684 JOURNAL, BOMBAY NATURAL HIST. SOCIETY, Vol. XXV. 



the main rachis at the base of the inflorescence is not so complete as to 

 prevent the development of a branch (umbel ray) from the axil of each 

 bract. The secondary branchings of the primary rays however often give 

 the appearance of being sympodial (cymose) especially in siich supra- 

 decompound forms as Fimhristijlis quinquangularis and Scirjms (/rossus. In 

 some of the Caricoidece the ultimate branchings, and even the spikelet 

 itself, seem to be certainly cymose. 



In the Caricoidece the morphology of the spikelet is often a matter 

 of dou'bt. For instance, the utricle of Care.v and the spike or spikelet of 

 Hypoli/trum are variously regarded by difi'erent authors. It would be 

 beyond the scope of this paper go into any detailed discussion of these 

 points, especially as these genera are poorly represented. 



The Cypemcece form a satisfactory family of fairly constant habit. 

 They are usually recognizable at once by their 3-angled stems, which dis- 

 tinguish them from grasses. The species with terete or polygonal stems 

 are recognizable by their o-ranked leaves or leaf-sheaths, the leaves of 

 grasses being always 2-ranked, by the sheath being perfect and not split 

 down the front as in grasses, and by the bracteate inflorescence. The 

 Juncacece ("rushes") can be distinguished by having a definite peri- 

 anth and 3-celled capsule of the Liliaceous type. The discrimination of 

 some species of Eriocaulon ("Hat-pin grasses'") from some Cyperace- 

 ous forms such as Eleocharis is a matter of more difllculty. As a rule 

 the Eriocaidacece can be distinguished by their dense rosettes of suc- 

 culent radical leaves. Failing this the inflorescence must be dissected. 

 The Eriocmilacece have a three-celled capsule containing separate seeds, 

 and the whole morphology of their flowers is quite difi'erent. 



The " sedges" are typically denizens of marshes, though some species 

 inhabit dry localities and even deserts, and some forests. 



Though so closely allied to grasses the "sedges," are of little economic 

 importance, being as a rule very poor fodder. The ancient "papyrus" was 

 made from the stems of Ci/perus papyrus, Linn., a Nile plant. Of the Indian 

 sedges a few have edible tubers or bulbous stem-bases. The stems and 

 leaves of some are used for mats. From the tubers of Cyperus rotundus, 

 Linn., is obtained the fragrant powder used in the " agarbattis", or fragrant 

 tapers burned in Hindu temples and houses. The same species is a notable 

 pest in agricultural land. 



KEY TO THE GENERA. 



I. SCIRPOIDE^^^. Primary branching of the inflorescence usually 

 umbellate {i.e., of several closely-placed rays subtended by several bracts), 

 or spikelets in a contracted head, or spikelet solitary (reduced umbels). 

 Spikelets always 2-sexual. Florets usually § . Nut free, but concealed 

 or partially concealed by its glume, even when ripe. 



A Flowering glumes all distichously arranged on a flattened or sub- 

 quadrangular rhachilla. Hypogynous bristles 0. (CYPEIIE^'F). 



1. llhachilla not disarticulating above the two 



lovvest glumes, but the fertile glomes and nuts 

 ultimately deciduous from the permanent rha- 

 chilla. Spikelets many-flowered, elongate .. 1. Cyperus. 



2. llhachilla disarticulating above the two lowest 



(empty) glumes, falling away with its fertile 



glumes and nuts, leaving a knob on the rhachis. 



(a) Spikelets 1 — many-flowered, elongate, usually 



spikately arranged on the branches of a true 



umbel, sometimes capitate ; style 3-fid . . 2. Mariscus. 



