338 BOTANICAL GAZETTE [NOVEMBER 
that portion of the more westerly tributary ravine lying between 
the road and the precipice over which the stream falls into the river, 
The locality is indicated by a star on the map (jig. 2). 
An important theoretical consideration is the relation of the locality 
to the whole range of the species and to the direction of its migration, 
What relation does the yari- 
ability of a species near its 
limit of range bear to that 
at itscenter? Is therea pro- 
gressive change of variable 
characters along the lines of 
migration radiating from the 
center of distribution of the 
species? ADAMS (1902) 
accepts such progressive 
change as one of his ten 
Fic. 2.—Map of Clifton and vicinity; station criteria, for Ti paper 
for Aster prenanthoides marked with a de tion of centers of cei 
tion. He concludes that the 
southeastern United States has been the principal center of post-glacial 
distribution for the eastern half of North America. The determimauio® 
of place-constants in various parts of the range would test the value of 
this criterion. It is to be hoped that investigators in other places W! 
make studies similar to this for the purpose of throwing a more certain 
light on these principles. 
To show the relation of Clifton, Ohio, to the total range 3 ree 
prenanthoides I present in fig. 3 a map showing the range 
species as represented at the present time by specimens in the leading 
American herbaria. Sucha map is always to some extent a pape? 
tary on the limitations of the herbaria, and does not correctly ais 
sent the relative frequency of the species in the different parts pe 
range. E.S. Burcess, the well-known specialist on the genus +1! 
writes that the stations in the Berkeley Hills, Mass., and mee 
Catskills and Shawangunk Mountains, N. Y., at hice 
limital, “the species becoming common 300 miles [48° fete Lake 
and reaching proper development along the southern ea of 
Erie and thence through western Pennsylvania.” be 
