Content uploaded by Zsuzsánna Horváth Igyártó
Author content
All content in this area was uploaded by Zsuzsánna Horváth Igyártó on Oct 26, 2022
Content may be subject to copyright.
Affiliative and disciplinary behavior of human handlers during play with
their dog affects cortisol concentrations in opposite directions
Zsuzsánna Horváth ⁎, Antal Dóka, Ádám Miklósi
Department of Ethology, Institute of Biology, Eötvös Loránd University Faculty of Science, Budapest, Hungary
Received 29 August 2007; revised 31 January 2008; accepted 1 February 2008
Available online 15 February 2008
Abstract
It has been shown that cortisol concentrations change characteristically in the course of agonistic interactions; our aim was to find out how a
playful situation may affect concentrations of this hormone in the saliva. We studied dogs' behavior and the changes of cortisol concentrations in a
play situation, where the dogs played with their handler for 3 min with a tug toy. In this experiment working dogs were divided into two groups by
the type of their work, namely police dogs and border guard dogs. We found that the cortisol concentrations of old police dogs significantly
increased, while the adult border guard dogs' hormone levels decreased, which shows that playing with the handler has an effect on both groups,
but interestingly this effect was opposite. Behavior analysis showed differences only in the behavior of the human handlers during the play
sessions, while the behavior analysis did not reveal significant differences in the two groups of dogs, except that old border guard dogs generally
needed more time to begin playing than old police dogs. During the play sessions police officers were mainly disciplining their dogs, while the
border guards were truly playing with them (including affiliative and affectionate behavior). Our results are in accordance with those of recent
studies, which show that behaviors associated with control, authority or aggression increase cortisol concentrations, while play and affiliative
behavior decrease cortisol levels.
© 2008 Elsevier Inc. All rights reserved.
Keywords: Play; Stress; Cortisol; Communicative signals; Working dogs
Introduction
Play is a characteristic behavior of young animals (Bekoff,
2001), which is composed of action patterns that are used in
various contexts, such as exploration, manipulation, locomo-
tion, predation, fighting and mating (Hol et al., 1994), or
combinations of these (Loizos, 1966). Although its evolutionary
origin is not clear, many hypothesized functions of play activity
have been proposed including its promotional effect on physical
and social skills (Byers and Walker, 1995; Drea et al., 1996;
Dugatkin and Bekoff, 2003; Holmes, 1995; Soderquist and
Serena, 2000).
Play is often used as an indicator of well being (for a review
see e.g. Boissy et al., 2007) as animals exposed to traumatic
stimulation during development, or live in impoverished
environments, show reduced play activity. If we define stress
as the non-specific response of the body to any demand for
change (Selye, 1950), namely as an increase in HPA axis
activity caused by any physical, psychological or environmental
stimulus (e.g. Sapolsky et al., 2000), then it seems that
environmental stress can have an inhibitory effect on the
expression of play behavior. However, there could also be an
opposite relationship between play and stress. Some recent
studies have shown that play activity may contribute to the
reduction of stress (Arelis, 2006). Rats stimulated by novel
objects increase their play activity and at the same time show
reduced anxiety (Darwish et al., 2001), and in chimpanzees the
intensity of playing increases before feeding times which are
usually associated with high levels of social tension (Palagi
et al., 2004). These latter authors assume that this behavior
A
vailable online at www.sciencedirect.com
Hormones and Behavior 54 (2008) 107 –114
www.elsevier.com/locate/yhbeh
⁎Corresponding author. Department of Ethology, Institute of Biology, Eötvös
Loránd University Faculty of Science, Pázmány P. s. 1/C 1117 Budapest,
Hungary. Fax: +36 1 381 2180.
E-mail address: horvath.zsuzsanna@gmail.com (Z. Horváth).
0018-506X/$ - see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.yhbeh.2008.02.002
might have a ‘preventive’effect to reduce subsequent social
stress elicited by the feeding situation. Humans also utilize play
routinely to familiarize children with novel situations including
research settings.
It is often stated that in comparison to their wolf (Canis lupus)
ancestors, the dog (Canis familiaris) is a very playful species,
since dogs play levels remain high throughout their life (Bekoff,
1972; Lorenz, 1950). Dogs are valued members of human social
groups partly as a result of their eagerness to engage in inter-
specific play (Bekoff, 1972; Fagen, 1981; Mitchell and Thomp-
son, 1993; Rooney et al., 2001; Russell, 1936), and many owners
spend a considerable amount of their time playing with their dogs
(Hart, 1995).
As playful interaction between dogs and humans represents a
natural activity for both species (see above), in the present study
we used this behavior to investigate the relationship between
play and stress by measuring saliva cortisol concentrations
before and after dog–human playing interaction.
In social situations other than play, it has been shown that
human behavior toward the dog has significant effects on its
physiological state (e.g. change of cortisol concentrations, heart
rate variability). For example, the presence of humans in a
shelter (Beerda et al., 1997; Hennessy et al., 1998; Tuber et al.,
1996) may be an effective means of reducing the cortisol
response of dogs. In a recent study, Jones and Josephs (2006)
found that after agility competitions there is a significant
correlation between humans' punitive behaviors (physically
pushing the dog and yelling at it) and the increase in dogs'
cortisol concentrations in the losing teams. We also recently
reported that threatening behavior shown by humans resulted in
increasing cortisol concentrations in dogs (Horváth et al., 2007).
