574 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 124. No. 3. September 2012 
constant calling and individuals moving in the 
same direction) without the presence of clumped 
food resources such as fruiting plants. Differen¬ 
tiation of mixed-species bird flocks from other 
bird groups, such as frugivorous birds foraging in 
trees, was straightforward due to the conspicu¬ 
ousness of mixed-species bird flocks enhanced by 
their constant movement (Moynihan 1962, Morse 
1977). We noted all times each species was 
observed during our fieldwork in 2005 to obtain 
information about species tendencies to partici¬ 
pate in mixed-species bird flocks (Arbeldez- 
Cortes el al. 201 lb) and considered if they were 
within Hocks or apart from them. We recorded the 
time of day. species composition, number of 
individuals per species, and the relative position 
of the mixed-species bird Hocks in the vegetation 
once a mixed-species bird flock was located. Data 
about foraging maneuvers, following behavior 
among species, and vocalizations were noted 
when possible. We checked the direction of 
movement of each mixed-species bird flock and 
compared it with the next mixed-species bird 
flock detected to assure it was coming from a 
different direction to avoid repeatedly recording 
data from the same mixed-species bird flocks. 
Statistical Analyses .—We used cluster analyses 
in PAST 3 (Hammer et al. 2001) and the Jaccard 
index to examine if composition of mixed-species 
bird flocks differed among three elevation hands 
of 150 m each. This also provided information if 
mixed-species bird flocks encountered consecu¬ 
tively were similar. We conducted multiple Chi- 
square and Fisher’s exact tests for species 
observed in >10% of the mixed-species bird 
flocks to identify which species pairs co-occurred 
more than expected by chance (Cole 1945). We 
designated a species as a potential nuclear species 
following Powell (1979) if its’ presence was 
correlated with the presence of at least two other 
species by Chi-square and Fisher's exact tests. We 
also considered qualitative trails that seem to 
be related with a cohesive role including: (I) 
contrasting or conspicuous plumages. (2) high 
occurrence, (3) large intraspecific groups, ( 4 ) 
conspicuous maneuvers. (5) regular calls, and 
(6) a tendency to be followed more than they 
followed others (Moynihan 1979, Powell 1985. 
Hutto 1994. Sridhar et al. 2009. Goodale and 
omT C u imP 201 °’ Harrison and Whitehouse 
2011). We calculated the proportion of occurrence 
m mixed-species bird flocks on each elevation 
hand for those species considered as potential 
nuclear species. Taxonomy and nomenclature 
follows Gill and Donsker (2012). 
RESULTS 
We recorded 42 species (Table 1) in 64 mixed- 
species bird flocks during 121 hr.s of fieldwork. 
Forty percent of all species observ ed participated 
regularly (>10% of all events) in mixed-species 
bird flocks (Table I) highlighting the significance 
of this behavior in this avian community. Main 
other species observed in flocks were occasional 
participants; 14 species (22%) were observed only 
in one mixed-species bird flock (Tabic 1). 
Tanagers (Thraupidae) were represented by die 
highest number of species (17) with the Scarlet- 
bellied Mountain-Tanager ( Anisognathns igniwn- 
tris) being most frequent (recorded in 35 ntixed- 
speeies bird flocks). This species was common in 
(he area blit was observed more times alone than 
in mixed-species bird flocks. In contrast, nine 
species were more often in mixed-species bud 
flocks than apart from them (Table I). The 
number of species and individuals per mixed- 
species bird flock ranged from two to 19 (mean - 
5.1) and from three to 66 (mean = 11.5), 
respectively. Some species were observed in 
groups of up to seven individuals but most were 
recorded in pairs or small groups (Table 1). 
Number of species and individuals per mixed- 
species bird flock were correlated i r = 0.91). 
Mixed-species bird flocks did not cluster by 
dittereni elevation bands (Fig. 2), and only a few 
flocks in the same elevation band had some 
similarity (Jaccard index >0.6). Cluster analysis 
also indicated the more similar mixed-species bird 
flocks were not observed consecutively, support¬ 
ing the observation that we did not repeatedly 
record the same flock. Mixed-species bird flock* 
were more frequent in the morning during 070% 
0800 hrs (17% of the mixed-species bird flocks). 
We observed mixed-species bird flocks moving 
principally in the sub-canopy and the understory 
I lie analysis of co-occurrence of species pair* 
( Table 2) indicated 15 species pair comparisons 
were significant (all P < 0.05) with 10 positive 
and five negative. Six co-occurrences could be 
deemed significant (y = 0.05) only by chance 
because of the multiple comparisons conducted 
Five species were observed in two or more 
positive co-occurrences (Table 2) and exhibited 
several attributes related to nuclear roles in flocks 
( Table 3). Those species considered as potential 
nuclear species were: Pearled Treenmner {Mar- 
