64 DEVELOPMENT OF THE FLOWER 
The summit appears of an uniform cellular grumy charac- 
ter, even the last pair of bracts offer no exception ; thence 
the next bracts will be found concave, provided with 
glandular hairs, those on the back being developed first : 
the axils of these shelter a convex grumous cellular body, 
presenting no traces of divisions. 
In the next axils, the first indications of the sepals may be 
seen, the posticous one taking the lead in development, and 
so on 
The calyx arrives at some size and development before 
the corollula parts become united, and it preserves this ratio 
for a long time. ` 
The corolla rapidly becomes monopetalous; the next 
change is irregularity, the lower 4 being much more develop- 
ed, so that a lateral view presents an obliquity of the top. of 
the corolla. The lower lip is outermost in these periods. 
The anther before the union of the petals have a tur- 
gescent quadrilocular look. "They are then nearly sessile, 
and broadly sub-reniform, but quite equal. 
The lower end of the lobes now appear to take an increased 
growth, no change taking place in the original line of union 
with the connectivum, at the same time the filament begins 
to form. These two growths going on, the next stage pre- 
sents the anther with separate loculi attached by their upper 
ends to the connectivum, and taking an unequal growth. 
The pistil is the last part to appear; in its first state it 
looks as if the edges of the somewhat depressed central part 
of disc was fashioned up into a rim. I have never seen in it 
two distinct leaves, although the orifice presents traces of 
binary composition. The irregularity takes place very 
early, but the parts are so minute that I cannot say whether 
it occurs before or after the union of the petals. 
At one time the corolla looks like a pentangular disc, with 
four elevations within four of the faces of the angles (the 
stamens); at this time there is no trace of pistil. 
If the stigmata and production of the placente be marginal, 
the Lindleyan definitions of the style will require alteration. 
But Lindley himself is not exactly consistent with regard 
