MB. C. B. CLABKE ON INDIAN SPECIES OF CTPEBUS. 19 



The difference between round-backed and carinate glumes 

 makes the difference between terete and compressed spikelets. 

 In many species there are no nerves except near the keel of the 

 glume. In other species the glumes are plicate-striate through- 

 out nearly their whole breadth. This character is useful, not 

 merely in separating some critical species as G. esculentus, but in 

 defining some sections, as the Nivei. 



The open cellular texture of the glumes of G. liyalinus, Vahl, is 

 strikingly unlike any other species of the genus. 



The glumes are generally quite smooth ; in a few species the 

 keel upwards is scabrous under a magnifier. 



More important points in the glume are the dorsal compression 

 in sect. Juncellus, accompanying the dorsally-compressed nut, 

 and its insertion, already considered under the rhachilla. 



(8) The Stamens. 



(a) Number. The number of stamens is 3, 2, or 1 ; in most 

 large species 3, in many small ones (whether the style be bifid or 

 trifid) 2-1. The sect. Luzuloideae (C. vegetus, G. virens, C. Luzula) 

 are almost the only large or medium-sized species in which the 

 stamen is 1 only. In many species the stamens are 2 in well- 

 developed, 1 in poorly-developed specimens, or mainly so. 



In Tycreus perhaps two thirds of the species have 2 stamens 

 (or 1), but this goes with the size mainly of the species. In the 

 small species of Eucyperus (as in sect. Aristati, and in G.flavidus 

 and small G. Raspan) the stamens are also 2-1 ; in Juncellus the 

 large species (as G. Monti, G. inundatus) have 3 stamens, the small 

 (as G. leevigatus, G . pannonicus) 2-1. The number of stamens rarely 

 varies 3 or 2 in one species ; so that we have in the number of 

 stamens a character often useful as a specific, sometimes as a sub- 

 sectional distinction. 



(b) Filaments. These are always more or less ligulate ; they are 

 exceedingly broad, somewhat widened upwards in the Conglome- 

 rati and in O. cephalotes and G. platystylis. They are usually 

 distinctly ligulate, and only obscurely so in the small-flowered 

 species with small stamens. 



The length of the filaments and exsertion of the anthers are 

 sometimes described ; but there is much uncertainty in the 

 value of the character, and the filaments (apparently from an 

 obscure tendency to unisexuality) sometimes remain short in 



c2 



