Physiology & Behavior, Vol. 60, No.

2, 389-396, 1996
Copyright 1996 Elsevier Science Inc.
Printed in the USA. All rights reserved
0031-9384/96 $15.00 + .00

ELSEVIER

PII S0031-9384(96)00013-9

Rearing Piglets in a Poor Environment:

Developmental Aspects of Social Stress in Pigs
F R A N C I E N H. D E J O N G E , t I E . A . M .

B O K K E R S , * W. G. P. S C H O U T E N * A N D F. A. H E L M O N D t

*Department of Animal Husbandry and ~Department of Human and Animal Physiology, P.O. Box 338, 6700AH,
Agricultural University, Wageningen, The Netherlands
Received 18 September 1995
DE JONGE, F. H., E. A. M. BOKKERS, W. G. P. SCHOUTEN AND F. A. HELMOND. Rearingpiglets in a poor
environment: Developmental aspects of social stress in pigs. PHYSIOL BEHAV 60(2) 389-396, 1996.--Effects of
rearing conditions on the development of social relationships were investigated in female pigs. For that purpose, one
group of piglets was reared in a poor environment (i.e., a commercially used indoor 4.2-m ~ standard farrowing crate)
and the other in an enriched environment (i.e., 0.5-ha outdoor pasture with half-open farrowing crates). After
weaning, all piglets were housed in pairs under standard conditions. Dominant and subordinate individuals were
distinguished within pairs and social relationships between members of a pair were studied in adulthood. The results
show that pairs reared in a poor environment behave more aggressively. The subordinates of these pairs also develop
symptoms indicative of chronic social stress exposure: i.e., they show a delayed onset of puberty, a smaller daily
weight gain, and elevated basal cortisol levels that are also higher 5 h after an acute restraint stressor (tethering stress).
No such deviations were found in subordinates reared in an enriched environment. It is concluded that rearing piglets
in a poor environment may facilitate the development of social stress in adult (subordinate) pigs. Two underlying
mechanisms may be proposed: a) deteriorated social skills lead to increased social stress, or b) a failure to cope with
stressors in general, and the stress of being subordinate in particular, occurs in animals reared under poor conditions.
Swine
Pig
Gilts
Social stress
Cardiovascular system
Puberty
Oestrus

Dominance
Tethering stress
Cortisol
Weight development
Individual differences

FERAL sows are highly social animals that wander around in

family groups with a strong and stable hierarchy and surrounded
by offspring for which they may exhibit communal care (11,12).
Such family groups are also highly intolerant to unfamiliar
conspecifics, and social relationships between members are very
important (4,45). Recent studies in the domestic pig kept in a
seminatural environment show that much of the feral behavioral
repertoire has been retained, and evidence in these pigs indicates
that social relationships are similarly important (22,32,33,42).
Under adverse conditions, social relationships between
(domestic) pigs may cause "social stress," operationally defined
as socially induced behavioral and physiological disturbances that
have been associated to chronic stress exposure (26,30,49). Such
disturbances have been found in group-housed pigs that are kept
for commercial purposes (17,18,44). Social stress may develop
when social interactions are unpredictable or uncontrollable (48),
and its development will therefore be facilitated under conditions
of crowding, scarcety of resources, or barren housing conditions.
In addition, social stress will develop when the species-specific
social skills of penmates are poorly developed (50).

Rearing conditions might be of critical importance for the

development of appropriate social skills (13,14,35). Experience
with play behavior seems essential in this respect (40). This was
further illustrated in recent studies by Hol in rats, showing that a
short period of deprivation from play behavior may result in an
increased cortisol response after exposure to a social stressor in
adulthood (19,20). Pigs are often reared under barren conditions
on commercial farms, and Lammers and Schouten showed that
deviant aggressive play behavior patterns develop in piglets
reared under such conditions (25,38). These authors suggested
that such deviant play behavior patterns might develop into poor
adult social skills. However, it is still unknown whether poor
rearing conditions may indeed contribute to the development of
social stress a n d / o r poor social skills in socially kept pigs in
adulthood.
The present study was designed to investigate whether poor
rearing conditions may cause social stress in adult female pigs,
possibly so by deteriorating effects on social skills. Standard
commercial farrowing crates similar to those used by Lammers
and Schouten (25) were used as " p o o r " rearing conditions. Such

1 To whom requests for reprints should be addressed. E-mail: Francien.deJonge@etho.vh.wau.nl

389

390

DE JONGE ET AL.

farrowing crates prevent contact with piglets of other litters and

offer minimal possibilities for play behavior and exploration.
Piglets of the "enriched" rearing condition were born in half-open
farrowing crates, offering possibilities for contact with piglets of
other litters 2 weeks after birth, and exercise with play behavior
and exploration through the adjacent 0.5-ha pasture [for a discussion on play behavior patterns in enriched as opposed to poor
rearing conditions, see (32,33)]. After weaning, all piglets were
housed in pairs under standard conditions, each pair consisting of
two sisters. Dominant and subordinate individuals were distinguished within pairs and social relationships between members of
a pair were studied in adulthood. Finally, stress-related physiological deviations were registered to investigate whether symptoms of "social stress" developed within pairs.
Social relationships were investigated in dominance tests and
social provocation tests in which aggressive tendencies were
provoked by presentation of a limited food amount. To investigate whether adult social skills were affected, the amount of
aggression and qualitative aggressive behavior patterns were also
registered in aggressive encounters with unfamiliar gilts. Social
stress was defined as socially induced behavioral and physiological disturbances that have previously been associated with chronic
stress exposure (26,30,49). These include altered function of the
hypothalamic-pituitary-adrenal axis (HPA axis), a decreased
activity of the hypothalamic-pituitary-gonadal axis, and reduced
metabolic efficiency (23). To investigate whether poor rearing
conditions contribute to the development of social stress in
dominant and subordinate pair-housed female gilts, weight development, onset of puberty, expression of oestrus behavior, and
basal cortisol levels were therefore investigated. In addition, the
behavioral, heart rate, and cortisol responses to an acute nonsocial restraint stressor (tethering) were studied. This challenge test
was chosen because chronic (social) stress exposure has been
shown to ameliorate the capacity for negative feedback after
activation of the HPA axis by an acute stressor (28,37). In this
test, resistance behavior, heart rate response, and cortisol response were studied as ethological, resp. cardiovascular, and
neuroendocrine parameters of the stress response, and these
parameters were studied before, during, and after tethering stress
in adult dominant and subordinate pigs, reared under poor as
opposed to enriched conditions.
METHOD

