604 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 123. No. 3. September 2011 
(2) show that acoustical communication has a 
major role in flock roost congregation. 
METHODS 
Study Area. —The study was conducted be¬ 
tween January 2006 and January 2007 at Brasilia, 
DF (15° 47' 07" S, 48° 02' 02" W). We sampled 
during early morning (0600-1000 hrs) and in the 
late afternoon (1600—1830 hrs). The area consists 
of sensu stricu Cerrado. savanna-like vegetation 
common in central Brazil. 
Flock Size.— We recorded flock sizes during the 
morning, when flocks were flying or feeding 
(feeding flocks), and also in the late afternoon, 
during roosting preparation (roosting flocks). We 
also recorded flock size during feeding (feeding 
flocks) at four different plants: Eriotheca pub- 
escens (Bombacaceae), Qua lea parviflora (Vo- 
chysiaceae), Caryocar brasiliense (Caryocaceae), 
and Mimosa claussenii (Mimosaceae). Mimosa 
claussenii, the smallest, has a maximum trunk 
diameter of II cm (Silva-Junior 2005) and 
maximum height of -3 m (CBdeA, pers. obs.) 
The other three plants had larger sizes, reaching 
~8 m. Eriotheca pubescens has a trunk diameter 
up to 85 cm, Qualea parviflora = 65 cm, and 
Caryocar brasiliense = 68 cm (Silva-Junior 
2005). We compared flock size of parrots using 
M. clausenii with flock sizes of parrots using the 
larger trees with a unilateral U -test (Biostat 5.0) 
(Ayres et al. 2007), assuming that large flocks are 
unsuitable in small trees and. therefore, flocks 
feeding on M. claussenii should be smaller. 
Acoustical Communication. —We recorded vo¬ 
calizations using a parabola, a Shure Beta-58 
cardioid microphone, and a Nagra-E analog 
recorder. Recordings were made on magnetic 
tape with a speed of 19 cm/sec, and digitized with 
a 24-bit and 44.1 kHz resolution. We used Cool 
Edit Pro 2000 for the analysis (FFT sample size of 
512). 
We were able to relate the physical character¬ 
istics of the calls with the message meaning 
through sonograms and response observations, 
based on Weaver’s (1949) broad definition of 
communication. We measured: call duration, 
minimum and maximum fundamental frequen¬ 
cies, and fundamental frequency band used to 
physically describe each call. Some calls present¬ 
ed pseudo pulses (very short sound with a low 
frequency resolution), in which case we also 
measured the number of pseudo pulses and 
pseudo-pulse rates. Pulse rates were measured in 
pulses per second (Hz). The recordings are 
deposited at the Neotropical Sound Archive 
(ASN) at the University of Campinas (Unicampi. 
We observed large variations in sound intensity 
during the analysis. The relative intensities were 
measured whenever the same individual emitted 
two different vocalizations. We only used record¬ 
ings of a single calling individual to assure that we 
were using the same individual for this analysis 
We used a single individual at first for measure¬ 
ments of congregation calls for comparison with 
the group congregation call. We amplified the 
track until its peak reached 0 dB, and then 
measured the intensities of each call type. Data 
are presented as means ± SD. 
RESULTS 
Flock Size. —Flocks of Yellow-faced Parrots 
arrived at the study site from Brasilia National 
Park (4 km distant) during early morning. Feeding 
flocks had 7.7 ± 8.2 individuals (n = 116 flocks), 
Flocks flew southeast of the area and we lost track 
of them at about 1000 hrs. Feeding flocks returned 
to the area and gathered during late afternoon. A 
single large roosting flock departed for Brasilia 
National Park just before nightfall. Roosting 
flocks could be as large as 90 individuals, and 
had 79 ± 10.5 individuals on average in = 7 
flocks). Feeding flocks varied from a single 
individual to 40 individuals feeding in the same 
tree (8.9 ± 8.8, n = 89 flocks). Feeding flocks 
were not as large as roosting flocks. 
Feeding flock size varied according to plant 
species. Flocks were larger during use of larger 
trees ( Qualea parviflora. Eriotheca pubescens. 
Caryocar brasiliense) when compared to Mimosa 
claussenii, (unilateral U -test P < 0.05 for all 
comparisons). 
Acoustical Communication. —We analyzed 
160 min of recordings which were visually 
selected based on degradation and signal to noise 
ratio. We ignored noisy and low-level recordings. 
Several repertoire components were found for the 
species. Communication was divided between 
long-range and short-range signals, based on high 
and low amplitude sounds, respectively. 
Flight Call. —This vocalization was common 
during observations (Fig. 1A). Individuals were 
often heard vocalizing during flight but also when 
perched. The call is equivalent to the functional 
song, which carries the species-specific recogni¬ 
tion code. The call has a fundamental frequency 
between 0.5 and 1.5 kHz, and duration of 240 ms 
