ArticlePDF Available

Behavioural and hormonal indicators of enduring environmental stress in dogs


Abstract and Figures

Four groups of dogs, which had been subjected to housing conditions of varying quality for years, were assumed to experience different levels of stress. The groups were compared for behavioural and hormonal parameters in order to identify measures that indicate chronic stress in the dog and which may help to identify poor welfare in this species. As a standard for comparison, one of the four groups was composed of privately owned dogs; we assumed that chronic stress levels were relatively low in this group (GI). The three remaining groups of dogs (GII, GIII and GIV) were kept under conditions of low to relatively high austerity, and had basal urinary ratios of cortisol to creatinine, adrenaline to creatinine and, to a lesser extent, noradrenaline to creatinine, that varied from low to high, respectively. Significant differences (P < 0.05) were found in cortisol to creatinine ratios when comparing GI to GII, GIII and GIV and when GII was compared to GIV. Statistical analyses indicated that the mean adrenaline to creatinine ratio in GI differed from that in the remaining groups and that the ratio in GII differed from that in GIII. Noradrenaline to creatinine ratios differed significantly only between GI and GIII. Dopamine to creatinine ratios and noradrenaline to adrenaline ratios did not differ significantly between groups. When dogs were not disturbed, those that were kept under the most austere conditions typically had high levels of locomotor activity, nosing, urinating and paw lifting. After mild disturbance by a slamming door or in the presence of a researcher these animals reacted actively, with increased locomotor activity, circling and nosing, and they showed high levels of behaviours that have previously been associated with acute stress: body shaking, yawning, ambivalent postures and displacement behaviours. Chronic stress in dogs may be identified by increased paw lifting when animals are not disturbed and by ample behavioural expressions of arousal when they are mildly stimulated. Since some behaviours may occur in contexts not related to stress, behavioural data are easily misinterpreted with regard to chronic stress. Interpretation will only be meaningful when physiological measures such as urinary adrenaline to creatinine ratios and, especially, urinary cortisol to creatinine ratios are also determined.
No caption available
Content may be subject to copyright.
B Beerda
M B H Schilder
JAR A M van Hooff
H W de Vries
Department of Clinical Sciences of Companion Animals, Utrecht University, The Netherlands
2Department of Ethology and Socio-Ecology, Utrecht University, Padualaan 14, PO Box 14,3508
TB, Utrecht, The Netherlands
Contact for correspondence and requests for reprints
Final Acceptance: 7 April 1999
Abstract Animal Welfare 2000,
Four groups of dogs, which had been subjected to housing conditions of varying quality for
years, were assumed to experience different levels of stress. The groups were compared for
behavioural and hormonal parameters in order to identifYmeasures that indicate chronic stress
in the dog and which may help to identifY poor welfare in this species. As a standard for
comparison, one of thefour groups was composed of privately owned dogs; we assumed that
chronic stress levels were relatively low in this group (GI). The three remaining groups of dogs
(GIl, GIIl and GIV) were kept under conditions of low to relatively high austerity, and had basal
urinary ratios of cortisol to creatinine, adrenaline to creatinine and, to a lesser extent,
noradrenaline to creatinine, that varied from low to high, respectively. Significant differences
0.05) werefound in cortisol to creatinine ratios when comparing GI to GIl, GIll and GIV
and when GIl was compared to GIV. Statistical analyses indicated that the mean adrenaline to
creatinine ratio in GI differed from that in the remaining groups and that the ratio in GIl
differedfrom that in GIll Noradrenaline to creatinine ratios differed significantly only between
GI and GIll Dopamine to creatinine ratios and noradrenaline to adrenaline ratios did not differ
significantly between groups. When dogs were not disturbed, those that were kept under the
most austere conditions typically had high levels of locomotor activity, nosing, urinating and
paw lifting. After mild disturbance by a slamming door or in the presence of a researcher these
animals reacted actively, with increased locomotor activity, circling and nosing, and they
showed high levels of behaviours that have previously been associated with acute stress: body
shaking, yawning, ambivalent postures and displacement behaviours. Chronic stress in dogs
may be identified by increased paw lifting when animals are not disturbed and by ample
behavioural expressions of arousal when they are mildly stimulated. Since some behaviours may
occur in contexts not related to stress, behavioural data are easily misinterpreted with regard
to chronic stress. Interpretation will only be meaningful when physiological measures such as
urinary adrenaline to creatinine ratios and, especially, urinary cortisol to creatinine ratios are
also determined.
Keywords: animal welfare, behaviour, chronic stress, dog, urinary catecholamines, urinary
© 2000, UFAW, The Old School, Brewhouse Hill, Wheathampstead, Herts AL4 8AN, UK
Animal Welfare2000,9: 49-62
Beerda et al
Animal shelters, breeding stations, animal boarding houses and laboratories are examples of
situations where living conditions may be austere and stressful to dogs. The assessmel1t of
welfare problems in dogs requires tools that measure poor welfare in this species. Humans who
experience reduced welfare show signs of stress (Silver & Wortman 1980; Hilton 1989; Kathol
et a11989; J0rgensen et a11990; Herbert &Cohen 1993). Following similar reasoning (for a
discussion see Stafleu et al [1992]), it can be argued that dogs also show expressions of stress
when they experience poor welfare. Thus, the measurement of stress parameters constitutes a
way of identifying welfare problems in dogs. Stressful situations that endure, say for several
days, may induce different stress responses, and will be more detrimental to a dog's state of
welfare, than when such situations are relatively short-lasting. Since welfare is most seriously
compromised in situations of enduring stress, parameters that indicate chronic stress are more
useful for detecting welfare problems than parameters that change only temporarily at the onset
of the stressor (ie acute stress parameters).
Previously, we have studied behavioural and physiological indicators of chronic stress in
socially and spatially restricted Beagles (Beerda et a11999a, b). It is unclear whether or not our
findings also apply to other, non-experimental, settings. The experimental animals that we used
in that study were socially and spatially restricted for only 6 weeks and they were homogenous
with regards to breed, age and life history. In any non-experimental setting, the dogs under study
are likely to differ from the animals that we studied. Moreover, the stressor to which they are
exposed may not resemble the treatment that we applied and it may last considerably longer than
6 weeks. As a result, dogs that are exposed to stressful situations may show responses that
deviate from the ones that we have reported. The present study aims to establish the practical
use of behavioural and physiological measurements for the assessment of chronic stress under
minimally controlled conditions.
Four groups of dogs that had been living for years under relatively enriched conditions or in
conditions of three degrees of austerity were assumed to have experienced levels of stress that
varied accordingly. The groups were compared with regards to behavioural parameters, urinary
cortisol and urinary catecho1amines. We aimed to identify behavioural and physiological
characteristics that typically manifest themselves under rather austere living conditions and that
indicate chronic stress in dogs. Behavioural parameters, urinary cortisol and urinary
catecholamines were measured because they indicate the activity of biological systems that are
known to respond in acutely stressed dogs, namely behaviour (Solomon
Wynne 1953; Corson
1971; Schwizgebel1982; Engeland et aI1990), the hypothalamic-pituitary-adrenal (HPA) axis
(Clower et a11979; Dess et a11983; Palazzolo & Quadri 1987; Assia et a11989; Bueno et al
1989; Gue et a11989; Engeland et a11990) and the sympathetic-adrenal-medullary (SAM)
system (Anderson & Brady 1972; Lown et a11973; Gaebelein et aI1977; Galosy et a11979;
Pagani et aI1991). Moreover, we have previously found that behaviour and cortisol levels may
also respond in chronically stressed dogs (Beerda et a11999a, b).
The groups of dogs that we studied were not balanced for factors like breed, gender and age.
However, intra-group comparisons were made to fmd out how variation in breed, gender and
age may have biased the measurement of chronic stress.
In summary, we wanted to know which behavioural and hormonal parameters distinguish
dogs which are kept poorly from those which are kept well, and, as such, have practical use as
indicators of poor welfare. For the same reason, we investigated correlations between cortisol
secretion, as an established indicator of stress, and the remainder of the parameters.
Animal Welfare 2000, 9: 49-62
Enduring stress in dogs
Data were obtained from 72 dogs that were categorized in accordance with the conditions under
which they were maintained. Preceding the collection of data, the dogs' housing conditions had
been unchanged for 1year or longer. Group sizes and statistics with regards to gender, breed and
age are summarized in Table I. Dogs in group GI were privately owned. On working days, from
0800h to 1700h, these dogs were maintained in outdoor kennels of 4.3m2. Between 1200h and
1400h the dogs were taken outside for a walk. GIl individuals were housed individually in
kennels that included an indoor section (2. 1m2) and an outdoor section (5.6m2). GIl dogs were
walked outside the premises on a regular basis between 1200h and 1330h. The dogs in GIll (all
female) were maintained in pairs in kennels measuring 2.4m2. From 0800h to 1400h these dogs
resided pairwise in outdoor kennels (3.6m2). GIV individuals were maintained individually in
kennels that measured 1.7m2. From 0800h to 1400h these dogs stayed individually in outdoor
kennels that were sized 3.6m2. GIV individuals were unique in that, over the years, they had
been subjected to a number of different, sometimes stressful, experiments. The animals in Gil,
GIll and GIV were fed once a day (brand DOKO, Trouw, Purten) between 0800h and llOOh.
The privately owned GI dogs were fed at around 0800h and around 1800h. Different owners
supplied different brands of food. All animals had free access to water.
Table 1
SEM age
Composition of the groups and the number of animals measured.
