10 MB. C. B. CLABKE ON INDIAN" SPECIES OF CYPERT/S. 



than the umbel ; but in leafy examples of C. Haspan or G. corym- 

 bosus the bracts will be long, far overtopping the umbel. The 

 most nearly leafless species in the whole genus is G. articulatus ; 

 and in this the bracts are 1-2 centim. long only. In breadth, 

 scabrousness (or pubescence), nervation, and texture the bracts 

 invariably resemble the leaves ; in many cases therefore, where a 

 critical species as G. Haspan is propped up by characters taken 

 from the bracts, there is no real combination of characters, and 

 the delimitation of the species is not materially strengthened. 

 In G. vaginatus, E. Br., of which I have seen very little material, 

 the leaves are reduced to short ovate acute terminations of the 

 sheaths, the bracts are 2-3 centim. long, liuear, rigid : this may 

 be a partial exception to the rule above laid down ; but I should 

 expect among a large quantity of G. vaginatus to find examples 

 with the leaves developed to 3-6 centim. long, and in such case I 

 feel confident they would closely resemble the bracts. 



The bracts are in most species apparently umbelled, but they 

 are always really alternate ; and in many species they are mani- 

 festly alternate, so that the umbel is said to be imperfect or 

 corymbose : Linnaeus named G. alternifolius from this character. 

 G. jeminicus (when the rhizome is wanting) can be separated 

 from G. rotundus (to which I find G. jeminicus without rhizome 

 very generally referred in herbaria) by the lowest bract being 

 clearly remote (2-3 millim., sometimes by 1 centim.) from the one 

 next above it — a trifling, but so far as seen invariable, mark. 



The bracts are usually divaricate or erect-patent ; but in 

 several species the lowest bract is nearly erect, sometimes quite 

 so, so as to appear a continuation of the culm, and the umbel 

 appears lateral, as in 0. leevigatus, G. stramineus, G. pauper, and 

 in several of the Conglomerate The character is useful in recog- 

 nizing some of these species rapidly ; but it is of small value, 

 for, of many species with spreading bracts, in smaller specimens 

 with 2 or 3 bracts the lowest bract becomes nearly erect, as in 

 G.fiavescens depauperated. From the axil of each bract springs 

 a ray : the number of rays is thus never less than the number 

 of bracts ; it is in general more, because, from the shortening 

 of the uppermost nodes of the axis, the bracts are suppressed 

 though the rays belonging to them are developed. This law can 

 be supposed true, however, only theoretically in the case of con- 

 gested or capitate umbels, which often appear as simple heads 

 with numerous bracts. 



