DISTRIBUTION OF TRIBES. 467 
scales and in the pappus, that it has been split up by Cassini and 
others into above fifty genera. Cassini’s, founded chiefly on the 
involucral scales, are, perhaps, the most natural; those of De 
Candolle, Spach, and Boissier, derived mainly from the pappus, 
may be rather more definite, but are very artificial, often widely 
separating species otherwise closely allied ; and very few of the 
groups formed on either grounds have any local character. The 
very few American species (North-western or Chilian) are con- 
nected with each other by their large capitula with broad fringed 
or torn scarious appendages to their involucral bracts ; but the 
nearest approach to these are from the diametrieally opposite 
limits of the general range—two Abyssinian species, which Boissier 
has even proposed to add to the genus Plectocephalus, founded on 
the American ones. The Abyssinian ones, however, are really, 
notwithstanding their pappus, more nearly connected with some 
of the European or Asiatic species of the Lopholoma group. A 
species still more remarkable for its distant outlying station is the 
Australian Leuzea australis, Gaudich., which we now find it 
necessary to associate with the section Rhaponticum of Centaurea. 
It is in some measure allied to the Abyssinian and West-American 
large-headed species above-mentioned ; but its closest affinity is with 
a Spanish species, the Leuzea rhaponticoides of Graells. 
A considerable number of the species have a rather wide range 
within the general area; and some appear to hybridize readily. 
C. nigra, belonging chiefly to temperate regions, extends over the 
greater part of Europe and extratropical Asia ; and two species, C. 
calcitrapa and C. melitensis, are frequently carried out in baliast or 
as weeds of cultivation to distant lands. A large number, how- 
ever, of the species are restricted to small areas. 
The small genera Crupina, two (or, aceording to some, five) 
species, Volutarella, four or five species, Zoegea, two species, and 
Leuzea, one species, all slightly diverging from Centaurea, belong 
to the same Mediterranean region taken in an extended sense E. 
and W., but do not spread northwards. 
Carbenia, one species, Carthamus, about twenty, and Cardun- 
cellus, about fourteen species, belong still to the Centaurea group 
and Mediterranean region, more abundant in the west than in the 
east, the chief character connecting all the above genera consisting 
in the very oblique or lateral scar at the point of attachment of 
the achenes. 
The same character of the achenes, though perhaps usually no 
