[Vor. 11 
424 ANNALS OF THE MISSOURI BOTANICAL GARDEN 
Comparing the mean number of flowers and fruits as given in 
tables ш and v, we note that the average flower number ranges 
from 23.5 to 29.9, whereas the mean number of fruits range from 
4.5 to 6.7. Thus only about 20 per cent of the flowers develop 
into fruits. 
We also require some measure of the relative fruit production 
of the several series. 
For this purpose we may make use of the coefficient of fecundity 
(Harris, '10), i. e., the ratio c/f.: 
The coefficients are given in table virt, which shows that 
about 22 per cent of the flowers develop into fruits. The probable 
errors are low, and the differences between the several series are 
TABLE VIII 
COEFFICIENTS OF FECUNDITY IN THE 4 SERIES 
Series Number of Coefficient of 
inflorescences fecundity 
eramec Highlands 1906.......... 300 . . 
Meramec Highlands 1907.......... 605 .2121--.0021 
Jefferson Barracks 1906........... 1250 .2117--.0016 
Jefferson Barracks 1907. .......... 1270 .2267+.0015 __ 
possibly significant in comparison with their probable errors. 
The actual magnitude of the differences is, however, very small, 
being .0124 + .004 for the 2 Meramec Highlands series, .0150 $ 
.002 for the 2 Jefferson Barracks series, .0128 + .003 for the 
series from the 2 habitats in 1906, and .0146 + .003 for the 
series from the 2 habitats in 1907. 
The standard deviations as given in tables ш and v for 
number of flowers range from 7.0 to 9.2, whereas those for number 
of fruits vary from 2.5 to 3.4 in the severalseries. Тһе signifi- 
cance of the differences between the several lots of material has 
been discussed above. "Тһе absolute variability (standard devi- 
ation) of number of fruits per inflorescence is much smaller than 
that for number of flowers per inflorescence. In general the 
fruits are about one-third the comparable values for the flowers. 
* The probable error of C. F., the coefficient of fecundity, is given by 
67449 с X (1—c/f)/f. 
