ME. C. B. CLAEKE ON INDIAN SPECIES OF CTPEETJS. 13 



tions arrangements like that in fig. 27 are described as " rays with 

 3 spikes, terminal spike elongate, lateral short or auriculate : " 

 this is a frequent occurrence in Cyperus, typical in G. Luzulce, 

 C. lucidus, and many others, and occurring occasionally in C. 

 umbellatus, Benth., and other species where it is not admitted by 

 the book description. It is so characteristic of the very nature 

 of Cyperus inflorescence, that I should not be surprised at its 

 occurrence anywhere. 



G. polystachyus is a well-known species excellently adapted 

 for a discussion of the inflorescence of Cyperus. In even mode- 

 rately-well developed examples I find the umbel invariably com- 

 pound ; the umbellules are often congested into dense heads, 

 but their axis is then evidently divided. In the vulgar form of 

 G. polystachyus, common near the sea in nearly all the continents 

 and islands of the tropical world, the umbel is usually (as in C. 

 rotundus and some other species when near the sea) greatly con- 

 tracted ; and this form, the commonest in collections, has been 

 treated as G. polystachyus type ; while a multitude of " species " 

 have been created by herbarium botanists out of the inland 

 forms with more or less open umbel. In nearly every example, 

 whether with open or close umbel, the smaller few-flowered 

 umbellules may appear perfectly simple spikes until they are 

 compared with all the other spikes on the same plant. 



But in C.filicinus, Yahl (=0. Nuttallii, Eddy), esteemed, not 

 without reason, a var. of G. polystachyus by Boeckeler, every spike 

 in every specimen I have is in appearance perfectly simple; 

 fig. 26 exhibits one of the largest. The spikelets themselves are 

 here considerably larger than in ordinary G. polystachyus ; but it 

 occurs in many species with various umbels that, besides the 

 standard form, there occurs another with the umbel divided one 

 time less, and the spikelets considerably larger. 



In this fig. 27 I have shown the two lowest glumes of the spike- 

 let A and B, alternate, smaller than those above them, and both 

 (or sometimes one only) empty ; the lowest empty glume (A) is (in 

 other species) sometimes truncate, sometimes with a long caudate 

 point bracteiform : this is a useful character in specific discrimi- 

 nation. In other species too, as a first indication that the spike 

 is beginning to become compound, there is a green caudate small 

 bract close below A and on the same side of the spikelet (it is 

 therefore I call it a bract, and not a glume). This is (at least I 

 hope I have, after much labour in distinguishing the bract from 



