No.422.]] QUANTITATIVE STUDY OF VARIATION I19 
by counting larger quantities of the same material These 
appear in the following table. 
TABLE B.— CONSTANTS OF ASTER NOV/E-ANGLLE L.! 


Bnacrs. | Rays, Disk FLORETS. 
| 
PG ee eee Ce 199 199 | 199 
NUR E 44.030150 42.874371 | 62.452261 
Mode PS Be 3549 a ae 1 35 aan st | 51, 60, 63, 70 
ad ou oe * 4-111310 5.048407 7.688088 
c PATRE NOKIA: 5.212961 6.308112 | 9-314270 
POR AI i so: + 3.516142 + 4.254822 | + 6.282475 
PREM ee + .249252 + .301616 + -445352 
PES Y 00 0 + .176248 + .213267 + -314912 
C. V 11.839527 14.710682 14.91422 

There is a high degree of correlation between rays and 
bracts (Fig. 9), which is the more apparent because of the 
close agreement in the number of rays and bracts; 10.54 per 
cent of all the capitula had the number of rays and bracts 
equal. The coefficient of correlation between these was 
.802388 + .012685. Between rays and disk florets the corre- 
lation (Fig. 10) was much lower, the coefficient being .594798 
= .024859. 
Aster puniceus L.— The polygons of distributign of the 
bracts (Fig. 11), rays (Fig. 12), and disk florets (Fig. 13) of 
Aster puniceus L. are all multimodal ; but it would not be fair to 
assume that this condition is a specific one, for, although the 
number of variates was 798, they represent only three individ- 
uals, and these individuals had an exceedingly wide range of 
variation and also differed widely in the values of their “ con- 
stants.” These facts will be best appreciated by a study of 
the table on the following page. 
1 The table of constants for Aster nove-anglie L. is presented because the multi- 
modal condition of the polygons of distribution of this species may be due to too 
limited material. If more material shows this same condition, then all the con- 
stants except the mean and mode must esca suco olus The same is 
ate mn a t Ae Sie alt At ch 
TOY ali OF Te sp B5* 

