The effect of mirrors on the behaviour of singly housed male and female laboratory rabbits

Abstract

It is widely recognised that single housing is detrimental to the welfare of social species. However, some experimental proce-dures dictate that laboratory animals are housed individually. There is evidence to suggest that, by mimicking social contact, mirrors are beneficial to the welfare of singly housed horses and heifers. However, experiments with singly caged laboratory mice have found mirrors to be mildly aversive. The present study investigated the behavioural response of singly housed male and female rabbits to a mirror in their cage. After a period of pre-trial behavioural observations, rabbits were provided with an acrylic mirror, either at the front or the back of their cage, for a period of seven days. This was followed by a post-trial period, at the beginning of which all mirrors were removed. Both sexes showed some changes in behaviour and in the use of space within the cage. The addition of a mirror significantly reduced grooming in females, which was previously considered to be at high levels. Both males and females showed an increase in investigatory behaviour, although the patterns of change differed between the sexes. Differences between males and females are attributed to differences in socio-sexual strategies between the sexes. It is concluded that, when single housing is unavoidable, mirrors might be appropriate to partially compensate for social contact in female laboratory rabbits. Further research using a wide range of welfare indicators is needed to establish whether the provision of a mirror could be used as a successful method of improving laboratory rabbit welfare.

Full text

461
© 2010 Universities Federation for Animal Welfare
The Old School, Brewhouse Hill, Wheathampstead,
Hertfordshire AL4 8AN, UK
Animal Welfare 2010, 19: 461-471
ISSN 0962-7286
The effect of mirrors on the behaviour of singly housed male and female
laboratory rabbits
JL Edgar*
†
and SC Seaman
‡
†
Division of Farm Animal Science, Department of Clinical Veterinary Science, University of Bristol, Langford, Bristol BS40 5DU, UK
‡
Royal (Dick) School of Veterinary Studies, Division of Veterinary Clinical Sciences, University of Edinburgh, Easter Bush Veterinary
Centre EH25 9RG, UK
* Contact for correspondence and requests for reprints: j.edgar@bris.ac.uk
Abstract
It is widely recognised that single housing is detrimental to the welfare of social species. However, some experimental proce-
dures dictate that laboratory animals are housed individually. There is evidence to suggest that, by mimicking social contact,
mirrors are beneficial to the welfare of singly housed horses and heifers. However, experiments with singly caged laboratory mice
have found mirrors to be mildly aversive. The present study investigated the behavioural response of singly housed male and
female rabbits to a mirror in their cage. After a period of pre-trial behavioural observations, rabbits were provided with an acrylic
mirror, either at the front or the back of their cage, for a period of seven days. This was followed by a post-trial period, at the
beginning of which all mirrors were removed. Both sexes showed some changes in behaviour and in the use of space within the
cage. The addition of a mirror significantly reduced grooming in females, which was previously considered to be at high levels.
Both males and females showed an increase in investigatory behaviour, although the patterns of change differed between the
sexes. Differences between males and females are attributed to differences in socio-sexual strategies between the sexes. It is
concluded that, when single housing is unavoidable, mirrors might be appropriate to partially compensate for social contact in
female laboratory rabbits. Further research using a wide range of welfare indicators is needed to establish whether the provision
of a mirror could be used as a successful method of improving laboratory rabbit welfare.
Keywords: animal welfare, behaviour, laboratory rabbit, mirror, single caging, social isolation
Introduction
Rabbits (Oryctolagus cuniculus) are a naturally social species
with social stability within the group maintained through
aggressive chasing and submissive retreat (Mykytowycz 1958;
Lehmann 1991). Despite decades of domestication, the rabbits’
social and sexual behaviour has changed little (Stodart &
Myers 1964; Vastrade 1986; Lehmann 1991; Held et al 1995).
An investigation by Chu et al (2004) found that pair-housed
laboratory rabbits spent 26.7% of observed time in direct
physical contact, indicating the social nature of the species.
According to UK Home Office statistics, in 2009
11,643 rabbits were used in 16,562 procedures (Home Office
2009); putting them in the top four most commonly used
laboratory mammals. Female laboratory rabbits are often
housed singly for experimental reasons, such as topical drug
administration (to avoid interference/ingestion of drugs by
conspecifics), a fast turnover of rabbits (which would result
in the mixing of unfamiliar animals), as well as aggression
(Seaman 2002). A review of the pharmaceutical industry in
1999 revealed that 100% of male rabbits were housed singly
(Seaman 2002). Indeed, group housing of mature males is
not advisable. Aggression is the major factor governing the
housing of males and is also one of the reasons for housing
female rabbits singly (Seaman 2002). Single cages also offer
several benefits over floor pens, including increased
hygiene, ease of monitoring individual health, efficient use
of vertical space and ease of identification (Chu et al 2004).
However, for highly social animals, it has been suggested
that single housing can compromise welfare and it is
therefore recommended that female laboratory rabbits be
housed socially where possible (Morton et al 1993).
