Species-Specific Differences and
Similarities in the Behavior of
Hand-Raised Dog and Wolf Pups
in Social Situations with Humans
Ma ´rta Ga ´csi
Borba ´la Gyori
A´da ´m Miklo ´si
Zso ´fia Vira ´nyi
Department of Ethology
Eo ¨tvo ¨s University
Budapest, Pa ´zma ´ny P. 1/c.
Jo ´zsef Topa ´l
Vilmos Csa ´nyi
Comparative Ethology Research Group
Hungarian Academy of Sciences
Budapest, Pa ´zma ´ny P. 1/c.
ABSTRACT: In order to reveal early species-specific differences, we observed the
behaviorof dog puppies (n¼11) and wolf pups (n¼13) hand raised and intensively
socialized in an identical way. The pups were studied in two object-preference
tests at age 3, 4, and 5 weeks. After a short isolation, we observed the subjects’
behavior in the presence of a pair of objects, one was always the subject’s human
foster parent (caregiver) and the other was varied; nursing bottle (3 weeks),
unfamiliar adult dog (3 and 5 weeks), unfamiliarexperimenter (4 and 5 weeks), and
familiar conspecific age mate (4 weeks). Dogs and wolves did not differ in their
general activity level during the tests. Wolf pups showed preference for the
proximity of the caregiver in two of the tests; Bottle-Caregiver at the age of 3 weeks
and Experimenter-Caregiver at the age of 5 weeks, while dogs showed preference to
the caregiver in three tests; conspecific Pup-Caregiver and Experimenter-Caregiver
at the age of 4 weeks and dog-caregiver at the age of 5. Compared to wolves, dogs
tended to display more communicative signals that could potentially facilitate
social interactions, such as distress vocalization, tail wagging, and gazing at the
humans’ face. In contrast to dog puppies, wolf pups showed aggressive behavior
toward a familiar experimenter and also seemed to be more prone to avoidance.
Our results demonstrate that already at this early age—despite unprecedented
intensity of socialization and the comparable social (human) environment during
early development—there are specific behavioral differences between wolves and
dogs mostly with regard to their interactions with humans.
? 2005 Wiley Periodicals, Inc. Dev Psychobiol 47: 111–122, 2005.
Keywords: hand-raising; domestication; inborn social preferences; communica-
tive signals; dog; wolf
In recent years, the evolutionary approach to under-
standing dog behavior has gained wide-spread interest
(Hare, Brown, Williamson, & Tomasello, 2002; Miklo ´si,
Topa ´l, & Csa ´nyi, 2004). This is partly due to the assump-
tion that dogs did not evolve simply by selection for
human proximity (e.g., as proposed by Coppinger &
Coppinger, 2002) but more broadly to the social relation-
ships that characterize human groups and societies.
Recent studies from different research groups suggest
that dogs show a specific attachment to caregivers from
very early age (Topa ´l et al., 2005), they can engage in
(McKinley & Sambrock, 2000; Soproni, Miklo ´si, Topa ´l,
& Csa ´nyi, 2001), are able to recognize minute behavioral
cues characterizing human visual attention (Call, Bra ¨uer,
Kaminski, & Tomasello, 2003; Ga ´csi, Miklo ´si, Varga,
Topa ´l, & Csa ´nyi, 2004; Vira ´nyi, Topa ´l, Ga ´csi, Miklo ´si, &
Csa ´nyi, 2004), and learn readily by observing humans
Received 19 July 2004; Accepted 10 May 2005
Correspondence to: M. Ga ´csi
Contract grant sponsor: OTKA
Contract grant number: T029705
Contract grant sponsor: Hungarian Academy of Sciences
Contract grant number: F01/031
Published online in Wiley InterScience
(www.interscience.wiley.com). DOI 10.1002/dev.20082
? 2005 Wiley Periodicals, Inc.
solving various problems (Pongra ´cz, Miklo ´si, Kubinyi,
Gurobi, & Csa ´nyi, 2001; Kubinyi, Topa ´l, Miklo ´si, &
Csa ´nyi, 2003). These observations provided support for
our earlier hypothesis that dogs have adapted to become
integrated into human social groups and they evolved
behavioral and cognitive skills to interact with us
(Miklo ´si, Polga ´rdi, Topa ´l, & Csa ´nyi, 2000).
Considering that the wolf is regarded to be the sole
ancestor of the dog (Vila ´ et al., 1997; Wayne, 1993) to
investigate the above hypothesis, it is critical to know the
extent to which predispositions are responsible for the
differences in the interspecific social behavior of dogs
and wolves. In light of recent findings, we need more
experimental data than former comparative studies that
provide to determine how domestic dogs have acquired
their unique skills. From the 1960s, a relatively large
number of observational and experimental studies have
investigated the similarities and species-specific differ-
ences in the social behavior of young dogs and wolves.
On the basis of a series of experiments, Scott & Fuller
(1965) arguedthat dog puppies have an ‘‘optimal’’ period
of socialization, but at the same time, even short periods
of social stimulation seem to counteract the effects of
relatively long-term isolation (Fuller, 1967). Further, Fox
& Stelzner (1966) found that puppies reared in isolation
from conspecifics or raised with cats (Fox, 1970) show
deficits in social behavior toward conspecifics, but in
both cases, normal social behavior could be reinstated
after 1–2 weeks of socialization with conspecifics.
