Animal Biology (2012) DOI 10.1163/157075612X631213 brill.

nl/ab

Habitat use by sympatric red and roe deer in

a Mediterranean ecosystem

Rita T. Torres1,2,∗ , Emílio Virgós3 , João Santos1 , John D.C. Linnell2 and
Carlos Fonseca1
1
CESAM & Department of Biology, University of Aveiro, Campus de Santiago,
3810-193 Aveiro, Portugal
2
Norwegian Institute for Nature Research, 7485 Trondheim, Norway
3
Biodiversity and Conservation Area, University of Rey Juan Carlos, c/ Tulipán s/n,
28933 Móstoles, Madrid, Spain

Accepted: February 2, 2012

Abstract
Much recent research has focused on understanding the environmental factors that limit the distribu-
tion and abundance of species. However, by no means all models consider the effects of interspecific
competition on species’ distributions. In this study, we explore the contribution of both environmental
factors and the presence of another ungulate species on the distribution of two sympatric deer species:
roe deer (Capreolus capreolus) and red deer (Cervus elaphus). To assess the importance of both biotic
and abiotic factors, we constructed models for each species incorporating environmental predictors
and an index of abundance of the other species. High density of shrubs and distance to roads had
a positive affect on roe deer occurrence while spatial heterogeneity, namely mean shape index, and
the presence of red deer had a negative effect. In contrast, the percentage of shrub cover, landscape
heterogeneity and the presence of roe deer all positively affected red deer occurrence. Our results
suggest that interspecific competition between these species might occur with asymmetric effects and
underlines the need to explore the nature of these interactions in greater detail in southern ecosystems.
© Koninklijke Brill NV, Leiden, 2012.

Keywords
Capreolus capreolus; Cervus elaphus; estimated competition coefficient; interspecific competi-
tion; pellet group counts; Portugal

∗) Corresponding author; e-mail: rita.torres@ua.pt

© Koninklijke Brill NV, Leiden, 2012 DOI 10.1163/157075612X631213

2 R.T. Torres et al. / Animal Biology (2012)

Introduction
Understanding the spatial and temporal distribution of organisms is a central topic
in ecology (Buskirk, 2005). Many studies have developed models to identify key en-
vironmental factors that influence species distribution (Guisan and Thuiller, 2005;
Hortal et al., 2006). However, some of these models fail to integrate biotic relation-
ships, in particular interspecific competition, that may affect the observed patterns
(Ritchie et al., 2009; but see Acevedo et al., 2007; Acevedo et al., 2010). Many stud-
ies have suggested that, when resources are limited, the potential for competition
is high among sympatric species of large herbivores, and may result in some type
of behavioural or ecological displacement to avoid competition before such com-
petition actually occurs (Putman, 1996; Prins et al., 2006). Niche differentiation is
often the basis for the coexistence of species (Putman, 1996; Richard et al., 2009).
Red deer (Cervus elaphus) and roe deer (Capreolus capreolus) are the most nu-
merous and widespread deer species in Europe with overlapping distribution ranges
in a wide variety of ecosystems (Apollonio et al., 2010). Red and roe deer are ex-
cellent model species to study interspecific interactions, because they occur in a
diverse array of habitats (Borkowski and Ukalska, 2008), exhibit considerable mor-
phological variation (Hofmann, 1989), have extensive distributions (Clutton-Brock
et al., 1982; Andersen et al., 1998) and shape ecosystems through the effects of
their herbivory (Latham et al., 1999). All these factors provide an opportunity to
increase knowledge on habitat use and potential interspecific interactions between
these two species on shared ranges.
Despite some studies of interspecific interactions between both species, the na-
ture of competition between them remains uncertain. Latham et al. (1997) anal-
ysed the variation in roe and red deer densities across many Scottish forests and
showed a negative effect of red deer density on that of roe deer, whereas the re-
verse was not observed (Latham et al., 1999). Conversely, Borkowski and Ukalska
(2008) found that red and roe deer pellet group densities were not negatively corre-
lated, suggesting that interspecific competition was not an important factor in their
study. Danilkin (1996) reported that red deer, sika deer (Cervus nippon) and fallow
deer (Dama dama) may displace roe deer from feeding grounds (behavioural in-
terference), suggesting that the roe deer are particularly vulnerable to competitive
displacement by other, larger, deer species’ (Latham et al., 1999). Furthermore, Fer-
retti et al. (2008, 2011) also showed active behavioural interference between fallow
and roe deer, with the former able to displace and exclude the latter from feeding
grounds, also through direct aggression. Even though these studies are good evi-
dence for actual displacement (or “interference”), they need careful interpretation
because, unless it actually prevents roe deer from obtaining crucial resources, they
demonstrate only interference, not interference competition. This underlines that
competition evidence is generally very difficult to obtain. However, Richard et al.
(2010) showed that the density of red deer had a negative effect on the body mass
of roe deer fawns, suggesting that the interaction does have a unidirectional effect
and that it is truly competition. However, most of these studies have been conducted
 Animal Biology (2012) DOI 10.1163/157075612X631213 3

