34 Rhodora [FEBRUARY 
from which seeds were obtained for a small culture (10 plants) which 
was started early enough so that it flowered normally during the sum- 
mer of 1911. "The culture of 1910, from the wild seeds, also flowered 
in 1911 and was identical with the F; generation, The same plant, 
which had flowered precociously in 1910, flowered again in 1911, and 
was self-pollinated. From the self-pollinated seeds of tliis plant a 
second F of ten plants was grown in 1912. One of them, No. 610-612 
in the writer's garden at Bethesda, Maryland, served as one parent 
of reciprocal crosses of which the other parent was a plant of true 
Oe. biennis L. in the garden of Prof. B. M. Davis at the University of 
Pennsylvania. The same parent plant, which was used for the crosses, 
was self-pollinated to continue the pure strain. In 1913, fifty plants 
were brought to maturity in the garden of the Bureau of Plant In- 
dustry at Glenn Dale, Maryland, and Davis grew the hybrids with 
Oe. biennis, as well as a considerable number of plants of the pure 
strain, at Philadelphia. Prof. Davis also had a culture of Oe. fran- 
ciscana from wild seed collected by Miss Alice Eastwood in San Mateo 
County. Since all the cultures, from both sources, have been reason- 
ably uniform, there is no reason to doubt that the species is a relatively 
stable type with a geographic range of considerable extent. The 
species has received its name from the fact that the material thus far 
seen has come from central California, the area covered by Greene's 
Flora Franciscana. 
The name Oenothera venusta 1s proposed for a species from the more 
southern part of California. Two varieties of it have been cultivated, 
differing from each other by the absence in one of a hair type which 
occurs in the other. It is obviously impossible to affirm that either 
variety is the parent form from which the other has been derived. 
In this case, however, and whenever a similar situation arises in the 
future, the writer will proceed on the hypothesis that the variety in 
which a character is absent is the derivative form, and the specific 
diagnosis will be drawn up to cover only the hypothetically older form. 
The varietal descriptions need then cover contrasting characters only. 
If a true specific diagnosis were drawn up to include all the varieties 
of the species, it would become necessary to define and name separately 
the "var. typica” of each species. Such a course would be logical, 
but in the present unsettled state of our knowledge of the relationships 
of the Oenotheras it seems undesirable to introduce any names which 
can be dispensed with. Consequently the diagnosis of the Oe. venusta 
