ArticlePDF Available

Abstract and Figures

Wolf (Canis lupus) kill rates, factors affecting their variation, and predation impact on ungulates were studied in the Polish part of Białowieża Primeval Forest (580 km2). With the mean size of hunting groups being 4.4 individuals, wolves killed, on average, 0.513 ± 0.04 prey · (pack)-1· d-1 (mean ± 1 SE); 63% of prey were red deer (Cervus elaphus), 28% were wild boar (Sus scrofa), and 4% were roe deer (Capreolus capreolus). Per capita kill rate averaged 0.116 ungulates ·(wolf)-1· d-1, and daily food intake was 5.58 ± 0.32 kg·(wolf)-1· d-1 . Kill rate on red deer was affected by snow cover (P -1· d-1), when piglets were present, than in autumn-winter (0.106 ± 0.04 boar · (pack)-1· d^{-1). Annually, wolves killed on average 72 red deer, 16 roe deer, and 31 wild boar over a 100-km2 area. Compared to prey densities, wolves were an important agent of mortality for red deer only, taking annually 12% of spring-summer (seasonally highest) numbers of deer, which was equivalent to 40% of deer annual increase due to breeding and 40% of their annual mortality. Compared to winter densities (3-6 deer/km2), percentage predation by wolves was inversely density dependent; thus wolves limited deer numbers but did not regulate prey population. By eliminating a substantial proportion of the annual production of the deer population, wolves hamper its growth and prolong the time until it reaches carrying capacity of the habitat. However, wolf predation alone is a poor predictor of deer population dynamics, because deer are also subject to lynx (Lynx lynx) predation and hunting harvest.
Content may be subject to copyright.
1341
Ecology,
83(5), 2002, pp. 1341–1356
q
2002 by the Ecological Society of America
KILL RATES AND PREDATION BY WOLVES ON UNGULATE
POPULATIONS IN BIAŁOWIEZ
˙A PRIMEVAL FOREST (POLAND)
W
Ł
ODZIMIERZ
J
E˛DRZEJEWSKI
,
1,3
K
RZYSZTOF
S
CHMIDT
,
1
J
O
¨RN
T
HEUERKAUF
,
1,2
B
OGUMI
Ł
A
J
E˛DRZEJEWSKA
,
1
N
URIA
S
ELVA
,
1
K
AROL
Z
UB
,
1
AND
L
UCYNA
S
ZYMURA
1
1
Mammal Research Institute, Polish Academy of Sciences, 17-230 Białowiez˙a, Poland
2
Wildlife Biology and Management Unit, Department of Ecosystem and Landscape Management,
Technical University of Munich, 85354 Freising, Germany
Abstract.
Wolf (
Canis lupus
) kill rates, factors affecting their variation, and predation
impact on ungulates were studied in the Polish part of Białowiez˙a Primeval Forest (580
km
2
). With the mean size of hunting groups being 4.4 individuals, wolveskilled, on average,
0.513
6
0.04 prey·(pack)
2
1
·d
2
1
(mean
6
1
SE
); 63% of prey were red deer (
Cervus elaphus
),
28% were wild boar (
Sus scrofa
), and 4% were roe deer (
Capreolus capreolus
). Per capita
kill rate averaged 0.116 ungulates·(wolf)
2
1
·d
2
1
, and daily food intake was 5.58
6
0.32
kg·(wolf)
2
1
·d
2
1
. Kill rate on red deer was affected by snow cover (
P
,
0.001). A pack of
wolves killed, on average, 0.264 deer/d in seasons with no snow and 0.587 deer/d when
snow was 17 cm deep. The increase in kill rates coincided with a decline in the condition
of juvenile (but not adult) deer in late winter (mean marrow fat content in the femur 66%
in October–January vs. 27% in February–March). Per capita kill rates decreased slightly
(not significantly) with the increasing size of wolf hunting group. However, the amount of
food acquired per wolf did not differ among groups containing 2–6 individuals, because
larger packs killed bigger prey more often and small prey less frequently than did small
packs. Wolf kill rates on wild boar were higher in spring–summer (0.242
6
0.06
boar·(pack)
2
1
·d
2
1
), when piglets were present, than in autumn–winter (0.106
6
0.04
boar·(pack)
2
1
·d
2
1
). Annually, wolves killed on average 72 red deer, 16 roe deer, and 31
wild boar over a 100-km
2
area. Compared to prey densities, wolves were an important agent
of mortality for red deer only, taking annually 12% of spring–summer (seasonally highest)
numbers of deer, which was equivalent to 40% of deer annual increase due to breeding and
40% of their annual mortality. Compared to winter densities (3–6 deer/km
2
), percentage
predation by wolves was inversely density dependent; thus wolves limited deer numbers
but did not regulate prey population. By eliminating a substantial proportion of the annual
production of the deer population, wolves hamper its growth and prolong the time until it
reaches carrying capacity of the habitat. However, wolf predation alone is a poor predictor
of deer population dynamics, because deer are also subject to lynx (
Lynx lynx
) predation
and hunting harvest.
Key words:
Canis lupus; Capreolus capreolus; Cervus elaphus;
deer population; density depen-
dence of predation; European pristine forests; hunting group size; snow cover;
Sus scrofa;
wolf kill
rates.
I
NTRODUCTION
The rate at which wolves kill their prey in relation
to prey density, known as functional response (Holling
1959), is an essential component of their predation on
ungulates. With different types of functionalresponses
and various numerical reactions of wolves to changes
in prey densities, the role of predation can range from
regulatory (density-dependent predation rates), to lim-
iting (density-independent predation), to depensatory
(inversely density-dependent predation rates) (Sinclair
1989, Messier 1991, 1995). However, both scarcity and
high variability of the available empirical data on wolf
kill rates often make it difficult to distinguish between
type II (asymptotic) and type III (logistic) functional
Manuscript received 20 October 2000; revised 7 August 2001;
accepted 6 September 2001.
3
E-mail: wlodek@bison.zbs.bialowieza.pl
responses, especially in the low range of prey densities
(Messier 1994, Boyce 2000, Hayes and Harestad 2000).
Aside from sampling errors, this variation in empiri-
cally assessed kill rates stems from the fact that the
rate at which wolves capture prey is governed by a set
of factors much more complex than just prey density.
Among these factors are snow cover (Huggard 1993),
wolf group size (Thurber and Peterson 1993, Hayes et
al. 2000), vulnerability of prey (Peterson and Allen
1974), and availability of alternative prey (Dale et al.
1994). All such external factors can add substantial
variation to wolf predation rates, independent of prey
density.
Unknown is the seasonal variation in kill rates by
wolves, especially the response to prolific abundance
of newborn ungulates in spring–summer. This results
from the restrictions of the method of detection used
(search for kills from aircraft), which is effective during
1342
WŁODZIMIERZ JE˛DRZEJEWSKI ET AL.
Ecology, Vol. 83, No. 5
snowy periods (e.g., Mech et al. 1995, Hayes et al.
2000). Finally, the rarely studied aspect of wolf–un-
gulate relationships is wolf kill rate in multispecies
prey communities. Most North American studies were
conducted in ecosystems where wolves coexisted with
one or two prey species (e.g., Fritts and Mech 1981,
Gasaway et al. 1983, 1992, Messier 1991). In Eurasia,
however, wolves usually coexist with three or four, and
exceptionally with up to five or six species of ungulates
(Okarma 1995). Moreover, the unique feature of Eur-
asian ungulate communities is the presence of the wild
boar (
Sus scrofa
), which are fertile animals, producing
large numbers of vulnerable young every spring (Je˛d-
rzejewska and Je˛drzejewski 1998). Estimating kill rates
in such communities is difficult, because small prey,
such as wild boar piglets, are usually eaten completely
by wolves. Thus, techniques other than aircraft search
for kill remains must be applied, especially in spring
and summer.
Kill rates of Eurasian wolves have not yet been stud-
ied directly, but data on wolf prey selection and diet
composition suggest that densities of red deer (
Cervus
elaphus
), wolves’ preferred prey, may be governing kill
rates on both deer and other co-occurring species of
ungulates (Je˛drzejewski et al. 2000). The influence of
physical environment, wolf sociality, and other extrin-
sic factors on kill rates of Eurasian wolf remains un-
known.
The role of predation in shaping ungulate popula-
tions has been discussed within the conceptual frame-
work that distinguishes between density-dependent fac-
tors involved in population regulation and other lim-
iting factors that may influence the population rate of
increase but cannot have any regulatory impacts (Sin-
clair 1989, Messier 1991, 1994, 1995, Boutin 1992).
In North America most research done on moose (
Alces
alces
) evidenced that wolf predation rates were density
dependent at low moose densities and changed to de-
pensatory at moderate to high ungulate densities
(Messier 1991, 1994, Hayes and Harestad 2000). Some
research on caribou (
Rangifer tarandus
) showed that
wolf predation rates were inversely density dependent
over the whole range of prey densities (Dale et al.
1995). Furthermore, in multipredator ecosystems, the
overall trend of prey population dynamics depends on
the combined (usually additive) effects of all predators
(e.g., Gasaway et al. 1992, Kunkel and Pletscher 1999).
The role of predation in shaping the numbers of Eu-
ropean ungulates was assessed indirectly, by analysis
of ungulate responses to wolf and lynx (
Lynx lynx
)
control by humans in the series of 100-yr population
dynamics of predators and ungulates (Je˛drzejewska and
Je˛drzejewski 1998). The red deer irrupted in the ab-
sence of wolves, and the roe deer (
Capreolus capreo-
lus
) in the absence of lynxes. Co-occurring European
bison (
Bison bonasus
), moose, and wild boar responded
to predator eradication weakly or not at all. Those un-
intended but well replicated ‘‘experiments’’ provideev-
idence for predation acting as a major limiting factor
for deer populations in Europe. Whether predators can
also regulate their prey remains yet unknown.
The study presented in this paper was conducted in
the last remnant of the natural temperate forests typical
of European lowlands (Białowiez˙a Primeval Forest, Po-
land), where wolves coexist with five species of un-
gulates (European bison, moose, red deer, roe deer,and
wild boar) and with the Eurasian lynx. The aims of our
study were as follows. (1) Based on a three-year (1996–
1999) intensive radio-tracking study of three to four
wolf packs, we endeavored to determine the kill rates
by wolves and to show how they were affected by the
following: a) social features of wolves, such as pack
size and coherence; b) ambient conditions, especially
severity of winter; c) characteristics of prey (prey size
and availability); and d) seasonal and between-year
variation in prey abundance and vulnerability. (2) We
calculated the magnitude of wolf predation on three
species of ungulates, and having combined the new data
with the earlier (1991–1995) information on wolf pre-
dation rates (Je˛drzejewski et al. 2000), we attempted
to explain the role of wolves in the ungulate community
of European temperate forests.
M
ATERIALS AND
M
ETHODS
Study area
Białowiez˙a Primeval Forest (BPF; at present
;
1450
km
2
) is located on the Polish–Belarussian border
(52
8
45
9
N, 24
8
E). Due to centuries-long protection as
a royal forest of Polish kings, Lithuanian dukes, and
Russian czars, BPF is the best preserved woodland of
its size, typical for the European lowland temperate
forests. The Polish part of BPF, where this study was
conducted (580 km
2
), consisted of exploited (480-km
2
)
and protected (100-km
2
) regions. The exploited part
(mean age, 72 yr) had timber harvest, reforestation,
and game hunting. In the protected area (Białowiez˙a
National Park [BNP]), most tree stands were of natural
origin (mean age,
.
100 yr). Neither exploitation of
timber nor game hunting was allowed in BNP. It has
been a Man and Biosphere Reserve of the United Na-
tions Education, Scientific, and Cultural Organization
(UNESCO) since 1977 and a World Heritage Site since
1979.
The most characteristic forest association in BPF is
oak–lime–hornbeam (
Quercus robur
,
Tilia cordata
,
Carpinus betulus
) with admixtures of maple (
Acer pla-
tanoides
) and spruce (
Picea abies
) growing on brown
and podzolic soil. Drier sandy soils are overgrown with
coniferous and mixed-coniferous forest dominated by
pine (
Pinus silvestris
) and spruce with admixtures of
oak. Wet places with stagnated water are covered by
black alder (
Alnus glutinosa
). Vicinities of small forest
rivers and brooks are habitats of riverside forests of
alder and ash (
Fraxinus excelsior
) with admixture of
elm (
Ulmus glabra
). The only open areas within the
May 2002 1343
WOLF KILL RATES AND PREDATION
woodland are marshes of sedges (
Carex
) and reeds
(
Phragmites
) in narrow river valleys (0.1–1 km wide)
and several glades with small villages. There are only
five bituminous roads of a total length
;
50 km acces-
sible for public transportation in the Polish part of BPF.
