Eg

Ja

b

AAA

KPRSSW

UT

(

0h

Applied Animal Behaviour Science 151 (2014) 13– 23

Contents lists available at ScienceDirect

Applied Animal Behaviour Science

jou rn al hom epage : w ww.elsev ier .com/ locate /applan im

ffect of rubber slat mats on the behaviour and welfare ofroup housed pregnant sows

ulia Adriana Calderón Díaza,b,∗, Laura Ann Boylea

Pig Development Department, Teagasc Animal and Grassland Research and Innovation Centre, Moorepark, Fermoy, Co. Cork, IrelandSchool of Agriculture and Food Science, University College Dublin, Belfield, Dublin 4, Ireland

a r t i c l e i n f o

rticle history:ccepted 27 November 2013vailable online 4 December 2013

eywords:ostural behaviourubber matspatial behaviourowselfare

a b s t r a c t

This study evaluated the effect of flooring, lameness, body and limb lesion scores on posturaland spatial behaviour of gestating sows. Sixty-four sows were kept in groups of four in penswith four solid concrete floored feeding stalls and a concrete fully slatted group area from4 weeks after service. The slats were either left uncovered (CON; n = 8 groups) or 10 mmthick rubber slat mats were affixed (RUB; n = 8 groups). Lameness (0 = normal to 5 = severe),limb (environmentally induced lesions, e.g. callus, wounds, swellings; scored according toseverity—0 = normal to 6 = severe) and body (aggression induced lesions; scored accordingto severity—0 = normal to 5 = severe) lesions were scored on days 1, 8, 25, 50 and 75 relativeto entering the experiment. Additionally video recordings were made of the groups for24 h on the same days which were sampled instantaneously every 10 min. An index of theproportion of time spent in (1) different postures (standing, ventral [VL] and lateral lying[LL] and total lying [VL + LL]); (2) locations (stalls or group area), (3) posture by location and(4) number of postural changes was calculated. Sows were categorized as non-lame (score≤1) or lame (score ≥2). Median scores were calculated for body and limb lesions and wereclassified as ≤median or >median. Lameness, limb and body lesions were analysed usinglogistic binomial regression. Behavioural variables were tested for normality and analysedusing mixed model equations. Flooring did not affect lameness, body lesion scores, timespent in each posture or the index of postural changes (P > 0.05). RUB sows spent more timein the group area (76.3 vs. 53.3 ± 5.8%; P < 0.01) and lay more there (80.0 vs. 62.4 ± 5.3%;P < 0.05) compared with CON sows. Sows with scores >median for wounds on the limbsspent more time LL (41.2 vs. 48.3 ± 3.6%; P < 0.05) and less time VL (36.3 vs. 29.9 ± 2.9%;P < 0.05). On the other hand, sows with body lesion scores >median spent more time VL(29.9 vs. 36.3 ± 2.9%; P < 0.05). Lame sows stood less and lay more (P < 0.05) in the feeding

stalls. When sows had access to rubber flooring they spent more time in the covered areaand lay more there compared with sows in pens where the concrete slats were bare. Thisreflects the preference of group housed sows for a comfortable surface for lying duringpregnancy. Other sow factors such as body and limb lesions and lameness status are alsorelated with lying behaviour.

© 2013 Elsevier B.V. All rights reserved.

∗ Corresponding author at: Department of Animal Sciences, Iowa Stateniversity, 337 Kildee Hall, Ames, Iowa, 50011, United States.el.: +1 515 294 32 28/+1 515 708 75 43; fax: +1 515 294 56 98.

E-mail addresses: jacalder@iastate.edu, juli.adri.c@gmail.comJ.A.C. Díaz), laura.boyle@teagasc.ie (L.A. Boyle).

168-1591/$ – see front matter © 2013 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.applanim.2013.11.016

1. Introduction

Comfort while lying is of vital importance for sow wel-fare as pregnant sows spend about 80% of their time lying(Ekkel et al., 2003). Hence, given that the majority ofpregnant sows worldwide are kept on concrete floors it is

mal Beh

14 J.A.C. Díaz, L.A. Boyle / Applied Ani

likely that their comfort needs are not being met (Elmoreet al., 2010). Floor type affects the incidence of lameness(Heinonen et al., 2006; Zurbrigg and Blackwell, 2006) andlimb lesions (Mouttotou et al., 1998; KilBride et al., 2008;von Wachenfelt et al., 2008). Straw bedding improves thephysical and thermal comfort of the floor (Barnett et al.,2001; Tuyttens, 2005); however, in modern pig productionsystems, the use of straw is unfeasible because of liquidmanure disposal systems and the associated increase inproduction costs and labour. Rubber slat mats could be analternative to bedding for pigs. They are more yielding andhave a lower thermal conductivity than bare concrete (Bøeet al., 2007) making them warmer to lie on. It also appearsthat they are less injurious than concrete as group housedpregnant gilts kept on rubber slat mats were less likely tobe lame and had a reduced risk of severe swellings andwounds on the limbs compared with gilts on concrete slats(Calderón Díaz et al., 2013). Research on the influence ofrubber flooring on sow behaviour is limited to two shortterm studies. Both Tuyttens et al. (2008) and Elmore et al.(2010) reported that when rubber mats were added to agroup housing system, sows preferred to rest on areas cov-ered with rubber mats compared to uncovered/concreteareas. Furthermore, sows spent more time lying laterallyin areas covered with rubber mats vs. areas of bare con-crete. In accordance with the findings of Boyle et al. (2000)for sows on rubber in farrowing crates, sows in groups onrubber slat mats also showed greater ease of changing pos-ture (Tuyttens et al., 2008; Elmore et al., 2010). The effectof rubber flooring on sow lying and spatial behaviour dur-ing her entire pregnancy is not known. Additionally, little isknown about the potential influence which welfare issuessuch as lameness and limb lesions may have on the posturaland spatial behaviour of group housed sows. Therefore theobjectives of this study were (1) to compare lameness, limband body lesion scores of sows housed on concrete slat-ted floor or rubber slat mats, and (2) to evaluate the effectof flooring type, lameness, body and limb lesion scores onpostural and spatial behaviour of gestating sows.

2. Materials and methods

2.1. Ethical statement

This trial was conducted in accordance with the Interna-tional Guiding Principles for Biomedical Research InvolvingAnimals as issued by the Council for the InternationalOrganizations of Medical Sciences in 1985 and the ethi-cal guidelines from the International Society for AppliedEthology. However, as at the time that the study was con-ducted Teagasc did not have an ethical committee, theprotocol did not undergo a formal ethical review. Theresearch farm on which this experimental work was con-ducted was in compliance with Statutory Instrument S.I.No. 311 of 2010 European Communities (Welfare of FarmedAnimals) Regulations 2000. Furthermore, as no invasive

measures were used, the experiment did not require licens-ing under Directive 2010/63/EU and S.I. No. 543 of 2012European Communities (Amendment of Cruelty to AnimalsAct, 1876) Regulations 2005.

aviour Science 151 (2014) 13– 23

2.2. Experimental design and husbandry

The study was conducted on the experimental farmof the Pig Development Department, Teagasc Animal andGrassland Research and Innovation Centre, Moorepark,Fermoy, Co. Cork, Ireland from March 2011 to June 2012. Atotal of 64 Large White × Landrace sows housed in groupsof four were included in the study.

