Landscape of fear in Europe: wolves affect spatial patterns of
ungulate browsing in Białowiez
.a Primeval Forest, Poland
D. P. J. Kuijper, C. de Kleine, M. Churski, P. van Hooft, J. Bubnicki and B. Je˛drzejewska
D. P. J. Kuijper (firstname.lastname@example.org), M. Churski, J. Bubnicki and B. Jędrzejewska, Mammal Research Inst., Polish Academy of
Sciences, ul. Waszkiewicza 1, PL-17-230 Białowieża, Poland. – C. de Kleine and P. van Hooft, Wageningen Univ., Resource Ecology Group,
Droevendaalsesteeg 3a, NL-6708PB, Wageningen, the Netherlands.
Large carnivores can either directly inﬂuence ungulate populations or indirectly aﬀect their behaviour. Knowledge from
European systems, in contrast to North American systems, on how this might lead to cascading eﬀects on lower trophic
levels is virtually absent. We studied whether wolves Canis lupus via density-mediated and behaviorally-mediated eﬀects
on their ungulate prey species inﬂuence patterns of browsing and tree regeneration inside the Białowieża National Park,
Poland. Browsing intensity of tree saplings (height class 150 cm), irrespective of tree species or forest type, was lower
inside a wolf core area (50.5%) where predator presence is highest, than in the remainder of the wolf pack’s home range
(58.3%). Additionally, browsing intensity was reduced when the amount of coarse woody debris (CWD), which can
act as a ‘ungulate escape impediment’, increased (within 5-m radius) inside the wolf core area. No relationship existed
outside the core area. As a result, the proportion of trees growing out of herbivore control increased more strongly with
increasing amount of CWD inside compared to outside the wolf core area. is suggests that next to direct eﬀects of
wolves on ungulate density caused by a higher predation pressure inside the core area, risk eﬀects are important and
are enhanced by habitat characteristics. ese results indicate that behaviorally-mediated eﬀects of predators on prey
can become more important than density-mediated eﬀects in aﬀecting lower trophic levels. is is the ﬁrst study we
are aware of, that shows CWD can create ﬁne-scale risk eﬀects on ungulates with the potential for cascading eﬀects of
large predators on patterns of tree regeneration for a European forest system. is knowledge broadens the discussion
on how the impact of large predators on ecosystem functioning depends on the physical landscape, by illustrating these
eﬀects for a system which largely contrasts in this respect to the North American systems.
Large carnivores can play an important role in structuring
ungulate communities with cascading eﬀects on lower
trophic levels (Terborgh and Estes 2010). In the classical
view carnivores directly modify these relationships by top-
down regulating herbivore populations and releasing plants
from herbivore control (Oksanen et al. 1981, Fretwell
1987, DeAngelis 1992). For several temperate forest sys-
tems these density-mediated eﬀects of carnivores on the
ungulate community have been illustrated both inside
(Jędrzejewski et al. 2002, Jędrzejewska and Jędrzejewski
2005) and outside Europe (McLaren and Peterson 1994,
Messier 1994, Ripple and Beschta 2005). Recently, there is
an increasing recognition that indirect, non-lethal predator
eﬀects are also important or even more important than
their direct lethal eﬀects (Schmitz et al. 1997, Creel and
Predators may indirectly inﬂuence ungulates by chang-
ing their distribution towards less risky habitat types (Lima
and Dill 1990, Kie 1999, Creel et al. 2005, Mao et al. 2005,
aker et al. 2011), change their movement patterns (Fortin
et al. 2005, Fischhoﬀ et al. 2007), or shift their activity
towards less risky times (Creel et al. 2008, Valeix et al.
2009). Besides or instead of changing their spatio-temporal
distribution, animals can reduce predation risk by changing
their behaviour. A common tactic for ungulates is to increase
vigilance levels while foraging (Pulliam 1973, Hunter and
Skinner 1998, Brown et al. 1999). Alternatively, group size
may increase to reduce individual predation risk due to the
dilution and confusion eﬀect (Bertram 1978, Dehn 1990).
An increase in group size may on its turn decrease individ-
ual vigilance time (Pulliam 1973, Bertram 1978, Pulliam
et al. 1982, McNamara and Houston 1992), and may
increase individual foraging eﬃciency (Lipetz and Bekoﬀ
1982, LaGory 1986, Lima and Dill 1990). Both avoidance
of high risk areas and behavioral changes can mediate brows-
ing pressure on plants growing in these areas (Ripple et al.
2001). Browsing should be reduced when ungulates avoid
high risk areas or increase their vigilance levels at the expense
of foraging. Alternatively, in case group size increases in
response to predation pressure, coinciding with a reduction
of individual vigilance level, browsing pressure may poten-
tially increase in high risk areas.
Predation risk eﬀects have played an important role
in explaining patterns of tree recruitment in Yellowstone
Ecography 36: 1263–1275, 2013
© 2013 e Authors. Ecography © 2013 Nordic Society Oikos
Subject Editor: Jean-Michel Gaillard. Accepted 3 April 2013
National Park (YNP) after the reintroduction of wolves
Canis lupus. Feeding behavior of elk Cervus elaphus
canadensis changed when wolves reappeared in the system
(Laundré et al. 2001, Ripple and Beschta 2003, Beyer
et al. 2007). Habitats with a high amount of escape impedi-
ments (objects which obstruct deer escape from a predator)
or low visibility are regarded to have higher perceived
predation risk and tend to be avoided by elk (Ripple and
Beschta 2006, Halofsky and Ripple 2008). Trees growing in
these high risk areas may experience lower herbivore top-
down eﬀects. Ripple et al. (2001) found evidence for that by
showing that sapling height of quaking aspen Populus
tremuloides was higher and elk dropping density lower inside
high wolf-use areas, suggesting an avoidance of deer of these
areas. Later studies, including other tree species, found simi-
lar indirect evidence for behaviorally-induced changes in
foraging patterns, with lower browsing and taller tree height
inside habitats with low visibility and/or the presence of
escape impediments (Ripple and Beschta 2003, 2006).
However, there is still an ongoing debate on how important
these indirect risk (non-lethal) eﬀects are relative to direct
(lethal) eﬀects of predators on their prey in explaining
patterns of tree regeneration (Creel and Christianson 2009,
Kauﬀman et al. 2010, Beschta and Ripple 2011). Moreover,
Winnie (2012) recently argued that there is currently little
undisputed evidence for the occurrence of ﬁne-scale risk
eﬀects created by escape impediments leading to cascading
eﬀects on tree recruitment inside YNP.
In contrast to the large body of literature from North
American systems, knowledge on whether large carnivores
(in)directly inﬂuence vegetation via eﬀects on ungulate prey
are virtually absent from other systems, and particularly
European systems. Obvious reasons for this gap of knowl-
edge is the lack of natural predators in many areas and
the scarcity of areas with undisturbed forest development.
e strictly protected parts of the Białowieża Primeval
Forest (BPF) in Poland oﬀer one of the rare examples in
Europe where these tri-trophic interactions can be studied
(Jędrzejewska and Jędrzejewski 2005). is area is unique as
it harbours the natural European assemblage of ungulate
species (European bison Bison bonasus, moose Alces alces, red
deer Cervus elaphus, roe deer Capreolus capreolus, wild boar
Sus scrofa) which co-occur with their natural predators (wolf,
lynx Lynx lynx). An additional unique feature is that it
belongs to one of the last remaining natural, temperate, low-
land forest systems in Europe where in part of the area
(Białowieża National Park, BNP) human intervention (hunt-
ing, forestry) has been excluded. e question is whether
knowledge from North American systems is directly appli-
cable to other systems which have largely contrasting
landscapes. e landscape of fear in YNP is strongly deter-
mined by the physical landscape with strongest risk eﬀects
occurring in, for example, large river valleys and mountain
ridges (Kauﬀman et al. 2007). e physical landscape of
Białowieża Primeval forest diﬀers largely from the well-
studied North American systems in at least two factors that
might aﬀect predator–prey interactions. Firstly, the area
lacks large river valleys, mountains and is mainly composed
of lowland forest as only 5% of the land consists of
open grassland. Secondly, the area is small (Polish part of
Białowieża Primeval Forest: 600 km2, Białowieża National
Park: 105 km2 ) compared to YNP (8980 km2). As a result,
there are virtually no parts of this area where wolves and
lynxes are absent, simply because their home ranges cover
the entire forest complex (Schmidt et al. 2009). erefore,
ungulate prey cannot reduce predation risk by moving to
predator-free areas as has been illustrated for the YNP (Creel
et al. 2005, Fortin et al. 2005, Mao et al. 2005, Kauﬀman
et al. 2007). However, there are gradients in risk related to
the frequency of large carnivore presence and ungulate prey
may shift from high-risk habitats towards low-risk habitats.