12 <Jand 12 ~ 9 <Jand6 ~ 20 ~ 7 <Jand 6 ~
Various breeds Various breeds Beagles Beagles
0.8 years 7.2
0.6 years 3.3
0.5 6.2
years years
Behavioural recordings (no. of
animals) when dogs were:
undisturbed (6 observation
sessions of lOmin) 8 12 18 11
- startled by a slamming door (6
observation sessions of 5min) 8 12 18 11
- approached (6min) 8 12 18 11
Hormonal measurements:
Urinary cortisol 21 15 17 11
Urinary catecholamines 13 15 16 11
We assumed that the levels of chronic stress increased progressively from GI to GIV in
accordance with increasingly austere conditions of maintenance.
Data collection
The dogs were alternating and/or limitedly available for experimental use. As a result,
measurements were performed with variable sets of animals. The numbers of animals used for
the different types of measurements are given in Table 1. The nature and diversity of the factors
that caused animals to be unavailable for testing makes it unlikely that any sort of unintended
selection of the experimental animals biased assessment of stress.
The undisturbed behaviour which the dogs displayed inside their home kennel was recorded
on video, using a Panasonic M7 video recorder (van Hulskamp, Nieuwegein) for IOmin on 6
different days between 1430h and 1700h. After a dog had been recorded for lOmin, it was
startled by slamming the entrance door of the housing facility. The dog's response to this
Animal Welfare 2000, 9: 49-62 51
Beerda et al
stimulus was recorded for 5min. On one of the 6 days the dogs were observed in the presence
of a female researcher. Having been in sight for 1min (t
1), the researcher moved towards the
door of a home kennel (t
2min), from where she tried to establish physical contact with the
animal (t
3min). Next, the researcher went inside the kennel (t
4min), tried to attract the
subject (t
5min) and, [mally, retreated to the starting position (t
6min). The female researcher
was familiar to the dogs in that she had frequently visited their housing facilities. On such
occasions she was in sight of the animals, sometimes handled them but did not purposely pet
Behaviour was analysed in terms of the frequency or duration of occurrence using the
Observer software package version 3 (Noldus Information Technology, 6702 EA Wageningen,
The Netherlands). Behavioural observations were conducted by one person according to the
following protocol:
Behaviours scored in terms of the frequency of occurrence
Autogrooming, body shaking, changes from one posture to another, changes from one state of
locomotion to another, circling, crouching, defecating, digging, drinking, eating faeces, floor
licking, intentions to change from one state of locomotion to another (scored when dogs
performed only the first motions of the normal repertoire that they performed when changing
from, for example, a standing position to a sitting position, thus maintaining the starting, in this
example standing, position), manipulating the environment, open mouth, oral behaviours, paw
lifting, sighing, stretching, urinating, vocalizing and yawning (see Beerda
et al
[1998] for
Behaviours scored in terms of the duration of occurrence
Nosing, panting, tail wagging, trembling, states of locomotion and postures were scored as
described previously (Beerda et aI1998). Postures were only recorded when the positioning of
a dog's ears, tail and legs could be readily assessed, ie when a dog stood or walked. Other
behaviours that we recorded in terms of their duration of occurrence were:
Ambivalent postures: a crouched body posture accompanied by a position of the tail that is
higher than the breed-specific position; or a high body posture accompanied by a position of the
tail that is below normal.
Contact: dog within a radius of approximately 30cm of the researcher.
Gnawing at the rest basket.
Latency to first contact.
Raising of the hairs on the withers.
Repetitive behaviour: motions that were repeated with minimal interruptions and that were
stereotyped in character.
Naturally voided urine samples were collected in the morning and stored, within 3h of being
voided, at -20°C until their creatinine and cortisol concentrations could be analysed. Half of each
6ml urine sample was adjusted to pH 3-4 using methanoic acid, stored at -20°C and analysed
for catecholamines.
Urinary cortisol and creatinine concentrations were determined following the procedures
described by Rijnberk et al (1988). Inter-assay variation was 9.5 per cent. Urinary catecholamine
concentrations were measured after solid-phase extraction by HPLC followed by radiochemical
detection, as validated for the dog (Beerda et aI1996). A commercially available kit was used
for urine extraction (Pharmacia LKB, Woerden). The HPLC system consisted of a model 112
52 Animal Welfare 2000, 9: 49-62
Enduring stress in dogs
solvent delivery system (Beckmann, Mijdrecht) and a model 460 electrochemical detector
(Waters, Milford, MA). The electrochemical detector data were analysed using a model 3396
integrator (Hewlett Packard, Amsterdam).
Data processing and statistical analysis
The data were normalized by means of logarithmic transformation and analysed by an analysis
of variance (one-way ANOV A). Pairwise comparisons between groups were conducted for
those parameters that showed a significant group effect as indicated by the F-statistic.
Differences between groups were tested for significance following the Tukey method when
group sizes were equal, and by the Tukey/Kramer approach when group sizes differed (for
details on the Tukey method and Tukey/Kramer approach see Stoline [1981]).
Statistical tests on locomotor activity and postures were performed with ratios in order to
achieve mutual independence between the different states of locomotion or postures. First, a
reference state of locomotion (or posture) was selected on the basis of it being performed by all
animals, under a given test condition, with relatively small differences between groups. Next,
the remaining states of locomotion (or postures) were expressed as a ratio to the selected
reference state oflocomotion (or posture). Lying with head rested and half-low were used as the
reference state of locomotion and posture, respectively, for the conditions when dogs were
startled by a slamming door or not disturbed. For the condition involving approach by a female
researcher, standing and neutral posture were used as references.
Per test condition, only the scores on the behaviours that significantly discriminated between
any two groups are presented. However, two exceptions are made. Firstly, in the case of
significant group effects in any of the locomotor states (or postures), the scores for the reference
locomotor state (or reference posture) are also shown. Secondly, scores on the behaviours that
significantly discriminated groups during the time that the dogs were startled by a slamming
door are also presented for the undisturbed condition. Thus, the extent to which the dogs
responded to the slamming door could be assessed.
Urinary levels of cortisol and catecholamines were expressed as ratios to creatinine.
Pooled Pearson's product-moment correlations (Morrison 1976) were calculated between
urinary cortisol to creatinine ratios and behaviours that significantly discriminated between
groups. By calculating pooled correlations, group effects, which are likely to inflate the
correlations, were taken into consideration.
The effects of gender were investigated by means of setting sex as an independent variable
when running ANOVAs with data from GI, GIl and GIV (GIll included only females). The
behaviour of 12 GIl individuals (six Beagles and six dogs from different breeds) was analysed
in an ANOV A for behaviours specific to Beagles. Age effects were investigated by means of
calculating Pearson's product-moment correlations (not pooled) between age and behavioural
or hormonal parameters. Gender, breed and age effects were only tested for the parameters that
differed significantly between the groups. Results are presented as mean values
standard errors
of the means (SEM). The level of significance was set at P
0.05. Our analytical techniques
corrected error rates for multiple between-group comparisons, but not for the number of
parameters studied.
Urinary cortisol and catecholamines
The urinary cortisol and catecholamine levels for each group are presented in Table 2. Urinary
cortisol to creatinine (CC) ratios were higher in GIV than in GIl or GI. CC ratios in GIll were
Animal Welfare 2000, 9: 49-62 53
Beerda et al
elevated in comparison to those in GI. Similarly, mean adrenaline to creatinine (AC) ratios
tended to decline from GIV to GI (see Table 2). Urinary noradrenaline to creatinine (NC) ratios
were significantly higher in GIII than in GI. Ratios of noradrenaline to adrenaline (NA) and
dopamine to creatinine (DC) were similar for all groups.
Table 2
SEM cortisol and catecholamine levels
in urine voided by dogs
that were maintained under different housin2 conditions.
4.8 ± OS 7.4 ± 0.6b8.7 ± 0.6ab 14.4 ± 3.4a
9.4 ± 3.7 2.3 ± 0.6 8.3 ±4.0 10.4 ± 3.5
6.5 ± LIb 8.7 ± 1.4ab 20.5 ± 4.4a16.6 ± 4.8ab
1.4 ± 0.6c2.2 ± 0.6b8.5 ± 2.3a11.5 ± 4.8ab
15.8 ± 5.7 7.9 ± 2.0 3.6 ± 0.8 6.9 ± 3.7
Urinary levels of cortisol (CC), dopamine (DC), noradrenaline (NC) and adrenaline (AC) are expressed as
ratios to creatinine concentrations x 10-6.Noradrenaline to adrenaline ratios are indicated by NA. Differences
between lettered superscripts within a row indicate significant (P
0.05) differences between groups.
instances h·1
0.6 ± 0.4
50.9:... 12.3bc 55.2 ± 12.3c90.0 ± 9.6a92.7 ± 16.6ab
71.3 ± 5.1 59.9 ± 7.2 50.5 ± 4.5 49.2±5.1 % obs. time
1.6 ± 1.2b13.2 ±4.4a17.3 ± 3.3' 20.1±5.7"
4.1 ± 2.0c5.2 ± 1.7bc 19.6 ± 3.4a12.6 ± 2.6ab
0.9±0.6b2.2 ± 0.8a5.3 ± 0.9" 5.0 ± 1.8a
21.9 ± 6.0 26.3 ± 8.2 28.5 ± 5.3 48.1 ± 9.0 instances h·1
1.0 ± O.4b2.4 ± 0.6b6.8 ± 1.0" 6.3 ± LIa % obs. time
0.3 ± 0.3b0.6 ± 0.2ab 2.0 ± 0.8ab 3.5 ± 1.3ainstances h·1
64.8 ± 13.2 64.7 ± 11.3 69.9 ± 6.7 70.6 ± 8.5 % obs. time
1.5 ± 1.0" 1.4 ± 0.7ab 2.5 ± 0.9ab 0.3 ± 0.3b
0.9 ± 0.5b2.9 ± 0.6" 0.2±0.lb0.1 ± O.lbinstances h·
1.6 ± 1.5ab 0.1 ± 0.8b1.3 ± O.4a1.5 ± 0.7ab % obs. time
Ob Ob O.I±O.lb0.8 ± O.3ainstances h-
1.3 ± 0.7 1.8 ± 0.4 2.4 ± 0.4 2.6 ± 0.6
Differences in the undisturbed behaviour of dogs that were maintained
under different hoUSin2conditions.
score GI GIl GIll
body shaking 0.4 ± 0.3 0.4 ± 0.2 0.8 ± 0.2
circling 0 1.3 ± 0.8 1.6 ± 0.6
lying head rested
oral behaviours
paw lifting
tail wagging
Table 3
Mean behavioural scores (± SEM) are expressed as frequencies or percentages of the observation time.
