OCEANIC DISPERSAL OF PLANTS. 313 



The flora of the Arrou Islands is absolutely Papuan, and although so poor, it possesses a 

 certain interest, because the means by which the islands were supplied with plants are 

 evident, A large proportion of the species have fleshy fruits, which are eaten by birds, 

 especially pigeons, which have perhaps contributed more than any other animals to the 

 diffusion of plants. And it is noteworthy that the majority of the birds of Arrou are 

 carpophagous. Cassowaries, too, are active agents in dissemination, for they swallow 

 every kind of pulpy fruit, and with the swiftness of a horse convey them long distances 

 undigested ; they are also excellent swimmers, and traverse considerable expanses of water. 

 The plants growing near the sea are spread throughout the Malayan Archipelago and 

 New Guinea, and the fruits of nearly all are spongy or corky, and float from place to 

 place on the surface of the water, and are cast ashore all round the coast. Some of the 

 plants growing in the interior also have fruits of this nature, and are thus easily trans- 

 ported. Among plants whose seeds may be conveyed by winds only Asclepiadese were 

 observed. 



Setting aside the possibilities of seeds being occasionally conveyed long distances by 

 birds at one stage, there remains the probability of a species being disseminated by birds, 

 not necessarily of the same kind, by degrees, and in various directions. Of the kinds 

 of seeds that might be conveyed by the sea and birds, in various ways, there seems almost 

 no limit. The evidence afforded by the available data is far too incomplete for any 

 generalisation ; but members of the most different natural orders occur in oceanic floras, 

 which, perhaps, more than any others, owe their existence to the agencies in question. 

 Whether the present distribution of Phylica nitida (see Part II., p. 14S) was brought 

 about by the agency of birds is highly problematical. The distribution of the genus, like 

 that of many others of the African region, points rather to a former greater land-connec- 

 tion. Dr H. B. Guppy, however, hazards the suggestion (Nature, xxvi. p. 12) that seeds 

 might be transported from South Trinidad to Amsterdam Island, a distance of between 

 5000 and 6000 miles ! One thing specially noticeable of the majority of the plants, con- 

 cerning which we have certain evidence that their present areas are in part, at least, due 

 to oceanic currents, is, that they have either exalbuminous seeds, or, if albuminous, the 

 albumen is oily ; yet Darwin proved that the seeds of various Graminese which have farina- 

 ceous albumen, will bear two or three months' immersion in sea-water without losing their 

 germinative power. Another point to be considered is, whether very many of the common 

 plants which are littoral throughout a great part of their present areas, are essentially 

 littoral, or only accidentally so, in consequence of their seeds being conveyed uninjured 

 to distant shores, where they are able to thrive. Some of them also occur inland, at 

 remote distances from the sea, in localities where there is a minimum of salt in the soil, 

 while others flourish wherever they are planted. An examination and fuller discussion 

 of various agencies in plant dispersal will be found in our general Introduction. 



(bot. CHALL. EXP. — PART III. — 1885.) C 40 



