132 MR. S. T. DUNN: A REVISION 
projections at the bases of these petals are hooked into the upturned auricles 
of the base of the standard. 
The staminal whorl, as has been said, is either monadelphous in this genus 
with the vexillary stamen free at the base * or diadelphous, the nine lower 
stamens being united by their filaments into a sheath, the upper one quite 
free, The cohesion of the stamens is nearly always constantly either 
diadelphous or monadelphous at anthesis throughout each section. Тһе 
character was much used by the earlier botanists in discriminating genera, 
but, in Millettia at least, it is not found useful to rely upon it, as it is not 
easy to ascertain whether the vexillary stamen is permanently attached to the 
rest or free at complete maturity. The vexillary stamen seems always to be 
inserted a little within the others, and where they are separated from the 
base of the ovary by a perceptible space is sometimes continued down to its 
base by a raised rib or line. Its lower extremity is usually broadly winged 
so as to close the very base of the staminal sheath, and the wing then 
overlaps the neighbouring portions of the sheath, being outside it on one side, 
inside on the other. In not a few species the petals are coherent with the 
sheath at its base, though free from the ealyx, so that the wings and keel 
appear to be inserted on the sheath while the vexillary stamen is adherent to 
the base of the claw of the standard. 
A dise is frequently present. In many sections, especially in those having 
stalked ovaries, it is a short fleshy tube closely investing the stalk or the 
lower part of the ovary. It may be truncate or oblique or lobed, or may be 
reduced to a ring. In some of the African sections it is adnate to the lower 
part of the calyx-tube, appearing as a plane or lobed rim at the base of the 
perigynous stamens. In the H/ulyentes the adnate part is longitudinally 
ribbed, the ribs running out at the top into a fringe of blunt linear lobes. 
The ovary is linear, containing from three to many ovules and ending in a 
curved, usually glabrous, and sometimes flattened style and a glabrous 
pitate stigma. 
The pod, in the great majority of cases, is elastically dehiscent into two 
valves at maturity, but in some species the dehiscence is so much delayed 
that the impression is given that no natural opening occurs along the sutures, 
as in the case of the species upon which the genus was founded by Wight 
and Arnott, who described it as having indehiscent pods. In a few species, 
such as JM. sericea, the dorsal suture does not usually open, and in a few 
others, such as M. pachycarpa, the valves are apparently soft and leathery at 
maturity and do not separate until the fruit dries up, but otherwise the 
whole genus is remarkably constant in the elastie woody nature of the pod 
valves and in their prompt dehiscence. 
* The vexillary stamen in that case usually becomes quite free as maturity is passed, and 
this may aid cross-fertilization, because the weight of the insect will rest on the anthers 
while the sheath is closed, but later when the sheath is split along the top the insect will 
push it down with the petals until its body rests on the stigma. 
