(152) 
With the exception of Vernonia altissima, which extends west- 
ward well across the Mississippi river, the species of the two 
centers do not overlap. In this one case, V. altisstma is in its 
western range confined entirely to the rich wooded alluvial bottom 
lands, a plant formation always recognized by phytogeographers as 
of southeastern origin. The southwestern species of the same 
region are invariably xerophytes of the uplands, or rarely hydro- 
phytes of open swamps, associations normally characterized by 
southwestern species and in which V. adéésséma isneverfound. The 
ecological distinction between the two is therefore complete, for it 
is well known that ecological barriers of this nature may be more 
nearly insurmountable than hundreds of miles of distance alone. 
The separation of these distributional centers dates back to the 
glacial period, and the two groups of Vernonieae have accordingly 
been separated since that time. It is also found that those species 
most nearly alike morphologically, such as V. flaccidzfolia and 
altissima, or V. fasciculata and V. Lettérmanni, are always dis- 
tributed about the same center. These two statements lead inevita- 
bly to the conclusion that these groups of nearly related species are 
all of inter-glacial or post-glacial origin, so that groups separated on 
a geographical basis are at the same time separated phylogenetically 
as well. Considered in this light it is seen that the western plant 
described by Gray as V. altissima grandiflora can not possibly be a 
variety of that species, and that V. ¢exaza can not be grouped with 
V. angustifolia as has heretofore always been done. 
Within the two divisions geographical evidence can not be given 
such a high rank, yet a study of distribution shows that of eleven 
species of Vernonza of the coastal plain but two are found inland, 
and that for but a minor part of their range; that of four Appa- 
lachian species two only extend a short distance into the coastal 
plain; and that the three Ozarkian species and the ten Texan species 
overlap in but two cases. Those are V. Lettermannz, endemic to 
the extreme southern part of the Ozarkian center, but clearly related 
to the Texan group Fasciculatae, and V. zzterior, which extends 
southward well into the Texan center. In each of these exceptions 
the morphological evidence is so apparent that there is in no case any 
doubt as to the relationship of the species concerned. 
Another point that can be satisfactorily studied by the geographi- 
cal method only is the relation of a species as a whole to its various 