For wolves, core areas of their pack territories can be deﬁned
where on a yearly basis 50% or more of all observations
of radio-collared individuals occurred (Jędrzejewski et al.
2007, Schmidt et al. 2009). During the reproductive season
(spring–summer) their spatial distribution is restricted
within this core area whereas outside this period they
regularly return to the core area (Jędrzejewski et al. 2001).
e average size of a wolf pack territory in BPF is 201 km2
with a core area of 35 km2 (Jędrzejewski et al. 2007). Taken
into account the average kill rate of their main prey species,
on a yearly basis one wolf pack kills 118 red deer per year
within its territory (Jędrzejewski et al. 2002), of which
half would be predated inside the core area based on the
amount of time spent by the wolf pack there. is would
result in 1.7 red deer km22 (59 deer in 35 km2) taken per
year inside the core, versus 0.36 red deer km22 (59 deer in
166 km2) taken in the rest of their territory. Due to this
5-fold higher predation pressure, wolves are predicted to
directly aﬀect red deer density leading to a lower density
inside compared to outside the wolf core area. Moreover,
because of more frequent predator presence and cues that
result from this, ungulates may perceive these areas as risky
and avoid them or change their behavior. Core areas of wolf
territories are most intensively scent-marked by means of
urination, scats and territorial scratchings (Zub et al. 2003)
and also howling activity is concentrated inside core areas
(Nowak et al. 2007). ese olfactory and acoustic cues,
next to visual cues, could be used by ungulates. Perceived
predation risk for red deer is predicted to be higher inside
wolf core areas resulting in higher vigilance levels at the
cost of time spent foraging. ese predator-mediated direct
and indirect changes may aﬀect spatial patterns in herbivore
top-down eﬀects by creating areas with reduced browsing
intensity of trees.
e present study was aimed at testing whether these
direct and indirect eﬀects of wolves on deer are visible in
browsing intensity of tree saplings and aﬀect patterns of tree
regeneration. We compared browsing intensity on tree
saplings growing inside a wolf core area with the area in the
remainder of the wolf pack’s territory. Due to a higher
predation pressure inside versus outside the wolf core area,
we predict that wolves via density-mediated eﬀects reduce
browsing intensity inside the wolf core area. To test
whether additionally behaviorally-mediated eﬀects of wolves
occurred, we related habitat characteristics associated with
perceived predation risk to browsing intensity. For this
we measured the amount of ‘escape impediments’ for deer
(sensu Halofsky and Ripple 2008) in the vicinity of each
tree sapling; coarse woody debris is the only physical
barrier which could act as such in our system (Fig. 2). When
(perceived) predation risk is additionally aﬀecting foraging
behavior of deer, the amount of escape impediments should
strengthen the density-mediated eﬀects and result in a stron-
ger reduction in browsing intensity especially inside the wolf
e Białowieża Primeval Forest (BPF, 52°45′N, 23°50′E)
situated in eastern Poland (600 km2) and western Belarus
(850 km2) is a large continuous forest composed of multi-
species tree stands. Since 1921, the best preserved central
part of the BPF (47 km2) have been strictly protected and
no human intervention (including forestry activities and
hunting) has been allowed. Human impact before this
period is considered as minimal (Jędrzejewska et al. 1997,
Samojlik et al. 2007). In 1996, the Białowieża National Park
(BNP), including the strictly protected area, was enlarged to
105.2 km2 (Fig. 1). In this enlarged area (acting as a buﬀer
zone of the strictly protected part), forestry activities are
minimized to sanitary cutting of diseased Norway spruce
Picea abies, and tree regeneration occurs naturally without
human intervention (euerkauf and Rouys 2008). Inside
the entire BNP hunting and access by motorized vehicles
have been prohibited and tourist access is only permitted
with a guide. e area outside BNP is managed by the state
forestry service, including wood exploitation and regulation
of ungulate numbers.
e present study was carried out inside the BNP, where
old-growth forest stands (81–120 yr) prevail and cover
42% of the area. Very old stands, dominated by trees aged
over 120 yr, account for 39%. Only 0.8% of the BNP
consists of open grassland (Michalczuk 2001). A mosaic of
diﬀerent forest types exists dominated by deciduous forest
(54% of the area) with Quercus robur, Tilia cordata
and Carpinus betulus as dominant species) and mixed
deciduous forest (23% with Picea abies, Quercus robur, Tilia
cordata and Carpinus betulus).
Along the small streams alder ash forest (dominated by
Alnus glutinosa and Fraxinus excelsior) occurs, whereas mixed
coniferous forest (Pinus sylvestris, Picea abies and Quercus
robur) and coniferous forest (Pinus sylvestris and Picea
abies) are found in the drier, more nutrient-poor parts (see
for more details Faliński 1986). e mean altitude of the
BNP is 158 m a.s.l. and the total altitudinal range is 23 m.
Mean annual air temperature is 6.8°C with the coldest
month in January with on average 24.7°C and the warmest
month is July with 17.8°C. Mean annual precipitation is
641 mm and snow cover lasts for an average of 92 d.
A unique feature of the BPF is that it is one of the few
areas in Europe, where the complete native assemblage
of forest ungulates still occurs (ﬁve species) together with
Figure 1. Map of the study area, the Białowieża National Park, showing the forest types (based on Michalczuk 2001) utilized most by
ungulates. Under ‘other forest types’, very wet forest types (such as swamps and bog forest) are included which have in general lower
ungulate numbers. Locations of 34 transects, 17 inside and 17 outside the wolf core area, on which saplings were measured are indicated.
e location of the core area of the annual territory of the wolf pack present is based on Jędrzejewski et al. (2007). Locations of wolf
dens and observations of wolf natural howling in the period 2000–2012 indicates continuous use of the core area.
at least 200 saplings per transect to get a proper estimate of
browsing intensity. For each tree sapling we measured its
height and estimated browsing intensity by measuring the
number of browsed branches (both old and new browsing
marks) of the 10 top branches. We measured the top branches
as they have highest chance to be browsed, and it is likely the
main factor slowing down tree growth. As we cannot distin-
guish browsing marks by the diﬀerent ungulate species, we
studied the browsing intensity on trees resulting from
the entire ungulate community. We recorded only tree sap-
lings between 10–200 cm as these are within reach of
the ungulates in the system and overlap with the preferred
height classes (50–150 cm) of red deer (Renaud et al.