Thirty-two first parity sows (or ‘gilts’) were artificiallyinseminated between 32 and 36 weeks of age by trainedfarm staff during their second heat. They were housedtogether in a single ‘all gilt’ group with an electronic sowfeeder and remained in this accommodation until assign-ment to the experiment at 28 days post-service. Thirty-twomultiparous sows were inseminated in individual stallswhen heat was detected after weaning. Sixteen of thesewere housed in a single dynamic group with an electronicsow feeder and the remaining 16 sows were housed in indi-vidual gestation stalls during their previous pregnancies.All sows were served for a second time 24 h after the firstservice and then they were moved to gestation stalls for28 days. A boar was present in the breeding barn for heatdetection at all times.

Four sows were grouped together according to parity(eight groups of gilts, and eight groups of sows parity ≥4;there were no 2nd or 3rd parity sows in the herd at the timeof the experiment) and assigned to the experiment at 28days post-service such that there were eight experimentalgroups per treatment. Within each parity group, sows werebalanced according to their lameness score on the day thetrial started. Thus, in each group there were non-lame andlame sows. Additionally, for the multiparous sows, sowswere balanced between flooring treatments according totheir previous housing system. Ultimately, thirty-two sowswere housed in pens with concrete flooring (parity average3.4 ± 2.64; BW average 213.7 ± 44.09) and 32 sows werehoused in pens with rubber slat mats in the group area(parity average 3.3 ± 2.52; BW average 199.7 ± 36.13).

The test pens used in this study had four free accessfeeding stalls (each 1.71 m L × 0.65 m W × 1.02 m H) withsolid concrete flooring. Behind the feeding stalls there wasa group area (3.20 m L × 2.68 m W) for exercise and dungingwith concrete slatted (slat width 14.5 cm, gap width 2 cm;void area = 8.2%) flooring which was either uncovered(CON) or covered with rubber slat mats (RUB) (EasyFixTM

Rubber Products, Ballinasloe, Co. Galway, Ireland). Four testpens (2 × CON and 2 × RUB) were re-used four times duringthe 16 month trial. Sows remained in the test pens until day110 of pregnancy. They were never confined in the feedingstalls and were free to move about the pen at all times.The RUB pens were identical to the CON pens except thatthe group area was covered with rubber slat mats (1.60 mL × 0.29 m W × 0.01 m H; void area = 8.2%). The rubber matsconsisted of a 10 mm thick two-strip system with circular-shaped patterns on the surface and wedges underneath.Each strip of rubber matting covered two slats and one gap.The rubber slat mats were attached to the concrete slats

by hammering the wedges underneath into the underlyinggaps (Fig. 1). No additional means of fixation was required.The house was ventilated by a cross-ventilation system,whereby fresh air entered the building through an opening

J.A.C. Díaz, L.A. Boyle / Applied Animal Behaviour Science 151 (2014) 13– 23 15

roup are

it3wsGvti

2

tlb5w

2

poaStt

2

pohwsKl(sbsntdot(bn

Fig. 1. Rubber slat mats used to cover the slatted g

n the lower part of the doors and was extracted by fans inhe roof. Sows were fed a liquid diet (water-to-meal ratio.5:1; 13.1 MJ digestive energy/kg) consisting mainly ofheat, barley and soya meal from a computerized feeding

ystem (Big Dutchman, Pig Equipment GmbH, Vechta,ermany) twice a day. Sows had ad libitum access to wateria a nipple drinker in the feed trough. Ambient tempera-ure in the gestation accommodation was recorded on thenspection/recording days and it varied from 9 ◦C to 22 ◦C.

.3. Scoring methodology

All measurements were taken by one trained researchero eliminate inter-observer variation. Locomotory ability,imb and body lesions and manure on the body (MOB) andehavioural observations were recorded on days 1, 8, 25,0 and 75 in relation to entering the trial. The dirtiness andetness of the floors was scored weekly.

.3.1.1. Locomotory abilityLocomotory ability was assessed using aspects of the

rocedure of Main et al. (2000), and included an evaluationf the sow’s standing posture and gait. Sows were given

score of 0 (not lame) to 5 (severely lame, cannot stand).ows were removed from their home pen and walked onhe concrete solid floor of the alleyway to evaluate locomo-ory ability.

.3.1.2. Limb lesionsLimb lesions (environmentally induced/arising from

ressure or traumatic contact with fixtures and fittings)n the front fetlock, carpal joint, humerus, elbow, carpus,ock, tarsus–metatarsus joint, hind fetlock and metatarsusere scored according to their severity using a ‘weighted

coring system’ adapted by Boyle et al. (2000) from deoning (1985). The lesions were classified under the fol-

owing categories: (i) score 0 = normal; (ii) score 1 = callusthickening of the epidermis and atrophy of glands); (iii)core 2 = swelling (abnormal enlargement of a part of theody, typically as a result of an accumulation of fluid); (iv)core 3 = wound (where the epidermis is interrupted butot ulcerated and there is no evidence of secondary infec-ion); (v) score 3 = bursitis (acquired fluid-filled sac thatevelops in the subcutaneous connective tissue; usuallyccurs on the hind legs below the point of the hock or on

he lateral sides of the elbow); (vi) score 4 = severe woundsthese ulcerated lesions may or may not be accompaniedy infection or (vii) severe swellings (characterized by red-ess and swelling accompanied by heat and pain) and (viii)

a. Pictures courtesy of EasyFixTM Rubber products.

score 6 = severe wounds plus severe swellings. The sum ofscores across all sites for each lesion type yielded a totalscore for each sow for each lesion type/inspection time.

2.3.1.3. Body lesionsBody lesions (i.e. scratches caused by the teeth of

another sow during an aggressive interaction) were exam-ined on the left and right sides on five regions of the sow’sbody: (1) ear; (2) neck and shoulder; (3) hindquarter; (4)belly and back and, (5) tail/ano-genital region. The lesionswere classified and scored as follows: (i) 0 = no lesions; (ii)1 = one small (approx. 2 cm), superficial lesion; (iii) 2 = morethan one small, superficial lesion or just one red (deeperthan score one) but still superficial lesion; (iv) 3 = one ormore than one big (2 to 5 cm) and deep lesions; (v) 4 = onevery big (>5 cm), deep, red lesion or more than one big,deep, red lesions and (vi) 5 = more than one very big, deep,red lesions. Addition of scores across all sites yielded a totalscore for each sow at each inspection.

2.3.2. Manure on the body (MOB)Manure on the body was assessed according to the

Welfare Quality® (2009) assessment protocol for pigs, sowsand piglets (scoring guide). The sows were made to standup for inspection and received a score from 0 to 2 where0 = up to 10% of the body surface was soiled; 1 = 10% to 30%of the body surface was soiled and 2 = more than 30% of thebody surface was soiled.

