190 PROCEEDINGS OF THE ACADEMY OF [April 



ered identical with Collocalia unicolor, from which, however, it may 

 readily be distinguished by its feathered tarsi, darker, less brownish 

 upper surface, with rump appreciably lighter than the back. It is 

 most closely allied to Collocalia fuciphaga, and by Dr. Hartert is con- 

 sidered a subspecies of this; but its large size and other characters, 

 combined with its isolated range, so far as C. fuciphaga is concerned, 

 quite clearly indicate its specific distinctness. Furthermore, there is, 

 so far as is known at present, no intergradation in size between Collocalia 

 brevirostris and any form of Collocalia fuciphaga; and the race of the 

 latter, C. f. elaphra, which is nearest in size is most different in color. 

 The so-called intermediate specimens from the western Himalayas, 

 mentioned by Dr. Hartert,^^ have proved to be examples of Collocalia 

 unicolor, a form belonging to the group without feathers on the tarsus, 

 and have thus no bearing on the question. Therefore, until actual 

 proof of intergradation with Collocalia fuciphaga be forthcoming, the 

 present form should stand as a full species. 



Collocalia lowi (Sharpe). 



Cypselus lowi Sharpe, Proc. Zool. Soc. Lond., 1879, p. 333. 



Cypselus labuanensis , Ibis, 1879, p. 116, in text {noinen nudurn). 



Chars, sp. — Like Collocalia innominata, but rump without a well- 

 defined light band. 



Wing, 127-134; tail, 50-57; exposed culmen, 5; tarsus, 10-12 mm. 



Type locality .—\^2ih\\2in Island, northern Borneo. 



Geographical distribution. — Northern Borneo; Palaw^an Island, Philip- 

 pine Islands; Anamba Islands; accidental in Sumatra and on Nias 

 Island (Hartert). 



This species resembles Collocalia leucopygia and Collocalia innomi- 

 nata in possessing dark shaft lines on the lower surface, and in this it 

 differs with them from all the other members of the subgenus. In 

 color above and below — except for the lack of a light rump band — 

 as w^ell as in size and proportions, it is identical with Collocalia innomi- 

 nata. It is so much larger, and usually has the tail so much less 

 deeply emarginate, -than Collocalia fuciphaga and all subspecies of the 

 latter that further comparison is unnecessary. One specimen in the 

 United States National Museum has the tail by no means square, but 

 considerably emarginated, from which it is evident that there is con- 

 siderable individual variation in this regard. 



Subgenus COLLOCALIA Gray. 



Chars, suhgen. — Tarsus entirely without feathers. 

 Type. — Hirundo esculenta Linnaeus. 

 15 Cat. Birds Brit. Mus., XVI, 1892, p. .502. 