in the temperate and continental parts of central and western Europe where densi-
ties are high and these species inhabit the core area of their distribution (Putman,
1996). Rather few studies have focused on Mediterranean ecosystems (San José et
al., 1997; Ferretti et al., 2011), which appear to be suboptimal habitats for these
species (Telléria and Virgós, 1997). Conditions in different parts of the range of a
species may lead to geographical variation in habitat use, depending on its ability
to adjust to environmental changes. The novelty of this study is that in Portugal,
roe and red deer are at the extreme edge of their distribution ranges (Mitchell-Jones
et al., 1999) and it has been shown that such populations can have a higher sus-
ceptibility to any stress, including interspecific competition (Hoffman and Blows,
1994).
In the Trás-os-Montes region, northeastern Portugal, red deer and roe deer oc-
cur sympatrically with a high degree of broad scale space overlap (Vingada, 1991).
Our objective was to build on early work on roe deer habitat association (Torres et
al., 2011) and to try to explore possible interactions between both species, e.g., the
increase of red deer densities in the area (Santos, 2009) can represent a negative
pressure on roe deer densities, which remain far below those found in central and
northern Europe (Vindada et al., 2010). Interspecific competition occurs most fre-
quently among species that have high overlap in diet and usually has asymmetric
effects, with larger species being competitively superior (Brown and Maurer, 1986).
Our aim was to estimate competition coefficients for roe and red deer on the
southwestern edge of their distribution. Our prediction, derived from previous stud-
ies (Latham et al., 1999), was that roe deer habitat use should be negatively affected
by red deer presence, whereas the reverse should not occur. Therefore, the best sup-
ported model of roe deer habitat use will include red deer presence, reflecting the
likelihood of interaction between the two species. Such information will contribute
to an understanding of interspecific competition among ungulates in a Mediter-
ranean habitat, and should assist in developing more appropriate management and
conservation strategies for these two deer species.

Materials and methods

Study area
The study was conducted in Montesinho Natural Park and Serra da Nogueira,
Trás-os-Montes, northeastern Portugal (6◦ 30 -7◦ 12 W and 41◦ 43 -41◦ 59 N), both
sites are part of the European Union’s Natura 2000 network, covering an area of
75.000 ha (fig. 1). The terrain consists of rolling hills with elevations ranging from
438 to 1481 m. The climate is Mediterranean with mean monthly temperatures
varying between 3°C in the coldest month and 21°C in the warmest month and
mean annual precipitation varies between 1000-1600 mm. The vegetation is var-
ied and characterised by oaks Quercus pyrenaica and Quercus rotundifolia, sweet
chestnut Castanea sativa, scots pine Pinus sylvestris, maritime pine Pinus pinaster.
Main understory species are heather Erica australis, Pterospartum tridentatum,
4 R.T. Torres et al. / Animal Biology (2012)

Figure 1. Map of Portugal (left), highlighting where the field survey was done (Trás-os-Montes region,
northeastern Portugal – right). On the right, there is the distribution of the sampling plots.

rockrose Halimium alyssoides, gum rockrose Cistus ladanifer and Lavandula sam-
paioana. The study area is crossed by a number of rivers and small streams and
the associated vegetation is mainly common alder Alnus glutinosa, narrow-leafed
ash Fraxinus angustifolia, black poplar Populus nigra and Salix salviifolia strongly
linked to mountain meadows. The area exhibits a mosaic of deciduous and conif-
erous forest, fragmented by small-cultivated fields. Roe deer are a native species
in the north of Portugal where populations have always persisted in a few patches
in the study area. Due to their low abundance, hunting is very restricted, occur-
ring in a few touristic hunting grounds only (Vingada et al., 2010). Vingada et al.
(2010) estimated that current populations of wild roe deer in Portugal as a whole
are somewhere between 3000 and 5000. Red deer populations were much reduced
in Portugal before the 1970’s, but there has since been an increase resulting from
natural dispersal process from Spain and re-introductions programs (Vingada et al.,
2010). According to the official game statistics, about 2000 red deer are harvested
annually (considering the last three years) in Portugal (Direcção-Geral dos Recur-
sos Florestais), mostly on intensively managed touristic hunting grounds (Vingada
 Animal Biology (2012) DOI 10.1163/157075612X631213 5

et al., 2010). In our study area the average red deer density varies between 2.27 and
4.70 red deer/100 ha (Santos, 2009). The region has a low human population den-
sity of 9.5 people per km2 , clustered in small villages. A number of national roads,
which provide connection between Portugal and Spain, cross the study area.

Data collection
Field work was conducted between October 2007 and August 2009, using pellet
group counts (Neff, 1968). Sampling plots were distributed along 6 transects, each
one consisted of a 1 × 1 × 1 km triangle (3 km in total); which is an efficient
field design (the start point and end point are at the same place). Transects were
randomly located with the help of technical staff from the Natural Park and were
drawn on a 1 : 10 000 map scale, distributed to provide an adequate coverage of
all the main habitat types in the study area (Mayle et al., 2000). In order to maxi-
mize spatial coverage and to mitigate sampling dependence, segments were evenly
spaced along the transect line. For each sample plot, we firstly assessed the pres-
ence of roe and red deer by recording the number of pellet groups and then, the
habitat variables that potentially could affect their distribution were recorded in a
10 m radius circle. A pellet group was defined as containing six or more individual
pellets, and identified as being produced at the same defecation (similar size, shape,
texture and colour) (Mayle et al., 2000). Pellet groups lying on the boundary of seg-
ments were alternately counted and ignored (Marques et al., 2001). Discrimination
between pellets of both species was possible by differences in size and shape (in
cases of doubt, pellets were neglected). Using experienced field staff and maintain-
ing one fixed observer throughout the field survey minimized observer bias. A total
of 120 sample plots (50 m × 2 m) were examined within our study area, which were
visited five times throughout the field survey. To avoid any complication regarding
different faecal decay rates, a pilot study on the study area was conducted to de-
termine decay rates and the mean number of days needed for pellets to disappear
was 182.5 ± 49.4 (S.E.) (R.T. Torres, unpubl. data). Therefore, the length of time
between visits was chosen to be smaller than the time to disappearance. Because of
the low number of pellet groups detected in all plots, data were pooled across years
to improve estimates.

