OF CEYLON AND INDIA. 
311 
No haptera were ever observed to form. The next stages are 
shown in PI. IX., figs. 5, 6. Here the hypocotyl has formed 
a kind of tuber, on the upper surface of which a few leaves 
have developed. There is no elongation of the epicotyl in 
the ordinary way. The order in which the early leaves 
appear is very difficult to make out ; only a small quantity of 
material in very few stages was at my disposal, and the 
plants are exceedingly small and delicate. I am inclined to 
think that the order is no very definite one ; sometimes, and 
perhaps most often, the first two leaves are more or less 
at right angles to the cotyledons, which are themselves 
either opposite or approximately so, but I have found cases 
in which the first two leaves were approximately parallel to 
the cotyledons. Sometimes there seems to be an indication 
of an approximate tristichous arrangement of the leaves, 
but it is not very close, and is probably accidental. The 
phyllotaxy is probably a complex many-ranked one (very 
likely the same as in Tristicha), and sufficient leaves are not 
formed to enable one to recognize it. In a very short time 
it is evident that the growth of the seedling is taking place 
in a lateral direction, that a marked dorsiventrality is appear- 
ing, and that there is no more development of leaves in the 
vertical direction at all. The direction of growth is nearly 
always at right angles to the plane of the cotyledons, and 
horizontal along the rock. All the newly-formed leaves are 
now seen to lean in one direction, that in which the apex is 
growing. PI. IX., fig. 7, shows a seedling with seven leaves 
besides the cotyledons, in which this growth is already 
becoming indicated, and in fig. 9 it is clearly evident. In 
this specimen, which is seen a little obliquely from above, 
one of the cotyledons is absent ; there are four larger lateral 
leaves somewhat divergent, and a considerable number of 
small linear leaves all pointing nearly in the same forward 
direction, while the tuberous hypocotyl is also evidently 
elongating. The same stage is shown in side view in fig. 8, 
and shows this growth of the hypocotyl into the “ shoot ” 
thallus. 
