462 Mason . — Grozvth and Abscission in Sea Island Cotton. 
proportion of flower-buds and bolls being shed, and that a stimulus of equal 
intensity occurring during the preceding period may lead to little or no 
shedding. 
Concerning the nature of the environmental factors which are to be 
associated with the modes in the shedding curves only the briefest mention 
is necessary at this stage. The waves of boll-shedding, it may be observed, 
are generally, but not always, preceded by periods of considerable precipita- 
tion and low rates of evaporation. Periods of heavy rainfall have, of course, 
long been associated in the mind of the West Indian cotton-grower with 
extensive waves of flower-bud and boll-shedding. It will be seen that the 
curves of flower-bud- and boll-shedding tend to synchronize. The error 
introduced into the results for bud-shedding by omitting to reckon a large 
number of flower-buds until the opening of the foliage leaf excludes the 
possibility of any close correspondence between the two. With this brief 
introduction, it is possible to proceed with the analysis of the problem. 
Susceptibility to Shedding. 
The internal factors responsible for the marked susceptibility to 
shedding which is noticeable during the later portion of the flowering 
period have apparently received in the past little or no attention. Balls ( 1 ), 
however, recognized two causes, constitutional and environmental, at work 
in effecting shedding. Lloyd (6), on the other hand, though emphasizing the 
fact that a progressive increase occurred in the proportion of flowers which 
were subsequently shed, ascribed this change to the gradual reduction in 
the moisture content of the deeper soil -layers. ‘Shedding’, he says, ‘is 
always a response to untoward conditions.’ 
The lower graphs in Fig. 5 represent the' approximate percentage of 
flowers produced on any day which were subsequently shed ; they were 
obtained from the observed rates of flowering and of boll-shedding by 
assuming that a period of ten days elapsed between the open flower stage 
and the completion of abscission ; ten days was, as a matter of fact, the 
mean period in the first group ; the records of the second group do not, 
however, permit of an exact estimate being formed. As the results are 
smoothed to four-day intervals the error introduced by this assumption is 
not of grave consequence. 
The mean daily growth-rates of the main axis of both groups have 
been smoothed for a similar period and are also reproduced in Fig. 5. 
Inspection of the two sets of graphs reveals the fact that the percentage 
of flowers which subsequently underwent abscission remained relatively 
insignificant until growth of the main axis had almost ceased. It is 
difficult to escape the conclusion that the fluctuations in the physical 
environment which generally herald waves of shedding are not of marked 
