IV. Correlations in Development . 273 
between the first and second flowers on a branch increases as the distance 
of the branch from the lowest fruiting branch increases. 
e.g. fruiting branches 1-20 20-30 30-40 
Average time interval be-) 
tween first and second j- 5.1 6.2 7.9 
flowers (days) j 
It seems clear that growth on the lowest branches proceeds at 
a maximum rate, because these branches draw first on the food supplies 
from the roots, the higher branches only receiving the residue. 
In addition, the first node on each branch flowers later than the first 
node on the branch below it, and these time intervals increase as one 
ascends the stem. Harland’s figures are: 
Fruiting branches 1-20 20-25 25-30 30-40 
Time interval between) 
first nodes on succes- j- 2*2 3.5 3.7 4*6 
sive branches (days) j 
The individual variation is much greater, the greater the distance from 
the lowest fruiting branch. 
Finally the number of nodes per fruiting branch decreases from below 
upwards, and the average percentage of fruits maturing to flowers produced 
decreases from 33-37*8 for nodes 1 to 3 on fruiting branches to 14*9-10 
for nodes 4 to 6, and is zero on nodes 7 and 8. 
The abortion of terminal buds on branches occurs along with flower- 
bud and fruit shedding, and may, according to Harland, be considered as 
being due to the same factors. There seems to be no reason why this and 
the other facts described should not be referred to decreasing growth rates, 
consequent upon the limiting of the food supply. I11 all the cases considered 
for cotton there are developing growing-points nearer the roots than are the 
later fruits and flowers or the branch apices. In the case of roots growing 
from seeds, the growth of the stem limits the food supply to the roots and 
hence reduces their growth rate. The diminution of growth rate observable 
in cotton as the growing organs get farther from the base of the stem seems 
to show an exactly comparable state of affairs. 
The effect of manurial treatment has been shown by both Balls ( 1 , 2 ) 
and Harland ( 3 ) to leave the form of the flowering and fruiting curves 
unaltered, but to increase the rate of flowering and fruiting throughout. If 
more of the limiting nutrient substances are available and are passing up the 
stem, then more will pass the lower growing-points, and the rate of growth 
of the later flowers and fruits will increase, so that the rate of flowering and 
fruiting as a whole will be increased. Thus the manurial results agree with 
the hypothesis developed here. 
Additional Note. 
Since writing these paragraphs a valuable statistical analysis of correla¬ 
tion in the cotton-plant has appeared (Mason, 5 ), which fully confirms the 