More information on BPF was given by Falin´ski (1986)
and Je˛drzejewska and Je˛drzejewski (1998).
Of five species of Białowiez˙a’s ungulates, the bison
is a protected species, but its population is kept stable
by yearly culling. The other four species are hunted in
the exploited part of BPF, but not in BNP. In the 1990s,
the mean winter densities of ungulates (all species
pooled) were
;
10 individuals/km
2
, equivalent to
;
820
kg crude biomass per square kilometer (Je˛drzejewski
et al. 2000). The guild of large predators was impov-
erished in the 19th century by extermination of brown
bear (
Ursus arctos
), and currently it contains the wolf
and the lynx. During the past 150 yr, the wolf popu-
lation ranged from temporary extermination due to
predator control to very high densities (Je˛drzejewska
et al. 1996). In the Polish part of BPF, wolves have
been protected since 1989, though some poaching does
occur. In the Belarussian part, wolves are still heavily
hunted. In 1996–1999, 5–16 wolves/yr were shot,
which constituted 10–64% of their estimated winter
numbers. In 1981, a wire fence was constructed in BPF
on the Polish–Soviet state border, but wolves are known
to cross it in some places. In the 1990s, typical home
ranges of wolf packs covered 100–300 km
2
, and wolf
densities were 2–2.6 individuals/100 km
2
(Okarma et
al. 1998).
The climate of BPF is transitional between Atlantic
and continental types, but continental features prevail.
During our study (1996–1999), mean daily temperature
in January was
2
3.9
8
C and 19.1
8
C in July. Snow cover
persisted on the ground for 60–157 d (mean, 96 d), and
maximal depth ranged 13–63 cm in the three study
winters. Mean annual precipitation was 622 mm.
Estimating the kill and consumption rates,
and predation impact by wolves
In 1996–1999, radio-tracking of 12 wolves belong-
ing to four packs was conducted in the Polish part of
BPF. Wolves were captured with ‘‘fladry’’ and nets
(Okarma and Je˛drzejewski 1997) or with foot snare
traps (Aldrich foot snare traps for black bears, modified
by authors). Foot snare traps were equipped with a
radio-alarm system (A. Wagener, Ko¨ln, Germany),
which allowed us to release the animal within 1–2 h
after capture. Wolves were immobilized with 1.2–1.8
mL of a xylazine–ketamine mixture (583 mg of Bayer’s
Rompun [Bayer, Leverkusen, Germany] dissolved in 4
mL of Parke-Davis Ketavet 100 mg/mL [Parke-Davis,
Munich, Germany]) and were fitted with radio collars
(models MOD500 and MOD505, Telonics, Mesa, Ar-
izona, USA; model LM Activity Collar, AVM Instru-
ments, Livermore, California, USA; model LT14-2TS-
HD-MS, Telemetry Systems, Mequon, Wisconsin,
USA; model 9D Advanced Telemetry Systems, Isanti,
Minnesota, USA). Most radio collars were equipped
with head position activity sensors, which helped to
identify if wolves were feeding, resting, or traveling.
Radio-collared wolves were located by following forest
roads with vehicle or bicycle. Sessions of two to nine
days (usually six) of continuous radio-tracking (24 h/d,
usually done by three to six people working in 8-h
shifts) were conducted for eight radio-collared wolves
belonging to four packs. The other four wolves were
tracked for periods too short to yield data suitable for
estimating kill rates, due to dispersal or death from
poaching. During continuous radio-tracking the mean
distance (
6
1
SD
) between the observer and the wolves
was 938
6
577 m, and the presence of an observer had
no effect on wolves’ activity or movements (J. Theuer-
kauf and W. Je˛drzejewski,
unpublished manuscript
).
In 1996–1999, we conducted 43 sessions of contin-
uous radio-tracking, which together covered 234 d
(from 2–9 d/session; mean
6
1
SD
, 5.4
6
1.8) and were
combined with searching for prey remains and wolf
scats. During the sessions, we strove to locate all po-
tential sites of kills and prey consumption by wolves
(based on their activity and duration of stay). After the
wolves had left such places, regular searches for prey
remains and wolf scats were conducted. In winters,
snow tracking was also conducted in an attempt to
cover as long a portion of wolves’ trail as possible.
When prey were found, the species, sex, and age were
determined, percentage consumption by wolves was
assessed visually, and jaws were extracted for a more
detailed age estimation. In 47 red deer killed by wolves,
marrow samples of 2–5 cm
3
were taken from the femur
(
n
5
34), tibia (
n
5
13), metatarsus (
n
5
7), or man-
dible (
n
5
19). The marrow samples were oven-dried
at 50
8
C to constant mass. The percentage of fat in a
sample was calculated as dry mass/fresh mass
3
100
(Neiland 1970). Based on highly significant correlation
between marrow fat in femur and all other bones sam-
pled (
r
, 0.96–0.97,
P
,
,
0.0005–0.002), we scaled the
samples for all deer to represent fat content in the fe-
mur.
Information from radio-tracking, tracks on snow, and
general appearance (freshness) of prey allowed us to
estimate when the prey had been killed by wolves (the
date and often the time of kill). If the prey was esti-
mated to have been killed before our radio-tracking
session, it was also included in the analysis, and an
adequate time (maximally 5 d) was then added to that
series.
As finding prey was usually difficult (due to inad-
equate precision of wolf localization, dense vegetation,
and inaccessible terrain in many places of BPF), we
applied an auxiliary method of prey detection based on
scat analysis. The assumptions were based on the work
by Floyd et al. (1978), who found that wolves defecated
the undigested prey remains 8–56 h (0.3–2.3 d) after
consumption. In BPF, wolves usually consumed alarge
1344
WŁODZIMIERZ JE˛DRZEJEWSKI ET AL.
Ecology, Vol. 83, No. 5
prey within not longer than three days (Je˛drzejewski
et al. 2000), so the time elapsed from killing a deer or
another large prey to depositing scats from the last
meals on that prey would be maximally five days. The
freshness of scats was estimated in the field and verified
by radio-tracking data. All scats found during radio-
tracking sessions and searches for prey were dried,
masses measured, washed through a 0.5-mm sieve, and
then dried and weighed again. Species, age (as adult
or juvenile), and number of prey individuals were iden-
tified whenever possible, based on undigested bone,
hoof, and hair remains according to methods specified
by Pucek (1981) and our own comparative procedures.
Microscope analysis of hair was conducted according
to Debrot et al. (1982) and Teerink (1991).
Based on radio-tracking data and scat freshness, the
date of defecation was estimated for each scat. Then,
in accordance with Floyd et al. (1978), we assumed
that all small prey (newborn deer, wild boar piglets,
hare, fox, beaver, dog) were killed and consumed 8–
56 h (0.3–2.3 d) prior to defecation, but for calculations
we applied a conservative estimate of two days for all
small prey to prevent an overestimate of the killing
rate. In the cases of larger prey detected in scats, we
considered the ‘‘wash-out’’ index, i.e., the ratio of a
scat mass before washing through a sieve to the mass
of washed, undigested prey remains. This index de-
pends on the proportion of amorphic mass (remains
from digestion of meat and other soft tissues), which
the scat contained. The wash-out index can vary from
one (if a scat contained only hard undigested parts,
such as bones and hair) to positive infinity (if wolves
consumed only soft tissues of a freshly killed prey).
To find out which values of the wash-out index indicate
consumption of a newly killed (fleshy) prey, we com-
pared two subsamples of wolf scats, both found in sit-
uations when the wolves were known to have recently
fed on a kill: (1) scats deposited 0.5–2 d after killing
that prey (
n
5
103), and (2) scats deposited
$
3 d after
killing (
n
5
14). Scats in subsample (1) had the wash-
out index varying within the range 1.2–140.7 (mean
6
1
SE
, 8.4
6
1.5), whereas those in subsample (2)
spanned 1.5–7.4, only (3.2
6
0.5). Therefore, we ac-
cepted a conservative criterion that prey recovered
from a scat with wash-out index
.
7 had been killed
and consumed two days prior to scat deposition. A prey
recovered from scat with index 1–6 was assumed to
have been killed five days prior to scat deposition. If
the kill date was estimated to be earlier than the be-
ginning of radio-tracking session, we added the appro-
priate number of days to that session for calculation of
wolf kill rate to prevent its overestimate.
The list of wolf kills found during any session of
radio- and snow tracking was supplemented with the
prey recovered from scats only in those cases when
prey specimens found in scats differed in species or
age from the prey remains already found by us in the
forest. Also, remains of prey of the same species and
age recovered from two or more scats, which had been
found in one session of radio-tracking, were conser-
vatively counted as one prey individual. In all, during
the continuous radio-tracking sessions, we found the
remains of 127 animals killed or scavenged upon by
wolves. In addition, we collected 328 scats. In 188 of
them, undigested prey remains were of the same species
and age class of prey as the prey already found in the
forest. In 174 scats, a total of 90 new prey were re-
vealed, half of them being of small size (e.g., wild boar
piglets, hare), which were completely eaten by wolves.
All prey analyzed (
n
5
217) were estimated to have
been killed (or scavenged upon) by wolves during 323
d, in 43 sessions of 4–15 d (mean
6
1
SD
, 7.5
6
2.1
d); 66 d in 10 sessions represented the spring–summer
season (1 May–30 September), and 257 d in 33 sessions
covered the autumn–winter season (1 October–30
April). Any other prey of wolves found by us or by
the forestry personnel outside the time frames of the
continuous radio-tracking and prey searching were not
included in the calculation of wolf kill rates, but were
added to the whole sample of prey remains found or
recovered from scats, and were used for calculating
species structure of wolf prey.
Kill rates by wolves were estimated by two methods:
(1) based on 46 time intervals between consecutive
kills (only the prey found in the forest were used in
this analysis); and (2) based on wolf kills found in the
forest and prey detected from wolf scats, both estimated
to have been killed and consumed by wolves during a
total of 323 d. In the first method, we considered only
those kills that we were reasonably sure had not been
separated in time by any other undetected prey. The
second method, relying on a much bigger sample than
the first one, gave us an opportunity to analyze vari-
ation in wolf kill rates.
During the field work, tracks of wolves on snow and
mud, cases of howling, and visual observations were
recorded, and attempts were made to count the number
of wolves in each pack and the number of wolves en-
gaged in hunting and utilization of a kill. In 89 kills
(47% of the total sample) an actual size of a hunting
group was determined. In other cases, we assumed that
the whole pack took part in a hunt and consumption
of prey, an exception being the BNP pack in the winter
season 1996–1997. That pack consisted of seven
wolves that had not been observed hunting together;
instead, that pack regularly split into fractions of three
and four individuals.
Food consumption by wolves was estimated based
on the degree of prey utilization by wolves (assessed
visually in the forest), body mass of prey, and mass of
uneaten remains. Body masses of various species,
along with sex and age classes of prey, were takenfrom
the following sources, reporting measurements from
Białowiez˙a Forest or other Polish populations of un-
gulates: Miłkowski (1970), Kozlo (1975), Pucek
(1981), Miłkowski and Wo´jcik (1984), Pielowski
May 2002 1345
WOLF KILL RATES AND PREDATION
F
IG
. 1. Frequency distributions of body masses of 167
prey specimens killed by four wolf packs in Białowiez˙a Pri-
meval Forest in 1996–1999. Means
6
1
SD
reported for red
deer and wild boar.
(1988), Bobek et al. (1992), and Dzie˛ciołowski and
Pielowski (1993).
In 13 cases of prey visually assessed as completely
eaten by wolves (7% of all kills analyzed), we were
able to measure the mass of the discarded parts. Based
on the mass of leftovers, we calculated that in large
prey (body mass
.
65 kg live mass) wolves consumed
65% of their live mass; in medium-sized prey (20–65
kg), 75%; and in small prey (
,
20 kg body mass) a
complete consumption by wolves meant the utilization
of 90% of their live mass. If the prey was visually
assessed as incompletely consumed by wolves (i.e.,
,
100%), we calculated the actual consumption ac-
cordingly.
In the cases of wolves’ scavenging on full corpses
of dead ungulates (bison, domestic cow, and domestic
pig), we estimated that the daily consumption was
;
4.5
kg/wolf (data obtained in the research on scavenging
in BPF; N. Selva, W. Je˛drzejewski,
unpublished data
).
On days when wolves returned to finish up the remains
of their old prey, we estimated (based on the mass of
leftovers,
n
5
6 cases) that they consumed 0.65 kg/
wolf daily.
