CSIRO PUBLISHING

www.publish.csiro.au/journals/ajea Australian Journal of Experimental Agriculture, 2006, 46, 777–782

Effects of dietary fibre and feeding frequency on wool biting and aggressive
behaviours in housed Merino sheep

S. VasseurA,B , D. R. PaullA , S. J. AtkinsonA , I. G. ColditzA and A. D. FisherA,C

A CSIRO Livestock Industries, FD McMaster Laboratory Chiswick, Locked Bag 1, Armidale, NSW 2350, Australia.
B INA P-G, Institut National Agronomique de Paris-Grignon, 16 rue Claude Bernard, 75 231 Paris Cédex 05, France.
C Corresponding author. Email: andrew.fisher@csiro.au

Abstract. Wool biting is a behaviour that can develop in housed sheep, in which sheep start to bite and eat the
wool of others. The aim of this study was to determine whether (i) supplementing the diet of housed sheep with
fibre and (ii) increasing feeding frequency would help to reduce wool biting, aggressive behaviours and wool
damage. In a 2 × 2 factorial experiment, 40 Merino wethers were either fed with lucerne-based pellets only, or
with pellets supplemented with barley straw. They received their pellets either on a low feeding frequency basis
(once a day Monday to Friday mornings, double ration on Friday afternoon, nothing Saturdays and Sundays), or
on a high feeding frequency basis (twice a day, every day). The sheep were housed in 4 treatment pens, each with
10 animals. Wool biting and aggressive behaviours were recorded through direct observation and the sheep were
scored for wool damage twice a week during the 15-week study. The provision of fibre had a significant effect on
reducing wool biting (P<0.001) and wool damage score (P<0.001). There was no consistent effect of feeding
frequency on wool biting or wool damage, and no fibre × feeding frequency interactions. Whereas wool biting in
general increased with time during the study (P<0.001), levels of aggressive behaviour showed no consistent time
trend, and there were no effects of fibre or feeding frequency treatments. It is concluded that wool biting is largely
a redirected behaviour in concentrate-fed housed sheep deprived of adequate levels of activity or oral stimulus, and
that the provision of roughage will reduce the development of wool biting and improve animal welfare in housed
experimental sheep.

Introduction evidence from animal house managers suggests that such

Housed sheep sometimes develop a behaviour known as
‘wool biting’, in which 1 or more sheep pick and bite at
the wool of other individuals within the group (Lynch et al.
1992; Sambraus 1985). Wool biting is unpredictable and can
develop within a few days of housing or after a longer period.
The behaviour often starts with a few animals biting a small
number of targeted sheep. Once started, however, it appears
to spread, and sheep that were not initially involved in wool
biting can adopt the behaviour. Anecdotal evidence would
suggest that, once established, wool biting appears relatively
intractable, and isolating the most active wool biters from the
pen may relieve the practice for only a short period.
Wool biting has not been studied in detail and its
causative factors remain a source of speculation. A lack of
environmental stimulation and dietary factors are thought to
be possible contributing factors (Lynch et al. 1992). Boredom
could be 1 cause for wool biting, and some techniques of
environmental enrichment have been developed to reduce
it, including hanging chains from the ceiling, introducing
basketballs, plastic bottles or chewing bars in the pens, or
playing music. Effects of these environmental stimuli on
wool biting have not been formally recorded, but anecdotal
strategies may only delay the appearance of the phenomenon
(Bradley, personal communication).
Another theory is that wool biting could result from
an inadequate diet, either in terms of quality or quantity
(Sambraus 1985). The effects of a reduced volume of food
intake on increased oral stereotypic behaviours have been
demonstrated in other species, especially pigs (Bergeron et al.
2000). This may be partly driven by a need for gut fill, as
evidenced by a greater water intake in feed-restricted pigs
(Barber et al. 1991). The use of concentrated feed in housed
sheep may have the same effect on the sensation of gut fill
and lead to the appearance of abnormal oral behaviours such
as wool biting.
The use of concentrated feed for housed sheep may also
be responsible for a lack of oral stimulation and rumination.
The time spent eating such feed is short and this rapid
consumption might not provide enough oral stimulation to
the sheep. At pasture, grazing requires extensive use of the
tongue and mouth, but in captivity the opportunities to do
so and the substrates upon which to develop this activity
are often limited. Any oral frustration that may result could
contribute to the development of redirected oral behaviours