Similar experiments have also been conducted with
the wolf (MacDonald & Ginsburg, 1981). Although,
the number of such observations is restricted, similar
flexibility of early learning has been documented in
relation to species-specific preferences, that is, in spite
of early social deprivation, the wolf pups’ behavior
In summary, these experiments suggest that both the dog
and the wolf have a flexible behavioral system that is
relatively buffered against environmental effects, and
strong inborn preferences for interaction with conspeci-
fics allow for the recovery of species-specific behavior
even after extensive periods without appropriate social
Although the two species seem to be similar in their
social developmental processes with regard to conspe-
cifics, some marked differences in their behavior toward
humans have been observed (Frank & Frank, 1982a,
1985). However, it should be pointed out that previously
there have been rather few research programs using
wolves and dogs with similar rearing history (Frank &
smallso itwas difficult toclearly establish thegenetically
based behavioral differences. Feddersen-Petersen (1986)
compared the intraspecific behavior of young wolves and
dogs having limited contact with humans. Others have
1967; Frank, 1980; Frank & Frank, 1982a; Frank, Frank,
Hasselbach, & Littleton, 1989; Woolpy & Ginsburg,
1967) in order to investigate motivational and cognitive
differences. Such comparative research usually assumes
that the revealed species characteristics and/or specific
differences reflect the influences of differential genetic
determination. With the same general assumption, our
research program is distinctive in three very important
and wolves should experience the same and especially
intensive socialization by humans. That is, each human
caretaker spent the first 2–4 months with one individual
by providing care for 24 hr a day. Observations that
humans can only socialize wolf pups successfully if they
are separated very early (before eye opening) from the
mother (Klinghammer & Goodmann, 1987) suggest
very early learning and/or strong genetic preference of
conspecifics in the wolf. So we planned our socialization
regime especially carefully to exclude results deriving
only from the differing sensitivity for early socialization
in the two species. Second, we performed numerous
specific investigations to observe (and compare) the
behavior of wolves and dogs toward humans with regard
to attachment (Topa ´l et al., 2005) and communication
(Miklo ´si et al., 2003, Vira ´nyi et al., 2005). Third, we
socialized a relatively great number of individuals in
order to have more chance to discriminate behavioral
traits that are exclusively specific to one of the species
(qualitative differences) from those that are present in
both species and possibly represent two extremes of the
same distribution (quantitative differences).
For the present study,three possible hypotheses can be
formulated: First, given the same environmental condi-
tions, no differences emerge between the species at this
early developmental stage. Second, dog-wolf differences
in behavior can be explained by wolves’different (faster)
developmental speed that has been documented earlier
(Frank & Frank, 1982a; Zimen, 1987). Third, the be-
selected for during domestication. A good example for
(approaching human hand) in ‘‘domesticated’’ foxes,
which resulted in a prolonged capacity of socialization
(Belyaev, Plyusnina, & Trut, 1985). These differences
could be based on various aspects of the developmental
process, for example, on a decreased specificity of the
learning constraints or on a change in the duration of the
their primary (human) caregiver. For the testing, we have
applied an object-preference test, a method often used to
Ga ´csi et al.
look for early effects of social experience. The subject
or nonsocial) simultaneously presented objects by spend-
ing more time with one or the other stimulus object (e.g.,
to use a natural set-up that interfered relatively little with
the behavior of the subjects.
Because of earlier indication for differences in tem-
perament traits already at this age (e.g., Frank & Frank,
1982b, 1987), we have also measured signs of aggressive
and avoidance behavior.
In previous studies, we have provided evidence that
dogs have an advantage to use face and eye-related
gestural cues (e.g., Soproni et al., 2001). Further, there
appears to be a species-specific difference in the use of
face/eye contact in social interactions with humans when
comparing 4-month-old wolves and dogs (Miklo ´si et al.,
2003). To address this question, we investigated whether
the species-specific differences can be traced back to an
to study whether these differences can be masked by a
In year 2000–2003, our group raised 13 gray wolf pups
(Canis lupus) born at Horatius Ltd. Animal Park (6 males
and 7 females, from five different litters) and 11 dog
puppies (Canis familiaris) from three shelters (6 males
the animals were socialized and tested the same way.
Not all of the subjects could participate in every test
(caregiverwas ill or subject had injury) and in some cases
video cassette). This means that the actual number of
individuals varies in the tests, so in each case we give the
exact number of the animals whose data were analyzed
(see Tab. 1).
Subjects (dogs and wolves)were individually hand raised
by humans after being separated from their mothers and
littermates on Day 4–6 after birth (before their eyes
were based on a detailed protocol (Ujfalussy, 2003).
Fentress (1967) reared one wolf most similarly to our
procedure, but the pup was already 4 weeks old when he
together with them at night. They were bottle-fed and
from the age of 4–5 weeks hand fed also with solid food.
The caregivers carried them in a pouch, so the pups could
transport, attending classes at the university, visiting
and at least twice a week they also met conspecifics of
about the same age and adult dogs. This way, they were
to familiar individuals. The basic principle of socializa-
tion was avoiding competitive situations and aggressive
interactions with the animals, that is, to behave rather like
a mother than a dominant conspecific.
Our team was licensed by the Department of Nature
(No.3293/2001) to hand rear and expose the wolf pups
to extensive socialization, and our department has also
been licensed by the Ethical Committee for Animal
Experimentation attheEo ¨tvo ¨sUniversitytoconductsuch
at the animal park where they could interact daily with
humans and other wolves. The caretakers carried on
visiting them once or twice a week, so the wolves
were regularly taken out of the pack for further testing,
Table 1.The Sequence of Testing at the Age of 3, 4, and 5 Weeks
3 Weeks (20–22 Days)4 Weeks (27–29 Days) 5 Weeks (34–36 Days)
5-min isolation in box
test (N(W)¼12, N(D)¼9)
Conspecific pup-caregiver object-
preference test (N(W)¼9, N(D)¼8)
10–15-min in pen
5-min isolation in box
preference test (N(W)¼12, N(D)¼11)
test (N(W)¼12, N(D)¼11)
test (N(W)¼12, N(D)¼11)
preference test (N(W)¼12, N(D)¼11)
Note. On each occasion, the procedure started with a 5-min-long isolation. Immediately after the isolation, subjects participated in the first object-
preference test. It was followed by a 10–15-min period when the animals rested in the pen. Then the pups were isolated again and the second object-
preference test came next. The number of participating subjects is indicated in case of each test (some of the tests could not be analyzed due to technical
problems; these are omitted from the table).
Social Behavior of Hand-Raised Dog and Wolf Pups with Humans
some regular training and other free social interactions.
they were successfully integrated into a pack (living in
caregivers, and we could find loving home for the other
four as well.
The subjects were presented with all together six object-
preference tests; two tests were conducted on every
subject at the age of 3, 4, and 5 weeks. On each occasion,
all subjects were observed first in the first test one-by-one
and then in the same order in the second test. The tests
were performed in the morning and the caregivers fed the
animals with milk at least 2 hr earlier. Before the tests,
cardboard box situated in an unfamiliar empty room.
