dilleniace^e. 17 



aware any fuller description of it has appeared, or that any other 

 species have since been published (except the doubtful C? paaciflora, 

 Zoll. & Moritz., from Java), although to the same genus undoubtedly 

 belongs No. 329 of Lobb's collection at Singapore, in 1846, as is noted 

 by Planchon, in the Hookerian herbarium, as well as the two species 

 here characterized. The fruit of these Feejeean plants unfortunately 

 was not collected ; but the half-grown carpels of the present species 

 plainly indicate a capsular fruit. As the genus has not been figured, 

 I devote a plate to its illustration, even from our imperfect materials. 

 The anthers, unfortunately, in all the flowers I could examine, had 

 been in a good degree destroyed by larvae. I think they open longi- 

 tudinally; while in the plant from Singapore, above-mentioned, the 

 cells open only at the top, by a pore or short slit.* 



Plate 1. — Capellia biflora: a branch in flower. Fig. 1. Ver- 

 tical section of an unopened flower. 2. A petal. 3. A separate sta- 

 men, enlarged. 4. The pistils divided transversely, moderately mag- 

 nified. 5. One of the pistils detached from the fleshy central axis, 

 its ovary longitudinally divided, more magnified. 6. Longitudinal 

 section of the half-grown fruit ; one of the carpels laid open, to show 

 the young seeds, moderately enlarged. 



2. Capellia membranifolia, Sp. Nov. 



C. foliis oblongis obovato-oblongisve repando-serratis membranaceis recte- 

 venosis costis subdistantibus ; pedunculis bir-tri-jloris ; floribus parvulis ; 

 ovariis 8-10. 



. 

 Hab. Ovolau, Feejee Islands. 



Of this we have only a single specimen, bearing one or two flower- 

 buds, which, so far as I can see, differs from the last principally in its 



* While this sheet was in type I have received, through Prof. Miquel, specimens of 

 the original Capellia multiflora, from Java. It bears a close general resemblance to the 

 present species, from which, however, it is clearly distinguished by the very narrow wings 

 or margins of the petiole, which do not wholly clasp the stem at the insertion, as well as 

 by the more numerous flowers and fewer ovaries. A few of the innermost anthers are 

 abruptly very much longer than the others, as described by Blume. The anthers open 

 only by two pores at the apex. 



5 



