vander Wal el al .AVIFAUNA IN AGROFORESTRY IN TABASCO. MEXICO 
561 
TABLE 2. Number ol bird species and individuals observed in eonnls in 32 fields of five agroforestry classes in 
Tabasco. Mexico in the rainy season. 2008 and the northern winds and dry season. 2009. AC = agroforestry class; nso = 
number of species observed in all points; ni = number of individuals in all points; as - average number ol species per 
point: pi = average number of individuals per point: ens estimated number of species; SI: sampling efficiency, A A — 
animal agroforestry; AF = agroforest; FA = linear agroforestry; SA = sequential agroforestry: C’T = crops under 
iree cover. 
Rainy season Season of northern winds 
AC 
a*» 
ni 
as 
ai 
nso 
ui 
as 
ai 
nso 
AA 
40 
432 
11.7 
48.0 
41 
329 
10.1 
36.6 
46 
AF 
27 
237 
8.4 
29.6 
31 
337 
11.6 
42.1 
40 
LA 
27 
99 
8.5 
24.8 
26 
103 
9,2 
20.6 
40 
SA 
28 
168 
11.0 
42.0 
23 
98 
10.5 
24.5 
24 
a 
25 
288 
7.9 
28.8 
40 
400 
10.4 
33.3 
46 
All 
64 
1.224 
9.4 
34.0 
66 
1.267 
10.4 
33.3 
76 
Dry season All seasons 
ni 
as 
ai 
nso 
ni 
as 
ai 
ens 
se % 
286 
12.5 
31.7 
64 
1.047 
11.2 
38.8 
81.3 
78.7 
280 
13.6 
35.0 
53 
854 
11.4 
35.6 
67.7 
78.6 
108 
11.6 
21.6 
59 
310 
9.6 
22.1 
85.9 
68.7 
77 
8.8 
19.3 
43 
343 
10.1 
28.6 
55.8 
77.1 
309 
12.0 
28.1 
60 
997 
10.2 
29.3 
73.6 
81.5 
.060 
12.1 
28.6 
102 
3.551 
10.6 
32.0 
123.8 
82.4 
and non-parametric Kruskal-Wallis and Mann- 
Whitney (7-tests in SPSS Version 17 to evaluate 
differences in richness and abundance, diversity, 
dominance, evenness, and cquitability among 
classes in all seasons. We used the raw data for 
comparing species richness among agrolorestry 
classes, and repeated the analysis with estimated 
richness for each agroforestry class in each 
season. We calculated the number of shared 
species and Chao-Jaccard abundance based sim¬ 
ilarity indices following Chao et al. (2000. 2005). 
We used Principal Component Analysis (PCA). 
and Analysis of Similarity (ANOSIM) and 
Similarity Percentage Analysis (SIMPER) in 
PAST (Hammer et al. 2001) to examine if 
agroforestry classes had different species compo¬ 
st lion and which species contributed most to 
observed differences. We also used ANOSIM and 
SIMPER to examine if agrolorestry classes had 
different abundance of feeding guilds. 
RESULTS 
The observation time per agroforestry class in 
each season varied from 4 to 12 hrs with a total 
observation time of 114 hrs (Table I). Sampling 
efficiency was >65% in all agroforestry classes 
and near 80% in all seasons. Overall sampling 
efficiency was 82.4%/ (Table 2), We observed 
3,551 individuals in the three seasons that 
belonged to 102 species (Appendix). Sixty-four 
species were observed in the rainy season. 66 in 
the season of northern winds, and 76 in the dry 
season. 
Analysis of raw data of all sampling points and 
transects (all seasons and all systems) indicated 
species richness did not vary among agroforestry 
classes (Kruskal-Wallis test. P = 0.527). Analysis 
of data of Jackknife I estimated species richness 
for all agroforestry classes revealed no differences 
(ANOVA. F = 1.374. P = 0.310). but abundance 
varied among classes (Kruskal-Wallis test. P = 
0.003). There were more birds at sampling points 
in animal agroforestry (Mann-Whitney (7-test. P 
- 0.002) and agroforests (Mann-Whitney (7-test, 
P = 0.001) than in linear agroforestry. There were 
no significant differences in abundance among 
animal agroforestry, agroforests, crops under tree 
cover, and sequential agrolorestry. The average 
species richness and abundance in the rainy 
season was not significantly different among 
agroforestry classes (Kruskal-Wallis lest, P = 
0.106 and P = 0.207). Species richness did not 
vary among classes in the season of northern 
winds and dry season (Kruskal-Wallis test, P = 
0.311 and P — 0.152), hut abundance varied 
(Kruskal-Wallis lest, P = 0.025). There were 
more birds in agroforests than in linear agrofor¬ 
estry in the season of northern winds (Mann- 
Whitney U- test. P — 0.007) and in the dry season 
there were more birds in agroforests than in linear 
and sequential agroforestry (Mann-Whitney 
(7-test. P = 0.040 and P = 0.006). The number 
of observed species differed among seasons 
(Kruskal-Wallis test. P = 0.001) and w'as higher 
in the dry season than in the rainy season and 
season of northern winds (Mann-Whitney (7-test. 
P = 0.001 and P = 0,020). 
Dominance. Shannon diversity index, and 
cquitability of the bird community, considering 
all sampling points and transects in all seasons, 
did not reveal differences among agroforestry 
classes (Kruskal-Wallis test. P — 0.829, 0.927, 
and 0.083), whereas some differences were 
observed in evenness (Kruskal-Wallis test. P = 
