60 THE ROYAL SOCIETY OF CANADA 
while those which increase the availability of soil-moisture, as by bring- 
ing more water into contact with the superficial root-system, or by 
decreasing the flow of water from the plant, ameliorate its condition 
for the time being. As the season advances, however, the mean of 
these fluctuations in shedding rates is always increasing, the change 
in the mean being the index of a steadily increasing shedding rate due 
to the gradual depletion of subterranean soil moisture. The end of 
this increase is determined by the material available for shedding 
and the stringency of conditions which obtain before the bolls have 
obtained a size which enables them to resist abscission. Even when 
this happens, the conditions imposed by the environment may still 
cause hindrance to the maximum development of parts as, e.g., the 
cotton fibre, although with this the present paper does not deal. 
The amount of boll shedding and that of square shedding appear 
to be nearly equal, relatively to the forms available for shedding, 
aside from that additional boll shedding due to inadequate fertilization 
and possibly also to other undetermined causes. It has been proved 
that rain falling in the day-time, especially at or near certain hours, can 
procure a large amount of boll shedding, amounting to 70 per cent. of 
the flowers in an observed instance. From data for the history of 
individual flowers (1912) it has also been shown, with a very small 
probability of error, that high maxima (involving the loss of 100 
per cent. of the flowers) do occur from this cause. In the absence, 
however, of the individual records of flowers and their subsequent 
history day-by-day, it is never possible to do more than estimate the 
aaximum possible effect, and this usually teaches us very little, that 
little becoming more perhaps when viewed in the light of concurrent 
evidence. Any study of shedding in a region where there is repeated 
rainfall, must, in order to be adequate, include the subsequent history 
of a sufficient number of the particular flowers subjected to day-time 
rains. 
The view that the limitation of the water supply, and therefore 
of the water content of the plant, is answerable for shedding, does not 
however rest merely on the data derived from shedding records alone. 
A study of the transpiration rates, water deficit in leaves and stem dur- 
ing rainless spells, daily cessation of growth or measurable shrinkage of 
tissues and the closure of stomata, all indicate quite clearly that there 
seldom occurs a day on which there is no minus water fluctuation in 
the plant. It is certain, however, that the water deficit is less on, 
or just after, a rain-day, but can become very pronounced in the course 
of a few days thereafter. When this takes place shedding at relatively 
high rates intervenes within the expected period. The effect is not, 
however, a direct one. The studies upon which these conclusions 
