570 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol 123. No. 3. September 2011 
Percent of time spent on stream 
FIG. I. The amount of time (transformed using arcsine 
square root to normalize the data) Louisiana Waterthrush 
foraged along streams in the Caribbean National Forest, 
Puerto Rico was negatively correlated with body condition 
(r = 0.57, l = 2.56, df = 15, P = 0.02). 
foraging along roads and in housing developmenis 
(X 2 3 = 10.7. P = 0.01), Second-year individuals 
spent more time foraging along roads and in 
housing developments than expected while forag¬ 
ing less along streams and muddy substrates than 
expected (r 3 = 10.7. P = 0.01). 
Mean body condition of individuals did not 
differ between years (Kruskal Wallis = 0.95, df = 
2, F = 0.62). Body condition did not differ 
between ASYs and SYs (Kolmogorov-Smimov Z 
= 0.936, P = 0.35). Individual body condition 
was not dependent upon prey availability within 
their home range (r = 0.07. t = 0.68, df = 7,p = 
0.53). Prey availability was not correlated with 
home range (r = 0.35, / = 1.21. df = 7, P = 0.26) 
or core area size (r = 0.27, / = 0.89, df = 7, p = 
0.40). Body condition was negatively correlated 
with amount of time foraging along the stream (r 
= 0.57, t = 2.56. df = 15, P = 0.02; Fig. I). 
Body condition was not correlated with amount of 
time foraging in muddy substrates (r = 0.13, / = 
0.49, df - 15, P = 0.63). Body condition was not 
influenced by the size of the home range (r 2 = 
0.12, t ~ 1.73, df = 14, P - 0.21) or core area (r 2 
- 0.13, / = 1.50, P - 0.16) overlap. 
Prey availability (AAIC c - 0.00) was the most 
parsimonious model that distinguished use from 
non-use locations. Increased prey availability and 
a lower percent of leaf litter cover (AAIC = 
TABLE I. I labital variables that characterized use ,inj 
random non-use locations of foraging Louisiana 
Wutcrthrush in the Caribbean National Forest Puerto 
Rico. The lowest AAIC, value indicates the model tliai 
best balances goodness-of-fil while minimizing the number 
of parameters in the model. tVi indicates the ranted model 
weights. Prey availability and percent leaf litter (LLi were 
selected as the two variables that best explain habitat use of 
Louisiana Waterthrush. 
Model 
AIC, 
aak; 
n< 
Prey 
13.09 
0.00 
0.542 
LL. Prey 
14.91 
1.82 
0.218 
Mud. Prey 
16.40 
331 
0.103 
LL. Water, Prey 
17.00 
3.91 
0.077 
Mud. LL, Prey 
18.17 
5.08 
0.043 
Mud. LL. Water, Prey 
20.43 
733 
0.014 
Mud, LL, Water, Prey, Veg 
Mud, LL, Fallen. Canopy. 
23.05 
9.% 
0.004 
Water. Prey, Vee 
27.27 
14.18 
0.000 
Mud, Canopy, Veg 
37.57 
24.48 
0.000 
Mud. LL. Water, Veg 
44.84 
31.74 
0.000 
1.82) also were parsimonious with respect to use 
and non-use locations (Tables 1, 2). Two addi¬ 
tional models that provided moderate support (2 
^ A AIC,. < 4) included higher percent of muddy 
subslrate and increased prey availability (A AlC c 
- 3.31), and lower percent of leaf litter cover, 
higher percent of standing water, and increased 
prey availability (A AIC,- = 3.91). Prey availabil¬ 
ity (1VV, = 1.00) was the single most important 
variable that distinguished use from non-use 
locations. Prey availability was higher along 
streams and in muddy substrates than non-use 
locations (F 2 = 7.9. P - 0.002; Fig. 2, Table 31- 
Prey availability did not differ between stream 
and muddy substrates (7-stat = 0.13, df = 16. P = 
0.90). The entire length of the Rio Sabana and its 
tributary was used by waterthrush during the 
study, and we were unable to compare use and 
non-use stream variables. 
DISCUSSION 
Waterthrush secured fixed home ranges that 
encompassed multiple habitat types. Home range* 
exhibited a wide range of overlap with neighbor¬ 
ing individuals although there was minima 
overlap of core areas. This overlap did not affec* 
body condition. There was greater overlap among 
ASYs despite having smaller home ranges. Thi' 
may be the result of older birds occupying high# 
quality habitats both along and away fro" 1 
streams. ASYs did not apportion greater percent- 