Experiments comparing group- and singly housed female
rabbits have revealed marked differences in behavioural
time budgets (Podberscek et al 1991; Chu et al 2004). Chu
et al (2004) compared the behaviour of singly housed and
pair-housed laboratory rabbits and found that singly housed
rabbits showed a significantly greater rate of development
of abnormal and stereotypic behaviours than their pair-
housed counterparts. In a similar comparison, Podberscek
et al (1991) found that stereotypic behaviours were only
seen in singly caged rabbits, and totalled 6.3% of observed
behaviours, compared to group-housed rabbits, which
Universities Federation for Animal Welfare Science in the Service of Animal Welfare

462 Edgar and Seaman
showed no stereotypic behaviours, but were housed in large
floor pens which differed physically to the cages.
Stereotypic behaviours performed by singly housed rabbits
include bar biting (Lidfors 1997), pawing of the cage walls
(Podberscek et al 1991; Chu et al 2004) and floor chewing
(Chu et al 2004). It has also been suggested that licking
objects and hair chewing in females may be due, in part, to
social deprivation, caused by the inability to allogroom
(Gunn & Morton 1995). Additionally, various studies have
indicated that female rabbits choose to have contact with
conspecifics when given the opportunity (Huls et al 1991;
Held et al 1995; Seaman et al 2008).
Castration of male laboratory rabbits has been considered as a
viable option to facilitate social housing where possible (Gunn
1994; Kalagassy et al 1999); however, some procedures
require intact males. It has been established that social housing
of mature males results in injurious aggression (eg Gunn
1994). However, it has not been determined whether social
contact for males is stressful per se, or whether, were the
aggression thwarted, for example via the use of a mirror panel,
the welfare of singly housed males would be improved. It is
plausible that rabbits might find mirror exposure stressful,
especially since there is little chance of escape from the
perceived social contact. In fact, mirror chambers have been
used to induce anxiety in rodents in human drug studies
(Toubas et al 1990). In keeping with this, in preference tests,
it has been found that singly housed mice find mirrors mildly
aversive, especially during feeding (Sherwin 2004).
Experiments have shown that mirror use has reduced the
incidence of stereotypic weaving and nodding in socially
isolated horses (McAfee et al 2002) and reduced the heart rate
of isolated heifers (Piller et al 1999). Jones and Phillips (2005)
found that mirrors increased behavioural complexity (as deter-
mined by the number of behaviours performed per minute), in
pet rabbits and so may offer some advantages to welfare that
might be applicable to the laboratory environment.
Previous attempts have been made to refine traditional
single caging in order to meet the social and physical needs
of laboratory rabbits (eg Gerson 2000). Replacing older
designs of caging with such new caging, or modifying
original caging may prove expensive. It is possible that an
inexpensive alternative may be the use of mirrors in existing
cages. This would potentially simulate avoidable visual
contact without the risk of injurious aggression.
In previous experiments on rabbits it has been difficult to
establish the importance of social contact independently of
other factors; social pens are generally larger than single
cages and are usually provisioned with physical enrichment
items, such as hanging devices or cardboard boxes (Chu et al
2004). The use of mirrors would determine the effect of
perceived social contact on the rabbits without the usual asso-
ciated increase in cage size or provisions or olfactory contact.
The aim of this study was to determine whether the
presence of mirrors in cages has an influence on the behav-
ioural repertoire of singly caged male and female laboratory
rabbits. The results may then be used to assess whether a
mirror could be appropriate as a practical method to
improve the welfare of singly housed laboratory rabbits.
Materials and methods
Subjects, housing and husbandry
Nine male and 15 female New Zealand White rabbits were
used in this study. The rabbits were 15–16 weeks old at the
start of the study. They had been singly housed from
weaning in standard metal laboratory rabbit cages
(740 × 740 × 450 mm; length × breadth × height), with solid
metal walls and plastic dimpled flooring (Figure 1). All
subjects were housed in the top two rows of a standard three-
row rack of cages, which were positioned facing a wall
which was at a distance of 1.5 m. The subjects were not in
visual contact, but had olfactory and auditory contact with
other rabbits in the room. Rabbits had ad libitum access to
standard rabbit pellets (Rabma pellets, SDS, Whitham,
Essex, UK) and water. Lights remained on between 0530
and 2330h giving a light: dark ratio of 18L:6D. Urine and
faeces were removed from trays underneath the cages twice
weekly. The rabbits were concurrently used for antibody
production and blood was collected every week from an ear
vein. They were not handled at any other time.
Procedure
The rabbits were continuously recorded using a 24-h
time-lapse video recorder (Panasonic AG-6024). Two
CCTV cameras (Panasonic WV-BP330/B) were used to
film four cages each, with a quad box (Panasonic Quad
System WJ410) allowing eight cages to be viewed at a
time. The 24 rabbits were therefore split into three groups
for observations. Observation group one consisted of
eight males, group two of one male and seven females,
and group three consisted of eight females. An infrared
lamp was used to allow filming during the dark period.
When rabbits were handled for experimental procedures
or the trays underneath the cages were changed, observa-
tions ceased for one hour. Observations were split into
pre-trial, mirror-trial and post-trial periods in order to
determine the effect of the mirror on the behaviour of the
rabbits before, during, and after mirror exposure.
Pre-trial observations — day 1-3
Pre-trial observations were carried out to enable each rabbit
to act as its own control and provide a baseline behavioural
time budget.
