Mr. J. Miers on the Winteracese. 115 



flesh-coloui'ed, two or three inches long ; the filaments are seta- 

 ceous and above half the length of the petals, with globose an- 

 thers ; the seeds are arillate, with the flavour of those of coffee, 

 only more bitter. This kind of fruit ill accords with the features 

 of the IVinteracecE. Until further evidence of its existence, and 

 better characters of its structure be ascertained, it ought to be 

 restored to the list of the " genera incertse sedis," in which it was 

 placed by Jussieu. I cannot help thinking there must have 

 been some confusion in the original notes of Molina : he de- 

 scribes his Temus moschata as furnishing an excellent and very 

 hard wood, which cannot be of rare occurrence, as he says it is 

 much used in Chile in various manufactures. The Luma Cnick- 

 shanksii, A. Gray {Eugenia Cruckshanksii, Hook.), which is abun- 

 dant in the midland provinces of Chile, is known among the 

 natives by the name of Temu; and it possesses a wood that 

 answers the above description. 



5. Trochodendron. 



This genus, established by Siebold and Zuccarini for a plant 

 which the former botanist brought from Japan, has been fully 

 described and figured in the ' Flora Japonica.^ It has been 

 referred to the neighbourhood of lUicium; but its characters 

 seem quite irreconcileable with those of the IVinteracea. The 

 features which appear to me incompatible are — its verticillated 

 leaves, which are serrated; its peculiar mode of budding; its 

 flowers destitute of both caljTC and corolla; its single pluri- 

 locular ovary ; and its capsular fruit, opening by five to eight 

 thick coriaceous valves ; to which may be added, the different 

 sort of integument by which its seeds are invested. On the 

 other hand, in the habit of the plant, its serrated leaves, its 

 numerous stamens, its many-celled ovary, with divided style, 

 its capsular many-valved fruit, and the form and texture of 

 its seminal integument, it approaches far nearer to the Tern- 

 stroemiacea ; and it has several features in common with Trocho- 

 stigma, also of Japanese origin {Adinidia, Lindl.), placed by 

 some botanists in Ternstrcemiacece, by others in Dilleniacece ; but 

 perhaps it comes still nearer to the Tasmannian Carjjodontos 

 (congeneric with Eucryphia from Chiloe), a genus of doubtful 

 |-<jsition placed between the Chlcenacece and Ternstroemiaceee. In 

 tiie latter genus the corolla is very deciduous, and its opercular 

 calyx at the period of aestivation falls off by a circumscissile line, 

 — a character approximating to the seemingly achlamydeous 

 flowers of Trochodendron. 



