FACTORS INFLUENCING FLOWER SIZE IN NICOTIANA 34 1 
Tables la and lb. For spread of corolla, the range shown in the dis- 
tribution designated "means for totals" is only 3 mm., from 34 to 
36 mm. The range of fluctuation of flowers was 17 mm., from 26 to 
42 mm. and an average of about 70 flowers was measured on each 
plant during a period of less than six weeks from August 9 to September 
19. The difference between the spread of flowers on plants at the 
beginning of the flowering season, at the middle of the season, and 
towards its end is very striking. The results show a decrease in size 
from 38.33 =b .11 mm. through 36.67 ± .44 mm. to 31.67 dz .26 mm. 
in less than six weeks. With reference to length of corolla this 
progressive decrease in size is not nearly so striking, the decrease 
being from 56.22 ± .09 mm. through 55.00 =±r .18 mm. to 53.44 ± .28 
mm. The number of plants measured in this and in other cases is 
very small, but the fact that the measurements refer to the same 
plants gives them an added importance which compensates for this 
defect. No attempt was made to arrange the individual means into 
a distribution of "means for different dates " ; but it is readily apparent 
from the data given that such an arrangement would have shown a 
more notable proportionate increase in the coefficient of variability 
than was shown in 13 22/07. 
As was above noted H38, the cross N. Tabacum var. macrophylla 
(U. C. B. G. 22/07) 9 X A^. sylvestris (U. C. B. G. 107/01) d', was 
made in 191 1 and flowers were first measured in 191 2. Tables Ilia 
and 1 1 15 show the size of flowers borne by plants from a second sowing 
Table Ilia 
Frequency Distributions for Spread of Corollas in 13(11) Fi Hz% 
Class Cen 
ters in Millimeters 
Designation 
No. 
Mean 
Coef. Var. 
31 
32 
33 
34 
35 
36 
37 
38 
39 40 
4. 
Means for totals 
2 
6 
2 
10 
35.OOzb.i3 
I.8lzb.27 
2 
2 
10 
39.IOzt.3i 
3-7o±.56 
Means for mid dates 
2 
I 
2 
2 
2 
10 
36.IOi.43 
5.6odb.85 
I 
2 
2 
5 
t 
10 
33.IOzb.22 
3-i5±48 
of the 191 1 seed which flowered in 1913. The number of plants is 
again small but as explained above this fact should not detract from 
the significance of the data here reported. The period of measurement 
here was greater than in the previous cases, from August 4 to November 
7, slightly over 3 months; and of the number of measurements taken, 