Environmental variables
Recent studies have emphasized the need to adopt a framework of multiple scales
(Rotenberry et al., 2006). Therefore, we used ecological descriptors that ranged
from patch to landscape scales: in the field, we measured micro-habitat variables
and we used GIS to derive several macro-habitat and landscape variables. Based on
the literature (Virgós and Tellería, 1998) and our own predictions, we measured 34
variables that could potentially influence the distribution of red deer (see table 1).
Habitat variables, abundance data, and field methods regarding roe deer are fully
described in Torres et al. (2011). These variables were classified into 4 classes:
6 R.T. Torres et al. / Animal Biology (2012)

Table 1.
Description of the variables used to model roe and red deer distribution, at patch and landscape scales,
in Trás-os-Montes region, northeastern Portugal (October 2007-August 2009). Cover of the main tree
and shrub species were visually estimated using a five-point scale.

Variables Description

HABITAT STRUCTURE
TREECOV Tree cover (%) (height > 2 m)
SHRCOV Shrub cover (%) (height 50-200 cm)
GRNCOV Ground cover (%) (height < 50 cm)
AGRICOV Agricultural field cover (%)
MEADCOV Meadow cover (%)
PINCOV Pinus sylvestris and Pinus pinaster cover (%)
OAKCOV Quercus pyrenaica cover (%)
DECCOV Deciduous cover (%)
ILEXCOV Quercus rotundifolia cover (%)
SATCOV Castanea sativa cover (%)
CISTCOV Cistaceae cover (%)
ERICOV Ericaceae cover (%)
THORNCOV Thorny shrubs (Rosa, Rubus, Crataegus, Prunus) cover (%)
LEGCOV Leguminosae (Pterospartum tridentatum, Cytisus sp.) cover (%)
PTERICOV Pteridium aquilinum cover (%)
POACOV Poaceae cover (%)
HALCOV Halimium alyssoides cover (%)
CANCOV Measure of mean overstory density (%)
VISINDEX Measure of lateral visibility (m)
HUMAN FACTORS
DIST_RIV Distance to the nearest river (m)
DIST_ROAD Distance to the nearest asphalted road (m)
DIST_POP Distance to population nuclei (m)
BROAD SCALE
PROP_1 Proportion of all habitats inside a 1.26 km radius-home range (%)
PROP_2 Proportion of all habitats inside a 12.6 km radius-pop. range (%)
LANDSCAPE STRUCTURE (from McGarigal and Marks, 1995)
NP Number of patches. Total number of patches in the home range
ED Edge density. Sum of length of all edge segments divided by total area.
All edge segments are defined.
MSI Mean shape index. Shape complexity. Equals 1 when all patches are
irregular (polygons) or square (grids)
SDI Shannon’s Diversity Index. Measure of diversity in community
ecology, applied at landscape level.
SEI Shannon’s Evenness Index. Is expressed such that an even distribution
of area among patch types results in maximum evenness. As such,
evenness is the complement of dominance.
 Animal Biology (2012) DOI 10.1163/157075612X631213 7

(1) micro habitat structure, (2) macro habitat composition; (3) human factors, and
(4) landscape structure (see table 1 for details).
(1) Micro habitat structure
Variables were estimated in the field on 10 m radius plots centred on each segment
in the field and were used to evaluate abundance of food and cover. Food was as-
sessed by visually estimating percentage of tree, shrub and ground cover within a
five-point classification scheme: 0 (0%), 1 (1-24%), 2 (25-49%), 3 (50-74%) and
4 (75-100%). Deer find concealment and cover in dense stands. Visual conceal-
ment, provided by cover, was estimated by measuring horizontal visibility (m). It
was estimated by placing a cover board (∼80 cm high) in the centre of the seg-
ment. Then, in a random direction, the minimum distance required for the board to
be completely hidden was noted (Mysterud, 1996). Canopy cover is also important
as protection from adverse weather and thermal extremes and was estimated using
Lemmon’s densiometer. It was measured against open sky in the north, south, east
and west, and the average constituted the cover (%).
(2) Macro habitat composition
Variables were obtained from CORINE Land Use/Land Cover database, with a spa-
tial resolution (pixel width) of 250 m. The available categories were aggregated into
five general land cover types: (1) agricultural fields, (2) scrubland, (3) meadows,
(4) deciduous, and (5) coniferous. The abundance of red and roe deer pellet groups
was investigated at a “population” spatial scale (buffer with 12.6 km) and a “home
range” scale (buffer with 1.26 km). The smallest buffer was calculated based on
home ranges values for roe deer, found by Carvalho et al. (2008), in Portugal. We
used ArcMap to build the buffers.
(3) Human disturbance
Red deer are hunted in the study area. Then, human disturbance factors can influ-
ence deer distributions as they may be considered as analogues to predation risk. To
analyze the effects of human disturbance, distances were measured from the cen-
tre of each segment to the nearest border of the following features: closest human
house and closed paved roads. We also measured the distance from the centre of the
plot to the closest water body.
(4) Landscape structure
Here we defined landscape as an area containing a mosaic of habitat patches in
which a particular target habitat patch is often embedded (Dunning et al., 1992).
Spatial arrangement of vegetation cover types can influence deer habitat use (Saïd
and Servanty, 2005). Both species are influenced by landscape structure and het-
erogeneous landscapes may favour high densities in contrast to more uniform ones
because of high food availability and proximity to cover (Patthey, 2003; Saïd and
Servanty, 2005). The buffers previously constructed were used to calculate indices
of landscape structure, e.g., each landscape characteristic was calculated based
on the two buffers. We used FRAGSTATS, in the Patch Analysis extension for
8 R.T. Torres et al. / Animal Biology (2012)

ArcView 3.2 (Elkie et al., 1999), to quantify landscape characteristics, specifically

number of patches (np), edge density (ed), mean shape index (msi), Shannon’s di-
versity index (sdi) and Shannon’s evenness index (sei).