2003). As red deer is the dominant browser in the system
(79% of the total number of browsing ungulate species
excluding wild boar), browsing intensity is largely inﬂuenced
by this species (see also Kuijper et al. 2010a).
e location of transects (n 34) was assigned randomly,
but stratiﬁed within forest types. We selected four forest
types which together cover 90% of the study area
(Michalczuk 2001): deciduous forest (n 12), wet decidu-
ous forest (n 10), mixed deciduous forest (n 6), mixed
coniferous forest (n 6). We divided the 34 transects
equally over the area inside and outside the wolf core area,
resulting in total in 17 transects inside and 17 outside
the wolf core area with equal numbers within each forest
type (Fig. 1). e number of transects within each forest
type is related to the areal extent of each type, with
most transects in the most common forest type. ese forest
types compose the prime habitat for ungulates (especially
red deer) in the forest (Jędrzejewska et al. 1994, Kamler
et al. 2008). We used the existing division into forest
compartments inside the BNP to randomly assign transects
per forest type inside and outside the wolf core area. Forest
compartments are equal-sized compartment blocks of
1.15 km2 (based on former Russian system). In every forest
compartment one transect per forest type was assigned. In
some cases two transects occurred within one forest
compartment, but within diﬀerent forest types and a mini-
mum distance of 300 m between transects.
is set-up allowed us to study the eﬀect of the wolf
core area on tree browsing level within each forest type and
hence taking into account diﬀerences in forest structure
and environmental circumstances. As both wolf and deer
activity are aﬀected by human presence (euerkauf et al.
2003b, euerkauf and Rouys 2008) we established the
transects only inside the borders of the National Park which
prohibits hunting and access by motorized vehicles and has
only restricted access for public. Besides, transects were
established in all cardinal directions from the core area to
exclude any gradients in human activity present in the study
area (Fig. 1).
Trees growing out of reach of ungulate browsing
inside and outside wolf core area
Within every transect, the number of trees per species
that had grown out of reach of browsing ungulates were
counted, measuring between 200 and 400 cm in height.
Trees above 200 cm are not browsed on the leader shoot
their natural predators. e most abundant species, both in
numbers and crude biomass (Jędrzejewska et al. 1997), is red
deer with a winter density evaluated at about 12 individuals
km22 during the most recent survey based on drive counts in
January 2010 inside the BNP (T. Borowik pers. comm.). e
second-most numerous ungulate is wild boar with a den-
sity evaluated at about 10 individuals km22 in 2010. Roe
deer were present at a density evaluated at about 2 indi-
viduals km22, whereas the larger species European bison
and moose occur in the lowest densities, evaluated at about
0.8 individuals km22 and 0.4 individuals km22 respec-
tively in the winter of 2010 (T. Borowik pers. comm.).
ese relative densities of ungulate species based on drive
counts are conﬁrmed by other census techniques carried
out during the last decades (Jędrzejewska et al. 1997).
Natural predators, wolf and lynx, are strictly protected
in BPF since 1989 and are not hunted throughout
Poland. In BPF, they occur with average densities around
1–5 individuals 100 km22 (wolf) and 1–3 individuals
100 km22 (lynx) (Schmidt et al. 2008).
Location of the wolf core area
Accurate estimates of size and core areas of wolf pack terri-
tories originate from earlier studies on collared wolves
(Jędrzejewski et al. 2007, Schmidt et al. 2009) which are
continuously being updated by new surveys. e present
four wolf pack territories inside the entire BPF cover the
entire forested area leaving no vacant areas (Jędrzejewski
et al. 2007, Schmidt et al. 2009). Likely as a result of this,
the location and size of wolf pack territories and core areas
in the BPF has been very stable (Jędrzejewski et al. 2007,
Schmidt et al. 2009) with dens located throughout the
years inside the same core areas (euerkauf et al. 2003a,
Zub et al. 2003, Schmidt et al. 2008).
As the location of only one territory of a wolf pack
overlaps with the BNP, which excludes forestry activities and
prohibits hunting, this was the only one that allowed us
to study patterns of tree regeneration undisturbed by
humans. Location of the core area of this wolf pack (contain-
ing 3–8 wolves) was based on Jędrzejewski et al. (2007). In
the present study we compared the wolf core area with the
area outside it (Fig. 1), but still within the boundaries of the
annual territory of the wolf pack (95% MCP, Jędrzejewski
et al. 2007). Previous studies showed that wolves inside
the BPF howled from the central parts of their territories
where the den for breeding was also located and not from
the peripheries (Nowak et al. 2007). As an indication for
the continuous use of the core area, we selected all wolf
howling observations and known den sites from the period
2000–2012 inside the BNP collected by park rangers or
scientists working in the area (Fig. 1).
Measuring browsing intensity inside and outside the
wolf core area
In April–May 2012 we established transects 150–300 m
long and 2 m wide and recorded all tree saplings
between 10–200 cm height. e exact length of the transects
depended on the density of saplings, the aim was to measure
ungulate, and forms an obstacle that a deer should step or
jump over while escaping. e minimum length for escape
impediments of 1 m within the 1–5 m radius was chosen as
we assumed that an object of 0.5 m in length cannot be con-
sidered as a relevant obstacle when it was located further
away. A red deer could easily run around it and it blocks
only a small fraction of the view of a foraging ungulate.
Possible confounding factors affecting browsing
inside and outside wolf core area
As tree canopy gap formation is a main driving factor in tree
regeneration inside old-growth forest (Runkle 1981) and
ungulate browsing is more intense in forest gaps (Kuijper
et al. 2009) we additionally measured canopy openness
as a potential confounding factor. Canopy cover was
recorded in June (leaves fully developed) with a spherical
densiometer with a convex mirror as described by Lemmon
(1956). For this we established additional transects in June
2012 during peak leaf cover of trees, one in each of the four
forest types. We estimated canopy cover into the four wind
directions and calculated an average percentage of canopy
cover at regular locations (n 50 inside and n 75 outside
the wolf core area) with an interval of 15 m along these
transects with a minimum of 10 measurements per transect.
As tree species composition or diversity potentially
diﬀers between inside and outside wolf core area because of
(micro)habitat preferences of wolves (euerkauf et al.
2003a) we calculated Shannon diversity indices (Pielou
1975) per transect for all tree saplings available for ungulates
(10–200 cm), and averaged them for inside and outside
wolf core area, according to:
where S is the number of species and p the proportion of
species i among all tree individuals within the height class.
e minimum value of the Shannon index is 0 when there is
only one species present and the index is highest when
each species is present in equal proportions. Low values indi-
cate a high dominance of a single species, either due to
low number of species or relatively high abundance of one
Tree sapling height were log-transformed and all percentage
browsing intensity data were arcsinus transformed before
testing to meet requirements for parametric testing
We used Chi-square test to test whether tree sapling
species composition diﬀered between inside and outside
the wolf core area. We used the percentages of each species
in the entire sample of recorded tree saplings, and com-
bined all species which accounted less than 5% in a rest
group. Subsequently, we carried out a multivariate GLM
in which inside and outside wolf core area and forest
type were entered as ﬁxed factors, and the number of sap-
lings of each species recorded per transect as dependent
(Renaud et al. 2003) and hence potentially can recruit into
the tree stand. We used a maximum of 400 cm, to exclude
regenerating trees from before the 1990s when no infor-
mation on the location of the wolf core area is available.
Browsing intensity in relation to escape impediments
We followed the deﬁnition by Halofsky and Ripple
(2008) and regarded large pieces of downed woody debris as
possible escape impediments which might block view or
escape routes for ungulates (Fig. 2). To study how browsing
intensity interacted with small-scale habitat characteristics,
we counted the presence of escape impediments in four
(wind) directions for each encountered tree sapling. We
measured this at two spatial scales; within a radius of 1 m
and 5 m around each tree sapling. We considered CWD
as a possible escape impediment when its size measured
minimally 0.5 m in height above the ground and 0.5 m in
length for escape impediments within a radius of 1 m, and
0.5 m in height and 1 m in length for escape impediments
between 1 and 5 m radius. A minimum height of 0.5 m was
used (similar to the height used by Halofsky and Ripple
2008) because this blocks the view of a head-down foraging
Figure 2. Large pieces of coarse woody debris may act as escape
impediments by obstructing escape routes of ungulates or increase
their perceived predation risk by blocking view on potential
predators (top, photo by Hidde Zemel). An example of a tree sap-
ling (Tilia cordata) growing in between two escape impediment
variable. We applied a univariate GLM to test for the
eﬀects of sapling height on browsing intensity inside and
outside the wolf core area. We entered both sapling height
and whether the sapling grew inside or outside the wolf
core area as ﬁxed factors. To test for diﬀerences in average
tree sapling height and browsing intensity between core
and non-core area an independent t-test was used.
