123 
known from Hongkong; an imperfectly known species from 
Formosa, an endemic species in the Fiji Islands and four in 
Australia, including a new tree species now described for the 
first time, complete the enumeration. 
CLASSIFICATION. 
For purposes of classification the Indian and Eastern species 
may be arranged in four groups or sections, two of which appear 
certainly to be natural subdivisions of the genus. 
The first section (Brevitubae) is distinguished by the very short 
corolla tube, so much so that in some species the corolla appears to 
be almost polypetalous. In the various species of this section the 
corolla may be glabrous or hairy within, the anthers glabrous or 
bearded at the base and the ovaryand style either hairyor glabrous. 
The second section (Lanigerae) also contains species hairy or 
labrous as to anihers, ovary and style, but is differentiated by 
the corolla being provided at the throat with long woolly hairs, 
and in the corolla tube being about equal in length to the lobes. 
These two sections contain some thiry-five species, but it is prob- 
able that they are not perfectly natural groups, and some of the 
short-tubed species with woolly hairs at the throat should per- 
haps be placed among those where the corolla tube approximates 
in length to the lobes. S. angustiflora from Hongkong finds a 
place in the key to this section, owing to the splitting of the 
corolla tube producing lobes equal in length to the tubular part 
of the corolla; but its proper affinity is no doubt among the long- 
tubed species near to S. Nu«-vomica. 
The third section (Penicillatae) appears to be quite a natural 
and well-defined group, being differentiated by the occurrence 
of a line of erect bristle-like hairs either across the base or 
across the middle of the inner face of the corolla lobes. Of the 
pare the inter-relationships between the Indian and Eastern 
species and those endemic.in Africa and America. This, how- 
ever, is unfortunately not possible until a complete re-examina- 
tion of the African and New World material has been made. In 
Africa alone some eighty species have been recognised, and many 
of those collected in Tropical America have not yet been pro- 
perly determined, — ve a 
e fourth section of the Indian and KEastern species 
a2 