2.3.2.1. Behavioural observations. Spatial and posturalbehaviours were observed. Sows were video recorded con-tinuously during 24 h using black and white cameras (SONYCCTV camera WV-BP130/B) installed on the wall of thegestation accommodation on days 1, 8, 25, 50 and 75. Atime-elapsed VCR (MITSUBISHI HS-1024) and video mul-tiplexer (PANASONIC WJ-FS 216) were used. Sows wereindividually marked with different numbers across theirback and sides using livestock paint (Coyle’s Animal MarkerSpray, Coyle Vet. Products Ltd., Co. Galway, Ireland) foridentification during video recording. Paint was reappliedas needed during the experiment. The videotapes (TDKTV180) were analysed using instantaneous sampling every10 min. An index of the proportion of time spent in differentpostures [‘dog’ sitting, standing (S), lateral lying (LL—lyingon the side with head, shoulder, belly, hindquarters and

flank in contact with floor and the four legs extended to oneside), ventral lying (VL—sow lies on her sternum/stomachwith her head up-right, two front legs extended in front ofher, one hind leg partially under the body and the other

mal Beh

16 J.A.C. Díaz, L.A. Boyle / Applied Ani

extended out to one side) and lying (i.e. lateral and ven-tral lying combined together—L)]; time spent in the twoareas of the pen (feeding stalls and group area); and pos-tures in each location were recorded. Additionally, changesin posture between each 10 min instantaneous scan samplewere counted to calculate an index of the number of pos-tural changes made for each location and in total on eachobservation day.

2.3.2.2. Flooring cleanliness. The feeding stalls and thegroup area were scored separately. Each of the locationsreceived a score between 0 and 4 where 0 = pen clean/dry;1 = 25% of the area covered with manure/wet faeces, wateror urine; 2 = 26–50% of the area covered with manure/wetfaeces, water or urine; 3 = 51–75% of the area covered withmanure/wet faeces, water or urine and, 4 = more than 75%of the area covered with manure/wet faeces, water or urine.

2.4. Statistical analysis

The group was considered the experimental unit; there-fore, average scores were calculated for each variable foreach day of inspection. Parity was classified as 1st and ≥4.Sows were categorized as non-lame (score ≤1) or lame(score ≥2). In addition, due to the low number of sowsaffected by severe wounds, severe swellings, and severewounds plus severe swellings on the limbs, the scores forthese lesions were summed and re-classified as a singlevariable (severe lesions). Medians were calculated for bodyand limb lesions. Medians were calculated for day 1 sepa-rate to the other days of inspection. For days 8, 25, 50 and 75medians were calculated across the 4 inspection days. Val-ues were classified as ≤median or >median lesion scores.MOB was classified as clean = score ≤ 1 and dirty = score 2.To analyse floor cleanliness, soiling and wetness were clas-sified into two categories: 0 = score ≤ 1 and 1 = scores ≥ 2.Parity and BW were highly correlated (results not shown);therefore, only parity was included in the analysis.

Locomotory ability, body and limb lesion scores, MOBand flooring cleanliness were analysed using logistic bino-mial regression analysis by the use of Wald statisticsto investigate their association with the predictor vari-able. The model for locomotory ability included floor type,day of inspection, parity and day 1 scores as covariates.The model for body and limb lesions included floor type,day of inspection, parity, manure on the body and day 1scores as covariates. The model for manure on the bodyincluded floor type, day of inspection and day 1 scoresas covariates. The model for flooring cleanliness includedfloor type and ambient temperature (included as a con-tinuous variable). However, only predictor variables withP < 0.35 remained in the final model except for flooringtype which was included in the model irrespective of itsP value. Additionally, univariate models were built to iden-tify the association between wounds, swellings and bursitison the limbs and locomotory ability. Data were analysed

using SAS V9.3 PROC GENMOD (SAS Inst. Inc., Cary, NC).Statistical differences were reported when P < 0.05. Resultsare reported as odds ratios (OR) with the associated 95%confidence intervals (CI).

aviour Science 151 (2014) 13– 23

Behavioural variables were tested for normality beforethe analysis using the Shapiro–Wilk test and examinationof the normal plot. Data were analysed using SAS V9.3PROC MIXED (SAS Inst. Inc., Cary, NC). ‘Dog’ sitting was notincluded in the analysis as there were very few observa-tions of this behaviour. The models for the time spent indifferent postures included floor type, day of inspection,parity, lameness, wounds on the limbs, bursitis, swellingson the limbs and body lesion scores and ambient tempera-ture (included as a continuous variable). The models for thetime spent in different locations included floor type, day ofinspection, parity, lameness, wounds on the limbs, bursitis,swellings on the limbs and body lesion scores, floor soilingand wetness scores and ambient temperature. The modelsfor the time spent in different postures by each pen loca-tion included floor type, day of inspection, the interactionbetween floor type and day of inspection, parity, lameness,wounds on the limbs, bursitis, swellings on the limbs andbody lesion scores, floor soiling and wetness scores andambient temperature The model for the index of posturalchanges included floor type, day of inspection, parity, lame-ness, wounds on the limbs, bursitis, swellings on the limbsand body lesion scores and ambient temperature. How-ever, only predictor variables with P < 0.25 remained inthe final model except for flooring type that was includedin the model irrespective of its P value. A Tukey–Krameradjustment was used to account for multiple comparisons.Statistical differences were reported when P < 0.05. Resultsare reported as least-square means ± standard error of themean.

3. Results

Descriptive statistics (number and percent) of groups ofgestating sows with lameness (i.e. locomotion score ≥2);body and limb lesion scores >median and MOB = 2 are pre-sented in Table 1. Two sows (1 CON and 1 RUB) died duringthe experiment. The CON sow had to be euthanized becauseof an injury to her back which caused paralysis. The RUBsow became trapped under the feed trough. Two sows(1 CON and 1 RUB) were removed from the experimentbecause they were not pregnant and needed to be insemi-nated.

3.1. Factors associated with locomotory ability

On average, 33.6% of CON and 27.5% of RUB groupsof sows were lame through the experiment. Having con-trolled for lameness score at the start of the study, therewas no association between locomotory ability and flooringtype (P > 0.05).

3.2. Factors associated with limb lesions

There was no association between flooring type andany of the limb lesions recorded (P > 0.05). There was adecreased risk of scores >median for wounds on the limbs at

day 50 (P < 0.05) and day 75 (P < 0.01) compared with day 8(Table 2). Sows with scores >median for swellings (P < 0.05)and bursitis (P < 0.01) on the limbs later in the study weremore likely to have had higher scores for these lesions on

J.A.C.

Díaz,

L.A.

Boyle /

Applied

Anim

al Behaviour

Science 151 (2014) 13– 23

17

Table 1Number and percentage of groups of gestating sows with lameness (i.e. locomotion score ≥2); body (i.e. aggression induced) and limb (alopecia, calluses, swellings, wounds, bursitis and severe limb lesions)lesion scores >median (M); and manure on the body score = 2, housed on concrete slats either uncovered or covered by rubber slat mats at each of five inspections.