Thus in this study we hypothesized that the actual motor and
communicative aspects of behavior of the human partners
during play affect playing experience and inner state of the
dogs. Since pilot observations showed that there is a variation in
the behavior of working dog handlers toward their dog
(policemen and border guards, see below), this offered a good
opportunity for testing our hypothesis. In our experiments we
asked handlers to interact in a playful manner with their dog. To
reveal how dogs respond to such a situation, we simultaneously
gathered behavioral data from both dogs and handlers and
measured the saliva cortisol concentrations of dogs before and
after the interaction. The experimental protocol was designed to
include situations such as playful struggle with the handler (tug-
of-war game), but also cooperative behaviors like the retrieval
of the toy or giving the toy up to the handler. Using this method
we aimed to document the changes of cortisol concentrations
that take place during the playful episodes in dogs, in parallel to
the actual behavior of the handlers.
Materials and methods
Subjects
The police dogs were purchased by the Hungarian National Police Training
School for Police Dog Handlers (Dunakeszi, Hungary). Dogs were acquired
between 1 and 3 years of age. Dogs were tested physically (i.e. for hip dysplasia)
and behaviorally (i.e. reaction to gun shot, bite work). Individuals were
purchased only if they did not show signs of hip dyplasia and fear of gun shot.
Thereafter they participated in a 12 week training course together with their
handlers. During this course the dogs were trained for guarding and obedience.
Dogs that participated in the present study were purchased between 1997
and 2003, and performed patrol service with their handlers on the streets for a
minimum of 1 year. All 84 subjects were male German Shepherd Dogs. The
dogs' age ranged from 2 to 11 years (mean age± SD: 6.91± 2.19 years) and the
subjects were categorized following Studzinski et al. (2006):Adult dogs were
2–7 years old (43 individuals; mean age ±SD: 5.18 ±1.45 years); Old dogs were
8–11 years old (41 individuals; mean age±SD: 8.76 ± 0.94 years). We also
divided the dogs into two work groups (‘occupation’): Police dogs (53
individuals; mean age±SD: 7.26 ± 2.05 years; 25 adult dogs (mean age ± SD:
5.52± 1.44 years) and 28 old dogs (mean age ± SD: 8.82 ± 0.94 years)), and
Border Guard dogs (31 individuals; mean age± SD: 6.42± 2.32 years; 18 adult
dogs (mean age± SD: 4.77± 1.47 years) and 13 old dogs (mean age± SD: 8.61±
0.96 years)). Dogs of both groups were selected according to the same criteria
initially and they also took part in identical training programs. The only
differences were that some dogs were handled by policemen and the others by
border guards and also in the nature of their work after the training period. Police
dogs and their handlers are assigned to patrol duty on the streets, while border
guard dogs and their handlers work the borders of Hungary in the countryside.
Policemen and border guards were also selected with similar criteria. Eighty two
of the handlers were men and 2 were women.
All procedures were approved by the Ethical Committee of Eötvös Loránd
University, Department of Ethology and conducted in accordance with the
Hungarian State Health and Medical Service (ÁNTSZ). There is a standing
agreementwith the HungarianPolice Force that permits testing their working dogs.
Date and premises
The experiments were carried out in 2005 and 2006 at the Hungarian
National Police Training School for Police Dog Handlers (Dunakeszi, Hungary).
The police dogs that were tested, participated in a special 2 week training course
with their handlers during the time of testing. The experiments were conducted
in an empty room (10 m long×10 m wide × 5 m high), where only the
experimenter (who was also the camera person), the handler and the dog were
present during the test. The experimental room was familiar for the subjects
(some of the training exercises take place at this location), as it has been shown
that introduction into a novel environment enhances HPA activity in the dog
(e.g. Beerda et al., 1997). All salivary samples were taken between 9 a.m. and 3
p.m. (e.g. Dreschel and Granger, 2005; Jones and Josephs, 2006). The dogs were
in their kennels resting for 30 min before starting the test.
Procedure
Dogs played with their handler for 3 min in a similar manner to that which is
described by Rooney and Bradshaw (2002). The handler could use a rag or a tug
toy (a piece of thick rope with two knots on both ends, 20 cm long) for inducing
play behavior in the dog. The handlers were instructed to play as intensively as
possible with the dog, and to adjust their behavior to the dog's reactions. Before
the test, the handler was asked to leave the dog's toys or food items outside the
experimental room. The handlers were asked to encourage the dog during the
entire trial, even if the dog displayed only slight or no inclination to play. We
imposed few restrictions in the type of play; however the handlers were instructed
to execute the following acts at least once during a play session: (1) they had to
throw the object and encourage the dog to bring it back; (2) they had to try to take
the object from the dog's mouth.
Measurement of saliva cortisol concentrations
Saliva samples were collected from the dogs before the play session and
20 min after the end of the play session (see also Beerda et al., 1998; Dreschel
and Granger, 2005; Jones and Josephs, 2006; Vincent and Michell, 1992).
Substances to stimulate saliva flow were not used. The saliva was collected
with cotton swabs by the handlers near to the location of the experiment.