Mirror-trial observations — day 4-10
An acrylic mirror (Westward Plastics, Bristol, UK)
(300 × 300 mm) was attached to the inside of each of the
eight cages using a standard metal nut and bolt at each
corner of the mirror. To avoid the effect of any positional
preference within the cage, four males were housed in cages
with a mirror fitted at the front of the cage and five at the
back, and eight females were housed in cages with a mirror
at the front and seven at the back. For both sexes, rabbits
were randomly assigned to the mirror-left or the mirror-
right position (see Figure 2). The size of the mirror meant
that rabbits could avoid their reflection by moving to
another part of the cage.
© 2010 Universities Federation for Animal Welfare

The effect of mirrors on rabbit behaviour 463
Animal Welfare 2010, 19: 461-471
Figure 1
Standard cage set-up during the study.
Figure 2
Mirror positions within the cage.

464 Edgar and Seaman
Post-trial observations — day 11-13
Mirrors were removed and behaviour was recorded for a
further three days.
Measurements
Instantaneous scan sampling at 10-min intervals was used to
record behaviour and position in the cage. An ethogram was
adapted from Gunn and Morton (1995). The behaviours
were subsequently divided into categories (Table 1). The
observation schedule and behavioural ethogram were
validated using a pilot study in which four male rabbits
were observed for two days ‘pre-trial’, 13 days ‘mirror-trial’
and seven days ‘post-trial’.
Position in the cage was recorded as either at the front or at
the back. This was used to determine time spent in front of
and away from the mirror. Position in the cage was deter-
mined by which half of the cage the rabbit’s head was in.
Statistical analysis
For each observation period, the percentage of time spent
performing each category of behaviour was used to
determine a behavioural time budget. A mixed between-
within subjects analysis of variance was conducted to assess
the impact of gender and mirror position on the rabbits’
behaviour across the three periods (pre-trial, mirror-trial and
post-trial periods) and to assess between-day variations in
behaviour during the mirror-trial period. All data were
checked to ensure they met the assumptions of ANOVA;
data did not violate Levene’s test of equality of error
variances or Box’s test of equality of covariance matrices.
Data were analysed using SPSS 17.0.
Results
Figure 3 shows the percentage of time rabbits spent
performing behaviours within each behavioural category,
© 2010 Universities Federation for Animal Welfare
Table 1 Ethogram of behaviour (adapted from Gunn & Morton 1995).
Behaviour Description
Inactive
Sit up Abdomen against floor, forelimbs straight so that thorax is clear of floor
Sit down Body against the floor, but still being supported by the fore and hind limbs
Lie down Body stretched against floor
Lie stretched Rabbit on its side with hind legs extended
Eat/drink
Eat Taking and chewing of food pellets from the hopper
Drink Licking water from the spout of the water bottle
Coprophagy Period in which rabbit has its mouth beneath its body and is chewing
Body maintenance
Groom Nibbling or biting of the coat
Wash Licking the coat with sweeping movements or face washing using the forepaws
Scratch The hind feet are used to scratch the body or head
Mobility
Ambulate Movement around the cage, with the exception of frisky hop
Frisky hop Rapid hopping around the cage
Stereotypy
Scratch corner Digging on the floor
Bar bite Chewing of the cage bars
Investigatory behaviour
Sniff cage Sniffing of the environment
Rear Both forepaws raised from the floor
Comfort behaviour
Roll Rabbit throws itself over onto its side (usually precedes lie stretched)
Stretch Prolonged extension of the front legs or arching of the back
Mirror-related behaviour
Scratch mirror Forepaws used to scratch mirror
Sniff mirror Sniffing of the mirror
Bite mirror Biting of the edges of the mirror

The effect of mirrors on rabbit behaviour 465
Animal Welfare 2010, 19: 461-471
Figure 3
Behavioural categories observed in (a) males and (b) females during pre-trial, mirror-trial and post-trial periods. INA: Inactive; ED:
Eat/drink; BM: Body maintenance; MB: Mobility; ST: Stereotypy; INV: Investigatory behaviour; CF: Comfort behaviour; MI: Mirror-related
behaviour. * Statistically significant difference between periods at P < 0.05.

466 Edgar and Seaman
according to treatment period. During each of the observa-
tions periods, both male and female rabbits spent the
majority of their time inactive (on average between 65 and
70%) and spent over 10% of their time in both
eating/drinking and body maintenance.
Inactive
There was no main effect for observation period (Wilks’
lambda = 0.934; F
2,11
= 0.453; P = 0.647; partial eta
squared = 0.076) and no significant interaction between
period and gender (Wilks’ lambda = 0.909; F
2,11
= 0.549;
P = 0.593; partial eta squared = 0.091) or period and mirror
position (Wilks’ lambda = 0.875; F
2,11
= 0.785; P = 0.480;
partial eta squared = 0.125).
During the seven days of the mirror-trial period there was a
significant main effect for day (Wilks’ lambda = 0.384;
F
6,15
= 4.009; P = 0.014; partial eta squared = 0.616), with
both genders showing an increase in inactive behaviours
over time. There was no significant interaction between day
and gender (Wilks’ lambda = 0.767; F
6,15
= 0.760;
P = 0.612; partial eta squared = 0.233) and day and mirror
position (Wilks’ lambda = 0.670; F
6,15
= 1.231; P = 0.345;
partial eta squared = 0.330).
