10 



pink also. In some species, e.g., ficifolia, m'miata, phceiiicea, it may be red, even 

 a verniiliou or orange-red. In a few species, e.g., pilulans, the filaments of dried 

 flowers turn red in course of time. 



Tlie pedicel is normally rounded, but owing to compression it is very often 

 strap-shaped, as in bolri/oides, and extreme cases are afforded by ohconlata {platypus) 

 and occidentaUs. 



Flower-bud. — The shape of the operculum was first used as a classification 

 character by Willdcnow in his Species Plantaruni, 1799. He divided the twelve 

 sj^ecies then known into two gi'oups — "operculo conico " and "operculo hemis- 

 phaerico." It is undoubtedly a useful character for the purpose, but variable, like 

 everything else about Eucalyptus. E. tcreticornis is usually looked upon as a 

 species to bo diagnosed by its operculum, 1)ut {Bull. Herb. Boissier, 1902, 579), I 

 have shown that this cl'.aractor breaks down completely as between that species and 

 E. rostrata. U. cctpitellata and _£'. mucrorrliyncha were at one time separated by 

 their opercula, but they pass into each other as regards those organs. As this work 

 progresses it will be obvious how very variable the operculum is. At the same 

 time, it Avill always remain, in the liands of a judicious observer, one of the most 

 practically iiseful diagnostic characters we have. 



Some species possess a double operculum, or membranous bract, enveloping 

 the whole of the young inflorescence. It was first observed by Robert Brown (see his 

 description of Eudesmia tetragon a), but a few years ago it was only recorded from 

 a very few species. In some it is very early deciduous and in others infrequent ; 

 but I have observed it in such a large number of species that I am inclined to the 

 opinion that extended research will show that it occurs in all. Brown's and 

 Jussieu's interesting observations on the single and double operculum will be found 

 supplementary to the former's descrij)tion of Eudesmia tetragona {Bot. App. to 

 Flinders' Voyage). 



Anther. — Bentham {Flora Anstraliensis) first grouped species according to 

 the shape and mode of dehiscence of the anthers. lie made five groups, but laid 

 no stress on the importance of tlie dehiscence on the top on the anther. He, 

 however, alludes (B.Fl. iii, ISO) to "truncate" anthers, and at page 189 to the 

 truncate anthers of E. leucoxylon. Mueller, finding that Bentham's five groups 

 could not be separately maintained, reduced them to tliree, viz. : — 



Fienanther<v, the anthers large and the cells divergent at the base. 

 This section mostly includes the stringybarks, although it includes several 

 white gums, — plants otherwise very different. 



Toranthcrcc, the anthers small and opening in pores. 



This section mostly includes boxes and some malices, and includes the silver- 

 leaved ironbark {melanophloia) , while E. crebra, \\\\w\\ is very closely allied to it, 

 is placed in another section. 