While the dog was standing still, the handler placed the swab into the mouth
of the dog and held it there until it absorbed the greatest amount of saliva
possible (lasting from 30 to 60 s). The soaked cotton swabs were temporarily
108 Z. Horváth et al. / Hormones and Behavior 54 (2008) 107–114
stored on dry ice in numbered Eppendorf tubes. For long term storage the
saliva samples were kept in a deep freezer (−80 °C). Before analysis the tubes
were warmed up to room temperature. The saliva was removed from the
cotton swabs by centrifugation (3000 rpm for 15 min) using special centrifuge
tubes with filters (Corning Spin-X; Sigma-Aldrich Kft., Budapest, Hungary).
After separation the saliva samples were analyzed for cortisol concentrations
using a highly sensitive (from 0.003 to 3.0 μg/dl) enzyme immunoassay kit
from Salimetrics (State College, PA, USA); the intra- and inter-assay coef-
ficients of variation as provided by the manufacturer, are below 10% and 15%
respectively (Salimetrics, 2005). The procedures were performed on the basis
of the protocol provided by Salimetrics. Before calculating concentrations,
log transformations were used to establish normal distributions. All analyses
used the log-transformed hormone values. However, non-transformed data are
reported in the figures to facilitate interpretation. All participating dogs
provided the required amount of saliva and all cortisol measurements could be
used for the statistical analysis.
Table 1
Behavioral variables coded in dogs and handlers, where 0 points are given (score 0) in case of the behavior that is least favorable from the point of view of our
cooperative situation, while 2 points (score 2) are given in case of the most favorable behavior
Behavioral variables Score 0 Score 1 Score 2
DOGS ‘Motivation’There is no physical contact between
the dog and the play object (e.g. the
dog never holds or chews the object
during the session)
The dog makes physical contact (with his
mouth or paw) with the object at least
once
There are two or more
physical contacts
made between the
dog and the object
‘Playfulness’The dog is passive, pays no attention
to play object
The dog shows interest (looks at, contacts,
etc.) in playing with the object at least one
time
The dog pays attention
to the play object
during the entire
session or more than
one time
‘Willingness to retrieve’The dog never brings the object to the
handler or there is no physical contact
between the dog and the object (e.g. the
dog never holds or chews the toy during
the session)
The dog makes steps to move towards
the handler with the object in its mouth,
but the handler cannot get the object
without approaching the dog
The dog usually
brings the object back
to the handler
‘Possessivity’(this behavior variable was
not coded when the dog was given Score
0at‘Motivation’and/or ‘Playfulness’)
The handler is unable to take the object
from the dog during the session even
with force
There are visible signs of resistance when
the handler tries to take the object or the
dog shows avoidance with the object in its
mouth, but finally the handler can take it
from the dog
The handler can take the
object from the dog's
mouth (without any
visible signs of force)
‘Aggression’(signs of aggression after
Kim et al. (2006), Kroll et al. (2004),
McLeod (1996), Netto and Planta (1997)
and Pal et al. (1998): tail is not tucked,
the ears are forward or ambivalent, stiff
body posture; growling, barking,
snarling, showing the teeth)
No signs of aggression can be observed
during the play session
The dog shows at least one of the behaviors
described during the play session
‘Fear’(signs of fear after Dreschel and
Granger (2005) and Kroll et al. (2004):
the dog's tail is tucked between its hind
legs, hides behind the human's leg, backs
away or retreats, ears are pinned or drawn
back or down)
No signs of fear can be observed during
the play session
The dog shows at least one of the
behaviors described during the play
session
‘Latency of starting to play’(sec; from
the handler's first call to play till the dog
takes the tug toy)
HANDLERS ‘Control commands’(number of verbal
orders issued by the handler during the
3 min session, e.g. using words of
command: sit —ül, stand —áll, lay —
fekszik, heel —lábhoz, listen —figyelj)
Does not occur Occurs between 1 and 5 times Occurs 6 or more times
‘Sound signal' (number of non-verbal
sound signals used by the handler during
the 3 min session to attract the dogs'
attention, e.g. whistle, clapping, etc.)
Does not occur Occurs between 1 and 5 times Occurs 6 or more times
‘Enthusiasm of handler’The handler does not laugh or smile
during the play session
The handler is cheerful during the play
session
‘Praising’(handler praises the dog or
speaks to the dog kindly, in a high pitched
or fluctuating voice)
Does not occur Occurs between 1 and 5 times Occurs 6 or more times
‘Petting head and body’(handler extends
hand to touch or pet the dog anywhere on
the head or body)
Does not occur Occurs between 1 and 5 times Occurs 6 or more times
109Z. Horváth et al. / Hormones and Behavior 54 (2008) 107–114
Analysis of behavior
The behavior of the dogs was videotaped and analyzed later. From the video
recordings, 11 different behaviors were scored based on a subjective assessment
of the intensity of the behaviors (frequency), with one variable measured as an
absolute latency. The behavioral variables are in Table 1.
The subsequent statistical analysis was based on the behavior scoring
recorded by an experienced person (Zs. H.), who also coded the entire
sample. However 20% of the recordings were also coded by another naïve,
independent observer. For data coded by both observers, Kappa coefficients
were calculated to measure inter-observer agreement and relatively high
values were calculated in all cases. The values of Kappa coefficients are as
follows: ‘motivation’:1.0;‘playfulness’:1.0;‘willingness to retrieve’:1.0;
‘possessivity’:0.91;‘latencyofstartingtoplay’:0.75;‘control commands’:
0.83; ‘sound signals’:0.85;‘enthusiasm of handler’:0.77;‘praising’:1.0;
‘petting head and body’:0.85.