A Kruskall–Wallis test was used to test for group diﬀer-
ences in browsing intensity between diﬀerent amounts
of escape impediments inside and outside the core area,
while the diﬀerences within groups with a given number of
escape impediments was tested using Mann–Whitney U
test (with Bonferroni correction of signiﬁcance level for
the number of tests performed). We used non-parametric
tests here because the sometimes small and unequal sample
size per group violated assumptions for parametric testing
(Kolmogorov–Smirnov test p 0.05). For a reasonable
estimate of the average browsing intensity considering the
existing variation, we calculated browsing intensity on a
minimum of 10 trees and when the number of trees per
impediment was lower than 10 in core or non-core it was
taken out of the analysis. A Jonckheere–Terpstra test was
used to test for a possible ordered pattern (trend) in
browsing intensity with increasing amounts of escape imped-
iments inside and outside the wolf core area. is test is a
nonparametric test and has more statistical power than a
Kruskall–Wallis test when ordered diﬀerences among
classes occur (Field 2009). e diﬀerence in Shannon diver-
sity index between wolf core and non-core area was tested
with a Mann–Whitney U test. To test for diﬀerences in
number of trees growing out of reach of ungulates between
inside and outside the wolf core area a Chi-square test
was used, whereas, we used Spearman correlation to test
whether the percentage of these trees which escaped ungu-
late control increased was related to the number of escape
impediments for inside and outside the wolf core area
separately. Subsequently, we used Spearman correlation to
test whether the diﬀerences between percentage of trees
which escaped ungulate control inside and outside the core
area was related to the number of escape impediments,
indicating that the diﬀerence between these two areas is
increasing or decreasing. All statistical analyses were per-
formed using SPSS 19.0. All maps were created using QGIS
software (Quantum GIS Development Team 2012)
Browsing inside and outside wolf core area
We measured in total 7414 tree saplings (10–200 cm in
height) equally divided over the area inside and outside the
wolf core area (Table 1). Species composition based on
proportion of each species in the samples tree sapling was
similar inside and outside the wolf core area (c2 2.137,
DF 4, p 0.711) and dominated by Carpinus betulus
( 50%), followed by Acer platanoides, Tilia cordata
and Picea abies (Table 1). Other tree species accounted less
than 5% of the sampled trees (with the exception of Sorbus
aucuparia outside the wolf core area). Number of recorded
individuals per species obviously varied between forest types,
Table 1. Characteristics of sampled area inside and outside the
wolf core area in Białowiez
.a National Park, Poland. Number of
recorded tree saplings, transect length, description of tree sapling
species composition (percentage of saplings, number of saplings)
in which only saplings which are available for ungulate browsing
are included ( 200 cm). Canopy cover refers to canopy closure of
tree stands inside and outside the wolf core area.
Number of trees saplings
Number of transects 17 17
Total transect lenght (m) 3570 3540
Tree sapling density (no. m22 SE) 0.52 ( 0.06) 0.53 ( 0.04)
Tree sapling height (in cm SE) 52.9 ( 0.59) 59.8 ( 0.61)
Carpinus betulus (%, no.) 58 (2135) 52 (1931)
Acer platanoides 12 (458) 14 (510)
Tilia cordata 8 (293) 9 (323)
Picea abies 5 (199) 7 (250)
Ulmus glabra 4 (134) 1 (45)
Betulus spp. 3 (123) 2 (87)
Sorbus aucuparia 3 (93) 8 (294)
Corylus avellana 2 (90) 3 (98)
Quercus robur 2 (82) 2 (62)
Fraxinus excelsior 2 (59) 2 (61)
Euonymus europaeus 1 (10) 1 (15)
Populus tremula 1 (6) 1 (26)
Pinus sylvestris 1 (5) 1 (5)
Salix spp. 1 (1) 1 (8)
Sambucus nigra 1 (1) 0 (0)
Malus sylvestris 0 (0) 1 (10)
Shannon diversity index ( SE) 0.51 ( 0.03) 0.54 ( 0.04)
Canopy cover (% SE) 78.1 ( 3.3) 77.4 ( 2.8)
but were highly similar between inside and outside wolf
core area (Table 1). Only some less abundant tree species
( 5% of sampled trees) showed a higher number of records
inside compared with outside the wolf core area; Sorbus
aucuparia (93 vs 294 saplings, F1, 68 4.65, p 0.035),
Populus tremulus (6 vs 26, F1, 68 5.70, p 0.020),
Malus sylvestris (0 vs 10, F1, 68 6.46, p 0.013). Tree sap-
ling density (0.52 0.06 vs 0.53 0.04 saplings m22,
t 0.380, DF 32, p 0.706) and Shannon diversity
index (0.51 0.03 vs 0.54 0.04, t 0.474, DF 32,
p 0.639) were not diﬀerent between the transects respec-
tively inside and outside the wolf core area. Besides,
canopy cover did not diﬀer between both areas (inside
78.1 3.3 vs outside 77.4 2.8, t 20.48, DF 6.259,
p 0.963). All this combined indicates that available
forage for ungulates was comparable between inside and out-
side the wolf core area. Only the average height of tree
sapling was slightly, but statistically signiﬁcantly higher out-
side (59.8 cm 0.61) than inside (52.9 cm 0.59) the wolf
core area (Table 1, t 29.491, DF 7412, p 0.001).
Browsing intensity, measured as the proportion of
branches browsed for all tree species combined, showed
an optimum between 50 and 150 cm of sapling height
both inside and outside the wolf core area (eﬀect of sapling
height: F10,7819 221.33, p 0.001, Fig. 3). Browsing inten-
sity for all height classes combined was lower on tree saplings
growing inside (44.7% 4.6) than outside (51.4% 5.2)
the wolf core area (F1,7819 45.64, p 0.001), but this
diﬀerence became smaller with increasing sapling height
(interaction between sapling height and core area:
than outside (from 59 to 52%) and was always lower inside
the wolf core area at each number of escape impediments
present (at p 0.0125 following Bonferroni correction).
Within a 5 m radius, the percentage of browsing on tree
saplings decreased with increasing number of escape impedi-
ments, however this eﬀect only occurred inside the wolf
core area (Jonckheere test J 2363964.00, z 4.731,
p 0.001, Fig. 5B). Outside the wolf core area the
browsing intensity was not related to the number of escape
impediments present (Kruskal–Wallis test, H (6) 7.725,
DF 1, p 0.259).
Trees growing out of reach of ungulate browsing
We explored the possible long term eﬀects of wolves on
forest regeneration by comparing the proportion of trees
that escaped browsing by ungulates (all trees 200–400 cm
in height). Overall we observed a higher number of
escaped trees (200–400 cm) inside (n 230) than outside
(n 197) the wolf core area. Inside the wolf core area, trees
that grew out of reach of browsing ungulate occurred
more on locations where escape impediments where pres-
ent than on locations lacking escape impediments (194
with escape impediments vs 36 without impediments,
(c2 108.54, DF 1, p 0.0001). Outside the wolf
core area there was no diﬀerence in the numbers of trees
above browsing height that were growing near escape
impediments (105) compared to the number of trees (92)
that did not (c2 0.858, DF 1, p 0.354). Although
the absolute number of trees above browsing height was
low (Fig. 6), the percentage they made up of the total
recorded trees within each class increased with the number
of escape impediments both inside (Spearman correla-
tion coeﬃcient, r 0.90, n 9, p 0.001) and outside
(r 0.667, n 9, p 0.05) the wolf core area. However,
the percentage of trees above browsing height increased
more strongly with increasing number of escape impediments
F10,7819 2.62, p 0.003). Above 160 cm there was no dif-
ference. When all trees within preferred foraging height
of red deer are combined (all trees 150 cm, Renaud
et al. 2003), overall browsing intensity was 8% lower
inside (50.5%) than outside (58.3%) the wolf core area
(t 211.95, DF 7156, p 0.001).
is overall pattern of lower browsing intensity (for all
species combined) was also observed when comparing
the transects inside and outside the wolf core area within
forest types (Fig. 4A). Browsing intensity of tree saplings
(height 10–200 cm) was lower inside than outside the wolf
core area for wet deciduous (t 23.131, DF 2379,
p 0.01), deciduous forest (t 28.207, DF 2867.50
(equal variances not assumed), p 0.001), mixed deciduous
forest (t 211.355, DF 909.83, p 0.001), mixed
coniferous forest (t –3.082, DF 1090, p 0.01). In
addition, browsing intensity was lower inside wolf core
areas than outside when comparing within the four
most common tree species (Fig. 4B); Acer platanoides
(t 22.308, DF 922.95, p 0.021), Tilia cordata
(t 22.565, DF 614, p 0.011), Carpinus betulus
(t 212.800, DF 4064, p 0.001), Picea abies
(t 22.224, DF 447, p 0.027).
