66 GRAMINE^E. 



VIII. TJiysaiwilcena NEES. 



Thysanolsena Agrostis NEES ; MATSUM. et HAYATA Enum. PI. Formos. 

 p. 516. 



IX. Chamceraphis E. BR. 



Chamaeraphis depauperata NEES ; MATSUM. et HAYATA Enum. PI. 

 Formos. p. 512. 



X. Spinifex LINN. 



Spinifex squarrosus LINN.; MATSUM. et HAYATA Enum. PI. Formos. 

 p. 513. 



XI. Oplismenus BEAUV. 



1. Spikelets solitary or clustered in a simple terminal spike. 



1. 0. undulatifolius. 



Spikelets in racemose spikes 2 



2. Stem usually tall with decumbent base, leaves 5-18 cm. long lanceolate, 



spikes 4-10, 3-10 cm. long, spikelets usually loosely imbricate, awns 



rather stout 2. 0. compositus. 



Decumbent, flaccid, leaves 1-3 cm. long, spikes 3-5 short, spikelets secund 



closely imbricate 3. 0. Burmanni. 



(The occurrence of the last species in Formosa is doubtful.) 



1. Oplismenus undulatifolius BEAUV.; MATSUM. et HAYATA Enum. 

 PI. Formos. p. 509. 



2. Oplismenus compositus ROEM. et Sen., MATSUM. et HAYATA Enum. 

 PI. Formos. p. 509. 



3. Oplismenus Burmanni BEAUV.; MATSUM. efc HAYATA Enum. PI. 

 Formos. p. 508. 



XII. Axonopus BEAUV. 



HOOK f. Fl. Brit. Ind. VII. p. 64; TRIM. Fl. Ceyl. V. p. 166. 



The present genus is not adopted by BENTHAM who regards it to be 

 referable to Paspalum (Gen. PL III. p. 1098). HACKEL too does not recognize 

 Axonopus. As far as I can judge from the study of Axonopus semialatus, the 

 structure of the spikelets resembles in some measure that of Panicum Crus- 

 Galli. The difference between them exists only in palea?. They are in 

 Axonopus 2-fid, but in Panicum entire. 



