356 Gates. 
good example of a variation which is clearly germinal and yet which 
can be made to fluctuate widely by varying the external conditions, 
thus obtaining a wide range of color variation even from a single 
individual. Several cuttings were made from this plant after it was 
taken into the greenhouse. Three of these were planted out the 
following year, after remaining in the greenhouse all winter, and two 
of them produced many flowers which came perfectly true to type, 
while the third remained in the rosette stage. Even a character which 
is so notoriously subject to environmental conditions as pigment 
production is thus seen in this case, to be under the immediate 
control of ‘heredity’ in contradistinction to ‘‘variation”’. 
From the seeds of this plant a large number of seedlings were 
raised in 1908 and they were soon observed to be distinguishable 
from O. rubrinervis by the presence of a conspicuous amount of red 
pigment on the under surface of the petioles of the young plantlets. 
Unfortunately, the seedlings which were sent to Wood’s Hole, failed 
to survive transit, except four, all of which proved to be O. rubricalyx. 
Eight other plants were kept at Chicago. Five of the latter when 
they bloomed, proved to be O. rubricalyx, one O. rubrinervis, while 
the other two were not observed in the flowering stage, and therefore 
remained undetermined. One individual, therefore, reverted to the 
O. rubrinervis condition. Circumstances prevented getting pure guarded 
seeds from the Chicago plants, but open-pollinated seeds yielded in 
1909 one hundred and thirteen plants, seventy-one of which were 
O. rubricalyx thirty-eight O. rubrinervis and four remained rosettes 
whose determination was doubtful. 
The O. rubricalyx plants (F,) at Wood’s Hole were carefully self- 
pollinated, and their seeds planted at the Missouri Botanical Garden, 
partly in 1909 and partly in 1910. The 1909 lot yielded 64 plants?), 
47 of which were O. rubricalyx and 17 O. rubrinervis, though four (two 
in each category) which remained rosettes were somewhat doubtful. 
This suggested a close approximation to a Mendelian ratio of 3:1. 
These 64 plants were the offspring of two individuals of the F,, the 
percentage of O. rubricalyx to O. rubrinervis in each being respectively 
14:6(?) and 33:11. In the open-pollinated plants above-mentioned, 
1) In a discussion last year, of the nature of Mendelian segregation (GATES, 1910) 
these plants were referred to as one of two (twin) types derived from O. nanella >< 
O. biennis. This was an error, owing to the loss of a label, as I have since shown 
by repeating the cross. The O. rubricalyx plants there referred to were therefore all 
derived from the original O. yubricalyx mutant, and not in part from a cross. 
