in my possession a monograph of Callitris (1), I investigated its distribution. 
The genus is confined to Australia. C. glauca occupies the whole range 
of the genus, covering Australia and Tasmania. Two species come into 
a second class, C. verrucosa (New South Wales to West Australia) and 
C. tasmanica (New South Wales to Tasmania), while a third class includes 
C. robusta , Drummondii , calcarata , intratropica , arenosa , rhomboidea , 
Macleayana , and propinqua . Lastly, a fourth class — those of very narrow 
distribution — includes C. oblonga , gracilis , Muelleri , tuberculata , Rod , 
Morrisoni , and a sp. nov. Considering the small size of the genus (eighteen 
species), this grouping fits in very well with my hypothesis, and goes to 
show that it applies also to the Coniferae. 
IV. The Ferns of New Zealand. 
I shall now go on to show, in two short essays, that the Ferns also 
follow the law of ‘ age and area In order to reduce the fern floras of 
New Zealand and Hawaii to a common denominator in a group in which 
the nomenclature has been so much in a state of flux, I have used the 
names as settled by Christensen ( 3 ), but the result is almost exactly the 
same as if I had left them untouched. 
