446 BROOKLYN BOTANIC GARDEN MEMOIRS 
that natural selection has presided over the appearance of these new 
forms, for groups which have developed few or no endemic species are 
apparently as successful elements of the vegetation as are those in 
which such species have been abundantly produced. In fact, Willis? 
has gathered evidence from the flora of Ceylon which seems to show 
that the non-endemic species are more successful, as a whole, than the 
endemic ones, a fact which militates strongly against the theory of 
selection. Of course we are confronted here, also, with one of the 
major difficulties urged against natural selection, namely that it can 
never create but can only eliminate. 
Nor do our figures support the theory that local forms owe their 
origin to the direct action of the environment, for such a theory can- 
not well explain the abundance of endemic species in some groups 
and their rarity in others. It may be argued that the vascular crypto- 
gams and glumaceous monocotyledons are more primitive and slow- 
changing types than the petaliferous groups, and are thus able longer 
to resist the pressure of the environment and to maintain their original 
characters. We have little evidence, however, that this is actually the 
case. Ferns under cultivation seem to be very plastic, and our knowl- 
edge of the genetics of the Gramineae, at least, does not indicate that 
they are a particularly rigid group. 
Both of these views look to the environment as the factor, either 
direct or indirect, which is chiefly responsible for the origin of new 
forms, and both are open to the objection (among others) that although 
the whole flora is subject to the same environment, these new forms 
develop only in certain groups. Our third alternative largely dis- 
regards the environment. It looks upon the actual production of 
new types as due to factors within the organism rather than in its 
surroundings, and considers that the locally developed species and 
genera in the floras under discussion would have appeared in these 
regions whether isolation existed or not. Isolation is thus regarded 
merely as the agency which keeps these new forms local and endemic 
by preventing their dispersal beyond the place of their birth. Of 
course such a theory allows for the play of selection in weeding out all 
new forms which were distinctly unsuited to the environment under 
which they appeared. 
But is not this view also open to the objection which we have 
offered to the others, that it cannot account for the rarity of endemism 
in certain groups and its extreme commonness in others? A study of 
the methods of reproduction in plants belonging to these two cate- 
gories suggests an answer to this question. Vascular cryptogams in 
the great majority of cases have bisexual gametophytes and are 
2 Willis, J. C., The evolution of species in Ceylon, with reference to the dying 
out of species. Annals of Botany 30: I-23. I9g16. 
