566 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol 124. No. .?, September 2012 
observed between resident (Fig. 2A) and migrant 
birds (Fig. 2B). Migrant birds were mostly-leaf 
gleaning insectivores and arboreal insectivores/ 
frugivores. Resident birds were distributed over 
all foraging guilds. 
DISCUSSION 
We observed 102 species of birds in the 
sampled agroforestry fields. —19% of the 539 
species documented for Tabasco (Winker et al. 
1999). This number included Keel-billed Toucan 
(Ramp host os sulfit rains ). which is on the Mexican 
list of threatened species, and Olive-throated 
Parakeet and Montezuma Oropendola (Psaroco- 
liits montezuma), which are on the list of protected 
species (SEMARNAT 2002). Thus, agroforestry 
systems support important avifauna as reported by 
several authors (Hughes et al. 2002, Cardenas cl al. 
2003, Harvey et al. 2006, Bhagwat el al. 200K). 
The average number of species in the three 
counts of each agrotorcstry class corresponding lo 
the three seasons in each field in our study varied 
from 9.6 to 11.4, and the number of individuals 
from 22.1 to 38.8 (Table 2). Reitsma et al. (2001) 
observed in two 10-min point counts in crops 
under tree cover an average of 8.6 and 8.5 
individuals and 5.2 and 5.9 species in abandoned 
and managed cacao in Talamanca, Costa Rica. 
Arriaga-Weiss (2008) observed 15.4 species and 
54.3 individuals in three 10-min counts in the 
matrix of pasture lands within a distance of 500 m 
from forest fragments. Our numbers are within the 
range of those reported in other studies. However, 
the community we observed is more depauperate 
than that observed by Arriaga-Weiss (2008) in 
pasture lands near forest fragments, indicating 
the importance of these fragments for conserving 
avifauna. 
We found no significant differences in the 
number of species between pairs of agroforestry 
classes. This confirms findings of Cardenas et al. 
(2003) in Costa Rica, who observed 45 species in 
sequential and animal agroforestry, and 42 species 
in live fences (linear agroforestry). Species 
richness in both animal and linear agroforestry 
in our study was higher than found by Cardenas 
et al. (2003). and equal in sequential agroforestry. 
This may be due to an overall low forest cover 
in Tabasco resulting in birds seeking trees in 
the agroforestry systems as there are no others 
available. Life zone differences may also have a 
role, tropical dry forest in Costa Rica versus 
tiopica! moist and wei forest in Tabasco. 
Shannon diversity, equitability, and dominance 
indices of pooled data and of data for each season 
were similar in all agroforestry classes (Table 3). 
Cdrdenas et al. (2003) also found similar Shannon 
diversity indices in animal, sequential, and linear 
agroforestry. Only evenness showed difference' 
among agroforestry classes and was larger in 
linear agroforestry than in animal agroforestry and 
crops under tree cover (Table 3). 
Abundance ot birds was higher in animal 
agroforestry and agrolorests than in linear agro- 
forestry. whereas no significant differences were 
observed in abundance among animal agrofor¬ 
estry. agroforests, crops under tree cover, and 
sequential agroforestry. Cardenas et al. (2003) 
also found higher abundance of birds in animal 
agroforestry than in linear and sequential agro- 
forestry. This indicates there is no direct relation¬ 
ship between bird diversity and structural com¬ 
plexity. and tree diversity of agroforestry systems. 
We would have found higher diversity and 
abundance ol birds in agrolorests and in crops 
under tree cover if there was a relationship. The 
absence ol differences may be due to movement 
of birds between systems. 
We found mostly common species in all 
agroforestry classes, as few' rare species were 
observed. Seventy-five percent of the species we 
observed also occurred in the matrix of pasture 
lands near forest fragments (Arriaga-Weiss cl al. 
2008). Nineteen of the 23 species reported by 
Najera and Simonetti (2010) in oil palm planta¬ 
tions in Guatemala were also present in the 
agroforestry systems in Tabasco. The prevalence 
of common and low frequency of rare species 
is not a reason to disregard the contribution of 
agroforestry systems, as these conserve biodiver¬ 
sity in otherwise open landscapes, enhance 
species movements between habitat remnants, 
reduce human pressure on remaining forests, and 
complement conservation in protected areas 
(Bhagwat et al. 2008). 
Most observed species were forest generalise- 
and specialists of open areas as few forest 
specialist species occurred in agroforestry fields. 
Only 1% of species and 0.03% of individuals 
were forest specialists in our total sample. Most 
open area specialist species (18 vs. 11-13 in the 
other system classes) were in animal agroforestry 
fields, which are the most open agroforestry 
systems (Fig. I). Habitat preferences of the bird 
communities in agroforestry classes were quite 
distinct from those observed in forest fragments 
