537 



of anthers, based on size, shape, dehiscence, gland, attachment of" filament — indeed, 

 everything we could think of. Later on it was found that there was a good deal of 

 variation in anthers attributed to the same species, and that they changed according 

 to age. With the experience we gained, it was found to be desirable to reduce the 

 sections to the present number, and although I am responsible for this. I have had 

 the advantage of the co-operation of Miss Flockton at every stage. Without her facile 

 pencil it would have been impossible to have taken comprehensive views of them, and 

 for this over 2,000 detailed drawings were necessary. The subject has not been absent 

 from our minds for any length of time during these twenty years, and we find it is 

 most puzzling, and there is a good deal of variation in anthers, as, indeed, in every 

 organ in Eucalyptus. The classification offered will form a basis for further 

 investigation, which should be made, as far as possible, on perfectly fresh material. 

 At the same time it wall be necessary to work upon material of various degrees of age. 



Shape. 



Bentham. Flora Australiensis. In the Group Renantherse, since it refers to 

 the shape of the anther, Bentham has not found it necessary to repeat this word or 

 '* reniform.''' Following are exceptions : — 



Reniform — 



E. cocci fern. 



E. diversifolia. 



Reniform or broad — 



E. pilularis. 



Small reniform — 



E. coriaeea. 



E. macrorrhyncha. 

 E. marginata. 



E. stellulata. 



In the other groups he only uses the words " ovate- 

 Ovate — ■ 



or " oblong." 



E. aVpi/na. 

 E. dealbata. 

 E. diversicolor. 

 E. etserta. 

 E. falcata. 

 E. globulus. 

 E. goniocalyx. 

 E. Gunnii. 

 E. orbifolia. 



Small ovate- 



E. goniantha.. 

 E. oleosa. 



E. pachyloma. 



E. pallidifolia. 



E. patens. 



E. phosnicea. 



E. Preissiana. 



E. pyriformis. 



E. saligna. 



E. vemicosa. 



E. virninalis. 



E. rostrata. 

 E. tereticomis. 



G 



