Inclusion of grass/clover silage in the diet of growing/finishing pigs – Influence on pig time budgets and social behaviour

Abstract

This study investigated how inclusion of grass/clover silage affected the time budgets and social behaviour of growing/finishing pigs. The behaviour of 128 pigs was observed through video recordings. Pigs fed intact or chopped silage spent a larger proportion of their time active compared with pigs fed silage in pelleted form or fed only cereal-based feed. There was a trend for pigs fed silage in pelleted form or fed only cereal-based feed to respond strongly to social interactions more often than pigs fed intact or chopped silage, irrespective of initial performing pig behaviour. Pigs in groups fed intact silage had a lower number of wounds from violent social interactions on their bodies. Thus, additional provision of silage in an environment enriched with straw can further improve pig welfare.

Keywords:

• Swine
• roughage
• time budget
• social interactions
• wounds

Introduction

In pig production, provision of straw as a substrate for performance of species-specific behaviours such as foraging and exploration is considered to be one of the most efficient ways to reduce the occurrence of injurious and potentially harmful behaviours, e.g. ear and tail biting (EFSA, 2007). However, additional access to roughage can increase the time pigs spend eating and provide them with better opportunities to perform foraging and exploration (Olsen et al., 2000).

Due to the ability of legumes to fix nitrogen from the atmosphere, a mixed ley crop is often included in the crop rotation. When ley forage is available on the farm, it can be used as a locally produced feed resource for pigs (Webrink, 1994; Sehested et al., 2004; Heyer et al., 2006; Høøk Presto et al., 2011). Besides adding to the nutrient supply of the pigs, grass/clover silage contributes to sustainable use of arable land without mineral fertilizers and may also improve pig welfare (Presto et al., 2009).

In enriched housing or outdoor housing systems, pigs spend a large proportion of their time rooting and foraging in different substrates (Beattie et al., 2000; Scott et al., 2006; Morrison et al., 2007). Pigs housed in conventional pens have limited possibilities for rooting and foraging and it has been suggested that pigs sometimes redirect their foraging and exploration behaviour towards pen fittings and other pigs (e.g. Vanputte, 1969; EFSA, 2007). In fact, pigs in barren environments are more involved in nosing, biting and aggressive encounters with other pigs than pigs housed in enriched environments (Lyons et al., 1995; Beattie et al., 2000).

Additional access to roughage increases the time pigs spend eating and increases their opportunities to perform foraging and exploration. However, the form in which the roughage is fed may influence pig behaviour. Therefore, the aim of the present study was to investigate how grass/clover silage (1) fed as intact silage separately from commercial cereal-based feed, (2) chopped and mixed with the commercial feed, or (3) mixed together with the commercial feed and pelleted, affect pig time budgets and social behaviour.

Materials and methods

The experiment was performed at the Research Station of the Swedish University of Agricultural Sciences at Funbo Lövsta outside Uppsala (latitude 60° north) from autumn to spring (September 2010 to May 2011). The experiment was performed according to Swedish regulations concerning animal use in experiments (Ministry for Rural Affairs, 1988).

Animals and housing

In total, 128 growing/finishing pigs (Swedish Yorkshire×Hampshire) from two production batches (batch one from September 2010 to January 2011; batch two from January 2011 to May 2011) were included in the experiment from approximately 30 kg live weight [30.7±4.55 kg and 70.7±2.89 days of age (Mean±SD)] until slaughter, which occurred at an approximate live weight of 110 kg [108.4±8.31 kg and 167.2±7.62 days of age (Mean±SD)]. At 30 kg live weight, pigs from 10 litters (birth litters, cross-fostering was not applied) of each of the two production batches were allocated to eight pens with eight pigs per pen, i.e. in total 32 pigs per treatment at the start of the experiment. The distribution of pigs was balanced with regard to litter and sex, e.g. no siblings in the same pen and four gilts and four castrates per pen. The pens had a concrete floor in the feeding and lying area, and a slatted dunging area at the back of the pen (1/3 of pen area). The total area was 14.8 m², giving a floor area of 1.9 m² per pig. The partitions between the pens were made of metal bars in the dunging area and solid walls in the eating and lying area. A 4.5 m long feed trough was provided along the front of the pen and four water nipples were placed in the slatted area. All pens were provided with approximately 1 kg straw per day.

