Floral Evolution. 
395 
the same idea is, that if a biological character be peculiar and well- 
defined, it may be of great value in determining relationships 
Examples of such characters are afforded by the pollen-mechanism 
of Asclepiads and Orchids, by the pollen-presentation device in 
Campanulas, and possibly by the mode of disposition of stereom 
in the plant-body. Indeed we may go so far as to say that the 
group Angiospermae is founded not upon the biological character of 
a closed ovary, but upon the coincidence that the closed ovary is 
formed, presumably, in all the Angiosperms, in a particular way, 
namely by marginal fusion of sporophylls or rolling of a single 
sporophyll. 1 This excludes such forms as Bennettiteae from the class 
of living Angiospermae (see chapter I, p. 78 in Vol. X). 
Group-Rank, The ideal system of classification is based solely 
upon the history of the various plant-forms. Groups of species have 
descended from a common stock representing the genus; groups of 
genera from a family-stock; and so on. The genus will comprise 
all the descendant species of the generic stock, no more and no less; 
the family, all the descendant genera of the family-stock. The genus, 
family, natural order, etc., are thus groups as “natural” as the 
species; but we cannot hope to attain more than the merest 
approximation to the complete display of these groups. 
Vegetative and Reproductive Characters. The geographical 
distribution of a genus is, generally speaking, wider than that of a 
species, and the higher the rank of the group the more extensive its 
distribution. The characters of the higher groups must tend there¬ 
fore to be such as are unaffected by particular environments, and 
their determinant tendencies such as make for increased efficiency 
irrespective of the environment. The vegetative characters, being 
concerned primarily with the life of the individual, are eminently 
susceptible—indirectly if not directly—to climatic and edaphic 
influences. The reproductive parts are virtually free from such 
influences ; hence the most drastic changes in the environment may 
leave these parts unaffected in the course of descent. 
It follows that, as regards biological characters, the higher- 
ranked groups of flowering plants are determined by floral and not by 
vegetative characters. Similarity in vegetative characters might be 
due to similarity in environment; whereas similarity in floral 
characters can only be due to their inheritance from a common 
1 This is not necessarily the invariable rule ; it is doubtful if the 
ovary of Najas, for example, is composed of sporophylls (See 
ch. I, p. 75, Vol. X). 
