258 BOTANICAL GAZETTE [OCTOBER 
different on different parts of the main spike of a plant and on differ- 
ent branches; and (2) are they different for annuals and biennials, 
provided that the individual strength is in both cases as great as 
possible? The following experiments will show that the first ques- 
tion is to be answered in the negative, but the second in the affirma- 
tive; or in other words, the percentage figures depend upon individual 
vigor of the plants, and this between the widest possible .im ts. 
The second generation of mutant no. 3, cultivated in 1913, was 
the most vigorous of all my annual cultures, as already mentioned. 
I chose for my experiment, therefore, the strongest individual of 
this group, having the largest supply of pollen in its anthers, and 
fertilized its flowers on the main spike and on a lateral branch in 
small bags, each with its own pollen. At the time of harvest I 
separated the fruits in groups of 10 each; there were 4 of these 
groups on the main spike and 2 on the branch. In the spring of the 
following year (1914) I sowed the seeds of these 6 lots separately. 
I counted the seedlings in the stage shown in fig. 1 without trans- 
planting them. The cana were easily distinguished from the 
Lamarckiana by their narrower leaves and gray color. There were 
a number of dwarfs, which combined with this character those of 
cana and will be called cana-nanella. I have planted them out 
after finishing the countings and found them true dwarfs of the 
cana type. About a dozen of them flowered as annuals, and some 
flowered the following year as biennials. The result of the count- 
ings is given in table III. 
TABLE III 
Number of | Percentage of | P f | Percentage 
seedlings | cana | cana-nanella | together 
A, main spike 
DAG sas a, 57 40 5 45 
second group......... 114 30 4 34 
Ba ee eee 121 31 4 35 
a ea Se Rs week ed 2 5 4° 
B, lateral branch ? a . 
OOS ee Fes ae 95 39 4 43 
nao eS ee 04 31 II 42 
Pee eee 
The means for the whole plant are 34, 6, and go per cent. It is 
easily seen that the deviations from the means fall within the 