Breeding and Housing Conditions Before Weaning

Fourteen 1-year-old primiparous sows of a Great Yorkshire x
British Landrace Breed (GY BL, Pig Improvement Co., UK)
were selected to breed the experimental subjects for this experiment. The sows were randomly assigned to one of two groups
(n = 7 each) and were inseminated within 1 week with commercially supplied sperm of Great Yorkshire boars. The first group
was tethered 30 days after insemination, and farrowed and nursed
their litters in the " p o o r " environment; the other group farrowed
and nursed their litters in the "enriched" environment. The poor
environment consisted of a stable with eight indoor 4.2-m 2
farrowing crates (two-thirds concrete, one-third slats). In each
farrowing crate, the sow was tethered and ringed in by two iron
railings at each side to limit freedom of movement. A heating
lamp was positioned in one of the corners to keep the piglets
warm and to reduce the chance that piglets were crushed by the
sow. No bedding material was present. The air was ventilated but
the stable was otherwise not temperature or humidity controlled.
Piglets from each litter were kept here until weaning, and the
piglets had no contact with piglets from other litters, other than

through odours and vocalisations. The enriched environment, on

the other hand, consisted of a 0.5-ha outdoor enclosure with eight
half-open farrowing crates of 8 m 2. A 30-cm high barrier separated the farrowing crate from the rest of the enclosure. Because
of the barrier, the piglets could leave the farrowing crate after
11-12 days to join and meet piglets from other litters. The sow
could tresspass the barrier whenever she liked. This set-up was
chosen to imitate the possibilities for social contact between
piglets of different litters as they occur in naturally behaving feral
swine (11,12,22). The enclosure contained feeding cages for the
sows, a mud pool, a straw-embedded shelter, and an additional
pasture that was accessible during a limited period of the day to
keep fresh grass available. All sows farrowed in June and within
3 weeks, and litters were weaned 6 weeks after farrowing (5-10
piglets per litter). Two sows of the " p o o r " environment were
discarded from the experiment: one because she was not pregnant, the other because she killed her offspring at birth. For the
remaining sows, no group differences in litter size, birth weights,
or perinatal mortality were observed. The sows were individually
fed twice a day with standard sow diet. Piglets were fed ad lib
with standard piglet diet until the age of 4 months and were fed
twice a day with standard sow diet thereafter.

Experimental Subjects
At weaning all piglets were weighed, and the two heaviest
female piglets from each litter were selected for the experiment
(poor, n = 2 x 5; enriched, n = 2 7). Each pair of sisters was
housed together in one sawdust-embedded 4-m 2 pen with two
feeding troughs separated by a 30 x 30-cm partition. The experimental pigs were housed under these conditions during the rest of
the experiment, except for weeks 6-11, when each pair was also
accompanied by two male littermates. Within pairs, dominant and
subordinate individuals were distinguished (see Behavioral Tests)
and analyses of data were performed on the following experimental groups: poor/dominant (n = 5), poor/subordinate (n = 5),
enriched/dominant (n = 7), and enriched/subordinate (n = 7).

Behavioral and Physiological Registrations

Weight development. The experimental subjects were weighed
at weaning (6 weeks), at 10, 23, 36, and 52 weeks of age, and at
first observed behavioral oestrus.
Onset of puberty. At 20 weeks of age, the gilts were transported to a neighbouring stall containing a mature, vasectomized
boar, which was used to asses the age of puberty (i.e., the age of
first observed behavioral oestrus). For that purpose, the boar was
led through the pens of the experimental gilts daily and behavioral oestrus was registered when the gilts showed a positive
response in the back pressure test in the presence of the boar (5).
Daily oestrus control continued until all gilts had shown behavioral oestrus (44 weeks of age).
Dominance tests. To assess dominant and subordinate individuals within a pair, feeding tests were run on 6 consecutive
nonoestrus days at 30 or 31 weeks of age. During feeding tests,
200 g of food pellets was offered in only one of the two feeding
troughs and at least 4 h after regular feeding time. Eating time of
both individuals of a pair was registered until the food was
finished (4 min on average). It was only possible for one animal
at a time to eat from the feeding trough. An individual was
assigned dominant if she ate during more than 85% of the time
during four consecutive tests. Only two pairs needed more than
four tests (i.e., six tests) to meet the criterium. In most of the
pairs, the entire amount of food delivered during the consecutive
feeding tests was only eaten by the dominant. To assess whether
dominance relationships were stable, the procedure was repeated

