118 THE AMERICAN NATURALIST. [Vor. XXXVI. 
The curve of the disk florets (Fig. 3) in this species shows a 
remarkable lack of variation in parts which, owing to their 
indeterminate character, might be expected to be the most 
variable. The mode is on 22, the mean on 22.053097, and the 
coefficient of variability is only 6.860830. It should be noted 
that in all these curves there is negative skewness,! though it 
is slight in the bracts and disk florets. The variability con- 
stants for Aster shortii Hook. are shown in the following 
table. 
TABLE A.— CONSTANTS OF ASTER SHORTII Hook. 


Bracts. Rays. Disk FLORETS. 
Noc auc 226 226 226 
Meu inco oS 36.800884 14 22.053097 
Moden c uve ue us 36-37 13 22 
MD serit 3.029211 1.176991 1.115200 
K aS E ry 3.947688 1.525592 1.513025 
FEA Dea) 5... + 2.662716 + 1.029012 + 1.020535 
FEM 3:7 + 77121 + .068448 + .067885 
PRO. | x .125243 + .048400 | + .o48oor 
e. v. 10.727157 10.897091 6.860830 



In Fig. 4 is represented a “correlation surface” with rays 
subject and bracts relative. The coefficient of correlation was 
found ‘to be .549555 (P. E. p= .025157). The correlation 
between rays and disk florets is shown in Fig. 5, in which the 
coefficient of correlation is -446645 (P. E. p = + .029861). 
Aster nove-angliz L. — In this species the range of varia- 
tion in all the parts was great, the least range, 30 to 60, 
occurring in the bracts. " 
All the frequency polygons, Figs. 6, 7, and 8, show a multi- 
modal condition, but this may be due to too scanty material. 
Although no dependence can be put upon the multimodal 
condition of these curves, the “varia 
perhaps not differ widely from results w 



* 1 The degree of Skewness has not been computed for any o£ the polygons of 
distribution because the range of material was too small to make this index of | 
any value. - | ura 
- 
