Influence of raw material on the occurrence of tail-biting in undocked pigs
2. Materials and methods 1. General aspects
The pigs in the present study were kept on the research farm of the Chamber of Agriculture of Schleswig-Holstein (Futterkamp), Germany, in accordance with EU Directive (2008/ 120/
EG) and in accordance with the Tierschutz-Nutztierhaltungsverordnung (TierSchNutztV, 2006). In case of tail-biting outbreaks, manipulable material was provided in control pens as well to avoid endangerment of animal welfare.
2.2. Animals and housing
Data collection was carried out between September 2013 and April 2014. Farrowing and rearing followed conventional farming practices; the farm size was 400 sows and 2,500 rearing places. In the present study, 721 crossbreed piglets (Pietrain x (Large White x Landrace)) from 60 litters were housed in ten batches. Each batch corresponded to a farrowing week. The piglets had an average birth weight of 1.4 ± 0.3 kg. The suckling period took place in conventional farrowing systems (5.2 m² per pen), tails were not docked and males were not castrated. From the second week of life until weaning, the piglets received a pre-starter diet (14.6 MJ ME, 17.5 % protein, 1.45 % lysine, 0.25 % sodium). The piglets were weaned with on average 28 days with an average weaning weight of 8.0 ± 1.7 kg.
Rearing lasted for 40 days until an average weight of 25.4 ± 2.3 kg. The piglets were housed in mixed gender groups consisting of one or two litters (12 or 24 piglets per pen) with an average space allowance of 0.38 or 0.42 m² per animal. According to the units, the feeding system was either mash or dry feed ad libitum with an animal to feeding place ratio of 2:1.
For the first two weeks of rearing the piglets received a starter diet (14.4 MJ ME, 18.0 % protein, 1.40 % lysine, 0.20 % sodium), thereafter the diet was gradually changed over the next four days and fed until day 40 of rearing (13.4 MJ ME, 17.0 % protein, 1.30 % lysine, 0.25 % sodium). The drinking system consisted of nipples and bowls, the floor was fully slatted and no bedding material was offered. Plastic sticks, plastic balls and hard wooden sticks were provided as enrichment material. The environmental temperature during rearing was automatically regulated by forced ventilation. It was set at 29.5 °C on day one of rearing and decreased stepwise until 22.0 °C on day 40. The animals had full artificial lighting between 06:00 h and 18:00 h.
2.3. Experimental design
In total, 721 piglets were divided randomly into three groups litter-wise: a control group (CG) with 231 long-tailed piglets housed without raw material, a dried corn silage group (SG) and an alfalfa hay group (AG) with 245 long-tailed piglets each. In the farrowing units, 20 litters were used for each treatment, two litters of each treatment group (n = 3) per batch (n = 10) respectively. After weaning, the piglets were housed either litter-wise or two litters were mixed, resulting in 14 pens for each treatment with two different group sizes in the rearing units (12 or 24 piglets per pen). A schematic view of the experimental set-up during rearing is given in Figure 1. Within each of the ten batches, the number of CG, SG, and AG pens was balanced and the locations of the treatment groups within the units were randomised.
Figure 1: Schematic view of the experimental set-up regarding rearing.
2.4. Treatments
The provision of raw material took place from the second week of life until the end of rearing twice a day (in the morning and in the afternoon) in the piglet nest (farrowing) or in a piglet bowl (Ø 40 cm) with an animal to occupation place ratio of 1.2:1 (12 piglets/ pen) or 2.4:1 (24 piglets/ pen) during rearing.
The amount of dried corn silage offered per day and pen was about 100 g, the amount of alfalfa hay about 120 g/ day/ pen, which corresponds to a handful of material per offer. In pens affected by tail-biting, the intervention scheme was a jute sack (fixed on the pen wall), a handful of long, chopped straw on the pen floor, grass silage or straw-peat mixture provided in piglet bowls. In severe cases, identified biters were removed from the pen. Treatment with intervention material was also applied in CG pens if an outbreak occurred.
