32 ME. 0. B. CLABKE ON INDIAN SPECIES OF CTPEETTS. 



margins greatly incurred when dry. The glumes of G. procerus 

 are more brightly somewhat yellow- or red-stained, less flat- 

 backed, less strongly plicate-striate, less incurved when dry. 

 The umbel in 0. procerus is compound, in G. inundatus decom- 

 pound in every example seen : from the analogy of other species 

 it is probable that poor specimens of C. inundatus would have a 

 compound umbel ; but no example of G. procerus (an abundant 

 plant) has the umbel decompound. G. procerus has the anthers 

 absolutely muticous, like those of C. pilosus ; but G. inundatus 

 has an obsolete crest to the anther exactly as has G. Monti. 



I have dwelt on these minute points because G. inundatus is 

 the only species of Gyperus in which there remains to me any 

 question that a much-compressed and clearly trigonous nut may 

 occur in one species. If this species be allowed to be distinct 

 from G. procerus, then we have an absolute distinction to separate 

 Pycreus and Juncellus from JEucyperus. 



Parag. 5. On the Genus Kyllinga. 



Benth. et Hook. f. G-en. PI. iii. p. 1039, separate the 1-flowered 

 (1-nutted) species of Gyperus from Kyllinga by the fact that the 

 upper (empty or male) glume is protruded from the flowering 

 (nut-bearing) glume. This is in most cases a working character ; 

 but Gyperus umbellatus, Benth., shows such a range of variation 

 in this character, that it is impossible to put it forward as the 

 chief diagnostic one : in one of the most abundant forms of G. 

 umbellatus, Benth., the upper (barren) glume does not overtop 

 the nut-bearing glume. In another frequent form, hardly distin- 

 guishable from this as a variety (G. pictus, Wall.), the spikelets 

 are exceedingly similar, but the barren glume is clearly protruded ; 

 and from this form we pass on to the linear -spikeleted forms of 

 G. umbellatus by insensible gradations. 



Boeckeler, in his ' Conspectus Gtanerum,' mentions the two 

 points which separate Kyllinga from Gyperus (though he does not 

 give contrasted differentiae), viz. "Spikelets one-flowered upheld 

 on discs, and style bifid." 



Kyllinga can be separated easily from each subgenus of Gype- 

 rus, but not (except by alternative cross-referenced diagnostic 

 points) from the genus as a whole. I should therefore have pre- 

 ferred to have made it a subgenus of Gyperus. 



The subgenus of Mariscus includes (for me) all those species 

 (Leptostachyi of Boeckeler and others) in which the glumes are 



