
904 The American Naturalist. [October, 
visited, and there is hardly any doubt that Nesomimus is extinct on 
Charles Island; and the same is true of Duncan. We have been all 
over the island, and not a singlespecimen was seen. On all the other 
islands Nesomimus was exceedingly common. It is highly interesting 
tosee that Tropidurus is greatly reduced on Charles Island. During my 
whole stay there I saw two specimens, of which I was fortunate enough 
to secure one. I do not need to state that I took the greatest pains to 
find more. On Duncan only twelve specimens could be secured after 
long hunting. On all the other islands Tropidurus is exceedingly 
frequent. Exactly the same we find in such genera which are repre- 
sented by different species on one island. 
One or two, perhaps three, of the ‘three original species’ 
may be extinct. This condition can only be explained by the 
subsidence theory, which, as I have stated in my article published 
in your journal, also gives the only satisfactory explanation of the 
harmonious distribution and differentiation of fauna and flora. 
My opinion is that at the time when these islands were still in 
connection there existed already a great number of species of certain 
genera. As soon as separation begun each of these species was differ- 
entiated for itself, just in the same way as if it were a genus. It 
might happen that one or more, even all, of the species of such a genus 
became extinct on certain islands, or were not present there during the 
time of separation.! 
Let us suppose the number of the species of Geospiza was four in 
. the time when these islands were still in connection. These four 
species may be called a, 4, c, d, and may be represented by G. fuliginosa, 
G. fortis, G. strenua, and G. dentirostris, which latter species I have 
not yet seen. a, 5, ¢, d is dependent on the different conditions on 
the different islands; taking three islands with the conditions x, y, Z, 
we can express the species on these three different islands, which 
formerly were all in connection, in this way: 
ist island, a x f(x) — 6 x f(x) — c x f(*) —d x f(a) 
ad island, a x f(y) — b Xx f(y) —¢ Xf) —@ xf) 
3d island, a x f (2) — b x f(z) —¢ x f(s) —adx f (2) 
The greater the difference between x, y, z, the greater the time of 
separation, which may be expressed by f(x), the greater will be the 
1] may state here the fact ee ae Oe e principal 
portion by barren lava fields, does not contain a single species of Camarhynchus, nor 
Nesomimus, nor Certhidea, 


iatea nei Alien dn eae Aak ERT 






















