APPENDIX 56 r 



inland species. This is an Old World genus containing some eighty species 

 mainly characteristic of tropical Asia and Malaya, and represented in the 

 South Pacific archipelagoes by two species, one Premna taitensis or 

 tahitensis, spread over the region and very near P. integrifolia, an Asiatic 

 species ; the other Premna serratifolia, an Asiatic plant found in Fiji, the 

 Marquesas, and other groups. Without endeavouring to give a precise 

 value to the Fijian plants, I will merely describe the prevailing forms, 

 which are, however, connected by intermediate varieties. These trees, I 

 may add, are known by the same name in the various Pacific groups, 

 " Avaro" or "Avalo " in Tahiti, " Alo-alo " in Samoa, " Yaro" and " Yaro- 

 yaro " in Fiji. 



The Fijian plants may be thus described . . . (a) Premna serratifolia, 

 an inland tree, growing in open woods and on the outskirts of the forest, 

 25 to 30 feet high, more or less hairy, leaves coarsely serrated with long 

 tapering points, putamen prominently tuberculated and thick-walled. 



(b) Premna taitensis or P. integrifolia, a low straggling coast tree or 

 shrub of the beaches, the coral islets, the swampy borders of the estuaries, 

 and the inland talasinga plains, its usual height being eight to ten feet, 

 except in the inland plains, where it is dwarfed, and three to five feet high. 

 It is more or less glabrous, the leaves being typically entire with obtuse 

 or retuse and mucronate apices. The putamen is thin-walled and rela- 

 tively smooth, (c) Intermediate forms found generally in the inland 

 plains or talasinga regions. 



On the Modes of Dispersal. Speaking generally, the small drupes of 

 both species float at first, but the soft parts are soon removed by decay, 

 and the stone is freed. In the case of the coast species, P. taitensis, the 

 stones float indefinitely and are often found afloat in rivers. In the case 

 of the inland tree, P. serratifolia, most of the stones sink at once, whilst 

 the others sink in a few days. It is probable that currents are one of the 

 effective agencies in distributing the coast species, but this could not apply 

 to the inland tree. The fruits of both the inland and the coast species 

 would attract birds, and the stones would resist injury in their crops. 

 This is the agency advocated by Prof. Schimper for the shore species, 

 P. integrifolia, of Indo-Malaya ; and fruits referred with a query to this 

 genus were found in the collection of seeds and fruits obtained by me 

 from the crops of pigeons in the Solomon Islands (Hot. Chall. Exped., 

 Introd. p. 46, part IV. p. 312). 



On the Cause of the Buoyancy of the Stone or Putamen of the Coast 

 Species. This is primarily connected with the empty seed-cavities, the 

 four-celled stone usually developing only one seed, the other cavities being 

 empty. This inference was established by the dissection of a large 

 number of stones, but it will be seen from the table below that one-seeded 

 stones are also frequent in the case of the inland tree (P. serratifolia), where 

 they as a rule sink. With either species the substance of the stone has no 

 floating power, but with the shore species, on account of the thin-walled 

 VOL. II 00 



