130 MR. S, T. DUNN: A REVISION 
greatly in number, shape, size, and pubescence, provide the greatest evidence 
of specific difference and are widely used for the superficial discrimination of 
Millettias. The hairs are always simple. In the Zwlgentes and a few other 
groups they become flattened and scale-like on the lower surface of the 
leaflets, giving them a fine metallic silver or golden lustre. 
The stipels, if present at all, are, in all cases known, persistent up to 
maturity and have proved a valuable guide to the divisions into which the 
genus naturally falls. 
Inflorescence.—True racemes, in the simplest sense of the word—v. e., single 
peduncles bearing pedicels scattered singly along their length,—do not occur 
in Millettia as here defined. No species has been included in the genus 
which has solitary racemes terminating a leafy shoot nor in the axils of a 
shoot terminating in leaves. This is the most effective distinction that I am 
able to discover between Wistaria (restricted as above) and Millettia, and 
also between the American Gliricidia and Millettia. Some species of the 
latter genus bear their flowers in nearly simple racemes, but in these all 
the minute nodes are two- or more-flowered. The Philippine Fragilifloræ 
contain many such species, and similar ones occur in the African Sericanthe. 
At the next step away from the simple raceme, fascicles of flowers occur on 
sessile globular nodes, as in the Afroscandentes in Africa and the Typicw in 
India. The above forms of inflorescence are generally (as in the diagnoses 
below) called racemes, because no branching in a general sense occurs. 
When the nodes become a little lengthened, so that the flowers are crowded 
upon a eylindrie spur (ramulus floriferus), as in most of the fulgentes, or at 
the top of a short peduncle, as in M. Brandisiana, the inflorescence is really 
panieulate, though little different in appearance from a raceme. It may 
then be called panicula racemiformis. When the branchlets are still further 
elongated and more similar to the panicle rachis, the flowers become more 
spaced and form panicles in the general sense, sometimes Jong and narrow as 
in M. leptobotrya, sometimes broad and spreading as in the Luryhotrye. 
Flower.—The calyx is usually five-toothed and stiffened on the upper side 
by the coalescence of the two anterior teeth nearly to the top. In the notch 
thus formed the standard rests at about the top of the claw, while the blade 
is strongly reflexed over the top of it as the flower fully opens. This ensures 
the maintenance of the standard base in close proximity to the bases of the 
lower petals and to the staminal sheath, which is essential for the proper 
working of the floral mechanism, if I rightly understand it, during pollina- 
tion. In the Fragiliforw and some other sections the calyx is wide and soon 
leaves the petals quite free. The bracteoles which enclose the young flower- 
bud are usually deciduous, but they persist in some sections at the base of the 
calyx or a little below it on the pedicel. In the Eyulgentes they are usually 
large and wide, and persist during flowering-time ; in some of the Serteanthw 
they are of unusual length, sometimes overtopping the flower. The pedicel 
