372 BOTANICAL GAZETTE [MAY 
disk florets of Aster Shortii, A. novae-angliae, A. puniceus, and A. 
prenanthoides appeared. In this he recognized clearly the decrease 
of the number of the bracts, rays, and disk florets of Aster prenan- 
thoides in the successive collections.’ 
The cause of the seasonal change of flowers in the multicipital 
Compositae is probably partly due to a change in nutrition, as HAACKE 
(4) and SHULL (12) have suggested; but an environmental condition 
seems to influence this decrease also. Recently Dr. Korrpa (16) 
of this laboratory found favorable material in the single-headed 
Arnica unalaschensis, which he collected on Mt. Hakkéda (Aomori), 
because the mean number of the rays changed exclusively from envi- 
ronmental changes. According to him, the mean increased from 
14.389 to 15.741 between July 29 and August 5, for this was the 
season of thawing snow, and hence a favorable condition; then it 
decreased continuously to 15.081 on August 30. 
During the summer of 1909 my attention was called to large patches 
of Aster jastigiatus on the banks of the Tone River, near Kokoku, 
Simosa Province, about thirty miles east of Tokyo. The number 
of rays and of disk florets was rather small, so that I was induced 
to calculate the coefficient of correlation between them. The present 
preliminary note seeks to verify Lupwrc’s rule in the variation of 
these rays and disk florets, and to find the seasonal change in rays and 
the range of correlation between rays and disk florets. 
The material was taken from three patches scattered through a 
grassy field (about 100 meters square) along the river bank. The 
environmental condition of these three patches was almost the same, 
and seems not to affect in any degree the variation of the flowers; 
therefore, only the difference of time seems to account for the differ- 
ences in the three collections. The first collection was made in the 
eastern part of the field, the second in the western, and the third in the 
middle. In collecting specimens care must be taken, for random col- 
lection seems to give a little larger value than picking all flowers eaten 
every individual (Lucas 14); and the latter method, adopted in this in- 
vestigation, may probably avoid personal error. All flowers that were 
injured by insects or that were extremely old or young were rejected. 
2 From September 27 to October 8 the mean number of rays decreased from — 
30.769 to 26.335. 