The number of red deer, roe deer, andwild boar killed
by wolves annually was calculated based on mean kill
rates of each pack in spring–summer and autumn–win-
ter of each year. In twocases where we lacked measures
of kill rates, mean estimates from the same respective
seasons, but from different years, were substituted.
Estimating the abundance of ungulates
Winter densities of red deer, roe deer, and wild boar
come from drive censuses conducted in the whole Pol-
ish part of BPF (580 km
2
), where the wolves were
studied (see details in Je˛drzejewska et al. [1994],
[1997], Kossak [1997], [1998], [1999], Okarma et al.
[1997]). Spring–summer numbers of ungulates were
calculated based on their densities in late winter, per-
centage of adult females in the population, and number
of juveniles per female (details in Je˛drzejewski et al.
[2000]). Also, we obtained the index of ungulate abun-
dance by documenting all observations of animals en-
countered during our fieldwork. Having recorded the
time spent in the forest by human observers, we were
able to calculate the encounter rates of ungulates (num-
ber of animals encountered per one hour spent in the
forest by a human observer). In all, during 1996–1999,
4889 ungulates were seen during 8722 h in the forest.
Meteorological data were obtained from the meteoro-
logical station located in the Białowiez˙a village, which
lies in the center of BPF.
R
ESULTS
Kill rates and consumption of prey by wolves
In 1996–1999, three to four packs of wolves, 15–18
individuals total (numbers collected in winter), lived
in the study area. Altogether, 269 ungulates killed by
wolves were found, mostly red deer (72%) and wild
boar (22%), rarely roe deer (4%), moose (1%), and
domestic cow (1%). Compared to the mean species
structure of wild ungulate community in the three years
(red deer 35.4%, wild boar 37.8%, roe deer 20.8%,
European bison 5.6%, and moose 0.4%), the red deer
appeared strongly selected by wolves, which confirmed
earlier findings (Je˛drzejewska et al. 1994, Je˛drzejewski
et al. 1992, 2000). During 323 d of our observations,
wolves killed 167 new prey, visited 22 remnants of
their own, previously killed prey, and scavenged on
carcasses of 28 dead animals (total
n
5
217). Among
animals killed by wolves during that period (
n
5
189),
we recorded red deer (63%), wild boar (28%), roe deer
(4%), brown hares (2%), beavers (1%), red foxes (1%),
and single specimens of moose, domestic cow, and do-
mestic dog. The wolves scavenged predominantly on
discarded carcasses of domestic animals (cattle and
pigs, 39% of all cases of scavenging,
n
5
28), and
carcasses of wild boar and European bison (32% and
25%, respectively). Based on approximate age of killed
ungulates, we prescribed them to classes of body mass
(Fig. 1). The mean body mass of a red deer killed by
wolves was 93 kg, and that of wild boar 23 kg. Gen-
erally, the mean body mass of all wolf prey was 67.2
kg (1
SD
, 52.2; mode, 63 kg; range, 2–300 kg).
Mean kill rate (estimated based on time intervals
between consecutive kills found by us in the forest)
was, on average, one prey per 1.9 d per pack (1
SD
,
1.5), i.e., 0.526 prey animals·(pack)
2
1
·d
2
1
. Prey-to-prey
time interval ranged from one hour to six days. Fur-
1346
WŁODZIMIERZ JE˛DRZEJEWSKI ET AL.
Ecology, Vol. 83, No. 5
F
IG
. 2. Time elapsed until the next kill in relation to the
body mass of prey. Open circles show time intervals (
n
5
46) between consecutive kills by wolves as documented by
accurately dated prey remains found in the forest. Horizontal
bars and numbers indicate mean values. Frequency distri-
butions of time intervals significantly differed among the four
classes of prey body mass (Kruskal-Wallis ANOVA,
H
5
9.53, df
5
3,
P
5
0.03).
F
IG
. 3. Number of wolves in a hunting group as recorded
at 89 instances of prey killed or carrion scavenged upon,
compared to a total number of wolves in a pack.
T
ABLE
1. Kill rates, scavenging rates, and consumption by wolves (
Canis lupus
) in Białowiez˙a Primeval Forest (eastern
Poland), 1996–1999.
Prey
Kill/scavenging
rate per pack
(no. prey·[pack]
2
1
·d
2
1
)
Mean 1
SD
Kill/scavenging
rate per wolf
(no. prey·[wolf]
2
1
·d
2
1
)
Mean 1
SD
Crude biomass
of prey per pack
(kg·[pack]
2
1
·d
2
1
)
Mean 1
SD
Consumption
rate per wolf
(kg·[wolf]
2
1
·d
2
1
)
Mean 1
SD
Live prey killed by wolves
Red deer
Wild boar
Roe deer
Moose
Cattle
Small prey
Subtotal
0.312
0.147
0.023
0.002
0.002
0.027
0.513
0.157
0.214
0.061
0.010
0.015
0.068
0.261
0.075
0.034
0.006
,
0.001
,
0.001
0.008
0.124
0.044
0.052
0.015
0.002
0.003
0.023
0.073
29.28
3.38
0.56
0.46
0.26
0.28
34.22
15.56
4.62
1.46
3.05
1.68
0.80
14.99
4.31
0.54
0.10
0.06
0.02
0.07
5.10
2.14
0.70
0.27
0.40
0.14
0.20
2.01
Utilization of own, old prey, and scavenging on carrion of dead animals
Old prey
Carrion
Subtotal
0.078
0.086
0.164
0.104
0.121
0.144
0.019
0.021
0.040
0.028
0.029
0.038
3.02
10.59
13.61
4.19
20.16
19.79
0.04
0.44
0.48
0.06
0.81
0.80
Total ··· ··· ··· ··· 47.83 25.10 5.58 2.11
Note:
Small prey included beavers, hares, red foxes, and a domestic dog; carrion were dead European bison, wild boar,
domestic cow, and domestic pig.
thermore, the time that elapsed from wolves’ killing
prey
n
to killing prey
n
1
1 markedly increased with
growing body mass of prey
n
(Fig. 2). The second
method of estimating kill rates (based on kills found
and prey detected from scats) yielded very similar re-
sults (Table 1). On average,each pack killed 0.513 prey
animals/day, including 0.486 ungulates and 0.027 small
animals (e.g., hare, beaver). Those figures are equiv-
alent to one prey killed per two days. As regards var-
ious species of prey, a pack of wolves killed one red
deer every 3.2 d, plus one wild boar every 6.8 d, one
small prey once a month, a roe deer every 44 d, and
one moose or domestic cow once per 8 mo. Moreover,
each pack of wolves visited the remnants of its own
previously killed prey once per 13 d and scavenged on
dead animals once per 12 d (Table 1). During the study,
the mean hunting group included, on average, 4.4
wolves (1
SD
, 1.0; mode, 4; range, 2–6). The appro-
priate per capita kill rates are given in Table 1.
In 1996–1999, the studied packs of wolves varied
from two to seven individuals (Fig. 3). A pair of wolves
was always traveling and feeding together. Packs con-
sisting of three to four wolves split into smaller sub-
groups for 18% of hunts, and those of five to six wolves
split on 41% of hunts. The largest pack (seven wolves)
was not observed hunting together. Instead, it regularly
May 2002 1347
WOLF KILL RATES AND PREDATION
T
ABLE
2. Percentage distribution of prey, in four classes of
body mass, killed by wolf hunting groups of various sizes.
No.
wolves
in group
No. prey
speci-
mens
Percentage of prey by body mass (kg)
1–50 51–100 101–150 151–300
4
5
6
85
34
26
45
26
19
40
44
46
14
15
23
1
15
12
Notes:
Samples of prey killed by fewer than four wolves
were too small for analysis (eight prey killed by one or two
wolves; 14 prey killed by three wolves). Differences in prey
frequencies among wolf hunting groups of various sizes were
statistically significant (
G
5
30.62, df
5
6,
P
,
0.001;
G
test).
F
IG
. 4. Seasonal, between-year, and between-pack variation in wolf kill rates for red deer and wild boar and scavenging
rate in Białowiez˙a Primeval Forest. Abbreviations: AW, autumn–winter (1 October–30 April); SS, spring–summer (1 May–
30 September).
divided into two subgroups (three and four wolves)
each hunting on its own. Still, however, the whole pack
lived and traveled together and held one territory. We
examined the frequency distributions of ungulate prey
killed by groups of various sizes. With increasing size
of a hunting group, wolves killed significantly fewer
small-sized prey and seized large prey more often (Ta-
ble 2).
Red deer, wild boar, and carrion of dead ungulates
made up 95% of wolves’ food biomass, so we analyzed
in detail wolves’ kill rates for deer and boar and their
scavenging rates. The densities of red deer were rather
stable in Białowiez˙a Forest during 1997–1999, which
was reflected by small between-year variation in wolf
hunting for red deer (Fig. 4). However, wolf kill rates
for red deer were constantly lower in spring–summer
(0.211
6
0.04 deer·(pack)
2
1
·d
2
1
, mean
6
1
SE
,) than
in autumn–winter (0.355
6
0.026), the difference being
statistically significant (Mann-Whitney
U
test,
U
5
305.5,
P
5
0.003). This was opposite to the seasonal
pattern of population dynamics of prey: deer numbers
were always higher in summer than in winter, so factors
other than deer number alone must have influenced
wolf hunting rates. Interestingly, the encounter rates of
deer by human observers were also lower in spring–
summer (on average 0.183 deer seen per hour) than in
autumn–winter (0.325 deer/h), which might have re-
sulted from lower visibility of animals in the period of
full development of vegetation. Between-pack varia-
tion in wolves’ kill rates for deer was negligible (Krus-
kal-Wallis ANOVA,
H
5
0.698, df
5
3,
P
.
0.8), but
it should be noted that we had only four packs with
relatively small variation in size (two to six individ-
uals).
An important factor affecting wolf kill rate for red
deer was the depth of snow cover (Table 3). It is worth
noting that it was indeed the depth of snow, and not
the advance of winter, that affected wolf kill rates; in
the analyzed material, snow depth was not correlated
with the consecutive day of winter (November 1
5
day
1;
r
5
0.02,
P
.
0.9). The number of wolves in a
1348
WŁODZIMIERZ JE˛DRZEJEWSKI ET AL.
Ecology, Vol. 83, No. 5
T
ABLE
3. Effects of snow cover and size of wolf hunting group on kill rates and consumption of red deerby wolves; results
of multiple regression analysis and ranking the set of alternative hypotheses.
Dependent variable
Multiple regression analysis
Total
variation
explained
(
R
adj2
)
Contributions by
independent
variables (
r
2
)
Snow
cover (
S
)No. wolves in
group (
N
)
Model selection
(AIC method)
Best model Aikake’s
weight
v
i
No. deer killed per pack per day (DP)
Crude biomass of deer acquired per pack daily (CBP)
No. deer killed per wolf per day (DW)
Deer biomass consumed per wolf daily (BW)
0.218**
0.236**
0.343***
0.365***
(
1
) 0.263***
(
1
) 0.253***
(
1
) 0.264***
(
1
) 0.344***
(
1
) 0.055
(
1
) 0.188*
(
2
) 0.031
(
2
) 0.004
DP
5
f
(
S
,
N
)
CBP
5
f
(
S
,
N
)
DW
5
f
(
S
)
BW
5
f
(
S
)
0.512
0.654
0.513
0.687
Notes:
Variation in snow cover, wolf hunting group, and the four dependent variables are shown in Figs. 5 and 6. The
r
2
values for contributions by each independent variable to the total variation explained (Tabachnick and Fidell 1983) are also
presented. A ‘‘plus’’ symbol (
1
) indicates positive correlation and a ‘‘minus’’ symbol (
2
) indicates negative correlation.
The best model was selected from the set of four alternative models (each of the dependent variables as a function of
S
,
N
,
S
1
N
,or
S
3
N
) by using the Aikake’s information criterion (AIC) method. Aikake’s weight (
v
i
) is the probability that a
given model is the best model in the set (Anderson et al. 2000).
*
P
,
0.05; **
P
,
0.01; ***
P
,
0.001.