Differences between lettered superscripts within a row indicate significant (P
0.05) differences between
changes from one state of locomotion to another.
changes from one posture to another.
Tables 3,4 and 5 sunnnarize the dogs' behaviour as recorded under three different conditions.
Locomotor activities and levels of nosing were higher in GIV and GIII than in GIl and GI. This
turned out to be consistent for the situations in which the dogs were not disturbed, when they
were startled by the slannning of a door and when they were approached by a familiar person.
GIV individuals stood out, especially in comparison to GI individuals, in that they showed high
54 Animal Welfare 2000, 9: 49-62
Enduring stress in dogs
levels of paw lifting and urinating when they were not disturbed (Table 3), and high levels of
circling, body shaking, yawning, urinating, tail wagging and oral behaviours when they were
startled (Table 4). Also, GlV individuals rarely showed a high posture. Only in GIV did the
slamming of a door induce increases in circling, body shaking and oral behaviours. Increased
nosing occurred in GlV and in GIll. When GlV individuals were approached by a familiar
person they typically showed ambivalent postures, changed often from one posture to another
and, together with GIll individuals, were the only dogs that drank: (Table 5). GI individuals that
were approached by a familiar person showed fewer oral behaviours and more growling than
the remainder of animals. At the same time, they appeared reluctant to contact the researcher and
instances 30min·1
4.9± 3.0'
0.8 ± 0.2ab
0.5 ± 0.2b
27.3 ± 5.5b29.2 ± 6.0b51.5 ± 6.]' 75.2 ± 11.0'
55.] ± 9.6 57.7 ± 7.5 46.0± 4.9 29.9 ± 3.8 % obs. time
6.] ± 4.5b5.6 ± 1.7b20.4±3.1' 24.2± 4.1"
1.4 ± O.4bc 2.0
0.6c5.2 ± 0.9ab 7.3
6.3 ± 2.2b14.5 ± 4.7b12.3
3.5b48.4 ± 10.2' instances 30min·
1.5 ± 0.5b2.1 ± 0.6b10.6 ± 1.2' 14.8 ± 2.4' % obs. time
46.4± 14.8 ]5.2 ± 13.6 75.3 ± 6.] 67.9 ± 9.0
9.4 ± 7.]- 2.6 ± 1.8- 1.9 ± LOb 0.9 ± 0.7b
0.5 ± 0.2b1.7 ± 0.3- 0.2 ± O.1b0.2 ± O.1binstances 30min·
1.7 ± 1.1ab 0.2±0.lb2.0 ± 0.8ab 3.6 ± 1.1- % obs. time
Ob O.]±O.lb0.1 ± O.lb0.9 ± 0.3- instances 3Omin·
0.3 ± 0.2ab 0.3 ± 0.2b1.3 ± 0.3- 2.] ± 0.8-
Differences in the response behaviour of dogs that were maintained under
different housing conditions and disturbed by slamming a door. For notes
and legend see Table 3.
score GI GIl
body shaking 0.] ± 0.] b 0.3 ± O.2b
circling Ob 0.4 ± 0.3b
lying head rested
oral behaviours
tail wagging
Table 4
% obs. time
instances 6min·1
instances 6min·1
% obs. time
instances 6min·
%obs. time
18.3 ± 7.4ab
16.]±3.9b28.6 ± 5.3b55.3 ± 3.9- 30.1 ± 5.4ab
29.5 ± 9.7' 11.2 ± 8.6ab Ob 1.3 ± 0.9b
27.0 ± 12.0- Ob Ob 2.1±I.4b
4.2 ± 2.0ab 26.5 ± 7.3- 6.7 ± 1.6b22.8 ± 7.6ab
32.9 ± 14.1 24.6 ± 5.7 44.9 ± 3.4 45.5 ± 10.6
2.9 ± 2.6ab 26.4 ± 8.3" 29.1 ± 3.7' 13.8±7.1b
5.9 ± 2.3b27.8 ± 5.8- 16.5±3.1- 29.1 ± 7.9'
4.1 ± 1.9b12.7 ± 2.5- 25.4 ± 2.5- 22.4 ±4.3'
26.0 ± 6.1b44.7 ± 3.8- 49.3 ± 2.6- 37.8 ± 4.7ab
23.7 ± 13.7b45.4 ± 7.6- 57.3 ± 6.0- 37.2 ± 6.3-
Differences in the response behaviour of dogs that were maintained under
different housing conditions and approached by a familiar person. For
notes and legend see Table 3.
score GI GIl GIll
ambivalent posture Ob 0.2 ± 0.2bOb
1.9 ± 1.2bc 1.8 ± 0.8c4.7 ± 0.8ab
drinking Ob Ob 0.6 ± 0.1"
growling 3.6 ± 2.7- Ob Ob
latency to contact 51.4 ± 11.9' 16.2 ± 6.4' 1.4 ± 0.4b
lying head rested
standing raised
oral behaviours
tail wagging
Table 5
Animal Welfare 2000, 9: 49-62
they showed relatively little tail wagging. In situations where the dogs were startled by a
slamming door or left alone, GIl individuals sighed relatively often and rarely wagged their tails
(Tables 3 and 4).
Correlations between urinary cortisol and behavioural or urinary catecholamine parameters
We investigated correlations of urinary cortisol levels with the remainder of parameters that are
listed in Tables 2 and 3-5. With regard to correlations between the basal urinary CC ratio and
behavioural parameters (see Table 6), we found positive correlations with changes from one
state of locomotion to another and walking when dogs were not disturbed. After being startled
by a slamming door, the CC ratios correlated with these same behaviours plus circling, tail
wagging and urinating. Basal urinary CC ratios and the behaviour of the dogs in the presence
of a female researcher were not correlated. With regard to catecholamine measures, the urinary
CC ratio was positively correlated with the basal urinary AC ratio (r
0.53, n
Table 6 Correlation coefficients between urinary CC ratios and behavioural
parameters (n
46 in all instances).
Pearson's correlation and significance
0.33 ns
0.53 P< 0.001
0.57 P< 0.001
0.86 P< 0.001
0.34 ns
0.38 P< 0.014
0.67 P< 0.001
(b) after disturbance by a slamming door;
changes from one state of locomotion to another
tail wagging
(a) when dogs were undisturbed;
changes from one state of locomotion to another
Pooled Pearson's correlation coefficients between basal urinary CC ratios and behavioural parameters when
dogs were (a) undisturbed and (b) startled by a slamming door.
Effects of gender, breed and age
Gender effects were studied in GI, GIl and GIV (GIll consisted only of females). Male dogs
showed more ambivalent postures than females when they were approached by a familiar
person. Ambivalent postures were shown during a (mean
SEM) of 1.6
0.8 per cent and 0
per cent of the observation time by males and females, respectively. Undisturbed females
wagged their tail during 1.5
0.8 per cent of the observation time, which was longer than the
0.3 per cent of the observation time that we recorded for male dogs.
With regard to stress hormone excretion, Beagles in GIl did not differ from group members
that were of other breeds, but they did deviate in their behaviour. GIl Beagles typically showed
high levels (mean
SEM) of oral behaviours after they were startled by a slamming door: 25.7
6.6 instances 30min-
vs 3.3
2.4 instances 3Omin-
in GIl individuals that were not Beagles.
When not disturbed, Beagles sighed more than dogs of other breeds: 4.2
0.9 instances h-
0.4 instances h-l.
Effects of age occurred in behavioural parameters only. In the situation that animals were not
disturbed, age correlated with sighing
0.30, n
48, P
0.04) and urinating
0.37, n
48, P
0.0 I). When animals were startled by a slamming door, age correlated with urinating
0.36, n
48, P
0.01), oral behaviours
0.42, n
48, P
0.003) and lying
0.38, n
48, P
0.007). Finally, when dogs were approached by a familiar person, relationships existed
Animal Welfare 2000, 9: 49-62
Enduring stress in dogs
between age and the performance of oral behaviours
0.34, n
48, P
0.02), drinking
-0.32, n
0.03), sitting
0.41, n
0.004) and changes from one state of
locomotion to another
-0.43, n
The present study aims to identify behavioural and hormonal parameters that measure chronic
stress in dogs and that have practical use in the assessment of poor welfare in this species. Dogs
maintained under relatively enriched (GI) or increasingly austere conditions (GIl, GIll and GIV)
for several years were assumed to experience higher levels of chronic stress, and were compared
for behavioural and hormonal parameters. The fmdings with regard to urinary cortisol indicated
that our experimental set-up was usable for the investigation of chronic stress parameters in dogs
and that, conforming to our assumptions, chronic stress levels increased progressively from
groups I to IV. There is ample evidence of increased glucocorticoid levels in acutely stressed
dogs (Clower
et a11979;
et a11983;
Palazzolo & Quadri 1987; Assia
et a11989;
et a11989;
et a11989;
et aI1990).
As in a number of other species (Sassenrath
1970; Craig
et a11986;
et a11986;
von Holst 1986; Roger
et aI1989),
it would appear
that increased glucocorticoid secretion also indicates dogs that experience enduring stress
et aI1999a).