DEVELOPMENT OF SOCIAL STRESS IN PIGS

at the end of the experiment (i.e., when the experimental subjects

were 70 weeks old). The outcome of the dominance tests at 70
weeks were exactly the same as that at 30 weeks of age except
for one pair reared in the enriched environment, which had
developed a place preference (i.e., one animal always ate from
the left food trough, the other from the right one). Dominance
tests at 30 weeks of age were used for further analysis for each
pair.
Aggression within pairs. Aggressive tendencies within pairs
were assessed in six consecutive daily tests (social provocation
tests), run on nonoestrus days at 32 or 33 weeks of age. For that
purpose, 200 g of food pellets was put in the middle of the pen,
outside the feeding trough, and at least 4 h after regular feeding
time. During the following 3 min, the number of headknocks and
headknocks with bites was registered for each individual
(11,12,25,32). In addition, the number of deviant aggressive
behaviors [bites and headknocks directed at the hindquarters of
the opponent (25)] was registered separately.
Expression ofoestrus behavior. Starting from 45-47 weeks of
age (i.e., when all gilts had reached puberty), qualitative aspects
of oestrus behavior were registered according to methods previously described in detail by de Jonge et al. (5). In short, after a
few adaptation trials, each gilt was guided to a straw-embedded
runway leading to a goalbox with a nonembedded concrete floor.
This goalbox was separated by a bar fence from a sexually active
stimulus boar. The gilt was daily tested in this runway during 10
min, and the time spent "presenting" in front of the boar was
registered. Presenting was registered when the gilt was standing
completely immobile with or without the ears "kocked". During
presenting behavior, the head was often slightly lowered and
tail-wagging was absent. The typical oestrous cycle of the pig is
around 21 days with a 2-4-day oestrous period. Before our tests,
oestrus synchrony was acquired by oral administration of a
synthetic progestagen (altrenogest, Hoechst Animal Health UK,
20 m g / 5 ml during 18 days).

Response to Acute Restraint Stress

Procedure. Resistance behavior, heart rate, and cortisol levels
in response to tethering stress were measured in week 52. Prior to
tethering stress, one individual of each pair was temporarily
removed (directly after the first basal saliva sample was taken)
and the cagemate only returned after the last saliva samples had
been taken. The other individual was then tethered in her home
pen during 1 h according to procedures described elsewhere (39).
Tethering always took place at 0900 h and four animals were
tested at a time (balanced between groups). Saliva samples for
cortisol analysis were taken 20 min before tethering (at 0840 h),
after 1 h of tethering stress (1000 h), and again 4 h later (1400 h).
Heart rate levels were measured 15 rain before tethering stress
(basal levels), and during tethering stress (1 h). Resistance to
tethering stress was measured during tethering only (1 h). Concentrations of cortisol in saliva were measured using essentially
the same single-antibody radioimmunoassay technique that has
been described previously for unextracted plasma samples (21).
Two modifications have been made in this procedure to measure
cortisol concentrations in saliva. Plasma samples were diluted 10
times before assaying and saliva samples were not diluted. The
standard curve for plasma samples was corrected for plasma
matrix effects by adding 50 ILl of cortisol-free plasma to the
standard concentrations of cortisol and the standard curve for
saliva samples was not corrected for these plasma matrix effects.
Heart rate was measured by means of a "Polar Sport Tester"
(Polar Inc, Finland). Mean heart rate values in beats per minute
were collected over 5-s periods. Behavioral resistance to tethering

391

TABLE 1
SUMMARY OF THE PROCEDURE
Week 1-6
Week
Week
Week
Week

6
20
30-31
32-33

Week 44
Week 45-47
Week 52
Week 60-63
Week 70-71

Rearing in poor
Environment

OR

Rearing in enriched
Environment

Weaning
six dominance tests
six aggression tests
(between pair members)

Start daily puberty tests

|

End puberty tests

six oestrous expression tests
Reaction to acute tethering stress
(resistance, heart rate, cortisol)
six aggression tests
(between unfamiliar pair members)
six dominance tests

stress was scored by scan sampling (one sample every 15 s),

using a PSION event recorder for the following behavioral
parameters: pulling the chain vigorously, dragging the chain,
standing (chain without tension), lying down. "Resistance" was
calculated as the lumped score of these parameters.
Agonistic encounters between unfamiliar pair members. Towards the end of the experimental period (weeks 60-63), the
hypothesis was tested that animals from the poor rearing condition would show more aberrant patterns of aggressive interactions
towards an unfamiliar opponent. Aberrant patterns were defined
as showing more headknocks and bites directed at the hind
quarters of the opponent [see (25,38)]. For that purpose, each pig
from the dominant group was confronted with another pig from
the subordinate group, which was unfamiliar (i.e., not a pair
member, not a littermate member, and not housed adjacent), and
reared in the same environment. The number of bites and headknocks directed at the front, at the middle part, and at the
hindquarters of the opponent was registered during 30-min confrontations.

Statistical Analysis
The data were presented as the mean + SEM and analysed by
analysis of variance (MANOVA, SPSS/PC + ) with the independent factor Group (two levels, Poor vs. Enriched), and the
dependent factor Dominance (two levels, Dominant vs. Subordinate). In some analyses, MANOVA with repeated measurements
included an additional dependent factor Age (for weight development) or Time (response to tethering). Nonparametric statistics
were used when criteria of normal distribution (estimated by
normal probability plots) or homogeneity of variance (Bartlettbox, p > 0.05) were not met. Data analysed by nonparametric
statistics are presented in the figures as median + SE of the
median (27). These data were analysed by Mann-Whitney U-tests
for independent samples, and with Wilcoxon matched pairs signed
ranks test for dependent samples.
RESULTS

Weight Development, Aggression, Puberty, and Oestrus Behavior

Weight development. Weight development of the different
groups is presented in Fig. 1 (upper left panel) and was analysed
by MANOVA for repeated measurements including the factors
Group (two levels), Dominance (two levels), and Age (five
levels). The results revealed a significant three-way interaction
effect of Group by Age by Dominance, F(4, 40) = 3.6, p < 0.05.
Subsequent separate analyses for the different groups involving

392

DE JONGE ET AL.

