GEOGRAPHICAL DISTRIBUTION OF THE GENUS CROCUS. 869 
other autumnal species. Two out of three vernal species, C. car- 
petanus and C. nevadensis, are essentially endemic and closely 
allied, especially in their singular leaf-structure; so that, with 
the exception of C. vernus, which is sparingly transgressive from 
the Alps, the whole of the West-European species are not only 
endemic, but the species of the vernal and of the autumnal respec- 
tively range themselves into compact groups. In Italy and the 
Italian islands the endemic vernal species form also a natural 
group; but they are associated with C. biflorus and C. vernus of 
wider range. Beyond these two cases, I know of no example in 
which either the bulk of the vernal or the bulk of the autumnal 
species predominate as natural groups; as a rule, the several 
types are intermixed. 
We find, however, many striking cases of the geographical 
isolation of individual species, the islands of the Mediterranean 
affording the most conspicuous examples. The remarkable 
C. Cambessedesii is limited to the Balearic Islands. Corsica and 
Sardinia and the neighbouring islets have two species, C. corsicus 
and C. minimus, which do not occur elsewhere. C. Crewei, with 
its singular black anthers, is limited to the little island of Syra; 
C. cyprius, the only species with a scarlet filament, is confined to 
Cyprus; and C. veneris to Cyprus and Crete. . 
Of insular varieties of species found on the mainland there are 
several striking examples: C. hadriaticus, which in Albania is 
either pure white or white with a purple throat, appears in the 
neighbouring island of Santa Maura with a golden-yellow throat ; 
C. vernusis represented on the Sicilian mountains by the diminu- 
tive C. siculus and C. Sieberi, which is self-coloured lilac on the 
mountains of Greece, appears in Crete, Andros, and some other 
neighbouring islands with variegated purple and white flowers. 
Of the passage in colour from white to blue, in one or two 
species in their ranging from west to east, I havealready referred 
to; and there are other somewhat similar cases of colour-varia- 
tion, running, as it were, parallel through several species within 
the same district. In illustration of this, I would refer to the 
general absence of striping and feathering of species occurring in 
Dalmatia, which occur elsewhere with feathered flowers. There 
are also even more marked cases than this of mimetic colouring, 
and of different species associated in the same habitat putting on 
some identical special form ofcolouring. I would especially name 
the exact identity in colour and markings of the Santa-Maura 
