862 MR. G. MAW ON THE CLASSIFICATION AND 
racters; but they are for the most part lost as the raphe, chalaza, 
and caruncle wither, and the dead-ripe seed parts with its distinc- 
tive colouring. 
The two most general forms are a buff glabrous seed and a 
papillose red seed. In Crocus vitellinus and the allied species the 
seed is oblong in form and bright crimson in colour, with a gla- 
brous shining surface. In C. sativus and its allies the seed is 
nearly spherical and of a madder-red colour. In C. aureus the 
seed is crimson and covered with minute hairs. In C. gargaricus 
the nearly spherical seed is bright orange. The buff seeds ripen 
to fawn-colour; and the red seeds pass through several changes 
of colour, commencing with bright rose, then red, and are, lastly, 
at maturity of a rich reddish-chocolate colour. 
The ripe seed, which is dispersed at about midsummer, in- 
variably remains dormant till the ensuing time of growth, of 
the autumnal species, in November, the vegetation of the vernal 
species being delayed till the spring. 
Contemporaneously with the maturity of the seed, the foliage 
attached to the new corm and the roots attached to the remnant 
of the previous year's old corm die away, and the life-cycle is 
completed with the new eorm in the condition of rest, from which 
the life-history of the plant was commenced. 
Classification and Seguence —In attempting a natural group- 
ing of the genus and sequence of the species, a difficulty at once 
presents itself in thelarge number of well-marked characters pos- 
sessed by individual species, which, as it were, interlace and 
overlap in other members of the genus. Community of character 
in one organ thus brings together a different set of species to 
those whieh would be associated by the common characters of 
another organ. For instance, if we take the structure of the 
corm-tunic as a basis for classification, it would group together 
species that would be widely separated by the character of the 
stigmata; again, the stigmatic characters bring together in close 
proximity species which have obviously little natural affinity. 
This difficulty, which presents itself in many genera, seems espe- 
cially prominent in the genus Crocus ; and when I come to deal 
with this subject in a tabular form in my monograph, I shall be 
able to show that the concurrent grouping of several common cha- 
racters is almost entirely absent, excepting very small and isolated 
sections; and it will be seen that no one method of classification 
will enable the species to be arranged in strictly natural sequence, 