the factors Dominance and Age revealed in the enriched group a

significant main effect for Age, F(4, 2 4 ) = 607.2, p < 0.0001,
but no significant main or interaction effects involving the factor
Dominance. For the poor group, on the contrary, a significant
interaction effect of Age by Dominance, F(4, 16) = 10.85, p <
0.001, indicated that subordinate pigs reared in a poor environment have a different weight development than dominant ones.
Post hoc contrasts indicated that dominant pigs are heavier than
subordinate ones when reared in a poor environment and at the
age of 23 ( p < 0.05), 36 ( p < 0.01), or 52 ( p < 0.05) weeks.
Aggression within pairs. The frequency of aggressive behaviors of dominant and subordinate individuals was calculated as
the total number of headknocks plus headknocks with bites
registered in the six consecutive aggression tests (see Fig. 1,
upper right panel). Mann-Whitney U-tests showed that pairs of
animals (frequency of dominant and subordinate individuals summated) from the poor environment showed higher levels than
pairs reared in the enriched environment (M-W, z = 2.20, p <

WEIGHT

0.05). Subsequent Wilcoxon tests showed that dominant animals

showed more aggression than subordinate ones when reared in an
enriched environment ( z = 2.11, p < 0.05), but this difference
was not significant for animals reared in the poor environment
(z = 0.67, p = 0.5). Mann-Whitney U-tests indicated that rearing piglets in a poor environment increased aggression frequencies of the subordinate individuals (M-W, z = 22.40, p < 0.05),
but not of the dominant ones. Deviant aggressive behavior patterns (bites and headknocks directed at the hind quarters) occurred very seldom and were therefore not analysed separately.
Puberty. Age of first behavioral oestrus is presented in Fig. 1
(lower left panel). Wilcoxon tests indicated that first behavioral
oestrus was late in subordinates compared to dominant ones,
when reared in a poor environment (z = 2.02, p < 0.05), but not
when reared in an enriched environment (z = 0.37, p = 0.7).
Mann-Whitney U-tests revealed no significant effect of rearing
condition on the dominant and subordinate groups separately. No
group differences were found with respect to the parameter

DEVELOPMENT

AGGRESSION
SO

" F

/r

/t

180

40

'Io
~,~

t2o

I=

.~

8o

o
r-

4o

o"

30
20

"k

Iml

t0

23

36

52

10

23

age

36

52

(weeks)

OESTRUS BEHAVIOR

PUBERTY
350

.0
"0

a)
E

Ae-

"k
I

305

75

(o

o1
c

6o

260

~.,
C

45

215

3o

>,
"0

Q.

170

125

is

O~

enriched

poor
dominant

enriched

poor

subordinate

FIG. 1. Effects of rearing conditions on physiological and behavioral parameters of dominant and subordinate subjects. Enriched rearing conditions: left
side of panels. Poor rearing conditions: right side of panels. Results are expressed as means + SEM for parametric data and as medians -I- SE medians for
nonparametric data. ,k p < 0.05, ~,k p < 0.01. ,k: dominant vs. subordinate, ~: enriched vs. poor rearing condition.

DEVELOPMENT OF SOCIAL STRESS IN PIGS

393

"weight at puberty." However, correlation coefficients calculated for age of puberty and weight at puberty indicated that
animals in general tended to be heavier when first behavioral
oestrus was later (Spearman rank, r = 0.72, p < 0.01, n = 24).
Expression of oestrus. Percent time presenting on consecutive
days was analysed by M A N O V A for repeated measurements
including the factors Group, Dominance, and Repeated. No significant main or interaction effects were found ( p > 0.05). Mean
maximal response levels for the different groups are presented in
Fig. 1 (right lower panel).

heart rate after tethering stress

250
200
"E~ 150
'~ 100

Reactions to Acute Restraint Stress

Cortisol levels. Mean cortisol levels are presented in Fig. 2 for
the different experimental groups. M A N O V A including the factors Group, Dominance, and Time (three levels) revealed a
significant main effect for Time, F(2, 20) = 4.48, p < 0.05, and
a significant interaction effect for Group by Dominance, F(10,
1) =5.55, p < 0.05. With respect to the main effect Time,
subsequent post hoc contrasts indicated that saliva cortisol levels
increase 1 h after tethering stress ( p < 0.05), and again decrease
after 5 h ( p < 0.001). With respect to the significant Group by
Dominance interaction effect, subsequent separate analyses per
group, including the factor Dominance, revealed a significant
main effect for Dominance in the poor group, F(1, 4 ) = 5.75,
p < 0.05, but not in the enriched group. No significant interaction
effects of Time by Dominance were found in the separate groups.
The results therefore indicate that subordinate animals reared in
the poor environment show, in general, higher cortisol levels than
their dominant cagemates, whereas such differences do not exist
in pairs of animals reared in enriched environments. No significant .correlations were found between cortisol levels and age or
weight at puberty, expression of oestrus behavior, frequency of
aggression, or resistence behavior to tethering stress. Correlation
coefficients were calculated on the entire group of experimental
subjects.
Heart rate. For the whole group, mean basal heart rates before
tethering stress were 97.3 + 7.35 bpm. Just after tethering, heart

cortisol levels after tethering stress

.3

enriched

poor

pre I l h r

5hr

I dominant

pre I l h r

5hr

subordinate

FIG. 2. Cortisol and restaint stress. Cortisol levels in saliva of dominant

and subordinate individuals after a 1-h period of tethering stress in
adulthood. Saliva samples were taken before, at the end of 1 h of
tethering stress (arrow), and again 4 h later (5 h). Data are expressed as
means + SEM for pigs reared in enriched (left aspect of panel) or poor
(right aspect of panel) conditions, ~ MANOVA, dominant vs. subordinate, F(1, 4) = 5.75, p < 0.05.

t'~ 50

- -

lt5

310

415

6tO

time(minutes)

FIG. 3. Heart rate during restraint stress. Heart rate before and during a
1-h period of tethering stress (start of tethering indicated by arrow),
presented for a dominant animal reared under enriched conditions. No
group differences were found. Sample frequency: 15 bpm. Data are
expressed as beats per minute (vertical axis) in relation to time in minutes
(horizontal axis).

