450 SEMI-CENTENNIAL OF TORREY BOTANICAL CLUB 
individual differences arise among plants of the same age that 
have been grown under as similar conditions as is possible under 
greenhouse and garden culture. 
The individual variations observed are to be considered as 
fluctuating and continuous. They indicate that the character of 
flower number is constantly varying, giving differences upon 
which selection may operate in the isolation of races. 
The performance of races with marked differences in vegetative 
habit of growth gives some clue to the extent to which the char- 
acter of flower number per head can be modified in correlation 
with different habits of growth. The various races range in habit 
from extremely robust and much branched races with annual pro- 
duction of large numbers of flower heads to small dwarf races which 
produce only a few flower heads in a season. The total number 
of heads produced and the length of the blooming period are greatly 
modified in such cases. The character of number per head, how- 
ever, remains more constant. In the most vigorous types we have 
found, as a rule, high values for the computed number of first day 
of bloom. Also, it is in the most vigorous races that partial 
variability is greatest in degree, for here lowest numbers per head 
(as low as 5) are found. 
The possibility of isolating races which will exhibit character- 
istic differences (although often slight) in flower number is well 
demonstrated, though rigid selection is here difficult because of 
the limitations imposed by self-incompatible plants. The con- 
tinued selection in self-fertilized lines of descent for extreme and 
unusual characteristics appearing within a race is impossible in 
chicory because of self-sterility and the limits of the effects of 
selection cannot be so rigorously tested as is possible in a species 
fully self-compatible. 
Hereditary variations of slight differences and of continuous 
gradation appear in every group of organisms with which ex- 
tended studies of heredity have been made. The two current in- 
terpretations of such phenomena are (1) those of the strict adherents 
of the genotype theory who see only recombinations of Mendelian 
units and (2) those of the selectionists who see evidence of con- 
tinuous alterations in constitution upon which selection may 
operate. 