Pigs were fed either commercial cereal-based feed + chopped silage mixed and fed together (SM); commercial feed + intact silage fed separately (SS); commercial feed + milled silage, mixed and pelleted (SP); or commercial feed alone (C), with four pens (pig groups) per treatment. The pigs were fed twice daily (7 am and 3 pm) according to Swedish energy recommendations (Simonsson, 2006). The silage used in the SM, SS and SP treatments contained 9.04 MJ metabolizable energy (ME) and 129 g crude protein (CP) per kg dry matter (DM) and constituted 20% of the diet on an ME basis. Representative samples of feed were collected and analyses for DM, CP, ash, crude fat, neutral detergent fibre (NDF) and gross energy and are presented in Table I.

Table I. Chemical composition of feeds and feedstuffs.

	Silage	SP^a feed	Commercial feed
Dry matter (DM, %)	31.5	89.4	90.6
Ash (g/kg DM)	100	62	54
Crude protein (CP, g/kg DM)	156	147	153
Crude fat (CF, g/kg DM)	28	35	31
Neutral detergent fibre (NDF, g/kg DM)	456	220	144
Gross energy (GE, MJ/kg DM)	18.6	17.8	–
Metabolizable energy (ME, MJ/kg DM)^b	9.3	12.8	13.7
^aCommercial feed + milled silage, mixed and pelleted.
^bEstimated according to Lindberg and Andersson (1998), where energy digestibility (dE, %) = 94.8+(–0.93×NDF%). Digestible energy (DE) = dE×GE and ME = 0.95×DE, ME values in commercial feed.

Of the 128 pigs entering the experiment, five were culled before slaughter. One pig in the SM treatment in batch one, one pig in the SP treatment in batch two and one pig in the C treatment in batch two were culled in the first week of the experiment due to injuries caused by fighting at mixing. Moreover, two pigs in the SM treatment (one in batch one and one in batch two) were culled in weeks six and eight of the experiment due to lameness.

Behaviour recording

The behaviour of the pigs was observed in video recordings from four occasions (48 hours per observation occasion) at approximately 30, 50, 70 and 100 kg live weight [corresponding to 79.5±3.66, 110.9±8.90, 128.7±3.41 and 149.7±3.41 days of age at the start of each respective observation occasion (Mean±SD)]. To estimate the time budget of the pigs, the number of pigs in the pen fulfilling the definitions of the variables for activity, body posture and location in the pen described in Table II were recorded with instantaneous scan sampling every 30 minutes. The video cameras were sensitive to infra-red light, thus scan samplings were performed day and night for 48 hours per observation occasion. Social interactions were recorded continuously from 15 to 5 minutes before and from 5 to 15 minutes after each feeding event (twice per day, am and pm). For each social interaction, the location in the pen and the behaviour of the performing and receiving pig were recorded according to the definitions presented in Table II. After a pause in an interaction for at least 30 seconds, a new approach from the performing pig was considered a new social interaction.

Table II. Ethogram of behaviours of pigs in the video recordings.