Statistical analyses

A combination of univariate and multivariate statistical approaches was used. Many

authors have developed statistical techniques to estimate competition coefficients
among sympatric species (MacArthur and Levins, 1967; Schoener, 1974). Fox and
Luo (1996) proposed a standardization procedure that eliminated effects of census
variance on competition coefficients. They developed a revised standardized regres-
sion technique to estimate competition coefficients. Competition coefficients were
estimated following the methods established by Fox and Luo (1996) and Ritchie et
al. (2009). The abundance data of each species was standardized (giving zero mean
and variance of one) before any regression analysis. We discarded a number of pre-
dictors with no significant association with deer pellet counts in order to avoid mul-
ticollinearity (Graham, 2003) by computing pairwise Pearson correlations among
all predictors themselves and each predictor and each deer species pellet groups.
Whenever a correlation exceeded 0.5, the variable with lower biological meaning
was dropped. The predictors which showed the lower correlation coefficient with
deer abundance were then related with the normal standardized abundance data
of each species (response variable) in a multiple regression-like approach, using a
Generalized Linear Mixed Model (GLMM) with a Poisson distribution error and
log-link function. Transects were introduced in the model as a random factor, to
deal with the greater degree of dependence of segments within them, arising from
nested data structures and to avoid pseudoreplication (Hurlbert, 1984). The compet-
itive coefficients of both deer species were estimated as the standardized coefficient
estimates for each appropriate variable, based on the standardized abundance data.
In order to select the best model, we evaluated the parsimony relative to predic-
tive efficiency of all possible subsets of uncorrelated candidate variables using an
information theoretical-approach (Burnham and Anderson, 1998). Such procedure
compares the suitability of a set of candidate models according to their AIC val-
ues. In this framework, we generated models and they were ranked according to
AIC values, where model with the lowest AIC is the best one. We also reported the
AIC value in order to compare the difference between each model and the best
model. As a rule of thumb, a AIC < 2 suggests substantial evidence for the model
(and consequently for the variables included), values between 3 and 7 indicate that
the model has considerably less support, whereas a AIC > 10 indicates that the
model is very unlikely (Burnham and Anderson, 1998). The level of significance
was set at 0.05 for all statistical tests. GLMM models were fitted using GLIMMIX
macro-implement in the SAS System for Windows 9.2. Pearson’s correlation was
assessed using the CORR procedure of the SAS System.
 Animal Biology (2012) DOI 10.1163/157075612X631213 9

Table 2.
Variables used in the first-step GLM testing the factors affecting roe
(Torres et al., 2011) and red deer relative abundance in sampling plots,
in northeast of Portugal, Trás-os-Montes, and the levels of significance
at the Pearson correlations between each explanatory variable and the
number of pellet groups in the segments (n.s. = not significant).

Variable Red deer Roe deer

Significance Significance

PROP_1 <−0.0001 n.s.

PROP_2 0.0015 n.s.
SHRCOV <0.0001 <0.0001
ERICOV <0.0001 n.s.
THORNCOV n.s. 0.097
LEGCOV 0.0004 n.s.
HALCOV n.s. 0.065
GRNCOV −0.0058 n.s.
POACOV −0.0015 n.s.
AGRICOV <−0.0001 n.s.
DIST_RIV −0.0155 n.s.
DIST_ROAD n.s. 0.034
MSI_2 0.0158 −0.007
SDI_2 −0.0128 *

* Not measured in previous work (Torres et al., 2011).

Results
Red deer
Of the initial set of 34 environmental factors that could explain red deer distribution,
the following factors were dropped from the analysis in step 1 owing to excessive
pairwise correlations: micro-habitat (TREECOV, MEADCOV, PINCOV, PYRCOV,
DECCOV, ILEXCOV, SATCOV, CISTCOV, THORNCOV, PTERICOV, HALCOV,
CANCOV, VISINDEX), human disturbance (DIST_ROAD, DIST_POP), and land-
scape structure (NP_1, NP_2, ED_1, ED_2, MSI_1, SDI_1, SEI_1, SEI_2). Before
performing the multivariate analysis, the mean values of a total of 11 variables were
related to the response variable by univariate tests (table 2). Variables selected in
step 2 were integrated into a final model. Of the initial set of 34 variables, the final
model retained only 3 variables: ROE_PRES; SHRUBCOV and MSI_2 (table 3, i).
The form of the relationships between red deer distribution and the environmental
factors varied with spatial scale and the final model incorporated both local scale
and landscape scale factors. At the largest scale, the key influence on red deer distri-
bution was spatial heterogeneity, where patches more irregular in shape were used
more often. At the smallest scale, high shrub cover was an important predictor in
red deer distribution. Our results show that across the north of Portugal, red deer
occurrence was positively correlated with areas with high cover of shrubs (GLMz
10 R.T. Torres et al. / Animal Biology (2012)

Table 3.
Number of model parameters (k) and the Akaike Information Criterion (AIC) de-
scribing roe and red deer relative abundance in sampling plots, in northeast of
Portugal, Trás-os-Montes (2007-2009).

