ENUMERATIO PLAXTARUM FORMOSANARUM. 233 



o-dentato, dentilnis lanceolatis. Corolla tomentosa, lobis 5 

 obtusis, erecto-patentibiis, ellij^ticis, imbricatis. Stamina 10, 

 filamentis inter se fere ad apices in tubum latum connatis, sub- 

 {Bqualibns, tubo stamineo basi corollse affixo ; antherarum loculi 

 lineares, adnati, parallel!, discreti. Ovarium omnino liberum, 

 ovatum tomentosum, o-loculare ; stylus crassiusculus, stigmate 

 terminali parvo capitato obscure 3-lobo ; ovula oblonga, in quo- 

 que loculo 5-8, medio axis 2-seriatim affixa, ascendentia. Fructus 

 oblongus, erectus v. cernuus, epiearpio subcarnoso, deinde deciduo, 

 endocarpio coriaceo in valvas 5 loculicide déhiscente, valvis erecto- 

 patentibus. Semina linearia parva, testa Crustacea, foveolata, 



The problem may be put as follows: — 



The characters for which the species 

 is affiliated to Sti/raceœ : 



1). External form of the tree. 



2). Stellate hairs covering: the 

 plant. 



3). External form of the flower 

 which resembles very much tliat of 

 tSti/rax. 



4). Absence of the disc which is 

 so verv common to Ericaceir. 



The characters for which the species 

 is affiliated to Ericacece : 



1). External form and internal 

 structure of the fruit. 



2). Seed in respect to its external 

 foi-m as well as internal structure. 



o). Embryo in its shape and size. 



4). Internal structure ot the ovary ; 

 in the shape of the ovule and 

 placenta, number of the cells and 

 ovules. 



5). Stamens in 2 series. 



Apart from the botli orders, this plant has an articulation in the middk' 

 of the pedicel — a character which does not come to either Ericacece or Styraceœ. 



On considering all the above mentioned points, I am forced to the conclu- 

 sion that the plant has in itself the characters of the two orders. AYe may 

 well think that the plant conies directly from the aucester, while l)oth Ericacece 

 and Stijraceiv come from it indirectly. It would be promising, therefore, to 

 establisli an oider for this plant, — a new order, Alniphyllacece, I should propose 

 to call it — and not to place such a peculiar one in some of the orders we have 

 at present. 



The branches through wiiich the three, Ericacece, Styraceœ, and Alniphyllacece, 

 may have been derived frotn the ancester, is shown as follows :— 