Category	Variable	Definition
Body posture	Lie	Lying on the side or sternum, straight or bent legs
	Sit	Front feet on the ground, back legs in lying position
	Stand	Standing on all four feet or walking
Activity	Eat	Snout in feed trough or silage rack
	Drink	Snout touching water nipple
	Nose/bite other pig	Snout touching other pig
	Nose/bite pen fittings	Snout touching pen fitting
	Nose/bite pen fitting	Snout touching pen floor
	Nothing	Snout in air
Location in the pen	Lying area	In the lying area
	In front of feed trough	At least one hoof/leg within 20 cm from the feed trough
	Slatted area	At least one hoof on the slatted area
Social interactions – performing pig	Nose	Snout touching other pig
	Bite	Biting another pig
	Head knock	Approaching other pig with rapid head movement and open mouth
	Lift	Snout on or under the body of another pig and lifting upwards
	Climb	At least one hoof/leg on top of another pig
Social interactions – receiving pig	No reaction	No change in body position or activity of the receiving pig
	Avoid	Head turned away or pig moving away from the performing pig
	Return approach	Receiving pig approaching the performing pig with head/snout
	Other pen	Interacting with pig from other pen, no reaction recorded

Recording of wounds and cleanliness

The number of wounds and the cleanliness of the pigs and the pen floor were recorded once, approximately one week before the first slaughter event. For each individual pig, the numbers of fresh and old wounds (partly healed or visual scars) were counted on the head, middle and hind part of the body as indicated in Figure 1. In addition, the dirtiness of each body part was scored according to a three-point scale, where one means that the whole body part was clean, two that half the body part was contaminated with faeces and three that the whole body part was contaminated with faeces. The cleanliness of the laying area of the pen was scored as the number of clean and dry quarters of the pen floor (1–4). All wound and cleanliness scorings were performed by the same person. Before data analyses, average number of wounds and cleanliness score per pig was estimated for each pen.

Figure 1.  Body sections on which wound numbers and pig cleanliness were recorded: head, middle of body and hind part.

Statistical analyses

Statistical analyses were performed using SAS software, version 9.2 (SAS, 1990). Descriptive statistics were calculated using procedure FREQ.

Differences between treatments in the time budgets of the pigs were analysed with ANOVA using procedure MIXED. Average percentage of time spent lying, eating, nosing or biting other pigs, nosing or biting pen fittings or floor and in the different functional areas of the pen (in front of feed trough, lying area and slatted dunging area) were estimated per pen and observation occasion. Pen averages were analysed with a model including the fixed effects of treatment (SM, SS, SP and C), batch (one and two), observation occasion (four occasions at approximately 30, 50, 70 and 100 kg of live weight), pen nested within treatment and batch (16 pens), the interaction between treatment and batch and the interaction between treatment and observation occasion. Observation occasion was included with REPEATED statement to model an unstructured covariance structure (choice based on smallest AIC-statistics) within pen. Pair-wise comparisons were adjusted for multiple comparisons according to the Tukey method.

For each pen and feeding event (am and pm at each observation occasion), the average number of social interactions per pig before and after feeding was estimated (expressed as average number per pig per 10 minutes of observation with pen as experimental unit). Differences between treatments were analysed with ANOVA using procedure MIXED. The model included the fixed effects of treatment, batch, observation occasion, time of feeding (am and pm), pen nested within treatment and batch, the interaction between treatment and batch, the interaction between treatment and observation occasion and the interaction between treatment and time of feeding. Observation occasion was included with REPEATED statement to model a compound symmetry covariance structure (choice based on smallest AIC-statistics) within pen. Pair-wise comparisons were adjusted for multiple comparisons according to the Tukey method.

Differences between treatments in average number of wounds and dirtiness scores per pig per pen (recorded once right before slaughter, pen as experimental unit) were estimated with ANOVA using procedure GLM and a model including the fixed effects of treatment and batch. The number of pigs in the pen at the time of scoring was included as a continuous covariate.

In all models where pen was included, the effect of pen included both the effect of the social environment in the unique pig group kept in the pen and the effect of the environment in the physical pen.

Residuals of all dependent production parameters in the ANOVA analyses were examined for normal distribution using PROC UNIVARIATE, considering the Shapiro–Wilks test for normality and a normal probability plot, and all were found to be normally distributed.

