448 
AMERICAN JOURNAL OF BOTANY 
[Vol. 9. 
the following description of the strains used gives their constitutional 
formulae as established by the various crosses into which they have entered. 
A review of the Oenothera cultures at the Botanical Garden of the 
University of Michigan in the summer of 1919 led to the selection of the 
following forms as especially likely to give interesting results: 
1. Oenothera pratincola Bartlett (4). This species is highly susceptible 
and was chosen because of the long period (seven years) that it had been 
grown in self-pollinated lines. Susceptibility of the chosen strain ("Lexing- 
ton C") had been observed for eight generations. This strain (originally 
from Kentucky) has, according to the above-mentioned hypothesis of a and /3 
gametes, the genetical constitution ai^i. 
2. ''Oenothera biennis Chicago.'' This is a hardly distinguishable strain 
of the preceding, and is referred to under the provisional name assigned to 
it by de Vries in Gruppenweise Artbildung (13). It was chosen because it 
was essentially identical with the foregoing, but of entirely different pro- 
venience. Both forms had been so extensively used in crosses that the 
opportunity was seized to see if their apparent specific identity would be 
verified by identical breeding behavior. ''Oe. biennis Chicago'' wa.s re- 
ceived from de Vries in 191 2, and had been mildewed every year for seven 
years. Its genetical constitution has proved to be the same as that of 
Oe. pratincola ("Lexington C"), namely, ai^\. 
3. Oe, mississippiensis Bartlett (5). This species had been grown for 
seven seasons under the tentative name " Cartersville," assigned to it by 
de Vries, who collected it at Cartersville, Mississippi, in 1904, and sent it 
to one of the writers in 1912. It has always been heavily mildewed. Geneti- 
cal constitution, ai/3i. 
4. Oe. pratincola hyb. immunis. The hybrid which we introduced into 
our mildew experiments under the name Oe. pratincola hyb. immunis had 
an interesting origin. Of the two Oenothera species known from Lexington, 
Kentucky, and extensively grown in experimental cultures for many years, 
one, Oe. pratincola, is always mildewed, whereas the other, Oe. numismatica, 
is very slightly infected, or not at all. When these species are hybridized, 
the cross with Oe. pratincola as the pistillate parent gives twin hybrids, 
both of which are mildewed. One of them is strictly like the maternal 
parent in all characters except one trivial one, namely, the presence of 
erect, thin-walled, viscid hairs on the flower buds, a character of the paternal 
parent. This matroclinic cross, known as Oe. pratincola hyb. viscida, 
behaves in every respect like Oe. pratincola, even to throwing the same 
mutations. The reciprocal cross, in which Oe. numismatica is the pistillate 
parent, is immune and in all other respects like pure Oe. numismatica. 
According to our hypothesis of a and jS gametes we explain these facts 
as follows: 
A. Both the a and /5 gametes of Oe. pratincola are carriers of the factor i 
(susceptibility to mildew). 