rate rapidly increased to average maximal peak values of 193.5 +

18.4. A typical example of heart rate registration during tethering
stress is presented in Fig. 3 for a dominant animal from the
enriched group. Some registration failures occurred at the end of
the registration. For that reason, two animals of the enriched
group were discarded from analysis and the total analysis was
limited to 15 min before tethering (averaged to one mean basal
value bl) and 15 rain during tethering (averaged over 3-min
periods to five values tl to t5). MANOVA including the factors
Group, Dominance, and Time (six levels) showed no significant
main or interaction effects, thus indicating that early environment
or dominance did not influence heart rate responding to tethering
stress.
Resistance behavior. Resistance behavior was calculated as
the relative percentage of scans an animal was either " p u l l i n g "
or "dragging" the chain. Resistence behavior was analysed by
M A N O V A including the factors Group and Dominance. No
significant main or interaction effects were found ( p > 0.05).
Agonistic encounters between unfamiliar pair members. Between encounters, there was a considerable variation in latency,
frequency, and duration of aggressive behaviors that was not
associated with rearing condition or weight difference between
opponents. The data were therefore analysed by nonparametric
statistics. The frequency of bites and headknocks directed at the
hindquarters was calculated as a percentage of the total frequency
of aggressive displays per agonistic encounter. Pigs reared in an
enriched environment tended to have more aggressive acts directed at the hindquarters of the opponent (median of 13%, range
37.5%) compared to the pigs reared in the poor environment
(median 6.3%, range 12%). However, differences were not significant (M-W, z = 1.39, p = 0.1645).
DISCUSSION

The present experiment shows that poor rearing conditions

facilitate aggression within familiar pairs of gilts in adulthood. In
addition, it was shown that poor rearing conditions facilitate the
development of social stress in such pairs: that is, the subordi-

394

nates develop physiological deviations associated to chronic stress

exposure such as delayed onset of puberty, elevated saliva cortisol levels, and a smaller daily weight gain. These results stress
the importance of early ontogenetic processes for the development of social stress in pigs in adulthood. However, these results
also have practical implications: the majority of group-housed
pigs kept for experimental or commercial reasons are reared
under conditions similar to our " p o o r " conditions and social
stress may therefore more easily develop in such systems.

Underlying Mechanisms
In the social provocation tests and tests for aggressive encounters between unfamilar pigs, no group differences in deviant
aggressive behavior patterns were found. We expected to find
such differences because Lammers and Schouten (25) previously
described that piglets reared in a poor environment develop
deviant agonistic play behavior patterns with increased aggression directed at the hindquarters of the opponent, less threatening
behavior, and decreased control of the subordinate to terminate
the aggression impinged on her. These authors suggested that
these aberrant play behavior patterns would pursue into adulthood, thereby culminating into social relationships that would be
less predictable and controllable and therefore more stressfull
(38,48). That such deviant behavior patterns were not found in
our subjects argues against the hypothesis that poor social skills
(i.e., showing more deviant behavior) might underly the development of stressful relationships in our animals. On the other hand,
the observed increase in aggressive behavior in pairs reared under
poor conditions corroborates the notion that poor social skills (in
terms of aggressive tendencies and decreased control of the
subordinate to terminate the aggression impinged on her) do play
a role. It is therefore concluded that the present experiment gives
additional, but no conclusive, evidence that poor social skills
contribute to the development of social stress in animals reared
under poor conditions.

Alternative Underlying Mechanisms

Previous results in other species have shown that stressful
early life events (prenatal or postnatal) may increase the individual's susceptibility to stressors in later life. Thus, prenatal stress
and/or the occurrence of stressful early life events have been
shown to delay onset of puberty (16,46) and to facilitate the
susceptibility for stressors later in development in a number of
different species (2,47), including pigs (15). Meaney and coworkers (28), in addition, showed that early handling procedures may
attenuate the susceptibility to adult stressors and, also, that such
handling procedures may act in a protective way to neural
hippocampal damage during repeated stressors in adulthood.
Other studies showed that prenatal stress or maternal deprivation,
peer deprivation, or early weaning may influence stress-induced
behavioral and cortisol reponses or facilitate stress-induced gastric lesions in rats (9,28,43). Finally, Ackerman noted a human
counterpart in this respect, showing that peptic ulcer formation in
human adolescents often appears to be preceded by separation or
loss (1).
In the present experiment, the poor rearing condition offered
the piglets only limted possibilities for locomotary and exploratory activity and limited possibilities for social play and
contact with the sow. However, also the prenatal environment of
the piglets differed: piglets from the poor environment were born
to mothers that were exposed to chronic tethering stress, whereas
piglets from the enriched environment were born to group-housed
sows with free access to a pasture. Tethering stress has been
extensively shown to induce stress symptoms like stereotyped

DE JONGE ET AL.

behavior, elevated cortisol levels, and reproductive disorders

(21,39,44). Pigs from the poor rearing condition have therefore
been exposed to a more "stressful" early environment, in particular to a more stressful prenatal environment, which could contribute to increased stress susceptibility in adult life. According to
this alternative underlying mechanism, the effect of being subordinate to a penmate as a social stressor may remain without
physiological consequences in animals reared in enriched conditions, whereas subordinates from the poor rearing condition will
develop symptoms of chronic stress exposure because they are
more susceptible to stress. The observed puberty delay, the
elevated basal cortisol levels, and the delay in weight gain are in
line with this hypothesis and may possibly work synergistically
with the presumed "behavioral" mechanism.

