550 



This crimpiness is also figured in fig. 96, Plate 88 (E. inlcranthero). 



This polygonal character of the filaments (hi contradistinction to the terete 

 or filiform appearance of those of most species) seems to be a consequence of tight 

 packing in the operculum, and it is accompanied by serrulations along the margins. 



This serrulatkm is not to be confused with the much coarser bendings known 

 as " flexuose " of terete filaments. Ben.th.am speaks of those of the Cornutse (as a 

 whole) being " flexuose."' and, in detail, as " slightly flexuose," but it seems to me 

 that the Cornutse should be treated as a group by themselves, with filaments 

 " polygonal-serrulate." 



Length. 



Bentliam, 18(i0. — ■" The dimensions given for the stamens refer to the outer 

 ones; the inner ones are almost universally gradually shorter." (B.F1. hi, 187.) 



Bentham seems to have been the only botanist who systematically mentions 

 the length of the filaments. The measurements have, however, limited diagnostic 

 value. 



Following are the measurements he gives, varying from a line to an inch :— 



1-2 lines. — 



E. tricolor. 



Not 2 lines. — 



E. Behriana. 



Not above 2 lines. — 

 E. slellulala. 



About 2 lines. — 



E. cneorifolia. 

 E. drepanophjlla. 

 E. exserta. 

 E. hemiphloia. 

 E. melliodora. 



2 to nearly 3 lines. — 

 E. Gunnii. 



2 or 3 lines. — 



E. buprestium. 

 E. capitelkUa. 

 E. cinerea. 

 E. coriacea. 

 E. doratoxylon, 

 E. odorata. 



2-3 lines or rather more. — 

 E.falcata. 



About 3 lines. — 



B. coccifera. 

 E. dealbaia. 

 E. yoniocalyx. 



E. microtheca. 



E. hptopoda. 



E. stricta. 



E. pallidifolia. 

 E. piperita. 

 E. rostrata. 

 E. uncinate. 



E. 



prmnosa. 



E. oleosa. 

 E. oligantha. 

 E. piluhris. 

 E. saiigna. 

 E. siderophloia. 



E. paniculala. 



E. patens. 

 E. Risdoni. 

 E. virgala. 



