TERMINOLOGY IN MONOCOTYLEDONS. 511 



phorum they are very numerous, and lengthen much after flower- 

 ing, becoming white and silky, and assuming that appearance which 

 has given to the plants the name of Cotton-rushes or Cotton-grass. 

 This irregularity in numbers has been one of the chief objections 

 raised to the assimilating these hypogynous setaD to perianth- 

 segments. It must, however, be borne in mind that in Mono- 

 cotyledons, as well as in Dicotyledons in genera or orders in which 

 a definite number of stamens and perianth-segments is the rule, 

 there are sometimes exceptional species where either the stamens 

 alone or the perianth-segments also are indefinitely multiplied. 



Amongst the bracteolate genera of Cyperaceae, Mapania, Pan- 

 danophyllum, Diplasia, Lepironia, and perhaps two or three others 

 which I have not had the opportunity of carefully examining, have 

 hypogynous scales usually long linear and narrow though flat, 

 and always within the two bracteoles. These, however, have not 

 been generally recognized as the homologues of the seta3 of 

 Scirpus and its allies, but the whole flower, with its stamens, 

 scales, and bracteoles, has been described as a secondary andro- 

 gynous spikelet formed of glumes, monandrous male flowers, and 

 a single terminal female flower, a view in which I am unable to 

 concur on the following grounds :— 1. The whole spike supposed 

 to be compound has precisely the structure and aspect of the 

 simple spikelet with its numerous closely imbricate glumes, each 

 completely enclosing what I should consider as a single flower. 



2. The scales within each glume are not arranged regularly on 

 the axis (all distichous, or all more or less spiral), as is always 

 the case in normal spikelets, but the two outer ones are nearly 

 opposite like bracteoles, and all the others inserted higher up and 

 as nearly as their mutual pressure will allow in a 1 -seriate or 

 2-seriate whorl, an arrangement never observed in a true spikelet. 



3. The stamens are'not regularly placed oue within every scale, 

 as is the case with the true androgynous spikelet of the Scleneae ; 

 but scales with and without stamens either alternate with each 

 other or are irregularly mixed in the whorl, or, as is very plain 

 in Diplasia, the statnens are in a distinct whorl within the whorl 



On the other hand, the genera Ghrysithrix and Chorisandra, 

 inaptly included by Boeckeler in the Hypolytrese, have the true 

 androgynous spikelets of the Sclerie*, and were on that account 

 more correctly placed by Kunth in the latter tribe. 



Tbe fine genus Pandanophyllum, probably too much subdivided 



