( 329 ) 



On comparing a series of seven specimens from Camacnsa anil the River 

 Carimang, we notice consideral)le variation in the colour of the back, wliich, as 

 fiir as I can see, does not in any way depend on sex, as supposed by Mr. 8clater. 

 In two specimens (marked " S " and " ? " by the collector) head and back are 

 uniform pale olive-green, only that portion of the lower back adjoining the snlphur- 

 yellow rump, being fiiintly washed with pale rufescent. The same coloration of 

 the upper parts is shown by the type of Lesson's A. brasiUensis. Another male 

 from the River Carimang has the entire middle portion of the back rather more 

 distinctly sntfnsed with rufescent-brown, while in a female from Camacnsa the whole 

 mantle is almost uniform dark rufescent-brown. The other extreme is reiiresented 

 by a female from Camacnsa, which has the back of a uniform dull greyish green 

 without the slightest trace of the brownish admixture. Exactly the same variation 

 is shown in a series from the Caura River, Venezuela. Two specimens (both 

 marked " S " by the collector) have not only the back, but also the head (including 

 cheeks and ear-coverts), strongly suffused with russet. Some peculiarities in 

 coloration — viz. the broader and more rufous apical margins on the wing-coverts, 

 the rusty suffusion on the sides of the body, rump and forehead, etc. — clearly 

 indicate their immaturity ; and I have, therefore, little doubt that the russet-backed 

 specimens are the young, the green-backed ones the adult birds. 



As to Lesson's type, there may be further stated that it agrees in coloration 

 of the under-parts with a female from River Carimang. Throat and foreneck are 

 dull grey with narrow whitish longitudinal streaks, the middle of belly white, sides 

 and crissum pale yellowish. In other specimens, throat and chest are decidedly 

 olive-green, streaked with pale yellow ; but there is every possible intergradation 

 between these two extremes in the series before me. 



In South-east Brazil occurs a closely allied form which presents the same 

 variation regarding the colour of the back, and differs from its northern repre- 

 sentative only by its considerably larger size. It ought to be called Attila 

 brasiUensis uropygiata (Wied). One of the specimens in the British Museum 

 (" Rio "-make) ex coll. Sclater is marked as having been compared with the type 

 of ^1. brasiUensis in the Paris Museum. As a matter of fact, however, the latter 

 belongs to the smaller race found in Guiana, and the locality " Bresil," assigned 

 to it by Lesson, had already been corrected by Pncheran, a statement that seems to 

 have escaped the notice of Mr. Sclater. It is very unfortunate that a bird which 

 does not occur in Brazil should be called " brasiUensis" but there is no way to avoid 

 this regrettable change of nomenclature. 



These two forms have, consequently, to stand as follows : 



1. Attila brasiUensis brasiUensis Less.* 

 Altila bnisUh-nah Lesson, TralU (TOrn. 1831. p. 36n [" Br&U "— errore ! Tbo type came from 



Cayenne] ; Salvin, Ibis, 188.5. p. 30. 

 A. urvjnjgialis Sclater, Cat. Binln xiv. p. 360 (Brit. Guiana). 



MEASUREMENTS. 



3 t?c?. Brit. Guiana. Wing 85—87 ; tail ()5— 67^ mm. 



3 ? ?. Brit. Guiana. Wing 8(1—81 ; tail 60—62 mm. 



2 Sd (with green back). Caura River, Venezuela. Wing 84, 87 ; tail 67 mm. 



* A. vii-idescem RUlgWiiy, P. U. S. Mns. x. 18S7 (18HS) p. 522 [type ex Santarem] belongs, most 

 probably, also to A. i, braiiliemw ; I have not been able, however, to examine a specimen from the 

 typical locality. 