Results

Time budgets

The pigs spent on average 76.5% of their time lying down and during 91.4% of that time they were inactive. Pigs fed pelleted feed (SP and C treatments) spent the largest proportion of their time in lying position and pigs fed chopped or intact silage (SM and SS treatments) the smallest proportion [72.6^a, 74.8^a, 79.3^b and 79.2^b% of time in lying position in treatments SM, SS, SP and C, respectively (least square means), standard error = 0.92, p=0.002, F=13.4, N=64]. Differences in pig time budgets between treatments are presented in Figure 2. Pigs in the SS treatment spent a larger proportion of their time eating (p<0.003, F=10.9, N=64) and pigs in the SM treatment spent a larger proportion of their time nosing or biting the pen floor than pigs in the other treatments (p<0.004, F=10.3, N=64).

Figure 2.  Proportion (%) of time (least square means±standard error) that pigs in treatments SM, SS, SP and C spent eating, nosing/biting the pen floor, nosing/biting pen fittings and nosing/biting other pigs. Different letters (a, b) for different treatments indicate pair-wise differences at p<0.1. N=64.

Differences in pigs' location in the pen between treatments are presented in Figure 3. Pigs in the SM and C treatments spent more time in front of the feed trough than pigs in the SS treatment (p<0.032, F=4.9, N=64). Instead, the pigs in the SS treatment spent more time in the lying area (p<0.011, F=7.33, N=64) with a larger proportion of time eating there, as indicated in Table III. In front of the feed trough, SM pigs spent a larger proportion of time nosing or biting the pen floor compared with pigs in the other treatments. Compared with SP pigs, SM pigs tended to spend a smaller proportion of their time in the slatted area of the pen (p<0.052, F=4.0, N=64) as shown in Figure 3.

Figure 3.  Proportion (%) of time (least square means±standard error) that pigs in treatments SM, SS, SP and C spent located in front of the feed trough, in the lying area and in the slatted area. Different letters (a, b) for different treatments indicate pair-wise differences at p<0.1. N=64.

Table III. Proportion (%) of time spent eating, nosing/biting pen floor, nosing/biting pen fittings, nosing/biting other pigs, in front of the feed trough, in the lying area and in the slatted area, per treatment.

Pen location, per cent of time	SM (%)	SS (%)	SP (%)	C (%)
Eating
In front of feed trough	99.9	48.7	100.0	100.0
Laying area	0.0	51.3	0.0	0.0
Slatted area	0.1	0.0	0.0	0.0
Nosing/biting pen floor
In front of feed trough	40.3	18.9	18.3	21.7
Laying area	56.9	76.4	76.5	72.3
Slatted area	2.8	4.7	5.2	6.0
Nosing/biting pen fittings
In front of feed trough	34.0	23.4	14.2	22.3
Laying area	9.2	22.3	23.6	16.1
Slatted area	56.8	54.3	62.2	61.6
Nosing/biting other pigs
In front of feed trough	11.0	6.9	8.6	10.9
Laying area	85.4	87.1	78.9	83.7
Slatted area	3.6	6.0	12.5	5.4

Observation occasion (at 30, 50, 70 and 100 kg of live weight) significantly (p<0.05) affected the time budgets of the pigs, indicating that the proportion of time spent eating and nosing other pigs decreased with age. Moreover, pigs spent more time lying down and nosing pen fittings at 50 kg compared with the other three observation occasions. The proportion of time spent in the slatted area and in front of the feed trough increased with age, while time spent in the lying area of the pen decreased. Production batch had a significant effect on time spent eating, lying down, nosing pen fittings, nosing other pigs and pig location in the pen. There was a significant interaction between treatment and observation occasion for proportion of time pigs spent lying and between treatment and batch for proportion of time pigs spent eating, lying, nosing pen fittings and time spent in front of the feed trough and in the lying area.

