501 



MB. G. BENTIIAM Off CLASSIFICATION AND 



bracts except the anthers and stigmatic tips, they are almost 

 always more or less stipitate, and on their stipes or pedicel, or on 

 the floral axis continuous with this pedicel, are usually 6 or fewer 

 hyaline scales in two series — the upper or inner series close under 

 the stamens in the males, or under the ovary in the females, the 

 outer or lower series often much lower down and sometimes at 

 different heights (see Plate VII.). In many species those of the 

 upper series are, in the male flowers, reduced to small teeth or 

 entirely wanting, and are also occasionally deficient in the females ; 

 those of the lower series are sometimes in one or both sexes re- 

 duced in number or entirely deficient, or are united in a single 

 spatha-like scale ; but their homology is the same in both sexes, 

 and they appear to be quite correctly termed perianth-segments. 

 Yet they are often called by other names, and differently so in 

 the male and female flowers. Thus Lindley and some others 

 have even described those of the outer series as glumes, although 

 they are never glume-like either in consistence or insertion ; and 

 most monographers describe the upper portion of the stipes or 

 axis of the male flowers as the solid tube of the inner perianth. 

 Setting aside the contradiction of terms implied in the expression 

 " solid tube," this interval between the outer and inner series of 

 segments is no more a part of the inner one in the males than in 

 the females, where it is never so described, or than in some species 

 of Campanumaa and a few other Dicotyledons, where the calyx is 

 attached to the stipes at some distance below the ovary and corolla. 

 It is true that in the female Eriocaulons the two series of segments 

 are sometimes close together, and the interval between the two 

 is generally shorter than in the males ; but they are sometimes 

 also as distant as in the males, or both series are at some distance 

 below the ovary. In some species one or both series are deficient. I 

 am not aware of any bracteoles having been observed in Eriocaulese 

 below the perianth-segments. 



The small order Centrolepidese, which, from the structure of its 

 ovary and seeds and other characters, must be placed in close 

 proximity to Eriocaulese and Restiacese, is nevertheless anomalous 

 as well as much diversified in its floral arrangement. The inflo- 

 rescence is a small terminal head or spike as in Eriocauleae, but 

 usually more or less compound, although the bracts or scales are 

 very much reduced in number and often in size. The nearest 

 approach to the normal spikelet is in Aphelia, where the bracts are 

 distichously imbricate, the upper ones, or nearly all, enclosing 



