1907] GATES—POLLEN DEVELOPMENT AND MUTATION 83 
By July 28, O. rubrinervis had begun blooming freely, and the other 
forms, including O. lata, began several days later. The earlier 
blooming of O. rubrinervis was very noticeable. 
The history of the growth of these plants, which was observed 
‘with some care, will not be taken up here. Numerous crosses were 
made and other flowers self-pollinated for a study of the next genera- 
tion, all necessary precautions being taken by means of bags, etc., to 
guard the flowers and to secure pure fertilization. 
As stated by DEVrtEs (33), the mutant O. /a/a does not mature its 
pollen and hence must be pollinated from another species, thus pro- 
ducing a hybrid in the second and succeeding generations. The 
seeds used were from a cross of O. lata with the pollen of O. Lam- 
arckiana. These are said by DeVries (/.c. 12168) to segregate in 
the next generation, giving about 15-20 per cent. of O. Jata plants 
and the remainder O. Lamarckiana, the characters of the latter thus 
being dominant. 
My garden of 1906 contained fifteen plants from seeds of O. 
laia pollinated by O. Lamarckiana, ten of which conformed more or 
less completely to the characters of the pollen parent and four to the 
lata type, which is easily distinguished even in the rosette stage. 
One plant (no. 79), however, differed markedly from either of these 
forms, and was clearly a “mosaic” hybrid, i. e., in some characters 
it resembled one parent and in some characters the other parent. 
The petals, however, were considerably larger than those of either 
parent species, and the sepals showed streaks of red, suggesting the 
sepals of O. rubrinervis but much paler. This character, however, 
is common to all the plants of O. lataxO. Lamarckiana having the 
Lamarckiana characters, and is occasionally seen to a less degree in O. 
Lamarckiana itself. The large ovaries and stout hypanthium, 
the greater pubescence on the young buds, and the broad leaves with 
their obtuse tips, are all characters of O. /ata. But the leaves were 
scarcely at all crinkled (the more or less complete absence of crinkling 
being a character of O. Lamarckiana which distinguishes it easily 
from O. lata), and the general habit of branching and greater uxuri- 
ance of growth also correspond with O. Lamarckiana. This plant 
is mentioned as showing that segregation of the parental characters 
is not always complete in this cross, as this individual was fairly 
