OF INDIA AND CEYLON. 
201 
first, afterwards with scaly bracts formed by a meta- 
morphosis of the leaf -bases, and equal-valved 8- (or 9- to 12-) 
ribbed fruit, both of whose lobes persist. Taking all 
characters together, the Dicræas are a very distinct group, 
probably, from their isolobous fruit, as old at least as the 
Podostemons proper, and I am inclined to retain the genus, 
not to merge it in Podostemon. Possibly P. Barberi comes 
near to Dicræa, but it has an unequal-valved fruit and no 
bracts, so that the resemblance is not so close as at first 
thought appears. The South American genus Ceratolacis 
comes very near to Dicræa, and perhaps should be united 
with it, but detailed study of the former is required. 
There remain now the two very peculiar species, selagi- 
noides and Hookerianus, placed under Podostemon in 
Hooker’s Flora, under Dicræa and Podostemon by Weddell. 
If either is to be placed in the former genus, its special 
characters require complete revision, because both have 
smooth fruits with one deciduous valve. Their characters 
differ in many other points from those of the Dicræas, and 
there seems to be no reason for placing them in that genus. 
Here, as so often in dealing with the Asiatic species, 
Weddell’s work bears marks of haste and want of due 
consideration, so necessary in a puzzling group like this. 
Bentham and Hooker place both of these species in Podoste- 
mon, apparently simply because Weddell placed them in 
Dicræa and Podostemon, genera which they unite. They 
differ as much from the rest of the Podostemons as from the 
Dicræas. Warming gets over the difficulty by giving each 
a separate genus (Griffithella Hookeriana, Willisia selagi- 
noides), and they are so distinct from most other forms of 
the family that this course is perhaps the best until we have 
a fuller knowledge of the group. They show relationships 
to Podostemon, Mniopsis, and perhaps other genera, and are 
probably ancient survivals. 
The genus Polypleurum, Wmg., is, I think, untenable, 
resting on insufficient knowledge of the morphology of the 
species proposed to be included in it. The position of the 