© CSIRO 2006 10.1071/EA05320 0816-1089/07/060777

778 Australian Journal of Experimental Agriculture
such as licking water troughs, chewing pen divisions and
wool biting. In addition, ruminating is a major component
of the sheep ethogram under most conditions, and the use
of concentrated feed with minimal roughage reduces this
behaviour. Campion and Leek (1996) showed that the removal
of hay from the diet of hay and concentrate-fed sheep reduced
true rumination to zero after a period of 8 days.
The rapid consumption of food and the absence of
rumination also result in large amounts of free time for the
sheep. Under Australian conditions, sheep graze for 6 to 9 h
per day (Arnold and Dudzinski 1978; Dudzinski and Arnold
1979). A study recording rumination times of sheep on a
variety of grazed forages found that the animals ruminated
for up to 6 h per day (Lofgreen et al. 1957). The reduction of
rumination time to zero and feeding time to 1 or 2 h potentially
liberates 13 to 14 h of time for an animal which, when housed
in conventional pens, has a relatively restricted array of
environmental interactions. Long periods without feed may
also increase boredom. In horses, a study by Cooper et al.
(2005) suggested that dividing stabled animals’ concentrate
ration into a number of smaller meals may be an effective
means of reducing oral stereotypies. Therefore, the increase
in free time and the reduction in oral stimulation which
result from reduced fibre in the diet could well lead to the
development of abnormal oral behaviours in sheep and be a
cause for wool biting.
Accordingly, this experiment examines the effect of
increased dietary fibre on wool biting, in order to determine
whether using more fibre in the diet can prevent or reduce
wool biting, and thereby enhance animal welfare in housed
experimental sheep. The effect of feeding frequency was
also examined, as increasing feeding frequency could help
to reduce boredom in housed sheep and thereby reduce
wool biting.
Materials and methods
Animals and treatments
The protocol and conduct of the study were approved by the institutional
Animal Ethics Committee, under the NSW Animal Research Act 1985.
The study utilised 40, 5-month-old Merino wethers from the same
flock. The sheep were brought from the paddocks to the animal house
1 week before the beginning of the 15-week study, and were individually
identified by dye-based numbers branded on their flanks. After a week
of acclimation to the animal house environment, the sheep were blocked
by weight, and within block, allocated at random to their treatments.
The experiment utilised a 2 × 2 factorial plan, with 2 levels of
roughage provision and 2 feeding frequencies. The sheep were either
fed a diet of lucerne-based concentrate pellets only (NS), or the same
pellets supplemented with barley straw (S). The animals received their
concentrate pellets on a low frequency basis (LFF) or on a high frequency
basis (HFF).
Sheep fed the NS diet received 700 g of pellets a day, whereas sheep
on the S diet were offered 650 g of pellets plus 300 g of barley straw
per day (Table 1). The pellets were made to a formulation designed to
supply all the nutrient requirements of maintenance and growing sheep
(11.6 MJ ME, 21.6% crude protein, and sufficient levels of calcium,
phosphorus, magnesium and sulphur). All feeding was on a group basis.
The feeding levels were determined from existing analyses of the feeds,
and were calculated to provide the same levels of energy and protein.
S. Vasseur et al.
Table 1. Composition of the two dietary treatments
Sheep were fed either lucerne-based concentrate pellets only (NS),
or the same pellets supplemented with barley straw (S)
Components (per animal) NS S
Lucerne pellets (g) 700 650
Barley straw (g) — 300
Dry matter (g) 632 717
Crude protein (g) 137 131
Metabolisable energy (MJ) 7.3 7.6
Samples of the feeds as offered during the experiment were collected and
analysed, and these results are presented in Table 1. All the pellets were
consumed each day, however there was approximately 50% wastage of
the straw by the sheep.
The feeding frequency treatments only concerned the pellets. The
LFF groups were fed once a day (0800 hours) during week days and were
given a double ration on Friday afternoon (1600 hours) for the weekend.
Sheep on the HFF treatment were fed twice a day (0800 and 1600 hours)
with half the total daily amount offered at each occasion, 7 days
per week.
Each treatment group of 10 sheep was housed in a 2.5 by 5 m pen
with a steel mesh floor and containing 1 drinking trough, 2 feed troughs
and 1 hay rack. The hay rack was used to deliver the barley straw
(treatment S) or remained empty (treatment NS). The steel mesh on the
hay rack was 5 by 5 cm. Shade cloth was placed on each panel between
pens to visually separate sheep from adjoining treatment pens.
On week 2 of the study, a break in the wool became apparent for
one sheep in the control group (NS and LFF). This animal was replaced
by another animal from the same original group that was housed in
the animal house for 1 week and then transferred to the treatment pen
on week 4. The replacement animal had received the same preparatory
feeding regimen prior to entering the experiment.
Measurements
The sheep were weighed at the beginning and at the end of the experiment
(week 15). Animal behaviour was recorded by direct observation. Each
pen was observed for 1 hour twice a week, once in the morning (0900
to 1000 hours), and once in the afternoon (1500 to 1600 hours).
The following behaviours were recorded: wool biting, aggression,
horn biting, ear licking and mounting. For each of these events, the
agonist and recipient sheep were identified and recorded. Wool biting
was defined as a sheep closing its mouth on another animal’s wool
and biting or tugging. Aggressive interactions included butts, agonistic
jumping and kicks. All the observations were conducted by the same
observer, except for weeks 9, 12 and 13, when a second, trained observer
was utilised.
To quantify the impact of wool biting, a measurement scale was used
to assess the state of the wool and its damage. The scale consisted of
6 levels from 0 to 5:
0: no sign of wool biting on any part of the sheep and no apparent
damage of the wool;
1: wool became fluffed — appearance of small tufts of wool protruding
above the surface of the fleece — score 1 from the first tuft;
2: most parts of the fleece contain protruding tufts — high density of
these tufts;
3: high density of tufts + appearance of a brighter area due to confluence
of protruding tufts or thinning of the fleece leading to greater
visibility of clean wool below the fleece tip (often on the rump or
legs) — skin was not visible and the bright area was less than 50%
of body surface;
4: spreading of the brighter area to more than 50% of body surface OR
skin visible in this brighter area;
Wool biting in housed sheep Australian Journal of Experimental Agriculture 779

