FLORA AMERIC/E BOREALI-ORIENTALIS. 359 
As we approach the flower a gradual change takes place in the leaf, 
and, more especially, in its size; but still we have even an inerea 
preponderance of the vascular system and its glands; in some cases 
the bracts being reduced to a mere stalk, with one or two large glands. 
But when we come to the floral envelope, a sudden change takes place, 
not only in their position on the axis, but in their texture; the vascular 
system, especially in their petals, is reduced to extreme tenuity, and the 
cellular tissue is proportionately more developed. Glands, in their 
quently, very irregular shapes and a petaloid consistence. In the 
leaves forming the andrecium, a partial return to the system of stem- 
leaves takes place, inasmuch as the filament is entirely reduced to the 
vascular system, its glands are conecte. into aane. mé the cellular 
onnectiv yum, 
ccasionally 
or slight membranous expansions of the filament. The carpellary 
leaves have, again, a more or less developed cellular parenchyma, as 
well as a strong vascular system, the glands becoming sometimes 
prominent, glanduliform stigmata, or sometimes papillose stigmatic 
surfaces of extreme tenuity. 
Such being the close morphological analogy between the stamen and 
the stem-leaf, we would next observe that multiplication takes place in 
the latter almost universally i» one plane (either horizontal, or, from 
various causes, more or less vertical), by the ramification or the separa- 
tion from the base of its ribs, each branch being more or less con- 
nected with the remainder of the leaf, or assuming the form of a 
distinct leaflet, Wherever a ¿uft of leaves occupies on the stem the 
place of an ordinary leaf, it is not by ramification of the leaf, but 
by the partial development of the axillary bud of which the axis is not 
elongated. Fd 
Proceeding now, upon these grounds, to compare the modifications 
of the stamens to those of the stem-leaves, taking the ordinary stamen 
to represent a petiole (the filament) with a gland on each side of its 
apex (the two cells of the anther), we see these glands 
united into one (the one-celled anther by confluence), or reduced to 
one by abortion (the dimidiate anther), or sometimes really single and 
terminal, or increased to four or more in pairs, or two or more super- 
posed or irregularly arranged as in the stem-leaf glands, so in the 