Browsing intensity and escape impediments
We tested the eﬀect of the presence of escape impediments
separately for the saplings inside and outside the wolf core
area at two spatial scales; within 1 m radius, and within
5 m radius (Fig. 5). When the number of escape impedi-
ments within 1 m radius of a sapling increased, the percent-
age of browsing on tree saplings (10–200 cm) decreased
both inside (Jonckheere test J 1039546.00, z –4.870,
p 0.001) and outside the wolf core area (Jonckheere
test J 934532.50, z 22.769, p 0.01, Fig. 5A).
However, browsing intensity decreased more strongly with
the number of escape impediments inside (from 51 to 27%)
Tree height class (cm)
(% top branches browsed)
215 1137 935 574 344 213 120 81 63 43 197 - Outside core
222 1546 880 408 223 146 108 65 45 45 231- Inside core
Figure 3. Browsing intensity of tree sapling per height class outside ( ) and inside (•) the wolf core area measured as the proportion of
browsed top 10 branches at diﬀerent classes of tree sapling height. Averages ( SE) are based on 17 transects inside and outside the wolf
core area. Sample sizes are indicated at the top of the graph.
Wet deciduous Deciduous Mixed deciduous Mixed coniferous
(% of top branches browsed)
Acer platanoides Carpinus betulus Tilia cordata Picea abies
(% of top banches browsed)
Figure 4. (A) Browsing intensity of tree saplings (10–200 cm in height) per forest type outside (grey bars) and inside (black bars) the wolf
core area. (B) Browsing intensity of the four most common tree saplings (10–200 cm) per tree species with more than 150 saplings outside
(grey bars) and inside (black bars) the wolf core area. Only the four most common tree species in the sample are presented with at least 150
saplings in both inside and outside the wolf core area (sample sizes are indicated above the bars). Signiﬁcant diﬀerences in browsing inten-
sity within forest types and within tree species are indicated with an asterisk (*p 0.05, **p 0.01, ***p 0.001).
inside the wolf core (correlation on diﬀerences inside and
outside core area r 0.883, n 9, p 0.002). With a high
number of escape impediments, up to 75% of tree saplings
grew above browsing height inside the wolf core area versus
Our study showed that ungulate browsing intensity of tree
saplings, irrespective of tree species, was reduced by 8%
inside a wolf pack core area compared to the rest of their ter-
ritory. is diﬀerence in browsing intensity between inside
and outside the core area strongly increased with the amount
of coarse woody debris (CWD) in direct vicinity of tree sap-
lings. As a result, more trees can grow above browsing height
with increasing amount of CWD inside compared to outside
the wolf core area. Whereas higher predation pressure can
explain the overall lower browsing intensity inside the core
area, behavioral changes of ungulates lead to a stronger
reduction in browsing than density-mediated eﬀects alone.
ese patterns strongly indicate that CWD can create ﬁne-
scale risk eﬀects for ungulate browsers. is is the ﬁrst
study we are aware of, that shows the potential for cascading
eﬀects of large predators on tree regeneration for a European
forest system. As the physical landscape of our study area
contrasts largely with already studied American systems,
this study increases our understanding of indirect predator
eﬀects on ecosystem functioning in diﬀerent landscapes.
Ungulate browsing inside and outside wolf core area
We observed lower browsing levels on trees growing inside
the core area of the territory of a wolf pack. Diﬀerent tree
Number of escape impediments in 5-m radius
Escaping trees (% of total trees)
1001372 1045 776 404 302 85 23 74 - Outside core
1817 997 547 273 158 67 37 18 8 - Inside core
Figure 6. Percentage of trees growing out of reach of ungulate browsing (200–400 cm) of total present trees growing in relation to the
amount of escape impediments within 5 m radius of a tree outside ( ) and inside (•) the wolf core area. Total number of trees present in
each category is indicated on top of the ﬁgure.
Number of escape impediments in 1-m radius
(% of top brnaches browsed)
70 3070 503 102 26 - Outside core
2927 545 181 30 - Inside core
Number of escape impediments in 5-m radius
(% of top brnaches browsed)
70 1725 936 530 266 148 64 34 - Outside core
1336 984 721 369 179 79 17 - Inside core
Figure 5. Browsing intensity on tree saplings (10–200 cm) outside ( ) and inside (•) the wolf core area in relation to the presence of varying
numbers of escape impediments only within 1 m radius (A) and within 5 m radius of each sapling (B). Sample sizes are indicated at the top
in each graph, browsing intensity is only determined for samples size of 10 trees. Asterisks indicate signiﬁcant diﬀerences between inside
and outside core based on Mann–Whitney test with Bonferroni correction of signiﬁcance level (p 0.0125 in (A) and p 0.007 in (B)).
most strongly indicated by the relationship we found with
the presence of coarse woody debris (CWD). Ungulates in
African systems have been shown to avoid and are more
vigilant in habitats with low visibility (Underwood 1982,
Shrader et al. 2008, Valeix et al. 2009). Earlier American
studies have found that deer reduce foraging and increase
vigilance levels in the vicinity of objects that reduce
habitat visibility or can serve as objects obstructing escape
(Halofsky and Ripple 2008, Liley and Creel 2008). ese
so-called escape impediments can be river valleys, ridges,
fallen trees or rocks (Halofsky and Ripple 2008). ere is
consensus that indirect, behaviorally-mediated eﬀects
of predators on ungulates occur in relation to habitat fea-
tures, however, the scale at which these eﬀects operate is
less clear. While several American studies have illustrated
eﬀects of predator presence on large scale distribution
of ungulates in relation to predation risk (Lima and Dill
1990, Kie 1999, Creel et al. 2005, Mao et al. 2005),
undisputed evidence that ungulates react to ﬁne-scale risk
factors aﬀecting foraging behavior or patch selection on a
small scale is lacking (Winnie 2012). at browsing inten-
sity on trees in the present study was more reduced with
increasing CWD inside the wolf core area compared to
outside the wolf core area, suggests that deer experience
higher predation risk eﬀects (perceived predation risk).
Outside the wolf core area, where the encounter rate of
wolves is lower, deer browsing intensity was not related
to the presence of CWD. Similarly, the number of trees
growing out of reach of herbivore control ( 200 cm)
increased with increasing CWD only inside the wolf core
area. is shows that a higher presence of wolves alone
is not the sole factor, but ﬁne-scale risk factors are impor-
tant related to habitat characteristics such as the presence
of CWD. e present study suggests that behaviorally-
mediated eﬀects of predators on prey can become in this
way more important than density-mediated eﬀects in
aﬀecting lower trophic levels.
According to the ‘risk allocation hypothesis’ vigilance
levels of deer may actually drop when the predator encoun-
ter increases (Lima and Bednekoﬀ 1999). Evidence for
this theory has also been found in wolf-red deer systems
(Creel et al. 2008). However, these behavioral eﬀects of red
deer in response to wolf presence have been observed in sys-
tems with an open landscape where deer often see wolves
approaching and can respond to it in advance (Creel et al.