F
IG
. 5. Wolf kill rate and daily consumption of red deer in relation to depth of snow cover. Points are values for 43
sessions of observations, each covering 4–15 d. Dotted lines show simple linear regressions. Solid lines and regression
equations show the influence of snow cover with the covarying impact of wolf group size held constant at its mean value
(4.4 wolves). See Table 3 for statistical significance.
hunting group appeared to predict a small amount of
variation in kill rates (Table 3). If we held the group
size constant (at the mean level of 4.4 individuals),
then on average kill rates for deer would increase 2.2-
fold (from 0.264 to 0.587 deer·[pack]
2
1
·d
2
1
) and daily
consumption of deer biomass per wolf 2.4-fold (from
3.5 to 8.7 kg·[wolf]
2
1
·d
2
1
) with snow cover varying 0–
17 cm (Fig. 5). If, in turn, snow cover were held con-
stant (mean depth, 2.53 cm), it appeared that, with wolf
hunting group growing from two to six wolves, the
number of deer killed daily by the group did not in-
crease significantly, but the crude biomass of deer ac-
quired by them grew 2.3-fold (from 16.5 to 37.6
kg·[pack]
2
1
·d
2
1
) (Fig. 6). This was caused by the fact
that larger groups killed large prey more often (see
Table 2). In effect, per capita kill rates and consumption
of deer biomass did not decline significantly among
groups of two to six wolves (Fig. 6).
What was the mechanism of higher kill rates by
wolves in periods with deep snow cover? We tested
whether the physical condition of killed deer, as ap-
proximated by marrow fat content, changed with an
advance of winter and severe ambient conditions. In
adult deer killed by wolves in October–March, fat con-
tent in femurs did not decline significantly (
P
.
0.5)
throughout the cold season (Fig. 7). They all had
.
50%
fat in bone marrow (80%
6
15%, mean
6
1
SD
;
n
5
18). In contrast, fat reserves of young deer (
,
1 yr old)
killed by wolves declined drastically throughout au-
tumn and winter (Fig. 7). Among 18 young deer killed
during October–January, 16 (89%) had
.
50% fat, and
the mean level for all was 66%
6
18% (
n
5
18). Of
May 2002 1349
WOLF KILL RATES AND PREDATION
F
IG
. 6. Wolf kill rates and daily consumption of red deer
in relation to number of wolves in a hunting group. The dotted
line shows a simple linear regression; the solid line and re-
gression equation show the influence of group size with the
co-acting role of snow cover held constant (at its mean value
of 2.53 cm). Statistical significance is given in Table 3. Scat-
terplots with no lines are shown when regressions were sta-
tistically nonsignificant. Other details are as in the Fig. 5
legend.
F
IG
. 7. Marrow fat content in the femur of young and adult red deer killed by wolves in Białowiez˙a Primeval Forest.
Each point denotes one individual, and all animals are plotted against the dates when killed. Scatterplot data of young deer
were smoothed (solid line) by Lowess method (Cleveland 1979) to show the most probable temporal pattern offat depletion.
In adult deer no significant decline in marrow fat content was detected.
11 deer killed in February–March, only 2 (18%) had
marrow fat content exceeding 50%, and mean for all
was 27%
6
21% (
n
5
11). That change was highly
significant (
G
test,
G
5
15.5, df
5
1,
P
,
0.001).
Indeed, the advance of winter (consecutive days;
D
)
and ambient temperature (
T
), with an interaction of
these two factors, explained 46% of the total variation
in the marrow fat (MF) of young deer (
5
0.455,
n
2
R
adj
5
28,
P
,
0.0005). When ranked by Aikake’s infor-
mation criterion (AIC) method (Anderson et al. 2000),
the model MF
5
f
(
D, T, D
3
T
) was the best (Aikake’s
weight
v5
0.671) among seven alternative models
involving consecutive days of winter, snow cover, am-
bient temperature, and the combinations as well as in-
teractions of these factors.
During the study, the wild boar population under-
went fluctuations in numbers, caused primarily by su-
perabundant crop of oak seeds in autumn 1996 (see
Hansson et al. 2000). Wild boar responded to extra food
supply by very early and prolific reproduction in spring
1997, which resulted in high abundance of piglets, a
segment of boar population most vulnerable to wolf
predation (see Fig. 1). However, wolves’ response by
higher kill rates for boar was manifested only in the
pack inhabiting Białowiez˙a National Park (BNP pack;
Fig. 4), where mature oak stands were widespread and
boar densities were much higher than in other parts of
Białowiez˙a Primeval Forest (BPF; see Je˛drzejewska et
al. 1994). Also, from spring 1997 through autumn–
winter of 1997–1998, the encounter rates of wild boar
by human observers were much higher in BNP (0.79
boar/h) than in other parts of BPF (0.13 boar/h).
Encounter rates of wild boar were generally somewhat
higher in spring–summer (on average, 0.307 boar seen
per hour) than in autumn–winter (0.190 boar/h). Cor-
respondingly, wolf kill rates for wild boar were higher
in spring–summer (0.242
6
0.058 boar·[pack]
2
1
·d
2
1
,
1350
WŁODZIMIERZ JE˛DRZEJEWSKI ET AL.
Ecology, Vol. 83, No. 5
F
IG
. 8. Daily food consumption (kg) per wolf in relation to snow cover, ambient temperature, and number of wolves in
a hunting group. Multiple regression analysis:
y
5
5.288
1
0.261(snow)
2
0.042(temp)
2
0.099
N
wolves
;
5
0.374;
n
5
2
R
adj
43,
P
,
0.0005. Snow cover was the most significant factor shaping wolf food intake; the thick line and regression equation
show the influence of snow with the two other independent variables held constant (at their mean values; temperature, 4.0
8
C;
no. wolves, 4.4). Other details are as in the Fig. 5 legend.
mean
6
1
SE
) than in autumn–winter (0.106
6
0.038
boar·(pack)
2
1
·d
2
1
), albeit the difference was not sig-
nificant due to the large variation among various packs
and between years (Fig. 4). Frequency of wild boar
killing by wolves positively correlated with the en-
counter rates of wild boars by human observers (
r
5
0.434,
P
5
0.004), but it was not related to snow cover
or the number of wolves in a hunting group (
P
.
0.6).
Wolf kill rates for red deer and those for wild boar in
the 43 series of intensive several-day samplings were
negatively correlated (
r
52
0.37,
P
5
0.01). Wolves
scavenged on carrion more readily in cold seasons than
in spring–summer (Mann-Whitney
U
test,
U
5
289,
P
5
0.006; Fig. 4). Furthermore, differences among
packs in the frequency of scavenging were significant
(Kruskal-Wallis ANOVA,
H
5
12.07, df
5
3,
P
5
0.007).
The total daily food intake by wolves was extremely
variable (Fig. 8). In multiple regression analysis, two
environmental factors (depth of snow cover and am-
bient temperature) and a social factor (size of wolf
hunting group) explained 37% of the total variation
observed. If we stabilized the covarying factors and
separated the sole impact of each of the three factors,
it appeared that snow cover had the strongest influence:
mean daily food consumption by a wolf increased by
80% (from 4.9 to 9.4 kg·[wolf]
2
1
·d
2
1
) with snow depth
ranging 0–17 cm (Fig. 8). Ambient temperature had
weaker impact; in the observed range of mean tem-
peratures from 20
8
Cto
2
10
8
C, daily food intake by
wolves increased by 26% (from an average of 4.9 to
6.2 kg·[wolf]
2
1
·d
2
1
; Fig. 8). The effect of group size
on wolf daily rations was negligible. The leading role
of snow in shaping the variation in daily food intake
by wolves was supported by the AIC method (Anderson
et al. 2000), which showed that food intake as a func-
tion of snow cover had the highest probability of being
the best model (Aikake’s weight
v5
0.459) among
May 2002 1351
WOLF KILL RATES AND PREDATION
T
ABLE
4. Predation impact by wolves on red deer, roe deer, and wild boar in Białowiez˙a
Primeval Forest (BPF) in 1991–1999, in relation to densities and population increases of
prey (estimated no. young born annually).
Year†
Ungulate density
(no. individuals/100 km
2
)
Late winter Spring–
summer
Annual
predation
by wolves
(no. killed/
100 km
2
)
Annual predation on ungulates
(%)‡
Spring
density Annual
increase Annual
mortality
Red deer (
Cervus elaphus
)
1991–1992
1992–1993
1993–1994
1995–1996
1997–1998
1998–1999
1999–2000
1991–2000
607
416
359
463
286
296
337
395
6
113
858
581
514
659
440
458
515
575
6
145
105
78
57
62
57
72
···
72
6
18
12
13
11
9
13
16
···
12
6
2
42
47
37
32
37
45
···
40
6
6
24
35
···
···
40
60
···
40
6
15
Roe deer (
Capreolus capreolus
)
1991–1992
1992–1993
1993–1994
1995–1996
1997–1998
1998–1999
1999–2000
1991–2000
492
425
288
324
115
182
240
295
6
132
820
702
501
517
229
313
414
499
6
208
25
21
21
19
5
4
···
16
6
9
3
3
4
4
2
1
···
3
6
1
8
8
10
10
4
3
···
7
6
3
6
5
···
···
11
5
···
7
6
3
Wild boar (
Sus scrofa
)
1991–1992
1992–1993
1993–1994
1995–1996
1997–1998
1998–1999
1999–2000
1991–2000
458
209
308
225
217
426
338
312
6
102
678
406
599
437
486
617
577
543
6
101
28
19
38
36
38
26
···
31
6
8
4
5
6
8
8
4
···
6
6
2
13
10
13
17
14
14
···
13
6
2
6
19
···
···
63
9
···
24
6
26
Notes:
Data for 1991–1996 were presented by Jedrzejewski et al. (2000); data for 1997–
1999 represent the current study. No data were available for 1994–1995 and 1996–1997.
† Measurements span March of the first year until the following February. Year entries of
1991–2000 indicate means
6
1
SD
for the entire period of study, but see
Notes
.
‡ Predation values are given as percentages of spring density, annual increase, and annual
mortality, respectively.
seven alternative models testing the effects of pack
size, snow cover, temperature, as well as combinations
and interactions of these factors.
Predation by wolves in relation to ungulate densities
Data on predation impact by wolves on red deer, roe
deer, and wild boar were available for the seven years
1991–1999 (Table 4; three years from this study and
four years from Je˛drzejewski et al. [2000]). In 1991/
1992–1995/1996, ungulates were subject to strongly
elevated hunting harvest aimed at substantial reduction
of their numbers. Since 1996/1997, the harvest was
markedly lower (Kossak 1999). The wild boar dynam-
ics was affected by two years of superabundant crop
of oak seeds (in 1989 and 1996, which resulted in high
densities in late winters and spring–summer seasons of
1990 and 1997, respectively).
Wolf predation was most important in the case of
red deer. Annually wolves took, on average, 72 deer
from 100 km
2
, which constituted 12% of spring (sea-
sonally highest) densities of deer, and was equivalent
to 40% of their annual increase due to breeding, and
40% of the annual mortality of deer (Table 4). Within
the observed range of red deer abundance, the number
of deer killed by wolves per annum was positively
related to deer densities (
5
0.556,
n
5
6yr,
P
5
2
R
adj
0.05 for spring densities, and
5
0.517,
P
5
0.06
2
R
adj
for winter densities). In 1991–2000, we observed a
two-fold decline of deer numbers (difference between
max and min recorded numbers:
G
5
118.03, df
5
1,
P
,
0.001,
G
test), followed by smaller (1.5–1.8-fold)
decline in the number of deer killed by wolves annu-
ally. In effect, percentage predation (number of deer
killed by wolves annually as percent of deer density)
showed a tendency to decline with growing density of
red deer, although the relationship was not statistically
significant (
r
52
0.404,
P
.
0.4 for spring densities;
r
52
0.542,
P
.
0.2 for late winter densities of deer).
In red deer, the estimated number of young born
annually increased with growing population density
(
5
0.847,
P
5
0.006), but the rate of population
2
R
adj
increase (number of young as percent of winter density)
1352
WŁODZIMIERZ JE˛DRZEJEWSKI ET AL.
Ecology, Vol. 83, No. 5
F
IG
. 9. Left panel: Annual population increase of red deer due to reproduction (no. young born per 100 km
2
per year),
and increase minus annual predation by wolves (number of deer killed per 100 km
2
per year), in relation to late winter
densities of red deer population. Right panel: percentage population increase of red deer in relation to late winter densities.
Annual predation by wolves is marked as a factor reducing increase rates of prey populations. Solid lines are mathematically
derived from the appropriate regression equations in the left panel. Dotted lines mark the hypothetical relationships beyond
the empirical range of prey densities. The bifurcated curve shows the range of all possible outcomes of wolf–deer interactions
(see
Results: Predation by wolves
for further explanations).
was inversely density dependent (
5
0.442,
P
5
2
R
adj
0.09). Wolf predation on deer (number of deer killed)
was positively related to deer population increase (
r
5
0.826,
P
5
0.04), and wolves exploited a fairly stable
proportion (32–47%) of deer population increase (Fig.