Models k AIC

Red deer
i) Alternative models
ROE_PRES + SHRUBCOV + MSI_2 3 3918.2
Roe deer
ii) Alternative models
RED_PRES + SHRUBCOV + DIST_ROAD + MSI_2 4 1659.0

For explanation of variables see table 1.

estimate parameter: c), areas where roe deer is present (GLMz estimate parame-
ter: 1.1015, t = 3.58, P = 0.0004), and areas where patches were more irregular in
shape (GLMz estimate parameter: 3.1909, t = 2.64, P = 0.0084) (table 4).

Roe deer
Univariate analyses are fully described in Torres et al. (2011). Our results show
that in our study area, at a local scale, roe deer occurrence was positively corre-
lated with areas with high density of shrubs (GLMz estimate parameter: 0.006,
t = 0.25, P = 0.08), distance to roads (GLMz estimate parameter: 0.0003, t = 4.52,
P < 0.0001) and negatively correlated with mean shape index (GLMz estimate pa-
rameter: −3.692, t = −4.62, P < 0.0001) and with the presence of red deer (GLMz
estimate parameter: −0.255, t = −3.98, P < 0.0001) (table 4).

Table 4.
Variables retained in the final model using a Generalized Linear Mixed Model (GLMM), explaining
red deer and roe deer distribution in northeast of Portugal, Trás-os-Montes, with the levels of signifi-
cance. The estimate value for the standardized coefficient, and the standard error, of each standardized
variable included in the model is shown including what is interpreted as the competition coefficient
for each species.

Red deer Roe deer

Estimated coefficient ± SE Estimated coefficient ± SE

INTERCEPT −2.590 ± 1.217 INTERCEPT 6.943 ± 1.561

SHRCOV 0.425 ± 0.116 SHRUBCOV 0.006 ± 0.024
MSI_2 3.191 ± 1.206 DIST_ROAD <0.001± < 0.001
ROE_PRES 1.102 ± 0.308 MSI_2 −0.369 ± 0.798
RED_PRES −0.255 ± 0.006
 Animal Biology (2012) DOI 10.1163/157075612X631213 11

Estimated competition coefficient

The competitive coefficient of roe deer on red deer was 1.102 while the competition
coefficient of red deer on roe deer was −0.255 (table 4). This difference suggests an
asymmetry in interspecific competition between these two deer species, indicating
that red deer did not suffer from a negative effect of roe deer presence, but that red
deer presence was detrimental to roe deer presence.

Discussion
Species’ habitat preferences and biotic interactions are tightly linked, shaping both
species distribution and community dynamics (Ritchie et al., 2009). To our knowl-
edge, this is the first study analysing interspecific interaction between sympatric
roe and red deer, at different spatial scales, in Portugal, a Mediterranean landscape,
where both species occupy the limits of their ecological envelope, live at low den-
sities and are the main prey species of the endangered Iberian wolf (Vingada et al.,
1991). Several parameters affected the occurrence of both roe (see Torres et al.,
2011) and red deer, but in somewhat different ways at the different scales.
Competition between large mammals is generally difficult to measure mainly
because of the logistic complexity of conducting experiments or manipulating pop-
ulations and also because predation is regarded as playing a major role in limiting
large herbivore populations (Mishra et al., 2004). However, the existence of a poten-
tial for competition, does not imply that competition actually occurs. Niche overlap
may be tolerated without competition if resources are not limiting. In fact, overlap
in fundamental niche implies a potential for competition (if resources are limiting),
while overlap in realized niche implies that competition may be occurring (Putman,
1996). Our results are evidence of overlap or separation of resources in sympatry.
However, overlap in fundamental niche does not necessarily result in overlap in
realized niche. In this case, competition can be avoided by niche separation. The
fundamental niche of each species can only be resolved in situations of the same
habitat availability, but where the species occur in allopatry future studies should
focus on exploring situations of allopatry versus sympatry.
Ritchie et al. (2009) suggested that the use of competition coefficients is a good
estimate in the study of competition among herbivore populations. The statistical
technique that we used in this study is a modification of that used by Fox and Luo
(1996) and Luo et al. (1998) and serves more as an exploration of interaction than
of formal competition, but it does serve as a starting point for exploring their rela-
tionship.
Consistent with predictions, our results suggest that the distribution of these sym-
patric deer is not only shaped by environmental variables but also by each others
presence. Although the nature of our evidence is only correlative, the strength of
the negative association between roe and red deer is consistent with asymmetric
competition (Ritchie et al., 2009). We observed a negative effect of red deer on roe
deer distribution suggesting that roe deer are vulnerable to competition, which is
12 R.T. Torres et al. / Animal Biology (2012)

