INVASION 233 
Darwin’, in speaking of convergence, has said: “If two species, belong- 
ing to two distinct though allied genera, had both produced a large number 
of new and divergent forms, it is conceivable that these might approach each 
other so closely that they would have all to be classified under the same 
genus; and thus the descendants of two distinct genera would converge 
into one.” The application of this statement to species would at once show 
the possibility of polyphylesis in the latter, and a further examination of 
the matter will demonstrate its probability. It is perfectly evident that a 
species may be split into two or more forms by varying the conditions, let 
us say of water-content, and that the descendants of these forms may again 
be changed into the parent type by reversing the process. This has, in 
fact, been done experimentally. Since it is admittedly impossible to draw 
any absolute line between forms, varieties, and species, it is at once clear that 
two distinct though related species, especially if they are plastic, may 
be caused to converge in such a way that the variants may constitute 
a new and homogeneous species. This may be illustrated by a concrete 
case at present under investigation. Kuhnistera purpurea differs from 
K. candida in being smaller, in having fewer, smaller, and more narrow 
leaflets, and a globoid spike of purple flowers in place of an elongated 
one of white flowers; in a word, it is more xerophytic.. This conclusion 
is completely corroborated by its occurrence. On dozens of slopes 
examined, Kuhnistera purpurea has never been found mingling with K. 
candida on lower slopes, except where an accident of the surface has resulted 
in a local decrease of water-content. The experiment as conducted is a 
simple one, consisting merely in sowing seed of each in the zone of the 
other, and in growing K. purpurea under controlled mesophytic conditions, 
and K. candida under similarly measured xerophytic conditions in the plant- 
house. 
While the polyphyletic origin of species is in a fair way to be decided by 
experiment, it receives support from several well-known phenomena. The 
striking similarity in the -plant body of families taxonomically so distinct as 
the Cactaceae, Stapeliaceae, and Euphorbiaceae, or Cyperaceae and Jun- 
caceae, indicates that a vegetation form may be polyphyletic. On the other 
hand, the local appearance of zygomorphy, of symphysis, and of aphanisis 
in the floral types of phylogenetically distinct families is a proof of the 
operation of convergence in reproductive characters. To be sure, the con- 
vergence is never so great as to produce more than superficial similarity, 
but this is because the groups are markedly. different in so many fundamental 
characters. The same tendency in closely related species would easily result 
*The Origin of Species, 186. 1859. 
