IN CEYLON. 
465 
and one would expect that a species would, by preference, 
pass through its leafless phase at that time. But what do 
we find ? In the first place some deciduous species are in full 
possession of their new foliage during the hot dry season, as 
Bassia longifolia ; in the second place some produce all their 
new leaf when the temperature and dryness of the air is at 
the maximum, as with Swietenia Mahogani, Careya arborea, 
Ficus asperrima, Gmelina arborea, Sterculia Balanghas, 
Phyllanthus indicus, Spondias mangifera, Terminalia 
belerica, Sterculia Thwaitesii, Lagerstrœmia Flos-reginæ, 
and others, and finally trees of Albizzia procera and often 
of Pterocarpus echinatus defer their leafless phase until the 
wet cold months of June and July, when transpiration and 
assimilation are usually near the minimum. Such species 
show that they are either able to carry on their inherent 
foliar periodicities in the face of a periodic and exacting 
climate or that they are less plastic than others. It maybe 
that they have, in the migration of species, found themselves 
at a place where the climate is not in harmony with their 
original periodicity, and time will be required for an 
adaptation to be accomplished. From these and other 
considerations, it may be concluded that the seeming irre¬ 
gularity is an expression of the autonomy of the species, 
and is not likely to be changed by the climate under con¬ 
sideration. 
There are several other observations which seem to 
suggestthat the internal factors are of great importance. 
If one tree be selected and the length of its leafless 
phase be observed from year to year it may transpire that 
there is a variation considerably greater than what one 
expects to find. One species may vary from 32 to 96 days 
for its leafless phase, another may vary from 2 to 14 days 
during different years, or a particular tree may not always 
pass through a perfectly leafless stage each year. In dealing 
with the frequency of defoliation several cases have been 
cited which thus indicate an independence of environ¬ 
ment. The differences in such behaviour cannot always 