2008). How red deer respond to predator presence in closed
forest systems, with low visibility and lower chance to detect
predators by sight, as our present study area, is still largely
unknown and may diﬀer from these patterns. e inter-
active eﬀects of (cues of ) predator presence and habitat
characteristics appear to inﬂuence in these systems spatial
patterns of tree regeneration and provide local spots with
reduced herbivore pressure and increased tree regeneration.
Landscape of fear or alternative explanation for
differences in browsing?
We explain the observed lower browsing intensity of tree
saplings inside the wolf core area as being the result of a com-
bination of direct (lethal) and indirect (non-lethal) predator
species showed a similar reduction in proportion of browsed
branches. Since red deer is the main browser in our system
(Jędrzejewska et al. 1997) and plays an important role in
determining patterns of tree regeneration (Kuijper et al.
2010a, b), patters in browsing are mainly determined by
this species. is ﬁts with the observed highest browsing
intensity on saplings in the preferred browsing height class
of red deer of 50–150 cm (Renaud et al. 2003). e reduc-
tion in browsing intensity inside the wolf core area is most
easily explained by density-mediated eﬀects of wolves
on red deer (Jędrzejewski et al. 2002). Based on the fact
that wolves spend more than 50% of their time annually
inside the core area of their territory (Jędrzejewski et al.
2007, Schmidt et al. 2009) and their spatial distribution
in spring–summer is restricted within this core area
(Jędrzejewski et al. 2001), we calculated a 5-fold higher
predation pressure (using Jędrzejewski et al. 2002) inside
the core area on a yearly basis. Outside the reproductive
period (autumn–winter), wolves increase their daily move-
ment distances and utilize their territory in a rotational
way (Jędrzejewski et al. 2001), hence killed prey can then
be found throughout their territory (Jędrzejewska and
Jędrzejewski 1998). is higher predation pressure in com-
bination with the strong range ﬁdelity of red deer living in
this area (Kamler et al. 2008) should lead to a lower red deer
density and lower browsing intensity inside the core area.
In addition to these density-mediated eﬀects, behaviorally-
mediated eﬀects can lead to a reduction in browsing inten-
sity by ungulates avoiding risky habitats or risky times
(Lima and Dill 1990, Mao et al. 2005, Creel et al. 2008,
Valeix et al. 2009, aker et al. 2011) or changing their
behavior (Pulliam 1973, Hunter and Skinner 1998, Brown
et al. 1999). Although, habitat characteristics between wolf
core areas and the remainder of their territories can diﬀer
(see below), forage availability for browsing ungulates
seems similar in the two areas. We observed a similar tree
sapling species composition, no diﬀerence in Shannon
diversity index and no diﬀerence in density of recorded tree
saplings inside and outside the wolf core area. We only
found that sapling height was 7 cm higher outside the wolf
core area, which unlikely has a large eﬀect on forage avail-
ability for red deer. e ﬁnding that inside the core area
still more than 50% of branches is being browsed of trees
in the height class 50–100 cm, shows that ungulates are
still abundantly present. Previous studies in the Yellowstone
National park showed that deer became more vigilant once
wolves returned to the area (Laundré et al. 2001). Also
other studies showed that deer may directly respond by
increased vigilance and lower foraging only when wolves
are present (Creel and Winnie 2005, Liley and Creel 2008).
ese behavioral eﬀects where observed already when
wolves were within 3 km distance (Liley and Creel 2008).
Similar direct eﬀects of predator presence on prey behavior
were also observed in African systems (Valeix et al. 2009).
Likewise, red deer in our study area may perceive wolf core
areas as more risky due to the higher frequency of occur-
rence of wolves (Jędrzejewski et al. 2001, Schmidt et al.
2009) and show behavioral changes leading to a lower
foraging eﬃciency when present inside the wolf core area.
at risk eﬀects are an important underlying mechanism
explaining our observed patterns in browsing intensity is
and random location outside the den area (euerkauf
et al. 2003a), suggesting that these are also not important
confounding factor explaining our results.
Finally, the observed patterns may be caused by the sec-
ond predator in the system, lynx, which also predates on
red deer (Okarma et al. 1997). Growing evidence suggests
that ambush predators, such as lynx, produce stronger
behaviourally-induced risk eﬀects than chase hunters such
as wolves (Schmitz and Suttle 2001, aker et al. 2011).
Home ranges of wolves and lynx in the BPF considerably
overlap and in some cases the core areas of their territories
overlapped (Schmidt et al. 2009). However, the wolf
territory used in the present study inside the Białowieza
National Park showed little overlap with the present lynx
territories. Whereas the lynx territory had a core area
situated in the south of the BNP, wolves had theirs in
the northern parts (Schmidt et al. 2009). Moreover, the loca-
tion of territories (and core area) of lynx are not stable
and changes each year (K. Schmidt pers. comm.). erefore,
the location of lynx territories does not provide a likely
explanation for browsing levels which are reduced for
longer periods, allowing eﬀects on the proportion of trees
escaping herbivore control.
To conclude, we are inclined to interpret the lower brows-
ing intensity of trees observed inside the wolf core areas as
being mainly the result of higher predation pressure and
increased perceived predation risk by ungulates rather than
other factors associated with the wolf core area.
Landscape of fear in European systems
Apex carnivores have been extirpated in many areas across
the globe, leading to cascading eﬀects on ecosystem func-
tioning (Estes et al. 2011). Especially in densely populated
areas, such as Europe, the function of large carnivores has
often been lost or their numbers have dropped below eco-
logical meaningful level. However, there has been a recent
expansion from wolf and lynx from thriving populations in
eastern Europe to areas in western Europe (Breitenmoser
1998, Enserink and Vogel 2006, Trouwborst 2010).
is recolonisation by apex predators may restore anti-
predator behaviour of ungulates with important conse-
quences for ecosystem functioning. e knowledge we have
from predator–ungulate interactions originating mainly
from American study systems is often applied to predict
how recolonisation or reintroduction of predators may
inﬂuence ungulate’s behaviour (Manning et al. 2009).
However, especially indirect predator eﬀects on ungulates in
European systems may diﬀer from those in American ones
because of several reasons, for example smaller size and con-
sequently lower landscape heterogeneity of nature reserves,
more fragmented landscape and higher human pressure.
erefore, studying indirect predator eﬀects on ungulate
behaviour in a variety of systems is needed to accurately pre-
dict their future role. e present study is the ﬁrst we are
aware of that has addressed the potential for how direct and
indirect predator eﬀects on ungulates can lead to cascading
eﬀects on tree performance and aﬀect spatial patterns of tree
regeneration in a European system. It illustrates that in
densely forested areas, with predators being present virtually
everywhere, indirect predator eﬀects seem to occur but also
eﬀects on their prey species. However, there might be several
confounding factors explaining our results.
As the core area and non-core area were not randomly
chosen, other confounding factors cannot be controlled
for. Previous studies in BPF showed that wolf den’s are
often found in parts of their territory which are associated
with the least human activity and often the farthest
away from any road or village (euerkauf et al. 2003b,
euerkauf and Rouys 2008). To minimise these potential
diﬀerences between inside and outside core areas we put
transects in all cardinal directions from the core areas
to exclude possible gradients in human activity. Besides
we restricted our study to the BNP where motorised vehi-
cles and hunting and forestry activities are prohibited.
Anyway, the studied wolf core area is the part of the
study area least visited by people mainly due to the concen-
tration of visiting tourists in the south-western corner of the
forest. is may have aﬀected the distribution patterns of
red deer and browsing intensity. Studies from other systems
showed that deer can concentrate in those parts most
frequently visited by humans to escape predation pressure
(Kloppers et al. 2005, Muhly et al. 2011, Rogala et al.