This leads us to assume that elevated cortisol levels are a strong indication
of chronic stress, although normal cortisol levels do not exclude chronic stress, since
physiological adaptations and stressor-specific responses cannot be ruled out. Our present data
show that dogs which are housed under rather austere conditions for several years show elevated
CC ratios and that these values become progressively higher as living conditions worsen. The
present investigations also show that adaptation of the HPA axis in chronically stressed dogs
probably does not imply a normalization of cortisol levels. This makes cortisol levels a very
useful measure of chronic stress.
Although the present findings on urinary cortisol indicate otherwise, we may have ranked the
groups incorrectly in terms of the austerity of their housing conditions. Differences in the way
GIl and GIll individuals were housed suggested to us that the former were 'better off than the
latter, but the opposite may have been true. Dogs in GIll, but not GIl, had the permanent
companionship of aconspecific. Dogs in GIl were privileged, in comparison to those in GIll,
in that they were kept in a more spacious and enriched home kennel and they were regularly
walked outside. The latter implies regular exposure to novel stimuli, contact with conspecifics
and interactions with humans. The importance of human companionship for dogs has been
highlighted by the work of Tuber et al (1996). Based on the way GI individuals were kept, we
assumed that these dogs had the lowest chronic stress levels. It carmot be excluded that, for
reasons unknown to us, some of the dogs in GI experienced chronic stress, although these
privately owned animals lived in a social, spacious and enriched environment. In contrast, from
all the dogs that we studied, those in GIV were subjected to the highest degree of social and
spatial restriction.
An important aspect of the present study is that we measured a large number of parameters
simultaneously. This complicated the statistical analysis of the results. A correction for the
number of parameters that were measured would have practically ruled out the possibility of
significant group effects. In our opinion, this would not have reflected the actual situation.
Instead, we chose to evaluate the results less conservatively and we adopted a comparison-wise
error rate of
0.05 as the level of significance. This necessitates a critical discussion of the
results since there is a fair chance that some of the significant results were accidental. Only with
great caution may the behavioural and hormonal differences between GIII dogs and the
remainder of the animals be interpreted in terms of signs of different levels of chronic stress.
Animal Welfare 2000, 9: 49-62 57
Beerda etal
Dogs in GIII were the only ones that were kept in pairs and direct influences of a cage
companion on a dog's behaviour may have obscured the effects of chronic stress. Also, these
dogs were relatively young and were all female. We investigated, and discuss hereafter, if
differences in the dogs' age and gender biased our assessment of stress: we cannot rule out that,
at least in part, the behavioural and hormonal differences between GIll and the other groups
were caused by differences in age and gender.
Hormonal measures
In accordance with our assumption that chronic stress levels increased from GI to GIV, we
found that basal urinary CC ratios were highest in GIV and declined progressively to the lowest
values in GI. Differences between groups were significant except when GIV was compared to
GIll, or GIII to GIl. Mean basal urinary CC ratios were comparable to those reported by others.
Van Vonderen et al
measured a mean ratio of
in morning urine samples of
pet dogs, which is somewhat lower than the 4.8 x
that we found for the privately owned GI
dogs. A mean basal urinary CC ratio of
has been reported for dogs that, like our GIV
dogs, were housed individually but unlike our GIV dogs were sometimes taken outside for a
walk (Jones et aI1990). This ratio suggests that at the time that Jones et ai's dogs were tested
these animals had stress levels that ranged between those experienced by our GIll dogs (a mean
ratio of
and our GIV dogs (a mean ratio of
Increased catecholamine
levels have been found in the plasma of acutely stressed dogs (Mekhedova & Ghadirian
Engeland et a11990; Parrilla et aI1990), but there is little evidence that increased catecholamine
levels occur in the urine of chronically stressed dogs. We found that basal urinary AC ratios
were highest in GIV and declined progressively to lowest values in GI. Differences were not
significant between GIl and GIV, or between GIII and GIV. NC ratios only discriminated GIll
individuals from GI individuals, the latter having the lowest ratios.
Our fmdings indicate that chronically stressed dogs have increased catecholamine secretion
and that this is reflected in urinary adrenaline levels and, to a lesser extent, urinary noradrenaline
levels, but not in basal urinary DC or NA ratios. Hormonal measures were not significantly
affected by gender, breed or age. The impression that arises when hormonal data from GIII
(pair-wise housing) and GIV (solitary housing) are compared is that housing dogs pair-wise in
a small kennel only moderately improves the situation for the dogs.
Behavioural measures in undisturbed dogs
The behavioural scores on undisturbed dogs confirm the findings on urinary cortisol and
catecholamines, in that behavioural differences were most pronounced between GI and GIV,
with GIl and GIII resembling GI and GIV, respectively. The dogs that we assumed were
experiencing the highest level of chronic stress, namely the GIV individuals, typically showed
increased locomotor activity, nosing, paw lifting and urinating, and they rarely exhibited high
postures. Past and present data indicate that increased paw lifting in dogs may indicate both
chronic (Beerda et a11999a) and acute stress (Beerda et aI1998). Measures of activity are
difficult to interpret with regard to chronic stress. Both inactivity (Hubrecht et a11992) and
increased movement (Hetts
et al
have been connected with chronic stress induced by
social isolation. In the present experiment, the relative inactivity of dogs in GI and GIl may have
been directly related to the fact that these animals were regularly walked out. The latter factor
may also explain why GIV individuals urinated more often than dogs in the other groups.
Contrary to the dogs in GI and GIl, those in GIV were never walked outside and, perforce, they
may have acquired a reduced inhibition to urinate inside their home kennel. Like GIV
individuals, those in GIII were not walked out. However, GIII included only bitches, which
58 Animal Welfare 2000, 9: 49-62
Enduring stress in dogs
urinate less often than male dogs (Beerda et aI1999a). We found positive correlations of age
with urinating and sighing, and a negative correlation with activity. Bitches wagged their tails
more often than male dogs. These effects of age and gender did not obscure the assessment of
chronic stress behaviour.
High incidences of repetitive locomotor behaviour (Hubrecht et aI1992), manipulations of
the environment, grooming, vocalizing (Hetts et al 1992; Beerda et al 1999a), bizarre
movements (Hetts et al 1992), coprophagy and a low posture (Beerda et al 1999a), have
previously been associated with chronic housing stress, but not in the present study. Differences
in the duration of the stress period, which spanned years in the present experiment but only 3
months or 6 weeks in the studies by Hetts et al and Beerda et ai, respectively (Hubrecht et aI's
study included variable periods of restricted housing), may have been a major cause of
discrepancies between the studies. Stereotyped behaviour rarely occurred in any of the groups
that we studied. Mertens and Unshelm (1996) reported stereotyped behaviour in 10 per cent of
109 dogs which were housed individually in an animal shelter. Perhaps the housing conditions
in the present study were not so austere that they induced stereotypies. Or perhaps we did not
observe the dogs at the moments that stereotypies are typically performed. Pigs, for example,
stereotype most abundantly around feeding time (Rushen 1985). It may be that dogs develop
stereotyped behaviour under conditions of social and spatial restriction (Mertens &Unshelm
1996; Beerda et aI1999a), but that such behaviour dissipates over the years. Dynamics in
abnormal behaviour have been reported by Clark
et al
(1997), who reported increasing
incidences of abnormal behaviour over a 3 month period in young, poorly kept Beagles. The
data that are currently available do not portray clearly how abnormal behaviour in dogs develops
over time.
Behavioural measures in dogs that were startled by a slamming door or approached by a
familiar person
Dogs that were mildly stimulated reacted differently according to the austerity of their living
conditions. High levels oflocomotor activity, changes from one posture (or state oflocomotion)
to another, circling, nosing, body shaking, oral behaviours, yawning, ambivalent postures and
displacement behaviour may characterize the responses of chronically stressed dogs. Reactions
to the sound of a slamming door typically involved a high activity (locomotor activity, circling
and nosing), and the expression of behaviours that we previously associated with acute stress
(body shaking, oral behaviours and yawning [Beerda et aI1998]), when dogs were maintained
under poor conditions. It can be argued that differences in response behaviour between GI dogs
and the remainder of animals merely indicate that dogs in the different groups learned different
associations with the sound of a slamming door. If so, one would expect relatively high, and not
low, levels of activity in GI dogs since they may have associated the sound of a slamming door
with being picked up by their owner and going home. Privately owned GI individuals typically
reacted minimally to an approaching female researcher. They sometimes growled, performed
a few oral behaviours or spent a little time nosing and tail wagging, but spent much time lying
and appeared reluctant to contact the researcher. In contrast, GIll and GIV individuals showed
relatively high levels of ambivalent postures, changes from one posture (or state of locomotion)
to another or drinking. The drinking of GIll and GIV individuals during the time that they were
approached by a person may be interpreted as displacement behaviour. The performance of oral
behaviours in stimulated dogs was positively correlated with age and appeared breed (Beagle)
specific. GIV individuals were, on average, 0.9 years older than GI individuals and were all
Beagles, contrary to GI individuals. This implies that in the present experiment, during mild
Animal Welfare 2000, 9: 49-62 59
Beerda et al
stimulation, high levels of oral behaviours may have been falsely interpreted as an indication of
chronic stress.
