OCEANIC DISPERSAL OF PLANTS. 289 



Barringtonia ami Cahphyllvm inophyllum, have established themselves by means of their 

 drifting seeds on a freshly dry coral islet, the fruit-pigeons alight in the branches in their 

 flight from place to place, and drop the seeds of all kinds of other trees with succulent 

 fruits. I have seen the pigeons thus resting on two or three small littoral trees, which as 

 yet form almost the only vegetation of Observatory Island, a very small islet at Nares 

 Harbour, Admiralty Islands." 



Jouan (Mem. Soc. Sci. Nat. Cherbourg, xi. 18G5, p. 101), remarks that, without being 

 exactly rare in the Marquesas, it is by no means so common there as in the Society 

 Islands, and a list of the trees remarkable for their size would not be a long one. He 

 adds: — "The poorest soils are not unfavourable to it, for one often sees seeds that have 

 been cast ashore on the coral islands germinating; and growing into trees." H. Mann 

 (Proc. Amer. Acad., vii. p. 156) includes it among the plants he regards as having been 

 introduced into the Sandwich Islands by the aborigines. 



The genus Calophyllum is also represented in the New World, and there are fruits of 

 one species in Mr Morris's collection of seeds and seed-vessels stranded in Jamaica. 



OCHNACE.E. 



Brackenridgea sp. (Plate LXI V., E.) 

 New Guinea drift. 



We are indebted to Dr Beccari for the determination of the genus of this singular 

 seed-vessel, of which there were a dozen or more in the collection, all of them more or less 

 encrusted. Brackenridgea was founded on a Fijian shrub, differing, among other things, 

 from Gomphia in having "the nearly annular ovule and seed curved around a large 

 projection into the cell of the ovary (in the manner of Menispermum), which arises from 

 its inner angle near the base " (A. Gray, Bot. U.S. Expl. Exped., i. p. 361, t. 42). A second 

 species, Brackenridgea zanguebarica, is figured in Hooker's Icones Plantarum, xi. p. 77, 

 t. 1096, the fruit of which is unknown ; but Oliver, loo. cit., says that he regards Bracken- 

 ridgea as a section of Gomphia rather than a good genus. Gomphia hooker i, Planchon 

 (Hook. Lond. Journ. Bot., vi. p. 3), from the Malayan Peninsula, should likewise be re- 

 ferred to Brackenridgea, if the genus be retained. In the present species the curvature 

 of the seed is not caused by a mere intrusion of the side of the carpel ; there are two 

 distinct cavities crossing and curving round each other at right angles, the one containing 

 the curved seed, the other empty. This empty cavity gives the fruit its buoyancy. 



So far as we are aware, no member of the Ochnacese has previously been recorded from 

 New Guinea. Neither Mueller, in his Descriptive Notes on Papuan Plants, nor Beccari, in 



his Notes on the Plants collected by D'Albertis, in the New Guinea of the latter, includes 



C 37 



(bot. CHALL. EXP. — PART III. 1885.) ^ ' 