(This was done in order to elicit similar motivational
levels in all pups to initiate social interactions during the
test.) The subjects was put into the box and taken out of it
by a familiar female experimenter.
Following the isolation period, the subject was carried
into another room (2.6 m?3.6 m) that was unfamiliar to
them at the first occasion (at the age of 3 weeks). All
In each case, two objects were placed in the room that
otherwise was empty, except the cameraman who stood
behind a 1.2 m high plastic screen and recorded the
behavior of the subjects. The position of the two objects
and the starting location of the subject formed an
equilateral triangle. The objects’ location (right or left
side) was counterbalanced on the same day. The familiar
the isolation box) placed the subject to the starting point,
held it for a second making it orient toward the objects,
and then let it go while stepping back two steps and
remained still from then on (see Fig. 1 for the testing
design). After placing the subject to the starting location,
90 s. In this case, at the 90th s she woke it up by gently
rubbing it for a few seconds and then placed it to
the starting point again. The subjects’ behavior was
observed for 5 min. At the end of the test, the familiar
experimenter slowly approached the subject from the
front, caught, and lifted it without talking to it.
In all tests, the caregiver of the subject was one of the
objects (‘‘reference’’) who were paired with different
kinds of other ‘‘social objects’’ with the only exception
of the nursing bottle in the very first test (see Tab. 1).
The human participants always sat cross-legged and
motionless on the floor quietly facing the subject. Their
hands were placed on the floor in front of them with
Bottle-Caregiver. There was some lukewarm milk (used
for feeding) in the nursing bottle that was placed on a
small cloth soaked with milk. The caregiver was looking
at the subject during the test.
Experimenter-Caregiver. Both humans were looking at
the subject during the test. In both tests, the same
female experimenter took part. At the first time, she
was unfamiliar to the subjects, and did not have any
contact with them betweenthe similar tests at the age of 4
and 5 weeks.
indicate the position of the two different types of objects that
experimenter, adult dog, conspecific pup, milk bottle according
to the type of the test). The gray-shaded areas represent the
proximity of the objects. At the beginning of the object-
preference test, the familiar female experimenter (E) put the
subject (S) on the starting point, which had been marked on
the floor. The two objects were 1.1 m from each other and 1.5 m
from the subject’s starting position. A cameraman, standing
behind a 1.2 m high plastic screen, recorded the tests with a
camera positioned on a tripod.
Ga ´csi et al.
Dog-Caregiver. The adult dog was positioned facing the
starting point of the subject. He was a well-trained, adult
Belgian shepherd male that at the time of the first test was
unfamiliar to the subjects and had no contact with them
apart from the tests. The dog laid calmly at his place for
5 min. This time the caregiver adjusted her behavior to
that of the adult dog, that is, she oriented to the subject
only when the dog did so.
Conspecific Pup-Caregiver. We always selected a same
age sleeping conspecific from the subject pool in order to
avoid interactions. Conspecific subjects knew each other
equally well, as they had the possibility to meet regularly
from the age of 2 weeks. The sleeping pup was gently
placed tothe floorand watched forafewseconds whether
it was lying calmly. The caregiver sat motionless looking
at the subject.
As the duration of the test sessions varied slightly, we
calculated the relative percentage of the time spent with
Activity (%). In this early age, social behavior obviously
cannot be analyzed without considering the animals’
general mobility or activity level. Taking into account the
immature motor behavior of the subjects, we assessed
activity observing the time (s) spent standing or moving
on four legs. (We only considered the time when the
subjects were not in physical contact with the objects.)
For statistical analysis, we used the mean activity
value measured in the two tests on the same day (i.e.,
activity level at the age of 3 weeks; (A1/t1þA2/t2)/2,
where A1¼time standing or moving on four legs during
the first test at the age of 3 weeks, t1¼total time of the
first test at the age of 3 weeks, A2¼time standing or
moving on four legs during the second test at the age of
3 weeks, t2¼total time of the second test at the age of
Proximity (%). We measured the relative duration of the
total time spent in proximity with each object closer than
the length of the subject’s own body (i.e., the subject is
closer than approximately 25–35 cm to any part of the
stimulus). For statistical analysis, a preference index was
calculated (see below).
Vocalization (%). We measured the relative duration of
the total time spent with any form of vocalization. For
statistical analysis, we used the mean vocalization value
measured in the two tests on the same day (same
calculation as in case of activity).
The occurrence of vocalization (score, 0–1) was also
recorded by the familiar experimenter just before she
entered the room where the subject was isolated.
Gazing at Face (Relative Frequency). Gazing was de-
was characterized by lifting of the head. This variable
was recorded only while the animal was in proximity
with a human (caregiver and unfamiliar experimenter).
The frequency of gazing at human face was calculated
by dividing the number of gazings by the time spent in
proximity with the humans. For statistical analysis, we
used the mean of gazing frequencies measured in the two
tests on the same day.(The adult dog’s eyeswere not high
enough to identify the subjects’ head movements as
Tail Wagging (Score, 0–1). The subject was given a
score 1ifitwagged itstailwhenapproachingeither ofthe
objects for the first time (getting closer than 30 cm) or
while first getting into physical contact with either of
them. A score 0 was given if this behavior was not ob-
served on these occasions or the animal did not get into
proximity with a stimulus during the test.
We have recorded signs of avoidance and aggressive-
ness of the subjects toward the approaching familiar
experimenter both before and after each test when the
animal was handled: first, at the end of the isolation
periods when the familiar experimenter took the subject
out of the box and for the second time, when she slowly
Avoidance (Score). The subject scored 1, if it showed
avoidance toward the familiar experimenter, or a score 0,
if it did not show avoidance (i.e., behaved passively or
Aggressiveness (Score). We categorized the reaction
of the animal as aggressive if it growled or tried to
bite the familiar experimenter. The subject was given
a score 0 if it showed no aggression toward the familiar
experimenter and a score 1, if it growled or tried to
All object-preference tests were videotaped and
analyzed later by one of the authors (B.G.). Interobserver
agreement between her and a naı ¨ve observer on the
behavior categories was assessed by comparing their
parallel coding of the same video records and evalua-
tion of the 22% of the data (eight wolves, eight dogs).