Correlations of urinary cortisol with behavioural and urinary catecholamine parameters
In the preceding discussion we compared dogs kept in four different housing conditions, in order
to identify parameters that indicate chronic stress. Chronic stress parameters may also be
identified by a positive correlation with an established sign of chronic and acute stress, namely
high cortisol secretion. The present correlations of behavioural and hormonal parameters with
basal urinary CC ratios suggest that chronically stressed dogs have increased basal urinary AC
ratios. Also, they respond actively to mild stimulation, as indicated by positive correlations of
the CC ratio with: i) changes from one state of locomotion to another and walking when dogs
were not disturbed; and ii) changes from one state of locomotion to another, walking, circling,
tail wagging and urinating when dogs were startled by a slamming door. This implies that the
increased activity we associated with chronic stress occurred both when dogs were and were not
disturbed. We do not believe that the increased activity in dogs kept under austere conditions
was the predominant cause of increased cortisol excretion because the groups that were most
intensely exercised (GI and GIl) showed the lowest CC ratios. Also, it has been reported that
exercise programmes in Beagles do not significantly affect the levels of plasma cortisol (Hughes
& Campbell 1990).
In summary, the present results substantiate the measurement of urinary cortisol, adrenaline,
and to a lesser extent noradrenaline, as a valid and non~invasive way to detect chronic stress in
dogs. Behaviourally, chronically stressed dogs may show increased paw lifting, locomotor
activity and nosing when they are not disturbed. During mild stimulation, such dogs may be
identified by relatively high levels of arousal as indicated by increased locomotor activity,
changes from one posture (or state oflocomotion) to another, circling, nosing, body shaking,
yawning, ambivalent postures and displacement behaviour. Because stress behaviour is rather
variable and often nonspecific to stress, it is readily misinterpreted. In many situations, where
pronounced behavioural aberrations are absent, behavioural observations may be of limited use
in the assessment of chronic stress and it may only help to interpret physiological measures like
urinary cortisol and adrenaline levels. The reported indications of chronic stress were associated
with rather austere keeping conditions and the results must be applied with some caution when
a completely different source of stress is involved.
We thank students Sandra Menting and Josefien C von Frijtag Drabbe Kunzel, who helped in
performing the experiments. The technical assistance of Elpetra A P Sprang and Ank van Wees
(Department of Clinical Sciences of Companion Animals, Utrecht University) was highly
appreciated. This work was supported by funds from the Ministry of Agriculture Nature
Management and Fisheries, the Sophia Vereeniging ter Bescherming van Dieren and the Bond
tot Bescherming van Honden.
Anderson D E and Brady
V 1972 Differential preparatory cardiovascular responses to aversive and
appetitive behavioral conditioning. Conditional Reflex 7: 82-96
Assia E, Epstein Y, Magazanik A, Shapiro Y and Sohar E 1989 Plasma-cortisol levels in experimental
heatstroke in dogs. International Journal of Biometeorology 33: 85-88
Animal Welfare 2000, 9: 49-62
Enduring stress in dogs
Beerda B, Schilder M B H, Janssen N S C R M and Mol J A 1996 The use of saliva cortisol, urinary cortisol
and catecholamine measurements for a noninvasive assessment of stress responses in dogs. Hormones and
Behavior 30: 272-279
Beerda B, Schilder M B H, van Hooff JAR A M, de Vries H Wand Mol J A 1998 Behavioural, saliva
cortisol and heart rate responses to different types of stimuli in dogs. Applied Animal Behaviour Science 58:
Beerda B, Schilder M B H, van Hooff JAR A M, de Vries H Wand Mol J A 1999a Chronic stress in dogs
subjected to social and spatial restriction. 1. Behavioral responses. Physiology and Behavior 66: 233-242
Beerda B, Schilder M B H, van Hooff JAR A M, de Vries H Wand Mol J A 1999b Chronic stress in dogs
subjected to social and spatial restriction. II. Hormonal and immunological responses. Physiology and
Behavior 66: 243-254
Bueno L, Gue M, Fargeas M J, Alvinerie M, Junien J Land Fioramonti J 1989 Vagally mediated inhibition
of acoustic stress-induced cortisol release by orally administered K-opioid substances in dogs. Endocrinology
124: 1788-1793
Clark J D, Rager D R, Crowell-Davis S and Evans D L 1997 Housing and exercise of dogs: effects on
behavior, immune function, and cortisol concentration. Laboratory Animal Science 47: 500-510
Clower B R, Andy 0 J, Montalvo M and Peeler D 1979 The hippocampus and stress induced 17-0HCS
elevations. Pavlovian Journal of Biological Science 14: 86-91
Corson S A 1971 Pavlovian and operant conditioning techniques in the study of psychosocial and biological
relationships. Society, Stress and Disease 1: 7-21
Craig J V, Craig J A and Vargas J 1986 Corticosteroids and other indicators of hens' well-being in four
laying-house environments. Poultry Science 65: 856-863
Dess N K, Linwick D, Patterson J and Overmier J B 1983 Immediate and proactive effects of controllability
and predictability on plasma cortisol responses to shocks in dogs. Behavioral Neuroscience 97: 1005-1016
Engeland W C, Miller P and Gann D S 1990 Pituitary-adrenal adrenomedullary responses to noise in awake
dogs. American Journal of Physiology 258: R672-R677
Gaebelein C J, Galosy R A, BotticeIli L, Howard J Land Obrist P A 1977 Blood pressure and cardiac
changes during signalled and unsignalled avoidance in dogs. Physiology
Behavior 19: 69-74
Galosy R A, Clarke L K and Mitchell J H 1979 Cardiac changes during behavioral stress in dogs. American
Journal of Physiology 236: H750-H758
GamaUo A, Villanua A, Trancho G and Fraile A 1986 Stress adaptation and adrenal activity in isolated and
crowded rats. Physiology
Behavior 36: 217-221
Gue M, Alvinerie M, Junien J L and Bueno L 1989 Stimulation of K opiate receptors in intestinal wall affects
stress-induced increase of plasma cortisol in dogs. Brain Research 502: 143-148
Herbert T B and Cohen S 1993 Stress and immunity in humans: a meta-analytic review. Psychosomatic
Medicine 55: 364-379
Hetts S, Clark J D, Calpin J P, Arnold C E and Mateo J M 1992 Influence of housing conditions on beagle
behaviour. Applied Animal Behaviour Science 34: 137-155
Hilton B A 1989 The relationship of uncertainty, control, commitment, and threat of recurrence to coping
strategies used by women diagnosed with breast cancer. Journal of Behavioral Medicine 12: 39-54
Hubrecht R C, Serpell J A and Poole T B 1992 Correlates of pen size and housing conditions on the
behaviour of kennelled dogs. Applied Animal Behaviour Science 34: 365-383
Hughes H C and Campbell S A 1990 Effect of primary enclosure size and human contact. In: Mench J A and
Krulisch E (eds) Canine Research Environment, Proceedings of a Conference, 22 June 1989, Bethesda pp
66-73. Scientists Center for Animal Welfare: Bethesda, MD, USA
Jones C A, Refsal K R, Lippert A C, Nachreiner R F and Schwacha M M 1990 Changes in adrenal cortisol
secretion as reflected in the urinary cortisol/creatinine ratio in dogs. Domestic Animal Endocrinology 7: 559-
2000,9: 49-62 61
Beerda etal
Jorgensen L S, Christiansen P, Raundahl U, 0stgaard S, Christensen N J, Fenger M and Flachs H 1990
Autonomic response to an experimental psychological stressor in healthy subjects: measurement of
sympathetic, parasympathetic, and pituitary-adrenal parameters: test-retest reliability. Scandinavian Journal
of Clinical and Laboratory Investigation 50: 823-829
Kathol R G, Jaeckle R S, Lopez J F and Meller W H 1989 Pathophysiology of HPA axis abnormalities in
patients with major depression: an update. American Journal of Psychiatry 146: 311-317
Lown B, Verrier Rand Corbalan R 1973 psychologic stress and threshold for repetitive ventricular response.
Science 182: 834-836
Mekhedova A Va and Ghadirian A M 1979 The influence of experimental neurosis on the conditional
reflexes and the content ofhlood catecholamines and acetylcholine in dogs. Pavlovian Journal of Biological
Mertens P A and Unshelm J 1996 Effects of group and individual housing on the behavior of kennelled dogs
in animal shelters. Anthrozoos IX: 40-51
Morrison D F 1976 Multivariate Statistical Methods 2nd edition. McGraw-Hill-Kogakusha Ltd: Tokyo, Japan
Pagani M, Rimoldi 0, PizzinelIi P, Furlan R, Crivellaro W, Liberati D, Cerutti Sand MalIiani A 1991
Assessment of the neural control of the circulation during psychological stress. Journal of the Autonomic
Nervous System 35: 33-42
Palazzolo D Land Quadri S K 1987 Plasma thyroxine and cortisol under basal conditions and during cold
stress in the aging dog. Proceedings of the Society for Experimental Biology and Medicine /85: 305-311
Parrilla P, Ramirez P, Muelas M S, Perez J M, Fuente T, Ruiz J M and Ponce J L 1990 Changes in small
intestinal motility in acute physical stress - an experimental study. Hepato-gastroenterology 37: 140-146
Rijnberk A, van Wees A and Mol J A 1988 Assessment of two tests for the diagnosis of canine
hyperadrenocorticism. Veterinary Record 122: 178-180
Roger G K, Anton R, Noyes R and Gehris T 1989 Direct comparison of urinary free cortisol excretion in
patients with depression and panic disorder. Biological Psychiatry 25: 873-878
Rushen J 1985 Stereotypies, aggression and feeding schedules of tethered sows. Applied Animal Behaviour
Science /4: 137-147
Sassenrath E N 1970 Increased adrenal responsiveness related to social stress in rhesus monkeys. Hormones
and Behavior I: 283-298
Schwizgebel D 1982 Zusammenhange zwischen dem Verhalten des Tierlehrers und dem Verhalten des
deutschen Schaferhundes im Hinblick auf tiergerechte Ausbildung. Aktuelle Arbeiten zur ArtgemiifJen
Tierhaltung 138-148
Silver R L and Wortman C B 1980 Coping with undesirable life events. In: Garber J and Seligman M E P
(eds) Human Helplessness pp 279-375. Academic Press: New York, USA
Solomon R L and Wynne L C 1953 Traumatic avoidance learning: acquisition in normal dogs. Psychological
Monographs: General and Applied 67(4): 1-19
Staneu F R, Rivas E, Rivas T, Vorstenbosch J, Heeger F Rand Beynen A C 1992 The use of analogous
reasoning for assessing discomfort in laboratory animals. Animal Welfare 1: 77-84
StoUne M R 1981 The status of multiple comparisons: simultaneous estimation of all pairwise comparisons in
one-way ANOVA designs. The American Statistician 35: 134-141
Tuber D S, Hennessy M B, Sanders S and Miller J A 1996 Behavioral and glucocorticoid responses of adult
domestic dogs (Canis familiaris) to companionship and social separation. Journal of Comparative
Psychology 110: 103-108
van Vonderen I K, Kooistra H Sand Rijnberk A 1997 Intra- and interindividual variation in urine osmolality
and urine specific gravity in healthy pet dogs of various ages. Journal of Veterinary Internal Medicine 11:
von Holst D 1986 Vegetative and somatic components of tree shrews' behaviour. Journal of the Autonomic
Nervous System (suppl.): 657-670
Animal Welfare 2000, 9: 49-62
... The females in our study had larger kennel than the males and had a back room, which provided full isolation of the external area and both characteristics could be contributing to lower FCM levels when compared to the males. Quality of the environment is considered a main criterion for the welfare of dogs housed in kennel environments (Beerda et al., 2000;Hubrecht et al., 1992) and space restriction is a known cause of chronic stress (Beerda et al., 1999). Furthermore, dogs in our study have restricted social contact and physical activity, which also are factors that decrease dogsẃ elfare and increase stress responses (Dalla Villa et al., 2013;Normando et al., 2014). ...