Eating and drinking
There was no main effect for observation period (Wilks’
lambda = 0.791; F
2,11
= 1.586; P = 0.245; partial eta
squared = 0.209) and no significant interaction between
period and gender (Wilks’ lambda = 0.702; F
2,11
= 2.548;
P = 0.120; partial eta squared = 0.298) or period and mirror
position (Wilks’ lambda = 0.843; F
2,11
= 1.119; P = 0.358;
partial eta squared = 0.157).
During the seven days of the mirror-trial period there was
no significant main effect for day (Wilks’ lambda = 0.312;
F
6,15
= 5.517; P = 0.85; partial eta squared = 0.093). There
was no significant interaction between day and gender
(Wilks’ lambda = 0.585; F
6,15
= 1.773; P = 0.172; partial eta
squared = 0.415) and day and mirror position (Wilks’
lambda = 0.585; F
6,15
= 1.777; P = 0.172; partial eta
squared = 0.415).
Body maintenance
There was a substantial main effect for observation period
(Wilks’ lambda = 0.54; F
2,11
= 4.64; P = 0.034; partial eta
squared = 0.458) and a significant interaction between
period and gender (Wilks’ lambda = 0.58; F
2,11
= 3.94;
P = 0.006; partial eta squared = 0.349), with females, but
not males, showing a significant reduction in body mainte-
nance across the three time periods. There was no signifi-
cant interaction between period and mirror position (Wilks’
lambda = 0.70; F
2,11
= 2.38; P = 0.138; partial eta
squared = 0.302).
During the seven days of the mirror-trial period there was
no main effect for day (Wilks’ lambda = 0.495;
F
6,15
= 2.553; P = 0.066; partial eta squared = 0.505) and no
significant interaction between day and gender (Wilks’
lambda = 0.698; F
6,15
= 1.081; P = 0.417; partial eta
squared = 0.302) or day and mirror position (Wilks’
lambda = 0.970; F
6,15
= 0.076; P = 0.998; partial eta
squared = 0.030).
Mobility
There was no main effect for observation period (Wilks’
lambda = 0.998; F
2,11
= 0.009; P = 0.991; partial eta
squared = 0.002) and no significant interaction between
period and gender (Wilks’ lambda = 0.981; F
2,11
= 0.106;
P = 0.900; partial eta squared = 0.019) or period and mirror
position (Wilks’ lambda = 0.887; F
2,11
= 0.698; P = 0.518;
partial eta squared = 0.113).
During the seven days of the mirror-trial period there was
no significant main effect for day (Wilks’ lambda = 0.795;
F
6,15
= 0.792; P = 0.590; partial eta squared = 0.241). There
was no significant interaction between day and gender
(Wilks’ lambda = 0.802; F
6,15
= 0.616; P = 0.719; partial eta
squared = 0.198) and day and mirror position (Wilks’
lambda = 0.670; F
6,15
= 1.231; P = 0.345; partial eta
squared = 0.330).
Stereotypy
Stereotypic behaviours that were seen were bar biting and
scratching the corner of the cage. Female rabbits spent on
average 0.8% of their time engaged in stereotypic behaviour,
and males spent 1.2% of their time in these activities.
There was no main effect for observation period (Wilks’
lambda = 0.882; F
2,11
= 0.738; P = 0.500; partial eta
squared = 0.118) and no significant interaction between
period and gender (Wilks’ lambda = 0.973; F
2,11
= 0.152;
P = 0.861; partial eta squared = 0.027) or period and mirror
position (Wilks’ lambda = 0.807; F
2,11
= 1.312; P = 0.308;
partial eta squared = 0.193).
During the seven days of the mirror-trial period, there was a
significant main effect for day (Wilks’ lambda = 0.457;
F
6,15
= 2.969; P = 0.041; partial eta squared = 0.543) and a
significant interaction between day and gender (Wilks’
lambda = 0.495, F
6,15
= 2.949; P = 0.042, partial eta
squared = 0.541). Figure 4 illustrates that males showed an
increase in stereotypic behaviour during days 1–2 of the
mirror-trial period, followed by a steep decline and then a
subsequent gradual increase into the post-trial period.
Females showed more a consistent level of stereotypic
behaviour, with a slight increase during days 1–2 of the
mirror-trial period, followed by a slight, gradual increase
into the post-trial period. There was no significant interac-
tion between day and mirror position (Wilks’
lambda = 0.640; F
6,15
= 1.405; P = 0.276; partial eta
squared = 0.360).
Investigatory behaviour
There was a main effect for observation period (Wilks’
lambda = 0.561; F
2,11
= 4.304; P = 0.042, partial eta
squared = 0.439), with both genders showing an increase in
investigatory behaviour. There was no significant interac-
tion between period and gender (Wilks’ lambda = 0.795;
F
2,11
= 1.417; P = 0.283; partial eta squared = 0.205) or
period and mirror position (Wilks’ lambda = 0.599;
F
2,11
= 3.675; P = 0.060; partial eta squared = 0.401).
During the seven days of the mirror-trial period, there was a
significant substantial main effect for day (Wilks’
© 2010 Universities Federation for Animal Welfare

The effect of mirrors on rabbit behaviour 467
lambda = 0.297; F
6,15
= 5.905; P = 0.002; partial eta
squared = 0.703) and a significant substantial interaction
between day and gender (Wilks’ lambda = 0.433;
F
6,15
= 3.276; P = 0.029; partial eta squared = 0.567).