The following Cohen Kappa results were obtained
(Martin & Bateson, 1986): activity¼.85; proximity¼
.93; avoidance¼1; aggressiveness¼1.
Social Behavior of Hand-Raised Dog and Wolf Pups with Humans
Reinforced Eye Contact
In case of three dog and four wolf pups, we tried to
increase the frequency of eye contact with a familiar
experimenter by reinforcing them with food for the re-
quired behavior. The 4-min test sessions were conducted
at a familiar place; first, when the subjects were 5 weeks
old (after the object-preference tests) and later when
of the food reward was tested by placing small pieces
on the floor 1 m from the subject. When we released the
subjects, each of them ran to the food and ate it up
immediately. Then a familiar experimenter (the same
person in all tests) sat on the floor facing the subject
continuously. A plate of food (small pieces of cold cut)
was placed on a table beside the experimenter out of
when it made eye contact with the experimenter, she
signed itwitha clicker (asmall device that givesasudden
snapping sound when pushed) and immediately threw a
1.5 m from the experimenter, she made noise with the
plate to redirect its attention. The caregiver sat still 2 m
away from the experimenter. The whole session was
videotaped. Reviewing the tapes, we counted the number
of ‘‘clicks,’’ which equaled the number of eye contacts
between the experimenter and the animal.
or t-tests were used. In the case of proximity, we have
calculated a preference index as follows: (relative
duration of time spent with caregiver?relative duration
of time spent with other object)/(relative duration of time
spent with caregiverþrelative duration of time spent
with other object). In case of the preference index, first
we tested for divergence from zero (i.e. no preference)
by one-sample t-tests. Then we compared the preference
index of dogs and wolves by one- and two-way ANOVA.
(no time spent in proximity of any of the two objects).
‘Gazing at face’ did not show normal distribution,
therefore nonparametric Mann–Whitney U-test was used
for analysis. ‘Tail-wagging’ ‘avoidance,’ and ‘aggres-
siveness’ were analyzed with Fisher exact test.
Activity. First of all, we wanted to determine whether
there was any difference in the motor ability or general
activity level of the species during the object preference
tests. This was measured by calculating the mean activity
level for the two tests at all three ages. The two-way
ANOVA (species?age, with repeated measures for age)
revealed no difference between the activity level of the
wolf and dog pups at any age (F1,20¼1.878, p¼.186).
However, the subjects spent more time with active
behavior as they got older (F2,40¼7.995; p¼.001)
(Fig. 2). The species-age interaction was not significant
Proximity (Preference Index). To assess the social
behavior after a short isolation, we compared the time
spent in close proximity to the objects presented. Com-
paring the preference index to zero (assuming no
preference for either object presented), we found that
at all or preference for the caregiver. Wolves preferred to
be in the proximity of the caregiver in two tests: at the
age of 3 weeks (Bottle-Caregiver) and at the age of 5
(Experimenter-Caregiver). Dogs tended to spend more
and at the age of 5 (dog-caregiver) (see Tab. 2).
The preference index of dogs and wolves was com-
pared by one-way ANOVA in case of the two pairings
that were carried out only once (Bottle-Caregiver and
conspecific Pup-Caregiver). Two-way ANOVA (with
repeated measures for age) was used to compare the
species in those two pairings (Experimenter-Caregiver
and Dog-Caregiver), which were tested two times (at
and wolves at the age of 3 weeks in the Bottle-Caregiver
spent in activity at the age of 3, 4, and 5 weeks averaged the
results ofthe two object-preference tests. The comparison ofthe
activity values at different age categories by repeated-measures
ANOVA revealed no significant difference between wolves and
was not significant.
Ga ´csi et al.
test (F1,14¼2.54, p¼.135) and also at 4 weeks of age in
Similarly, no effect of the species has been found in
the Experimenter-Caregiver tests (F1,17¼.05; p¼.824),
lacking also the effect of age (F1,17¼.113; p¼.74) and
interaction (F1,17¼1.45; p¼.24). The Dog-Caregiver
tests also showed no overall difference in the social
preferences of the two species (F1,16¼2.16; p¼.16) and
no effect of age was found either (F1,16¼.04; p¼.84).
However, the significant interaction (F1,16¼7.08; p¼.02)
indicates that compared to wolves dogs showed more
pronounced preference for the caregiver when they were
5 weeks old.
of which object showed they less preference toward
the caregiver. Comparing the preference index values in
the two tests at the age of 5 weeks, we found that dog
puppies tended to prefer the caregiver less if the other
object was the unfamiliar human (t8¼2.77, p¼.024)
while wolf pups showed less (actually no) preference
toward the caregiver when the other choice was the
unfamiliar adult dog (t11¼?3.84, p¼.003) (Fig. 3).
Vocalization. During the isolation period, distress voca-
lization was characteristic mainly for dogs. While seven
of eight dogs vocalized (high pitched sounds or howl) at
the room, only three of nine wolves showed similar
behavior. (Unfortunately, data of some subjects could not
be analyzed due to the lost video cassette.)
We have never observed growling and barking during
the object-preference tests with regard of the objects. All
vocalizations were high-pitched sounds (e.g., whining
or yelping: see Cohen & Fox, 1976; Ohl, 1996) thus
reflecting most probably signs of distress because the
animals vocalized mainly when they were not in the
proximity of the objects. Using two-way ANOVA (with
repeated measures for age), we have found that dog
than wolf pups did (F1,20¼11.24; p¼.003) without an
effect of age (F2,40¼1.954; p¼.115). However, sig-
nificant interaction (F2,40¼3.912; p¼.028) indicated
that the tendency for vocalization decreased with age in
dog puppies, while in case of wolf pups, no such change
was evident (Fig. 4).