consistent with previous studies (Danilkin, 1996; Latham et al., 1996; San José et
al., 1997; Richard et al., 2010). Possible mechanisms of this alleged competition
can be that the diet of red deer can also include plants eaten by roe deer (dietary
overlap), suggesting a potential for exploitation competition (as has been showed
in other areas, e.g., Latham et al., 1999).
On the other hand, roe deer may be intolerant to the proximity of bigger species
(habitat overlap) (e.g., Ferretti et al., 2008, 2011, for roe/fallow deer behavioural
interactions) and their density appeared to be negatively affected by that of other
deer species (e.g., by red deer: Latham et al., 1996; by fallow deer: Focardi et al.,
2006; and muntjac: Hemami et al., 2004; Ferretti et al., 2011). Nevertheless, Melis
et al. (2009) showed that competition with red deer did not appear to play a limiting
role for roe deer abundance. By contrast to the negative effect of red deer pres-
ence on roe deer density, our results suggest that the main effect is unidirectional:
red deer presence itself may be positively correlated with roe deer presence. How-
ever, this might be due to the fact that red deer are evenly distributed throughout
the area and the presence of roe deer does not seem to have any displacement ef-
fect on red deer. Red deer occurred more often in patches with higher densities of
shrub cover, areas with higher spatial heterogeneity and areas where roe deer were
present. Red deer are classified as an intermediate feeder (Hofmann, 1989), which
adapts to seasonal changes in the quantity and quality of available food by feeding
alternatively on grasses and shrubs. Therefore, red deer seem to fulfill their nutri-
tional requirements in the shrub layer. Additionally, such use can also be regarded
as an anti-predator response. Our results show that red deer make use of areas with
less spatial heterogeneity, as their occurrence was negatively correlated with mean
shape index (MSI). MSI needs careful interpretation as its meaning can switch from
an index of complexity to an index of compactness. Some factors can decrease land-
scape complexity such as landscape elements in agricultural landscapes (rectangles
with straight and distinct boundaries) and decrease of natural or semi-natural areas
(Mander et al., 1999). Roe deer model results are fully described in Torres et al.
(2011).
Animals of smaller body size are more selective in their food uptake than larger
species (Gordon and Illius, 1996). Our results point in this direction, since roe deer
use more patches with richer food resources, which are, in the study area, the pre-
ferred food items (Faria, 1999).
Pellet group counts is a technique widely applied in studies of ungulate habitat
use (e.g., Neff, 1968; Latham et al., 1996; Telléria and Virgós, 1997; Borkowski and
Ukalska, 2008; Torres et al., 2011) and provides a valid approach to allow an initial
coarse scale assessment of habitat associations. Although it has been criticized by
some authors (Collins and Urness, 1981), when it is compared with other methods,
such as radio-telemetry, to infer habitat use patterns, it has been found that data are
similar (Guillet et al., 1995) and other authors have found that pellet group counts
accurately indicate which habitats receive the greatest and least amount of use (Loft
and Kie, 1988; Edge and Marcum, 1989; Månsson et al., 2011). Moreover, other
 Animal Biology (2012) DOI 10.1163/157075612X631213 13

authors have reported the use of pellet group counts in the Iberian Peninsula (e.g.,
in Portugal for roe deer – Torres et al., 2011; in Spain for roe deer – Telléria and
Virgós, 1997; Virgós and Telléria, 1998 and for red deer – San José et al., 1997).
Despite the difficulty in analysing the interaction between two species with dif-
ferent spatial behaviours and life history traits, the results of this study reinforce
the need for further studies. Our work identifies key biological and environmental
factors that affect the distribution of these ecologically, economically and cultur-
ally significant species, in the south-western part of their distribution range. Future
research should compare competitive coefficients across different points of the dis-
tribution range of these species. Our findings not only add a new understanding of
the influence of interspecific competition in roe and red deer in a Mediterranean
area, but also have implications for the conservation of roe deer, a species in the
initial stages of recovery in Portugal. Previous studies (see above) suggest that if
red deer continue to increase, roe deer might be replaced by the previous. Conse-
quently, management actions in the study area (Trás-os-Montes) should be directed
towards protecting and improving large patches of heterogeneous scrubland habitats
to benefit roe deer (Torres et al., 2011). Furthermore, roe and red deer are important
prey species of the Iberian wolf (Vingada et al., 2010) and since this carnivore has
recently re-colonized this Mediterranean ecosystem, detailed models of ungulate
abundance and habitat use are required. More information is needed to understand
the mechanisms of interspecific relationships that act on these two species on a
Mediterranean ecosystem.

Acknowledgements
We are grateful to Agostinho Gomes, Francisco Gonçalves, Jan van Mil, Manuel
Borges and Ricardo Ramalho for the valuable assistance in the field. Likewise, sev-
eral institutions provided invaluable support: Parque Natural de Montesinho, Insti-
tuto da Conservação da Natureza e Biodiversidade, Autoridade Florestal Nacional
and the Serviços Florestais da Direcção Regional de Agricultura de Trás-os-Montes,
especially Núcleo Florestal de Bragança. We would also like to express our thanks
to Professor Rory Putman whose constructive comments improved the quality of
this article. Financial support was provided by Fundação para Ciência e Tecnologia
(SFRH/BD/28310/2006). The involvement of JDCL was funded by the Research
Council of Norway and the Norwegian Directorate for Nature Management.

References
Acevedo, P., Cassinello, J., Hortal, J. & Gortázar, C. (2007) Invasive exotic aoudad (Ammotragus
lervia) as a major threat to native Iberian ibex (Capra pyrenaica): a habitat suitability model ap-
proach. Diversity and Distrib., 13, 587-597.
Acevedo, P., Ward, A.I., Real, R. & Smith, G.C. (2010) Assessing biogeographical relationships of
ecologically related species using favourability functions: a case study on British deer. Diversity
and Distrib., 16, 515-528.
14 R.T. Torres et al. / Animal Biology (2012)