Figure 5 illustrates that males showed an increase in inves-
tigatory behaviour during days 1–2 of the mirror-trial
period, and then a subsequent decline. Females showed a
similar, but less dramatic, pattern of behaviour. There was
no significant interaction between day and mirror position
(Wilks’ lambda = 0.487; F
6,15
= 2.631; P = 0.060; partial eta
squared = 0.513).
Comfort behaviour
There was no main effect for observation period (Wilks’
lambda = 0.796; F
2,11
= 1.411; P = 0.285; partial eta
squared = 0.204) and no significant interaction between
period and gender (Wilks’ lambda = 0.795; F
2,11
= 1.398;
P = 0.288; partial eta squared = 0.203) or period and mirror
position (Wilks’ lambda = 0.866; F
2,11
= 0.854; P = 0.452;
partial eta squared = 0.134).
During the seven days of the mirror-trial period there was
no significant main effect for day (Wilks’ lambda = 0.915;
F
6,15
= 0.232; P = 0.960; partial eta squared = 0.085). There
was no significant interaction between day and gender
(Wilks’ lambda = 0.813; F
6,15
= 0.573; P = 0.746; partial eta
squared = 0.187) and day and mirror position (Wilks’
lambda = 0.847; F
6,15
= 0.451; P = 0.833; partial eta
squared = 0.153).
Mirror-related behaviour
Overall, female rabbits were observed sniffing, scratching
and biting the mirror 0.21% of the time. Males were
observed sniffing the mirror only and spent on average
0.15% of their time engaged in this activity.
During the seven days of the mirror-trial period there was
no significant main effect for day (Wilks’ lambda = 0.621;
Animal Welfare 2010, 19: 461-471
Stereotypic behaviour during the pre-trial
(Pre 1-3), mirror-trial (M1-7) and post-
trial (Post 1-4) periods.
Investigatory behaviour during the pre-
trial (Pre 1-3), mirror-trial (M1-7) and
post-trial (Post 1-4) periods.
Figure 4
Figure 5

468 Edgar and Seaman
F
6,15
= 1.524; P = 0.237; partial eta squared = 0.379). There
was no significant interaction between day and gender
(Wilks’ lambda = 0.748; F
6,15
= 0.844; P = 0.556; partial eta
squared = 0.252) and day and mirror position (Wilks’
lambda = 0.933; F
6,15
= 0.178; P = 0.979; partial eta
squared = 0.067).
Time spent at the front and back of the cage
There was a substantial main effect for observation period
(Wilks’ lambda = 0.534; F
2,11
= 4.799; P = 0.043; partial eta
squared = 0.466) and a significant interaction between
period and gender (Wilks’ lambda = 0.534; F
2,11
= 6.084;
P = 0.017; partial eta squared = 0.525), with a decrease in
time spent at the front of the cage during the mirror-trial
period for females (25% pre-trial vs 31% mirror-trial) and
an increase for males (34% pre-trial vs 33% mirror-trial).
There was no significant interaction between period and
mirror position (Wilks’ lambda = 0.894; F
2,11
= 0.651;
P = 0.541; partial eta squared = 0.106).
During the seven days of the mirror-trial period there was
no main effect for day (Wilks’ lambda = 0.833;
F
6,15
= 0.502; P = 0.797; partial eta squared = 0.167) and no
significant interaction between day and gender (Wilks’
lambda = 0.524; F
6,15
= 2.270; P = 0.093; partial eta
squared = 0.476) or day and mirror position (Wilks’
lambda = 0.805; F
6,15
= 0.607; P = 0.721; partial eta
squared = 0.195).
Discussion
Both the behavioural repertoire of the rabbits and the use of
space within the cage were affected by the presence of a
mirror. This was influenced by the gender of the rabbits but
not the position of the mirror within the cage.
Male and female rabbits showed a number of differences in
behavioural response to the addition of a mirror to their cage.
Females showed a decrease in body maintenance, an increase
in investigatory behaviour and spent more time at the back of
the cage. The decreased levels of body maintenance and time
spent at the back of cage continued into the post-trial period.
Males showed an increase in investigatory behaviour and an
increase in time spent at the front of the cage. For males,
investigatory behaviour increased further during the post-
trial period, and there was also a further increase in time
spent at the front of the cage during this period.
Perhaps the most interesting change in terms of welfare,
which occurred with the provision of a mirror, was a
decrease in grooming which, during pre-trial observations,
was at similar levels to those recorded in previous studies
(see Gunn & Morton 1995). Females in the present study
spent 15% of their time budget grooming, whilst males
spent 13% of their time in this activity. Gunn and Morton
(1995) reported levels of 10% for singly caged males and
females. They suggested that behaviours such as hair
chewing in singly caged rabbits might be due to social
deprivation, ie an inability to allogroom. It has also been
reported that singly caged male and female rats spend more
time grooming than their group-housed counterparts (Hurst
et al 1997). In addition to social deprivation, excessive
grooming has been attributed to a barren environment
(Berthelsen & Hansen 1999) and the absence of hay (Leslie
et al 2004).