Questionnaire information
After the test, the handlers were asked to fill in a questionnaire which
included questions about the dog and the dog–human relationship. Based on this
questionnaire we could gain information about the handlers (age, experience
with this dog in years), the living arrangements of the dogs, how much time the
dog–handler pair spends patrolling and how often the handlers play with the
dogs. The exact questions regarding this information are presented in Table 2.
Statistical analysis
Shapiro–Wilk's test was used to assess whether or not the sample is
consistent with a specified distribution function.
Cortisol concentrations were log transformed for analyses. The salivary
cortisol data were analyzed using ANOVAs for repeated measures followed by
Bonferroni's post-hoc test. Three-way ANOVAs with repeated measures were
carried out with ‘occupation’(two levels: police dogs, border guard dogs) and
‘age’(two levels: adult, old) as between-subjects factors, and ‘sampling’(two
levels: baseline, 20 min post) as within-subjects factor. We used paired t-tests for
comparing the ‘baseline’and ‘20 min post episode’cortisol concentrations
within work groups, and independent sample t-test for comparing the ‘baseline’
between working groups.
The behavior of police and border guard dogs, adult and old dogs, policemen
and border guards, was calculated using Mann–Whitney U-tests. Spearman's
Rank correlation was used to search for relationships between non-parametric
behavior variables measured in the case of policemen and border guards. We
used also Spearman's Rank correlation to search for relationships between
humans' behavioral variables and dogs' cortisol concentrations. We used the
Chi-Square test for analyzing the questionnaire data provided by the policemen
and border guards.
Statistical analyses were performed using SPSS (Version 13.0). A significance
level of 0.05 was adopted throughout.
Results
The analysis of salivary cortisol concentrations
The three-way ANOVA revealed that in general, the ‘occu-
pation’did not have a significant effect on the cortisol concen-
trations (F(1,80)= 0.18; p=0.67), nor was there an effect of
‘sampling’(F(1,80)= 0.03; p=0.8), or ‘age’of the dog (F(1,80)=
0.15; p=0.7). Conversely, we found significant interaction be-
tween ‘sampling’and the ‘occupation’(F(1,80) =9.0; p=
0.004). The cortisol concentrations after the test were signi-
ficantly higher than before the test in police dogs (t=−2.93;
p=0.005), while cortisol concentrations after the test were
significantly lower than before the test in border guard dogs
(t=2.04; p= 0.05). When comparing the ‘baseline’cortisol
levels of police and border guard dogs, we found no significant
differences between the two groups (t= 1.94; df= 82; p=0.055).
We also found no significant interaction between ‘sampling’
Table 2
Summary table for the questionnaires filled in by the handlers
Answers to the questionnaire Police Border guard
Where is the dog kept (dog's accommodation (%))?
Block of flats 0 0
Flat with common garden 5.66 3.22
Family house with garden 62.26 96.77
Farm 3.77 0
Kennel at the police station 28.3 0
How many times per week does the handler play with the dog? (%)
Never 3.77 3.22
Once/week 11.32 12.9
Three times/week 16.98 38.7
Five times/week 41.5 35.48
Every day 26.41 9.67
How much time is spent with the dog in
patrol duties? (mean ±SD hours/day)
5.87± 2.87 9.26 ± 3.83⁎
Accommodation and playing with the dog are exclusive categories (tick boxes
on the questionnaire) and duration of the duty had to be judged by the handler.
The bold numbers signify that more border guards kept their dogs at the family
house with a garden than policemen, while more policemen kept their dogs in
the kennel at the police station than border guards. Also border guards played
three times per week with their dogs which was more than policemen, while
more policemen played with their dogs daily which was more than border
guards. The star (⁎) shows that border guards spend significantly more time on
duty with their dogs than policemen do.
Fig. 1. The cortisol concentrations after the test of adult border guard dogs was
significantly lower than before, while the cortisol concentration increased
significantly in old police dogs (the connecting lines indicate the interactions).
Post hoc analysis showed that the baseline cortisol concentrations were
significantly higher in adult border guard dogs than in police dogs. Non-
transformed data are presented as median±quartiles in μg/dl. Sample size of the
subgroups: APD: N=25 (mean age± SD: 5.52± 1.44 years); OPD: N=28 (mean
age± SD: 8.82 ± 0.94 years); ABGD: N=18 (mean age ± SD: 4.77± 1.47 years);
OBGD: N=13 (mean age ± SD: 8.61 ±0.96 years). BGD —border guard dogs;
PD —police dogs; A —adult dogs; O —old dogs.
110 Z. Horváth et al. / Hormones and Behavior 54 (2008) 107–114
and ‘age’groups (F(1,80)= 0.14; p=0.7). The interaction
was significant between work (‘occupation’)and‘age’groups
(F(1,80)= 4.47; p=0.038), indicating that adult border guard
dogs had higher basal cortisol concentrations than the adult
police dogs, while the old police dogs had higher post-playing
cortisol concentrations than the old border guard dogs
(Fig. 1). The three-way interaction between ‘sampling’,the
work groups (‘occupation’)andthe‘age’groups, was not
significant (F(1,80)= 0.28; p=0.6).