Effects of Rearing on Reproductive Parameters

Only a few studies until now have dealt with the effects of
different rearing conditions on adult reproductive performance in
pigs. Poor rearing conditions were reported to decrease litter size
in the first and second parity (44). In addition, Schouten (38)
reported a longer parturation time together with increased biting
and snapping to the piglets at first parity, and Hemsworth et al.
(15) reported an increased conception rate at puberty after "good"
as opposed to " p o o r " handling in young piglets. In our animals,
the subordinates from the poor rearing condition showed a delayed puberty. This is in line with other studies reporting a
puberty delay after chronic stress exposure such as crowding
(8,44) or stress exposure during the prenatal or preweaning
period (15,16,46).
It may be suggested that the observed puberty delay is secondary to the delay in weight gain. However, in pigs, puberty
appears to be more determined by age than by weight (7). This
was furthermore corroborated by our data showing that puberty at
an older age was associated with a higher weight at puberty. The
mechanism underlying the puberty delay therefore does not seem
to be the delayed weight gain in these animals, although the
possibility cannot be ruled out that the subordinates from the
poor rearing group had a delayed puberty as a result of different
fat to body weight ratios.

Effects of Acute Tethering Stress

Tethering is widely used as an experimental stressor in pigs.
In this experiment, it was used to investigate the cortisol response
to an acute, nonsocial restraint stressor. This reaction was investigated because chronic stress exposure has been shown to ameliorate the capacity for negative feedback after activation of the
HPA axis (28). In accordance with previous reports, our animals
reacted to tethering stress with fierce resistence behavior, a
drastic, temporal elevation in heart rate, and an increase in saliva
cortisol levels. In addition, saliva cortisol levels were higher,
both before and after exposure to the stressor, in subordinates
from the poor rearing condition.
Although the results clearly show that subordinates from the
poor rearing condition show elevated basal cortisol levels, it
remains difficult to decide whether they also show a slower
quenching of the cortisol response after the stress-induced increase: the cortisol values after exposure to the stressor did differ
from other groups, but not from the prestress levels. Thus far, it
may therefore only be concluded that subordinates from the poor
rearing condition show elevated basal cortisol levels.

Decreased Daily Weight Gain

Also with respect to weight gain, the subordinates from the
poor rearing condition deviated from the three other groups. It is

DEVELOPMENT OF SOCIAL STRESS IN PIGS

395

important to note that differences in weight between dominant

and subordinate subjects developed only at 23 weeks of age.
Thus, differences in weight at weaning were not predictive for
later dominance/subordinate relationships within pairs. Two
mechanisms may underly the delay in weight gain in the subordinates: first the dominant may have eaten a great deal of the daily
portion of the subordinate. However, the animals were fed in
separate feeding troughs from which only one animal at a time
could eat. If the dominant decided to change place, the subordinate could always switch place as well, and this would only be
advantageous for the dominant when she would eat faster. However, if the dominants would eat a greater portion than the
subordinates, they would outgrow the animals from the "enriched" group, which they did not. A more likely explanation is
that the subordinates were less efficient in food conversion,
which is one of the effects that may accompany elevated basal
cortisol values (17,18,23). A lower daily weight gain as a result
of social stress was previously reported in a number of studies on
social stress in group-housed pigs (8,17,18,44).

Social Stress
In accordance with many studies, "social stress" was operationally defined in this study as socially induced behavioral and
physiological disturbances that are associated to chronic stress
exposure. In this experiment, it was found that chronic subordination acts as a chronic stressor, using weight development, onset
of puberty, oestrus expression, and reaction to an acute stressor
(behavioral response, heart rate response, and cortisol reponse) as
stress-associated parameters (23,49). Effects of subordination
stress of shorter or longer duration have mainly been studied in
male rodents, but also in several other species, and an elevated
HPA axis activity, reproductive failure, altered immune function,
behavioural disorders , gastric ulcers, or sudden death, among
others, have been reported (29,30,35-37,41,44). In (female) pigs,
chronic subordination stress has been reported to cause reproductive failure, delayed weight gain, altered immune function, and
increased disease susceptibility (18,19,29,31,44). It should be
taken into consideration, however, that the former results in pigs
were obtained in studies in which rearing conditions were either
"standard" or not specified at all. Most pigs, whether they are
bred commercially or for scientific reasons, are reared under
conditions comparable to our " p o o r " rearing condition and the
aforementioned results may therefore be typical for pigs reared
under " p o o r " conditions, rather than for pigs in general. In
addition, subordinate individuals in these studies are mostly
defined in terms of the low ranking individuals from grouphoused, unfamiliar pigs with a space allowance of about 1 m 2 per
pig. Thus, both aspects of " c r o w d i n g " and group composition do
contribute considerably to the effects of "subordination stress"
in the aforementioned studies.

In our experiment, social stress was studied in pairs of

familiar pigs. Are the same mechanisms underlying the development of social stress in groups? Incidental suggestions indicate
that social stress is more easily developed in pairs than in groups
of 8 - 1 0 animals or more (44). It is assumed that pigs in such
groups have more possibilities to evade aggressive dominants and
to search social support with others. The social context then
becomes more "controllable" than in a pen with only one other
pair member. Indeed, Mendl, studying aggressive strategies in
groups of pigs, found that among the pigs that regularly loose
conflicts, some simply shy away from the aggressor ( " No-success
pigs"), whereas others participate in many different conflicts
( " L o w success pigs"). Apparently, different aggressive strategies emerge in groups (29). Again, however, the importance of
ontogenetic processes should be stressed, because poor rearing
conditions may contribute in an important way to the development of the forementioned " l o w " as opposed to "no"-success
pigs.
In the present study, social relationships were studied in
familiar littermate pairs that were housed together during the
entire observation period. Under these conditions, effects of
being chronically subordinate to a familiar penmate were only
found in pigs reared under poor conditions. In this respect, it is
important to realize that dominance and submission are aspects of
the social organization of the feral pig that primarily functions to
keep group cohesion. It should therefore be anticipated that
subordination in females of this species actually does not involve
the development of stress-associated deviations under more optimal breeding conditions. In this respect, female pigs do differ,
probably considerably, from male rodents, the species in which
most of the physiological consequences of subordination stress
has been studied.

