790 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 123. No. 4. December 2011 
(2 sites) and parking lot (2 sites). Each site was 
embedded within a larger area of the same type of 
land use. The parking lots consisted of large areas 
of pavement for parking spaces, interspersed with 
buildings and islands of mowed grass, shrubs, or 
trees. The residential neighborhoods consisted of 
a mosaic of roads, houses, and yards. The yards 
had variable amounts of mowed grass and 
ornamental shrubs and trees that were both native 
and nonnative. The pastures consisted of scattered 
trees and large areas of grass that were periodi- 
cally grazed by cattle. Fences in the pastures were 
often covered with vines and shrubs. The wildlife 
preserve consisted of scrub trees and shrubs 
scattered among areas of short grass and bare 
ground with an occasional large tree. These open 
areas were variably surrounded by pine (Pinus 
spp.) forests or xeric forests. Study sites were 
between 2.7 and 52.9 km apart. We calculated the 
percentage of ground covered by buildings, 
pavement, grass, open areas (sum of pavement 
and grass), trees, and other (water or undeter¬ 
mined surface) for each study site based on 
Google Earth satellite images. We designated 
wildlife preserve and pasture as rural land-use, 
and parking lot and residential areas as urban 
land-use. We only sampled nests from rural 
habitats in 2007. The highest densities of 
mockingbirds were in residential areas, followed 
by parking lot and wildlife preserve; pasture had 
the lowest densities (Stracey 2010). 
Spatial Pattern of Prevalence and Intensity 
of Parasitism .—We searched each study site for 
mockingbird nests. Active nests were monitored 
every 1-3 days until nestlings fledged or were 
depredated. No botflies were detected on the 
nestlings during handling. We collected nests after 
nestlings fledged and placed them in sealed plastic 
bags. We only collected nests from which at least 
one nestling fledged to ensure that if there were 
parasites in the nest they had sufficient time to 
pupate. We also noted the date when clutches 
were initiated, nest height, and the type of plant 
containing the nest. 
We dissected each nest and counted the number 
of pupae, pupal cases, and adult flies in the 
nesting material. Adult flies were identified by G. 
J. Steck (Florida Department of Agriculture and 
Consumer Services) and voucher specimens were 
deposited in the collection of the Florida Depart¬ 
ment oi Agriculture and Consumer Services. All 
pupae, pupal cases, and adults were identified as 
• porten except a few pupal cases which 
belonged to a parasite (a tachinid fly) of P. 
porteri. These pupal cases were excluded from the 
study. 
The parasite intensity in each nest was defined 
as the number of pupae plus the number of pupal 
cases in the nest. The number of nestlings in a nest 
varied from one to four, and we defined the 
average parasite intensity per nestling as the nest 
intensity divided by the number of nestlings at 
hatching day. Average parasite intensity per 
nestling, however, is only an estimate of parasite 
load because parasites may not be spread evenly 
among nestlings and may concentrate on one ora 
few nestlings (Christe et al. 1998). We calculated 
the proportion of parasitized nests in each site and 
for each year. 
Statistical Analyses .—We used R Software 
(2010; Version 2.12.1) to test the effect of year, 
month in which the nest was initiated, clutch size, 
land-use category (urban, rural), habitat (parking 
lot. residential, pasture, wildlife preserve), study 
site, nest height, and plant type (tree, shrub, vine 
or huilding/object) on parasitism status of the nest 
(parasitized or not parasitized) using logistic 
regression. The different study sites were also 
defined by ground-cover variables, and we tested 
the effect of these variables (percentage of 
buildings, trees, and open areas) on parasitism 
status of the nest using logistic regression. We 
used a stepwise selection process for both model 1 ' 
based on A 1C to remove useless variables. P 
values for the effect of the different variables were 
obtained by performing a Type III sums-of-squarc 
analysis on the final models. We also analyzed the 
effect of these variables on the number of 
parasites per nestling for parasitized nests using 
ANC'OVA. We log-transformed the data on 
parasite intensity per nestling to meet assumptions 
of normality. We used a stepwise selection 
process for both models based on AIC to remove 
useless variables and P values for the effect of the 
different variables were obtained by performing a 
Type III sums-of-square analysis on the final 
models. 
RESULTS 
We collected 73 nests in 2008 and 26 nest* 
exclusively from rural areas in 2007. Thirty-eight 
percent of the nests were parasitized in 2008 
versus 42% in 2007 (Tables 1, 2). The number ol 
parasites in a nest ranged from 0 to 85 with a 
mean ± SE of 22.85 ± 3.99 parasites in the 
parasitized nests in 2008, and from 0 to 88 with a 
