CRUCIATE-FLOWERED OENOTHERAS OF SUBGENUS ONAGRA 23 1 
The writer believes that the two Sandy Hill types (and probably 
also the New Hampshire type) are distinct taxonomic units which for 
the present at least must be treated as species. Neither of them is 
the same as Oe. cruciata Nutt. There is a possibility that they are the 
cruciate-flowered varieties of thus-far undescribed broad-petaled 
species, and that further collections may discover their true relation- 
ships in this direction. The cruciate character has probably arisen 
many times in totally unrelated strains, so there is no reason to believe 
that our three types are of necessity derivatives of any common parent 
species, even if they are true species in the sense that cruciateness is 
only one of a group of correlated characters by which each differs 
from its nearest ally. 
During the summer of 19 13, the writer had the two Vermont 
strains in cultivation from seeds furnished by Professor de Vries, 
but not the New Hampshire strain. The latter has not been grown 
recently at Amsterdam. As a matter of fact, de Vries has continu- 
ously maintained only one of the Sandy Hill types in his cultures; the 
seeds of the other which he sent were old and germinated poorly, 
although a sufhcient number of plants were obtained for a taxonomic 
study. It turned out that neither of the Vermont types corresponded 
with the original Oe. cruciata of Nuttall, the type specimen of which 
"is preserved in the CandoUean Herbarium. Through the kindness of 
M. Cas. de Candolle a photograph of this specimen was obtained, which 
is reproduced as figure i. There is also one of Nuttall's original 
specimens in the Herbarium of the Philadelphia Academy of Natural 
Sciences, which the writer has carefully compared with the types sent 
by de Vries. Not only are the latter distinct from Oe. cruciata, but 
in the writer's opinion they are specifically distinct from one another. 
After the writer had reached this conclusion, and had planned to 
publish the two Lake George types as new species, he learned, quite 
by accident, that Shull had likewise assigned names to them, which 
he would soon have published in connection with a discussion of 
certain genetic experiments. In order not to cause Shull any incon- 
venience, the writer proposed that the species be published with 
technical diagnoses in this paper, but under the names which Shull 
had provisionally given them and had attached to a great many speci- 
mens and permanent photographic records. It is with Shull's approval 
that the names Oe. atrovirens and Oe. venosa are here proposed under 
our joint authority. In justice to Doctor Shull, however, it is only 