Social interactions

Irrespective of treatment, pigs performed a higher number of social interactions in the period 15 to 5 minutes before feeding than in the period 5 to 15 minutes after feeding as indicated in Figure 4a and b. The most frequent social interaction performed before feeding was nosing other pigs and the most frequent social interaction performed after feeding was lifting other pigs. More detailed analyses of the different behaviours performed before feeding (presented in Figure 4a) showed that pigs in the SP treatment tended to lift other pigs more frequently than pigs in the other treatments (p<0.052, F=4.01, N=128). Pair-wise differences also indicate that pigs in the SP treatment nosed other pigs more frequent than pigs in the SS and C treatment and bit or head knocked other pigs more frequent than pigs in the SM treatment, even though the overall effect of treatment was not significant. After feeding however (presented in Figure 4b), lifting of other pigs was instead performed more frequently by SM pigs (p<0.083, F=3.2, N=123), and pigs in the SS treatment showed a lower frequency of nosing other pigs (p<0.068, F=3.5, N=123) than those in the other treatments. After feeding, pigs in the SM and SP treatments displayed a higher frequency of climbing other pigs (p<0.012, F=7.2, N=123) compared with pigs in the other two treatments.

Figure 4.  Number of nose, lift, bite/head knock and climb interactions (least square means±standard error) per pig in the period: (a) 15–5 minutes before and (b) 5–15 minutes after feeding in treatments SM, SS, SP and C. Different letters (a, b) for different treatments indicate pair-wise differences at p<0.1. N=123–128.

Batch affected the frequency of performed bites/head knocks before feeding (p<0.1) and the frequency of lifting other pigs before feeding (p<0.05). Observation occasion affected the frequency of nosing other pigs before and after feeding, bite/head knock and lifting other pigs after feeding and climbing other pigs before feeding (p<0.05 for all). There were significant interactions between treatment and batch for frequency of climbing other pigs after feeding (p<0.05) and between treatment and observation occasion bite/head knock (p<0.05) and climbing other pigs (p<0.1) after feeding.

As presented in Figure 5, receiving pigs reacted less commonly when nosed by a performing pig than when climbed on, lifted, bitten or head-knocked by a performing pig. Although there were no major differences between treatments in the reactions of the receiving pigs, there was a trend for pigs in the SP and C treatments to exhibit stronger reactions (return approach to social interactions) more frequently than SM and SS pigs, irrespective of the behaviour of the performing pig.

Figure 5.  Reaction of receiving pigs: proportion (%) of social interactions per treatment (SM, SS, SP and C) with each specific reaction (return approach, avoid, no reaction and receiving pig in other pen) when the performing pig nosed, lifted, bit/head-knocked or climbed on the receiving pig. N=123–128.

Wounds and dirtiness

Pigs in the SS treatment tended to have a lower total number of wounds on their bodies at the end of the growing/finishing period than pigs in the SM and C treatments [10.5^a, 5.72^b, 8.74^ab and 13.2^a wounds per pig in treatments SM, SS, SP and C, respectively (least square means), standard error = 1.74, p<0.068, F=3.25, N=16]. More detailed investigations presented in Figure 6 showed that the pigs in the SS treatment had the lowest average number of wounds on the head (p<0.2177, F=1.76, N=16), middle (p<0.036, F=4.22, N=16) and hind part of the body (p<0.105, F=2.67, N=16) and that the pigs in the C treatment had the highest number of wounds on the middle and hind part of the body. The pigs had on average a cleanliness score of 1.2±0.20 (mean±standard deviation) per body part and the pens had on average 2.5±0.63 clean and dry quarters of the concrete floor. There were no significant differences between treatments in the cleanliness of either the pigs or the pen floors.

Figure 6.  Number of wounds per pig (fresh and old) on the head, middle and hind part of the body in the SM, SS, SP and C treatments (least square means±standard error). Different letters (a, b) for different treatments indicate pair-wise differences at p<0.1. N=16.