Tail Wagging. Wolf pups never wagged their tails while
approaching the objects for the first time or while first
getting into physical contact with them. (Even if we
consider the entire period of tests, therewas only one pup
Table 2. Results of the Comparison of Preference Indexes in all Tests
3 Weeks4 Weeks5 Weeks
t5¼3.83, p .05
t9¼2.72, p .05
t9¼12.131, p .01
were corrected with False Discovery Rate adjustment (Benjamini, Drai, Elmer, Kafkafi, & Golani, 2001). In all
(&, filled bars) and the ‘‘Caregiver-Experimenter’’ (, dotted
bars) object-preference tests. Index values different from zero
The index values were calculated as: (relative duration of time
spent with caregiver?relative duration of time spent with
other stimulus)/(relative duration of time spent with caregi-
verþrelative duration of time spent with other stimulus).
A comparison of the index values from the two types of tests
at the same age reveals in which test subjects showed greater
less when the other object was the experimenter, while wolves
showed actually no preference toward their caregiver in the
presence of an adult dog (paired t-tests). Significant differences
are indicated with asterisks (?p<.05,??p<.01).
Preference index values of 5-week-old dogs (two
Social Behavior of Hand-Raised Dog and Wolf Pups with Humans
that showed this behavior orienting toward the adult dog
and also toward the caregiver in one test at the age of
tail wagging in wolves during the observed periods, the
two groups did not differ statistically in the tendency for
showingtailwagging inrespectofanyobject attheageof
3 weeks. In the proximity of the nursing bottle and the
adult dog, none of the animals showed tail wagging
and only few dog puppies (two subjects in the Bottle-
Caregiver and three in the Dog-Caregiver test) wagged
their tail when approached or contacted the caregiver.
However, 4- and 5-week-old dog puppies significantly
differed from wolf pups in case of all objects. Compared
to wolves, more 4-week-old dog puppies showed some
tail wagging toward the caregiver both in the Experi-
menter-Caregiver (Fisher exact test; p<.001) and in the
Pup-Caregivertest (p¼.029). We found the same species
difference in tail wagging in Experimenter-Caregiver test
Pup-Caregiver test toward the pup (p¼.029). This
difference was also characteristic in case of 5-week-old
subjects as dogs wagged their tail more often toward all
objects (Dog-Caregiver test: caregiver: p<.001; adult
p¼.001; caregiver: p<.001). In all but one occasions,
dogs wagged their tails horizontally or in a high position
rather than holding it low or between the legs.
Gazing at Face. As we found no difference between the
their face (dogs: z¼?1.69, p¼.09; wolves: z¼?.45,
p¼.66 by Wilcoxon test), and considering the relative
rare occurrence of this behavior, we added up the number
of gazings at any human face during the tests at a certain
Three- and 4-week-old animals gazed rarely at the
human face, and no difference was found between the
species (Nd,w¼8;11, U¼38.50, p¼.241; Nd,w¼11;11,
U¼40.00, p¼.104, respectively). At the age of 5 weeks,
however, dog puppies gazed at the humans’ face more
often than wolf pups did (Nd,w¼11;12, U¼31.00;
p¼.02) (Fig. 5).
Avoidance. Considering all three age categories together
only one dog showed avoidance on 1 occasion (1%) of 98
interactions (taking up the puppy by the experimenter)
in contrast to eight wolves displaying such behavior on
19 occasions (17%) of 106 interactions. Due to the small
sample size and the relatively rare occurrence of this
behavior, however, this did not mean significant differ-
ence between the species in avoidance shown toward a
familiar experimenter at the age of 3 and 4 weeks (Fisher
exact test; p¼.214 and .242, respectively) and only a
Aggressiveness. None of the dog puppies behaved ag-
gressively in the 99 interactions with the familiar
experimenter during the tests. Among the 13 wolf pups,
humans’ face in the three age categories during the two tests.
The medians of nonparametric data are represented by bold
lines, and boxes indicate the 50% of the data (lower and upper
interquartile range). Whiskers extend to the smallest and largest
values excluding outliers and extremities. Mann–Whitney U-
tests showed that 5-week-old dogs gazed at the humans’ face
significantly more frequently than wolves did. Significant
differences are indicated with asterisks (?p<.05).
Frequency (number/min) of gazings at the
spent with vocalization at the age of 3, 4, and 5 weeks averaged
the results of the two object-preference tests. The values of the
two species at different age categories were compared by
repeated-measures ANOVA. Dogs spent more time vocalizing
than wolves, with no effect of age. Significant interaction
indicated that vocalization tended to decrease with age in the
Mean value (þSE) of relative durations of time
Ga ´csi et al.
however, there were nine individuals that one or more
times growled at the familiar experimenter and/or tried
to bite her (in 29 cases of 112 interactions, 26%).
Comparing the species by age-categories, we have
found that aggressive behavior was more pronounced
in wolves than in dogs at each age (Fisher exact test:
3-week-olds, p¼.04; 4-week-olds, p¼.013; 5-week-
Eye Contact. This experiment was carried out only with
be regarded as preliminary. Subjects were tested both at
the age of 5 and 9 weeks, but most 5-week-old wolf pups
could be analyzed for this age.
Comparing the number of eye contacts at the age of
9 weeks, no difference was found between the perfor-
mance of the species in the first minute (t5¼.985,
p¼.370). At the beginning, all animals made intensive
attempts to take the food directly from the plate, which
was unreachable for them. As the session went on,
however, dogs tended to gaze more at the experimenter’s
face. Wolves, on the contrary, kept mainly orienting
from the experimenter if they happened to gaze at her.
Compared to wolves, dogs achieved significantly
more eye contact with the experimenter during the
fourth minute (t5¼4.811, p¼.005).
To look for any learning effect across the two ages in
dogs, we compared their performance (difference be-
tween number of eye-contact in min 4 and min 1) at the
age of 5 and 9 weeks, and found very similar pattern of
performance (t2¼?.615, p¼.601) (Fig. 6).
the social behavior of 13 wolves and 11 dogs reared in
identical human social environment and aiming to reveal
the species-specific differences during the course of
addressed in past studies (e.g., Zimen, 1987), the analysis
of experimental data on the social relationship and
interaction between the socialized animal and its human
great importance to understand specific behavioral
characteristics indogs, which are assumed tobe the result
of the domestication process.
