404 BOTANICAL GAZETTE [MAY 
of the fruits, but it may be of some interest to determine the corre- 
lation between the number of ovaries and the number of ovules 
per ovary for the three series of intact inflorescences gathered in 
the spring of 1907.5 The data for 7, are given in tables XVIII 
and XIX.° The coefficients of correlation? are 
Tree 1, r= —0.007+0.023 
Tree 2,7= 0.030+0.021 
Tree 3,7= 0.1340.024 
In the first two cases the correlations are certainly insignificant. 
In the third tree there is a slight correlation which is about six 
times its probable error. Ordinarily this would be considered 
trustworthy, but the actual number of ovaries instead of the 
TABLE XVIII, CERCIS 
TREE I TREE 2 TREE 3 
ie hehe tarmac Number of Total Number of Total Number of Total 
4 ies ovules ovaries ovules ovaries ovules 
Be reitstiy ice eek 5 27 25 125 
Gee ees oki 6 44 Te 112 532 
vs peters ae, raat 84 647 ase 139 672 
NS eer araes s c 197 1515 15 76 214 1043 
se eee ee 176 1344 26 146 181 5 
WOES rcs, So as 209 1585 130 6790 49 243 
hated ola Te calle ace 142 1095 283 1495 II 60 
LA aS eae Geen Te 47 3590 207 1596 ses 
1 & STE ah Sa rN Satie as 168 807 
Pee ye ae 27 207 139 763 14 75 
Pee sia ys ees ss 15 89 
105 Six eee ets eos rank a 
17) ask. Soe Sia vas Pas ue eu 17 g2 
WO. o 5s: 888 6796 1078 5768 762 37937 
5 Harris, J. ArtHur, Is there a selective oo of ovaries in the fruiting 
of the Leguminosae ? mee Nat. 43: odie Le 
ad in the diff ily weighted with the number 
of flowers which they bear, and the poeaais and Mandatd deviations used in deter- 
mining the coefficients of correlation are calculated from these weighted frequencies. 
t used. The single case of no ovules is probably due 
In dissecting the ovaries out of the flowers, 
clearing, and examining under the lens, some accidents are rene rages There is no 
reason to believe that the 35 ovaries which ‘were broken or ruined in cl earing differed 
from those which could be counted. They are simply omitted in a calculations. 
7 Calculated by method described in Amer. Nat. 44:693-699. 1910. 
SHEPPARD’S correction was no 
to one of the ovaries being still too young. 