Conclusions
Being chronically subordinate to a dominant penmate may
have adverse consequences for weight development, onset of
puberty, and basal cortisol levels in adult pigs. However, this is
only true for pigs reared in a poor (i.e., barren) environment. This
effect is most likely accomplished by a "behavioral mechanism,"
implicating that disturbed agonistic behavior patterns evolve as a
result of physical and social restrictions in play behavior during
rearing. However, it may also be suggested that increased susceptibility to the stress of being subordinate occurs in pigs reared
under poor conditions.
ACKNOWLEDGEMENTS
We wish to thank Marinus C. P. Verkerk, Andr6 A. M. Jansen,
Jacobus W. M. Tijnagel, and Ben W. van den Top for excellent animal
care and valuable suggestions during experimentation, Mm C. P. J.
Oudenaarden for analysis of saliva samples, and Prof Dr T. Schuurman
for his comments on an earlier draft of the manuscript.

REFERENCES
1. Ackerman, S. H.; Manakker, S.; Cohen, M. I. Recent separation and
the onset of peptic ulcer disease in older children and adolescents.
Psychosom. Med. 43:305-310; 1981.
2. Ackerman, S. H.; Hofer, M. A.; Weiner, H. Early maternal separation
increases gastric ulcer risk in rats by producing a latent thermoregulatory disturbance. Science 201:373-376; 1978.
3. Clarke, A. S.; Schneider, M. L. Prenatal stress has long-term effects
on behavioral responses to stress in juvenile rhesus monkeys. Dev.
Psychobiol. 26:293-304; 1993.
4. Dardaillon, M.; Teillaud, P. Ethogramme du sanglier adulte et du
marcassin (Sus scrofa L.). Monitore Zool. Ital. 21:41-68; 1987.

5. De Jonge, F. H.; Mekking, P.; Abbott, K.; Wiepkema, P. R. Proceptive and receptive aspects of oestrus behaviour in gilts. Behav.
Process 31:157-166; 1994.
6. Delcroix, I.; Mauget, R.; Signoret, J. P.; Jouventin, P. Fostering in a
group of free-living wild boars (Sus scrofa L.) ?l?(in press).
7. Den Hartog, L. A.; Vesseur, P. C.; Kemp, B. Nutrition-reproduction
interactions in sows. In: Cole, D. J. A.; Wiseman, J.; Varley, M. A.,
eds. Principles of pig science. Nottingham: Nottingham University
Press; 1994:215-224.
8. Diekman, M. A.; Green, M. L.; Clapper, J. A.; Pusateri, A. E.
Environment and reproduction. In: Cole, D. J. A.; Wiseman, J.;

396

9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.

21.

22.
23.
24.

25.
26.
27.
28.

29.
30.

DE J O N G E ET AL.

Varley, M. A., eds. Principles of pig science. Nottingham: Nottingham University Press; 1994:319-333.
Glavin, G. B.; Pare, W. P. Early weaning predisposes rats to exacerbated activity-stress ulcer formation. Physiol. Behav. 34:907-909;
1985.
Glavin, G. B. Prenatal maternal stress: Differential effects upon male
and female offspring responses to restraint stress as adults. Pav. J.
Biol. Sci. 10:157-159; 1984.
Graves, H. B. Behavior and ecology of wild and feral swine (Sus
scrofa). J. Anim. Sc. 58:482-492; 1984
Gundlach, H. Bmtftirzoge, Bmtpflege, Verhaltungsontogenese und
Tagesperiodik beim Europe'ischen Wildschwein (Sus scrofa L.) Z.
Tierpsychol. 25:955-995; 1968.
Harlow, H. F.; Harlow, M. Social deprivation in monkeys. Sci. Am.
207:136-146; 1962.
Harlow, H. F.; Seay, B. M. Mothering in motherless mother monkeys. Br. J. Social Psychiatry 1:63-70; 1966.
Hemsworth, P. H.; Coleman, G. J.; Barnett, J. L. Reproductive
performance of pigs and the influence of human-animal interactions.
Pig News Information 12:563-566; 1991.
Herrenkohl, L. R. Prenatal stress reduces fertility and fecundity in
female offspring. Science 206:1097-1099; 1979.
Hessing, M. J. C.; Scheepens, C. J. M.; Schouten, W. G. P.; Tielen,
M. J. M.; Wiepkema, P. R. Social rank and disease susceptibility in
pigs. Vet. Immunol. Immunopathol. 43:373-387; 1994.
Hessing, M. J. C.; Schouten, W. P. G.; Wiepkema, P. R.; Tielen, M.
J. M. Implications of individual behavioural characteristics on performance in pigs. Livestock Prod. Sci. 40:187-196; 1994.
Hol, T.; Van Ree, J. M.; Spmijt, B. M. Behavioural and pharmacological interventions during a critical period of play behaviour;
consequences for adult social behaviour of rats.?3?(in press).
Hol, T.; Everts, H.; Van Ree, J. M.; Koolhaas, J. M.; Spmijt, B. M.
Short-term isolation after weaning selectively alters behavioral and
neuroendocrine reactions to stress during adulthood in rats. ?4?(in
press).
Janssens, C. J. J. G.; Helmond, F. A.; Wiegant, V. M. Increased
cortisol response to exogenous adrenocorticotropic hormone in chronically stressed pigs: Influence of housing conditions. J. Anim. Sci.
72:1771-1777; 1994.
Jensen, P.; Stangel, G.; Algers, B. Nursing and suckling behavior of
semi-naturally kept pigs during the first 10 days postpartum. Appl.
Anim. Behav. Sci. 31:195-210; 1991.
Johnson, E. O.; Kamilaris, T. C.; Chrousos, G. P.; Gold, P. W.
Mechanisms of stress: A dynamic overview of hormonal and behavioral homeostasis. Neurosei. Biobehav. Rev. 16:115-130; 1992.
Kalin, N. H.; Shelton, S. E.; Barksdale, C. M. Separation distress in
infant rhesus monkeys: Effects of diazepam and Ro15-1788. Brain
Res. 408:192- 198; 1987.
Lammers, G. J.; Schouten, W. G. P. Effect of pen size on the
development of agonistic behaviour in piglets. Neth. J. Agric. Sci.
33:305-307; 1985.
Levine, S.; Ursin, H. What is stress? In: Brown, M. R.; Koob, G. F.;
Rivier, C., eds. Stress, neurobiology and neuroendocrinology. New
York: Marcel Dekker Inc.; 1991:3-23.
McGill, R.; Tukey, J. W.; Larsen, W. A. Variations of box plots.
Am. Stat. 32:12-16; 1978.
Meaney, M. J.; Mitchell, J. B.; Aitken, D. H.; et al. The effects of
neonatal handling on the development of the adrenocorticol response
to stress: Implications for neuropathology and cognitive deficits later
in life. Psychoneuroendocrinology 16:85-103; 1991.
Mendl, M.; Zanella, A. J.; Broom, D. Physiological and reproductive
correlates of behavioural strategies in female domestic pigs. Anim.
Behav. 44:1107-1121; 1992.
Miczek, K. A.; Thompson, M. L.; Tornatzky, W. Subordinate ani-

