Neiker-Tecnalia, Department of Animal Production, 01080 Vitoria-Gasteiz, SpainNavarra Public University (UPNA), Campus de Arrosadia, 31006 Pamplona, SpainIKERBASQUE, Basque Foundation for Science, 48011 Bilbao, Spain
a r t i c l e i n f o
rticle history:ccepted 6 November 2013vailable online 14 November 2013
Provision of adequate housing is essential to assure the welfare of farm animals. One rele-vant aspect is space allowance, due to potential consequences on the behaviour and welfareof animals. The objective of this study was to determine the effect of space allowanceon the behaviour and potential welfare implications of dairy ewes during gestation. Withthis purpose, 54 pregnant ewes were randomly allocated to groups of 6 ewes with spaceallowances of 1, 2, or 3 m2/ewe (three replicates per treatment). Behaviour was observedbetween gestation weeks 9 and 19. Data collection included focal observations (4 gestationperiods of 2 consecutive weeks, 3 days/week) and scan samplings (twice/week during 11gestation weeks). The effects of space allowance, point of gestation, and their interactionon the percentage of time, frequency (focal observations) and percentage of occurrence(scan samplings) of behaviours were determined with a mixed model ANOVA, with pointof gestation as repeated measure. In general, less activity as indicated by reduced move-ment (P < 0.05) and higher percentages of time at the feeder (P < 0.05) were observed at1 m2/ewe as compared to 2 and 3 m2/ewe. Occurrences of negative, and especially posi-tive social interactions, were also higher at 1 m2/ewe (P < 0.05), perhaps as result of thehigher chances to find another individual in the path of movement rather than of increasedsocial conflict. The initial period of study was characterized by a higher frequency of posi-
tive social interactions (P < 0.001) and explorative behaviours (P < 0.005), likely due to thenovelty effect. On the contrary, the decline in activity, as indicated by a reduction in thepercentage of time standing (P < 0.05) and increased resting (P < 0.05), was evident as ges-tation progressed. A clear effect of space allowance was detected, although it was mostlyrestricted to moving and eating behaviours.
. Introduction
Provision of adequate housing is essential to assure theelfare of production animals, but also to maintain the
efficiency and assure the sustainability of any productionsystem. Although many factors must be considered whenevaluating the adequacy of housing conditions, one that isof primary relevance is the availability of space, as spacelimitations can have major impact in terms of welfare andperformance of the animals.
Space allowance is generally defined as the averagearea offered per animal (Petherick, 2007; Petherick andPhillips, 2009), and it is considered a defining featureof all animal production systems due to its economic
implications. Provision of large enclosures for animal hous-ing implies larger land requirements, higher constructionand maintenance costs, and possibly manpower require-ments. Therefore, commercial animal production systemsusually limit, to some degree, space allowance.
However, it is well known that space limitations canhave negative consequences for the welfare of productionanimals (Fraser and Broom, 1997; Estevez et al., 2007) aswell as on performance. For example, space limitationshave been associated with increased behavioural prob-lems and aggression, and a reduction of the performancein pigs (Averós et al., 2010a,b, 2012; Gonyou et al., 2006;Turner et al., 2000), cattle (Ingvartsen and Andersen, 1993;Krawczel et al., 2012; Wechsler, 2011), and poultry (Bessei,2006; Cornetto et al., 2002; Estevez, 2007; Lay Jr. et al.,2011).
Dairy sheep production systems have traditionally beencharacterized by the utilization of extensive grazing areas(Boyazoglu and Morand-Fehr, 2001), where spatial restric-tion may be considered small, or virtually non-existent.Nonetheless, the progressive transition from traditional tomodern production systems, due in part to an attempt toprotect animals from adverse weather, has led to the use ofmore intensive husbandry procedures and higher produc-tive efficiency in sheep (Shrestha, 2011).
The transition to more intensive systems can imposestrong limitations on the space available to dairy sheep.Previous studies have already shown that a reduction inlying area altered the resting patterns of pregnant ewes(Bøe et al., 2006). Reduced available space also impaired theperformance of lambs (Gonyou et al., 1985; Horton et al.,1991) and altered ewesı́ spatial distribution and behaviourdespite the higher space allowances as compared to moreintensive housing conditions (Sibbald et al., 2000). Onthe other hand, increasing space allowance from 1.5 to3 m2/ewe has been associated with an increased proportionof animals walking and a reduced incidence of aggressions(Caroprese et al., 2009). In addition an improvement inhealth condition, including enhanced immune response,and a quantitative and qualitative improvement in milkproduction in dairy ewes has been observed when spaceallowance increased from 1 to 2 m2/ewe (Sevi et al.,1999).
The effects of spatial restriction for dairy sheep mayalso have consequences for their reproductive efficiency,and the quality of their offspring. This is suggested by thefinding that management-related stressors (such as ther-mal, nutritional, hormonal and other) may compromiseewes’ reproductive efficiency (Dobson et al., 2012), andultimately, the survival and performance of their offspring(Shelton, 1964; Brown et al., 1977). It is therefore clear thata successful pregnancy in sheep will be, at least partially,determined by the ability of ewes to cope with any poten-tial stress source affecting them throughout gestation,possibly including space limitation. Literature regardingchanges in the behaviour of ewes across gestation is scarce,but Caroprese et al. (2009) found that the behaviour of ewes
changed during a 1-year experiment. It may, therefore, beexpected that space allowance would alter the behaviouralpatterns of pregnant ewes across gestation. However,the impact of space allowance provided under intensive
our Science 150 (2014) 17– 26
housing on the activity and social behaviour of gestatingewes and potential impacts on welfare are still unknown.Investigation of the effects of space allowance would,therefore, provide information that would contribute to theoptimization of dairy sheep production systems in terms ofanimal welfare.
