174 THREE NEW CHINESE ASTERACEX. 
plicibus vel paniculatis interruptis, capitulis sessilibus 2-7 aggregatis 
circ. 3 lineas longis, involucri squamis flores eequantibus linearibus 
acutis exterioribus brevibus omnibus extus dense et pulcherrime ame- 
thystino-villosis intus glaberrimis nitidis, floribus fem. numerosis 
hermaphroditis circ. 20, corollis luteis, achzeniis nitidis parce pilosulis, 
pappo albo.—Ad fauces Shiu-hing, secus fl. West River, prov. Can- 
toniensis, Febr. 1866-7, collegit T. Sampson.  (Exsice. n. 12815.) 
An exceedingly beautiful plant, remarkable for the brilliant amethyst- 
coloured involucres, the tint frequently extending to the whole 
pubescence on the inflorescence; all the specimens gathered in two 
successive years quite alike. Its relationship is with B. hieracifolia, 
De Cand., and its allies. This genus may, in Asia, be compared for 
intricacy to Hieracium in Europe. I feel persuaded that Mr. Bentham’s 
ideas (Fl. Hongk. p. 78) of the instability of the majority of the 
species are erroneous, for many of the South Chinese forms appear to 
me quite constant, though certainly not easily characterized in words. 
adeo 2. Gnaphalium Amoyense, n. sp.; sesquipedale, floccoso-album, 
caule erecto subsimplici, foliis confertis linearibus 14-pollicaribus 
lineam latis semi-amplexicaulis non decurrentibus apice. mueronulo 
calvo eastaneo apieulatis supra dense tomentosis subtus densissime 
albo-pannosis, corymbo composito terminali laxo corymbulis densius- 
culis, involucri squamis oblongo-linearibus apice obtusis sepe eroso- 
denticulatis inferne testaceis lana alba obductis supra medium nitide 
eitrinis.—In collibus cirea Amoy, ipse legi, m. Octobri 1857. (Exsice. 
n. 1420. 
I had taken this for the Japanese and Martaban G. confusum, 
De Cand., and sent it to friends under that name ; Mr. Bentham, 
however, informed me of the error, and has since then, in the ‘ Flora 
ongkongensis,' referred De Candolle's species as a synonym to G. 
multiceps, Wall., which, again, F. Mueller has recently (Fragm. Phyt. 
Austr. v. 150) reduced to G. luteo-album, Linn., which to me seems 
quite distinct. The Amoy plant, in Mr. Bentham's judgment, is 
closest to G. hypoleucum, De Cand. 
ugut 3. Senecio (Obajace) exul, n. sp.; glaberrimus, caule a basi decum- 
bente erecto ramoso angulato striato-sulcato, foliis inferioribus petiola- 
tis superioribus sessilibus oblongo-lanceolatis semiamplexicaulibus 
pinnatipartitis laciniis remotis dentatis, corymbo composito laxiusculo, 
involucri campanulati esphacelati phyllis lineari-subulatis acutis 
