HYPERANTHERA. "573 
. *minais a good deal analogous to that of Moringa, in which if 
10 were developed the uppermost would as in Leguminosz be 
the smallest. "The opposition of the two lowermost to the 
carina may be explained by assuming the obliteration of the 
proper Ist stamen, and the approximation from pressure of 
the two normally alternating with the carina, so that they 
become opposed to.it, of this opposition by pressure other 
instances occur. 
- _ Of course this view is only reasoning on the idea, that the 
petals are properly developed, if not so, then the opposition. 
of these is at once accounted for. . 
X. flavescens approaches Moringa in its pluriovular ova- 
rium, 
. No stipulz exist, unless they are represented by glands. 
_ The great differences it has with Leguminosze exists in the 
uniloeular anthers, the position of the 5th petal which is an- 
. ticons, the compound nature of the pistillum, and the nature 
of the ovula, which in Leguminose are campulitrope, and 
in this anatropa, that is the base of the nucleus corresponds 
to the apparent apex of the ovulum. 
The great objection to the idea of DC. that Moringa is a 
.. Constantly tricarpellar Leguminousa, I take to be that, if this 
.. ally were the case, each carpellum would be distinct, since 
 inallcases in which an additional carpella is formed in 
Leguminose, it is always so, hence a tricarpellar Leguminosa . 
-Ìt ever be found, will be apocarpo-tricarpellar, not as in 
: this genus, syncarpo tricarpellar. . 
3 . With Samydacez it has, but little in common, although 
: *xternally there are no signs indicating ternary composition 
: t the style, yet the shape of its cavity as well as the dis- 
tribution of its vessels into three fascicles, point this 
: fat. The stigma is cup-shaped, or rather none exists, 
the communication with the interior of the ovary, is free and 
na inct, and the pollen comes in immediate contact with 
Mucous conducting tissue by which it is filled; such a 
ure is compatible with placentary stigma, but incom- 