This work is framed within a larger study designedto determine the effects of space allowance over thebehaviour, welfare and reproductive performance of ges-tating dairy ewes, and the potential impact that such spacelimitation may have over the quality of their offspring. Thisstudy aimed to determine the effects of space allowance,while maintaining constant group size, on the activityand the social behaviour of dairy ewes between weeks9 and 19 of gestation. We hypothesized that low spaceallowance (high densities) will constrain the activity andbehavioural repertoire in pregnant dairy ewes, and wouldpotentially increase the occurrence of negative social inter-actions as result of increased competition for space. Theseeffects might, consequently, have a negative impact ontheir welfare than could be exacerbated through gesta-tion.
2. Materials and methods
The experiment was approved by the NEIKER-Tecnalia Animal Experimentation Committee (ReferenceAFA 2011 02), and was carried out according to theEuropean Directive 86/609/ECC regarding the protectionof animals used for experimental and other scientificpurposes.
2.1. Animals and pre-treatment facilities
The present study was carried out at the experimen-tal dairy sheep farm of Neiker-Tecnalia (Vitoria-Gasteiz,Spain) between August 2011 and January 2012. One hun-dred and thirty five primiparous and multiparous Latxaewes (Ovis aries), one to eight years old, were artifi-cially inseminated (AI) at the end of August. Latxa breedis traditionally used for dairy purposes, and is generallymanaged under semi-extensive conditions. Ewes were ini-tially housed as a single flock in a barn (14 m × 32 m)with solid walls and windows allowing natural lighting.Artificial lighting was available for necessary manage-ment duties at any hour. Straw bedding was provided,and fresh straw was added periodically as required tomaintain adequate bedding conditions. Ewes had dailyaccess (about 20 h/day) to an adjacent outdoor pasture(5 ha) until housed in the experimental pens. Pasture wascomplemented with ad libitum access to vetch and oatsand water through automatic nipple drinkers inside thebarn.
Forty-six days after AI, pregnancy and number of foe-tuses per ewe were confirmed by ultra-sound (Ovi-scan6, BCF, Australia). Simultaneously, ewesı́ body condition
was determined using a 5-point scale (Russel et al., 1969).Among the ewes with confirmed pregnancy, fifty-four wererandomly selected for the experiment; they remained inthe flock until the start of the experimental period.
l Behavi
2
dsafttuta−oms
ttb2Gbsiahslwn
8tpw1wmpha(Fbsm
2
etDpttmtbw
X. Averós et al. / Applied Anima
.2. Experimental treatments
The experiment was carried out in the previouslyescribed barn. Fifty days after AI the selected ewes wereheared, and sixty-two days after AI, they were weighednd divided into nine groups of six individuals, balancingor age, body condition score, and number of viable foe-uses. Each group was then randomly assigned to one ofhe nine experimental enclosures where ewes were keptntil week 19 of gestation. The rest of the flock continuedhe standard management procedures. Outdoor temper-tures during the period of data collection ranged from5.0 ◦C to 21.1 ◦C, with a daily average temperature (±SE)f 6.8 ± 0.4 ◦C. Animals in each group were individuallyarked with a number on the back with purple livestock
pray (Multi-line, Ukal, France).Three space allowances 1, 2, and 3 m2/ewe, with
hree replicates per treatment, were tested. Experimen-al pens were built with PVC walls to avoid visual contactetween experimental units, and pen dimensions were.7 m × 2.25 m, 2.7 m × 4.5 m, and 3 m × 5.9 m, respectively.roup size was kept constant at 6 ewes/enclosure. Strawedding was available in all the enclosures, and freshtraw was periodically added to maintain bedding qual-ty. Feed was provided in one automatic feeding trough,nd each pen had eight independent feeding spaces withead protection available/enclosure, so that all ewes couldimultaneously access the feed, and feeding space was notimited and was identical in all treatments. Drinking water
as provided ad libitum through an automatic drinkingipple installed in each enclosure.
From the beginning of the study, at the end of week through to gestation week 15 ewes were feed silagewice/day, at 08:30 and at 15:00 (about 1.5 kg in totaler ewe per day), while from 15 to 18 weeks fescue hayas provided twice/day at the same time as before (about
.5 kg in total/ewe and day). The diet was complementedith 400–500 g of a barley and wheat mix/ewe in theorning meal, and with ad libitum access to oat hay and
eas in the afternoon meal. From week 18 to 20 fescueay was provided twice/day (about 1.5 kg in total/ewend day), complemented with about 500 g of concentrate1.101 UFL/kg; 168 g PB/kg) per ewe in the morning meal.rom week 12 to 16 pregnant ewes had free access to saltlocks (TIMAC SAS, St Malo, France) after which time thealt blocks were substituted by a cube containing vitamin-ineral corrector (INAFORM, Timac Agro, Orcoyen, Spain).
.3. Behaviour observations
Data collection started 69 days after AI, 7 days afterwes were grouped together in pens (week 9 of gesta-ion), and lasted for 11 weeks (end of week 19 of gestation).ata were collected by two observers, one of whom com-leted the focal observations and the second performedhe scan samplings. On each observation day, data collec-ion started after the morning meal, around 09:30 when
ost of ewes had already finished their meal. One ewe fromhe 1 m2/ewe treatment died during week 18 of gestation,ut observations on the remaining individuals in the groupere collected until the end of the study.
our Science 150 (2014) 17– 26 19
Focal observations were carried out during threeconsecutive days/week (Tuesday to Thursday) and twoconsecutive weeks, with a week break in between obser-vation periods, for the duration of the study. Therefore thedata collection consisted of 4 observation periods, 1–4, cor-responding to weeks 9–10, 12–13, 15–16, and 18–19 ofgestation, respectively. Two ewes per enclosure were ran-domly chosen for each data collection day, and each ofthem was observed for 10 min using the Noldus Observersoftware (v. 10.5), installed on a portable computer. Datacollection was organized in 2 rounds per day of observa-tion. In this form, all 54 ewes were observed twice duringeach period. Behaviours considered and their definitionsare included in Table 1. Behaviours were divided into states,for which both the frequency and duration of the behaviourwere collected, and events, for which, due to their shortduration, only frequency of occurrence of the behaviourwas recorded.