2011). Whether deer in our study area also show similar
behavioral responses is unknown. However, in case they do,
lower browsing levels inside the wolf core area can still
be interpreted as resulting from indirect predator eﬀects
creating a higher perceived predation risk inside core
areas. Alternatively, deer in our study area may respond dif-
ferently to human presence as they are intensively hunted
outside the strictly protected area. Earlier studies from
Białowieża Primeval forest showed that also deer avoid areas
of human activity (euerkauf and Rouys 2008). In that
case, the opposite trend in observed browsing levels of
trees would have been expected, with most browsing in the
least disturbed area hence inside the wolf core area. Since
this did not occur we do not regard human activity as
an important confounding factor explaining our results.
In addition the location of wolf core areas may be related
to other habitat characteristics. Our results showed that
wolf core and non-core area are comparable in tree density,
Shannon diversity index and canopy cover. Only the
average sapling height was higher (7 cm) outside the wolf
core area compared to inside the core area. is diﬀerence
cannot be explained by browsing intensity as this was
higher outside the wolf core area. As there was no diﬀerence
in canopy cover, as a proxy of light availability, a higher
soil fertility outside the wolf core area might potentially
result in higher growth rate of tree saplings and explain
these results. Earlier studies showed that wolves dens
and resting sites in our study area are more often located
in coniferous forest compared to random locations, but
occurred in all other present forest types as well (euerkauf
et al. 2003a). We observed that the lower browsing intensity
inside the wolf core area occurred within each forest type.
is indicates that diﬀerences in the occurrence of forest
type between inside and outside the wolf core area are not a
confounding factor in our analyses. Other habitat features
related to perceived predation risk such as habitat visibility
(Shrader et al. 2008, Valeix et al. 2009) and the amount of
escape impediments (Halofsky and Ripple 2008) were not
diﬀerent between location of dens (in the wolf core area)
Faliński, J. B. 1986. Vegetation dynamics in temperate lowland
primeval forests: ecological studies in Białowieża forest.
– Dr. W. Junk Publishers.
Field, A. P. 2009. Discovering statistics using SPSS. – SAGE
Fischhoﬀ, I. R. et al. 2007. Habitat use and movements of plains
zebra (Equus burchelli) in response to predation danger
from lions. – Behav. Ecol. 18: 725–729.
Fortin, D. et al. 2005. Wolves inﬂuence elk movements:
behavior shapes a trophic cascade in Yellowstone National
Park. – Ecology 86: 1320–1330.
Fretwell, S. D. 1987. Food-chain dynamics – the central theory
of ecology. – Oikos 50: 291–301.
Halofsky, J. S. and Ripple, W. J. 2008. Fine-scale predation risk
on elk after wolf reintroduction in Yellowstone National Park,
USA. – Oecologia 155: 869–877.
Hunter, L. T. B. and Skinner, J. D. 1998. Vigilance behaviour in
African ungulates: the role of predation pressure. – Behaviour
Jędrzejewska, B. and Jędrzejewski, W. 1998. Predation in vertebrate
communities: the Baiłowieża Primeval Forest as a case study.
Jędrzejewska, B. and Jędrzejewski, W. 2005. Large carnivores and
ungulates in European temperate forest ecosystems: bottom-up
and top-down control. – In: Ray, J. C. et al. (eds), Large
carnivores and the conservation of biodiversity. Island Press,
Jędrzejewska, B. et al. 1994. Eﬀects of exploitation and protection
on forest structure, ungulate density and wolf predation
in Białowieża Primeval Forest, Poland. – J. Appl. Ecol. 31:
Jędrzejewska, B. et al. 1997. Factors shaping population densities
and increase rates of ungulates in Białowieża Primeval Forest
(Poland and Belarus) in the 19th and 20th century. – Acta
eriol. 42: 399–451.
Jędrzejewski, W. et al. 2001. Daily movements and territory use
by radio-collared wolves (Canis lupus) in Białowieża Primeval
Forest in Poland. – Can. J. Zool. 79: 1993–2004.
Jędrzejewski, W. et al. 2002. Kill rates and predation by wolves on
ungulate populations in Białowieża Primeval Forest (Poland).
– Ecology 83: 1341–1356.
Jędrzejewski, W. et al. 2007. Territory size of wolves Canis lupus:
linking local (Białowieża Primeval Forest, Poland) and
Holarctic-scale patterns. – Ecography 30: 66–76.
Kamler, J. F. et al. 2008. Home ranges of red deer in a European
old-growth forest. – Am. Midl. Nat. 159: 75–82.
Kauﬀman, M. J. et al. 2007. Landscape heterogeneity shapes
predation in a newly restored predator–prey system. – Ecol.
Lett. 10: 690–700.
Kauﬀman, M. J. et al. 2010. Are wolves saving Yellowstone’s aspen?
A landscape-level test of a behaviorally mediated trophic
cascade. – Ecology 91: 2742–2755.
Kie, J. G. 1999. Optimal foraging and risk of predation: eﬀects on
behavior and social structure in ungulates. – J. Mammal.
Kloppers, E. L. et al. 2005. Predator-resembling aversive condition-
ing for managing habituated wildlife. – Ecol. Soc. 10: 31.
Kuijper, D. P. J. et al. 2009. Do ungulates preferentially feed in
forest gaps in European temperate forests? – For. Ecol. Manage.
Kuijper, D. P. J. et al. 2010a. Fluctuating ungulate density shapes
tree recruitment in natural stands of the Białowieża Primeval
forest, Poland. – J. Veg. Sci. 21: 1082–1098.
Kuijper, D. P. J. et al. 2010b. Bottom-up versus top-down control
of tree regeneration in the Białowieża Primeval Forest, Poland.
– J. Ecol. 98: 888–899.
LaGory, K. E. 1986. Habitat, group size, and the behaviour of
white-tailed deer. – Behaviour 98: 168–179.
might be more subtle (i.e. not leading to large shifts in
spatial distribution or habitat selection) than observed
in large scale American national parks. e underlying
mechanism might also diﬀer, for example olfactory cues
might be an important attribute indicating predator risk
eﬀects in these low visibility habitats.
Whereas ungulate management often focuses only on
the direct eﬀects of predators in inﬂuencing ungulate
numbers, the potential for indirect eﬀects in European sys-
tems should be acknowledged (see also Cromsigt et al.
2013). e present study is a ﬁrst step to show how preda-
tors structure ecosystems in a European context and might
help to broaden the discussion on predator eﬀects in a
variety of systems.
Acknowledgements – Part of the work of DPJK has been supported
by a Marie Curie European Reintegration Grant under the
7th framework program (project PERG06-GA-2009-256444). In
addition, the work of DPJK was supported by funding from
the Polish Ministry of Science and Higher Education (grant no.
2012/05/B/NZ8/01010), whereas the work of MC and BJ was
ﬁnanced by grant no. 5P06H 03418.
Beschta, R. L. and Ripple, W. J. 2011. Are wolves saving
Yellowstone’s aspen? A landscape-level test of a behaviorally
mediated trophic cascade – comment. – Ecology doi:10.1890/
Bertram, B. C. R. 1978. Living in groups: predators and prey.
– In: Krebs, J. R. and Davies, N. B. (eds), Behavioural ecology.
Oxford Univ. Press, pp. 64–96.
Beyer, H. L. et al. 2007. Willow on Yellowstone’s northern range:
evidence for a trophic cascade? – Ecol. Appl. 17: 1563–1571.
Breitenmoser, U. 1998. Large predators in the Alps: the fall
and rise of man’s competitors. – Biol. Conserv. 83: 279–289.
Brown, J. S. et al. 1999. e ecology of fear: optimal foraging,
game theory, and trophic interactions. – J. Mammal. 80:
Creel, S. and Winnie, J. A. Jr. 2005. Responses of elk herd size
to ﬁne-scale spatial and temporal variation in the risk of
predation by wolves. – Anim. Behav. 69: 1181–1189.
Creel, S. and Christianson, D. 2008. Relationships between
direct predation and risk eﬀects. – Trends Ecol. Evol. 23:
Creel, S. and Christianson, D. 2009 Wolf presence and increased
willow consumption by Yellowstone elk: implications for
trophic cascades. – Ecology 90: 2454–2466.