The effects observed in this study can be grouped into
three main categories. First, the behavior differences
could be affected by maturation. According to our data,
the activity level of dog and wolf pups did not differ
significantly. Therefore, it is unlikely that species
differences are related to differing motor abilities;
however, we cannot exclude that wolf pups may develop
faster in other respects.
Second, preference to stay in the proximity of the
caregiver could be the result of the interaction of various
and not necessarily exclusive processes. The behavior
of the subjects could be controlled by their familiarity
to the objects, revealing itself in a tendency to approach
the familiar, or alternatively, to explore the (relatively)
novel, less familiar stimulus. In case of attachment,
it is generally assumed that for young animals, it is
more advantageous to ‘‘choose’’ the familiar stimulus
(‘‘parent’’) in a novel and hazardous environment
is, the subject identified one of them as being the same or
(more) similar toits‘‘parent.’’Onewaytoshowthe effect
of learning is to manipulate the ‘‘parent’’ and look for
preference changes in choice tests. In our case, both dogs
and wolves had human ‘‘parent,’’ so the question was
whether similar experience affected their choice behavior
differently, or in other words even similarly intensive
socialization could not compensate for different learning
Wolves preferred the caregiver to the bottle at the age
of 3 weeks, which affirms the findings that in some
contexts, social stimuli have higher incentive value than
food reinforcement for wolf pups (Frank & Frank, 1988).
Dogs displayed preference for the caregiver starting from
the 4th week, and this preference disappeared only in the
experimenter during the 4-min long operant conditioning
test sessions of dogs at the age of 5 and 9 weeks versus that of
be evaluated.) Comparing the number of eye contacts in the
presence of a plate with food, 9-week-old dogs’ and wolves’
behavior did not differ during the first minute. During the last
minute, however, dogs initiated more eye contacts than wolves.
the age of 5 and 9 weeks.
Mean number of eye contacts with a familiar
Social Behavior of Hand-Raised Dog and Wolf Pups with Humans
last test when they were tested with two humans. This
could suggest that dogs develop a preference toward
humans as they get older, and additionally this preference
is not restricted to the caregiver but becomes more
generalized by the age of 5 weeks. It is also possible that
dog puppies are more sensitive to being ‘‘ignored’’ by
their motionless caregiver than are wolf pups, thus
are more prone to seek contact (try to initiate social
interaction) alsowith the experimenter.Wolf pups show a
clear preference for the human individual recognized as
when the other human is either a novel stimulus or is a
canid (adult dog or wolf pup). It could be assumed that
‘‘noncaregiver’’ social stimuli evoked exploratory beha-
vior on the part of the wolf. This is supported by the
observation that wolves showed no preference for the
for this time the (relatively) novel stimulus may evoke
some fearfulness besides the exploratory behavior in
Taken together, this suggests that both species are able
to learn about its heterospecific (human) foster parent but
there might be differences in the ability to generalize its
characteristics to other similar ‘‘objects.’’ If we assume
set of features of the parent then the difference may be
that wolves are more restrictive in their choice when
the caregiver is highly different from the natural parent.
the ‘‘default’’ state of wolves’ learning system, domes-
tication might have changed this in dogs by making the
recognition process less precise, that is, recognition can
occur in greater ranges of these characteristic features.
This relaxed (or generalized) recognition in case of
nonconspecificsmighthavebeen advantageous foryoung
dogs living in human setting, which have been raised
a group of humans (‘‘family’’).
Third, in the case of some communicative signals,
is supported generally by the observation that dogs are
morevocal in comparison towolves (Fox, 1971a). Young
wolves could be more prone to situations of being left
alone in comparison to dogs, or vocalization while alone
would make a wolf pup more vulnerable to predators,
but selection against this trait has been relaxed in dogs
living in the human social environment lacking predators
(Fox, 1971a, see also Frank & Frank, 1982a).
Wolf pups showed more avoidance and aggression
toward a familiar human, though the observed difference
humans. The greater number of growls and attacks (e.g.,
attempted biting) in wolves and the absence of these
behaviors in the case of dogs could be best explained by
supposing that wolves either did not like to be touched
or constrained in their movements, or they had a lower
threshold for the elicitation of aggressive behavior.
Humans must have successfully selected against such
behaviors in the case of dogs. A recent comparative study
observing young wolves’ and dogs’ attachment behavior
toward their caregiver in the Strange Situation procedure
has also shown differences in temperament traits such
as approach-avoidance behaviors (Topa ´l et al., 2005).
These findings might also help in finding new methods
for wolf–dog hybrid identification (e.g., comparing the
aggressive–avoidance behavior of animals of unknown
origin with that of dogs) even at very early age.
Although tail wagging is listed as a behavior unit in
the ethogram of the wolf (McLeod, 1996; McLeod &
observed usually in the context of active submission.
Although this behavior has been observedin wolf pups as
early as 3 weeks of age (Fox,1971a; McLeod & Fentress,
1997; Zimen, 1987), and was present also in our pups’
behavior during interactive social contexts, we did not
found tail wagging in the wolves during the observed
periods. It is also interesting to note that tail wagging
became also more frequent in the foxes selected for
tameness (Belyaev, 1978). Considering the context of our
test arrangement, dog puppies are either more prone to
show submission toward passive social stimuli at the age
of 4–5 weeks, or may use tail wagging for a somewhat
broader communicative intention to facilitate interaction.
give some support to the latter explanation.
Further differences were found in relation to gazing
behavior toward humans. Despite extensive socialization
with humans, wolves seem to avoid looking at the face
of the experimenter, which was revealed in the low
frequency of gazing in the object-preference tests. Their
behavior in the eye contact test 4 weeks later supports the
idea that this difference cannot be explained simply by a
delay in the development of their social communication
As wolves proved to be motivated by the food when
they could get it for ‘‘free’’ prior to the eye contact test
at both ages, we suggest that their reaction could be
explained by the strategy observed in wolves in other
situations as well when responding to ‘‘unsolvable’’
had a rest (Miklo ´si et al., 2003).
In the conditioning test, the required behavior was a
very simple one and food reinforcement always fol-
lowed the eye-contact right away. There are two possible
Ga ´csi et al.
explanations for increased tendency to gaze at the
human’s face in dogs; either they learned the association
between eye-contact and food reward very quickly, or
they might have not learned much, but in this moderately
only because ‘‘solicitation’’ came more natural to them.
