Denes el al. • LEPTODON FORBESI IS A VALID SPECIES 
327 
coverts, secondaries, and inner primaries in both 
the type and MZliSP specimens of Leptodon 
turksi have conspicuous white tips which are 
absent in L cayanensis. Photographs taken in the 
field survey clearly show this character in al least 
one wild specimen (Fig. 2A). White tips of 
feathers seem to wear quickly in the wild. 
The holotype of L forbesi has a broad distal 
white tail band (contra pale gray in L. cayanensis) 
both dorsal and ventrally with a narrower white 
band hidden under the tail coverts. The gray band 
that separates the two white bands on the ventral 
side Is also paler in color. The broad white band is 
wider in the MZUSP specimen, and on the ventral 
'ide of the tail the distal black band is narrower 
and discontinuous in some feathers, and the 
proximal black band is absent. Ten of the L. 
forbesi recorded in the field had one distal and one 
proximal black band and one white band on the 
upper side of the tail, while four had a third, 
medial black band and two white bands. Six 
individuals had a single distal black band on the 
underside of the tail (Fig. 2B). while 13 had one 
distal and one medial black band. It is difficult to 
see the proximal tail bands in the field, as they are 
usually hidden under the tail coverts. We recorded 
°ne juvenile molting into adult plumage with 
■hree visible dark bands in the undertail. The tail 
bunds in L forbesi are white, and those in L 
cayanensis are pale gray. 
Pam-.—Beak black. Cere and eye-ring 
c °l°r varies from ashy gray to a lighter pale pearl 
gray, Eyes dark. Tarsi show yellowish to gray 
color. 
Distribution. —Recent museum specimens of 
Leptodonforbesi were collected in the eastern part 
01 the State of Alagoas in northeastern Brazil 
'big- 3) in Sinimbu, Murici. and Sao Miguel dos 
Campos (Teixeira et al. 1987a). We found L. 
f"rbesi in all localities surveyed (Fig. 3). Addi- 
'tonal surveys in the forests north and south of the 
'umpled areas, including the states of Paraiba. 
Ser gipe. and Bahia arc needed to ascertain the 
distribution limits of L forbesi and L cayanensis. 
Habitat,—Leptodon forbesi occurs in lowland 
; md highland (0 to 585 m) Atlantic Forest. Birds 
have been observed soaring over open country 
Ku gar cane fields), when moving between forest 
fragments, but otherwise appear to avoid such 
‘toas. Birds were also recorded flying over 
mangroves (Rhizophnra spp.). 
Behavior and Ecology’.—We gathered Iife- 
bistory information on L. forbesi during field 
surveys in Alagoas and Pernambuco. Most birds 
were detected between 1 and 2 hrs after sunrise, 
for periods not longer than 30 min, when they 
soured over the forest. These kites frequently flew 
in pairs in October 2007 and November 2008, 
using the Butterfly Display flight (Thorstrom 
1997). This behavior is best known for L. 
cayanensis and is an indicator of reproductive 
activity. We recorded feeding once, when a 
distant perched bird flew down, then returned to 
the same perch seconds later, and began to eat its 
prey. Prey identification was not possible, even 
using a telescope. Detection frequency in Febru¬ 
ary 2008 was lower and display flights were rare 
suggesting the breeding season had ended. No 
nests or juveniles were recorded in either survey. 
Voice. —Vocalizatiotts of the White-collared 
Kite heard during field surveys apparently show 
no meaningful differences from known Grey¬ 
headed Kite calls. Mostly heard was the wuh-wuh- 
wuh..., or caw-caw-caw..., in series of 10-20, in 
flight, which may elicit response from individuals 
of nearby territories. We also heard the cat-like 
eeeAAW. often during the Butterfly Display flight 
(Thorstrom 1997). Conclusions on the vocal 
repertoire of these species should be viewed as 
preliminary, despite this apparent similarity, until 
several better quality vocalization recordings are 
made, allowing comparative sonogram analysis. 
Subspecific Review of Grey-headed Kite.— 
Plumage color patterns arc quite variable within 
and among subspecies and do not follow the 
geographic patterns that would allow delineation 
of the subspecies as proposed by Swann (1922, 
1945), which was also noted by Hcllmayr and 
Conover (1949). This variation should be consid¬ 
ered as variation among individuals within 
species, as is common in the Family Accipitridae 
(Grossman and Hamlet 1964, Brown and Amadon 
1968. Thiollay 1994, Ferguson-Lees and Christie 
2001 ). 
The morphometric variation analysis shows that 
wing length and tail length follow Bergmann’s 
Rule (Fig. I )• Although restricted to wing and tail, 
this might explain the greater size of the southern 
populations of L. cayanensis as described by 
Hellmayr and Conover (1949). No significant 
variation was found, when controlled lor latitude, 
to justify separating the subspecies, including the 
beak variables (culmen and width, which did not 
require correction for latitude). Tail band width 
variations also do not support the subspecies 
described by Swann (1922) for L. cayanensis. The 