... Past studies of kennelled dogs found that inactivity is connected to increased cortisol levels (Hiby et al., 2006). Conversely, increased cortisol levels do not always change behavioural patterns in dogs (Beerda et al., 2000;Hennessy et al., 2001;Rooney et al., 2007), hence the importance of using different investigative techniques to understand the effects that kennelling has on dogs. Although we have not been able to link the fluctuation of FCM levels with the expressions of abnormal behaviours -a clear indication of failure to adapt to conditions present -there is an agreement among animal welfare researchers that dogs which remain a long time in kennelled environments are more prone to develop chronic stress and reduced welfare (as reviewed in Polgár et al., 2019). ...
Studies with humans and some other animal species have shown that sleep is compromised when the presence of external factors such as light, sound, and temperature surpass normal levels. This study investigated the effects of these environmental conditions on 13 kennelled laboratory dogs, assessing whether each variable interfered with their sleep behaviour and/or increased stress responses, which could further compromise sleep quality. The behaviour of dogs was video recorded for eight months. Diurnal and nocturnal behaviour were recorded, along with naturally occurring levels of temperature, light and sound in the dogs’ kennel environment. Faecal cortisol metabolites (FCM), from samples collected every morning, were used to monitor the dogs’ adrenocortical activity. GLMM models and non-parametric tests were conducted to evaluate the relationship between sleeping patterns, environmental variables, and stress on the studied dogs. Nocturnal sleep decreased in response to increases in temperature and in day light duration. No effects of sound and FCM levels on dogs’ sleep were observed. However, diurnal sleep was affected by sound and FCM levels, decreasing when both factors increased. Additionally, noisier days increased stress responses, especially in male dogs. Increased FCM levels were associated with changes in the diurnal behaviour of dogs; for example, decreased activity. The decrease in daily activities and increased physiological stress responses could be associated with maladaptation to the environment, which could indicate poor welfare. Our study suggests that mitigating the impact of environmental conditions in the kennels could improve sleep quality and the overall quality of life of the dogs.
... In some regions, such as Lazio, animal rights associations can take over the management and responsibility for these dogs, after obtaining the agreement of the LVHU. TNR is considered an alternative to a long-term shelter, which does not necessarily provide an optimum environment for dogs (confinement combined with novel experiences and changes in routine can be stressful for the sheltered animals) [3,4,[6][7][8][9][10][11]. TNR is viewed as a way to lower the costs of maintaining long-term dog shelters that burden administrations. ...
... These behaviours provide information on social interactions and emotions. Agonistic behaviours, such as aggression to conspecifics, have been reported to be associated with stress and poor welfare [6,31,32]. In contrast, affiliative behaviours play an important role in the formation of bonds and alliances among individuals, allowing social interaction, and are essential for maintaining complex social groups [33][34][35]. ...
Full-text available
A descriptive analysis, inter-observer and test–retest reliability of the animal-based measures (ABMs) included in the protocol were performed. This study aimed at the development of a welfare assessment protocol for dogs recruited in the trap-neuter-release (TNR) programmes and the description of the implantation of these programmes in Italy. Nine Italian regions carried out TNR programmes. A varied scenario, along with some critical issues, emerged. Fifty dogs were recruited and assessed simultaneously by two assessors to determine the reliability of ABMs included in the protocol. A subsample of ten dogs were assessed three times to assess test–retest reliability. All females were neutered against 36% of males. Most dogs were adults (58%) and of a large size (68%). Vaccine prophylaxis and parasitic prevention were regular in 13% and 76% of dogs, respectively. Few dogs showed lameness, evidence of pain, other clinical problems, or thermal discomfort. Overall, 82% of dogs did not show fear or aggression to unfamiliar people. The level of agreement between the two assessors was quite high, ranging from substantial (0.61–0.80) to perfect (1) for the majority of measures. This study highlighted some critical issues in TNR implementation and the suitability of the protocol as a tool for animal welfare assessment.
... For example, after 'aversive training' dogs show increased saliva cortisol (a glucocorticoid linked to the stress response, specifically the activation of HPA axis-hypothalamic-pituitary-adrenal axis 37 ) and a higher frequency of lips licking, yawning, and body shaking in comparison to dogs that experienced 'positive training' sessions 38 . However, other studies did not find a direct correlation between cortisol concentrations (measured in blood, urine, or saliva) and the performance of displacement behaviours in putatively stressful situations 39 . Beerda and colleagues 40 found that dogs' saliva cortisol increased just after non-social stressors (e.g. ...
Full-text available
In the present study we investigated the influence of positive and negative arousal situations and the presence of an audience on dogs’ behavioural displays and facial expressions. We exposed dogs to positive anticipation, non-social frustration and social frustration evoking test sessions and measured pre and post-test salivary cortisol concentrations. Cortisol concentration did not increase during the tests and there was no difference in pre or post-test concentrations in the different test conditions, excluding a different level of arousal. Displacement behaviours of “looking away” and “sniffing the environment” occurred more in the frustration-evoking situations compared to the positive anticipation and were correlated with cortisol concentrations. “Ears forward” occurred more in the positive anticipation condition compared to the frustration-evoking conditions, was positively influenced by the presence of an audience, and negatively correlated to the pre-test cortisol concentrations, suggesting it may be a good indicator of dogs’ level of attention. “Ears flattener”, “blink”, “nose lick”, “tail wagging” and “whining” were associated with the presence of an audience but were not correlated to cortisol concentrations, suggesting a communicative component of these visual displays. These findings are a first step to systematically test which subtle cues could be considered communicative signals in domestic dogs.
... Assessment of therapy dog welfare has primarily focused on measurements of stress-related physiology, particularly the peripheral concentration of the hormone cortisol, and/or dog behaviors found to be previously associated with stressful situations (13)(14)(15), such as panting, paw-lifting, repetitive liplicking, and avoidance behaviors (6-8, 10, 16). However, focusing on the identification of poor welfare indicators only partially addresses the animal welfare concept. ...
Full-text available
To date, investigations of the welfare of therapy dogs have focused largely on examining physiological and behavioral measures that could indicate if the animal is experiencing stress or distress. However, this approach does not fully address the definition of welfare which is often described as existing on a continuum from negative (or stressful) to positive . With therapy dogs, it would be worth addressing if they experience positive emotional affect while working since the quality and efficacy of animal-assisted interventions for the human recipient is likely to be influenced by the animal's emotional state during the interaction. The purpose of this review is to articulate how objective measurements of the HPA axis and measurements of behavioral observations and standardized questions can be used to evaluate positive welfare in therapy dogs. A potentially relevant indicator of positive welfare is the peripheral concentration of the neurohormone oxytocin, which has been found to increase in systemic circulation within a variety of species during positive social and affiliative contexts, including during human-dog interaction. Oxytocin is also a negative-feedback regulator of the Hypothalamic-Pituitary-Adrenal (HPA) axis, which culminates with the production of the stress hormone cortisol. Cortisol is widely used as a physiological indicator to assess negative welfare states in animals, including therapy dogs. Observable behavior during interactions with humans that may convey enjoyment could provide indicators of positive welfare in dogs such as engagement in play, or human-directed affiliative behaviors including leaning against, nudging, or licking the patient. However, in assessing positive welfare, it is also critical to consider that all animal behavioral displays and physiological responses are dependent on the dog's individual (and breed) temperament. Temperament directly drives how the animal copes and responds to its current physical and social environment, including during stressful situations such as when therapy dogs interact with unfamiliar humans in novel healthcare settings. Coupled with both positive and negative physiological and behavioral welfare indicators, questionnaire data can provide further context to, and enhance interpretations of, therapy dog welfare assessment results. Overall, to date, no studies have measured all of these factors to assess therapy dog welfare.
