525 



In regard to the above, it may be pointed out : — 



1. That D. includes E. and F. 



2. That this 1916 classification has special application to New South Wales 

 species, and therefore omits altogether, e.g., Cornutse, which. are confined to "Western 

 Australia. 



Maiden (1922). I now submit an improved tentative classification: — 



A. Renanthera. with two sub-sections— 



(«) Alpina?.* 



(b) Brachyandra? * 



B. Renantlieroidece. 



C. Porantheroidecp. 



D. Terminales. 



E. Platyantherce, with two sub seefciniiH 



(a) Graciles.* 

 (6) .Pyriformes.* 



F. Maerantherce. 



1. Tereticornes.* 



Miscellaneous. 



2. Longiores. 



I. Corymbosa?, 

 II. Xon-Corymbosa?. 



(a) Miscellaneous, embodving extra characters. 



(b) Papery fruits (Clavigera?).* 



(c) Serrulate filaments (Cornuta?). 



(d) Stamens in four clusters (Eudesmia?). 



A. Pvenaxtheb.e (46). with two sub-sections — ■ 



(a) Alpina?. 



(6) Brachyandrse. 



ftenantliera?.- — The anthers are kidney-shaped and very uniform. Like all 

 anthers, they have two pollen cells, but eventually, by absorption of the dissepiment, 

 they form a single cavity. Bentham was fond of the term " confluent at the apex." 

 There is a medium-sized gland in the front. 



* After well-known spscies in the respective sub-seetions. 