2.5. Data collection 2.5.1. Scoring
Scoring of tail lesions and tail losses took place weekly during farrowing and rearing. The scoring scheme (modified from Abriel and Jais, 2013) classified the severity of tail lesions with a four-point score consisting of “no visible damage” (0), “scratches, light bite marks”
(1), “moderate damage” (2) and “severe damage” (3). A tail-biting outbreak was defined as a point in time when at least one piglet showed a freshly bleeding tail wound or a loss of the
Total number
tail. Tail losses were classified by “original length of tail” (0), “loss of tail tip” (1), “partial loss” (2) and “total loss” (3). Furthermore, the gender and the size of the animals (small, medium, large, in relation to pen mates) were recorded.
2.5.2. Weight gain
Weight was collected at pen level in the beginning and in the end of rearing. Due to data transmission problems group weights of only six batches (27 pens) were available. The pens were equally distributed over the treatment groups.
2.5.3. Video surveillance
To investigate the activity behaviour of the animals and the duration of occupation with the raw material, three farrowing units and three rearing units were equipped with colour cameras (Santec, VTC-249IRP/ W or VTC-279/ IRPWD). In total, 99 piglets during farrowing (five pens of AGs, three pens of SGs) and 188 piglets during rearing (four pens for each treatment group) were video recorded 24 hours every day. The HeitelPlayer software (Xtralis Headquarter D-A-CH, HeiTel Digital Video GmbH, Kiel, Germany) was used to watch the videos. Six different behavioural patterns during farrowing and ten different behavioural patterns during rearing were coded. The ethogram is given in Table 1. The time budgets of each animal were investigated by instantaneous scan sampling with a 10 min interval from 06:00 h to 18:00 h. Additionally, a one-minute sampling interval was used for the first 10 minutes after raw material provision to determine the number of piglets occupied by the material.
Table 1: Ethogram used for video observation in the 10 min sampling frame.
2.6. Statistical procedures
The software package SAS 9.2® was used for statistical analysis (SAS, 2008). The fit statistics AICC “Akaike’s information criterion corrected” (Hurvich and Tsai, 1989) and the BIC “Bayesian information criterion” (Schwarz, 1978) were used to evaluate the fitting of the models. Fixed effects were added stepwise to the models. The model with the smallest AICC and BIC was chosen for the analysis.
Behaviour Description
Lying Lying on the side or ventrally
Sitting Body supported by hind-quarters and stretched front legs Standing Body supported by four stretched legs, includes
locomotion
Feeding Head positioned in the feeder
Occupation with raw material Sniffing, nosing or rooting the raw material in piglet bowl/ nest
Occupation with toys Head on the toys Additionally observed during
rearing
Tail exploration Head on the back side of a pen mate, includes tail-in-mouth and tail-biting behaviour
Drinking Head on the water nipple
Belly nosing Manipulation of the abdomen of a lying pen mate Occupation with intervention
material
Head on the additional offered material (in case of tail-biting outbreaks)
2.6.1. Tail lesions and tail losses
The data of tail lesions and tail losses followed a multinomial distribution (Score 0-3).
Therefore, the procedure GLIMMIX was used assuming a cumulative logit link function for the multinomial distributed data. Due to the fact that tail-biting did not occur during farrowing, the data for the statistical analysis was limited to the rearing period. The experimental unit was the pen. The fixed effects group (CG, SG, AG), batch (1-10), week after weaning (1-6) and the interaction of group and batch were used in the final model for tail lesions. The pen (nested in batch) was included as a random effect. Only the last observation at the end of rearing was taken into consideration regarding tail losses. The group (CG, SG, AG) and the batch (1-10) were used as fixed effects in the final model for tail losses.
