312 The American Naturalist. [April, 
Of all the forms from the islands, the genus Tropidurus is best 
known. The most divergent species of this genus we find on 
Abingdon on one side, and on Hood on the other. On Abing- 
don we also find a peculiar species of Nesomimus, of Certhidea, 
of Cactornis, of Testudo, and probably of Geospiza. On Hood 
Island we also find a peculiar species of Nesomimus, of Certhidea, 
of Geospiza (Cactornis has not yet been discovered), and of Testudo. 
These forms are entirely different from each other, and different 
from the forms of the central islands. What is the reason 
of this difference? The fact is that a// forms of an island 
become modified, and not alone a single species; the plants 
and the different groups of animals all at the same time. 
There must be a common cause which produces this effect. And . 
this cause can only be looked for in the surroundings, in the physical 
conditions of each island. That there is a difference among 
these islands is evident. All the lower islands do not reach the 
damp region; they must be therefore in quite a different physical 
state. Some of these islands are without a drop of fresh water; 
others are furnished with this element. This difference must have - 
a different effect on the same forms of animal and vegetable life. 
I have expressed the opinion that when these islands were 
still in connection, forming one large island, there was probably 
only asingle species of Tropidurus, Testudo, Nesomimus, and so 
on. There were probably certain small local variations, but they 
were not so expressed, being not separated; and besides that any 
new characters appearing were checked by intercrossing. We® 
could imagine, for instance, that the large island had the higher 
moist region over its whole extent; the effect would have been 
that humidity was spread more equally over the whole island. , 
If a certain portion ‘became separated, and lost that upper horizom, 
it was at once in a fundamentally different condition. This affect 
ed the flora and fauna; and the flora again the fauna. All these 
changes, of course, must have gone on very gradually. te 
From these considerations we may proceed to get an expe 
tion of the variation. Hoffmann and others have succeeded, i 
the course of several years, in changing wild plants by cultiva, 
tion in gardens. Thus Hoffman could change the wild carrot 