Variable Inspection day

Day 2 (n = 16) Day 8 (n = 16) Day 25 (n = 16) Day 50 (n = 16) Day 75 (n = 15)

Concrete Rubber Concrete Rubber Concrete Rubber Concrete Rubber Concrete Rubber

n % n % n % n % n % n % n % n % n % n %Lameness 4 25.0 2 12.5 3 18.7 3 18.7 2 12.5 2 12.5 1 6.2 2 12.5 3 20.0 2 13.3Body lesions 7 43.7 6 37.5 5 31.2 6 37.5 3 18.7 4 25.0 2 12.5 4 25.0 0 0.0 2 13.3Limb lesions

Alopecia (M = 0.15) 3 18.7 4 25.0 4 25.0 4 25.0 3 18.7 3 18.7 3 18.7 3 18.7 7 46.7 4 26.7Calluses (M = 0.25) 4 25.0 4 25.0 4 25.0 4 25.0 4 25.0 5 31.2 4 25.0 2 12.5 1 6.7 0 0.0Swellings (M = 0.08) 5 31.2 5 31.2 4 25.0 4 25.0 4 25.0 5 31.2 4 25.0 2 12.5 3 20.0 3 20.0Wounds (M = 0.17) 7 43.7 4 25.0 5 31.2 3 18.7 4 25.0 1 6.2 3 18.7 1 6.2 2 13.3 1 6.7Bursitis (M = 0.21) 5 31.2 3 18.7 6 37.5 2 12.5 5 31.2 2 12.5 6 37.5 2 12.5 4 26.7 3 20.0Severe lesions (M = 0) 4 25.0 1 6.2 2 12.5 1 6.2 4 25.0 1 6.2 2 12.5 2 12.5 0 0.0 1 6.7

Manure on the body 6 37.5 7 43.7 7 43.7 4 25.0 4 25.0 4 25.0 6 37.5 5 31.2 3 20.0 3 20.00

Table 2Differences between parity, day of inspection, lesion scores on day 1 and manure on the body scores on body, limb and lesion scores greater than the median and manure on the body score = 2 of sows duringgestation.

Parity1 Inspection day2 Scores on day 13 Manure on the body4

Day 25 Day 50 Day 75

Odds ratio 95% CI Odds ratio 95% CI Odds ratio 95% CI Odds ratio 95% CI Odds ratio 95% CI Odds ratio 95% CI

Body lesions5 5.7 1.41–23.07 0.2 0.03–1.18 0.1a 0.02–0.90 0.01a 0.002–0.11 3.0 0.78–11.49 2.3 0.65–7.97Limb lesions6

Callus 18.6a 3.17–109.12 6.5 0.74–54.13 1.0 0.12–9.33 0.4 0.07–0.48 2.9 0.45–18.32 0.05a 0.01–0.26Swelling 3.6a 1.25–10.71 NI7 NI NI NI NI NI NI NI NI NIWound NI NI 0.5 0.10–2.02 0.2a 0.03–0.84 0.1a 0.02–0.56 2.2 0.67–7.35 3.0 0.84–10.75Bursitis NI NI NI NI NI NI NI NI 11.9a 3.24–44.44 NI NISevere lesions NI NI 1.9 0.35–10.58 1.5 0.17–12.52 0.2 0.02–2.60 12.0 0.63–54.93 2.4 0.53–11

Manure on the body NI NI 1.4a 0.98–56.24 2.6a 0.70–10.03 11.4a 2.54–51.65 21.7a 1.69–278.02 – –

1 Odds ratios reported for sows parity ≥4 compared to sows parity 1.2 Odds ratios reported for days 25, 50 and 75 compared to day 8.3 Odds ratios reported for scores > median compared to scores ≤median on day 1.4 Odds ratios reported for sows with manure on the body score 2 compared with sows with manure on the body score ≤1.5 Cummulative score of lesions observed on the ear, neck/shoulder, hindquarter, and belly/back for both left and right sides of the sow, as well as the tail/ano-genital region (0 = normal to 5 = severe).6 Lesions observed on the front fetlock, carpal joint, humerus, elbow, carpus, hock, tarsus-metatarsus joint, hind fetlock, and metatarsus (0 = normal to 6 = severe injuries).7 NI = not included in the final model; P > 0.35.a Significantly different from reference category; P < 0.05

mal Beh

18 J.A.C. Díaz, L.A. Boyle / Applied Ani

day 1 compared with sows with lesion scores ≤median onday 1 (Table 2). Sows with scores >median for swellings(P < 0.05) and wounds (P < 0.01) on the limbs were at anincreased risk of lameness compared to sows with scores≤median for swellings and wounds on the limbs (Table 2).

3.3. Factors associated with body lesions

There was no association between body lesion scoresand flooring type (P > 0.05). There was a decreased risk ofbody lesion scores >median at day 50 (P < 0.05) and day 75(P < 0.01) compared with day 8 (Table 2).

3.4. Factors associated with manure on the body

There was no association between MOB scores andflooring type (P > 0.05). Sows had an increased risk of beingdirty at day 25 (P < 0.05), day 50 (P < 0.01) and day 75 com-pared with day 8 (P < 0.05) (Table 2).

3.5. Behavioural observations

3.5.1. PosturesThere was no association between the time spent in

the different postures and flooring type (P > 0.05). On day 1sows spent more time S (P < 0.01) and VL (P < 0.01) and timespent in these postures decreased as pregnancy progressed(Fig. 2A). Additionally, sows spent less time LL (P < 0.01)and L (P < 0.01) on day 1 and time spent in these posturesincreased as pregnancy progressed (Fig. 2A). First paritysows spent less time S (P < 0.01) and more time VL (P < 0.01)and L (P < 0.01) compared with older sows (Table 3). Lame-ness score did not affect the time spent in the differentpostures observed (P > 0.05). Sows with scores >median forwounds on the limbs spent more time LL (P < 0.05) and lesstime VL (P < 0.05) compared to sows with scores ≤medianfor wounds on the limbs (Table 3). Sows with body lesionscores >median spent more time VL (P < 0.05) compared tosows with body lesion scores ≤median (Table 3). Ambi-ent temperature was not related to any of the posturesrecorded.

3.5.2. LocationsRUB sows spent more time in the group area (77.8 ± 6.3%

vs. 52.4 ± 6.3%; P < 0.01) compared with CON sows. Firstparity sows spent more time in the group area com-pared with older sows (P < 0.01; Table 3). Sows with scores>median for wounds on the limbs spent less time inthe group area (P < 0.01) compared to sows with scores≤median for wounds on the limbs (Table 3). Sows withbody lesion scores >median spent more time in thegroup area (P < 0.05) compared to sows with body lesionscores ≤median (Table 3). Time spent in the feeding stallsincreased with an increase in 1 ◦C in the gestation accom-modation (P < 0.05). Lameness score did not affect thetime spent in the group area or in the feeding stalls(P > 0.05).