Andersen, R., Duncan, P. & Linnell, J.D.C. (1998) The European Roe Deer: The Biology of Success.
Scandinavian University Press, Oslo, Norway.
Apollonio, M., Andersen, R. & Putman, R. (2010) European Ungulates and Their Management in the
21st Century. Cambridge University Press, Cambridge.
Borkowski, J. & Ukalska, J. (2008) Winter habitat use by red and roe deer in pine-dominated forest.
For. Ecol. Manag., 255, 468-475.
Brown, J.H. & Maurer, B.A. (1986) Body size, ecological dominance and Cope’s rule. Nature, 324,
248-250.
Burnham, K.P. & Anderson, D.R. (1998) Model Selection and Inference: A Practical Information-
Theoretic Approach. Springer-Verlag, New York, USA.
Buskirk, J.V. (2005) Local and landscape influence on amphibian occurrence and abundance. Ecology,
86, 1936-1947.
Carvalho, P., Nogueira, A.J.A., Soares, A.M.V.M. & Fonseca, C. (2008) Ranging behaviour of translo-
cated roe deer in a Mediterranean habitat: seasonal and altitudinal influences on home range size
and patterns of range use. Mammalia, 72, 89-94.
Clutton-Brock, T., Guiness, F.E. & Albon, S.D. (1982) Red Deer: Behavior and Ecology of Two Sexes.
University Chicago Press, Chicago.
Collins, W.B. & Urness, P.J. (1981) Habitat preferences of mule deer as rated by pellet-group distri-
butions. J. Wildl. Manag., 45, 969-972.
Danilkin, A.A. (1996) Behavioural Ecology of Siberian and European Roe Deer. Chapman & Hall,
London.
Dunning, J.B., Danielson, B.J. & Pulliam, H.R. (1992) Ecological processes that affect populations in
complex landscapes. Oikos, 65, 169-175.
Edge, W.D. & Marcum, C.L. (1989) Determining elk distribution with pellet-group and telemetry
techniques. J. Wildl. Manag., 53, 621-624.
Elkie, P.C., Rempel, R.S. & Carr, A.P. (1999) Patch analyst users manual: a tool for quantifying
landscape structure. NWST technical manual, Ontario.
Faria, A.M.S. (1999) Dieta de corço (Capreolus capreolus L.) no Centro e Nordeste de Portugal.
M.Sc. Thesis. University of Coimbra, Coimbra.
Ferretti, F., Sforzi, A. & Lovari, S. (2008) Intolerance amongst deer species at feeding: roe deer are
uneasy banqueters. Behav. Process., 78, 487-491.
Ferretti, F., Sforzi, A. & Lovari, S. (2011) Behavioural interference between ungulate species: roe are
not on velvet with fallow deer. Behav. Ecol. Sociobiol., 65, 875-887.
Focardi, S., Aragno, P., Montanaro, P. & Riga, F. (2006) Inter-specific competition from fallow deer
Dama dama reduces habitat quality for the italian roe deer Capreolus capreolus italicus. Ecogra-
phy, 29, 407-417.
Fox, B.J. & Luo, J. (1996) Estimating competition coefficients from census data: a re-examination of
the regression technique. Oikos, 77, 291-300.
Gordon, I.J. & Illius, A.W. (1996) The nutritional ecology of African ruminants: a reinterpretation.
J. Anim. Ecol., 65, 18-28.
Graham, M.H. (2003) Confronting multicollinearity in ecological multiple regression. Ecology, 84,
2809-2815.
Guillet, C., Bergström, R., Cederlund, G., Bergström, J. & Ballon, P. (1995) Comparison of telemetry
and pellet-group counts for determining habitat selectivity by roe deer (Capreolus capreolus) in
winter. Gibier Faune Sauvage, 12, 253-269.
Guisan, A. & Thuiller, W. (2005) Predicting species distribution: offering more than simple habitat
models. Ecol. Lett., 8, 993-1009.
 Animal Biology (2012) DOI 10.1163/157075612X631213 15