2.6.2. Weight gain
In order to estimate the effect of tail-biting on total weight gain during rearing the MIXED procedure was used. For every treatment group and scoring scheme (tail lesions and tail losses) nine pens were ranked regarding the total number of piglets with a score higher than Score 0 within each pen, resulting in three pens for a low, respectively medium and high level of tail lesions and tail losses (only last observation was taken into consideration). The level of tail lesions/ tail losses (low, medium, high), the treatment group (CG, SG, AG) and the interaction between level of tail lesions/ tail losses and treatment group were used as fixed effects. Significant differences in the least-square-means were adjusted with the Bonferroni-correction (p < 0.05) (Westfall et al., 2011).
2.6.3. Video analysis
The trait occupation with the raw material provided was investigated with two models (farrowing, rearing). Occupation was coded as a binary trait (0: not occupied, 1: occupied).
The procedure GLIMMIX was used with the link function logit. The fixed effects group (SG, AG), batch (1-3), day after first raw material provision (1-16) and the interaction between group and day after first raw material provision were used in the final model for farrowing.
The fixed effects group (SG, AG), batch (1-3), day after weaning (1-28) and the interaction between group and day after weaning were used in the final model for rearing. The pen (nested within batch) was included as a random effect in both models.
The trait activity behaviour was analysed with one model. Due to the low frequencies of the behavioural patterns occupation with raw material and toys, tail exploration, drinking and belly nosing, they were summarised as “active”, together with sitting, standing and feeding.
Lying behaviour is counted as “inactive” in further analysis. The activity behaviour was coded as a binary trait, the procedure GLIMMIX was used with the link function logit. The fixed effects group (CG, SG, AG), batch (1-3), day after weaning (1-28), daytime (6-18h), the interaction of group and day after weaning, as well as the interaction of batch and day after weaning were used in the final model. The pen was included as a random effect.
3. Results
3.1. Tail lesions and tail losses
The effect of group, batch, week after weaning and the interaction of group and batch had highly significant influences on tail lesions (p < 0.001). Tail-biting started on average two to three weeks after weaning with a tail lesion score higher than 0 in 24.3 % of the piglets in the third week (Fig. 2) and occurred in all pens.
Figure 2: Estimated frequencies of tail lesions over six weeks after weaning.
100 80 60 40 20 0
1 2 3 4 5 6
No lesions (0) Small lesions (1)
Moderate lesions (2) Severe lesions (3)
Estimated frequencies of tail lesions
Week after weaning Tail lesions
In the fifth week after weaning tail lesions reached a peak (73.7 % of the piglets with scores higher than 0) and decreased approximately 8.5 % until the last week of rearing. The first tail losses occurred one to two weeks after the first tail lesions had become visible.
The effect of batch and group had highly significant influences on tail losses at the end of rearing (p < 0.001). The highest number of tail losses (Score > 0) occurred in batch one (96.4 %) and batch five (84.4 %), whereas the lowest number of tail losses was documented in batch nine (9.4 %) and batch ten (7.4 %). Tail losses in other batches ranged from 63 % to 20.9 % (Fig. 3).
Figure 3: Estimated frequencies of tail losses over ten batches at the end of rearing.
At the end of rearing, piglets had lost their tails to the highest extent in CGs (50.4 %), followed by AGs (49.2 %) and SGs (30.2 %), (Fig. 4).
100 80 60 40 20 0
1 2 3 4 5 6 7 8 9 10
Original length (0) Loss of tail tip (1)
Partial loss (2) Total loss (3)
Estimated frequencies of tail losses
Tail losses Batch
Figure 4: Estimated frequencies of tail losses comparing the treatment groups at the end of rearing.
3.2. Weight gain
The level of tail lesions and tail losses and the treatment group had no significant influences on the total weight gain during rearing (p > 0.05). The interaction of the number of tail lesions and treatment group, as well as the interaction of number of tail losses and treatment group had significant effects on the total weight gain during rearing (p < 0.05). Piglets in CGs gained less weight in pens with high levels of tail lesions/ tail losses (LS-Mean ± se: 16.9 ± 1.0 kg/ 15.6 ± 1.0 kg) in comparison to pens with a low level (19.9 ± 1.0 kg / 18.9 ± 1.0 kg).
