2 CLXXIII. GRAMiNB-ffi. (J. D. Hooker.) 



(1883), a work of remarkable completeness and accuracy, considering' the chnotic 

 condition in which the author found the Order; and Hackel's admirable monograph 

 of the Andropogonece in A. de Candolle " Monogr. Phanerog." vol. vi. (1889%- -the 

 largest and most difficult Tribe of grasses. 



I have cited synonyms to a much greater extent in this than in any other Order 

 described in the "Flora of British India" (except the " Cyperacese" by Clarke) 

 with the view of aiding the researches of future authors. Many of these syuofiyuis 

 had never been fixed with any approach to exactness, and many were for the first 

 time identified during Mr. Jackson's and my labours on the " Index Kewensis." L 

 cannot expect that all will prove to have been satisfactorily reduced, aud still less 

 that all have been brought to light. 



Deferring to the classification of the Indian genera, I have been compelled to 

 diverge somewhat from Bentham's arrangement, and to abandon some of his Tribes 

 and Subtribes, in many cases following Hackel's more recent and well-considered 

 views. These divisions throughout the Order are of very unequal value, and are as 

 difficult of delimitation as of definition upon any fixed principles. They are, indeed, 

 in some cases arbitrary, and in others more or less artificial. In this matter, and in 

 the description or revision of some of the very difficult genera, I have derived great 

 aid from Dr. Stapf, Assistant in the Herbarium of the Eoyal Gardens, who has 

 la tely been instructed by the Director of Kew to study and name, the materials in that 

 Herbarium critically. As Dr. Stapf shares my views as to the wide ranges of the 

 species, and the undue multiplication of their synonyms, his labours may be expected 

 to yield far-reaching results, taxonomic, morphological and geographical, aud to 

 modify some of the conclusions arrived at in the following attempt to. methodize 

 the Indian Qraminece. As affecting the numbers and synonyms taken from" Wul- 

 lich's List," I have to observe that for some of them I have had to depend on the 

 ticketed duplicates in Herb. Hook, and Benth., the types being wanting in the 

 Wullichian collection in the rooms of the Linnaean Society. 



Series A. Panicaceae. Spikelets articulate on their pedicels, or 

 deciduous with them, 1-2- fld., upper fi. alone (if 2) fruiting. 



Exceptions. Spikelets inarticulate on their pedicels in Arundinella and Isaclme. 

 Spikelets articulate on their pedicels, or with articulate pedicels occur in Poacc<i\ in 

 Alopecurus, Cyathopus, Polypogon, Fingerhuthia^ and Lopkat/ierum. Upper 

 flower alone fertile in Tribe fifialuridece of Poacece, and in Tristaehya. 



L Rachis of inn. inarticulate, or subarticulate in Stenotaphrum. Fl. 

 gl. usually coriaceous or herbaceous. 



Tribe I. PANICEJE. Spikelets 2-fld., upper fl. bisexual, lower mal- 

 neuter, rarely. both fertile. (Styinifex is dioecious.) 



* Spikelets articulate on their pedicels, except Isacline and Arundinella. 



f Spikrlef-s not invoHicolliitc by bristles. 



Bpikelatt not awned (yl. II and III leaked or awned in Pan. Crus-gaifi). 

 (See also Trickolcena and Arundinella.) 

 Spikelets dorsally flattened, base not thickened ; gls. 3, 



with very rarely a minute fourth (lowest) . . . 1. PASPALUW. 

 Spikelets of Paspalum, but with a thickened basal callus . 2. EKIOCHLOA. 

 Spikelets subglobose, panicled; gls. 4, I and II sepa- 

 rately deciduous, subequal ..... 3. ISACHXE. 

 Spikelets pauicled or spicate, gls. 4 (2 in P. subeglume) 



I and II very rarely subequal, II and III awned in 



P. Crus-yalli, IV rarely very shortly awned . . 4. PAXICFM. 

 Spikelets of Panicum, but gl. IV. narrowed into a short, 



flattened stipes, or with 2 appendages near the base . 5. JCIIXAXTII rs. 

 Spikelets innumerable, very minute, hairy, densely 



crowded in the capillary branches of a very largo 



panicle G. THYSAXOT..KXA. 



