52 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 124. No. 1. March 2012 
habitats suitable for successful breeding and 
survival (Soderstrom 2001, Golawski and Meiss¬ 
ner 2008). 
My objectives were to examine one population 
ot Red-backed Shrikes breeding in high altitude 
meadows (>900-1,100 m asl) and another in low- 
hill farmlands to compare environmental charac¬ 
teristics of these two habitat types. 
Methods 
Study Area .— The study was conducted during 
2009 in the Marche Region of central Italy, which 
is characterized by a temperate climate, high 
spring and summer temperatures, and a marked 
summer drought (Tomaselli et al. 1972). The 
study area contained two breeding zones of 
5,573 ha total (farmland: 43° 47' 27.81" N, 12" 
39 39.91 E; 4,975 ha; high altitude meadows: 
43" 37' 51.84" N. 12 45’ 48.18" E; 598 ha). 
Monitoring Frequency.— The study area was 
monitored for breeding Red-backed Shrikes every 
5 days from early May to mid-July 2009. 
Occupied territories were surveyed at least once 
every 3 days during the breeding period to find 
nest sites (Bibby et al. 1997). Red-backed Shrikes 
are particularly sensitive to disturbance during 
nest building, egg-laying, and early incubation 
(Olsson 1995a. Tryjanowski and Kuzniak 1999), 
and the number of visits to each nesting site was 
kept to a minimum. Location of each nest site 
was recorded using a Global Positioning System 
(GPS) and numbers of fledged young per pair near 
a nest were recorded. The density of Red-backed 
Shrike territories in the two habitat types was 
calculated as breeding pairs per 10 ha. 
Environmental Parameters.- The typical size 
ot a Red-backed Shrike territory varies from I-? 
to 5-6 ha (Cramp and Perrins 1993. Lefranc I993~ 
Olsson 1995b. Harris and Franklin 2000). A 3-ha 
territory size around the nest site was standardized 
to enable territory preferences to be studied 
ArcGIS 9 software (ESR1 2009) was used to 
delineate a polygon with a 100-m radius around 
(AAVV w'„f nd a llmd ' USC "' ap " :l0 -00<» 
I i ." .' ) was llsccl to characterize the 
habitat in the polygons. 
The environmental parameters studied for each 
nest site were: ( 1 ) nest shrub: the plant species 
where breeding pairs placed their nest and w e 
very high (>5() items); (3) the altitude of the 
nesting site (m above sea level); and (4) com¬ 
position and percentage of land-use types around 
the nesting area (cultivated, uncultivated vine 
yard, shrub, forest, reforestation, grassland, river, 
buildings, and roads) 
The features and structural characteristics 
studied for each nest site were: (I) ‘open space' 
around the nesting site: percentage of cover of 
roads, grasslands, uncultivated land, and shrubs: 
(2) 'edge density' as the sum of the perimeters of 
all polygons in the buffer zone per number of 
land-use types/100 (as a surrogate of the habitat 
fragmentation level in the buffer zone); (3) road 
type (paved, unpaved); (4) distance from the 
nearest road (m); and (5) distance to the nearest 
building (m). 
Statistical Analysis. —Differences in the plani 
species used for nesting in the two habitat type-' 
were compared with Chi-square tests using 
standardized residual analyses to highlight the 
associations of plant species with habitat tjpes. A 
Mann-Whitney U- test was also performed to ex¬ 
amine differences in environmental parameters of 
breeding sites in farmland areas versus those in 
high altitude meadows. The structural cbaiacler- 
islics (open space and edge density) between the 
two habitats were further compared with a Chi- 
square test. 
I also performed a logistic regression analysis to 
identity which variables best explained the pre¬ 
sence or absence of Red-backed Shrikes in the 
farmland study area. Presence and absence of 
active nests were used as response variables, while 
composition and percentage of land-use types 
around the nesting area, structural characteristics, 
and road and building distance were the explana¬ 
tory variables. Internal validation was performed 
by calculating the non-parametric confidence 
interval of the AUC (Area Under the Curve 1 
values. The AUC generally ranges from 0,5 lor 
models with no discrimination ability to 1.0 l° r 
models with perfect discrimination. Data are pre¬ 
sented as means ± SE and all tests were conducted 
with SPSS lor Windows software. Version I ■ 11 
(® SPSS Inc.. Chicago. 1L. USA. www.spss.com 
RESULTS 
I monitored and characterized 150 Red-backed 
Shrike nest sites in two breeding zones: 132 m 
farmland (mean altitude = 347.2 ± 138.8 m asl) and 
18 in high altitude meadows (mean altitude = 837.6 
— 115.5 m asl) (Fig. 1). The density o! breeding 