5: spreading of the white area to 80–100% of body surface OR skin
visible on more than 10–20% of body surface. At this stage, decisions
were made to avoid any further damage (removal of the animal,
covering with a sheep coat, etc.).
The scale was objectively recorded using a series of photographs
corresponding to each level. All 40 sheep were scored twice a week,
throughout the study. All the scoring was done by the same observer,
except for weeks 12 and 13, when a trained, second scorer was used.
Statistical analyses
The statistical analyses were performed using Genstat, 7th Edition
(Payne et al. 2003), with individual sheep as the experimental unit.
Behavioural observation data were analysed by split–split plot ANOVA,
with the main plot including the effects of fibre content, feeding
frequency and their interaction. The 2 subplots included: (i) the
effect of time of day (morning or afternoon) and its interaction with
fibre content, feeding frequency and fibre content × feeding frequency;
and (ii) the effect of week (time repetitions) and its interaction with
fibre content, feeding frequency and fibre content × feeding frequency,
as well as the interaction between time of day and week and the
interaction of this term with fibre content, feeding frequency and fibre
content × feeding frequency.
Logarithmic transformations of the behavioural data were used.
Because there was insufficient frequency of occurrences of horn biting,
ear biting and mounting to permit robust analysis, these variables were
not analysed. Thus, the variables analysed were: log(number of wool
biting events as attacker + 1), log(number of wool biting events as
victim + 1), log(number of aggressions as attacker + 1) and log(number
of aggressions as victim + 1). Unless otherwise stated, all behavioural
results are presented as the log values.
Wool damage score data were analysed by split plot ANOVA, with
the main plot including the effects of fibre content, feeding frequency
and their interaction, and the subplot including the effects of time
and its interaction with fibre content, feeding frequency and fibre
content × feeding frequency. The average daily gain for the duration of
the study was calculated for each animal, and analysed by ANOVA for a
2 × 2 factorial plan. Within fibre treatments, correlation analyses were
performed between average daily gain and mean wool damage score for
each animal.
the mean number of wool biting events per hour for each
sheep over the duration of the study, both as attacker and
victim, and indicates that some sheep were more commonly
filling 1 role compared with the other. All subsequent results
are presented based on the analyses of wool biting and
aggressive events recorded for the attacking sheep. The
time trends for wool biting events for the main treatment
effects are presented in Figure 3. There was a treatment
effect of fibre provision (P<0.001), but not for feeding
frequency (P>0.05), however, there were treatment × time
effects for both fibre provision (P<0.001) and feeding
frequency (P<0.001). The key finding was that as the study
progressed, the S sheep displayed significantly less wool
biting than the NS sheep. The interaction of fibre provision
and feeding frequency was not significant (P = 0.34). The
effect of time of day was significant (P<0.01), with more
wool biting events recorded in the afternoon.
Unlike wool biting, aggressive behaviours were present
throughout the experiment (Fig. 4), and remained relatively
As aggressor As victim
NS
LFF
S
LFF