3.5.3. Postures by location—Group areaRUB sows spent less time S (P < 0.05) and more time

L (P < 0.05) in the group area compared with CON sows

aviour Science 151 (2014) 13– 23

(Fig. 3). Sows spent more time S (P < 0.01) on day 1 withreduced levels of S as pregnancy progressed (Fig. 2B). Sowsspent less time LL (P < 0.01) and L (P < 0.01) on day 1 and LLincreased as pregnancy progressed (Fig. 2B). There was nointeraction between flooring treatment and day of obser-vation for any of the postures (P > 0.05). First parity sowsspent less time S (P < 0.01) and more time VL (P < 0.01)and L (P < 0.01) compared with older sows (Table 3). Sowswith scores >median for wounds on the limbs spent less(P < 0.05) time VL compared to sows with scores ≤medianfor wounds on the limbs (Table 3). Sows with body lesionscores >median spent less (P < 0.05) time S and more timeVL (P < 0.05) compared to sows with scores ≤median forwounds on the limbs (Table 3). There was no difference inthe proportion of time spent in the different postures bylame and non-lame sows (P > 0.05). Ambient temperaturewas not related to any of the postures recorded in the grouparea.

3.5.4. Postures by location—Feeding stallsRUB sows spent less time LL (P < 0.05) in this location

compared to CON sows (Fig. 3). Sows spent more time S(P < 0.01) and VL (P < 0.01) on day 1 but time spent in thesepostures decreased as pregnancy progressed (Table 2). Con-versely, sows spent less time LL (P < 0.01) and L (P < 0.05) onday 1 but LL increased as pregnancy progressed (Table 2).There was no interaction between flooring treatment andday of observation for any of the postures (P > 0.05). Firstparity sows spent more time S (P < 0.05) and less timeLL (P < 0.01) and L (P < 0.05) compared with older sows(Table 3). Sows with body lesion scores >median spent lesstime (P < 0.05) lateral lying compared to sows with scores≤median for wounds on the limbs (Table 3). Lame sowsspent less time S (P < 0.05) and more time LL (P < 0.05),VL (P < 0.01) and L (P < 0.01) compared to non-lame sows(Table 3). Ambient temperature was not related to any ofthe postures recorded in the feeding stalls.

3.5.5. Postural changesThere was no association between flooring treatment

and the index of total posture changes. Additionally, therewas no difference between flooring treatment in the indexof postural changes made in the group area or in the feed-ing stalls (P > 0.05). First parity sows showed a higher indexof postural changes (27.3 ± 1.6% vs. 22.1 ± 1.5%; P < 0.05)compared with older sows throughout the experiment.Lameness, limb and body lesions and ambient tempera-ture were not related to the total index of postural changesor to the index of postural changes made in the differentlocations.

3.6. Factors associated with flooring cleanliness

There was no association between flooring type andsoiling or wetness scores of the floor in the group area(P > 0.05). Feeding stalls in the RUB pens had an increased

risk of soiling (OR = 1.7; CI = 1.24–2.47; P < 0.01) and wet-ness (OR = 2.5; CI = 1.38–4.66; P < 0.01) compared with thefeeding stalls in CON pens. There was no associationbetween soiling or wetness and ambient temperature.

J.A.C. Díaz, L.A. Boyle / Applied Animal Behaviour Science 151 (2014) 13– 23 19

Fig. 2. (A) Total time (lsmeans ± SEM%) spent standing, lateral and ventral lying and total time spent lying by gestating sows during gestation. (B) Timespent in different postures in the group area (observation days was not included in the model for ventral lying in the group area (B) as theP value was>0.35) and (C) the feeding stalls on each observation day. a,b,cSignificant differences (P < 0.05) between days within each posture.

0

20

40

60

80

100

Standing Lateral

lying

Ventral

lying

Lying Standing Lateral

lying

Ventral

lying

Lying

sllats gnideeFaera puorG

Tim

e sp

ent

%

*

*

*

Con crete

Rubb er

Fig. 3. Time (lsmeans ± SEM%) spent standing, lateral and ventral lying and total time lying by gestating sows housed on concrete slats either uncoveredor covered by rubber slats mats in the different areas of the pen. *Values differ significantly; P < 0.05

20 J.A.C. Díaz, L.A. Boyle / Applied Animal Behaviour Science 151 (2014) 13– 23

Table 3Differences between parity, lameness, wounds on the limbs and body lesion scores in the time (lsmeans ± SE %) spent in different postures (standing, lateraland ventral lying and total time lying), locations (group or stall area) and postures by location by gestating sows.

Parity Lameness Wounds on the limbs Body lesions

Parity 1 Parity ≥4 Non-lame Lame Score ≤median Score > median Score ≤median Score >median

PosturesStanding 13.7 ± 1.2a 25.5 ± 1.5b NI1 NI NI NI NI NILateral lying 38.0 ± 4.5 51.4 ± 4.7 46.9 ± 3.4 42.5 ± 3.9 41.2 ± 3.6a 48.3 ± 3.6b NI NIVentral lying 47.7 ± 3.6a 18.5 ± 3.6b NI1 NI 36.3 ± 2.9a 29.9 ± 2.9b 29.9 ± 2.9a 36.3 ± 3.0b

Lying 85.1 ± 1.8a 71.9 ± 1.7b NI NI 78.0 ± 1.6 78.9 ± 1.7 77.4 ± 1.7 79.5 ± 1.8Locations

Group area 79.7 ±5.9a 49.9 ± 5.6b NI NI 56.16 ± 5.3b 72.2 ± 5.5b

Feeding stalls 19.2 ± 6.1a 54.6 ± 6.0b 32.3 ± 4.1 41.4 ± 5.5 41.59 ± 4.8 33.0 ± 4.9Postures by location

Group areaStanding 17.2 ± 5.1a 37.4 ± 4.9b NI NI 33.9 ± 4.3 31.9 ± 4.5 32.8 ± 4.2a 21.8 ± 4.6b

Lateral lying 35.5 ± 5.1 43.9 ± 5.1 NI NI NI NI 36.2 ± 4.1 43.2 ± 4.3Ventral lying 48.9 ± 5.0a 18.0 ± 4.9b NI NI 37.5 ± 4.1a 29.4 ± 4.2b 29.5 ± 4.0a 37.4 ± 4.2b

Lying 81.8 ± 5.4a 60.7 ± 5.1b NI NI 75.5 ± 4.4 67.0 ± 4.6 65.1 ± 4.3a 77.4 ± 4.7b

Feeding stallsStanding 65.2 ± 7.4a 38.2 ± 7.3b 61.6 ± 5.4a 41.9 ± 6.7b 56.0 ± 6.1 47.4 ± 5.9 NI NILateral lying 15.4 ± 5.6a 40.4 ± 5.8b 22.4 ± 4.1a 33.4 ± 5.1b 24.4 ± 4.6 31.4 ± 4.5 34.1 ± 4.6a 21.6 ± 4.6b

Ventral lying NI NI 13.6 ± 2.1a 24.7 ± 2.8b NI NI 17.0 ± 2.6 21.3 ± 2.5Lying 32.1 ± 7.4a 59.6 ± 7.4b 34.8 ± 5.4a 56.8 ± 6.8b 41.4 ± 6.2 50.3 ± 6.0 NI NI

ounds

1 NI = not included in the final model; P > 0.25.a,b Significant differences (P < 0.05) between predictor parity, lameness, wby location.

