gene pool or repeated hybridization. The point is of 
significance to this work because it affects the taxonomic 
status of these populations. Because backcrossing is more 
likely between these plants and S. magnicamporum than 
between them and the usually spatially isolated S.cernua, 
and because the two live plants that I have seen and the 
majority of herbarium specimens examined appear more 
characteristic of S. magnicamporum than WS. cernua, in 
the absence of wider experience with these plants it seems 
best to treat them as a vaguely defined race of S. magni- 
camporum rather than to consider them in a more defini- 
tive manner. 
That such hybridization can occur is of interest in 
addition to its relevance to the problems inherent in 
determining a specimen in this complex. I have seen a 
few specimens from east of the range of S. magnicampo- 
rum that exhibited various characters of this species but 
which were otherwise referable to S. cernwa. Occasion- 
ally these were collected with normal specimens of the 
latter species. It would appear that during the xerother- 
mic period S. magnicamporum ranged much further east 
than at present, following the availability of dry prairie. 
As the climate moderated, this species retreated west- 
ward, but some hybridization with S. cernua did occur 
as environmental barriers deteriorated. Genes of S. mag- 
nicamporum were then thoroughly incorporated in the 
gene pools of various populations of S. cernwa, so that 
now recombination produces rare plants suggesting SS. 
magnicamporum in various respects. 
It is unfortunately not certain how S. magnicamporum 
behaves with regard to var. odorata. I have not had live 
material of this taxon available for hybridization experi- 
ments. Apparently the only contact between the ranges 
of these taxa is in the black belt. They cannot, however, 
be truly sympatric because of their different habitat re- 
[ 296 | 
