PECULIAR SPECIES AND OCEANIC ISLANDS 439 
other causes of peculiarity, as a preponderance of species, 
genus or higher group, or insulation of individuals, &c., 
&c., must be secondary considerations. Except Brown and 
Humboldt, no one has attempted this, all seem to dread the 
making Bot. Geog. too exact a science ; they find it far 
easier to speculate than to employ the inductive process. 
The first steps to tracing the progress of the creation of 
vegetation is to know the proportion in which the groups 
appear in different locahties, and more particularly the 
relation which exists between the floras of the localities, a 
relation which must be expressed in numbers to be at all 
tangible. 
Edinburgh : July 1845.i 
Bother variation, development and all such subjects ! 
it is reasoning in a circle I believe after all. As a Botanist 
I must be content to take species as they appear to he, not 
as they are, and still less as they were or ought to be. You 
see I am annoyed at my own incapacity to fathom or follow 
the subject to any good purpose (open confession is good 
for the soul). 
I think I can give you plenty of instances of pecuhar 
genera with several good species in very small islands. [A 
list follows.] 
I have always felt opposed to Bory's (who is a great 
Gascon ! but not to be despised) views of the variableness 
of insular species. I certainly have no good evidence in 
favour of the loose statement I made and which corresponded 
with a vague idea I held, of insects being scarce on islands ; 
yet 13 species is surely very few for Keeling if size is to be 
regarded ; how often may you not find 13 on yom' own 
window ? Kerguelen Land has only 3. New Zealand and 
V.D.L. are certainly poor — in Trinidad (of Brazils) I saw 
only 3, I think, a Hemerobius and the House flies and Cock- 
roach, introduced from a wreck : Canaries and Madeira are 
poor, I think ; Cape de Verds are too dependent on the W. 
coast of Africa to judge from. Nothing struck me as so 
marvellous as the appearance of 4 Insecta and many Arach- 
nida you mention as on St. Paul's rocks. Still I agree with 
you on the main point that such few as there are would be 
enough for impregnation if they only went to work about it. 
* For Darwin's answer, see More Lettcn:, i. 5]. 