9). Thus, wolves notably lowered the rate of population
increase in red deer.
Annually wolves took, on average, 16 roe deer/100
km
2
, a number constituting only 3% of the roe deer
spring densities, 7% of their annual increase, and 7%
of the average yearly mortality (Table 4). During the
observed 4.3-fold decline of roe deer numbers (differ-
ence between maximum and minimum recorded den-
sities:
G
5
252.16, df
5
1,
P
,
0.001), the number of
roe deer killed by wolves per year declined faster (over
six times). This produced a positive correlation be-
tween the number of deer killed by wolves per annum
and roe deer densities (
5
0.789,
n
5
6yr,
P
5
2
R
adj
0.01 for spring densities; and
5
0.783,
P
5
0.01
2
R
adj
for winter densities). However, percentage predation
on roe deer was always small with not much variation
(1–4%), and it was not related to roe deer density (
P
.
0.4). In roe deer, the estimated number of young
born annually grew with increasing density of popu-
lation (
5
0.953,
P
5
0.001), but the percentage
2
R
adj
rate of increase showed a tendency to be inversely den-
sity dependent (
5
0.419,
P
5
0.1). Predation,
2
R
adj
amounting to 3–10% of roe deer annual production of
young, could only slightly lower the rate of population
increase in roe deer.
From the wild boar population, wolves annually took
a mean value of 31 animals/100 km
2
, which constituted
6% of the average spring densities of wild boar, and
was equivalent to 13% of their yearly increase due to
reproduction and 24% of their annual mortality (Table
4). The number of wild boar killed by wolves annually
was not related to boar densities (
P
.
0.9 for spring
densities, and
P
.
0.6 for winter densities), but it
showed a tendency to grow with higher annual pro-
duction of young (
r
5
0.78,
n
5
6yr,
P
5
0.07).
Percentage predation was inversely related to densities,
although the relationship was not significant (
r
5
2
0.527,
P
.
0.2 for spring densities; and
r
52
0.783,
P
5
0.07 for winter densities of wild boar). Wolves
took a fairly stable and small proportion of wild boar
annual production of young (10–17%).
In summary, wolves exerted the strongest impact on
red deer population and much smaller impacts on roe
deer and wild boar. In all species, however, percentage
predation was either inversely density dependent or did
not vary with changing densities of prey.
D
ISCUSSION
Kill rate estimates from Białowiez˙a cannot be easily
compared to measures obtained in numerous North
May 2002 1353
WOLF KILL RATES AND PREDATION
American studies (because kill rates strongly varied
with prey size and wolf pack size), but per capita
amount of food eaten by wolves can be compared. The
value from Białowiez˙a Primeval Forest (BPF; 5.6
kg·(wolf)
2
1
·d
2
1
) is similar to or higher than those re-
ported elsewhere (4.4–6.3 kg·[wolf]
2
1
·d
2
1
, Mech
[1966]; 0.5–7.0 kg·[wolf]
2
1
·d
2
1
, Fritts and Mech
[1981]; 1.6–2.8 kg·[wolf]
2
1
·d
2
1
, Messier and Crete
[1985]; 4.5–14.9 kg·[wolf]
2
1
·d
2
1
, Ballard et al. [1987];
2 kg·[wolf]
2
1
·d
2
1
, Fuller [1989]; 4.4–7.8 kg·[wolf]
2
1
·d
2
1
,
Thurber and Peterson [1993]; 4.1–6.4 kg·[wolf]
2
1
·d
2
1
,
Hayes et al. [2000]). According to Schmidt and Mech
(1997), the lowest of the values listed here represent
underestimates. In the United States and Canada, data
on predation rates were usually collected by locating
wolf kills from aircraft. The accuracy of that method
strongly depends on visibility (the aerial search for prey
would be inapplicable in closed forests) and prey size
(large prey, such as moose, are more probable to be
detected than small prey). In our study area, a sub-
stantial number of prey were those of small size, often
completely eaten by wolves. Searching for prey re-
mains alone would not reveal the true kill rate, espe-
cially in spring–summer. By combining ground radio-
tracking, prey searching, and identification of prey in
scats, we were able to better estimate the kill rate and
its seasonal variation. As many as 41% of wolf prey
were traced out in scats only, and most of them were
small prey (e.g., wild boar piglets, young deer, bea-
vers). Although this method may still somewhat un-
derestimate the kill rates (two or more prey of the same
species and age, killed and consumed by wolves in a
short time sequence, would be counted as one prey if
recovered from scats only), it is highly recommended
or even necessary to supplement the search for prey
remains with the analysis of scats in studies on wolf
predation in dense woodlands and in regions where
wolves consume many small and medium-sized un-
gulates (roe deer, piglets of wild boar). Our method
also yielded reliable data for all seasons, and not only
for the period with snow cover. Higher kill rates in
winter than in summer did not result from higher prob-
ability of detecting prey on days with snow. First, kill
rates on deer increased but those on wild boar declined
during winter. Secondly, wolf kill rates were similar in
seasons with no snow (spring, summer, autumn) and
in winter periods with shallow snow cover, which al-
ready enabled us to do snow tracking.
In North America, major factors affecting kill rates
by wolves were prey density, depth of snow cover, and
wolf pack size. Having reviewed more than a dozen
studies on wolf kill rates on moose, Messier (1994)
found that per capita kill rate was positively related to
moose abundance within the observed densities of 0.2–
2.5 moose/km
2
(
;
40–500 kg crude biomass/km
2
).
Compared to those data, winter densities of ungulates
in Białowiez˙a during our study were much higher and
less variable (6.2–10 individuals/km
2
and 500–800 kg/
km
2
, three species pooled). Interestingly, the density of
wolves in BPF (2–2.6 wolves/100 km
2
) was within the
range of densities cited by Messier (1994) for the Unit-
ed States and Canada (0.2–7.8 wolves/100 km
2
, mean
1.75). Thus, we think that in our study wolves were
not limited by prey resources. Low rates of wolves’
scavenging on carrion (although carrion was very abun-
dant in the cold season; Je˛drzejewska and Je˛drzejewski
1998; N. Selva and W. Je˛drzejewski,
unpublished data
)
support this conclusion. We did observe, however,
some variation in wolf kill rates due to changing prey
availability. Wolves killed more wild boar in a year
when piglets were very abundant, and nearly left off
hunting roe deer when roe deer numbers declined.
These cases, however, concerned secondary prey of
wolves and not their main prey, red deer.
Positive influence of deep snow cover on wolf kill
rates is well documented. Huggard (1993) reported that
in Banff National Park, Canada, kill rates by wolves
increased from 0.185 wapiti/d in periods with no snow
to 0.910 wapiti/d when snow was 60 cm deep. Peterson
and Allen (1974), who observed the same phenomenon
in a wolf–moose system in Isle Royale, USA, proposed
that lower mobility of moose and, thus, their poorer
access to forage in deep snow brought about a higher
incidence of malnutrition (especially in calves) and, in
consequence, greater vulnerability to predation. On the
other hand, Post et al. (1999) showed that the mean
size of wolf packs increased during snowy winters, and
Fuller (1991) evidenced that in severe, snowy winters
wolves spent more time with other pack members than
during mild winters. As kill rates per pack normally
increase with pack size (Thurber and Peterson 1993,
Schmidt and Mech 1997), the observed relationship
between snow depth and wolf kill rates can result from
two factors: higher vulnerability of prey and stronger
coherence of wolf packs in periods with deep snow
cover.
In this study, we attempted to disentangle the roles
of snow and pack size. In Białowiez˙a Primeval Forest,
snow appeared the most important factor affecting wolf
kill rates on deer, and it acted through greater vulner-
ability to predation of young deer with depleted fat
reserves. Earlier studies on food habits of wolves in
Białowiez˙a Forest documented that the share of red
deer in wolf diet increased in severe, snowy winters
(Je˛drzejewski et al. 1992). These findings are in agree-
ment with other studies on European red deer, which
showed that even in the absence of large predators,
severe winter is a critical time for calf survival. In
Norway, Loison et al. (1999) found that body masses
of red deer calves decreased with increasing snow
depth in winter. Okarma (1991) revealed that in Biesz-
czady Mountains (southeastern Poland), fat reserves of
red deer calves (but not adults) killed by wolves were
markedly smaller in late winter (February–March) than
in early winter (December–January).
In Białowiez˙a, per capita kill rates by wolves de-
1354
WŁODZIMIERZ JE˛DRZEJEWSKI ET AL.
Ecology, Vol. 83, No. 5
clined slightly (and not significantly) with the growing
size of wolf hunting group, whereas the amount of food
acquired per wolf practically did not decline, because
larger groups tended to select large-sized prey more
often than smaller groups. This finding is intriguing,
as the studies conducted in North America showed that
albeit large packs exhibited higher kill rate, the amount
of food acquired per wolf declined significantly with
growing pack size (Thurber and Peterson 1993,
Schmidt and Mech 1997, Hayes et al. 2000). There are,
however, two essential differences between wolf situ-
ation in European temperate forests and in the boreal
zone of North America. First, in European deciduous
and mixed forests, wolves coexist with and prey on
three to five species of ungulates ranging in size from
,
2 kg (wild boar piglets in spring) to
.
300 kg (moose,
European bison) (review in Okarma [1995]). Thus, the
richness of species and size classes of prey allow for
great flexibility in prey choice and kill rates by wolves
in order to satisfy food demands of all pack members.
Second, the mean size of wolf hunting groups (4.4
wolves in this study) and the maximal recorded pack
size (seven wolves) were small compared to North
American packs, which usually comprised more than
six (and up to 20) individuals (Thurber and Peterson
1993, Schmidt and Mech 1997, Hayes et al. 2000).
We propose that it is the size of the most frequently
taken prey that sets the upper limit for wolf pack size.
Wolves neither guard nor hide their kill (as the lynx
do; Je˛drzejewski et al. 1993), so the optimal utilization
of a kill (i.e., with minimal losses to scavengers) would
be to consume the kill immediately. In Białowiez˙a, a
63-kg deer calf, the most common prey of wolves,
would be eaten completely by four to five wolves with-
in few hours. By contrast, a killed moose would support
a pack of 15–20 wolves with daily food rations. This
explains why packs of six to seven wolves began to
disintegrate in Białowiez˙a Forest, and packs of more
than eight individuals were practically not recorded
during recent decades (Je˛drzejewska et al. 1996). In
BPF, scavenging on ungulate carcasses (by
.
30 species
of birds and mammals) was widespread, and scavengers
were able to usurp a fair proportion of wolf kills (Je˛d-
rzejewska and Je˛drzejewski 1998; N. Selva and W. Je˛-
drzejewski,
unpublished data
). Thus we think that
group living in wolves cannot be explained by simply
easier acquisition of larger prey by bigger groups. (In
Białowiez˙a, wolves were able to seize adult male deer
even when hunting singly.) Instead, we propose that
the upper limit for a social group size is set by the
need to optimally utilize a dominant class of prey.
What is the role of wolf predation in functioning of
ungulate populations, especially red deer, in European
temperate forests? There is historical evidence that, in
BPF, wolves and lynxes limited deer numbers (Je˛drze-
jewska et al. 1997, Je˛drzejewska and Je˛drzejewski
1998). During the period 1850–2000, wolves had been
exterminated twice, and in both periods red deer pop-
ulation irrupted and reached densities at which further
growth was halted by intraspecific competition for
food. Furthermore, in the whole series of
.
100-yr data,
the percentage population increase of deer from year
to year was negatively correlated with combined den-
sities of wolves and lynxes (Je˛drzejewska et al. 1997).
Notably, however, wolf numbers were negatively cor-
related with red deer densities (Je˛drzejewska and Je˛-
drzejewski 1998), a situation opposite to those reported
for wolf–moose or, generally, wolf–ungulate systems
in North America, where positive correlation of wolf
numbers and ungulate density or biomass was found
(Keith 1983, Fuller 1989, Messier 1994). In BPF, the
negative relationship resulted from human impelling
the fluctuations of wolves (via control) and partly also
deer (by heavy poaching in years of political instability
and economic regress) (Je˛drzejewska et al. 1997).
Worth noting is the fact that Białowiez˙a’s wolves
were able to maintain high densities (up to 7–9 indi-
viduals/100 km
2
) even when red deer were scarce or
temporarily absent from the community (Je˛drzejewska
et al. 1996). This situation may only occur where
wolves are sustained by several prey species, and then,
as suggested by Seip (1992) for a wolf–moose–caribou
system, the most vulnerable prey species may suffer
heavy depensatory predation or even become com-
pletely eliminated. In European temperate forests, wild
boar seem very important in maintenance of wolf pop-
ulations. Piglets and young boar are the most important
alternative prey to wolves, and they are most numerous
and vulnerable to predation in spring–summer, when
wolves have pups.
