342 
T. H. GOODSPEED AND R. E. CLAUSEN 
an average of over 80 for each plant, about half were taken on the 
last date of measurement. Here we have for spread the distribution 
Table lUb 
Frequency Distributions for Length of Corolla in 13(11) Fi Hz% 
Class Centers 
in Millim 
eters 
Designation 
No. 
Mean 
Coef. Var. 
51 
52 
53 
54 
55 
56 
57 
58 
Means for totals. 
4 
4 
2 
10 
53.8Ozt.l6 
i.39±.2i 
Means for first dates 
I 
2 
2 
5 
10 
57.IOdz.22 
i.83=t.28 
Means for mid dates 
I 
4 
3 
2 
10 
54.IOi.29 
2.54±-38 
Means for last dates 
3 
2 
2 
2 
I 
10 
52.6O1t.29 
2.58±.39 
of the "means for totals" showing a range of but 3 mm., from 34 to 
36 mm.; although the range of flower size on these plants amounts to 
21 mm., from 23 to 43 mm. The mean of the ''means for totals" 
distribution is 35.00 =t .13 mm. as contrasted with 39.10 =b .31 mm. 
for the "first dates," 36.10 d= .43 mm. for the "mid dates," and 
33.10 ± .22 mm. for the last dates. This means for the spread of 
corolla a progressive decrease of 6 mm. in two steps within the period 
of measurement. For the flowers of individual plants we have in 
some cases a greater decrease in mean flower size during the period. 
Thus in Plant 4 there is a decrease in flower size of 8.1 1 mm., from 
40.00 to 31.89 mm., within one month, and in the distribution of 
means for some of the other individual plants we have nearly as great 
a decrease in size during the period of measurement. The measure- 
ments on length of corolla show, in a lesser degree, the same situation. 
Within the period of measurement, the lengths decreased 4.50 mm. 
and the measurements on some of the individual plants on different 
dates show a correspondingly larger decrease in length. Thus for 
Plant 4 the decrease was 6.44 mm., from 58.00 to 51.56 mm. during 
one month. The increase in the coefficient of variability during the 
period of measurement is again very marked and would obviously 
be still more so, if the system of arrangement of "means for different 
dates" followed in Tables la and Ih had been carried out. 
Tables IVa, IWh, Va, and Vb, have to do with two Fi hybrids made 
between N. Tahacum varieties and N. sylvestris. 12 Fi Hss and 
13 Fi H38 described in Tables IVa and IV5, do not differ genetically 
from 13 (11) Fi H38 with which Tables Ilia and Wlh are concerned. 
They represent measurements on flowers of the plants from the first 