... Therefore, measuring behaviours is essential when examining whether environmental interventions can be considered successful [20]. Relaxation, alert and stress behaviours in dogs are widely used indicators of psychological welfare [21][22][23]. Despite the importance of EE, there are few controlled studies assessing the effect of multiple enrichment activities on dog welfare in different environments [24]. ...
Full-text available
Environmental enrichment (EE) can be used to enhance the environment of various animals. The aim of this pilot study was to determine the effects of seven EE activities (Bonding, Bubble machine, Conspecific play, Interactive toy, Playhouse, Stuffed food toy and Tug play) on dog behaviour, pre- and post-EE for dogs housed in an office environment during training as part of an assistance dog training programme. EE activities resulted in a significant increase in the frequency of relaxation behaviours (p < 0.01) and a significant reduction in alert (p < 0.01) and stress behaviours (p = 0.02). Results suggest various benefits of the different activities with Conspecific Play and Playhouse activities having the greatest overall positive behaviour change when compared to the other activities. The food-based EE activities (Interactive toy and Stuffed food toy) had the least behaviour change of all the activities provided. Findings will be of interest to pet owners, animal rescue centres, dog trainers and working dog organisations.
... This has a consequence on the dog's behaviour (lower posture, more oral behaviour). In accordance with previous research by the author (Haverbeke et al., 2008a) but contrary to other studies in primates (Deputte, 1994) and dogs (Beerda et al., 1998(Beerda et al., , 2000Hennessy et al., 1998Hennessy et al., , 2002, this study does not show any significant relationship between stress or insecurity and yawning. ...
... However, during the resting period, the eye temperature related to rectal temperature was less accurate than the ear temperature. It should also be kept in mind that the stress factors may not always be related to behavioral changes due to variability in gender, breed, experience or type of stressful situation [63][64][65][66], and observation of spontaneous behavior is needed to interpret physiological data [67,68]. In that vein, Pasing et al. [68] have reported that stallions increased their heart rate in response to semen collection (p < 0.001); however, it was seen that stallions remained calmer during the collection. ...
Full-text available
This study aimed to describe the thermal variation of external reproductive tracts during ejaculation in relation to sperm quality in dogs. Forty-six adult fertile dogs were monitored using a thermal camera before, during and after the semen collection, taking into account penile and scrotal temperatures as reproductive thermal patterns while eye and perianal temperatures were recorded as complementary thermal patterns of behavioral response. The parameters were classified depending on age (≤4 years and >4 years), body weight (BW) (≤75 kg and >75 kg), sperm concentration (CON) (≤300 million and >300 million), total testicular volume (TTV) (≤600 cm3 and >600 cm3) and total ejaculation time (TET) (≤800 s and >800 s) of the animals from which semen was collected successfully. Heavier males (p < 0.05) that have more consistent testicles (p < 0.01) as well as quicker ejaculate responders (p < 0.001) and lower scrotal temperature had better semen (Δ motility) freezability. The lower eye temperature prior to the ejaculation (p < 0.01), lower scrotal temperature following ejaculation (p < 0.01), and conversely, higher penile temperature during the ejaculation (p < 0.001) had a higher sperm concentration. Furthermore, the sperm freezability was negatively correlated with total ejaculation time (r = −0.39, p < 0.05) and sperm abnormalities were lower in the ejaculate of dogs having a higher temperature of the scrotum, bulbus and penis. In conclusion, infrared monitoring throughout semen collection in dogs can provide information on behavioral reactions during human manipulation, as well as semen quality and testicular functionality.
High-volume breeders often produce purebred or “designer breeds” to supply the pet industry. Some high-volume commercial breeding facilities (CBEs) include extreme conditions such as overcrowding, little human contact, exposure to harsh weather conditions, inadequate food and water, and untreated medical conditions. Previous research with puppies born in these extreme conditions reports that this early life history of deprivation is associated with multiple behavioral and psychological abnormalities once adopted. Still, few studies to date have examined whether impoverished conditions affect the physiological systems of domestic dogs. The following study examined hormonal and behavioral differences in dogs from adverse early-life environments (i.e., puppy mills, unlicensed commercial breeders, hoarding situations) and a control group of dogs during social interactions with an unfamiliar human while in the shelter environment. Dogs from known early adverse backgrounds displayed significantly higher salivary cortisol levels than other dogs found at the shelter, F(1, 23) = 4.51, p = 0.045. Neither time spent in adverse living conditions nor time spent in the shelter correlated with cortisol output. During social interactions with an unfamiliar human, dogs from adverse backgrounds exhibited more fear behavior (p = 0.022) and displayed lower levels of affiliative behavior (p = 0.039). Dogs from adverse backgrounds spent more time in stationary positions (i.e., sitting, standing, freezing) than other shelter dogs (p = 0.043), and we were unable to complete a food-based social-cognitive task because of this. Using a moderated multiple regression model, we found dogs’ background was a significant moderator of the relationship between total cortisol output and panting during interactions with an unfamiliar human, R² = 0.39, F(1, 11) = 6.94, p = 0.023. In contrast, no relationship between these variables was seen in other shelter dogs. The findings of this study demonstrate that even when residing in the same shelter environment, a history of living in extreme conditions may lead to higher cortisol levels and altered behavior in a novel social context. While studies have examined dogs while still in the CBEs or once they have been adopted, this study examines their behavior and cortisol levels once removed and placed in a shelter environment where the public might interact with them and consider them for adoption. Gaining a better understanding of the behavioral and hormonal outcomes of dogs that have experienced extreme early life deficiencies may assist in the development of effective interventions that will facilitate rehabilitation and enhance the welfare of these dogs as pets.
Providing behavioral care to animals in special circumstances, such as following a natural disaster or after removal from a cruelty or neglect situation, presents a variety of unique challenges. Following disasters, animals are often held in rudimentary field shelters until they are reunited with their owners or considered unclaimed. Cruelty cases involve populations of animals, such as dogs from organized dogfighting operations and animals from hoarding situations, that present with behavioral needs for safe and humane sheltering. Long‐term holds, often due to legal cases, compound shelter stress over time, which can lead to behavioral decline. These special circumstances represent substantial challenges to maintaining animal welfare. Even when faced with less‐than‐ideal conditions and other limitations, best efforts should be made to prevent, mitigate, or eliminate negative welfare and to facilitate psychological well‐being.
Stereotypic behaviour is correlated to stress in wild animals in captivity. When wild animals are exposed to stress repeatedly, they may experience a chronic elevation of glucocorticoids (G.C.s), especially for captive sun bears. This study aimed to investigate the stress in captive sun bears by analysing the cortisol level using Faecal Glucocorticoid Metabolite (FGM) analysis before and after introducing environmental enrichments. 57 samples of faecal captive sun bears were collected from four selected study sites around Malaysia. The behavioural data were collected with an instantaneous sampling method. While for FGM analysis, glucocorticoids were analysed using enzyme immunoassays. As predicted, the stereotypic behaviour increased in pre and post enrichment and decreased during the enrichment period. For FGM analysis, the cortisol level was high in pre-enrichment and low in the enrichment period. In conclusion, the outcome of this study indicated that environmental enrichment might help reduce chronically elevated cortisol concentrations that are generally associated with distress in mammals.
Full-text available
The dog (Canis familiaris) has been domesticated for thousands of years but the effects of different housing regimens on canine behaviour are poorly understood. This study presents behavioural data collected from solitary and group-housed dogs housed in animal shelters and laboratories. The dogs differed greatly in their behaviour under the different housing regimens. Solitary dogs were more inactive (72-85% of observed time compared with group-housed dogs 54-62% of observed time), and spent more time in non-social repetitive locomotory behaviour categories (4-5% compared with group-housed 0.9-2% of observed time). Group-housed dogs were not only able to interact socially, but also spent more time investigating the floor of their pens, presumably because of the increased olfactory stimuli. Group-housed laboratory dogs provided with kennels used them for: rest, play and the control of social interactions. Single-housed dogs, which were housed in smaller pens, had low overall activity and tended towards stereotyped circling rather than pacing. At all the sites the opportunities for interactions with humans were limited (0.24-2.52% of the time observed). The results are discussed in terms of cage design and animal husbandry.
Full-text available
Six weeks of social and spatial restriction were used as a model to induce chronic stress in Beagles. Behavioral and physiological measurements were performed during a period of enriched spacious outdoor housing in groups (GH) and during a subsequent period of solitary housing in small indoor kennels (IH). Behavioral parameters that may indicate chronic stress in dogs are reported. During IH, the dogs showed significantly (comparison-wise error rate <0.05) lower postures than during GH. IH induced enduring increments in frequencies of autogrooming, paw lifting, and vocalizing, and was associated with incidents of coprophagy and repetitive behavior. So far, we interpret the behavioral changes as signs of chronic stress. Relatively low levels of walking, digging, intentions to change from one state of locomotion to another, and increments in circling are conceived as obvious adaptations to the specific features of the IH system. By challenging the dogs outside their home kennel we tested whether the dogs’ coping abilities were affected by IH. Dogs that were challenged were introduced into a novel environment, given the opportunity to escape from their home kennel, restrained, walked down an unfamiliar corridor, presented a novel object, exposed to loud noise, given food, or confronted with a conspecific. During IH, challenged dogs exhibited higher postures, showed more tail wagging, nosing, circling, urinating, and defecating, and changed more often from one state of locomotion (or posture) to another than during GH. These behavioral changes were observed across the different types of challenges, with the exception of the noise administration test. In the presence of conspecifics, the socially and spatially restricted male dogs behaved more dominantly and aggressive than during the time that they were kept in groups. Such behavior manifested as increased performances of raised hairs, growling, paw laying, and standing over. Both sexes showed increases in paw lifting, body shaking, ambivalent postures, intentions to change from one state of locomotion to another, and trembling in any of the challenges, excluding the walking down the corridor test. In short, during a variety of challenges, socially and spatially restricted dogs exhibited a heightened state of aggression, excitement, and uncertainty. Behavioral differences between dogs that had experienced pleasant and bad weather conditions during GH, suggested that “pleasant-weather individuals” had experienced early stress during the control period, and, as a result, responded to the subsequent period of IH differently. Regardless of the housing conditions, challenged bitches showed stronger indications of acute stress than male dogs. Gender did not affect the chronic stress responses to social and spatial restriction. A low posture and increased autogrooming, paw lifting, vocalizing, repetitive behavior, and coprophagy may indicate chronic stress in dogs, and as such, can help to identify poor welfare. When challenged, chronically stressed dogs may show increased excitement, aggression, and uncertainty, but the nonspecificity of such emotional behavior will complicate its practical use with regard to the assessment of stress.