In the first case, the difference in the performance
there might be some other type of learning difference
between the two species in this situation (i.e., dogs were
simply more quick in this conditioning task).
However, comparative studies on the problem solving
abilities and species-specific constraints on learning in
the two species have revealed no inferior performance of
wolves (Frank & Frank, 1985, 1988; Frank et al., 1989).
Moreover, in other tests, we found some plasticity in
young wolves’ willingness to look into the eyes of a
familiar experimenter, which supports that after exten-
sive training, learning can overcome initial differences
(Vira ´nyi et al., 2005).
When comparing the gazing behavior of the two
species, it is important to note that in the wolf, gazing
plays a crucial role in agonistic communication, that is,
dominants express threat by gazing at the subordinates,
and both the dominant and the subordinate can avoid
conflict by trying to avoid gaze contact (Fox, 1971a,
Schenkel, 1967). Although no direct quantitative beha-
vioral comparisons are known, it is generally assumed
that the use of gazing in intraspecies aggressive interac-
wolf (Fox, 1971b), and in dogs, direct staring can release
very easily aggressivebehavior or attack. Additionally, in
humangazingaccompanied with otherbehavioralcues as
indicationofthreat(Vas,Topa ´l,Ga ´csi,Miklo ´si,&Csa ´nyi,
2005). However, with respect to nonaggressive commu-
nicative interactions, we have described that wolves and
dogs differ in their willingness to look at the humans face
(Miklo ´si et al., 2003). At present, two different, not
exclusive processes can account for the difference
between the wolf and the dog. Dogs could have been
selected for being more resistant to gazing, that is, only
more extended gazing would induce subordinate or
agonistic behavior, or dogs could be selected for gazing
preferentially at humans. Interestingly, on the basis of
comparativeinvestigations, Frank and Frank (1985) seem
to support the latter view when noting that dogs might
have been selected for soliciting human intervention.
Elsewherewe have also proposed that increased gazing at
social skills in dogs that are able to utilize this visual
communicative channel used predominantly by our
species (Miklo ´si et al., 2003, Miklo ´si et al., 2004).
learn about the human caregiver, their social preferences
differentiate during early development mainly because
human and conspecific social stimuli affect their prefer-
In addition, the behavior of the dog puppies is char-
acterized by less aggression and avoidance toward
humans in parallel with the increase in communicative
signals such as vocalization, tail wagging, and gazing,
which can provide a basis for positive feedback on inter-
specific dog–human interaction.
The authors thank the hand raising caregivers for their devotion
and assistance during the project. We are grateful for the
cooperation of Zolta ´n Horkai in helping the hand raising of the
wolf pups and providing home for them at his wolf-park after
Belyaev, D. K. (1978). Destabilizing selection as a factor in
domestication. Journal of Heredity, 70, 301–308.
Belyaev, D. K., Plyusnina, I. Z., & Trut, L. N. (1985).
Domestication in the silver fox (Vulpes fulvus): Changes in
physiological boundaries of the sensitive period of primary
Benjamini, Y., Drai, D., Elmer, G., Kafkafi, N., & Golani, I.
(2001). Controlling the false discovery rate in behavior
genetics research. Behavioural Brain Research. 125, 279–284.
Call, J., Bra ¨uer, J., Kaminski, J., & Tomasello, M. (2003).
Domestic dogs are sensitive to the attentional state of
humans. Journal Of Comparative Psychology, 117, 257–263.
Cohen, J.A., & Fox, M. W. (1976). Vocalization in wild canids
and possible effects of domestication. Behavioral Processes,
Coppinger, R. P., & Coppinger, M. (2002). Dogs. A new
understanding of canine origin, behavior and evolution.
Chicago: University of Chicago Press.
Feddersen-Petersen, D. (1986). Observations on social play in
some species of Canidae. Zoologischer Anzeiger, 217, 130–
Fentress, J. C. (1967) Observations on the behavioral devel-
opment of a hand-reared male timber wolf. American
Zoologist, 7, 339–351.
Fox, M. W. (1970). Behavioral effects of rearing dogs with cats
during the ‘critical period of socialisation. Behavior, 35,
Fox, M. W. (1971a). Behavior of wolves, dogs and related
canids. London: The Trinity Press.
Fox, M. W. (1971b). Integrative development of brain and
behavior in the dog. Chicago: University of Chicago Press.
Fox, M. W., & Stelzner, D. (1966). Approach/withdrawal
variables in the development of social behavior in the dog.
Animal Behavior, 14, 362–366.
Social Behavior of Hand-Raised Dog and Wolf Pups with Humans
Frank, H. (1980). Evolution of canine information proces-
sing under conditions of natural and artificial selection.
Zeitschrift fu ¨r Tierpsychologie, 59, 389–399.
Frank, H., & Frank, M. G. (1982a). On the effects of domestica-
tion on canine social development and behavior. Applied
Animal Ethology, 8, 507–525.
Frank, H., & Frank, M. G. (1982b). Comparison of problem-
solving performance in six-week-old wolves and dogs.
Animal Behavior, 30(1), 95–98.
Frank, H., & Frank, M. G. (1985). Comparative manipulation-
test performance in ten-week-old wolves (Canis lupus) and
Alaskan malamutes (C. familiaris): A Piagetian interpreta-
tion. Journal of Comparative Psychology, 99, 266–274.
Frank, H., & Frank, M. (1987). The University of Michigan
Canine Information Processing project (1979–1981). In
H. Frank (Ed.), Man and wolf (pp. 275–292). Dordrecht, The
Netherlands: W. J. Publishers.
Frank, M. G., & Frank, H. (1988): Food reinforcement versus
social reinforcement in timber wolf pups. Bulletin of the
Psychonomic Society, 26, 467–468.
Frank, H., Frank, M.G., Hasselbach, L.M., & Littleton, D. M.
(1989). Motivation and insight in wolf (Canis lupus) and
Alaskan malamute (Canis familiaris): Visual discrimination
learning. Bulletin of Psychonomic Society, 27(5), 455–458.