Messier (1994), having analyzed moose–wolf inter-
actions over a broad spectrum of moose densities in
North America, concluded that percentage predation
by wolves was density dependent in the range 0–0.65
moose/km
2
and inversely density dependent at higher
moose densities. Based on a detailed analysis of the
historical data of wolf and deer population numbers in
BPF, Je˛drzejewska and Je˛drzejewski (1998) proposed
that, with a negative numerical relationship between
wolf and deer and a type II (logarithmic) functional
response, percentage predation by wolves on deer
should be density dependent at low densities of deer
(
,
1 individual/km
2
) and inversely density dependent
at higher densities. In this study, as expected, per-
centage predation by wolves was inversely density de-
pendent over deer densities 3–6 individuals/km
2
. Thus,
wolves limited deer numbers, but were unable to reg-
ulate their prey (sensu Sinclair 1989, Messier 1991).
By eliminating a fair number of deer relative to their
annual production, wolves can hamper population
growth of red deer and prolong the time until they reach
the carrying capacity of habitat.
Empirical data on wolf–deer interactions at low den-
sities of deer are not yet available. The range of pos-
sible outcomes of those interactions is shown by a bi-
furcated curve in Fig. 9. They can vary from heavy,
May 2002 1355
WOLF KILL RATES AND PREDATION
depensatory predation, leading eventually to elimina-
tion of deer by wolves (descending section of a curve
in Fig. 9), to declining predation rates at low densities
of deer ensuring coexistence of predator and prey (as-
cending section of a curve, Fig. 9). We think that in
the Palearctic region, where deer is the most preferred
prey to wolves (Okarma 1995), both situations may
happen, depending on a set of extrinsic factors. The
most important of them are as follows: (1) abundance
of alternative prey species, which can support high den-
sities of wolves, (2) unfavorable climatic conditions
(severe winters), and (3) heavy exploitation by humans,
which causes elevated mortality of deer, additive to that
inflicted by wolves.
Yet wolf predation alone is a poor predictor of deer
population dynamics in BPF. Je˛drzejewski et al. (2000)
and Okarma et al. (1997) documented that predation
by lynx and hunting harvest by humans were also im-
portant factors limiting deer numbers in Białowiez˙a
Forest. Those three agents of deer mortality were ad-
ditive (Je˛drzejewski et al. 2000). This study docu-
mented that only in the case of deer fawns were wolf
predation and mortality caused by severe winter con-
ditions found to be partly compensatory. Wolves killed
fawns with very poor fat reserves, some of which might
have died anyway.
A
CKNOWLEDGMENTS
This study was financed by the Polish State Committee for
Scientific Research (grant 6 P04F 026 12), the Mammal Re-
search Institute budget, the European Natural Heritage Fund
(Euronatur), the Flaxfield Nature Consultancy (The Nether-
lands and UK), the German Academic Exchange Service (to
J. Theuerkauf), and the German Donor’s Association for the
Promotion of Sciences and Humanities (to J. Theuerkauf).
Permissions to capture and radio-collar wolves were issued
by the Ministry of Forestry and Nature Protection, and the
Director of Białowiez˙a National Park. We thank all the as-
sistants and volunteers, who helped us during the field work.
L
ITERATURE
C
ITED
Anderson, D. R., K. P. Burnham, and W. L. Thompson. 2000.
Null hypotheses testing: problems, prevalence, and an al-
ternative. Journal of Wildlife Management 64:912–923.
Ballard, W. B., J. S. Whitman, and C. L. Gardner. 1987.
Ecology of an exploited wolf population in south-central
Alaska. Wildlife Monographs 98:1–54.
Bobek, B., K. Morow, K. Perzanowski, and M. Kosobucka.
1992. The red deer (
Cervus elaphus
)—its ecology and man-
agement. [In Polish.] Wydawnictwo S
´wiat, Warszawa, Po-
land.
Boutin, S. 1992. Predation and moose population dynamics:
a critique. Journal of Wildlife Management 56:116–127.
Boyce, M. S. 2000. Modeling predator–prey dynamics. Pages
253–287
in
L. Boitani and T. K. Fuller, editors. Research
techniques in animal ecology—controversies and conse-
quences. Columbia University Press, New York, New York,
USA.
Cleveland, W. S. 1979. Robust locally weighted regression
and smoothing scatterplots. Journal of the American Sta-
tistical Association 74:829–836.
Dale, B. W., L. D. Adams, and R. T. Bowyer. 1994. Functional
response of wolves preying on barren-ground caribou in a
multiple-prey ecosystem. Journal of Animal Ecology 63:
644–652.
Dale, B. W., L. D. Adams, and R. T. Bowyer. 1995. Winter
wolf predation in a multiple ungulate prey system, Gates
of the Arctic National Park, Alaska. Pages 223–230
in
L.
N. Carbyn, S. H. Fritts, and D. R. Seip, editors. Ecology
and conservation of wolves in a changing world. Canadian
Circumpolar Institute, Occasional Publication 35. Edmon-
ton, Alberta, Canada.
Debrot, S., C. Mermod, G. Fivaz, and J.-M. Weber. 1982.
Atlas des poils de mammiferes d’Europe. Universite de
Neuchatel, Peseux, Switzerland.
Dzie˛ ciołowski, R., and Z. Pielowski. 1993. Łos´ [The Moose].
Wydawnictwo Anton-5, Warszawa, Poland. [In Polish.]
Falin´ski, J. B. 1986. Vegetation dynamics in temperate low-
land primeval forest. Dr W. Junk Publishers, Dordrecht,
The Netherlands.
Floyd, T. J., L. D. Mech, and P. A. Jordan. 1978. Relating
wolf scat content to prey consumed. Journal of Wildlife
Management 42:528–532.
Fritts, S. H., and L. D. Mech. 1981. Dynamics, movements,
and feeding ecology of a newly protected wolf population
in northwestern Minnesota. Wildlife Monographs 80:1–79.
Fuller, T. K. 1989. Population dynamics of wolves in North-
Central Minnesota. Wildlife Monographs 105:1–41.
Fuller, T. K. 1991. Effect of snow depth on wolf activity and
prey selection in north central Minnesota. Canadian Journal
of Zoology 69:283–287.
Gasaway, W. C., R. D. Boertje, D. V. Grangaard, D. G. Kel-
leyhouse, R. O. Stephenson, and D. G. Larsen. 1992. The
role of predation in limiting moose at low densities in Alas-
ka and Yukon and implications for conservation. Wildlife
Monographs 120:1–59.
Gasaway, W. C., R. O. Stephenson, J. L. Davis, P. E. K.
Shepherd, and O. E. Burris. 1983. Interrelationships of
wolves, prey, and man in interior Alaska. Wildlife Mono-
graphs 84:1–50.
Hansson, L., B. Je˛drzejewska, and W. Je˛drzejewski. 2000.
Regional differences in dynamics of bank vole populations
in Europe. Polish Journal of Ecology 48(Supplement):163–
177.
Hayes, R. D., A. M. Baer, U. Wotschikowsky, and A. S.
Harestad. 2000. Kill rate by wolves on moose in the Yukon.
Canadian Journal of Zoology 78:49–59.
Hayes, R. D., and A. S. Harestad. 2000. Wolf functional
response and regulation of moose in the Yukon. Canadian
Journal of Zoology 78:60–66.
Holling, C. S. 1959. The components of predation as revealed
by a study of small-mammal predation of the European
pine sawfly. Canadian Entomologist 91:293–320.
Huggard, D. J. 1993. Effect of snow depth on predation and
scavenging by gray wolves. Journal of Wildlife Manage-
ment 57:382–388.
Je˛drzejewska, B., and W. Je˛drzejewski. 1998. Predation in
vertebrate communities. The Białowiez˙a Primeval Forest
as a case study. Springer Verlag, Berlin, Germany.
Je˛drzejewska, B., W. Je˛drzejewski, A. N. Bunevich, L.
Miłkowski, and Z. A. Krasin´ski. 1997. Factors shaping
population densities and increase rates of ungulates in
Białowiez˙a Primeval Forest (Poland and Belarus) in the
19th and 20th centuries. Acta Theriologica 42:399–451.
Je˛drzejewska, B., W. Je˛drzejewski, A. N. Bunevich, L.
Miłkowski, and H. Okarma. 1996. Population dynamics of
wolves
Canis lupus
in Białowiez˙a Primeval Forest (Poland
and Belarus) in relation to hunting by humans, 1847–1993.
Mammal Review 26:103–126.
Je˛drzejewska, B., H. Okarma, W. Je˛drzejewski, and L.
Miłkowski. 1994. Effects of exploitation and protection on
forest structure, ungulate density and wolf predation in
Białowiez˙a Primeval Forest, Poland. Journal of Applied
Ecology 31:664–676.
Je˛drzejewski, W., B. Je˛drzejewska, H. Okarma, and A. L.
1356
WŁODZIMIERZ JE˛DRZEJEWSKI ET AL.
Ecology, Vol. 83, No. 5
Ruprecht. 1992. Wolf predation and snow cover as mor-
tality factors in the ungulate community of the Białowiez˙a
National Park, Poland. Oecologia (Berlin) 90:27–36.
Je˛drzejewski, W., B. Je˛drzejewska, H. Okarma, K. Schmidt,
K. Zub, and M. Musiani. 2000. Prey selection and pre-
dation by wolves in Białowiez˙a Primeval Forest, Poland.
Journal of Mammalogy 81:186–202.
Je˛drzejewski, W., K. Schmidt, L. Miłkowski, B. Je˛drzejews-
ka, and H. Okarma. 1993. Foraging by lynx and its role
in ungulate mortality: the local (Białowiez˙a Forest) and the
Palaearctic viewpoints. Acta Theriologica 38:385–403.
Keith, L. B. 1983. Population dynamics of wolves. Pages
66–77
in
L. N. Carbyn, editor. Wolves in Canada and Alas-
ka: their status, biology, and management. Canadian Wild-
life Service Report Studies No. 45. Edmonton, Canada.
Kossak, S. 1997. Wildlife management and game protection
in Białowiez˙a Primeval Forest in 1991–1997. [In Polish
with English summary.] Sylwan 141(9):55–61;141(11):67–
81.
Kossak, S. 1998. Raport o stanie zwierzyny w Puszczy
Białowieskiej w pierwszym kwartale 1998 roku. [In Pol-
ish.] Forest Research Institute, Białowiez˙a, Poland.
Kossak, S. 1999. Raport o stanie zwierzyny w Puszczy
Białowieskiej w pierwszym kwartale 1999 roku. [In Pol-
ish.] Forest Research Institute, Białowiez˙a, Poland.
Kozlo, P. G. 1975. Dikii kaban [The wild boar]. [In Russian.]
Izdatelstvo Uradzhai, Minsk, Belarus Republic.
Kunkel, K. E., and D. H. Pletscher. 1999. Species-specific
population dynamics of cervids in a multipredator ecosys-
tem. Journal of Wildlife Management 63:1082–1093.
Loison, A., R. Langvatn, and E. J. Solberg. 1999. Body mass
and winter mortality in red deer calves: disentangling sex
and climate effects. Ecography 22:20–30.
Mech, L. D. 1966. The wolves of Isle Royale. Fauna of the
National Parks of the United States, Fauna Series 7:1–210.
Mech, L. D., T. J. Meier, J. W. Burch, and L. G. Adams. 1995.
Patterns of prey selection by wolves in Denali National
Park, Alaska. Pages 231–243
in
L. N. Carbyn, S. H. Fritts,
and D. R. Seip, editors. Ecology and conservation of
wolves in a changing world. Canadian Circumpolar Insti-
tute, Occasional Publication 35. Edmonton, Alberta, Can-
ada.
Messier, F. 1991. The significance of limiting and regulating
factors on the demography of moose and white-tailed deer.
Journal of Animal Ecology 60:377–393.
Messier, F. 1994. Ungulate population models with predation:
a case study with the North American moose. Ecology 75:
478–488.
Messier, F. 1995. On the functional and numerical responses
of wolves to changing prey density. Pages 187–197
in
L.
N. Carbyn, S. H. Fritts, and D. R. Seip, editors. Ecology
and conservation of wolves in a changing world. Canadian
Circumpolar Institute, Occasional Publication 35. Edmon-
ton, Alberta, Canada.
Messier, F., and M. Crete. 1985. Moose–wolf dynamics and
the natural regulation of moose populations. Oecologia
(Berlin) 65:503–512.