Scan samplings were collected twice per week for all11 weeks of observation (Wednesday and Friday). Datacollection was performed with the Chikitizer software(Sanchez and Estevez, 1998) installed on a portable com-puter. For each observation day the experimental penswere observed in a random order, starting at 9:30. Eachexperimental pen was observed for 15 min through a seriesof continuous scan samplings. During each 15 min scansampling, the identity, location (in XY coordinates), andbehaviour of all ewes within the pen were sequentially col-lected. An average of 12 scan samplings was collected per15 min observation. The ethogram used for scan samplingcomprised: drink, explore pen, eat, lick mineral resource,move, negative social interactions, positive social interac-tions, rest, self-groom, stand, and stereotypies as definedin Table 1. Data related to use of space will be presentedseparately.
2.4. Statistical analysis
The percentages of time that each ewe spent performingthe different behaviours defined in the ethogram were cal-culated from the 10 min focal observation sequences. Thesevalues were then averaged for all 6 ewes per enclosure andgestation period. The frequency (number of episodes per10 min. observation period) for both behavioural states andevents were calculated similarly.
For scan sampling data, the percentages of occurrenceof behaviours were calculated, individually per ewe andobservation day, according to the total number of scans col-lected during the 15 min observation. These values werethen averaged per enclosure, and mean enclosure valueswere finally averaged per week of observation.
The distribution of all variables was tested using theShapiro–Wilk test, and natural logarithms used when nec-essary to normalize variables. In the latter case, a constantwas added to zero values. The effects of space allowance,gestation period (for focal observations) or gestation week(for scan sampling), and their interaction was determined
for all behavioural variables by means of a mixed modelanalysis of variance using the SAS software (version 9.3, SASInstitute Inc., Cary, NC, USA), with point of gestation (periodor week) as a repeated measure. Enclosure was included in
States Drink Standing by the drinker and using itEat Standing by the feeder, with the head completely inside one of the
feeder holesExplore pen Nose interaction with the pen wall of other physical issueForage Standing, with the head down, interacting with the floor beddingLick mineralresource
Stand by the mineral resource licking it
Move Change position within the pen, either walking or runningRest Lie down on the floorSelf-groom Groom, either by self-licking or by rubbing against a physical pen
objectStand, static Stand with the four feet on the floor, doing nothing elseStereotypies Floor kicking Repeatedly kicking the floor with one of the forelegs, with no apparent
intentionalityStar-gazing Arching the head and neck over the backPen licking Repeatedly licking the walls or the physical environment (except for
the mineral resource) with no apparent intentionalityTongue playing Swinging the tongue outside of the mouth, contorting it, or rolling it
inside the mouthGrind teeth Clenching and grinding the teeth, making a particular noise
Events Negative socialinteraction
Butting Sudden, strong head contact with another ewe
Threatening Directing the forehead towards another ewe with no physical contactKicking Hitting another ewe with the forelegsDisplacing from resources Force another ewe to leave the feeder, drinker, or the mineral resourceDisplacing from resting place Force another ewe to leave the resting placePushing Press the head against another ewe to force the pass
Positive socialinteraction
Sniffing Smell another ewe without physical contact
Nosing Slightly contact another ewe with the noseGrooming Clean the wool of another ewe using the mouth
Nudging
Licking
all models as a random factor. Least square means werecomputed for all significant effects (P < 0.05).
For those behaviours where data could not be normal-ized because of their low incidence (i.e. drink and grindteeth for focal observations; drink, stereotypies, and neg-ative social interactions for scan samplings), the effects ofspace allowance and gestation period were tested usingKruskal–Wallis tests. In the case of a significant interactionbetween space allowance and point of gestation, tests ofsimple effects (Winer, 1971) were performed to determinefor which time period/week significant statistical differ-ences between space allowances existed (P < 0.05).
3. Results
In those cases where statistical tests were calculated ontransformed variables, log-transformed values (mean ± SE)and untransformed mean values are given for easiness ofinterpretation of their biological meaning.
3.1. Space allowance
3.1.1. Focal observations
A clear effect of space allowance was detected, although
it was mostly restricted to moving and eating behaviours.Ewes housed at 1 m2/ewe spent less total time moving(LSM ± SE: 1.49 ± 0.14, 1.06 ± 0.14, and 0.68 ± 0.14% for 3,
Slightly, gently push another eweLick any part of another ewe’s body
2, and 1 m2/ewe, respectively; F2,6 = 8.36, P = 0.018), andtended to initiate fewer movements than those at 3 m2/ewe(LSM ± SE: 2.51 ± 0.23, 1.89 ± 0.23, and 1.49 ± 0.23 timesfor 3, 2, and 1 m2/ewe respectively; F2,6 = 4.99; P = 0.053).On the other hand, time spent eating was higher at1 m2/ewe as compared to other space allowances (mean:5.17, 5.52, and 9.45% for 3, 2, and 1 m2/ewe respectively;F2,6 = 8.33; P = 0.019). Similar results were observed for thefrequency of eating bouts (mean: 0.27, 0.38, and 0.57 timesfor 3, 2, and 1 m2/ewe respectively; F2,6 = 8.68, P = 0.017).