4. Discussion

We hypothesised that sows with access to rubber floor-ing would have a decreased risk of lameness comparedwith sows on concrete; however, the prevalence of lame-ness did not differ between flooring treatments. This resultwas unexpected as in a previous study we found a bene-ficial impact of rubber slat mats on lameness in a similargroup housing system (Calderón Díaz et al., 2013). How-ever, in that study sows were kept on rubber from serviceuntil farrowing whereas in the current study they werekept in stalls for the first 28 days post service. The shortertime on experiment could explain why the rubber flooringwas not associated with an improvement in lameness inthe current study. Furthermore, Calderón Díaz et al. (2013)focused on young sows in which the benefits of a softersurface underfoot are likely to be more obvious as they areless likely to have chronic locomotory problems (Dewey,2006). In the current study half of the sows used wereparity four or above. These animals had all been housedon slatted concrete (in stalls or a in dynamic group) dur-ing their entire productive life. It is possible that a shortperiod (10 weeks) on rubber floor was insufficient to over-come any chronic locomotory disturbances and/or injuriesthese animals had sustained. Nevertheless, if the differ-ences between the findings on lameness in this study andthat of Calderón Díaz et al. (2013) were because of theinclusion of older sows in the former then we would haveexpected a significant effect of parity on locomotory ability.The fact that there was none suggests that the causal factorsfor lameness in this study may have been different to thosein the study by Calderón Díaz et al. (2013) and were such

that they could not be resolved/ameliorated by provid-ing more comfortable flooring. Finally other managementfactors such as diet composition and group size differed

on the limbs and body lesions within each posture, location and posture

between the studies and this could also have contributed tothe conflicting findings. In accordance with Calderón Díazet al. (2013) we found that higher scores for wounds andswellings on the limbs were risk factors for lameness. It ispossible either that the discomfort associated with limblesions caused alterations in locomotion (KilBride et al.,2009, 2010) or that such lesions developed in lame sowsbecause of prolonged lying (Bonde et al., 2004). Furtherresearch is required to elucidate this relationship.

Bursitis is associated with lying on hard surfaces(Mouttotou et al., 1998; Scott et al., 2006; Gillman et al.,2008; KilBride et al., 2008). The fact that sows with scoresgreater than the median at the beginning of the study con-tinued to have higher scores for this lesion throughoutthe experiment reflects the chronic nature of such limblesions. In fact the chronic nature of bursitis also explainswhy rubber flooring did not have a beneficial impact onthese lesions and is in agreement with Calderón Díaz et al.(2013). Introduction of pigs to a new environment canresult in acute injuries to the limbs (Gonyou, 1986; Boyleet al., 2002). This explains why sows had higher scores forwounds on the limbs on day 1 than on any other inspectionday. However, there was no beneficial effect of rubber floor-ing on this limb lesion which contrasts with the findings ofCalderón Díaz et al. (2013). If introduction to a new envi-ronment is accompanied by aggression between unfamiliarsows then traumatic collisions with fixtures and fittingsor slips and falls on concrete exacerbates injuries/woundsto the limbs (Philipot et al., 1994). Anecdotally there wasmuch less aggression at mixing in the current study than inthe study by Calderón Díaz et al. (2013) probably becauseall of the gilts were familiar with one another and the older

sows which were previously group housed also had priorexperience of one another. This may explain why no protec-tive effect of rubber on wounds to the limbs was detected. It

mal Beh

mwwflfr

bthrissTactfoim(

eidorsrwsSesh

f(btoaboPwphstjBwsifmws

J.A.C. Díaz, L.A. Boyle / Applied Ani

ust also be emphasised that the sample size of this studyas relatively small and so the results should be treatedith caution. Further testing of different types of rubberooring with, for instance, different group sizes and dif-

erent group feeding arrangements is required before theesults can be generalised.

In spite of there being less aggression than in the studyy Calderón Díaz et al. (2013) there was some fightingo establish a new social hierarchy between sows whichad no prior experience of one another and this waseflected in higher body lesion scores in general, after mix-ng (Sadler et al., 2011). In accordance with our previoustudy (Calderón Díaz et al., 2013), there was no relation-hip between body lesion scores and flooring treatment.his suggests that flooring has little impact on aggressiont mixing. Nevertheless body lesion scores do not alwaysorrelate well with aggression (Stukenborg, 2011) and sohis would need to be confirmed by observations of therequency and intensity of aggressive behaviour at mixingn rubber and concrete flooring. Particularly as there is anndication that sows will terminate fights earlier on rubber

ats than on concrete when they become wet and slipperyBoyle and Llamas Moya, 2003).

In agreement with Tuyttens et al. (2008) and Elmoret al. (2010), flooring type influenced the time sows spentn different areas of the pen and the time they spent in theifferent postures in those areas but it did not have an effectn the total time spent in the different postures. Sows onubber spent more time in the group area and importantly,pent more time lying there than sows on concrete. Thiseflected their preference for physical comfort while lyinghich is in accordance with findings for other species (e.g.

heep; Gordon and Crockram, 1995; Faerevik et al., 2005.oftness plays an important role in physical comfort (Bøet al., 2007) and even relatively hard and inflexible rubberuch as used in the current study is likely more yielding andence more comfortable than concrete.

Our results confirm sows’ preference for a more com-ortable surface to lie on as reported by Tuyttens et al.2008) and Elmore et al. (2010) and extend their findingsy illustrating that this preference persists throughout ges-ation. Nevertheless, they are in conflict with the findingsf both studies regarding the effect on postural changess we found no increase in posture changing in the rub-er flooring treatment. This also conflicts with findings forther species (e.g. dairy cows: Haley et al., 2001; cattle:latz et al., 2008) where the improved comfort associatedith rubber flooring was associated with higher levels ofosture changing. Prolonged and uninterrupted lying onard surfaces causes compression lesions and leg diseasesuch as carpal bursitis and periarthritis (inflammation ofhe tissues around the joint) and overworks the carpaloints during lying down and standing up (Stanek, 1997;onde et al., 2004; Schulze Westerath et al., 2007). Comforthile lying is likely to be particularly important to pregnant

ows as they spend up to 80% of their time lying. Indeedt is possible that a healthier locomotory system arising

rom more frequent posture changing was one of the main

echanisms through which rubber flooring was associatedith improvements in the locomotory ability of sows in the

tudy by Calderón Díaz et al. (2013), especially as rubber

aviour Science 151 (2014) 13– 23 21

flooring actually had a negative impact on claw health inthat study. Following on from this the absence of an effectof rubber flooring on posture changing may explain whythere was no beneficial effect of rubber on lameness scoresin the current study. However, it is more likely that the wayin which posture changes were determined was deficient.An index of posture changing was calculated by countingchanges between each 10 min sampling interval. This is incontrast to both Tuyttens et al. (2008) and Elmore et al.(2010) who counted the total number of posture changes.