Miłkowski, L. 1970. Łowiectwo w Puszczy Białowieskiej.
Przewodnik dla mys´liwych [Hunting in Białowiez˙a Pri-
meval Forest. Guide for hunters]. [In Polish.] Wojewo´dzki
Os´rodek Informacji Turystycznej, Białystok, Poland.
Miłkowski, L., and J. M. Wo´jcik. 1984. Structure of wild
boar harvest in the Białowiez˙a Primeval Forest. Acta Ther-
iologica 29:337–347.
Neiland, K. A. 1970. Weight of dried marrow as indicator of
fat in caribou femurs. Journal of Wildlife Management34:
904–907.
Okarma, H. 1991. Marrow fat content, sex and age of red
deer killed by wolves in winter in the Carpathian Moun-
tains. Holarctic Ecology 14:169–172.
Okarma, H. 1995. The trophic ecology of wolves and their
predatory role in ungulate communities of forest ecosys-
tems in Europe. Acta Theriologica 40:335–386.
Okarma, H., and W. Je˛drzejewski. 1997. Livetrapping wolves
with nets. Wildlife Society Bulletin 25:78–82.
Okarma, H., W. Je˛drzejewski, K. Schmidt, R. Kowalczyk,
and B. Je˛drzejewska. 1997. Predation of Eurasian lynx on
roe deer and red deer in Białowiez˙a Primeval Forest, Po-
land. Acta Theriologica 42:203–224.
Okarma, H., W. Je˛drzejewski, K. Schmidt, S. S
´niez˙ko, A. N.
Bunevich, and B. Je˛drzejewska. 1998. Home ranges of
wolves in Białowiez˙a Primeval Forest, Poland, compared
with other Eurasian populations. Journal of Mammalogy
79:842–852.
Peterson, R. O., and D. L. Allen. 1974. Snow conditions as
a parameter in moose–wolf relationships. Naturaliste can-
adien 101:481–492.
Pielowski, Z. 1988. Sarna [The roe deer]. [In Polish.] Pan´-
stwowe Wydawnictwo Rolnicze i Les´ne, Warszawa, Po-
land.
Post, E., R. O. Peterson, N. Ch. Stenseth, and B. E. McLaren.
1999. Ecosystem consequences of wolf behavioural re-
sponse to climate. Nature 401:905–907.
Pucek, Z., editor. 1981. Keys to vertebrates of Poland. Mam-
mals. PWN—Polish Scientific Publishers, Warszawa, Po-
land.
Schmidt, P. A., and L. D. Mech. 1997. Wolf pack size and
food acquisition. American Naturalist 150:513–517.
Seip, D. R. 1992. Factors limiting woodland caribou popu-
lations and their interrelationships with wolves and moose
in southeastern British Columbia. Canadian Journal of Zo-
ology 70:1494–1503.
Sinclair, A. R. E. 1989. Population regulation in animals.
Pages 197–241
in
J. M. Cherret, editor. Ecological con-
cepts. Blackwell Scientific, Oxford, UK.
Tabachnick, B. G., and L. S. Fidell. 1983. Using multivariate
statistics. Harper and Row, New York, New York, USA.
Teerink, B. J. 1991. Hair of West-European mammals. Atlas
and identification key. Cambridge University Press, Cam-
bridge, UK.
Thurber, J. M., and R. O. Peterson. 1993. Effects of popu-
lation density and pack size on the foraging ecology of
gray wolves. Journal of Mammalogy 74:879–889.
... he formula for winter mortality was each Age class for the summer population multiplied by the RB upper and lower survival rates. The following years' spring population stage carried over all age and sex class cells from the results of the winter population, and the annual cycle was ran for 50 years. Nores, et. al., 2008 Medium Predation 11.9 19.5 Jedrzejewski, et. al., 2002 High Predation 25 40 Heptner, et. al., 1988 Percentage of Wolf Diet on Boar Age Class Piglet (Age 0) 0.83 0.95 Mattioli, et. al., 2011 Adult (Age 1+) 0.05 0.17 Mattioli, et. al., 2011 Wolf predation rates were calculated by; ...
... Medium-Estimate that 31 to 39 wild boars are killed annually by wolves in a 100km 2 area in Poland. Also, was estimated that 2-2.6 wolves live in a 100km 2 area. Those estimates were used to calculate 11.9 lower and 19.5 upper predation rate (Jedrzejewski, et. al., 2002). ...
... Restoring predator-prey interactions is another complementary yet controversial measure to reduce the impacts of overabundant ungulates (Apollonio et al. 2017;Marrotte et al. 2022;Nilsen et al. 2007). Large carnivores predate on large ungulates (Jedrzejewska and Jedrzejewski 1998;Jedrzejewski et al. 2002;Ripple and Beschta 2012;van Beeck Calkoen et al. 2023) and modulate prey behaviour through landscapes of fear (Epperly et al. 2021;Gaynor et al. 2019). The grey wolf (Canis lupus) -the apex predator with the largest distribution in Europe (Chapron et al. 2014) -can reduce the risk of disease transmission from wild boars to livestock by removing diseased individuals (Tanner et al. 2019) and the incidence of deer-vehicle collisions by keeping ungulates away from roads (Raynor et al. 2021). ...
Article
Full-text available
Increasing abundance of large ungulates is raising human-wildlife impacts and the effectiveness of recreational hunting to reduce their population growth is increasingly questioned. We report on long-term trends (> 15 years) in wildlife-vehicle collisions (WVC) and hunting bags, and on associations between the annual growth rate of WVC and that of hunting bags for three ungulates – the wild boar, the red deer, and the roe deer – and the grey wolf in northwest Spain to evaluate the regulating capacity of recreational hunting at large spatial scale. Wildlife-vehicle collisions increased by 332% in 16 years and 91% of all traffic accidents were caused by collisions with these three ungulates. All ungulate species showed significant positive trends in WVC and hunting bags, but we did not observe a negative association between annual growth rate of hunting bags and that of WVC except for the wild boar. Results suggest that recreational hunting was unable to reduce ungulate population growth at the regional scale. There was no upward trend of vehicle collisions with wolves over the study period, possibly reflecting stable wolf populations. Natural mortality due to predation could be promoted through the protection of apex predators, but the lethal management of apex predators, often based on sociopolitical pressures rather than damage levels, can conflict with the strategy for mitigating ungulate impacts. Ungulate management needs to be reconsidered from an ecological perspective that integrates human management measures, including recreational hunting, based on the population dynamics and the recovery of predator–prey interactions by favoring the expansion of apex predators.
... Studies on diets are important to predict the viability of each species in its habitat (Wang, 2002). Diet analyses of free-ranging carnivores can improve our understanding of their potential effects on prey populations, ecology, and potential competition with other predators (Jedrzejewski et al., 2002;Wachter et al., 2012). ...
Article
Full-text available
Co-occurring carnivore species that are phylogenetically related or of similar size, morphology, and ecological needs often reduce competition by partitioning shared resources through temporal, spatial, and dietary niche segregation via behavioral adaptations. Caracals (Caracal caracal) and jungle cats (Felis chaus) co-occur in portions of their geographical ranges and are expected to display resource segregation in these ranges. We compiled scat, stomach content, and prey remains found data from published and unpublished sources to summarize information on the diets of caracals and jungle cats across their geographical ranges during 1842–2021. We obtained 63 sources from 26 countries in Europe, Asia, and Africa, in which caracal diet included 151 species while jungle cat diet included 61 species. We found that caracals and jungle cats did not exhibit dietary niche partitioning and had greater dietary similarities in areas of range overlap. We also found that caracals consumed more diverse prey species including prey with greater average body mass compared to jungle cats. Our results suggest that greater prey diversity in areas of range overlap, caracal predation on wide range of prey, and opportunistic feeding behavior that facilitates consumption of more diverse prey species compared to jungle cats, may facilitate co-occurrence between these two felid species.
... In Germany, the diet of the wolf population in the period of increase after initial immigration included 21% C. elaphus and 55% roe deer, Capreolus capreolus [116]. From [117], we may presume that wolf predation on C. elaphus increases in natural forests, which is not the case in Lithuania. ...
Article
Full-text available
The history, current population status and (un)sustainable management of red deer (Cervus elaphus) in Lithuania are reviewed on the basis of 57 publications and, additionally, analysis of data on numbers and hunting bag. After the extinction of the species at the beginning of the 19th century, red deer were reintroduced into the northern part of the country during World War I. Population re-establishment was further fueled after WWII by immigration from neighboring countries and local translocations. After the introduction of the Law on Hunting in 2002, which enabled hunting organizations to rent hunting grounds for longer than 10 years, local wildlife management strategies were targeted at increasing the number of animals. However, a scientifically-based target population size for C. elaphus, agreed by all major stakeholders, has still not been defined. In the last 20 years, population growth has been exponential, deer numbers in Lithuania being estimated as 68,816 individuals in 2021. An increase in the size of the hunting bag is necessary for the sustainable management of the species at the national scale and to avoid risks of over-population as observed elsewhere in Europe.
Article
Full-text available
Through killing and instilling fear in their prey, large terrestrial carnivores shape the structure and function of ecosystems globally. Most large carnivore species have experienced severe range and population declines due to human activities, and many are now threatened with extinction. Consequently, the impacts of these predators on food webs have been diminished or lost completely from many ecosystems. Kill rates provide a fundamental metric for understanding large carnivore ecology and assessing and comparing predation within and across ecological communities. Our systematic review of large terrestrial mammalian carnivore kill rates reveals significant positive geographic (North America, Europe, and Africa) and taxonomic (grey wolf Canis lupus, puma Puma concolor, lion Panthera leo, and Eurasian lynx Lynx lynx) bias, with most studies apparently motivated by human–carnivore conflict over access to ungulate prey and wildlife management objectives. Our current understanding of the behaviour and functional roles of many large carnivore species and populations thus remains limited. By synthesising and comparing kill rates, we show that solitary carnivores (e.g. brown bears Ursus arctos and most felids) exhibit higher per capita kill rates than social carnivores. However, ungulate predation by bears is typically limited to predation of neonates during a short period. Lower per capita kill rates by social carnivores suggests group living significantly reduces energetic demands, or, alternatively, that group-living carnivores defend and consume a greater proportion of large prey carcasses, or may acquire more food through other means (e.g. scavenging, kleptoparasitism) than solitary hunters. Kill and consumption rates for Canidae – measured as kilograms of prey per kilogram of carnivore per day – are positively correlated with body mass, consistent with increasing energy costs associated with a cursorial hunting strategy. By contrast, ambush predators such as felids show an opposite trend, and thus the potential energetic advantage of an ambush hunting strategy for carnivores as body mass increases. Additionally, ungulate kill rates remain relatively constant across solitary felid body sizes, indicative of energetic constraints and optimal foraging. Kill rate estimates also reveal potential insights into trophic structuring within carnivore guilds, with subordinate carnivores often killing more than their larger counterparts, which may be indicative of having to cope with food losses to scavengers and dominant competitors. Subordinate carnivores may thus serve an important role in provisioning food to other trophic levels within their respective ecosystems. Importantly, kill rates also clarify misconceptions around the predatory behaviour of carnivores (e.g. spotted hyaenas Crocuta crocuta and wolverines Gulo gulo are often considered scavengers rather than the capable hunters that they are) and thus the potential impacts of various carnivore species on their ecological communities. Despite the importance of kill rates in understanding predator–prey interactions, their utility is not widely recognised, and insufficient research limits our ability to fully appreciate and predict the consequences of modified predation regimes, justify current management actions affecting carnivores, or inform effective conservation measures. Together with other important research on predator–prey interactions, robust kill rate studies that address the research deficiencies we highlight will provide a deeper understanding of the foraging behaviours and potential ecosystem impacts of many of the world's carnivores, thus aiding effective conservation and management actions.