The average frequency of resting bouts tended todecrease with lower space allowances (LSM ± SE:0.80 ± 0.07, 0.63 ± 0.07, and 0.53 ± 0.07 times for 3, 2,and 1 m2/ewe respectively; F2,6 = 4.35; P = 0.068), similarto the tendency found for the time spent grinding teeth(mean ± SE: 0.11 ± 0.05, 0.08 ± 0.06, and 0.00 ± 0.00%for 3, 2, and 1 m2/ewe respectively; X2, 2 df = 4.70;P = 0.096). None of the other variables were significantlyaffected by space allowance (P > 0.05). Mean values (±SE)were 43.20 ± 2.24, 37.70 ± 1.84, 6.53 ± 0.91, 6.03 ± 0.75,1.01 ± 0.22, 0.90 ± 0.17, 0.76 ± 0.10, and 0.21 ± 0.07% ofthe time spent resting, standing, mineral resource licking,foraging, performing stereotypies, pen exploring, self-
grooming, and drinking, respectively. Mean values (±SE)for the frequencies of occurrence were 4.23 ± 0.22,1.29 ± 0.11, 0.59 ± 0.06, 0.43 ± 0.06, 0.41 ± 0.07,0.36 ± 0.05, 0.21 ± 0.04, 0.11 ± 0.05, 0.11 ± 0.02, and
l Behavi
0pni
3
ewa1ab1tfPqiaF(
TRs
W
X. Averós et al. / Applied Anima
.05 ± 0.01 times for standing, foraging, self-grooming,en exploring, stereotypies, mineral resource licking,egative social interactions, teeth grinding, positive social
nteractions, and drinking respectively.
.1.2. Scan sampling observationsSpace allowance affected the mean percentage of times
wes were observed eating (F2,6 = 12.07; P = 0.008), whichas higher at 1 m2/ewe as compared to other space
llowances (mean: 9.45, 13.26, and 12.97% for 3, 2, and m2/ewe respectively). There was also an effect of spacellowance on the occurrence of social interactions, withoth positive (mean: 0.81, 0.47, and 2.44% for 3, 2, and
m2/ewe respectively; F2,6 = 7.69; P = 0.022) and nega-ive (mean ± SE: 0.21 ± 0.07, 0.30 ± 0.09, and 0.71 ± 0.15%or 3, 2, and 1 m2/ewe respectively; X2, 2 df = 7.02;
= 0.030) occurring more frequently at 1 m2/ewe. The fre-uency of resting tended to decrease with the decline
n space allowance (LSM ± SE: 48.57 ± 4.67, 37.41 ± 4.67,nd 28.51 ± 4.67% for 3, 2, and 1 m2/ewe respectively;2,6 = 4.63; P = 0.061). The frequency of both drinkingmean ± SE: 0.18 ± 0.09, 0.52 ± 0.18, and 0.37 ± 0.09% for 3,
able 2esults of the Mixed model ANOVA for the effect of gestation period (LSM ±SE)tates and events performed by gestating ewes.
Gestation period
Period 1 (weeks9–10)
Period 2 (weeks12–13)
Mean SE Mean SE
Stand % 46.57a 5.37 42.81ab 5.37n 5.06a 0.49 4.87ab 0.49
ithin each row, different letters (a–c) indicate statistically significant differencea For behaviours were log transformed values were analyzed, untransformed mb Kruskal–Wallis tests; P-values obtained by means of Chi-square tests with 3 dc Gestation period, F2,39.
our Science 150 (2014) 17– 26 21
2, and 1 m2/ewe respectively; �2, 2 df = 5.28; P = 0.071), andmineral resource licking (mean: 2.60, 3.82, and 2.41% for 3,2, and 1 m2/ewe respectively; F2,6 = 5.02; P = 0.052) tendedto be highest at 2 m2/ewe. None of the other behaviourswere affected by space allowance, and overall meanvalues (±SE) were 41.66 ± 2.12, 0.99 ± 0.13, 0.75 ± 0.08,and 0.53 ± 0.14% for standing, self-grooming, moving, andstereotypies respectively.
3.2. Point of gestation
3.2.1. Focal observationsResults for the effects of point of gestation period
are presented in Table 2. The results show a generaltrend towards lower activity levels as gestation pro-gressed regardless of space availability. This tendency wasindicated by a decline between periods 1 and 4 in the per-centage of time and frequency of standing, a decline in
the frequency of pen-exploring events after period 1, alongwith a trend for less time spent exploring. The frequency ofpositive social interactions also decreased from period 1 toperiod 3, although a later increment was observed during
for the percentage of time and frequency of occurrence of behavioural
s (P < 0.05).ean values are presented.egrees of freedom.
l Behaviour Science 150 (2014) 17– 26
esta
tion
wee
k
on
the
per
cen
tage
of
occu
rren
ce
(mea
n
±
SE)
of
obse
rved
beh
avio
urs
du
rin
g
scan
sam
pli
ngs
.
Ges
tati
on
wee
k
9
(n
=
9)
10
(n
=
9)
11
(n
=
9)
12
(n
=
9)
13
(n
=
9)
14
(n
=
9)
15
(n
= 9)
16
(n
=
9)
17
(n
=
9)
18
(n
=
9)
19
(n
=
9)
F 10,
60P-
valu
e
Mea
n
SE
Mea
n
SE
Mea
n
SE
Mea
n
SE
Mea
n
SE
Mea
n
SE
Mea
n
SE
Mea
n
SE
Mea
n
SE
Mea
n
SE
Mea
n
SE
62.8
9a
5.73
45.5
6bc
5.73
48.8
1abc
5.73
34.1
9cd
5.73
52.6
4ab
5.73
46.4
1bc
5.73
23.8
6d
5.73
39.8
3bc
5.73
22.8
4d
5.73
40.9
1bc
5.73
40.3
7bc
5.73
5.01
<0.0
001
0.64
1.18
1.56
0.59
0.69
0.48
0.97
0.77
0.37
0.66
0.39
1.98
0.05
116.
15d
7.24
39.1
8bc
7.24
25.0
4cd
7.24
42.4
0bc
7.24
32.3
0c
7.24
29.6
3c
7.24
55.0
0ab
7.24
37.0
7bc
7.24
70.4
1a
7.24
39.8
4bc
7.24
42.7
7bc
7.24
5.38
<0.0
001
4.62
3.26
2.55
1.26
2.55
0.95
0.65
1.24
0.09
0.24
0.00
14.0
6
<0.0
001
1.24
1.46
0.99
0.15
1.31
0.51
0.85
1.38
1.64
0.83
0.55
1.40
0.20
4220
.59a
6.31
bc
11.1
8abc
13.8
6abc
5.21
cd
17.1
0ab
16.3
2ab
11.2
2abc
2.23
d
13.0
1abc
13.8
1abc
3.82
0.00
050.