Fuller, J. L. (1967). Experiential deprivation and later behavior.
Science, 158, 1645–1652.
Ga ´csi, M., Miklo ´si, A´., Varga, O., Topa ´l, J., & Csa ´nyi, V.
(2004). Are readers of our face readers of our minds? Dogs
(Canis familiaris) show situation-dependent recognition of
human’s attention. Animal Cognition, 7(3), 144–153.
Hare,B., Brown,M.,Williamson, C.,& Tomasello,M. (2002). The
domestication of cognition in dogs. Science, 298, 1634–1636.
Klinghammer, E., & Goodmann, P. A. (1987). Socialization and
management of wolves in captivity. In H. Frank (Ed.), Man
and wolf (pp. 31–59). Dordrecht: W. J. Publishers.
Kubinyi, E., Topa ´l, J., Miklo ´si, A´., & Csa ´nyi, V. (2003) Dogs
learn from their owner via observation in a manipulation
task. Journal of Comparative Psychology, 117, 156–165.
MacDonald, K., & Ginsburg, B. E. (1981). Induction of normal
prepubertal behavior in wolves with restricted rearing.
Behavioral and Neural Biology, 33, 133–162.
McKinley, J., & Sambrock, T. D. (2000). Use of human-given
cues by domestic dogs (Canis familiaris) and horses (Equus
caballus). Animal Cognition, 3, 13–22.
McLeod, P. J. (1996). Developmental changes in associations
among timber wolf (Canis lupus) postures. Behavioral
Processes, 38, 105–118.
McLeod, P. J., & Fentress, J. C. (1997). Developmental changes
in the sequential behavior of interacting timber wolf pups.
Behavioral Processes, 39, 117–136.
Martin, P., & Bateson, P. (1986). Measuring behavior. Cam-
bridge: Cambridge University Press.
Miklo ´si, A´., Polga ´rdi, R., Topa ´l, J., & Csa ´nyi, V. (2000). Inten-
tional behaviour in dog-human communication: An experi-
mental analysis of ‘‘showing’’ behaviour in the dog. Animal
Cognition, 3, 159–166.
Miklo ´si, A´., Kubinyi, E., Topa ´l, J., Ga ´csi, M., Vira ´nyi, Zs., &
Csa ´nyi, V. (2003). A simple reason for a big difference:
Wolves do not look back at humans but dogs do. Current
Biology, 13, 763–766.
Miklo ´si, A´., Topa ´l, J., & Csa ´nyi, V. (2004.) Comparative social
cognition: What can dogs teach us? Animal Behavior, 67,
Ohl, F. (1996). Ontogeny of vocalisation in domestic dogs,
breed standard-poodle. Bonner Zoologische Beitrage, 37,
Pongra ´cz, P., Miklo ´si, A´., Kubinyi, E., Gurobi, K., & Csa ´nyi, V.
(2001). Social learning in dogs I: The effect of a human
demonstrator on the performance of dogs (Canis familiaris)
in a detour task. Animal Behavior, 62, 1109–1117.
Rajecki, D. W., Lamb, M. E., & Obmascher, P. (1978). Toward
a general theory of infantile attachment: A Comparative
review of aspects of the social bond. Behavioral and Brain
Sciences, 3, 417–464.
Sackett, G. P., Porter, M., & Holmes, H. (1964). Choice
behavior in rhesus monkeys: Effect of stimulation during the
first month of life. Science, 147, 304–306.
Schenkel, R. (1967). Submission: Its features and function in
wolf and dog. American Zoologist, 7, 319–329.
Scott, J. P., & Fuller, J. L. (1965) Genetics and the social
behavior of the dog. Chicago: University of Chicago Press.
Soproni, K., Miklo ´si, A´., Topa ´l, J., & Csa ´nyi, V. (2001).
Comprehension of human communicative signs in pet dogs.
Journal of Comparative Psychology, 115, 122–126.
Topa ´l, J., Ga ´csi, M., Miklo ´si, A´., Vira ´nyi, Zs., Kubinyi, E., &
Csa ´nyi, V. (2005) Attachment to humans: A comparative
study on hand-reared wolves and differently socialized dog
puppies. Animal Behavior (in press).
Ujfalussy, D., Vira ´nyi, Zs., & Kubinyi, E. (2003) Individual and
intensive hand-rearing of wolves: The method and its results
in handling the wolves. ASAB 2003 Summer Meeting,
Gru ¨nau, Austria.
Vas, J., Topa ´l, J., Ga ´csi, M., Miklo ´si, A´., & Csa ´nyi, V. (2005) A
friend or an enemy? Dogs’ reaction to an unfamiliar person
showing behavioral cues of threat and friendliness alter-
nately. Applied Animal Behavior (in press).
VilA˜, C., Savolainen, P., Maldonado, J. E., Amorim, I. E., Rice,
J. E., Honeycutt, R. L., Crandall, K. A., Lundeberg, J., &
Wayne, R. K. (1997). Multiple and ancient origins of the
domestic dog. Science, 276, 1687–1689.
Vira ´nyi, Zs., Topa ´l, J., Ga ´csi, M., Miklo ´si, A´., & Csa ´nyi, V.
(2004). Dogs respond appropriately to cues of humans’
attentional focus. Behavioral Processes, 66, 161–172.
Vira ´nyi, Zs., Miklo ´si, A´., Ga ´csi, M., Kubinyi, E., Topa ´l, J.,
Bele ´nyi, B., Ujfalussy, D., & Csa ´nyi, V. (2005). Comprehen-
sion of human pointing gestures in young socialized wolves
and dogs (submitted).
Wayne, R. K. (1993). Molecular evolution of the dog family.
Trends in Genetics, 9, 218–224.
Woolpy, J. H., & Ginsburg, B. E. (1967). Wolf socialization:
A study of temperament in a wild social species. American
Zoologist, 7, 357–363.
Zimen, E. (1987). Ontogeny of approach and flight behavior
towards humans in wolves, poodles and wolf–poodle
hybrids. In H. Frank (Ed.), Man and wolf (pp. 275–292).
Dordrecht, The Netherlands: W. J. Publishers.
Ga ´csi et al.