The decrease in grooming seen in the present experiment
supports the theory that increased levels of grooming may be
as a result of a lack of social stimulation, although no differ-
entiation could be made in this study between grooming and
hair chewing. In the present study, a decrease in body main-
tenance in rabbits with mirrors cannot be attributed to an
opportunity for allogrooming. Excessive grooming
behaviour in mice can be induced by a stressful situation
such as a novel cage (Dunn et al 1981) and also by an
injection of ACTH (Guild & Dunn 1982), the precursor of
the stress hormone, corticosterone. The fact that grooming
levels for both sexes were slightly higher in the present study
than previously reported for singly caged rabbits could be
attributed to the fact that in the present study the rabbits had
no enrichment or hay and so may not represent the ‘typical’
laboratory situation. In the present study, grooming
remained relatively high after the addition of a mirror, and
was unchanged in the case of male rabbits, and this supports
the hypothesis that excessive grooming might also be caused
by the stress of a barren environment, or under-stimulation
caused by a barren environment (Hansen & Berthelsen
2000). In fact, excessive grooming seems to be a complex
issue, and has also been ameliorated with the addition of hay
(Berthelsen & Hansen 1999) and a coarse mix diet (Leslie
et al 2004). The fact that rabbits in the present study had no
access to hay or enrichment and were fed on pelleted food
may explain, to a certain extent, why grooming did not
decrease further after the addition of a mirror. Further studies
are needed to reveal the relationships between social contact,
diet, enrichment and grooming levels.
Both males and females showed a gradual increase in inac-
tivity throughout the seven-day mirror-trial period. A compar-
ison of the behaviour of individually and pair-housed female
rabbits revealed that resting was higher in pair-housed rabbits
(Chu et al 2004). Chu and colleagues suggested that this
might be as a result of social facilitation of resting. Social
facilitation has been used to explain increases in food pecking
in domestic chicks provided with mirrors (Montevecchi &
Noel 1978), and might have been the motivation for the
increase in inactive behaviours in this study.
The fact that a mirror reduced the amount of grooming in
female, but not male rabbits, adds further weight to the
hypothesis that rabbits might perceive the mirror reflection
as a conspecific (Anderson 1994). In agreement with this,
Mills and Davenport (2002) compared the effect of a neigh-
bouring conspecific versus a mirror panel in stabled horses
and found that they induced the same reduction in weaving
behaviour. Such a study with rabbits would help to confirm
whether the mirror was actually being perceived as social
contact. Although in the present study males showed a
different behavioural response to the presence of a mirror
than females, it is highly likely that they still view the
reflection as a conspecific. In fact, the difference in
response between males and females could be due to differ-
ences in social strategies. This was discussed in rats by
© 2010 Universities Federation for Animal Welfare

The effect of mirrors on rabbit behaviour 469
Hurst et al (1998, 1999), where differences between the
sexes in behavioural response to group housing was attrib-
uted to differences in strategies of social and reproductive
competition, with previously singly housed males
responding to their cage as a territory and singly housed
females seeking to re-establish social contact.
Indeed, the peak in time spent engaged in investigatory
behaviour in male rabbits during days 1–2 of the mirror-trial
period may have represented an increase in vigilance,
caused by the perceived conspecific. Taking this into
account, it is possible that the male rabbits in the present
experiment could have been highly stressed by the addition
of perceived social contact, due to limited space and scant
opportunity for avoidance. This could be exacerbated by the
fact that the reflected image was the same size and therefore
potentially of equal social status to the rabbit itself. In an
experiment carried out by Seaman et al (2008), rabbits were
given the opportunity to work to gain access to social
contact with a conspecific through a mesh panel. A correla-
tion was found between the difference in weight between
test and stimulus rabbits and time spent in front of the mesh
panel by the test rabbit, with test rabbits that were heavier
than their stimulus rabbit spending more time in front of the
mesh. Mixing of males of similar weight and age provokes
aggression, and is intensified by a close similarity in
dominance ability (Parker 1974). With this in mind, it
would be expected for males to be aggressive towards the
mirror, especially during the first few days. Indeed, interac-
tions with the mirror panel were seen, but only rarely
(0.15% in males) and the fact that females showed slightly
more interactions (0.21%) indicates that these may have
been thwarted attempts to socially interact with the
perceived conspecific.
It can be argued that the rabbits’ behavioural response to the
mirror may be due to an increase in the novelty of their envi-
ronment, rather than perceived social contact. This was
suggested by McAfee et al (2002) as a possible mechanism
for a mirror-induced reduction of weaving behaviour in
horses. If this were the case then males would likely respond
to the mirror in a similar way to females. Additionally, inter-
actions with the mirror and investigatory behaviour would be
expected to peak only in the period immediately following
introduction of the mirror and would then decrease over time.
In the present study, this does seem to be the case for the
males. However, for the females, only a slight peak is
observed and then investigatory behaviour remains relatively
consistent for the duration of the mirror-trial period.
Additionally, there was no significant decrease in time spent
interacting with the mirror for both sexes. This was in accor-
dance with Jones and Phillips (2005), who found that rabbits
did not habituate to the presence of a mirror throughout a
seven-week period. This supports the hypothesis that singly
housed animals’ responses to mirrors (eg in horses; McAfee
et al 2002) are indeed as a result of perceived social contact
and not of novelty as some authors have suggested.