Some of the sows used in this study were previouslyhoused individually and hence initially spent more time inthe feeding stalls. However, irrespective of flooring treat-ment, sows tended to spend more time in the group areaas pregnancy progressed. This may have been becausethe sows’ movement inside the feeding stalls becameconstrained by their increasing body size as pregnancy pro-gressed (O’Connell et al., 2007). Additionally, sows spentmore time lateral lying as pregnancy progressed, proba-bly to facilitate advancing gestation (Marchant-Forde andMarchant-Forde, 2004). First parity sows included in thisstudy were reared in groups since birth and had no previ-ous experience of close confinement which could explainwhy they spent more time in the group area. However, wecannot offer an explanation for the higher proportion oftime spent ventral lying by first parity compared to oldersows.

Bonde et al. (2004) reported that lame sows spend moretime lying than non-lame sows. However, lameness scorehad no effect on the time spent in the different posturesin the current study or on the proportion of time spent inthe different locations of the pen. Nevertheless, while therewas no difference in behaviour between lame and non-lame sows in the group area, lame sows spent less timestanding and more time lying in the feeding stalls. It is pos-sible that being isolated from the rest of the group, the lamesows had a better chance of undisturbed rest in the stalls.They were also better protected in the feeding stalls fromincurring further injury while lying which could have exac-erbated their lameness. These findings suggest that it mayimprove the welfare of lame or otherwise compromisedsows in groups to have the option of protection or isolationfrom their penmates while lying.

Sows with higher body lesion scores spent less timelying on their sides (i.e. laterally). Lateral lying increasesthe body area in contact with the floor (Aarnink et al., 2006)but as the majority of sows had only mild body lesions,it is unlikely that they were trying to minimize contactbetween their skin and the floor to reduce discomfort orpain. Sows with higher body lesion scores are usually thesubjects of more aggression than sows with lower scores(Barnett et al., 1992). Therefore, such sows may need tobe more vigilant regarding potential attacks by aggressorswhile lying. Ventral lying could be considered a more vigi-lant posture and less vulnerable posture than lateral lyingfor resting. However, this theory conflicts with the findingthat sows with higher body lesion scores spent more time

in the group area where they would be more susceptibleto aggression than in the feeding stalls (Arey and Edwards,1998). Sows with scores higher than the median for woundson the limbs spent more time lateral lying. As the limbs are

mal Beh

Main, D.C., Clegg, J., Spatz, A., Green, L.E., 2000. Repeatability of a scoring

22 J.A.C. Díaz, L.A. Boyle / Applied Ani

somewhat tucked underneath the body while ventral lyingthis posture could apply more pressure to wounds on thelimbs and therefore cause more pain/discomfort comparedto lying with the limbs extended.

As previously reported by Tuyttens et al. (2008) andElmore et al. (2010) sows in the rubber covered slats treat-ment spent a higher proportion of time in the uncoveredarea of the pen (i.e. the feeding stalls) when ambient tem-peratures increased. In this study the ambient temperatureranged from 9 ◦C to 22 ◦C indicating that sows were alwayswithin their thermoneutral zone (Noblet et al., 1989) andwere unlikely to have experienced heat stress. Neverthe-less our results suggest that sows preferred to spend moretime in the concrete floored feeding stalls rather than onrubber mats at even moderately warm ambient tempera-tures to increase heat loss. This emphasises the need thatpregnant sows kept in groups have for a choice of surfacesto lie on and suggests that in warmer climates covering theentire pen with rubber slat mats should be avoided as itmay contribute to heat stress in pregnant sows.

In agreement with Calderón Díaz et al. (2013), pens cov-ered with rubber slat mats were at higher risk of being moresoiled and wetter in the feeding stalls. Sows with rubberflooring in the group area spent more time lying there andso may have been more likely to keep that area clean andtherefore used the feeding stalls as places to defecate andurinate. Also in accordance with Calderón Díaz et al. (2013)there was no difference in the dirtiness of the sows’ bodybetween floor types. As discussed in that paper, the scoringsystem employed may not have been sensitive enough todetect potential treatment differences. These were likelygiven the differences reported in the dirtiness of the floor,at least in the feeding stalls, between treatments. Sowsbecame dirtier as the experiment progressed irrespectiveof flooring treatment. This was possibly due to the fact thatthe pens were not cleaned during the course of the study.

5. Conclusion

When sows had access to rubber flooring in the grouparea they spent more time there and lay more there com-pared with sows in pens where the concrete slats werebare. This confirms the preference of group housed sowsfor a comfortable surface for lying during pregnancy. Thisstudy also illustrates that there are relationships betweenother welfare indicators such as the severity of body andlimb lesions and the postural behaviour of gestating sows.

Acknowledgements

This work was administered by Teagasc’s Walsh Fel-lowship Scheme via funding from Enterprise Ireland andEasyFixTM Rubber Products. The authors thank the farmstaff of the experimental pig unit at Moorepark and to Siob-han Collins and Caroline La Marle for their help with datacollection.

References

Aarnink, A.J.A., Schrama, J.W., Heetkamp, M.J.W., Stefanowska, J., Huynh,T.T.T., 2006. Temperature and body weight affect fouling of pig pens.J. Anim. Sci. 84, 2224–2231.

aviour Science 151 (2014) 13– 23

Arey, D.S., Edwards, S.A., 1998. Factors influencing aggression betweensows after mixing and the consequences for welfare and production.Livest. Prod. Sci. 56, 61–70.

Barnett, J.L., Hemsworth, P.H., Cronin, G.M., Newman, E.A., McCallum, T.H.,Chilton, D., 1992. Effects of pen size, partial stalls and method offeeding on welfare-related behavioural and physiological responsesof group-housed pigs. Appl. Anim. Behav. Sci. 34, 207–220.

Barnett, J.L., Hemsworth, P.H., Cronin, G.M., Jongman, E.C., Hutson, G.D.,2001. A review of the welfare issues for sows and piglets in relationto housing. Aust. J. Agric. Res. 52, 1–28.

Bøe, K.E., Andersen, I.L., Buisson, L., Simensen, E., Jeksrud, W.K., 2007.Flooring preferences in dairy goats at moderate and low ambienttemperature. Appl. Anim. Behav. Sci. 108, 45–57.

Bonde, M., Rousing, T., Badsberg, H., Sørensen, H.T., 2004. Associationsbetween lying-down behaviour problems and body condition, limbdisorders and skin lesions of lactating sows housed in farrowing cratesin commercial sow herds. Livest. Prod. Sci. 87, 179–187.

Boyle, L.A., Regan, D., Leonard, F.C., Lynch, P.B., Brophy, P., 2000. The effectof mats on the welfare of sows and piglets in the farrowing house.Anim. Welfare 9, 39–48.

Boyle, L.A., Leonard, F.C., Lynch, P.B., Brophy, P., 2002. The influence ofhousing system on skin lesion scores, behaviour and responses to anACTH challenge in pregnant gilts. Ir. J. Agric. Food Res. 41, 181–200.

Boyle, L.A., Llamas Moya, S., 2003. Effect of covering slatted floors withmats on the behaviour and welfare of loose housed sows at mixing.In: Proceedings of the 37th International Congress of the InternationalSociety for Applied Ethology, p. 198.