The analysis of behavior
Behavior of dogs
When comparing the behavior of all police and border guard
dogs, we found no significant differences in any of the
behavioral variables (Table 3).
Comparisons of the behavior of adult police and border
guard dogs did not reveal any significant differences in any of
the behavioral variables. However, old border guard dogs
generally needed more time to begin playing, than old police
dogs (‘latency of play’:Z=−2.2; p= 0.038).
In the case of police dogs, we found no significant
differences in behavior between adult and old dogs. However,
old border guard dogs generally needed more time to begin
playing than younger border guard dogs (Z=−2.46; p=0.022)
(Fig. 2).
Handlers' behavior
The police officers used more ‘control commands’in the
play situation (Z=−2.8; p=0.005), while border guards ‘pet’
the dog more often (Z=−3.7; pb0.001), and generally showed
more ‘enthusiasm’in the play situation (Z=−2.35; p= 0.018)
(Fig. 3).
There was no difference between adult and old dogs in either
of the behavioral variables measured, and both policemen (P)
and border guards (BG) showed similar behavior toward their
dog's, independent from their dog's age (Table 4).
We found positive correlation between ‘control commands’
and ‘praising’(r
S
=0.349; p= 0.009) in policemen. The police-
men gave a command and if the dogs obeyed it, they praised
Table 3
Statistical results of behavioral variables in the case of police and border guard
dogs
Behavioral
variables
Statistical results
PD vs.
BGD
Adult PD
vs. adult
BGD
Old PD
vs. old
BGD
Adult PD
vs. old PD
Adult BGD
vs. old BGD
Motivation Z=−1.05,
p=0.293
Z=−0.49,
p=0.62
Z=−1.49,
p=0.13
Z=−0.72,
p=0.47
Z=−1.51,
p=0.13
Playfulness Z=−1.94;
p=0.052
Z=−1.21;
p=0.22
Z=−1.78;
p=0.07
Z=−0.86;
p= 0.39
Z=−1.32;
p=0.18
Willingness
to retrieve
Z=−0.65,
p=0.511
Z=−0.58,
p=0.56
Z=−1. 7,
p=0.08
Z=−0.22,
p=0.83
Z=−1.90,
p=0.06
Possessivity Z=−0.30,
p=0.75
Z=−0.28,
p=0.77
Z=−0.53,
p=0.59
Z=−0.73,
p=0.47
Z=−1.09,
p=0.27
Latency of
starting to
play
Z=−1.9;
p=0.057
Z=−0.88;
p=0.38
Z=−2.19;
p=0.03⁎
Z=−1.76;
p=0.08
Z=−2.46;
p=0.02⁎⁎
Aggression –––––
Fear –––––
The stars for the significance show that old border guard dogs needed
significantly (⁎) more time to begin playing, than old police dogs; and old border
guard dogs needed significantly (⁎⁎) more time to begin playing than younger
border guard dogs. We have not mentioned the results concerning fear and
aggression, as fear was not coded in any of the subjects (–), while aggressive
behavior was shown by only one subject (–). BGD —border guard dogs; PD —
police dogs.
Fig. 2. We found significant (⁎) differences in the latency of starting to play
between adult and old dogs within the groups of border guard dogs (□white
boxes), and between old border guard dogs (□white box) and old police dogs
(grey filled box), as well. The old border guard dogs needed more time to begin
playing. Sample size of the subgroups: APD: N= 25 (mean age ± SD: 5.52±
1.44 years); OPD: N=28 (mean age ± SD: 8.82 ±0.94 years); ABGD: N=18
(mean age± SD: 4.77 ± 1.47 years); OBGD: N= 13 (mean age±SD: 8.61 ±
0.96 years). Data are presented as median± quartiles and differences are
considered statistically significant if pb0.05.
Fig. 3. The policemen ( grey filled boxes) use significantly (⁎) more ‘control
commands’with their dogs (□white boxes), while the border guards pet their
dogs significantly (⁎) more. Data are presented as median ± quartiles and
differences are considered statistically significant if pb0.05.
111Z. Horváth et al. / Hormones and Behavior 54 (2008) 107–114
their dogs similarly as they did during training. In the police
dyads we found a weak negative correlation between ‘enthu-
siasm of handler’and ‘latency of starting to play’of the dog
(r
S
=−0.302; p= 0.025), that is, if the handler was more
cheerful, the playing of the dog began sooner. The border
guards that were more enthusiastic (‘enthusiasm of handler’),
tended to pet (‘petting head and body’) their dogs more often
than the others (r
S
=0.413; p= 0.019).
Relationship between cortisol concentrations and behavior
In case of the police dogs, no significant correlations were
found between ‘baseline' cortisol concentrations and the dogs'
behavioral variables. However, we found a slight positive
correlation between dogs' ‘post playing' cortisol concentrations
and the amount of ‘control commands' used by policemen
(r
S
=0.272; p= 0.049).
In border guard dogs we found a positive correlation
between ‘baseline’cortisol concentrations and ‘motivation’
(r
S
=0.515; p= 0.003), and negative correlation between ‘base-
line’levels and ‘latency of starting to play’(r
S
=−0.428;
p=0.014). Border guard dogs with high ‘baseline’cortisol
concentrations were more motivated to play than others and
they started to play immediately. In addition, we found
negative correlation between dogs' ‘post playing’cortisol
concentrations and the ‘frequency of physical praising’(r
S
=
−0.37; p=0.041).
