Stout & Boas: STATISTICAL STUDIES IN CICHORIUM 349 
of Caltha palustris which indicate that the number of stamens and 
carpels is larger in terminal flowers than in laterals. 
Haacke ('96) made a detailed study of ray-flower number of 
Tanacetum corymbosum, taking into account the position of the 
heads on the plant. He studied 81 plants and presents data for 
each plant separately. The largest number of heads for a single ` 
plant recorded is 14. There is one flower head at the end of the 
main stem, which he calls primary. There are, оп the average, four 
or five unbranched branches each bearing a terminal head. These 
he calls secondary heads. There usually follow several branched 
branches, which bear secondary and several tertiary flower heads. 
The lower branches of the plant are longest, thickest, most 
branched and the secondary heads on these branches are about 
the same distance from the ground as the primary head. In other 
words, the lower branches have the greatest similarity to the plant 
as a whole. He found that the primary head had on the average 
the highest number of ray-flowers, the secondary head of the first 
branch the lowest. There was then an increase in the number of 
ray-flowers of successive secondary heads. From the tenth branch 
downwards the number of ray-flowers of the secondary heads was 
the same as that of the primary head. There was a correlation 
between ray-flower number in the primary head and of the secon- 
dary and tertiary heads; those plants having high or low numbers 
in the primary had correspondingly high or low numbers in the 
others. 
MacLeod ('99) made a similar study of the flower heads of 
Centaurea atropurpurea. He does not, however, keep the data 
for individual plants apart and his results are not as clear as those 
of Haacke. He first counted the terminal heads of the main 
axes, 424 in all. The average total number of flowers (disk and 
ray) in these was 47.7. He then examined the heads on the 
branches. Of these the first group consisted of 524 heads in 
bloom between July ro and 12. These had an average flower 
number of 39.2, considerably lower than the average for terminal 
heads. He calls this a "'bud-generation"  (knopgeneratien). 
After seven days he cut off all the open flower heads, and his 
second ‘‘bud-generation” consisted of 656 heads blooming from 
July 21-25, and with an average flower number of 34.4 flowers per 