Studies comparing singly and group-housed rabbits have
revealed a greater likelihood of stereotypic behaviour in
singly housed rabbits (eg Chu et al 2004). It would be
expected then, if the rabbits in this study perceived the
mirror-image as a conspecific, that the addition of a mirror
would significantly reduce stereotypic behaviour. Neither
males nor females showed a significant difference in the
performance of stereotypic behaviour after the addition of
the mirror. However, there is evidence to suggest that once
stereotypic behaviour patterns are ingrained, even when an
animal is removed from the situation, the stereotypic
behaviour pattern remains (Mason 1991). Although there
was no significant difference in stereotypic behaviours in
the present study, as Figure 4 illustrates, males do show an
increase in stereotypic behaviour during the first two days
of the mirror-trial period. In fact, bar biting and scratching
the corner of the cage might have represented attempts to
escape by the males. There was no evidence to suggest that
male rabbits were trying to escape, or that they found the
mirror exposure stressful (such as a ‘freeze’ response), but
validation of behavioural with physiological measures
could give a clearer indication, and remains a topic
requiring further study.
Another gender difference in the response to the addition of
a mirror was that males spent more time at the front of the
cage while females spent more time at the back. It is
difficult to speculate possible motives for this behaviour
without adequate knowledge of the rabbits’ perception of
the front and back of the cage. A variety of factors could
potentially cause differences between the front and back of
the cage. These include differing light levels or physical
qualities which may have affected the perceived mirror
image. If males felt threatened by the mirror then they may
have been motivated to escape, causing them to spend more
time at the front of the cage. Indeed, bar biting is a
behaviour which can only be performed at the front of the
cage, and males did show a peak in stereotypic behaviour
during days 1–2 of the mirror-trial period. Conversely, the
back of the cage might be viewed as a refuge, and so this
also needs to be considered. Perhaps if both sexes found the
mirror exposure stressful, the differences in space use in the
cage might represent different coping strategies, with males
becoming active and attempting to escape, and females
opting for a more passive strategy at the back of the cage.
In 1990, Toubas and colleagues introduced the mirror
chamber, a novel test of anxiety for mice. This test was
based on the assumption that many species show approach-
avoidance behaviour when they are confronted by a mirror.
Mice showed an extended latency to enter the mirrored
chamber and the anxiolytic drug, diazepam, reduced this
latency. In keeping with this, Sherwin (2004) found that
mice find mirrors mildly aversive when they are placed
adjacent to the feeder. In Sherwin’s study, the presence of
the mirror significantly reduced feeding. In the present
study, however, time spent feeding remained unaffected by
the presence of a mirror.
The fact that the present study found differences in response
to the mirror between males and females highlights the
importance of studying both sexes when assessing the
welfare implications of new enrichment or caging. Previous
Animal Welfare 2010, 19: 461-471

470 Edgar and Seaman
studies have looked at sex differences in behavioural
response to enrichment (eg Hansen & Berthelsen 2000).
However, attempts to improve the physical and social envi-
ronment are largely dominated by studies using only female
rabbits (eg Chu et al 2004), although some attempts have
been successful by providing physical enrichment (hay and
grass cubes) to males (Lidfors 1997).
Animal welfare implications and conclusion
The results of the present study indicate that when single
housing cannot be avoided, mirrors might be appropriate to
partially compensate for social contact in female rabbits.
The provision of mirrors has the potential to be a cost-
effective, practical method of refining current housing for
singly housed females in accordance with the ‘Three R’s’
(Russell & Burch 1959). However, further investigation and
validation with physiological parameters would help to
determine whether mirrors could also be used to improve
the welfare of singly housed males.
Acknowledgements
The authors would like to thank the staff at the animal
facility for all their help during the period of research.
Thanks also to Dr Toby Knowles for his advice on the statis-
tics. This project was carried out in 2004 as part of the MSc
in Applied Animal Behaviour & Animal Welfare at the
University of Edinburgh, UK.
References
Anderson JR 1994 The monkey in the mirror: a strange conspe-
cific. In: Parker ST, Mitchell RW and Boccia ML (eds) Self
Awareness in Animals and Humans pp 315-329. Cambridge
University Press: Cambridge, UK
Berthelsen H and Hansen LT 1999 The effect of hay on the
behaviour of caged rabbits (Oryctolagus cuniculus). Animal Welfare
8: 149-157
Chu LR, Garner JP and Mench JA 2004 A behavioural com-
parison of New Zealand White rabbits (Oryctolagus cuniculus)
housed individually or in pairs in conventional laboratory cages.
Applied Animal Behaviour Science 85: 121-139
Dunn AJ, Guild AL, Kramarcy NR and Ware MD 1981
Benzodiazepines decrease grooming in response to novelty but
not ACTH or β-endorphin. Pharmacology, Biochemistry and
Behavior 15: 605-608
Gerson P 2000 The modification of ‘traditional’ caging for exper-
imental laboratory rabbits and assessment by behavioural study.
