200 NATURAL HISTORY OF PLANTS. 
Spatholobus. Excepting the last, which though rendered by its 
fruit and vegetative organs inseparable from Butea, presents the 
general organization of the Galactica, all these genera have the 
inflorescence of the Galactiee, or of the Huphaseolee or Vignee, with 
flowers whose standard is shorter than the wings or keel; or else 
very large, with the wings very short, or in any case shorter than 
the keel. The style is beardless and the floral bracts are small or 
caducous. 
The small group Dioclee has the same inflorescence and bracts. 
The calyx is usually four-lobed, with the posterior lobe broadest, or 
else it is equally bilabiate. The vexillary stamen, free at the base, 
joins the rest above; the style is beardless. The five genera Dioclea, 
Camptosema, Cleobulia, Pueraria, and Canavalia constitute this sub- 
series. 
In the group Glycinee the secondary axis of the inflorescence loses 
Kennedya prostrata, the knot-like appearance of the preceding sub- 
series. The vexillary stamen may be either free 
or united with the rest. The standard is 
altogether without appendages, or else its edges 
are only slightly inflexed at the base. The style 
is glabrous except in Clitoria. This group con- 
tains the eleven genera: Glycine, Shuteria, 
Fie. 156. Fra. 157. Teramnus, Kennedya (figs. 156, 157), Dumasia, 
Fruit. Longitudinal Amphicarpa, Cologania, Periandra, Centrosema, 
section of fruit. Ofjtoria, and Platycyamus. 
In Cajanee, too, the nodes of the inflorescence are wanting. The 
flowers have no lateral bractlets ; the vexillary stamen is free; and 
the style is beardless with a terminal stigma. The leaves are 
covered, at least underneath, with resinous dots, and the stipules of 
the leaflets are small or absent. In this last subseries we have 
the eight genera: Cajanus, Fagelia, Atylosia, Dunbaria, Cylista, 
Rhynchosia, Eriosema, and Flemingia. In the four last the number of 
ovules and seeds is always very limited.’ 
2 The characters separating these six sub- artificial and in no way absolutely distinguished 
series are very far from being absolute, as, indeed, from its neighbours. 
might be expected in a series which is itself quite 