Questionnaire information
There were no significant differences between the policemen
and border guards in their age (Z=−0.022; p=0.982), or in
their experience with their dog (Z=−0.252; p= 0.801). We
found significant differences in living arrangements, time spent
in service and time spent weekly with play. The border guards
only kept their dogs at home (‘dog's accommodation’:
Chi
2
=13.29; df =3; p= 0.004), while policemen often kept
their dogs in a kennel at the police station. Border guards spent
more time with their dogs in service (‘patrol duties’:Z=
−3.619; pb0.001), in contrast to policemen. However, police-
men reported that they play more with their dogs (Z=−2.305;
p=0.021) than border guards, and we found a negative
correlation between ‘time in patrol’and ‘time spent with play’
(r
S
=−0.365; p=0.001).
Discussion
The aim of this study was to examine the effect of human
communicative and social behavior on the inner state (measured
in terms of saliva cortisol concentrations) of dogs in a play
situation. To this end, dogs and their handlers were observed in
a playful interaction, before and after which we measured
changes of the dogs' cortisol concentrations.
We found that on the whole, short term play interaction with
the handler did not significantly change the cortisol concentra-
tions of the dogs at a population level. This was in contrast to
our previous study in which we observed an overall increase in
cortisol concentration after being exposed to an approach by a
threatening human (Horváth et al., 2007). However, important
insights emerged when we compared the results of our two
groups, police and border guard dogs, in which the direction of
change in the cortisol concentration was the opposite. More-
over, the age of the dogs seemed to also have an influence, as in
adult border guard dogs, cortisol concentration decreased after
play, whereas in the old police dogs, the cortisol concentration
increased by the end of the play session.
We also found a relationship between border guard dogs'
baseline cortisol concentrations and motivation and latency of
starting to play, namely border guard dogs with higher baseline
levels are more motivated to play with the handler than those
with a lower levels, as they generally start to play immediately.
Notably, in contrast to the changes in cortisol concentrations,
the play behavior of police dogs was similar to play of the
border guard dogs (i.e. they appeared just as playful and
motivated as border guard dogs). Although the behavioral
analysis of the play behavior may have not revealed initial
differences in the motivation, we believe there were different
causal factors behind the play behavior of the dogs. That is,
police dogs may have executed playful behavior as part of a
training exercise (‘they were commanded to play’), while border
guard dogs may have played spontaneously with their handler.
As a result, the social play may have mediated an inner state
which was associated with lower concentrations of cortisol.
The analysis revealed that the policemen continually
disciplined (‘controlled’) their dogs and used sound signals to
gain their dogs' attention during play sessions, while border
guards were more empathetic and enthusiastic, and also pet and
praised their dogs more often. The positive correlation between
enthusiasm of handler and dogs' latency of starting to play in
the case of policemen, also suggested that the more enthusiasm
shown by handlers, the sooner their dogs started to play. Thus
we suppose that the differences in behavior, mood and
motivation of the handlers not only influenced their dogs'
motivation and behavior during the play session (O'Farrell,
1997), but also had an effect on the post-play cortisol
concentrations, which is in accordance with the findings of
previous studies (e.g. Coppola et al., 2006; Jones and Josephs,
2006; Tuber et al., 1996).
Accordingly, the disciplinary behavior of policemen resulted
in higher cortisol concentrations in the case of old police dogs,
while the friendlier attitude and petting by the border guards,
reduced cortisol concentrations in the case of adult border guard
Table 4
Statistical results of behavioral variables in the case of policemen (P) and border
guards (BG)
Behavioral variables Statistical results
PBG
Control commands Z=−0.02; p=0.18 Z=−1.66; p= 0.09
Sound signals Z=−1.44; p=0.15 Z=−1.13; p= 0.26
Enthusiasm of handler Z=−0.32; p=0.75 Z=−1.2; p= 0.23
Praising Z=−0.13; p=0.89 Z=−0.89; p= 0.37
Petting head and body Z=−0.99; p=0.32 Z=−1.0; p=0.31
112 Z. Horváth et al. / Hormones and Behavior 54 (2008) 107–114
dogs. Police dogs might have assessed this situation as a
training session due to the behavior of the handlers, rather than
play time. Our results are similar to conclusions of Jones and
Josephs (2006), who have revealed a relationship between
larger increased cortisol concentrations and punitive behavior
and lesser increased cortisol concentrations and affiliative
behavior, respectively after an agility competition.
It is important to stress that other factors might have also
contributed to this effect, such as the age of the dogs, previous
keeping conditions and the established relationship between
dogs and their handler.
In the case of age, we found that adult border guard dogs
with higher baseline cortisol concentrations are more motivated
to play than old ones, which in general is in agreement with age-
related decrease in playing activity in dogs (e.g. Rooney et al.,
2000). In our case, the more affiliative behavior of the handler
contributed to the reduction of cortisol concentrations in the
adult dogs, suggesting a calming effect of social interaction. In
contrast, the more disciplinary attitude of the policemen affected
the cortisol concentrations in old dogs in the opposite direction.
