164 CLXxiii. GRAMiNE^. (J. D. Hooker.) [Coelarthm. 



a solitary erect spike ; lowest pair on the spike imperfect ; joints of 

 rachis shorter than the sessile spikelet, clavate, tubular, tip truncate 

 open, mouth entire compressed. Sessile spikelets dorsally flattened ; calla3 

 very short, bearded; gl, I ovate-oblong, tip narrowly truncate, doraallr 

 glabrous with a deep median channel, margins narrowly inflexed above 

 broadly incurved below, keels subpectinately ciliate above the middle; 11 

 cymbiform, 1-nerved, awned ; III oblong, 1 -nerved ; IV very small, 2-na, 

 3-nerved, awn slender, twisted; palea minute, Lodicules 2, cuneate^- 

 quadrate. Stamens 3. Styles 2. Grain fusiform, Pedicelled spikelets -^^i 

 the sessile or less, male or neuter gls. 2-3, not awned ; pedicel linear, flat. 



A remarkable grass, differing from Andropogon in the secund spikelets, holl<>* 

 joints of the rachis, deeply cliannelled gl. I of the sessile spikelet (as in Sect. Seh\)ii» 

 of Isckcemum) J and the reduced pedicelled spikelets. 



C. Brandisil, Hoolcf. in Hook, Ic, Plant, ined. 



KuRMA ; high ground, generally with Teak, Brandts, ^ . , 



Steyyis very weak, branching, struggling, rooting at tbe nodes, white i ff' 

 nscending filiform 1-2-leaved tiowering branches. Leaves 1-2 by ^-i m-* '^^^^ 

 on the sheath by a mere point ; acute or jicuminate at both ends, membranODj. 

 fiac<'id, surfaces and margins quite smooth, midrib and nerves very slender; sbea 

 glabrous; ligule oblong, obtuse. Spike 1-H in., strict or incurved, quite wm <*• 

 Sesiile spilcelets i in. ; gl. I with a strong median keel on the ventral ^"^^^^^'ttt 

 notched at the tip, awn as long as the gl. or shorter, very slender, recurved; 

 flat, ciliate; awn of IV f in., very slender, geniculate, column brown twisted, upp 

 half very pale. 



53. ANDR0P060N, Linn 



id 



Grasses of varions habit. Leaves nsnally narrow, base rarely roun 

 or cordate. Ivflorescence of solitary binate digitate fascicled or P^^J^. 

 spikes (which, are sometimes reduced to 3 spikelets), rachis *^^'^ tli- 

 nsnally fragile. Spikelets binate, a sessile fern, or bisexual, and ^ P j- 

 celled male or neuter. Sessile spikelet l-fld-, glnmes 4, I dorsally ^^^^^^ 

 laterally more or le3s compressed, margins inflexed or incurved, ^^^%^[ 

 keeled ; II as long as I, usually laterally compressed, awned or not; ^^ 

 hyaline, empty, 2-nerved or nerveless, epaleate ; IV" hyaline, ^"^^^j^g 

 narrow, 2-fid and awned in the cleft or reduced to an awn more or ^^ 

 dilated at the base, paleate or not. Lodicules cuneate-quadrate. ^ • ^^ 

 3. Stigmas laterally exserted. Grain free, Pedicelled spikelets vari 

 gls. 3^, IV awnless. — Species about 200, chiefly tropical. 



venturing to depart far from his grouping of the Indian species, to attempt "> 

 with any prospect of success would demand a revision of the whole genus. /^* -j^jj 

 two principal groups j(Series), Isozygi and Heterozygi, unfortunately ®*^P;'^^ we^ 



subgenera of clo?e 



In this most difficult genus I have followed Hackel as closely as I conl ^.^.^^ 



\ ~ — ^ "^^ ^ ^ ^^ ^ MM ^'^ ^^ ^^ ''^ ^1 ^^ ^ ^^ ^L^ "^jm ^^ ^L ^^ V ^^ J ^^ JW %M \J V A ^A M^ & ^J & Iff %^ ^^^ m m ^ -^ - _^j m 1 jih 4'%T VlU 



affinity in habit and in all other characters, except that ^ ^^ 

 to distinguish between these groups, Thns the first two or even 4 ^^"^^^ ^hQst 

 Isozygi are closely related to Cymhopogon amongst Helerozygi, and not to ^^^ 

 that immediately follow them in the series. Added to this, on ^be one h« ^^^ 

 characters drawn from the form, sexuality, &c., of the lower sessile ®F^^^ j yerf 

 Heterozygt as distinguiahing them from those of Isozygi are often variable an ^^ 

 difficult of detection ; and on the other hand the same characters do occur !^^ gfj, 

 there in several subgenera of the latter group. With regard to Hackel's subg ^^^ 

 they appear to me to be on the whole well constructed and natural, thoug ^^^ 

 depending on characters, some inconstant, others difficult of detection or *PP*^.j.ggie<i 

 Of the best defined I should, following older authors, have preferred to hate 





A 



t- - 



