Birds of Celebes; Dicaeidae. 451 



in the Himalayas is known, that it is impossible to doubt that they are very 

 closely related (see figures of both in Ibis 1874, pi. I): the Ceylonese bird has 

 changed in wing-structure, but very little in coloration, showing how a generic 

 difference may be evolved, while specific characters remain con- 

 stant'). Acmonorhynchus quadricolor of the Philippines has the back red, but 

 the subterminal or basal part of the feathers black, "producing a mottled appear- 

 ance" (Sharpe), and, if Mr. Keeler's hypothesis — that the tip shows the 

 more recent and the basal part the former colour of a feather — be correct {?), 

 then A. quadricolor once had a black back, which would make it somewhat like 

 A. vincens and Pachyglossa. It would be of interest to know if the immature 

 Philippine bird has a black back. The forms of the Celebes area and Flores 

 with their olive-green upper surface (without black bases) are much more dis- 

 tinct. An investigation of the pigments in them and in Pachyglossa might be 

 made with advantage. 



From the foregoing it appears that Pachyglossa has some claim to be re- 

 garded as the ancestor of Acmonorhynchus. In the East Indies it has undergone 

 much change, whether by reason of different food, climate, closer interbreeding, 

 different animal foes, or whatever else may be supposed to have caused it to 

 develop in fresh directions. This unknown quantity has not been encountered 

 by the bird in the Himalayas, or it must have changed there also. Mr. Wallace 

 has already (c 2) spoken of -1. aureolimbatus as a Himalayan type in Celebes, and 

 though it cannot now be regarded as one "unknown in any Malay island", since 

 Mr. Biittikofer shows that the forms ofSangi, Flores, one of the Philippines, 

 and Ceylon are of one genus, there is certainly some reason to think that it 

 and the others originally came from the Himalayas. 



Acmonorhynchus, with the closely allied genera Piprisoma, Pachyglossa and 

 Prionochilus, does not occur in the Australian Region, being unknown east of 

 Timor and the Molucca Strait. The relationship of A. aureolimbatus with Dicaeum 

 may be seen in the finely serrated cutting edges of the terminal third of its 

 bill, a character which, as Gates shows, connects the true Dicaeidae (a family 

 in which some other genera are wrongly included by Dr. Sharpe) with the 

 Nectariniidae . 



jf * 178. ACMONORHYNCHUS SANGIRENSIS (Salvad.). 

 Sangi Yellow-sided Flower-pecker. 

 Plate XXVn. 



a. Prionochilus sanghirensis (1) Salvad., Ann. Mus. Civ. Gen. IX, 1876, 59; ('S; Meyer, 

 Isis, Dresden 1884, 6; (3j Sharpe, Oat. B. X, 1885, 71; (4) W. Bias., Omis 

 1888, 591. 



I; See, also, our speculations on the origin of the Femes and Spizaeti under Pernis celebeitsis, and of 

 certain Kingfishers vmder Ceycopsis. 



57* 