Calderón Díaz, J.A., Fahey, A.G., KilBride, A.L., Green, L.E., Boyle, L.A., 2013.Longitudinal study of the effect of rubber slat mats on locomotoryability, body, limb and claw lesions and dirtiness of group housedsows. J. Anim. Sci. 91, 3940–3954.

de Koning, R., 1985. On the well being of dry sows. In: Ph.D. Dissertation.University of Utrecht, Utrecht, Netherlands.

Dewey, C.E., 2006. Diseases of the nervous and locomotor systems. In:Straw, B.E., Zimmerman, J.J., D’Allaire, S., Taylor, D.J. (Eds.), Diseases ofSwine. , ninth ed. Blackwell Publishing Professional, Ames, Iowa, USA,pp. 87–112.

Elmore, M.R.P., Garner, J.P., Johnson, A.K., Richert, B., 2010. A flooring com-parison: the impact of rubber mats on the health, behaviour, andwelfare of group-housed sows at breeding. Appl. Anim. Behav. Sci.123, 7–15.

Ekkel, E.D., Spoodel, H.A.M., Hulsegge, I., Hopster, H., 2003. Lying charac-teristics as determinants for space requirements in pigs. Appl. Anim.Behav. Sci. 80, 19–30.

Faerevik, G., Andersen, I.L., Bøe, K.E., 2005. Preference of sheep for differenttypes of pen flooring. Appl. Anim. Behav. Sci. 90, 265–276.

Gillman, C.E., KilBride, A.L., Ossent, P., Green, L.E., 2008. A cross-sectionalstudy of the prevalence and associated risk factors for bursitis inweaner, grower and finisher pigs from 93 commercial farms inEngland. Prev. Vet. Med. 83, 308–322.

Gonyou, H.W., 1986. Assessment of comfort and well-being in farm ani-mals. J. Anim. Sci. 62, 1769–1775.

Gordon, G.D.H., Crockram, M.S., 1995. A comparison of wooden slatsand straw bedding on the behaviour of sheep. Anim. Welfare 4,131–134.

Haley, D.B., de Passilleı́, A.M., Rushen, J., 2001. Assessing cow comfort:effects of two floor types and two tie stall designs on the behaviour oflactating dairy cows. Appl. Anim. Behav. Sci. 71, 105–117.

Heinonen, M., Oravainen, J., Orro, T., Seppa-Lassila, L., Ala-Kurikka, E., Viro-lainen, J., Tast, A., Peltoniemi, O.A.T., 2006. Lameness and fertility ofsows and gilts in randomly selected loose-housed herds in Finland.Vet. Rec. 159, 383–387.

KilBride, A.L., Gillman, C.E., Ossent, P., Green, L.E., 2008. A cross-sectionalstudy of the prevalence and associated risk factors for capped hockand the associations with bursitis in weaner, grower and finisher pigsfrom 93 commercial farms in England. Prev. Vet. Med. 83, 272–284.

KilBride, A.L., Gillman, C.E., Green, L.E., 2009. A cross-sectional study of theprevalence of lameness in finishing pigs, gilts and pregnant sows andassociations with limb lesions and floor types on commercial farms inEngland. Anim. Welfare 18, 215–224.

KilBride, A.L., Gillman, C.E., Green, L.E., 2010. A cross-sectional study ofthe prevalence and risk factors for foot lesions and abnormal posturein lactating sows on commercial farms in England. Anim. Welfare 19,473–480.

system for finishing pigs. Vet. Rec. 147, 574–576.Marchant-Forde, R.M., Marchant-Forde, J.N., 2004. Pregnancy-related

changes in behavior and cardiac activity in primiparous pigs. Physiol.Behav. 82, 815–825.

mal Beh

M

N

O

P

P

S

S

S

tocol for Pigs, 2009. Welfare Quality® Consortium, Lelystad, TheNetherlands.

J.A.C. Díaz, L.A. Boyle / Applied Ani

outtotou, N., Hatchell, F.M., Green, L.E., 1998. Adventitious bursitis of thehock in finishing pigs: prevalence, distribution and association withfloor type and foot lesions. Vet. Rec. 142, 109–114.

oblet, J., Dourmad, J.Y., Le Dividich, J., Dubois, S., 1989. Effect of ambienttemperature and addition of straw of alfalfa in the diet on energymetabolism in pregnant sow. Livest. Prod. Sci. 21, 309.

’Connell, M.K., Lynch, P.B., Bertholot, S., Verlait, F., Lawlor, P.G., 2007.Measuring changes in physical size and predicting weight of sowsduring gestation. Animal 1, 1335–1343.

hilipot, J.M., Pluvinage, P., Cimarosti, I., Sulpice, P., Bugnard, F., 1994. Riskfactors of dairy cow lameness associated with housing conditions. Vet.Res. 25, 244–248.

latz, S., Ahren, F., Bendel, J., Meyer, H.H.D., Erhard, M.H., 2008. What hap-pens with cow behaviour when replacing concrete slatted floor byrubber coating: a case study. J. Dairy Sci. 91, 999–1004.

adler, L.J., Johnson, A.K., Lonergan, S.M., Nettleton, D., Dekkers, J.C.M.,2011. The effect of selection for residual feed intake on general behav-ioral activity and the occurrence of lesions in Yorkshire gilts. J. Anim.Sci. 89, 258–266.

chulze Westerath, H., Gygax, L., Mayer, C., Wechsler, B., 2007. Leg lesionsand cleanliness of finishing bulls kept in housing systems with differ-ent lying area surfaces. Vet. J. 174, 77–85.

cott, K., Chennells, D.J., Campbell, F.M., Hunt, B., Armstrong, D., Tay-lor, L., Gill, B.P., Edwards, S.A., 2006. The welfare of finishing pigs in

aviour Science 151 (2014) 13– 23 23

two contrasting housing systems: fully-slatted versus straw-beddedaccommodation. Livest. Sci. 103, 104–115.

Stanek, C., 1997. Tendons and tendon sheaths. In: Greenough, P.R., Weaver,A.D. (Eds.), Lameness in Cattle. WB Saunders Company, Philadelphia,PA, pp. 188–194.

Stukenborg, A., 2011. Investigations on agonistic behaviour in pigs keptunder commercial farm conditions. In: Ph.D. Dissertation. Christian-Albrechts-Universität, Kiel, Germany.

Tuyttens, F.A.M., 2005. The importance of straw for pig and cattle welfare:a review. Appl. Anim. Behav. Sci. 92, 261–282.

Tuyttens, F.A.M., Wouters, F., Struelens, E., Sonck, B., Duchateau, L., 2008.Synthetic lying mats may improve lying comfort of gestating sows.Appl. Anim. Behav. Sci. 114, 76–85.

von Wachenfelt, H., Pinzke, S., Nilsson, C., Olsson, O., Ehlorsson, C.J., 2008.Gait analysis of unprovoked pig gait on clean and fouled concretesurfaces. Biosystems Eng. 101, 376–382.

Welfare Quality® Consortium, 2009. Welfare Quality® Assessment Pro-

Zurbrigg, K., Blackwell, T., 2006. Injuries, lameness, and cleanliness of sowsin four group-housing gestation facilities in Ontario. Swine HealthProd. 14 (4), 202–206.