Teta et al • VARIATION IN WESTERN BARN OWL DIETS 
591 
RESULTS 
We identified 5,152 prey items, mostly native 
sigmodontine and exotic murid rodents (Table I). 
Mammals were present at all sites, whereas birds 
and amphibians were present at 79.1 and 50. (Ur ot 
the sites, respectively (Table I). We identified 
4.815 mammals of which 4,729 (98.2%) were of 
species with mass <250 g. Larger mammals 
(>500 2 ) occurred in low frequencies and 
nduded young lagomorphs ( Lepus europaeiis), 
caviomorphs ( Cavia aperea ), and marsupials 
(Didelphis albiventris). 
The native sigmodontine rodents Akocloit 
azarae, Calomys spp. (including C. laucha and C. 
nmculinus), and Oligoryzomys flavescens com¬ 
posed 85.8% of the total prey items (Table I). 
These species were prominent toward periurban 
{Akodon azarae, Oligoryzomys flavescens) and 
rural ureas (Calomys spp.). and were replaced bv 
exotic murid rodents, such as Mus musculus and 
Rottus spp. in urban settlements (Table 1. fig. 2). 
Birds, especially passerines, were present in low to 
moderate proportions (0.4-40.0%) in most sam¬ 
ples. increasing in frequency al urban and periur¬ 
ban areas (Table 1, Fig. 2). 
FNB varied between LOO and 2.92. while FNBsl 
varied between 0 and 0.48; both parameters 
increased in values at intermediate levels oi built 
Habitat, decreasing towards the extremes ol the 
gradient and describing second order polynomial 
functions (R : = 0.655 and R : - 0.468, respectively). 
The two first axes generated by PCA analysis 
accounted for 99.3% of the variance in the diet 
'Fig, 2). Representation of sites and prey categories 
i.e.. Orders Anura. Columbiformes, Passeriformes. 
Didelphimorphia. Chiroptera. Rodentia [exotic]. 
Rodemia [native], and Lagomorpha) defined by 
'he two first factors segregated urban from periurban 
and rural samples. Prey in urban sites included 
mostly exotic murid rodents, while in periurban and 
rural areas the most common prey were native 
sigmodontine rodents (Fig. 2. Table I). Birds were 
more abundant at urban and periurban areas. 
DISCUSSION 
Bam Owls feed primarily on micromammalian 
prey with weights between 10 and 150 g (Taylor 
2004). Stenophagy and specialization, which are 
characteristics of these owls, decrease when the 
diversity and abundance of its main prey decreas¬ 
es (Taylor 2004). Birds, amphibians, arthropods, 
and bats have been reported as prey of Barn Owls 
when preferred micromammalian prey species 
declined (Vargas ct al. 1984. Bose and Guidali 
2001). such as in urban areas (Charter et al. 2007). 
However, our study documented that small 
mammals represented the main prey items, and 
were only partially replaced by other taxa at some 
sites (e.g.. by birds in some urban and periurban 
areas [e.g., sites 12. 13, 16; Fig. 11 or amphibians 
near wetlands or water courses |e.g.. sites 4. 15: 
Fig. I|). Barn Owls in our study, unlike other 
birds of prey that switch their diet from small 
mammals in rural areas to birds in cities (e.g.. 
Yalden 1980. Pikula et al. 1984. Kubler et al. 
2005). maintained their selectivity as micromam¬ 
mal predators. Dominance of small mammals in 
the diet at all sites explained the low values tor 
FNB and FNBst. Overall, our results are in 
agreement with those ot Bellocq (1998). Leveau 
ct al. (2006), and Gonzalez-Fischer et al. (2011) 
who studied food habits ot Barn Owls at similar 
latitudes in central-eastern Argentina: they also 
found high predation on micromammalian picy 
and low values of food niche breadth (at the level 
of Order) for this species. 
Urban development produces some of the 
greatest extinction rates and frequently eliminates 
the large majority of native species (McKinney 
2002 and references therein). This is certainly true 
in the case of native sigmodontine rodents, which 
were almost completely replaced by exotic rats and 
mice in urban areas of central-eastern Argentina 
(e.g.. Massoia and Forties 1967, Cavia et al. 2009). 
Hercolini (2007) described in detail the micro- 
mammal communities along the same gradient that 
we studied and suggested that exotic rodents, such 
as Rati us spp. and Mus musculus prevail at the 
urbanized extreme, while mice of the genus 
Calomys spp. were the most frequent species in 
rural areas. High proportions of some native 
species (e.g.. Akodon azarae , Oligoryzomys Jlaves¬ 
cens) occur at middle portions ot the gradient in 
periurban areas surrounded by large patches of 
parklands and spontaneous vegetation (Hercolini 
2007: Tabic 1). This ‘periurban peak' in species 
diversity is a relatively well-known pattern, 
previously reported for taxa such as birds, lizards, 
bumblebees, and butterflies (Racey and Euler 
1982. Pawlikowski and Pokomiecka 1990. Blair 
2001. Germaine and Wakeling 2001). 1 he increase 
of FNB and FNBst values at intermediate urban¬ 
ization levels in our study agrees with this pattern. 
It is usually accepted that Bam Owls may 
capture commensal rodents in low frequencies. 