35
0.20
0.08
0.08
0.18
0.18
0.14
0.10
0.63
0.56
0.47
0.24
0.64
0.26
0.08
0.08
0.17
0.11
0.43
0.19
0.77
0.29
15.4
8
0.11
57–
–
6.40
a
5.48
a
2.39
bc
2.63
c
0.57
e
4.94
b
1.77
cde
1.93
cd
0.40
de
6.49
<0.0
001
1.99
1.23
0.88
0.52
0.96
0.57
0.16
0.11
0.34
0.20
0.32
0.13
0.08
0.08
0.79
0.22
0.00
0.00
0.35
0.14
0.00
0.00
21.6
7
0.01
69l sa
1.05
1.71
2.24
1.63
1.62
1.11
0.54
1.80
0.18
1.22
0.53
1.20
0.31
03
al sb0.
46
0.26
0.37
0.28
0.10
0.10
0.14
0.09
0.32
0.16
0.39
0.21
0.52
0.25
0.88
0.29
0.31
0.16
0.59
0.27
0.40
0.24
9.30
0.50
35
row
, dif
fere
nt
lett
ers
(a–e
)
ind
icat
e
stat
isti
call
y
sign
ifica
nt
dif
fere
nce
s
(P
<
0.05
).io
urs
wer
e
log
tran
sfor
med
valu
es
wer
e
anal
yzed
, un
tran
sfor
med
mea
n
valu
es
are
pre
sen
ted
.l s
ign
ifica
tion
of
mai
n
effe
cts
was
test
ed
usi
ng
Kru
skal
–Wal
lis
test
s;
P-va
lues
wer
e
obta
ined
by
mea
ns
of
Ch
i-sq
uar
e
test
s
wit
h
10
deg
rees
of
free
dom
.
48.
22 X. Averós et al. / Applied Anima
period 4, reaching similar values to those observed duringperiod 2. Both the percentage of time spent licking mineralresources and the frequency of licking events declined afterthe second observation period. In contrast, the percentageof time spent resting increased for periods 3 and 4 withrespect to 2, with a similar trend seen for the frequencyof resting events for all groups. Teeth grinding was onlyobserved during the last two observation periods, and thepercentage of time spent exhibiting stereotypic behaviourvaried but with no clear pattern (Table 2).
3.2.2. Scan sampling observationsIn agreement with the results obtained from the focal
observations, there was a relative decline in activity withadvancement of the gestation period (Table 3). Mean per-centage of standing between weeks 14 and 19 was lowerthan during week 9, whereas the mean percentage of timesobserved resting between weeks 12 and 19 of gestationwere consistently higher than during week 9. The meanpercentage of time observed licking mineral resources fromweek 13 of gestation until the end of the experiment wassignificantly lower than in weeks 9 and 10, except dur-ing week 16 of gestation, which were similar to those ofthe 2 initial weeks after the first introduction of mineralresources. There was a general decrease in the occur-rence of stereotypic behaviours between weeks 9 and 15of gestation, with values remaining between 0 and 0.35%afterwards until the end of the experimental period, withthe sole exception of a peak during week 16.
Space allowance and week of gestation had an inter-active effect on the occurrence of explorative behaviours(F20,60 = 3.51; P < 0.001). Although lower exploration wasobserved at 2 m2/ewe during week 9 of gestation than atthe other space allowances, more exploration subsequentlyoccurred at 3 m2/ewe than at the lower allowances duringweeks 10, 13 and 15 (Fig. 1). After week 15 values remainedlow for all treatments, with no differences between spaceallowances. The occurrence of eating behaviour showedsignificant changes throughout the study period, althoughno clear pattern could be actually observed. Changes inthe occurrence of moving behaviour throughout pregnancytended towards the statistical signification, although noclear pattern could be determined either.
4. Discussion
The results of this study indicate that both spaceallowance and point of gestation affected the behaviour ofpregnant ewes although, contrary to what was expected,their interaction was almost inexistent.
4.1. Space allowance
When studying the effects of space allowance it isessential to understand that variations can be achievedby either altering the enclosure area while maintaininga constant number of animals, or by altering the number
of animals if the enclosure area remains unchanged. Bothapproaches result in changes in the space allowance perindividual. In practice, however, whether the experimen-tal choice is to maintain constant enclosure size and vary Ta
ig. 1. Interaction between the effects of space allowance and the gestaampling. P < 0.10; *P < 0.05; **P < 0.01; ***P < 0.001.
roup size or the opposite, approaches will lead to con-ounding effects between space allowance and group size,oth with strong implications on the behaviour and welfaref animals (Estevez et al., 2007; Leone and Estevez, 2008).onsidering the impossibility of simultaneously control-
ing for enclosure size and group size while elucidatinghe effects of space allowance, for this study we chose to
aintain group size constant.The results of this study evidenced that ewes housed at
m2/ewe spent less time moving (walking and running)nd tended to move less often compared to ewes housedt 3 m2/ewe. It is possible that reduced movement mighte caused by difficulties in movements due to the phys-
cal proximity of their pen mates, as reported elsewhereAverós et al., 2013). A similar effect of limited space haseen observed in other domestic species (Newberry andall, 1990; Estevez et al., 2007).