Hemami, M.R., Watkinson, A.R. & Dolman, P.M. (2004) Habitat selection by sympatric muntjac
(Muntiacus reevesi) and roe deer (Capreolus capreolus) in a lowland commercial pine forest. For.
Ecol. Manag., 194, 49-60.
Hoffmann, A.A. & Blows, M.W. (1994) Species borders: ecological and evolutionary perspectives.
Trends Ecol. Evol., 9, 223-227.
Hofmann, R.R. (1989) Evolutionary steps of ecophysiological adaptation and diversification of rumi-
nants: A comparative view of their digestive system. Oecologia, 78, 443-457.
Hortal, J., Borges, P.A.V. & Gaspar, C. (2006) Evaluating the performance of species richness estima-
tors: sensitivity to sample grain size. J. Anim. Ecol., 75, 274-287.
Hurlbert, S.H. (1984) Pseudoreplication and the design of ecological field experiments. Ecol. Monogr.,
54, 187-211.
Latham, J., Staines, B.W. & Gorman, M.L. (1996) The relative densities of red (Cervus elaphus) and
roe (Capreolus capreolus) deer and their relationship is Scottish plantation forests. J. Zool., 240,
285-299.
Latham, J., Staines, B.W. & Gorman, M.L. (1997) Correlations of red (Cervus elaphus) and roe
(Capreolus capreolus) deer densities in Scottish forests with environmental variables. J. Zool.,
242, 681-704.
Latham, J., Staines, B.W. & Gorman, M.L. (1999) Comparative feeding ecology of red (Cervus ela-
phus) and roe deer (Capreolus capreolus) in Scottish plantation forests. J. Zool., 247, 409-418.
Loft, E.R. & Kie, J.G. (1988) Comparison of pellet-group and radio triangulation methods for assess-
ing deer habitat use. J. Wild. Manag., 52, 524-527.
Luo, J., Monamy, V. & Fox, B.J. (1998) Competition between two australian rodent species: a regres-
sion analysis. J. Mammal., 79, 962-971.
MacArthur, R. & Levins, R. (1967) The limiting similarity, convergence, and divergence of coexisting
species. Am Nat., 101, 377-385.
Mander, Ü., Mikk, M. & Külvik, M. (1999) Ecological and low intensity agriculture as contributors
to landscape and biological diversity. Landsc. Urban. Plan., 46, 169-177.
Marques, F.F.C., Buckland, S.T., Goffin, D., Dixon, C.E., Borchers, D.L., Mayle, B.A. & Peace, A.J.
(2001) Estimating deer abundance from line transect surveys of dung: sika deer in southern Scot-
land. J. Appl. Ecol., 38, 349-363.
Mayle, B.A., Putman, R.J. & Wyllie, I. (2000) The use of trackway counts to establish an index of
deer presence. Mamm. Rev., 30, 233-237.
McGarigal, K. & Marks, B. (1995) FRAGSTATS: spatial pattern analysis program for quantifying
landscape structure. U.S. Forest Service General Technical Report.
Melis, C., J˛edrzejewska, B., Apollonio, M., Bartoń, K.A., J˛edrzejewski, W., Linnell, J.D.C., Kojola,
I., Kusak, J., Adamic, M., Ciuti, S., Delehan, I., Dykyy, I., Krapinec, K., Mattioli, L., Sagaydak,
A., Samchuk, N., Schmidt, K., Shkvyrya, M., Sidorovich, V.E., Zawadzka, B. & Zhyla, S. (2009)
Predation has a greater impact in less productive environments: variation in roe deer, Capreolus
capreolus, population density across Europe. Glob. Ecol. Biogeogr., 18, 724-734.
Mishra, C., Van Wieren, S.E., Ketner, P., Heitkönig, I.M.A. & Prins, H.H.T. (2004) Competition be-
tween domestic livestock and wild bharal Pseudois nayaur in the Indian Trans-Himalaya. J. Appl.
Ecol., 41, 344-354.
Mitchell-Jones, A.J., Amori, G., Bogdanowicz, W., Krystufek, B., Reijnders, P.H.J., Spitzenberger, F.,
Stubbe, M., Thissen, J.B.M., Vohralik, V. & Zima, J. (1999) The Atlas of European Mammals. T. &
A.D. Poyser, London.
Mysterud, A. (1996) Bed-site selection by adult roe deer Capreolus capreolus in southern Norway
during summer. Wildl. Biol., 2, 101-106.
16 R.T. Torres et al. / Animal Biology (2012)

Neff, D.J. (1968) The pellet-group count technique for big game trend, census, and distribution: a re-
view. J. Wildl. Manag., 32, 597-614.
Patthey, P. (2003) Habitat and corridor selection of an expanding red deer (Cervus elaphus) popula-
tion. Ph.D. Thesis. University of Lausanne, Lausanne.
Prins, H.H.T., De Boer, W.F., Van Oeveren, H., Correia, A., Mafuca, J. & Olff, H. (2006) Co-existence
and niche segregation of three small bovid species in southern Mozambique. Afr. J. Ecol., 44, 186-
198.
Putman, R.J. (1996) Competition and Resource Partitioning in Temperate Ungulate Assemblies. Chap-
man Hall, London.
Richard, E., Gaillard, J.-M., Saïd, S., Hamann, J.-L. & Klein, F. (2010) High red deer density depresses
body mass of roe deer fawns. Oecologia, 163, 91-97.
Ritchie, E.G., Martin, J.K., Johnson, C.N. & Fox, B.J. (2009) Separating the influences of environ-
ment and species interactions on patterns of distribution and abundance: competition between large
herbivores. J. Anim. Ecol., 78, 724-731.
Rotenberry, J.T., Preston, K.L. & Knick, S.T. (2006) GIS-based niche modeling for mapping species’
habitat. Ecology, 87, 1458-1464.
Saïd, S. & Servanty, S. (2005) The influence of landscape structure on female roe deer home-range
size. Lands Ecol., 20, 1003-1012.
San José, C., Braza, F., Aragón, S. & Delibes, J.R. (1997) Habitat use by roe and red deer in Southern
Spain. Mis. Zool., 20, 27-38.
Santos, J.P.V. (2009) Estudo populacional do veado (Cervus elaphus L.) no Nordeste Transmontano.
M.Sc. Thesis. Universidade de Aveiro, Aveiro.
Schoener, T.W. (1974) Some methods for calculating competition coefficients from resource-
utilization spectra. Am. Nat., 108, 332-340.
Torres, R.T., Santos, J., Linnell, J.D.C., Virgós, E. & Fonseca, C. (2011) Factors affecting roe deer
occurrence in a Mediterranean landscape, Northeastern Portugal. Mamm. Biol., 76, 491-497.
Tellería, J.L. & Virgós, E. (1997) Distribution of an increasing roe deer population in a fragmented
Mediterranean landscape. Ecography, 20, 247-252.
Vingada, J.V. (1991) Estudo da população de veados (Cervus elaphus) da Lombada (Bragança).
Bachelor Thesis. Universidade do Porto, Porto.
Vingada, J., Fonseca, C., Cancela, J., Ferreira, J. & Eira, C. (2010) Ungulates and their management
in Portugal. In: M. Apollonio, R. Andersen & R.J. Putman (Eds.) European Ungulates and Their
Management in the 21st Century. Cambridge University Press, Cambridge.
Virgós, E. & Telléria, J.L. (1998) Roe deer habitat selection in Spain: constraints on the distribution
of a species. Can. J. Zool., 76, 1294-1299.