( 53 ) 



greyish, the axiUaries and nnder wing-coverts dirty whitish (never pale yellow as 

 in E. f. fiiscatus), and the upper parts light rufescent brown, decidedly clearer 

 and more reddish than in the typical race. No. 2136 slightly points towards 

 E. f.fuscatus by having a faint yellowish tinge on the chest and flanks, while 

 in other respects it agrees with the birds from the Rio Aragnaya. 



The typical form, E. f. fuscatus* inhabits the forest region of the 

 states of Bahia, Rio de Janeiro, and S. Panlo, has much darker and duller 

 upper parts, pale yellow breast and belly, the chest more washed with brownish, 

 etc., etc. 



84. Myiarchus tyrannulus bahiae Berl. & Lev. 



[Musckapa tijrannulus P. L. S. Miiller, Natursysl. Siipjil. p. 169 (1776. — ex Daubenton, PI. Eid. 541, 



fig. 1 : Cayenne).] 

 Myiarchus bahiae Berlepsch & Leverkuhn, Ornis vi. p. 17 (1890. — Bahia). 

 M.ferox (nee Gmelin) Pelzeln, I.e. p. 116 (part. : Golaz). 



Nos. 2018, 2044. Si ad., Goyaz, April 1906.— Wing 90, 93 ; tail 87 ; bill 

 20, 21 mm. 



No. 2020. ? ad., Goyaz, April 1906.— Wing 94J; tail 90j ; bill 20 mm. 



No. 2283. S ad., Rio Aragnaya, Jnly 1906.— Wing 93 ; tail 88^ ; bill 21 mm. 



Besides these, I have examined a female, obtained by Count Castelnan (Mus. 

 Paris), and an adnlt male secured by Natterer, in the Vienna Museum. 



These six specimens agree with a series from Bahia in having no rnfous border 

 to the inner web of the outermost rectrix, but are much paler, more olivaceous 

 grey, instead of dull brown, on the upper parts, resembling in this respect 

 M. ti/rannulus chlore.piscius Berl. & Lev., from Mattogrosso and Eastern Bolivia. 

 The few examples of the latter form, which I have seen, show a broad rufous 

 border to the inner vane of the outermost rectrix, like M. t. tyrannulus from 

 Cayenne, Venezuela, etc. 



In view of the variability of this character in other species of Mtjiarckus, 

 however, I do not lay much stress on the markings of the tail-feathers, inasmuch 

 as, according to Dr. Allen, t in M. t. chlorepiscius from Mattogrosso the rufous 

 border to the external rectrix is also sometimes absent. Therefore I should not 

 be surprised if a larger series would prove the Goyaz form to be referable to 

 M. t. chlorepiscius. 



85. Tyrannus melancholicus melancholicus Vieill. 



Tyrannus melancholic^ Vieillot, Noiiv. Did. xxxv. p. 84 (1819.— ex Azara No. 198 : Paraguay); 

 Pelzeln, I.e. p. 117 (Goiaz). 



Nos. 2077, 2078, 2093, 2094. (?(???, Goyaz, May 1906. 

 Nos. 2275, 2298. ? ?, Rio Aragnaya, July 1906. 



86. Muscivora tyrannus (Linn.). 



Muscicapa lyraimus Linnaeus, Syst. Nat. xli. 1. p. 325 (1766.— ex Brisson : "Canada" errore !— and 

 Cayenne). 



No. 2442. i ad., Agua Suja, Minas Geraes, alt. lOOO metr., October 1906. 



* Mmcipetafuscata Wied, Beitr. Naturg. Braiil. 3. ii. 1831. p. 902 (Hio de Janeiro), 

 f BuU. Amer. Mtts. iv. 1892. pp. 3 12-3 : " M. tyrannulus." 



