1874.] BIRDS FROM THE ANDAMAN ISLANDS, 259 



CoUocalia nidifica (Lath.), Jerdon, B. of India, i. p. 182. no. 103 (1862). 

 Collocalia fuciphaga (Thunb.), \Yallace, P. Z. S. 1863, p. 384. no. 6. 



CoUocalia nidifica, G. E. Gray, Ann. & Mag. N. H. (3) xvii. p. 118, "Java, Sumatra, and 

 other isl. E. archipel." (186G). 



CoUocalia innominate, Hume, Str. Feath. i. p. 294, "Andamans" (Feb. 1873). 

 CoUocalia spodiopi/gia, Peale, Hume, torn. cit. p. 296, " Andamans." 

 CoUocalia inexpectata, Hume, /. c, " Andamans." 



" S. Andaman : March 1, iris brown, bill black, legs brown, feet darker ; March 24, s ." 

 The large number of Andaman specimens I have been enabled to examine, collected by 

 both Messrs. Wardlaw Eamsay and Wimberley, in no material respect differ from Sikim and Ceylon ibig, 1874 

 individuals ; nor am I enabled to find any important character whereby they can be separated P- 134. 

 from Seychelles, Mauritius, or Reunion examples (Ilirundo francica, Gm.). 



The dorsal feathers in all examples from the above-named localities have the tips of the 

 basal portion of the webs pure white. This can only be detected by parting the feathers ; for the 

 overlapping terminal and exposed part of the dorsal feathers is uniform smoke-brown. The 

 extent of white on the edging of the webs increases as the feathers descend the back, so that 

 those which clothe the uropygiura have more of the edges of their webs, both in length and 

 breadth, coloured white. The result is that the white sometimes becomes partially exposed. In 

 some of the shorter of the upper tail-coverts the white colour of the webs is still more developed, 

 occasionally forming a conspicuous w^hite edging ; but no covert is entirely white, the tip and 

 central part of each being of a varying shade of mouse-colour. It is thus that the albescent or 

 pale mouse-coloured band on the rump observable in many examples of this species is produced ; 

 and it is frequently made more prominent in the dried skins by the mode of preparation of the 

 specimens. In three examples of true C. francica from Mauritius and Pieunion, kindly lent me 

 by Professor Newton, a pale band is discernible ; in another from the Seychelles it is absent. In 

 a Ceylon individual in the collection of Mr. Holdsworth it is also entirely wanting. My Sikim 

 specimens have the band as much developed as in those from Mauritius ; Andaman birds are not 

 to be distinguished, all of them exhibiting, more or less, a pale band on the rump. That there 

 is a tendency in this section of the genus CoUocalia to evolve a pure white band on the rump is 

 shown in C. troglodytes and other more eastern species, in which we find it a permanent and well- 

 determined character. But in none of the races of the species under notice does it appear to be 

 stable, or sufficiently and constantly developed to make it a trustworthy differential character. 

 In all other essential respects birds from the localities alluded to are identical ; and I therefore 

 adopt Gmelin's title as being the oldest. 



Since writing on CoUocalia affinis (Ibis, 1873, p. 302)* I have had an opportunity of Ibis, 1874, 

 comparing it with Horsfield's type specimen of C. linclii, and I find that it in no way differs. ^' ■'^^' 

 Hirundo fuciphaga was described by Thunberg from Javan examples ; and I have no doubt 

 whatever that Horsfield's C. linchi-=H. fuciphaga, Thunb. The diagnosis oi H. fuciphaga, ample 

 in its details, applies in every respect to C. linchi ; while the last phrase, " differt ab H. esculenta 

 Cauda tota atra immaculata," of itself marks the species ; for with the exception of the white 

 spots on the lateral rectrices, Moluccan C. esctdenta (Linn.) is barely to be distinguished from 

 Javan C. linchi. The synonymy of C. fucipthaga will therefore be as follows : — 



* lAntea, p. 239.— Ed.] 