Results
Wool biting was virtually absent at the start of the experiment,
but increased with time through to Week 15, with more
sheep undertaking the behaviour (Fig. 1). Figure 2 presents NS
HFF

10
No. of wool biters

6 S
HFF
4 NS, HFF
NS, LFF
S, HFF
2
S, LFF 30 20 10 0 10 20 30

0 Mean no. of events/hour

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Week Fig. 2. Mean number of wool biting events per hour for sheep (n = 40)
either fed pellets only at low feeding frequency (NS, LFF), supplemented
Fig. 1. Cumulative number of sheep observed wool biting within each with straw and fed pellets at low feeding frequency (S, LFF), fed pellets
treatment during the 15-week study. The sheep were either fed pellets only at a higher feeding frequency (NS, HFF), or supplemented with
only at low feeding frequency (NS, LFF), supplemented with straw straw and fed pellets at a higher feeding frequency (S, HFF). Open
and fed pellets at low feeding frequency (S, LFF), fed pellets only at a bars indicate individual sheep and solid bars represent the treatment
higher feeding frequency (NS, HFF), or supplemented with straw and means. Data are untransformed values and represent the mean over the
fed pellets at a higher feeding frequency (S, HFF). 15 weeks.
780 Australian Journal of Experimental Agriculture S. Vasseur et al.

** *** **
0.7
(a) *** ***
3.0
0.6 (a)

2.5
0.5 Pooled s.e.

2.0
Pooled s.e.
0.4
NS
0.3 S
1.5

0.2
1.0
Log(wool biting events/h)

0.1
0.5 NS

Wool damage score

0 S
5 10 15
0
5 10 15
0.7 (b)
* * *** ** 3.0 (b)
0.6

0.5 Pooled s.e. 2.5

0.4 2.0 Pooled s.e.

0.3
1.5
0.2
LFF 1.0
0.1
HFF
LFF
0.5
HFF
0
5 10 15
Week 0
5 10 15
Fig. 3. Wool biting events per sheep during the 15-week study period
Week
for the main effects of (a) fibre provision (NS, low fibre; S, fibre
supplement) and (b) feeding frequency (LFF, low feeding frequency; Fig. 5. Wool damage score during the 15-week study period for the
HFF, high feeding frequency). *P<0.05; **P<0.01; ***P<0.001. main effects of (a) fibre provision (NS, low fibre; S, fibre supplement)
and (b) feeding frequency (LFF, low feeding frequency; HFF, high
feeding frequency). **P<0.01; ***P<0.001.

consistent throughout the study period, although the effect

0.7 of time was significant (P = 0.002). There were no effects
of fibre provision, feeding frequency, or their interaction
Log(aggressive events/h)

