396 
H. F. Weriiham. 
ancestry: and it is only by the efficiency of the floral characters 
that inheritance is possible. 1 
The greatest biological need of the flowering-plant apart from 
its particular environment is the formation, protection, and efficient 
germination, of its seeds. This involves the functions of cross¬ 
pollination, 1 fruit-specialization, and fruit- and seed-dispersal. The 
part that these have played in the evolution of the Sympetalse we 
have attempted to exhibit throughout our studies of each group. 
We have seen that the main lines of descent have been determined 
by two broad principles, economy and progressive efficiency for 
cross-pollination. Particular stocks, like the Apocynal, Campanal, 
Umbellifloral, and Rubialian, have been determined by the stage of 
advance attained in the course of particular tendencies on the lines 
of these principles; and actual cohorts (natural orders) reflect the 
realization of those tendencies. Lines of inferior rank have been 
determined by fruit-specialization (Diovulatae, Acanthaceae, Umbelli- 
ferae), by highly complex floral mechanisms (Asclepiadaceae, 
Candolleaceae), and even by peculiar and definite vegetative 
characters (Rubiaceae, Lentibulariaceae). 
It is not to be wondered at if vegetative characters, necessarily 
so plastic, seldom constitute the basis of any very general tendency. 
We have noticed, however, that the most advanced members of 
particular evolutionary lines are prevailingly herbaceous; and that, 
in the van of the flowering plant-class there is a general movement 
towards the herbaceous habit (Higher Tubiflorse; Campanulatae; 
Dipsacales). This, we have seen, is explicable on biological grounds. 
Natural plant-groups are often characterized by that complex 
of external features, chiefly vegetative, which constitutes the general 
facies or aspect. Such groups are exemplified in the tribe Ericoidese 
of Ericaceae, in Epacridaceae, Cactaceae, Gentianaceae, Gramineae, 
etc. This facies is often very distinctive, and, in combination with 
floral characters, affords valid evidence of descent from a common 
ancestry. 
The interpetiolar stipules of Rubiaceae provide, as we have seen, 
yet another example of the phyletic value of a purely vegetative 
character in relation to a wide circle of affinity. The climbing 
habit, again, characterizes certain extensive natural groups,— 
Cucurbitaceae-Passiflorales, Convolvulaceae, etc. In these cases 
it is not improbable that the vegetative character reflects a definite 
1 The assumption is here made that extensive groups of plant- 
races cannot persist either by asexual methods of reproduction 
or by self-pollination for indefinite periods. 