A recent meta-analysis found greater cortisol response to a
challenge in older people (Otte et al., 2005) and we have
reported similar observation in old police dogs as a response to
the approach of a threatening human (Horváth et al., 2007).
Notably, the behavior of the handlers was similar toward adult
and old dogs, therefore this cannot explain the present findings.
According to the handlers (questionnaire), many of the
police dogs live in kennels, while all the border guards take their
dogs home after work. In the course of the present training
course, all of the dogs spent their day in kennels, which might
have been a more stressful experience for the border guard dogs,
as police dogs are used to living in such quarters. This
difference could have contributed to the different reaction of the
dogs to the play situation.
Furthermore, both previously mentioned factors could have
also influenced the nature of the actual social relationship
between dogs and handlers, which might in some form, mirror
intraspecies relationships in dogs or wolves. In this case, the
larger social distance in terms of dominance could also
influence the calming effect of the interaction, partly because
dominants may use more forceful and physically challenging
behavioral actions during the interactions, whereas partners
having a similar social status may induce interactions (including
play) by using more communicative signals including signals
for play invitation (e.g. play bow).
In summary, police dogs playing with their handlers showed
some ‘similarity to’experience obtained during general
training. This could be partly attributed to the behavior of
policemen, but other factors cannot be excluded. It follows that
despite the session's playfulness, the interaction with police
handlers increased cortisol concentrations in the dogs, suggest-
ing increased level of stress. Contrasting processes have been
revealed in border guard dogs where we found support for the
stress reducing effect of play.
Further studies may reveal whether dogs living with families
might be also affected by the play behavior of their owner. It
seems that affiliative communicative signals during play might
indeed contribute to stress reduction (at least in terms of
decreased cortisol concentrations) supporting earlier findings
(Arelis, 2006; Palagi et al., 2004). Thus regular play with dogs
could contribute to their well being in general and also reduction
of stress in certain situations.
Acknowledgments
This study was supported by the Hungarian Academy of
Sciences (F 226/98) EU FP-6 NEST012787 and an OTKA grant
(T049615). The authors are grateful to the Hungarian National
Police Training School for Police Dog Handlers (Dunakeszi,
Hungary) for their cooperation and to the all participating
policemen; special thanks to Pál Marsi police colonel and
Ferenc Suszter police major. We would like to express our
gratitude to Professor Imre Oláh, head of the Developmental
Biology and Immunology Lab (Semmelweis University) for his
cooperation. We would like to thank Professor Sergio M. Pellis,
Canadian Centre for Behavioural Neuroscience for his valuable
help concerning the literature. We thank Dorottya Ujfalussy for
her valuable comments on previous versions of the manuscript.
We would like to express our appreciation to Celeste R. West-
Pongrácz for being kind enough to correct the language of our
manuscript as a native English speaker. We are most grateful for
the supportive comments of two anonymous referees.
References
Arelis, C.L., 2006. Stress and the power of play. A Thesis Submitted to the
School of Graduate Studies of the University of Lethbridge in Partial
Fulfillment of the Requirements for the Degree. Department of Neu-
roscience, University of Lethbridge, Lethbridge, Alberta, Canada.
Beerda, B., Schilder, M.B.H., van Hooff, J.A.R.A.M., de Vries, H.W., 1997.
Manifestations of chronic and acute stress in dogs. Appl. Anim. Behav. Sci.
52, 307–319.
Beerda, B., Schilder, M.B.H., van Hooff, J.A.R.A.M., de Vries, H.W., Mol, J.A.,
1998. Chronic stress in dogs subjected to social and spatial restriction. II.
Hormonal and immunological responses. Physiol. Behav. 66, 243–254.
Bekoff, M., 1972. The development of social interaction, play, and meta-
communication in mammals: an ethological perspective. Q. Rev. Biol. 47,
412–434.
Bekoff, M., 2001. Social play behaviour. J. Conscious. Stud. 8, 81–90.
Boissy, A., Manteuffel, G., Jensen, M.B., Moe, R.O., Spruijt, B., Keeling, L.J.,
Winckler, C., Forkman, B., Dimitrov, I., Langbein, J., Bakken, M., Veissier,
I., Aubert, A., 2007. Assessment of positive emotions in animals to improve
their welfare. Physiol. Behav. 92 (3), 375–397.
Byers, J.A., Walker, C., 1995. Refining the motor training hypothesis for the
evolution of play. Am. Nat. 146, 25–40.
Coppola, C.L., Grandin, T., Enns, R.M., 2006. Human interaction and cortisol:
can human contact reduce stress for shelter dog? Physiol. Behav. 87,
537–541.
Darwish, M., Korányi, L., Nyakas, C., Almeida, O.F.X., 2001. Exposure to a
novel stimulus reduces anxiety level in adult and aging rats. Physiol. Behav.
72, 403–407.
Drea, C.M., Hawk, J.E., Glickman, S.E., 1996. Aggression decreases as play
emerges in infant spotted hyenas: preparation for joining the clan. Anim.
Behav. 51, 1323–1336.
Dreschel, N.A., Granger, D.A., 2005. Physiological and behavioural reactivity
to stress in thunderstorm-phobic dogs and their caregivers. Appl. Anim.
Behav. Sci. 95, 153–168.
Dugatkin, L.A., Bekoff, M., 2003. Play and the evolution of fairness: a game
theory model. Behav. Processes 60, 209–214.