These results would also agree with those of Caropreset al. (2009), who observed a greater proportion of ewesalking at 3 m2/ewe than at 1.5 m2/ewe. The proportion
f ewes walking in such study was low as compared tother behaviours, and this would agree with the fact that,n our study, the average percentage of time spent moving
as low (about 1.5% of the observation time at 3 m2/ewe).hese authors attributed this finding to a reduction in thehysical interactions among ewes under reduced spacellowance. In contrast, our results show higher frequen-ies of both positive and negative social interactions forwes maintained at 1 m2/ewe. In particular, most times theigher frequency of interactions between ewes at lowerpace allowance was positive (almost 80% of total socialnteractions). Therefore, it seems unlikely that space limi-ation resulted in reduced activity because of higher levels
f ‘ocial conflictı́ per se. However, it could be argued that,n order to sort out the conflict between movement needsnd the presence of other pen mates hindering move-ent, ewes used social interactions. Alternatively, it is also
ek on the mean frequency (±SE) of pen exploring observed during scan
possible that the higher level of social interactions mightresult simply from a higher chance of finding another indi-vidual in the path of movement at lower space allowance.The observed reduction in the frequency of positive socialinteractions, as standing and exploration, after the first ges-tation period in all groups would support the latter. Thepartial recovery in the number of positive social inter-actions observed during the last gestation period couldprobably be explained by the more frequent changes inposture during late gestation, which would have madephysical encounters with other pen mates more likely aswell.
The importance of space allowance, among other hous-ing aspects, on the resting patterns of animals has beensuggested for other domestic species such as growing-finishing pigs (Averós et al., 2010a). In our study, althoughno significant effects of space allowance on restingbehaviour were detected, there was a trend towards adecline in the frequency of resting bouts (focal obser-vations) and the mean percentage of times ewes wereobserved resting (scan samplings) with lower space. Theseresults are in agreement with those of Bøe et al. (2006)for sheep, and would suggest that movement difficultiesencountered at 1 m2/ewe would, to some extent, inter-fere the resting patterns of ewes in comparison to higherspace allowances. The frequency of resting bouts tendedto decrease with decreasing space allowance but no dif-ferences were detected regarding the total % of time spentresting. This means that ewes with less space tended to restfor longer periods before getting up, while those with morespace tended to have shorter resting bouts. This might bedue to ewes resting at higher space allowance getting morefrequently disturbed by the movement of other pen mates.
An interesting result was the higher proportion of timespent eating and higher frequency of visits to the feederobserved (both in focal and scan sampling) in ewes main-tained at lowest space allowance. Stressful conditions, such
l Behavi
24 X. Averós et al. / Applied Anima
as space restriction, are known to trigger adaptations in thefeeding strategies of animals, thereby negatively affectingtheir performance, as it has been described for growing-finishing pigs (Averós et al., 2012). Weight gains were notmonitored in the present study, and all behaviour observa-tions started once ewes had finished their morning meal,although it is possible that some feed leftovers might havebeen still be present at the feeder. In addition, there wereeight individual feeders available at each experimentalpen to prevent any competition for food, and they wereequipped with individual head protectors. This type offeeders reduces the prevalence of feeder aggression in pigs(Baxter, 1989). Therefore, although it is possible that ewesate these leftovers during visits to the feeder, it might bespeculated that the presence of individual head protectionsmade the feeder be perceived as a safe place to visually getaway from other pen mates. In this sense, the effect of visualbarriers as an effective protective device has been proved inseveral studies in poultry (Newberry and Shackleton, 1997;Cornetto et al., 2002; Leone and Estevez, 2008).
Following a similar trend to that observed for the useof feeders, the trend towards a smaller use of drinkersat the highest space allowance might be related to theclose relationship between feed and water intake in rumi-nants (Silanikove, 1989), so that this trend could be, atleast partially, a consequence of the lower use of thefeeder in pens with higher space allowance. Caroprese et al.(2009) interpreted the highest proportion of ewes observeddrinking at lower space allowance in terms of behaviouralredirection, as it has also been suggested for growing-finishing pigs (Averós et al., 2012). This might apply toour results, so that this explanation cannot be discardedeither.
4.2. Point of gestation
Resting behaviour, both in terms of percentage of timeand frequency, became more predominant towards the endof gestation, and this occurred with a simultaneous reduc-tion in standing activity. These results might be attributedto an increase in the physical discomfort and/or fatigueexperienced by ewes during the later stages of gestation, sothat they gradually prioritized lying down. However, dueto scarcity of information, comparisons with other studiesare limited. Caroprese et al. (2009) studied the behaviourof ewes during one year, but did not mention the existenceof a dependence of changes in the behaviour of ewes acrossthe experiment according to space allowance treatments.Irrespective of space allowance, they found that the great-est proportions of ewes eating and walking were observedin January (end of gestation) and February, although it isnot clear whether ewes were observed during the feedingperiods.
Similarly to resting, the reduction in the frequency ofpen exploring further support the global decrease in theactivity levels of ewes as lambing approached, which is tobe expected due to the associated increase in their body
weight towards the end of pregnancy. The drop in explo-ration that occurred after period 1 (gestation weeks 9–10)would additionally suggest initial restlessness due to thenovelty of the new enclosure and social environment. The
our Science 150 (2014) 17– 26
drop in the frequency of explorative behaviours was partic-ularly steady for ewes housed at 1 m2/ewe, while trends forthe other space allowances were less clear, and thereforemore difficult to interpret. Despite this, results suggest thatunder highly restricted spatial conditions, initial explo-ration of the physical environment might be due to thecloser proximity of pen mates, in the sense that ewes closerto the walls or to other physical items, due to difficul-ties in moving within the pen, would have redirected theirinterest towards pen walls and/or physical items. Oncehabituated to the new environment, ewes closer to thewalls or other physical features lost interest in them, andsimply stood.
Free access to mineral resources is a common practiceaiming to avoid mineral deficiencies in sheep (McDowell,1996). The use of mineral resources may be affected bysocial dynamics within the flock (Lawrence and Wood-Gush, 1988) and boredom, as it might happen in the caseof barren environments, has been suggested to increasetheir utilization (Crosby et al., 2004). No clinical signs ofmineral deficiencies were observed in our experiment,therefore it is speculated that the high initial use of mineralresource after the blocks were introduced in the enclo-sures would most likely be due to a novelty effect. Thisfact is further supported by the peak in the mineral blocklicking frequency observed during week 16 of gestation,when salt blocks were replaced by vitamin-mineral cor-rector cubes. Decreased levels of mineral resource lickingobserved between weeks 17 and 19 of gestation might beattributable to the global reduction in the activity levels ofewes at the end of pregnancy.
