CYPERACE^E. 



117 



Fig. 34, Cladium jamaicense CEANTZ ; a, a 

 cluster of spikelets. 



Near R. glauca Void; but se- 

 parable from it in the very much 

 longer seise and longer spikelets. 

 Cladium P. BE. 



Cladium jamaicense 

 CRANTZ (Fig. 34) ; HAYATA Mater. 

 Fl. Formos. p. 377. Our speci- 

 mens are a little different from 

 the typical form of the species. In. 

 the former, the spikelets are arrang- 

 ed in the less numerous clusters 

 and the inflorescence is looser than in the latter. 



Seleria BERG. 



Scleria pubigera MAKING. 



HAB. Toyen, leg. T. SOMA, Oct. 1914; Daijiorin, leg. Y. SHIMADA, Oct. 

 1914. 



Compared with a specimen so labelled in the Herbarium at Tokyo. Our 

 specimens are more patently or sometimes retrorsely hairy, while the Japanese 

 are less patently hairy. 



Carex LINN. 



On working up the collections of Carex, I have principally followed the 

 system of G. KUKENTHAL appearing in ENGLER'S " Das Pflanzenreich : 

 Cyperacese-Caricoidese." Among the subgenera, Primocarex is not represented in 

 Formosa, while Vignea is represented by 1, Indocarex also by 5, and Eucarex 

 by 28 species. As to the arrangement of the sections of the subgeuus Eucarex, I 

 have found it extremely difficult to follow KUKENTHAI/S system. According 

 to this eminent authority, the subgenus is divided into two groups on the basis 

 whether utricles are " erostrati vel brevirostres raro longirostres, rostrum ore 

 truncatum vel emarginatum raro leviter bidentatum," or "longius vel 

 brevius rostrati, rostrum ore bidentatum raro integrum." This arrangement 

 is not proved to be practicable in the Formosan species. The latter group is 

 divided into two divisions on the basis whether leaves are " nunquam septato- 

 nodosa " or " septato-uodosa." I am very much puzzled to understand 