Animal Technology 51: 13-36
Guild AL and Dunn AJ 1982 Dopamine involvement in ACTH-
induced grooming behavior. Pharmacology, Biochemistry and
Behavior 17: 31-36
Gunn D 1994 Evaluation on welfare in the husbandry of laboratory
rabbits. PhD Thesis, University of Birmingham, UK
Gunn D and Morton DB 1995 Inventory of the behaviour of
New Zealand White rabbits in laboratory cages. Applied Animal
Behaviour Science 45: 277-292
Hansen LT and Berthelsen H 2000 The effect of environmen-
tal enrichment on the behaviour of caged rabbits (Oryctolagus
cuniculus). Applied Animal Behaviour Science 68: 163-178
Held SDE, Turner RJ and Wooton RJ 1995 Choices of labo-
ratory rabbits for individual or group-housing. Applied Animal
Behaviour Science 46: 81-91
Home Office 2009 Statistics of Scientific Procedures on Living
Animals, 2009. Her Majesty’s Stationery Office: London, UK
Huls WL, Brooks DL and Beanknudsen D 1991 Response of
adult New Zealand white rabbits to enrichment objects and
paired housing. Laboratory Animal Science 41: 609-612
Hurst JL, Barnard CJ, Tolladay U, Nevison CM and West
CD 1997 Housing and welfare in laboratory rats: welfare implica-
tions of isolation and social contact among caged males. Animal
Welfare 6: 329-347
Hurst JL, Barnard CJ, Tolladay U, Nevison CM and West
CD 1998 Housing and welfare in laboratory rats: welfare implica-
tions of isolation and social contact among caged females. Animal
Welfare 7: 121-136
Hurst JL, Barnard CJ, Tolladay U, Nevison CM and West
CD 1999 Housing and welfare in laboratory rats: effects of cage
stocking density and behavioural predictors of welfare. Animal
Behaviour 58: 563-586
Jones SE and Phillips CJC 2005 The effect of mirrors on the
welfare of caged rabbits. Animal Welfare 14: 195-202
Kalagassy EB, Carbone LG and Houpt KA 1999 Effect of
castration on rabbits housed in littermate pairs. Journal of Applied
Animal Welfare Science 2: 111-121
Lehmann M 1991 Social behaviour in young domestic rabbits under
semi-natural conditions. Applied Animal Behaviour Science 32: 269-292
Leslie TK, Dalton L and Phillips CJC 2004 Preference of domes-
tic rabbits for grass or coarse mix feeds. Animal Welfare 13: 57-62
Lidfors L 1997 Behavioural effects of environmental enrichment for
individually caged rabbits. Applied Animal Behaviour Science 52: 157-169
Mason, GJ 1991 Stereotypies and suffering. Behavioural Processes
25: 103-115
McAfee LM, Mills DS and Cooper JJ 2002 The use of mirrors
for the control of stereotypic weaving behaviour in the stabled
horse. Applied Animal Behaviour Science 78: 159-173
Mills DS and Davenport K 2002 The effect of a neighbouring
conspecific versus the use of a mirror for the control of stereotypic
weaving behaviour in the stabled horse. Animal Science 74: 95-101
Montevecchi WA and Noel PE 1978 Temporal effects of mir-
ror-image stimulation on pecking and peeping in isolate, pair-reared
and group-reared domestic chicks. Behavioural Biology 23: 531-535
Morton DB, Jennings M, Batchelor GR, Bell D, Birke L,
Davies K, Eveleigh JR, Gunn D, Heath M, Howard B,
Koder P, Phillips J, Poole T, Sainsbury AW, Sales GD,
Smith DJA, Stauffacher M and Turner RJ 1993 Refinements
in rabbit husbandry: Second report of the BVAAWF/FRAME/
RSPCA/UFAW Joint working group on refinement. Laboratory
Animals 27: 301-329
Mykytowycz R 1958 Social behaviour of an experimental colony
of wild rabbits (Oryctolagus cuniculus L) 1. Establishment of the
colony. CSIRO Wildlife Research 3: 7-25
Parker GA 1974 Assessment strategy and the evolution of fight-
ing behaviour. Journal of Theoretical Biology 47: 223-243
Piller CA, Stookey JM and Watts JM 1999 Effects of mirror-
image exposure on heart rate and movement of isolated heifers.
Applied Animal Behaviour Science 63: 93-102
Podberscek AL, Blackshaw JK and Beattie AW 1991 The
behaviour of group penned and individually cages laboratory rab-
bits. Applied Animal Behaviour Science 28
: 353-363
Russell W and Burch R 1959 The Principles of Humane
Experimental Techniques. Methuen: London, UK
Seaman SC 2002 Laboratory rabbit welfare: an investigation of the social
and physical environment. PhD Thesis, University of Edinburgh, UK
Seaman SC, Waran NK, Mason G and D’Eath RB 2008
Animal economics: assessing the motivation of female laboratory
rabbits to reach a platform, social contact and food. Animal
Behaviour 75: 31-42
© 2010 Universities Federation for Animal Welfare

The effect of mirrors on rabbit behaviour 471
Sherwin CM 2004 Mirrors as potential environmental enrich-
ment for individually housed laboratory mice. Applied Animal
Behaviour Science 87: 95-103
Stodart E and Myers K 1964 A comparison of behaviour,
reproduction, and mortality of wild and domestic rabbits in con-
fined populations. CSIRO Wildlife Research 9: 144-159
Toubas PL, Abla KA, Cao W, Logan LG and Seale TW 1990
Latency to enter a mirrored chamber – a novel behavioural assay for
anxiolytic agents. Pharmacology, Biochemistry and Behavior 35: 121-126
Vastrade FM 1986 The social behaviour of free-ranging
domestic rabbits (Oryctolagus cuniculus). Applied Animal Behaviour
Science 16: 165-177
Animal Welfare 2010, 19: 461-471