The lack of a clear pattern in the occurrence of eat-ing behaviour observed during scan samplings may relateto differences in food composition and characteristics dueto diet changes throughout gestation. However, no differ-ences were found either in the proportion of time spentand the frequency of visits to the feeder through gestationduring focal sampling observations.
There was a significant effect of gestation period on thetime spent (focal observations) and frequency (scan obser-vations) of stereotypies. The proportion of time performingbehavioural stereotypies was highest during period 2, witha tendency to decline afterwards. Despite the existenceof a peak during week 16, the frequency of stereotypicbehaviours markedly decreased during week 12, one weekafter the introduction of salt blocks. During week 11, whensalt blocks were novel, the mean percentage of stereo-typic behaviours performed remained high. This might beexplained by the fact that pen licking and tongue playing,which were considered as stereotypic behaviours, wouldhave become more frequent as a consequence of the uti-lization of mineral resources. A similar explanation canbe offered for the peak observed during week 16, whenvitamin-mineral corrector cubes were first introduced.Nevertheless, the overall decrease of stereotypies as ges-tation progressed remains unexplained.
In sheep, teeth grinding is usually associated with
pain and, in the case of gestating ewes, to pregnancytoxaemia, particularly during late gestation (Duncanson,2012). Although the average time spent teeth grinding andthe average number of episodes observed in this study may
l Behavi
gftsipm
5
orpwelstsvgebgttot
bawodas
A
UseDAadtat
R
A
A
X. Averós et al. / Applied Anima
enerally be considered low, some episodes were detectedrom gestation week 15 until the end of gestation. No addi-ional signs of pregnancy toxaemia were observed, ando that cause can be ruled out. Therefore, teeth grind-ng episodes might be better explained by discomfort andotentially pain caused by the normal foetuses’ develop-ent as pregnancy progressed.
. Conclusion
This study provides evidence of the behavioural changesccurring in gestating dairy sheep as consequence ofeduced space allowance. Over the gestation period,regnant ewes were clearly affected by space allowance,ith their ability to move within their pen being appar-
ntly thwarted, and with a trend to rest less often but foronger periods at 1 m2/ewe. Ewes housed at 1 m2/ewe alsohowed higher frequencies of social interactions, mainlyhrough the use of positive ones. On the other hand, lowerpace allowances resulted in longer and more frequentisits to the feeders, suggesting an adaptive strategy toet away and seek visual protection from pen mates. Novidences were found in terms of the interactive effectsetween space allowance and gestation, suggesting thatestation per se did not influence the effect of space restric-ion on behaviour. After disappearance of the pen and/orhe social environment novelty effects, the strongest effectsn behaviour seem to be caused by the progressing of ges-ation, likely associated to physical discomfort/fatigue.
In terms of practical implications of our study,ehavioural differences suggest that reducing space avail-bility to 1 m2/ewe has a potential negative impact inelfare. This was particularly evidenced in the restriction
f movement. However, it is possible that the behaviouralifferences observed in this study may, in part, obey todaptive strategies to better cope with the negative con-equences of spatial restriction.
cknowledgements
The authors gratefully acknowledge the Europeannion financial support, to conduct the study and provide
alary for X. Averós and J. Marchewka, under the Sev-nth Framework Programme for Research, Technologicalevelopment and Demonstration Activities, of the Projectnimal Welfare Indicators (AWIN; FP7-KBBE-2010-4). Theuthors also thank Imanol Etxebarría and Juan Carlos Ochoae Zuazola for the excellent care of the animals all alonghe experiment. We want to express our gratitude to twononymous Reviewers than contributed to largely improvehe final version of this manuscript.
eferences
verós, X., Brossard, L., Dourmad, J.Y., de Greef, K.H., Edge, H.L., Edwards,S.A., Meunier-Salaün, M.C., 2010a. Quantitative assessment of theeffects of space allowance, group size and floor characteristics on thelying behaviour of growing-finishing pigs. Animal 4, 777–783.
verós, X., Brossard, L., Dourmad, J.Y., de Greef, K.H., Edge, H.L., Edwards,S.A., Meunier-Salaün, M.C., 2010b. A meta-analysis of the combinedeffect of housing and environmental enrichment characteristics onthe behaviour and performance of pigs. Appl. Anim. Behav. Sci. 127,73–85.
our Science 150 (2014) 17– 26 25
Averós, X., Brossard, L., Dourmad, J.Y., de Greef, K.H., Edwards, S.A.,Meunier-Salaün, M.C., 2012. Meta-analysis on the effects of the phys-ical environment, animal traits, feeder and feed characteristics on thefeeding behaviour and performance of growing-finishing pigs. Animal6, 1275–1289.
Averós, X., Lorea, A., Beltrán de Heredia, I., Arranz, J., Ruiz, R., Estevez, I.,2013. Space availability in confined sheep during pregnancy; effectsin movement patterns and use of space. Submitted for publication.
Baxter, M.R., 1989. Design of a new feeder for pigs. Farm Build. Progr. 96,19–22.
Bessei, W., 2006. Welfare of broilers: a review. World Poult. Sci. J. 62,455–466.
Bøe, K.E., Berg, S., Andersen, I.L., 2006. Resting behaviour and displacementin ewes-effects of reduced lying space and pen shape. Appl. Anim.Behav. Sci. 98, 249–259.
Boyazoglu, J., Morand-Fehr, P., 2001. Mediterranean dairy sheep and goatproducts and their quality: a critical review. Small Rumin. Res. 40,1–11.
Brown, D.E., Harrison, P.C., Hinds, F.C., Lewis, J.A., Wallace, M.H., 1977.Heat stress effects on fetal development during late gestation in theewe. J. Anim. Sci. 44, 442–446.
