Segura and Reboreda • BOTFLY PARASITISM OF RED-CRESTED CARDINALS 
109 
Data Collection — Nests were found by search¬ 
ing systematically in potential nest sites and by 
observing nesting behavior of territorial pairs 
(Martin and Geupel 1993) of Red-crested Cardi¬ 
nals. We found 367 nests {n = 108, 120, and 139 
for the breeding seasons of 2005-2006, 2006- 
2007, and 2007-2008, respectively). Nearly 50% 
of the nests (n = 111) were found during 
construction and laying with the remainder found 
during incubation (n = 152) and after hatching 
(n = 38). We used 131 nests that survived at least 
6 days after the first nestling hatched (n = 36, 45, 
and 50 for 2005-2006, 2006-2007, and 2007- 
2008, respectively). We used this criterion in our 
study because botfly parasitism occurred while 
nestlings were between 1 and 6 days of age. 
Inclusion of nests depredated before nestlings 
were 6 days of age would result in underestima¬ 
tion of parasite prevalence. 
Nests were checked daily until all eggs hatched 
and then every 2 days until the nestlings fledged 
or the nest failed. Nestlings were marked after 
hatching on the tarsus with black ink for 
individual identification and color banded after 
day 6. We recorded: (1) day of hatching for each 
nestling, (2) number of nestlings hatched, (3) day 
we found the first larvae in the nestling, (4) day 
the nest failed or fledged young, and (5) number 
of young fledged for each nest. We recorded: (1) 
body mass, (2) lengths of the beak, right tarsus, 
a nd wing, and (3) parasite intensity (number of 
botfly larvae/nestling) for each nestling at each 
nest visit. 
We measured body mass with 30 and 50 g 
Pesola spring scales (accuracy ± 0.2 and ± 0.5 g, 
respectively), length of the tarsus and beak with a 
dial caliper (accuracy ± 0.1 mm), and length of 
the wing with a ruler (accuracy ±0.1 mm). We 
minimized the effect of daily variation in body 
mass and size by collecting these data between 
1600 and 1900 hrs. 
We analyzed the structure of the vegetation 
surrounding the nest by measuring vegetation 
characteristics at three different scales: (1) nest 
tre e, (2) vegetation surrounding the nest tree, and 
( 3) landscape. We measured (1) tree species, (2) 
nest height, (3) distance from the nest to the edge 
°f the canopy, and (4) cover of the canopy at the 
nest tree scale. We measured the cover of tree 
canopy within a 15-m radius of the nest at the 
surrounding nest tree vegetation scale, and 
whether the nest tree was in the continuous strips 
°f forest parallel to the river or in small isolated 
forest patches more distant from the river at the 
landscape scale. We used images QuickBird (5 m) 
extracted from Google Earth (Digital Global 
Coverage, 6 October 2008) to calculate the cover 
of individual nest trees and proportion of canopies 
in the nest surrounding area using Program 
IDRISI Kilimanjaro 14.01 (Clark Labs 2003). 
Data Analysis.—We assumed a nest was 
successful if it fledged at least one young and 
depredated if all nestlings disappeared between 
two consecutive visits. We did not observe 
abandonment of nests with nestlings in circum¬ 
stances other than botfly parasitism. We assumed 
that a nestling died as a result of botfly parasitism 
if it was previously parasitized and found dead or 
disappeared between visits with no evidence of 
attack by predators (i.e., feathers or blood in the 
nest). 
We estimated the lethal effect of botfly 
parasitism by comparing nestling survival (pro¬ 
portion of nestlings that fledged) between non- 
and parasitized nests excluding nests that were 
depredated. We estimated the sub-lethal effects of 
botfly parasitism by comparing growth rates of: 
(1) body mass, (2) tarsus length, (3) beak length, 
and (4) wing length between non- and parasitized 
chicks that survived. We used brood means to 
avoid pseudoreplication. We calculated growth 
rates as the slope of a linear regression of the 
values of each variable versus age of nestlings 
between 2 and 8 days of age (hatching day = age 
0). Growth rates of all the variables were almost 
linear for nestlings 2-8 days of age (body mass, y 
= 3.0 x + 1.6, r = 0.99, P < 0.001; tarsus length: 
y = 2.1 x + 5.9, r = 0.99, P < 0.001; beak length: 
y = 0.56 x + 4.9, r = 0.99, P < 0.001; and wing 
length: y = 4.7 x - 0.79, r = 0.99, P < 0.001; n 
= 222 data points from 66 non-parasitized nests). 
We only considered nests in which we had three 
or more measurements in that period (16 parasit¬ 
ized and 66 unparasitized nests). We used nests 
with nestlings during January and February only 
for analysis of the association between vegetation 
characteristics and botfly parasitism, as the 
occurrence of parasitism in nests with nestlings 
during the previous months was practically zero. 
We used parametric tests for normally distrib¬ 
uted data only, and nonparametric tests with 
corrections for ties. We used Mann-Whitney U or 
Ki'uskal-Wallis tests for independent compari¬ 
sons. We used logistic regressions to analyze the 
association between botfly parasitism (binary 
dependent variable) and one or more independent 