The total weight gain of piglets in SGs was comparable. Piglets in AGs gained less weight in medium levels of tail lesions/ tail losses (14.8 ± 1.0 kg) in comparison to pens with low and high levels (18.5 ± 1.0 kg vs. 16.5 ± 1.0 kg). Differences in total weight gain between the treatment groups within the levels of tail lesions and tail losses were not significant (p > 0.05).
3.3. Video analysis
Concerning farrowing, the effects of group and batch, as well as the interaction of group and batch had no significant influences on the occupation with raw material (p > 0.05). The day after the first raw material provision had a highly significant influence on the number of
100 80 60 40 20 0
Control Alfalfa hay Corn silage Original length (0) Loss of tail tip (1)
Partial loss (2) Total loss (3)
Tail losses
Estimated frequencies of tail losses
piglets occupied with the raw material (p < 0.001). No clear trend was visible in the course of farrowing. During rearing, the group, the day after weaning and the interaction of group and day after weaning had a significant effect on the occupation with raw material (p < 0.01). The batch had no significant effect (p > 0.05). Corn silage attracted the piglets’ attention during the whole observation period, whereas the attractiveness of the alfalfa hay decreased towards the end of rearing (Fig. 5).
Figure 5: Least-square-means of the occupation with raw material during ten minutes after provision regarding the interaction of group and day after first raw material provision in three batches.
The batch, the day after weaning, the daytime, the interaction of group and day after weaning, as well as the interaction of batch and day after weaning had highly significant influences (p < 0.001) on the overall activity behaviour during rearing. Group had no significant influences on the overall activity (p > 0.05). The data showed a two-phase activity curve with a peak between 10:00-11:00 h and 14:00-15:00 h. The activity of the piglets was highest on the first day after weaning, as well as between two and three weeks after weaning and decreased towards the end of rearing (Fig. 6). The activity behaviour between the batches followed no clear trend during the observation period.
Estimated percentage of occupied animals
Day after first raw material provision
Farrowing Rearing
Figure 6: Least-square-means of the overall activity during rearing regarding the interaction of group and day after weaning in three batches.
4. Discussion
4.1. Tail lesions and tail losses
No tail-biting was observed in none of the treatment groups during farrowing. This finding is in line with Munsterhjelm et al. (2009), who did not observe tail lesions either in enriched or in barren pens in the first four weeks of life. Tail-biting occurred on average in the second until third week after weaning for the first time, followed by tail losses one until two weeks later. This finding contributes to the work of Abriel and Jais (2013), who found increasing tail-biting behaviour in the second week after weaning. An explanation for the beginning of tail-biting in the early rearing phase could be the number of conversions the piglets are faced with during the weaning process. When separated from the sow, they need to adapt to a new environment and feeding, at the same time their immune system is forced to deal with a new germ environment. Under “natural conditions” weaning is a gradual process in piglets and is not complete until 10–12 weeks of age (Lallès et al., 2007). Furthermore, the mixing of piglets and therefore, rank order fights contribute to a stressful situation for weaned piglets (Hötzel et al., 2011). In intensive housing systems, however, the animals often fail to change
Estimated percentage of active animals
Day after weaning
aversive situations by using evolved coping strategies, and it is argued that abnormal behaviour can originate from unsuccessful coping behaviour (Wechsler, 1995).
The interaction of group and batch had a highly significant influence on tail lesions.
Nevertheless, there was no clear trend between the treatment groups and the batches. The variation in the amount of tail lesions supports the assumption that environmental conditions (e.g. climate, pen structure, group size, feeding spaces, ventilation and health statues) play an important role in the occurrence of behavioural disorders (Taylor et al., 2012). The group sizes and the feeding and drinking systems differed between batches in the present study. No clear connection between group sizes and unit facilities could be drawn for the levels of tail lesions between the batches. Daily raw material provision seemed to play a subordinated role in the occurrence of tail lesions, a finding which is not in line with (Zonderland et al., 2008), who concluded that tail-biting is best prevented with a small amount of straw (20 g/ animal/
day), provided twice daily.
