MB. C. B. CLA.BKE ON INDIAN SPECIES OF CTPEBTJS. 21 



Cyperus either 3-fid or 2-fid, without exception ; and therefore 

 the primary subgenera of Cyperus have always been founded on 

 this character. This was the view of Kunth. Regarding the 

 character as absolute, Nees, Eeichenbach, and others have elevated 

 Pycreus to generic rank. But Boeckeler has blown upon the 

 character a little by uniting certain species into composite species, 

 each of which may have either a 3-fid style with triangular nut, 

 or 2-fid style with complauate nut. These species are C. alope- 

 curoides, Rottb., cum C. dives, Delile ; C. inundatus, Roxb., cum 

 Cprocero, Rottb. ; C. pygmceus, Linn., cum Isolepide Micheliand, 

 Roem. et Sch. In each of these cases I consider two remote 

 species have, from general aspect, been mixed, and that the more 

 carefully they are examined, the more essentially and structurally 

 they are found to be distinct. This I have attempted to explain 

 fully below. 



In G. stoloniferus, Retz., Boeckeler says, "stylo elongato (quan- 

 doque abbreviato) profunde bi- v. trifido aut indiviso." I find the 

 style always deeply trifid. To examine the style, a flower should 

 be dissected just before expansion (this gives you the stamens by 

 the same dissection) ; it is dangerous to trust to older flowers, as 

 the brittle style-branches often break off exactly at their point of 

 separation, in which case, even with the microscope, it is difficult 

 to see the scar. The casual cases of suppression of a style-branch 

 that occur are excessively few, I should say in my experience not 

 1 in 1000 for the genus. The stamens (ecristate &c.) are certain 

 proof that Boeckeler is in error in placing this species with 0. 

 Icsvigatus &c. It is not far from C. rotundus, but perhaps nearer 

 C. arenarius. In nearly all the critical species it is less trouble 

 to examine the nut than the style (». e. in mere sorting and 

 naming). The nuts of G. alopecuroides, Rottb., and C. dives, 

 Delile, are as unlike (not merely by two rounded or three sharp 

 angles) as any two in the genus. 



The style-branches are usually long-linear, shortly or distinctly 

 exserted from the glume, but sometimes very much exserted, so 

 that the spicula are said to be comose. This character is the only 

 point left to keep G. diandrus distinct from C. rivularis; and some 

 (even of the American botanists) do not think the two distinct. 

 There is a strongly comose Cyperus, resembling C. Monti in all 

 other points but this, among Griffith's collections ; and there 

 is a fragment, collected by V. Ball in Chota Nagpore, which 

 does not differ much from C. pilosus except in being comose. I 



