540 MR. G. BENTHAM ON COMPOSITA. 
represented both in west Mediterranean Africa and in Persia, 
there are so many evidences of ancient continuity and interchange 
of races, that I have been unable to suggest any division useful 
for our present purpose. 
The flora of the greater part of this region has now been very 
fairly investigated; or, at least, copious materials for working it 
out have accumulated in our herbaria; but much remains to be 
done before they can be made properly available for geographical 
purposes, before any real accuracy can be given to the figures 
given in the preceding Table. It has often been a matter of great 
uncertainty to me as to which of the species are so far limited to 
the southern declivities of the great east-and-west chains of 
mountains as to belong to the Mediterranean, not to the Euro- 
pxo-Asiatic, region, or how far a few mountain species which 
extend into the higher ranges of the three peninsulas should or 
should not be included in the Mediterranean flora. The stations 
given in the floras of Spain, Italy, and the Caucasus will require 
to be much more carefully tabulated than I have had time to do it. 
Boissier’s most valuable Oriental Flora has not yet reached the 
Composite; and even when that is done, there will still remain 
the Turkish peninsula and a great portion of Mediterranean 
Africa, of which the Composite statistics are very vague. So local 
also are many of the most marked species and genera that we may 
still expect that a considerable number have been overlooked in 
the southern and eastern districts, though, probably, the additions 
will not be so large as in the case of the Mexican region. 
This Mediterranean region, however, is undoubtedly by far the 
richest in Composite of all those into which I have divided the 
area of the order. It surpasses the South-African by 500, the 
Mexican by 600, species, and contains nearly two fifths of the 
whole number of Old-World Composite. It may not be so diver- 
sified as either of the two just mentioned; for the genera, about 
the same in number as the South-African, are not much more than 
half those of the Mexican region, and more of the thirteen tribes 
are scarcely or not at all represented; but a large propertion of 
the genera (about half) are endemic, and about half of these are 
almost or quite monotypic. On the other hand, so large is the 
number of species of some endemic or prevailing genera, such as 
Cousinia, Centaurea, &c., that the general average is brought up 
to full fourteen to a genus, or much higher than in any other region. 
