OCEANIC DISPEKSAL OF PLANTS. 293 



MYRTACE^E. 

 Barringtonia speciosa, Linn. f. 



Fruits in the New Guinea drift. 



There can be no question about the big square fruits of this tree being perfectly at 

 home in salt water. As we have set forth at p. 152, it reaches the most rLote islands 

 of the Pacific, and it is one of the first arboreal occupants of newly-raised coral islets. 

 Mr Moseley picked up on the beach of Arrou a large succulent flower of a species of this 

 genus. 



RUBIACE^E. 

 Guettarda speciosa, Linn. 1 



A fruit from the New Guinea drift. Belongs apparently to this very widely diffused 

 littoral shrub or small tree. 



APOCYNACEyE. 

 Chilocarpus sp. ? 



Much decayed fruit in the New Guinea drift. 



Cerbera odollam, Gcertn. 



Decayed fruits containing dead seeds in the New Guinea drift. The fibrous fruit 

 of this common sea-shore tree would doubtless bear long immersion without injury to the 

 exalbuminous seeds, though the seeds of these are quite dead. There are several unrecog- 

 nisable fruits in the collection which probably belong to this Order. 



Tabernaemontana sp. ? 



This is the vermilion-red fruit mentioned by Moseley in the Journal of the Linnean 

 Society (xv. p. 77). He found it in the Admiralty Islands, and abundantly at Arrou, 

 thickly scattered on the mud beneath the trees on the sea-shore, as well as in the drift off 

 the coast of New Guinea. Although the pericarp is completely decayed, the seeds appear 

 to be sound, and the embryo quite perfect. The complicate seeds, with a somewhat corky, 

 rugose testa, do not answer to the description given in Bentham and Hooker's Genera 

 Plantarum, but the seeds of comparatively few species are known. Indeed, this is very 

 near, if not the same as, Tabernceniontana aurantiaca of Gaudichaud in Freycinet's 

 Voyage Botanique, pp. 50 and 55, and tab. Gl, which was collected in the island of Rawak. 

 Gaudichaud's analysis of the seed, however, represents the embryo nearly as long as the 

 seed, with a radicle longer than the cotyledons, whereas in ours the embryo was found in 

 the axis of the albumen near one end of the seed, and not more than half its length, 

 while the radicle was very short. 



