271 
IV. Correlations in Development. 
one series of experiments Balls also gives the average length of the stem. 
This set of results has, therefore, been used, since it possesses the additional 
advantages of having been obtained in a long and favourable season 
(1913)- 
In the cotton-plant the flowers are borne successively on lateral 
branches. The oldest and longest branches at the base are the first to pro¬ 
duce fruits, which appear while the upper branches are still producing 
flowers. Since the general arrangement of the plant is as in Fig. 5, the 
appearance of flowers ought to curtail the growth of the stem apex, and the 
development of fruits should reduce the rate Qf flowering, if the supply of 
Fig. 6 . Stem length, flowering and fruiting curves in Egyptian cotton (from the data of 
Balls and Holton), giving average increase per plant per day. The continuous fruiting curve is put 
forward three weeks (see text). 
food material from the roots is the limiting factor in the total growth rate of 
the plant. 
The results of Balls and Holton (1, 2 ) for the Middle Period sowings in 
the sowing date experiments of 1913 appear to justify this assumption. 
They are given as weekly averages (Fig. 6), for increase in stem length, 
flowers opening, and fruits maturing per plant per day. It is apparent that 
the decreased rate of growth of the stem after June 13 coincides with the 
development of flowers nearer the base of the plants. After July 18 the 
flowering rate also decreases, and this is followed by the development of 
the fruits. The flowering and fruiting curves do not actually represent the 
effects of the developing fruits on the flowering curves. The flowers give 
rise to mature fruits about seven weeks later (Balls, loc. cit.), but the fruits 
grow most rapidly about two (or two and a half) weeks after flowering. 
Since the flowers take four weeks to develop, we can assume they are 
growing most rapidly about a fortnight before they open. Hence the most 
T 
