STOUT & Boas: STATISTICAL STUDIES IN CICHORIUM 453 
and the lighter seeds planted were beyond the limits of the entire 
range of the seeds of the original parent plant. The actual ranges 
of the successive crops, which might show multimodal irregularity, 
are not given; only the average weights of all above or below the 
mean are presented. Here, however, the data show that the 
range of variability in offspring of a single plant may far exceed 
that of an original parent. While this increased variability is 
interpreted by Johannsen as within the range of variability of the 
genotype such results are not at all inconsistent with the view that 
actual changes in hereditary constitution are in evidence which 
when subjected to such careful selection as that employed by 
Castle and Phillips may result in the isolation of further genotypes 
within the “риге line.” 
"The necessity of considering the performance of individuals 
as units in judging a progeny, or a line of progenies, and of repre- 
senting on this basis the individual variations rather than the 
partial variations occurring within the individual, when such 
fluctuating characters as weight of bean seeds and number of 
flowers per head in composites are involved, is well shown in the 
studies of chicory reported above. The greatest difference be- 
tween individuals and between races as such (in respect to flower 
number) is seen in the production of higher numbers per head rather 
than of lower numbers. The tendency of the variabilities is to 
extend the range of flowers per head to higher numbers keeping 
the range in lower numbers much the same. There is not a de- 
cided shifting of the entire range to higher and lower values for 
the plant or the race as a whole. 
This characteristic variability among races, to extend or to 
limit the range in high numbers rather than to shift the entire 
range, indicates that racial differentiation is here quite different 
from that which appears to have prevailed in the races which de 
Vries ('or) isolated in Chrysanthemum segetum. However, de 
Vries confined his attention chiefly to terminal flowers and 
practiced a rigid selection for extremes only. In isolating the 
race with 21 rays per terminal head, various intermediates ranging 
as low as 13 for a maximum were discarded. At any rate, the 
isolation of two extreme races by de Vries is no proof that other 
and intermediate races could not also be isolated quite as they 
have been in chicory. 
