Australian Birds in the Collection of the Linncan Society. 203 



out of the genus Ampelis, Linn., and some conterminous birds. 

 He however still assigned it a station in the series of his genera, 

 immediately subsequent to the Linnean genus Coracias. The 

 bird evidently exhibits strong relations to both these groups. In 

 its colours, its size and strength, and its general form, it bears a 

 resemblance to the Rollers ; while in the wide gape of the bill 

 it equally seems to approach the broad-billed Berry-eaters. 

 Judging, however, even from external characters, we have no 

 hesitation in stating our opinion, that these relations are merely 

 analogical, and that the natural station of the bird is in the 

 tribe of Fissirostres, but at that extremity of the group where it 

 joins the Dcntirostres. The shape of the wing, evidently formed 

 for a rapid flight, — the comparative shortness and weakness of 

 the legs, indicating the pursuit of an aerial rather than of a ter- 

 restrial prey, — and the notched apex of the bill, denoting its 

 food to be rather animal than vegetable, — all are characters 

 Avhicli point out its affinity to the Fissirostral group, whose per- 

 fection consists in their powers of flight, and of inclosing their 

 prey within their gape as they pursue it on the wing. The 

 brilliant colours of this bird, which have hitherto led to its being 

 placed among the Rollers, equally associate it with the present 

 family of Todidœ and with the neighbouring Halcyonidœ : while 

 the bill appears decidedly formed according to the same model 

 as that of Eurylaimus* and the conterminous group oi Podargits. 

 We had much pleasure in finding that our conjectures respecting 

 the affinities of this group, which we had drawn merely from its 

 external characters, were confirmed by the accounts of its habits 

 and mode of life, which we have extracted above from IMr. Ca- 

 ley's manuscripts. We must observe in addition, that this natu- 

 ralist, arranging his birds without any predilection for systems, 



* See Linn. Trans, vol. xiv. p. 430; and ZooL Researches in Java, no. '2. 



2 D 2 and 



