Stout & BOAS: STATISTICAL STUDIES IN CICHORIUM 393 
These facts have a special significance in suggesting that any 
evolutionary change that may have occurred or that may be now 
in progress in respect to flower number is affecting the high num- 
bers more than the low. The numbers in the various classes, as. 
revealed in TABLES 21, 22, and 23, indicate that there are few 
plants with extremes of highest flower value and that the greater 
number of plants of the general population as grown have inter- 
mediate values for the highest flower number. In other words, 
highest flower number exhibits fluctuating variation of greater 
extent than lowest flower number and is to a large measure 
independent of the latter. Data will be given later regarding 
the heredity of high values as expressed for a plant as a whole by 
values of a, and also as to the effect of selection on high or low 
values, 
TABLE 23 
RANGE OF VARIABILITY OF FLOWER NUMBER PER HEAD FOR 219 ONE-YEAR-OLD 
PLANTS OF THE F4 GENERATION 
Number of | Maximum flower Minimum flower number Difference between 
plants number | maximum and 
Range Average average minimum 
I 26 I2.0 14.0 
4 25 16-ІІ 14.0 II.O 
2 24 т4-та I3.5 10.5 
7 23 ІЗ-ІІ 13.4 9.6 
19 22 16-12 14.1 7.9 
45 2I 12-9 13.4 7.6 
бт 20 25-7 13.3 6.7 
55 I9 1sS-II I2.9 6.1 
20 18 I3-II 2.3 ET 
5 17 14-12 12.8 4.2 
Lucr md Da irons Sows атан шы ы ышы 
3. RELATION OF FLOWER NUMBER PER HEAD TO POSITION OF 
HEADS 
A. Descriptive studies regarding position. 
Early in the collection of data, it was observed that the first 
heads to bloom are, as a rule, situated on the uppermost branches, 
and also that the first head which opened on a branchlet or in a 
cluster of flower heads is the terminal one. This at once suggested 
that the seasonal decrease so uniformly observed may be related 
to a succession of bloom involving a periodicity between develop- 
ment of different main laterals and also between different second- 
ary branches of main laterals. 