0.6 (P>0.05). There was an effect of fibre provision × time

of day, whereby sheep that received fibre displayed less
0.5
aggression during the afternoon [0.22 and 0.19 log(events/h)
for NS and S, respectively, s.e. = 0.029; P = 0.013], but more
0.4
aggression during the morning [0.16 and 0.22 log(events/h)
0.3 for NS and S, respectively].
The main effects of treatment on wool damage are
0.2 presented in Figure 5. The wool damage score increased
(P<0.001) with time during the experiment. There was
0.1 an effect of fibre provision (P<0.001) and a fibre
0 5 10 15
provision × time interaction (P<0.001). Sheep on the S
Week
treatment displayed a slower increase in wool damage than
Fig. 4. Aggressive events per sheep during the 15-week study period. NS sheep during the course of the study. There were no effects
Wool biting in housed sheep
(P>0.05) of feeding frequency, feeding frequency × time,
or fibre provision × feeding frequency on the wool
damage score.
The average daily gain was 70.2 g for S sheep and 42.4 g
for NS sheep (pooled s.e. = 7.72; P = 0.015). There was no
difference in average daily gain between LFF and HFF sheep
(54.8 and 57.8 g, respectively; pooled s.e. = 7.72; P = 0.78).
The fibre × feeding frequency interaction was not significant
(P = 0.91). Within fibre treatments, there were no significant
correlations between average daily gain and mean wool
damage score for either S sheep (r = 0.02; P = 0.93) or NS
sheep (r = –0.16; P = 0.51).
Discussion
The main findings of this study were that wool biting
developed and increased in sheep from the fifth week of
housing, and that the provision of fibre significantly delayed
and reduced the occurrence of wool biting. Furthermore,
the provision of increased dietary fibre provided protection
against wool damage.
Although previous theories on wool biting have included
nutritional deficiencies as a possible cause (Sambraus 1985),
the results of the present study indicate that the wool
biting was more related to either a need for fibre and oral
stimulus, or a need for activity that was satisfied by oral
stimulus. The diets were designed to be nutritionally suitable
for growing sheep, and there was less than 5% difference
between the two dietary treatments in the major components
of energy and protein. Furthermore, all treatment groups
gained weight during the study. A study by Yurtman et al.
(2002) showed that increasing the daily protein intake in
lambs under energy restricted conditions may have additive
effects on the development of stereotypic oral behaviours,
such as licking and chewing inanimate objects. In the present
study, the wool biting behaviour developed under conditions
of adequate energy provision, although we did not examine
the effect of differences in protein content. In pigs, the
provision of straw was shown to reduce the incidence of
anomalous oral behaviours by sows on otherwise isoenergetic
diets (Whittaker et al. 1998). In the present study, the large
difference in daily gain between the sheep on the NS and
S treatments is presumably due to the nuisance and stress
caused by wool biting. Given that most sheep were both
aggressors and victims in wool biting, and that there was
no correlation between average daily gain and wool damage
score within fibre treatments, it is likely that it was the
performance and receiving of wool biting that was disruptive
to the animals, rather than any physical harm caused.
The increase in wool biting during the course of the study
was not only due to a fixed number of sheep increasing their
habit, but also to new sheep undertaking the behaviour. This
suggests that either the sheep may learn from each other by
visual observation, or that there is individual variation in the
critical point from which the wool biting behaviour starts.
Australian Journal of Experimental Agriculture 781
The effect of feeding frequency on wool biting was less
obvious. There were effects of feeding frequency on wool
biting behaviour at some time points, but this was not
consistent and did not result in differences in wool damage
scores. Contrary to our hypothesis, the effects of feeding
frequency, where present, suggested that higher feeding
frequency was associated with more wool biting behaviour.
This finding may result from animals being able to consume
all their allocated feed at once when it is presented in smaller
portions, providing them with less oral stimulation than when
they are presented with feed in larger quantities, and return to
it several times during the day. However, given that any effects
of feeding frequency on wool biting were not consistently
present, it is not possible to draw robust conclusions.
From the results of this study, the 2 behaviours of wool
biting and aggression are different and the motivations
resulting in their development are not the same. Incidents