Furthermore, the effect of group and batch had highly significant influences on tail losses at the end of rearing. The decreasing number of tail losses at the end of the study could be explained by enhanced and more precise animal observation by stable staff and points out the learning process in the course of the batches. The staff members reacted faster in case of tail-biting outbreaks, using tail lesions as an indicator, and offered for example jute sacks or a straw-peat-mixture to the pigs as additional occupation material. Thus, tail lesions were able to heal again and did not result in tail losses. In addition to that, the European Food Safety Authority pointed out good stockman ship and intervention before severe outbreaks become established as useful (EFSA, 2007). The behavioural disorder occurred to a smaller extent in SGs, but differences between CGs and AGs were not significant. Curative measurements were also carried out in CGs to avoid severe injuries and welfare problems in the case of tail-biting outbreaks. Thus, CGs were falsified, which could have led to an approximation to the raw material groups. Tail-biting in CGs might have provoked higher numbers of tail losses if intervention had not been carried out. According to Petersen et al. (1995), pigs from barren environments had higher frequencies of biting floor and walls, nudging and tail-biting litter mates than piglets in enriched conditions. Nevertheless, the additional provision of raw material in piglet bowls in SGs and AGs could not prevent tail-biting in the present study.
4.2. Weight gain
The level of tail lesions and tail losses in dependence on the treatment group influenced the total weight gain of the piglets during rearing. The results of CGs and SGs were in line with Camerlink et al. (2012), who found that pigs that received more tail-biting, ear-biting and paw-biting, grew less well (p < 0.05). Moreover, Wallenbeck and Keeling (2013) stated that tail-biting victims had decreased daily feed intakes during and after the tail-biting outbreaks.
Nevertheless, the results of AGs were in contrast to this and the limited weight data (27 pens) made it difficult to draw general conclusions.
4.3. Video analysis
Corn silage stayed attractive during the whole observation period. A possible explanation could be a better palatability of the material due to a higher concentration of carbohydrates and lower dry-matter content in comparison to alfalfa hay, which contains more fibres. The preference of pigs for roughage with a low dry-matter content and the attractiveness of glucose/ sucrose for pigs was shown by Olsen et al. (2000) and Kennedy and Baldwin (1972).
The higher acceptance of corn silage and therefore sustainable occupation could have led to lower tail losses in SGs in comparison to AGs and CGs at the end of rearing (see Fig. 4). In the present study, a fresh replacement twice a day should remain the raw material of interest to the animals because of the novelty aspects (Van de Weerd et al., 2003; Wood-Gush and Vestergaard, 1991). However, provision in the piglet bowl did not allow every piglet to reach the raw material at the same time, which could have forced the development of the behavioural disorder. According to Van de Weerd et al. (2006), the limited size of a point source may restrict access to enrichment causing competition, aggression or restlessness in groups of animals. Especially in larger group sizes (> 12 piglets), more than one piglet bowl
The higher acceptance of corn silage and therefore sustainable occupation could have led to lower tail losses in SGs in comparison to AGs and CGs at the end of rearing (see Fig. 4). In the present study, a fresh replacement twice a day should remain the raw material of interest to the animals because of the novelty aspects (Van de Weerd et al., 2003; Wood-Gush and Vestergaard, 1991). However, provision in the piglet bowl did not allow every piglet to reach the raw material at the same time, which could have forced the development of the behavioural disorder. According to Van de Weerd et al. (2006), the limited size of a point source may restrict access to enrichment causing competition, aggression or restlessness in groups of animals. Especially in larger group sizes (> 12 piglets), more than one piglet bowl