Fagen, R., 1981. Animal Play Behavior. Oxford University Press, New York.
113Z. Horváth et al. / Hormones and Behavior 54 (2008) 107–11 4
Hart, L.A., 1995. Dogs as companions: a review of the relationship. In: Serpell,
J. (Ed.), The Domestic Dog: Its Evolution, Behaviour and Interactions with
People. Cambridge University Press, Cambridge, pp. 161–179.
Hennessy, M.B., Williams, M.T., Miller, D.D., Douglas, C.W., Voith, V.L., 1998.
Influence of male and female petters on plasma cortisol and behaviour: can
human interaction reduce the stress of dogs in a public animal shelter? Appl.
Anim. Behav. Sci. 61, 63–77.
Hol, T., Koolhaas, J.M., Spriujt, B.M., 1994. Consequences of short-term
isolation after weaning on later adult behavioural and neuroendocrine
reaction to social stress. Behav. Pharmacol. 5, 88–89.
Holmes, W.G., 1995. The ontogeny of littermate preferences in juvenile golden-
mantled ground squirrels: effects of rearing and relatedness. Anim. Behav.
50, 309–322.
Horváth, Z., Igyártó, B.-Z., Magyar, A., Miklósi, Á., 2007. Three different
coping styles in police dogs exposed to a short-term challenge. Horm. Behav.
52, 621–630.
Jones, A.C., Josephs, R.A., 2006. Hormonal interactions between man and the
domestic dog. Horm. Behav. 50 (3), 393–400.
Kroll, T.L., Houpt, K.A., Erb, H.N., 2004. The use of novel stimuli as indicators
of aggressive behavior in dogs. J. Am. Anim. Hosp. Assoc. 40 (13), 19.
Loizos, C., 1966. Play in mammals. Symp. Zool. Soc. Lond, 18, 1–9.
Lorenz, K., 1950. Man meets dog. Routledge Classics, Great Britain.
Mitchell, R.W., Thompson, N.S., 1993. Familiarity and the rarity of deception:
Two theories and their relevance to play between dogs (Canis familiaris) and
humans (Homo sapiens). J. Comp. Psychol. 107 (3), 291–300.
O'Farrell, V., 1997. Owner attitudes and dog behaviour problems. Appl. Anim.
Behav. Sci. 52, 205–213.
Otte, C., Hart, S., Neylan, T.C., Marmar, C.R., Yaffe, K., Mohr, D.C., 2005. A
meta-analysis of cortisol response to challenge in human aging: importance
of gender. Psychoneuroendocrinology 30, 80–91.
Pal, S.K., Ghosh, B., Roy, S., 1998. Agonistic behaviour of free-ranging dogs
(Canis familiaris) in relation to season, sex and age. Appl. Anim. Behav. Sci.
59, 331–348.
Palagi, E., Cordoni, G., Borgognini Tarli, S.M., 2004. Immediate and delayed
benefits of play behaviour: New evidence from chimpanzees (Pan
troglodytes). Ethology 110, 949–962.
Rooney, N.J., Bradshaw, J.W.S.,2002. An experimental study of the effects of play
upon the dog-human relationship. Appl. Anim. Behav. Sci. 75, 161–176.
Rooney, N.J., Bradshaw, J.W.S., Robinson, I.H., 2000. A comparison of dog–
dog and dog–human play behaviour. Appl. Anim. Behav. Sci. 66, 235–248.
Rooney, N.J., Bradshaw, J.W.S., Robinson, I.H., 2001. Do dogs respond to play
signals given by humans? Anim. Behav. 61, 715–722.
Russell, E.S., 1936. Playing with a dog. Q. Rev. Biol. 11 (1), 1–15.
Salimetrics, L.L.C., 2005. Expanded Range High Sensitivity Salivary Cortisol
Enzyme Immunoassay Kit [Brochure]. State College, PA.
Sapolsky, R.M., Romero, L.M., Munck, A.U., 2000. How do glucocorticoids
influence stress responses? Integrating permissive, suppressive, stimulatory,
and preparative actions. Endocr. Rev. 21 (1), 55–89.
Selye, H., 1950. Stress. Acta, Inc. Medical Publishers, Montreal, Canada.
Soderquist, T.R., Serena, M., 2000. Juvenile behaviour and dispersal of chuditch
(Dasyurus geoffroii) (Marsupialia: Dasyuridae). Aust. J. Zool. 48, 551–560.
Studzinski, C.M., Christie, L.-A., Araujo, J.A., Burnham, W.M., Head, E.,
Cotman, C.W., Milgram, N.W., 2006. Visuospatial function in the beagle
dog: an early marker of cognitive decline in a model of human aging and
dementia. Neurobiol. Learn. Mem. 86 (2), 197–204.
Tuber, D.S., Hennessy, M.B., Sanders, S., Miller, J.A., 1996. Behavioral and
glucocorticoid responses of adult domestic dogs (Canis familiaris)to
companionship and social separation. J. Comp. Psychol. 110 (1), 103–108.
Vincent, I.C., Michell, A.R., 1992. Comparison of cortisol concentrations in
saliva and plasma of dogs. Res. Vet. Sci. 53 (3), 342–345.
114 Z. Horváth et al. / Hormones and Behavior 54 (2008) 107–114