of aggressive behaviour were present throughout the study at
a consistent level, whereas wool biting took time to develop
and then increased. Furthermore, wool biting was not the
sole preserve of 1 or 2 dominant animals in a pen, with
all but 2 sheep acting as both initiators and victims of
wool biting.
Rather than a stereotypy, which is defined as a behaviour
that is repetitive, invariant and has no obvious function
(Odberg 1978), wool biting fits more closely with the concept
of a redirected behaviour. The results of this study would
support the argument that wool biting is at least partially
initiated by the absence of a natural substrate for grazing and
the non-satisfaction of the need for stimulation of sheep’s oral
organs by eating and ruminating.
With regard to its efficacy in preventing wool biting, we
did not test the effect of the type or source of fibre, but
our previous experience with housed sheep suggests that
generally coarser fibre was better. For instance, when chaff
had been fed as a supplement in the animal house before
this experiment, wool biting still occurred. We also would
recommend that fibre be given in a way which stimulates
the need for sheep to use their oral organs and to browse.
Increasing the work sheep have to undertake to extract the
straw may be a good way of achieving this, and requires
the sheep to be browsing for longer without increasing the
amount of straw provided. This may also help to counteract
any contribution of boredom to wool biting.
In conclusion, the provision of fibre in the form of straw
within 5 by 5 cm metal mesh hay racks had a beneficial
effect in delaying and reducing wool biting in housed
concentrate-fed sheep. Additional research is required to
identify whether the contribution of the fibre to altering
the time budgets of the sheep was part of its beneficial
effect, through reducing time available for redirection of oral
behaviour to wool biting. Similarly, it would be useful to
determine whether the provision of fibre can reduce wool
biting once it has commenced.
782 Australian Journal of Experimental Agriculture
Acknowledgments
We thank Christine Leger, Troy Kalinowski, Peter Bradley,
Eliane Duflou, Jim Lea, Mandy Choice, Kerri Tyrrell and
Sandra Eady for their assistance in the conduct of the study, as
well as Laurie Piper and Brad Hine for their help in statistical
analyses.
References
Arnold GW, Dudzinski ML (1978) ‘Ethology of free-ranging domestic
animals.’(Elsevier: Amsterdam)
Barber J, Brooks PH, Carpenter JL (1991) The effect of four levels of
food on the water intake and water to food ratio of growing pigs.
Animal Production 52, 602.
Bergeron R, Bolduc J, Ramonet Y, Meunier-Salaun MC, Robert S (2000)
Feeding motivation and stereotypies in pregnant sows fed increasing
levels of fibre and/or food. Applied Animal Behaviour Science 70,
27–40. doi: 10.1016/S0168-1591(00)00142-8
Campion DP, Leek BF (1996) Mechanical stimulation of rumination in
sheep by the intraruminal addition of inert fibre particles. Animal
Science 62, 71–77.
Cooper JJ, Mcall N, Johnson S, Davidson HPB (2005) The short-term
effects of increasing meal frequency on stereotypic behaviour of
stabled horses. Applied Animal Behaviour Science 90, 351–364.
doi: 10.1016/j.applanim.2004.08.005
Dudzinski ML, Arnold GW (1979) Factors influencing the grazing
behaviour of sheep in a Mediterranean climate. Applied Animal
Ethology 5, 125–144. doi: 10.1016/0304-3762(79)90084-1
http://www.publish.csiro.au/journals/ajea
S. Vasseur et al.
Lofgreen GP, Meyer JH, Hull JL (1957) Behavior patterns of sheep and
cattle being fed pasture or soilage. Journal of Animal Science 16,
773–780.
Lynch JJ, Hinch GN, Adams DB (1992) ‘The Behaviour of Sheep:
Biological Principles and Implications for Production.’ (CAB
International: Wallingford)
Odberg FO (1978) Abnormal behaviours: stereotypies. In ‘Proceedings
of the First World Congress on Ethology Applied to Zootechnics’.
(Ed. J Garsi) p. 475. (Industrias Graficas: Espana)
Payne R, Murray D, Harding S, Baird D, Soutar D, Lane P (2003)
‘GenStat® for WindowsTM (7th Edition) Introduction.’ (Lawes
Agricultural Trust: Rothamsted, UK)
Sambraus HH (1985) Mouth-based anomalous syndromes. In ‘Ethology
of farm animals’. (Ed. AF Fraser) pp. 391–422. (Elsevier:
Amsterdam).
Whittaker X, Spoolder HAM, Edwards SA, Lawrence AB, Corning S
(1998) The influence of dietary fibre and the provision of straw
on the development of stereotypic behaviour in food restricted
pregnant sows. Applied Animal Behaviour Science 61, 89–102.
doi: 10.1016/S0168-1591(98)00183-X
Yurtman IY, Savas T, Karaagac F, Coskuntuna L (2002) Effects of
daily protein intake levels on the oral stereotypic behaviours in
energy restricted lambs. Applied Animal Behaviour Science 77,
77–88.
Received 17 November 2005, accepted 2 April 2006