138. CYPERAGEM. 



Spkts. all reduced to 1 flower, in the axils of spicate glume- 

 like bracts, upper bracts in a spike with male flowers 

 and the lower female or lower spikes entirely female, 

 uppermost entirelj'^ male. Nut enclosed in a utricle . 16. Carex. 



1. CYPERUS, L. 



Annual or with perennial rhizomes. Leaves mostly towards the 

 base of the erect stems, rarely reduced to sheaths, bracts under the 

 inflorescence usually resemble ordinary foliage leaves. Spikelets 

 more or less laterally compressed, with many glumes, in heads or 

 spikes which are arranged in simple or compound umbels or sometimes 

 reduced to a single head or spike. Rhachilla persistent (exc. in 

 aristatus), sometimes winged (see note below). Glumes distichous, 

 at least the loAver deciduous in fruit, 2 lowest empty, next above 

 several or many 2 -sexual (very rarely reduced to 1 or 2 flowering and 

 then the lower glume not enclosing the upper), all subequal, upper- 

 most 1-3 male or empty. Hypogynous bristles 0. Stamens 3-1, 

 with linear or oblong often apiculate anthers. Style branches 

 (" stigmas ") 3, base not tumid nor articulate on the nut which is 

 usually equally 3-gonous and 3-sided (a few species have unsymmet- 

 rical or slightly compressed or quite globose nuts). 



Cyperus as defined above excludes Pi/creus. Jiou-eUx.s and Mariscus as in the 

 Flora of British India and as I understand is being done in the Flora of the Upper 

 Gongefic Plain. 



Much use has been made in the F.B.I, of the wings of the rhachilla, and it seems 

 to me that the distinction of narrow and broad wings (in itself a purely relative 

 distinction difficult to work without a suite of specimens) widely separates some 

 closely allied species. The rhachilla may be winged in various ways, either the 

 wings are the very thin continuous margins of the compressed rhachilla and lie 

 in the same plane as that of the rhachilla, or the wings are at right angles to this 

 plane and discontinuous, but form no part of the glumes or, thirdly, the wings 

 appear to be the detached bases of the glumes themselves and can therefore only 

 be seen after these fall. Cooke (wrongly as I think) states as a universal rule that 

 the wings form part of the glumes. 



Of other characters I am a hit doubtful of the value of the length of the bmcts 

 but that of the glumes appears to be a very good character and to vary within 

 singularly narrow limits. 



a. Stigmas 3, minute or tapering, nut unsymmetrical and 

 angles corky. Tank floaters:— 

 Style long with minute stigmas. Infl. capitate . . .1. cephalotes. 

 Style short slightly compressed and widening upwards, 



stigmas as long, somewhat tapering. Infl. sub-capitate 2. plati/gti/lis. 

 fi. Stigmas 3 filiform. Nut symmetrically 3-sided (exc. some- 

 times in fiavidus and difformis and exceptionally in 

 others) :— 

 I. Slender or ca^spitose annuals (sometimes perennial in 

 haspan). Glumes small in all (under "07" except No. 

 11). Spp. 1-11 :— 

 A. Spikelets clustered {i.e. not spicate). Spp. 1-8 :— 



1. Glumes very short, as broad as long. Spkts. in 



dense globose heads, hds. umbelled. Nut sub- 

 equally 3-gonous : — 

 Gl. '025" long, obovate, truncate. Nut '02" , . 3. difformis. 



2. Glumes slightly longer than broad, •02-'04" long. 



Spikelets numerous tligitate. Nuts sub- 

 globose :— 



Rhizome 0. Nut globose or sub-compi-essed, 



obscurely 3-sided. St. 1 i. fiavidus. 



Rhizome .slender. Nut globoselj'-obovoid, 3-sided. 



St. 2-3 5. haspan. 



890 