Caroprese, M., Annicchiarico, G., Schena, L., Muscio, A., Migliore, R., Sevi,A., 2009. Influence of space allowance and housing conditions on thewelfare, immune response and production performance of dairy ewes.J. Dairy Res. 76, 66–73.
Cornetto, T., Estevez, I., Douglass, L.W., 2002. Using artificial cover toreduce aggression and disturbances in domestic fowl. Appl. Anim.Behav. Sci. 75, 325–336.
Crosby, T.F., Boland, T.M., Brophy, P.O., Quinn, P.J., Callan, J.J., Joyce, D.,2004. The effects of offering mineral blocks to ewes pre-mating andin late pregnancy on block intake, pregnant ewe performance andimmunoglobulin status of the progeny. Anim. Sci. 79, 493–504.
Dobson, H., Fergani, C., Routly, J.E., Smith, R.F., 2012. Effects of stress onreproduction of ewes. Anim. Reprod. Sci. 130, 135–140.
Duncanson, G.R., 2012. Nutrition and metabolic conditions. In: Dun-canson, G.R. (Ed.), Veterinary Treatment of Sheep and Goats. CABInternational, Wallingford, Oxfordshire, UK, pp. 117–126.
Estevez, I., 2007. Density allowances for broilers: where to set the limits?Poultry Sci. 86, 1265–1272.
Estevez, I., Andersen, I.-L., Nævdal, E., 2007. Group size, density and socialdynamics in farm animals. Appl. Anim. Behav. Sci. 103, 185–204.
Gonyou, H.W., Stookey, J.M., McNeal, L.G., 1985. Effects of double deck-ing and space allowance on the performance and behaviour of feederlambs. J. Anim. Sci. 60, 1110–1116.
Gonyou, H.W., Brumm, M.C., Bush, E., Deen, J., Edwards, S.A., Fang-man, T., McGlone, J.J., Meunier-Salaün, M., Morrison, R.B., Spoolder,H., Sundberg, P.L., Johnson, A.K., 2006. Application of broken-lineanalysis to assess floor space requirements of nursery and grower-finisher pigs expressed on an allometric basis. J. Anim. Sci. 84,229–235.
Horton, G.M.J., Malinowski, K., Burgher, C.C., Palatini, D.D., 1991. The effectof space allowance and sex on blood catecholamines and cortisol, feedconsumption and average daily gain in growing lambs. Appl. Anim.Behav. Sci. 32, 197–204.
Ingvartsen, K.L., Andersen, H.R., 1993. Space allowance and type of housingfor growing cattle: a review of performance and possible relation toneuroendocrine function. Acta Agr. Scand. A – Anim. Sci. 43, 65–80.
Krawczel, P.D., Klaiber, L.B., Butzler, R.E., Klaiber, L.M., Dann, H.M., Mooney,C.S., Grant, R.J., 2012. Short-term increases in stocking density affectthe lying and social behavior, but not the productivity, of lactatingHolstein dairy cows. J. Dairy Sci. 95, 4298–4308.
Lawrence, A.B., Wood-Gush, D.G.M., 1988. Influence of social behaviouron utilization of supplemental feedblocks by Scottish hill sheep. Anim.Prod. 46, 203–212.
Leone, E.H., Estevez, I., 2008. Use of space in the domestic fowl: separatingthe effects of enclosure size, group size and density. Anim. Behav. 76,1673–1682.
McDowell, L.R., 1996. Free choice mineral supplements for grazing sheep
in developing countries. In: Masters, D. (Ed.), Detection and Treatmentof Mineral Nutrition Problems in Grazing Sheep. ACIAR, Canberra,Australia, pp. 81–94.
Newberry, R.C., Hall, J.W., 1990. Use of pen space by broiler chickens:effects of age and pen size. Appl. Anim. Behav. Sci. 25, 125–136.
Newberry, R.C., Shackleton, D.M., 1997. Use of visual cover by domesticfowl: a Venetian blind effect? Anim. Behav. 54, 387–395.
Petherick, J.C., 2007. Spatial requirements of animals: allometry andbeyond. J. Vet. Behav. 2, 197–204.
Petherick, J.C., Phillips, C.J.C., 2009. Space allowances for confined live-stock and their determination from allometric principles. Appl. Anim.Behav. Sci. 117, 1–12.
Russel, A.J.F., Doney, J.M., Gunn, R.G., 1969. Subjective assessment of bodyfat in live sheep. J. Agric. Sci. 72, 451–454.
Sanchez, C., Estevez, I., 1998. The Chickitizer Software Program. Universityof Maryland, College Park, MD (USA).
Sevi, A., Massa, S., Annicchiarico, G., Dell’aquila, S., Muscio, A., 1999. Effect
of stocking density on ewes’ milk yield, udder health and microenvi-ronment. J. Dairy Res. 66, 489–499.
Shelton, M., 1964. Relation of environmental temperature during ges-tation to birth weight and mortality of lambs. J. Anim. Sci. 23,360–364.
Sibbald, A.M., Shellard, L.J.F., Smart, T.S., 2000. Effects of space allowanceon the grazing behaviour and spacing of sheep. Appl. Anim. Behav. Sci.70, 49–62.
Silanikove, N., 1989. Interrelationships between water, food and digestibleenergy intake in desert and temperate goats. Appetite 12, 163–170.
Turner, S.P., Ewen, M., Rooke, J.A., Edwards, S.A., 2000. The effect of spaceallowance on performance, aggression and immune competence ofgrowing pigs housed on straw deep-litter at different group sizes.
Livest. Prod. Sci. 66, 47–55.
Wechsler, B., 2011. Floor quality and space allowance in intensive beefproduction: a review. Anim. Welfare 20, 497–503.
Winer, B.J., 1971. Statistical Principles in Experimental Design, 2nd Edi-tion. Mc-Graw-Hill, Inc., NY, USA.