31.

32.
33.
34.
35.
36.
37.

38.
39.
40.
41.

42.
43.
44.
45.
46.

47.

48.

49.
50.

mals: Behavioral and physiological adaptations and opioid tolerance.

In: Brown, M. R.; Kooh, G. F.; Rivier, C., eds. Stress, neurobiology
and neuroendocrinology. New York: Marcel Dekker Inc.; 1991:3-23.
Morm~de, P.; Lemaire, V.; Castanon, N.; Dullac, J.; Laval, M.; Le
Moal, M. Multiple neuroendocrine responses to chronic social stress:
Interaction between individual characteristics and situational factors.
Physiol. Behav. 47:1099-1105; 1990.
Newberry, R. C.; Wood-Gush, D. G. M. Social relationships of
piglets in semi-natural environment. Anim. Behav. 34:1311-1318;
1986.
Newberry, R. C.; Wood-Gush, D. G. M.; Hall, J. W. Playful
behaviour of piglets. Behav. Process 17:205-216; 1988.
Pearce, G. P.; Paterson, A. M. The effect of space restriction and
provision of toys during rearing on the behaviour, productivity and
physiology of male pigs. Appl. Anim. Behav. Sci. 36:11-28; 1993.
Sachser, N. The ability to arrange with conspecifics depends on
social experiences around puberty. Physiol. Behav. 53:539-544;
1993.
Sapolsky, R. M. Stress in the wild. Sci. Am. 262:106-113; 1990.
Sapolsky, R. M. Effects of stress and glucocorticoids on hippocampal
neuronal survival. In: Brown, M. R.; Koob, G. F.; Rivier, C., eds.
Stress, neurobiology and neuroendocrinology. New York: Marcel
Dekker Inc.; 1991:3-23.
Schouten, W. G. P. Effects of rearing on subsequent performance in
pigs. Pig News Information 12:245-247; 1991.
Schouten, W.; Rushen, J.; De Pasillr, A. B. Stereotypic behavior and
heart rate in pigs. Phys. Behav. 50:617-624; 1990.
Smith, P. K. Does play matter? Functional and evolutionary aspects
of animal and human play. Behav. Brain Sci. 5:139-184; 1982.
Stem, J.; Winokur, A.; Eisenstein, A.; Taylor, R.; Sly, M. The effect
of group vs. individual housing on behavior and physiological responses to stress in the albino rat. J. Psychosom. Res. 4:185-190;
1960.
Stolba, A.; Wood-Gush, D. G. M. The identification of behavioural
key features and their incorporation into a housing design for pigs.
Ann. Reck Vet. 15:287-299; 1984.
Sucheki, D.; Mozaffarian, D.; Rosenfeld, P.; Levine, S. Effects of
maternal deprivation on the ACTH stressresponse in the infant rat.
Neuroendocrinology 57:204-212; 1993.
Varley, M.; Stedman, R. Stress and reproduction. In: Cole, D. J. A.;
Wiseman, J.; Varley, M. A., eds. Principles of pig science. Nottingham: Nottingham University Press; 1994:277-297.
Von Beuerle, W.; Freilanduntersuchungen zum Kampf- und Sexualverhalten des Europiiischen Wildschweines (Sus scrofa L.). Z.
Tierpsychol. 39:211-258; 1975.
Ward, I. L. Sexual behavior: The product of perinatal hormonal and
prepubertal social factors. In: Gerall, A. A.; Moltz, H.; Ward, I. L.,
eds. Handbook of behavioral neurobiology, vol. 11, sexual differentiation. New York: Plenum Press; 1992:157-175.
Weinstock, M.; Matlina, E.; Maor, G. I.; Rosen, H.; McEwen, B. S.
Prenatal stress selectively alters the reactivity of the hypothalamicpituitry-adrenal system in the female rat. Brain Res. 595:195-200;
1992.
Wiepkema, P. R.; Schouten, W. G. P. Mechanisms of coping in
social situations. In: Zayan, R.; Dantzer, R., eds. Social stress in
domestic animals. Curr. Topics Vet. Med. Anim. Sci. Boston: Kluwer
Academic; 1990:8-25.
Zayan, R. The specificity of social stress. Behav. Process 25:81-95;
1991.
Zeeb, K.; Bock, Ch.; Heinzler, B. Control of social stress by consideration of suitable social space. In: Zayan, R.; Dantzer, R., eds.
Social stress in domestic animals. Curr. Topics Vet. Med. Anim. Sci.
Boston: Kluwer Academic; 1990:275-282.