Members of the Genus Ptilopus. 567 



followed, and consequently Mr. Elliot's statements that there 

 are, for instance, four green Ptilopi on the island of Tonga- 

 tabu alone, four in the Samoa Islands, and five in the Fiji 

 Islands"^, render it probable that some errors have crept in 

 here, since we have the repeated assertions of Mr. Layardf 

 and of the collectors of the Museum Godeffroy J that there is 

 but one species in the Tonga Islands, which is peculiar to the 

 group, only one in Samoa, and one in Fiji. Canon Tristram, 

 indeed, can hardly make the Tongan and Fijian birds distinct 

 (Ibis, 1878, p. 261), and the orange under tail-coverts of the 

 former, when adult, as against the yellow ones of the latter, 

 seem to constitute almost the only (^'3erence. 



But, apart from the statements of collectors, most natu- 

 ralists will find a diflficulty in believing that several forms so 

 closely resembling one another as these Central-Polynesian 

 Pigeons could possibly exist together on a small island like 

 Tonga-tabu, or on a group like the Fijis, with every island 

 within sight of others §, and yet keep distinct. One may 

 feel pretty sure that, if a bird of one species should stray to 

 an island inhabited by another, it would find a partner, and 

 its racial characters would become lost in its descendants, 

 for it is almost too much to suppose that the barrier of 

 sterility would be very strong between such closely allied 

 forms. It is only through separation and those other reasons 

 which we do not yet know sufficiently that these species 

 have developed the characters which entitle them to rank 

 as species. 



There are other reasons for believing that the three island- 

 groups of Central Polynesia have never possessed more than 

 one green Ptilopus each, for few species, probably, have been 



* I deduct the pale-coloured P. perousei, which is common to all three 

 groups, and the three Chrysosnce of the Fiji Islands, from the numbers 

 given by Mr. Elliot. 



t Layard, P. Z. S. 1876, pp. 495, 502 ; id. op. eit. 1877, p. 464 ; id. Ibis, 

 1878, p. 261 ; id. P. Z. S. 1879, p. 385. 



X Schmeltz, Verh. Ver. Hamburg, 1877 (1879), pp. 177-180, 



§ Taviuni, one of the most isolated, is only about 30 miles from Vatu 

 Lole, which is again about 15 miles only from Viti Levu. 



