372 159. GRAMINE. 
margine exteriore, in parte racemi superiore, piloso, pedicelli 
2 ad 3 articulorum longitudini &quantes, utroque margine cum 
pilis patentibus instructi. Spicula sessilis lineari-lanceolata 
acuta, 2-23 lin. longa; gluma i. dorso subtiliter tuberculata, et 
suleata, apice breviter acuminata et biaristulata, marginibus 
implieatis superne ciliolatis, callo brevissimo (vix d lin. longo) 
barbato ; gl. ii. eymbiformis, primam subequans, apice leviter 
aristata, arista vix 2 lin. longa; gl. iii. minuta (4 lin. longa), 
hyalina, late elliptiea, obtusa ; gl. iv. minuta (vix 1 lin. longa), 
hyalina, breviter oblonga, apice in aristam debilem, imperfectam, 
purpureo-tinctam, 6-7 lin. longam, producta; palea absens; 
caryopsis anguste ovalis, pallide brunnea, glabra, 1 lin. longa. 
Spicula pedicellata quam sessilis paullo minor, ceterum similis, 
mox tabescens. 
Formosa: on cliff, Takow (A. Henry, 1167!). Herb. Kew. 
Near A. vagans, Rendle, resembling in habit the more delicate 
form of that species. It is distinguished by the narrower 
racemes and spikelets, the awned second glume, and the larger 
but weak and imperfect awn of the fertile glume. 
Var. minor, Rendle, n. var. — Rhacheos articuli pedicelli et 
spicule quam in specie breviores; spicula sessilis cum | callo 
brevissimo vix 12 lin. longa; gluma 1. obtusa. 
HowcKowsa: Climbing in hedges near Aberdeen ( Hance, 
10746 in part!). Mus. Brit.; Herb. Kew. 
9. Andropogon fragilis, R. Br. Prodr. p. 202. 
Var. sinensis, Rendle, n. var. Planta annua quam typus 
minus robusta, eum spathis florentibus paullo angustioribus. 
Spicula sessilis gluma i. dorso ad medium pilosa, non velut in 
specie glabra. 
KwaNvarUNG : French Island, Whampoa (Hance, 1388 (?) !). 
Herb. Kew. 
The type occurs in Queensland and North Australia. 
The species is included by Hackel as a variety (8) of Andro- 
pogon brevifolius, from which, however, it is, I think, quite 
distinct. Brown founded his species on plants collected during 
Captain Cook's first voyage by Banks and Solander at Endeavour 
River. Professor Hackel has probably not seen the original 
specimens at the British Museum ; they differ from A. brevi- 