Thirty tethered sows were observed for 5 min every half-hour for 9 h spanning the two feeding periods. Activity, consisting largely of food searching behaviour and drinking, was largely restricted to two 2-h periods following each feed. Three categories of stereotyped behaviour were observed and these were closely tied to the feeding periods. Short-duration bouts of rubbing, head-waving and bar-biting occurred during food delivery, while long-duration bouts of highly stereotyped and idiosyncratic sequences of rubbing and drinking were shown by older sows immediately after feeding. Vaccuum chewing tended to occur slightly later. I suggest that frustration of feeding motivation rather than under-stimulation underlies stereotypies in pigs, and that the different forms may represent stereotype of the appetitive and consummatory phases. Aggression was rare and was not closely related to the feeding periods or to stereotypies.
The effects of different spatial areas and different social conditions on behaviours of beagles maintained in a laboratory were evaluated. Eighteen female purpose-bred beagles were divided into six groups of three, and housed individually for 3 months each in six different housing conditions: (A) a 6.1 m × 9.1 m outdoor pen; (B) a 1.8 m × 6.1 m outdoor run; (C) a 1.2 m × 3.66 m indoor run; (D) a 0.9 m × 1.2 m × 0.84 m cage; (E) a 0.9 m × 1.2 m × 0.84 m cage with 30 min of forced treadmill exercise, 5 days week-1; (F) a 0.71 m × 0.86 m × 0.69 m cage. Behaviours of six dogs housed in pairs in Conditions A and C were also compared. Behaviours studied were movement, vocalisation, lying down, sleep, object manipulation, barrier manipulation, barrier jumping, fence running, agonistic and affiliative activities, and proximity. Behavioural effects were compared among housing conditions, order of rotation through each housing condition, and behavioural changes over time during each 3 month rotation. Dogs spent more time moving in pens and runs than in cages. Dogs housed in the greatest degree of social isolation spent the most time moving, exhibited the greatest number of bizarre movements, and vocalised the most. Dogs housed in the smallest cages spent more time grooming and in manipulation of enclosure barriers than those housed in any other conditions. Forced treadmill exercise did not significantly alter behaviours. When housed in pairs, dogs spent more time sleeping and showed a tendency to spend less time vocalising than when housed singly. The results indicate that spatial area and activity are not likely to be the most important factors to be considered when evaluating psychosocial well-being of dogs. In assessing the psychosocial well-being of dogs, social isolation may be as harmful or more harmful than spatial restriction.
To emphasize the effects of group- and single housing of kennelled dogs, the behavior of 211 dogs in two German animal shelters was tested and observed. After being placed, 197 of the dogs' new owners were interviewed. Although 51% of the German animal shelters already keep dogs in groups, there is strong prejudice against group housing because of the fear of fights. This study demonstrates that this apprehension is unfounded. Ninety-one percent of the social confrontations between dogs housed together were settled by the use of behavioral rituals. Keeping dogs in groups, furthermore, leads to a significant reduction in noise emission (p<.001). Group housing fulfills the dog's need for social interaction and the need to move. Dogs that were housed in groups displayed a closer human-animal relationship (80%) than those that had been kept individually (43%). A high percentage of individually housed dogs suffered from behavioral problems (31%) and 10% developed stereotypes. The percentage of behaviorally disturbed dogs observed in group housing was 11%, and stereotyped forms of behavior did not occur. Dogs who had been kept in groups were, on average, placed within 10 days, and were returned to the animal shelter less often (9%) compared to those housed individually (25%). Dogs that were housed separately needed an average of 17 days to be placed. Even after being placed, there is a correlation between the animal shelter's type of housing and the dog's behavior. Within four weeks after picking up their pet, 88% of the owners of dogs that had been housed individually complained of problems compared to the owners of the dogs that had been kept in groups, 53% of whom were completely satisfied with the adoption. Despite the fact that these results might be influenced by the small number of shelters examined, the study leads to the conclusion that keeping dogs in groups is a suitable alternative for dog housing in animal shelters and, for the animals' welfare, is preferable to individual housing.
Two mixed-sex groups of subadult M. mulatta, maintained under chronic crowding stress, were characterized re adrenal responsiveness to ACTH (via urinary 17-hydroxycorticosteroids) and intragroup social behavior over a 2-year period which included a series of changes in social environment.Under conditions of unchanging dominance structure, ACTH-response levels of individual cagemates remained relatively constant, with 3- to 10-fold differences between the lowest (dominant male) and the highest (most subordinate female cagemate). Elevated response levels of subordinate cagemates were lowered by changes which reduced social stress, such as transfer to individual caging, removal of the dominant male, or consort pair formation involving the dominant male. Conversely, removal of the most subordinate cagemate resulted in elevation of ACTH-response levels of some remaining subordinates. Covariance was demonstrated between ACTH-response levels, aggressive stimulation received, and fear-anxiety behavior for two female subordinate cagemates.The data suggest that the prime determinant of ACTH-response levels is the fear- or anxiety-evoking aspects of group interaction. Elevated adrenal responsiveness is assumed to be a function of chronically elevated ACTH output reflecting the characteristic level of fear-evoking or “stressful” stimulation sustained by each member of the social group.
The physiologic contribution of the limbic brain to emotionally induced stress is still poorly understood. The present study is designed to more specifically evaluate the role of the hippocampus in stress induced plasma 17-OHCS elevations. The conditional reflex to a sequential presentation of tone and shock was used as the stress agent in adult mongrel dogs. Plasma 17-OHCS levels were determined by the Porter-Silber method. Control and stress levels of 17-OHCS were determined before and after unilateral (left) hippocampectomy, and subsequent contralateral (right) hippocampectomy. A unilateral posterior hippocampal lesion partially attenuated (20%) the normal 17-OHCS stress response. In contrast to unilateral lesions, equivalent bilateral posterior hippocampal lesions abolished the normal 17-OHCS stress response. These observations support the thesis that the elevated 17-OHCS levels in response to the conditioning paradigm is dependent on the hippocampus. Furthermore, it is dependent upon the continuity of the hippocampal circuit and not upon the volumetric steroid binding capacity of the hippocampus. These studies also suggest that a unilaterally functioning hippocampus may be adequate to meet the physiologic needs of stress, as reflected by the 17-OHCS response. Presented at the 1978 meeting of the Pavlovian Society, St. Petersburg, Florida.
In this study, we used spectral analysis of short-term R-R and systolic arterial pressure (SAP) variabilities to estimate the changes in neural control of the circulation produced by psychological stress. The 0.1 Hz low-frequency (LF) component of R-R and SAP variabilities provided a quantitative index of the sympathetic activity controlling heart rate and vasomotion. Conversely the high-frequency (HF) respiratory component of R-R variability provided an index of vagal tone. In conscious dogs we used the seemingly stressful situation of being accompanied for the first time to the experimental laboratory as a stimulus. In human subjects we used mental arithmetic. In both cases LF of R-R and SAP variabilities increased significantly suggesting enhanced sympathetic activity both to the SA node and the vasculature. In man, the index α, a measure of the overall gain of baroreceptor mechanisms, was found to be reduced during mental arithmetic. Spectral analysis of cardiovascular variabilities thus suggests that in man and in conscious dogs psychological challenges induce a profound re-arrangement of neural control of the circulation, which appears to be characterised by sympathetic predominance and which can be monitored by this technique.
Stress parameters that can be measured noninvasively may help to identify poor welfare in dogs that live in private homes and institutions. Behavioural parameters are potentially useful to identify stress, but require further investigation to establish which behaviours are appropriate. In the present study, behaviours were recorded and analysed for signs of acute stress in dogs. Simultaneously, saliva cortisol and heart rate were measured to support the interpretation of the behavioural data with regard to stress. Ten dogs of either sex, different ages and various breeds were each subjected to six different stimuli: sound blasts, short electric shocks, a falling bag, an opening umbrella and two forms of restraint. Each type of stimulus had been selected for its assumed aversive properties and was administered intermittently for 1 min. The stimuli that could not be anticipated by the dogs, sound blasts, shocks and a falling bag, tended to induce saliva cortisol responses and a very low posture. The remainder of the stimuli, which were administered by the experimenter visibly to the dog, did not change the cortisol levels but did induce restlessness, a moderate lowering of the posture, body shaking, oral behaviours, and to a lesser extent, yawning and open mouth. Pronounced increases in the heart rate were nonspecifically induced by each type of stimulus. Heart rate levels normalized within 8 min after stressor administration had stopped. Saliva cortisol levels decreased to normal within the hour. Correlations between behavioural and physiological stress parameters were not significant. From the present results, we conclude that in dogs a very low posture may indicate intense acute stress since dogs show a very low posture concomitant with saliva cortisol responses. Dogs may typically show increased restlessness, oral behaviours, yawning, open mouth and a moderate lowering of the posture when they experienced moderate stress in a social setting. The nonspecific character of canine heart rate responses complicates its interpretation with regard to acute stress.