The present paper is an attempt to throw some light on two Composite genera,
Cavea W. W. Smith et Small and Nannoglottis Maxim., of which the generic features
and the systematic position were heretofore rather badly known.
Cavea, a monotypic genus based upon Saussurea salwinensis Drumm., was described
by W. W. Smith and Small as monocious. From the ample material now in our com-
mand, it is found that the genus is either monocious and dioecious, the capitula in the
latter case being homogamous. The individuals exclusively provided with sterile flowers
are usually smaller in size, sometimes even stemless. Such repartition of sexes was
known in the subtribe Plucheinae (tribe Inuleae), to which W. W. Smith and Small were
correct to ascribe their genus.
Nannoglottis was considered by Maximovicz as un abnormal genus of the tribe
Inuleae, having heterogamous capitula with one series of peripheric fertile ligulate pistil-
late flowers. Having examined the type specimen of N. carpesioices Maxim. and the ex-
tensive collection from Kansu and Tsinghai, we incline to say that the genetic characters
as discerned by Maximovicz are doubtlessly due to his inaccurate observation. The
flowers of Nannoglottis are in fact trimorphous, of which the pistillate ones are 2-3-
seriate, the outer being shortly ligulate and the inner, short-tubulate with truncate apex.
Handel-Mazzetti, much later, indicated that, in another species, the pistillate flowers are
2-3-seriate, but no mention was made about the inner tubulate pistillate flowers.
Franchet, appearing to be unware of Maximovicz's genus, founded his Stereosanthus
Franch. and considered it as a genus intermediate between the tribes Inuleae and Sene-
cioneae. The genus was misinterpreted by Franchet as having dimorphous flowers, the
pistillate ones being all ligulate. Curious enough, one of the major generic features, viz.,
the presence of inner tubulate pistillate flowers, was neglected by all early botanists and
was first recognized by Handel-Mazzetti fourty years later.
However, Handel-Mazzetti was still unacquainted with the characters of Nannoglot-
tis and proposed therefore his new genus Vierhapperia Hand.-Mzt. on the reason that the
flowers are trimorphous. Moreover, he compared his genus with Conyza L. and Erigeron
L. of the tribe Astereae, but the characters marked by him appear not to be different
from those of Nannoglottis.
From the above morphological comparaison, it is evident that Nannoglottis and two
other genera in question possess in common important characters in the general appea-
rance, the flower-forms and the structure of achenes and pappus. The differences in the
disposition of involucral bracts and the length of ligules are criteria good for separating
species, but can hardly be regarded as sufficient for generic delimitation. Moreover,
these three genera are almost similar in their geographical distribution. These points are
adequate to warrant that Stereosanthus and Vierhapperia are in reality congeneric with
Nannoglottis.
Nannoglottis appears to have been correctly placed by Hoffmann in the subtribe
Senecioninae (tribe Senecioneae) though in that subtribe it finds no close relatives. On
account of the outstanding characters of the genus, a further study of its systematic posi-
tion is required.
The present paper is brought to close by a tentative scheme of classification of the
genus so amplified, of which two sections, namely, sect. Stenolepis Ling et Y. L. Chen
(Stereosanthus Franch., p. p. min.) and sect. Nannoglottis (Stereosanthus Franch., p. p.
maxim.; Vierhapperia Hand.-Mzt.), altogether including 9 species, are being proposed.