Technical Report
Full-text available
Factors Affecting Moose Population Declines in British Columbia May 2022 Progress Report
Article
Full-text available
The aim of the study was to assess which kill site characteristics were selected by a lone wolf living in a protected Mediterranean coastal area near the city of Pisa, Italy, where both wild and domestic ungulates were available as potential prey. Between 2017 and 2019, we monitored the wolf’s predatory behaviour through a combination of camera trapping and active search for kill sites and prey carcasses. The main prey found was the fallow deer (n = 82); only two wild boars and no domestic ungulates were found preyed upon. The features and habitat of kill sites were modelled to test for selection by the wolf. The habitat type of kill site was composed of meadows and pastures (89.3%), woods (7.3%), degraded coastal areas (1.9%), roads and rivers (1.1%), and marshes (0.5%). We calculated their distance from landscape features and ran a binomial generalised linear model to test the influence of such landscape variables. The distance of kill sites from landscape elements was significantly different from random control sites, and a positive selection for fences was found. In fact, the wolf pushed fallow deer towards a fence to constrain them and prevent them from escaping. We also analysed the body condition of predated fallow deer as a percentage of fat content in the bone marrow of the hind legs. Our results revealed the selection of the lone wolf for deer in good body condition. This is a possible outcome of the habitat selection shown by fallow deer in the study area, where fenced open pastures are the richest in trophic resources; therefore, our findings suggest a high efficacy for the lone wolf hunting strategy, but also the adoption of a high risk feeding strategy by deer. This study suggests that a lone predator can take advantage of human infrastructures to maximise its predatory effectiveness.
Book
Full-text available
Roe deer (Capreolus capreolus) is an important component of natural ecosystems, the most common object of hunting among the species of the Deer family. In Belarus, the number of the species has shown a steady increase in recent years, also due to the absence of severe and snowy winters. Despite the growth in numbers, the roe deer still remains quite vulnerable to various biotic and abiotic factors. In addition, when using erroneous management principles, roe deer can be significantly exposed to negative anthropogenic influence. In turn, an excessive and uncontrolled increase in the number of roe deer can lead to an increase in conflict situations with human interests - an increase in damage to forest plantations, an increase in road traffic accidents (hereinafter referred to as road accidents) involving wild animals, the emergence and spread of various diseases common to agricultural and wild animals. One of the activities of the state program "Belarusian Forest" for 2016-2020* was the development of a management plan for the roe deer population in Belarus, which is a very urgent task aimed at optimizing the protection and management principles of the population in order to develop hunting and ecological tourism.
Article
Tree cavities are an essential resource for cavity-dwelling mammals, birds, invertebrates and fungi, and so are important for maintaining forest biodiversity. In North American forests, woodpeckers (Picidae) play a keystone role in cavity creation by excavating holes. However, in European forests many hole-nesting songbirds rely on non-excavated cavities that are formed by fungal decay and compartmentalization after tree damage. Several factors are recognised in initiating non-excavated cavities that are used by hole-nesting birds, including loss of a tree branch or stem breakage, but this topic is poorly studied. Here, we propose that bark stripping by large herbivores (e.g. Red Deer Cervus elaphus and European Bison Bison bonasus) could be another important, and previously overlooked, mechanism for initiating tree cavities that are used by hole-nesting birds. We suggest that, after the initial damage from herbivore bark-stripping, fungal decay can create specific elongated, slit-like cavities, which are particularly important as nest sites for some common forest songbirds. We outline this idea using original observations and evidence from the literature, primarily from the primeval forest in Poland's Białowieża National Park. We also use studies from elsewhere in Europe to show a generally low usage of slit cavities by birds where large herbivores are scarce or absent. We suggest that restoring such animals in European forests could help to restore the abundance and diversity of the tree cavity resource for hole-dwelling species. We encourage future research to investigate this proposal of large herbivores being important agents of tree cavity formation that could enhance biodiversity
Article
Full-text available
During September 1980-December 1986, 81 radio-collared wolves (Canis lupus) were monitored in and near the 839-km2 Bearville Study Area )BSA) in north-central Minnesota. Each year winter-territory size averaged 78-153 km2; no territories had road densities >0.72 km/km2. From zero to 30% of radiomarked pup, yearling, or adult wolves left their territories each month. Pups left natal packs during January-March and older wolves left frequently during September-April. Wolves temporarily leaving territories moved 5-105 km away and were absent 3-118 days; up to 6 exploratory moves were made prior to dispersal. Dispersing wolves traveled 5-100 km away during periods of 1-265 days. One disperser joined and established pack, but 16 others formed new packs. Annual dispersal rates were about 0.17 for adults, 0.49 for yearlings, and 0.10 for pups. Each year mean pack size ranged from 5-9 in November/December to 4-6 in March. Annual wolf density (including 16% lone wolves) ranged from 39-59 wolves/1,000 km2 in November-December and 29-40 wolves/1,000 km2 in March. Annual immigration was 7%. The observed mean annual finite rate of increase was 1.02, and annual rates of increase were correlated with mean number of pups per pack in November. Litters averaged 6.6 pups at birth and 3.2 pups by mid-November, at which time pups made up 46% of pack members. Annual survival of radio-marked wolves >5 months old was 0.64. Despite legal protection, 80% of identified wolf mortality was human caused (30% shot, 12% snared, 11% hit by vehicles, 6% killed by government trappers, and 21% kill by humans in some undetermined manner); 10% of wolves that died were killed by other wolves. During sample periods in 2 winters, wolves were located twice daily to estimate predation rates on white-tailed deer (Odocoileus virginianus). Estimated minimum kill rates during January-February (x = 21 days/kill/wolf) did not differ between winters with differing snow depths. Winter consumption averaged 2.0 kg deer/wolf/day (6% body wt/day). Scat analyses indicated deer were the primary prey in winter and spring, but beaver (Castor canadensis) were an important secondary prey (20-47% of items in scats) during April-May. Neonatal deer fawns occurred in 25-60% of scats during June-July whereas the occurrence of beaver declined markedly. Overall, deer provided 79-98% of biomass consumed each month. Adult wolves consumed an estimated 19/year, of which 11 were fawns. A review of North American studies indicates that wolf numbers are directly related to ungulate biomass. Where deer are primary prey, territory size is related to deer density. Per capita biomass availability likely affects pup survival, the major factor in wolf population growth. Annual rates of increase of exploited populations vary directly with mortality rates, and harvest exceeding 28% of the winter population often result in declines. Management decisions concerning wolf and ungulate density and ungulate harvest by humans can be made using equations that incorporate estimate of wolf density, annual ungulated kill per wolf, ungulate densities, potential rate of increase for ungulates, and harvest.
Article
Full-text available
Home-range size, its seasonal variation, and pattern of home-range use of wolves (Canis lupus) were studied in Bialowieza Primeval Forest (BPF) located on the Polish-Belarussian borderland in 1994-1996. Tn the Belarussian part of BPF where wolves were hunted, their winter density was 0.9-1.5 individuals/100 km(2), and mean pack size was 2.7-3.2 wolves. In the Polish part of BPF where wolves were protected, their densities were 2-2.6 individuals/100 km(2), and mean pack size was four to five wolves. In spring-summer, wolves usually moved singly or in pairs (65% of observations), but in autumn-winter, 51% of seen or snowtracked groups were whole packs. In 1994-1996, four wolves belonging to two neighbouring packs of five to seven individuals each were radiotracked in the Polish part of BPF for 4-18 months. Their total home ranges, estimated by the minimum convex polygon method with 100% of locations, covered 173-294 km(2). Core areas of home ranges, comprising 50% of locations, were small: 11-23 km(2), or 5-13% of the total home ranges. Packs hunted both in core areas and peripheral parts of the ranges, but the majority of their diurnal resting sites were located in core areas. Home ranges of wolves were 141-168 km(2) in spring-summer (May-September) and 99-271 km(2) in autumn-winter (October-April). There was nearly no overlap of the two packs' home ranges (0-3% in various seasons). Variation in the size of seasonal hone range was observed for two breeding females. During parturition and early nursing in May-June, they confined their activity to an average of 17 km(2), Literature on sizes of home ranges of Eurasian wolves was reviewed. Home ranges of wolves increased from 80-240 km(2) in southern and central Europe to 415-500 km(2) in northern Scandinavia. Smallest ranges were reported from regions where red deer (Cervus elaphus) were common. The population status of wolves affected size of their ranges, they were large in low-density colonizing populations and small in established populations.
Article
Within the 2700km² Beltrami Island State Forest, near the W edge of the primary range of Canis lupus in Minnesota, wolf population density was low at the start of the study in 1972 but increased substantially up to 1977 (end of study). At least 8 of 13 social units present in mid-1976 had formed since 1972. Size of litters of established packs averaged 4.6 pups, and those of newly-formed pairs averaged 4.1. Mortality decreased over the study period, and recruitment of young wolves exceeded mortality following legal protection. A high rate of dispersal of young from packs was documented. Dispersal peaked in autumn. Most wolves paired within a few days of leaving their packs. Average territory size decreased as both population and pack numbers increased. Behaviour of alpha males, alpha females and subordinate members of the packs is discussed. Deer and moose comprised 94% of animal biomass eaten by wolves, with deer along accounting for 67%. Seasonal differences in food taken and energy requirements are noted.-P.J.Jarvis
Article
Moose Alces alces and caribou Rangifer tarandus populations increased following a wolf Canis lupus reduction program in the 1950's and reached peak abundance in the 1960's. Deep snow and heavy browsing caused an initial crash of moose in 1965-66. Moose continued to decline until 1976, primarily due to periodic deep snow, harvest by man, and wolf predation. The long-term effect of moose mortality from deep snow was to increase the impact of predation by lowering moose/wolf ratios. Hunting and wolf predation were the principal causes of moose mortality from 1971-75. Harvests removed 6-19% of the moose population annually; mean harvest rate equalled mean yearling recruitment. After 1974, harvest removed 2% of the moose. Predation by wolves removed 13-34% of moose during winters 1973-74 and 1974-75 and a high proportion of calves during summer. Mortality from predation during winter exceeded recruitment of calf moose. Hunting and wolf predation were also the primary proximate mortality causes in the decline of caribou, but calf recruitment was so low from 1971-75 that a significant decline would have occurred without hunting. After 1973 when hunting was stopped, predation limited the population. Following a 61% reduction in wolves in 1976, survival of calf and yearling moose increased 2- to 4-fold, adult mortality declined, and the moose population increased. Survival of caribou calves also increased significantly, and the population grew rapidly. -from Authors
Article
Canis lupus activity and interactions with white-tailed deer Odocoileus virginianus were monitored during 6 winters in which mean January-February snow depth alternated between shallow (19-26 cm) and relatively deep (40-47 cm) and winters alternated between mild and moderately severe. Wolves travelled farther and more often and spent less time with other pack members in mild than in severe winters. Wolves and deer used conifer cover less, and fewer deer were killed there, when snow was shallow. Similarly, fewer wolf-killed deer were found in and near deer concentration areas during mild winters. Of the 74 deer killed by wolves, the proportion that were fawns (54%) differed from the proportion of fawns in the winter population (27%), but neither varied with winter severity. Few deer killed by wolves appeared debilitated. Carcass consumption was high in all winters, regardless of their severity, but wolves scavenged less in mild than in severe winters (10 vs. 29% of deer carcasses observed). -from Author
Article
A decline in the gray wolf (Canis lupus) population in Isle Royale National Park prompted an intensive study of radiocollared individuals in 1988-1991, complementing an ongoing study begun in 1958. During winter, 1959-1991, the proportion of lone wolves was higher when the population was at low levels, whereas average size of packs declined with density of wolves. Sizes of territories (1971-1991) were not related to size of packs. In winter, food availability (kg wolf-1 day-1) and kill interval (days/kill) varied inversely with size of packs. Social relationships of radiocollared lone wolves were flexible; they often formed temporary groups of two or three wolves. Small groups (less than four) and lone wolves roamed large areas of the island (300 km2 and 540 km2, respectively), frequently moving through territories of established packs. Solitary wolves and pairs readily killed adult moose, in contrast to a common belief that larger packs benefit from cooperative hunting.
Article
Analyses of 34 femur marrow samples from barren ground caribou, Rangifer tarandus granti, revealed that this tissue is a three-component system comprised of water, fat and non-fat residue. Both the water and fat, and residue and fat are linearly and inversely related. Because the residue amounts to only 2-6 percent of the net weight, the dry weight of the marrow, except at low fat levels, is essentially equal to the weight of fat in the marrow. For maximum accuracy at low fat levels, the dry weight:fat inequality can be corrected by subtracting the average residue value for that fat level from the dry weight. Such corrected dry weights adequately measure femur fat content at all fat levels for many studies concerned with marrow fat reserves.
Article
In 9 trials, captive wolves (Canis lupus) were fed prey varying in size from snowshoe hares (Lepus americanus) to adult deer (Odocoileus virginianus), and the resulting scats were counted. Field-collectible scats were distinguished from liquid, noncollectible stools. In collectible scats, the remains of small prey occurred in greater proportion relative to the prey's weight, and in lesser proportion relative to the prey's numbers, than did the remains of larger prey. A regression equation with an excellent fit to the data (r2 = 0.97) was derived to estimate the weight of prey eaten per collectible scat for any prey. With this information and average prey weights, the relative numbers of different prey eaten also can be calculated.