Creel, S. et al. 2005. Elk alter habitat selection as an antipredator
response to wolves. – Ecology 86: 3387–3397.
Creel, S. et al. 2008. Time and space in general models of
antipredator response: test with wolves and elk. – Anim.
Behav. 76: 1139–1146.
Cromsigt J. P. G. M. et al. 2013. Hunting for fear: a promising
tool to manage ecosystem functioning? – J. Appl. Ecol. doi:
DeAngelis, D. L. 1992. Dynamics of nutrient cycling and food
webs. – Chapman and Hall.
Dehn, M. M. 1990. Vigilance for predators: detection and dilution
eﬀects. – Behav. Ecol. Sociobiol. 26: 337–342.
Enserink, M. and Vogel, G. 2006. Wildlife conservation – the
carnivore comeback. – Science 314: 746–749.
Estes, J. A. et al. 2011. Trophic downgrading of Planet Earth.
– Science 333: 301–306.
Ripple, W. J. and Beschta, R. L. 2005. Linking wolves and plants:
Aldo Leopold on trophic cascades. – Bioscience 55: 613–621.
Ripple, W. J. and Beschta, R. L. 2006. Linking wolves to willows
via risk-sensitive foraging by ungulates in the northern
Yellowstone ecosystem. – For. Ecol. Manage. 230: 96–106.
Ripple, W. J. et al. 2001. Trophic cascades among wolves, elk and
aspen on Yellowstone National Park’s northern range. – Biol.
Conserv. 102: 227–234.
Rogala, J. K. et al. 2011. Human activity diﬀerentially redistributes
large mammals in the Canadian Rockies National Parks.
– Ecol. Soc. 16: 16.
Runkle, J. R. 1981. Gap regeneration in some old-growth forests
of the eastern United States. – Ecology 62: 1041–1051.
Samojlik, T. et al. 2007. Man in the ancient forest. – Academia 4:
Schmidt, K. et al. 2008. Reproductive behaviour of wild-living
wolves in Białowieża Primeval Forest (Poland). – J. Ethol. 26:
Schmidt, K. et al. 2009. Spatial interactions between grey wolves
and Eurasian lynx in Białowieża Primeval Forest, Poland.
– Ecol. Res. 24: 207–214.
Schmitz, O. J. and Suttle, K. B. 2001. Eﬀects of top predator spe-
cies on direct and indirect interactions in a food web. – Ecology
Schmitz, O. J. et al. 1997. Behaviorally mediated trophic cascades:
eﬀects of predation risk on food web interactions. – Ecology
Shrader, A. M. et al. 2008. Do free-ranging domestic goats show
‘landscapes of fear’? Patch use in response to habitat features
and predator cues. – J. Arid Environ. 72: 1811–1819.
Terborgh, J. and Estes, J. A. 2010. Trophic cascades: predators,
prey, and the changing dynamics of nature. – Island Press.
aker, M. et al. 2011. Minimizing predation risk in a landscape
of multiple predators: eﬀects on the spatial distribution of
African ungulates. – Ecology 92: 398–407.
euerkauf, J. and Rouys, S. 2008. Habitat selection by ungulates
in relation to predation risk by wolves and humans in
the Białowieża Forest, Poland. – For. Ecol. Manage. 256:
euerkauf, J. et al. 2003a. Selection of den, rendezvous, and
resting sites by wolves in the Białowieża Forest, Poland. – Can.
J. Zool. 81: 163–167.
euerkauf, J. et al. 2003b. Spatiotemporal segregation of wolves
from humans in the Białowieża Forest (Poland). – J. Wildl.
Manage. 67: 706–716.
Trouwborst, A. 2010. Managing the carnivore comeback: interna-
tional and EU species protection law and the return of lynx, wolf
and bear to western Europe. – J. Environ. Law 22: 347–372.
Underwood, R. 1982. Vigilance behaviour in grazing African ante-
lopes. – Behaviour 79: 82–107.
Valeix, M. et al. 2009. Behavioral adjustments of African herbivores
to predation risk by lions: spatiotemporal variations inﬂuence
habitat use. – Ecology 90: 23–30.
Winnie, J. A. Jr. 2012. Predation risk, elk, and aspen: test of
a behaviorally mediated trophic cascade in the Greater
Yellowstone Ecosystem. – Ecology 93: 2600–2614.
Zar, J. H. 1984. Biostatistical analysis. – Prentice Hall.
Zub, K. et al. 2003. Wolf pack territory marking in the Białowieża
Primeval Forest (Poland). – Behaviour 140: 635–648.
Laundré, J. W. et al. 2001. Wolves, elk, and bison: reestablishing
the “landscape of fear” in Yellowstone National Park, U.S.A.
– Can. J. Zool. 79: 1401–1409.
Lemmon, P. E. 1956. A spherical densiometer for estimating forest
overstory density. – For. Sci. 2: 314–320.
Liley, S. and Creel, S. 2008. What best explains vigilance in elk:
characteristics of prey, predators, or the environment? – Behav.
Ecol. 19: 245–254.
Lima, S. L. and Dill, L. M. 1990. Behavioral decisions made under
the risk of predation: a review and prospectus. – Can. J. Zool.
Lima, S. L. and Bednekoﬀ, P. A. 1999. Temporal variation in
danger drives antipredator behavior: the predation risk
allocation hypothesis. – Am. Nat. 153: 649–659.
Lipetz, V. E. and Bekoﬀ, M. 1982. Group size and vigilance in
pronghorns. – Z. Tierpsychol. 58: 203–216.
Manning, A. D. et al. 2009. Restoring landscapes of fear
with wolves in the Scottish Highlands. – Biol. Conserv. 142:
Mao, J. S. et al. 2005. Habitat selection by elk before and
after wolf reintroduction in Yellowstone National Park. – J.
Wildl. Manage. 69: 1691–1707.
McLaren, B. E. and Peterson, R. O. 1994. Wolves, moose, and
tree rings on Isle Royale. – Science 266: 1555–1558.
McNamara, J. M. and Houston, A. I. 1992. Evolutionarily stable
levels of vigilance as a function of group size. – Anim. Behav.
Messier, F. 1994. Ungulate population models with predation:
a case study with the American moose. – Ecology 75:
Michalczuk, C. 2001. Forest habitats and tree stands of the
Białowieża National Park. – Phytocoenosis 13, Supplementum
Cartographiae Geobotanicae 13, Białowieża Geobotanical
Station of Warsaw Univ., Warszawa.
Muhly, T. B. et al. 2011. Human activity helps prey win the
predator–prey space race. – PLoS One 6: e17050.
Nowak, S. et al. 2007. Howling activity of free-ranging
wolves (Canis lupus) in the Białowieża Primeval Forest and
the western Beskidy Mountains (Poland). – J. Ethol. 25:
Okarma, H. et al. 1997. Predation of Eurasian lynx on roe deer
and red deer in Białowieża Primeval Forest, Poland. – Acta
eriol. 42: 203–224.
Oksanen, L. et al. 1981. Exploitation ecosystems in gradients of
primary productivity. – Am. Nat. 118: 240–261.
Pielou, E. C. 1975. Ecological diversity. – Wiley.
Pulliam, H. R. 1973. On the advantages of ﬂocking. – J. eor.
Biol. 38: 419–422.
Pulliam, H. R. et al. 1982. e scanning behaviour of juncos:
a game theoretical approach. – J. eor. Biol. 95: 89–103.
Quantum GIS Development Team 2012. Quantum GIS
Geographic Information System. Open Source Geospatial
Foundation Project. – http://qgis.osgeo.org.
Renaud, P. C. et al. 2003. Damage to saplings by red deer
(Cervus elaphus): eﬀect of foliage height and structure. – For.
Ecol. Manage. 181: 31–37.
Ripple, W. J. and Beschta, R. L. 2003. Wolf reintroduction,
predation risk, and cottonwood recovery in Yellowstone
National Park. – For. Ecol. Manage. 184: 299–313.