There were significant effects (p<0.05) of batch and number of pigs in the pen on total number of wounds and numbers of wounds on the head, middle and hind part of the body. There was a higher number of wounds in batch two compared with batch one and an increase in number of wounds with an increasing number of pigs in the pen.

Discussion

Under feral conditions, pigs spend most of their day foraging and exploring the environment in their search for feed (Studnitz et al., 2007). However, in commercial pig production the pigs are usually fed a restricted amount of a concentrated diet twice or three times per day. In this experiment the pigs fed silage, either chopped and mixed with commercial feed, or intact and fed separately from the commercial feed were more active and spent a larger proportion of their time foraging than the pigs fed the pelleted feed. This was despite the fact that the SP diet had the same ingredients as the SM and SS diets. However, in the SP treatment the roughage was dried, milled and pelleted together with the commercial feed. Pigs kept in environments enriched with straw are reported to be more active than pigs in barren environments (Beattie et al., 2000). The results from the present study show that further enrichment with grass/clover silage stimulates pigs to be even more active (i.e. spend less time lying down).

Pigs lacking opportunities to forage or explore their surroundings may redirect these behaviours towards other objects, usually their pen mates and the pen equipment (Lyons et al., 1995; Beattie et al., 2000). In the present experiment, the pigs fed only pellets spent on average a larger proportion of their time nosing and chewing pen fittings and nosing and biting other pigs than the pigs fed wet silage which could indicate that increased possibilities to perform foraging behaviour can lower the frequency of manipulation of other pigs. Even though this trend was clear, the difference was not statistically significant.

Pigs fed chopped silage mixed with commercial feed spent approximately twice as much time nosing and biting the floor in front of the feed trough compared to pigs in the other treatments (Table III). This might be because these pigs moved a lot of their feed out onto the floor where they manipulated the feed mix, probably sorting out and consuming the most desirable parts. The pigs fed intact silage separately spent more time in the lying area than those in the other treatments. In these pens, the feed-rack was placed on the side wall between the pens, thus the pigs were probably sorting and consuming the silage that fell down from the feed-rack.

One reason for more lifting and bite/head knocking towards other pigs among the pigs in the SM treatment might be that they manipulated and separated out desirable parts of the feed on the pen floor. This in turn probably resulted in more interactions between the pigs. While the type of social interactions (nose, bite, etc.) was recorded, it was not possible to assess the severity of the behaviours by only observing the performing pigs. The severity of social interaction can be assessed by the response of the receiving pig, as suggested by Brunberg et al. (2011). As expected, the reactions were stronger (avoidance or return approach) when the performing pig climbed on, lifted, bit or head-knocked the receiving pig than when it only nosed (Figure 5). Moreover, the reaction to the social interactions seemed to be stronger (i.e. a higher proportion of interactions were met with a return approach from the receiving pig) among pigs fed only pellet feed, irrespective of the type of the interaction performed.

Another way to estimate the amount and severity of aggressive interactions is to count the number and severity of wounds on different parts of the pig body (Broom, 1991). In the present experiment, pigs fed intact silage had the lowest number of wounds on the head, middle part and hind quarters of the body. This indicates a relatively low level of violent social interactions in those pig groups. Pigs fed only commercial cereal-based feed had the highest number of wounds on the middle and hind part of the body. This is in agreement with previous studies showing that housing in barren environments causes more body damage from aggression than housing with straw bedding and more space allowance (Lyons et al., 1995; Guy et al., 2002). Our results provide additional evidence that in an environment already enriched with straw, provision of intact silage has the potential to further reduce damaging social behaviour. Comparisons between the pigs in the two pelleted feed treatments showed a higher number of wounds among pigs fed cereal-based feed, indicating that even when roughage is milled and pelleted, it influences the social environment in a pig group. This finding is probably not explained by better opportunities for foraging among pigs fed pellet silage, but can possibly be explained by a higher fibre content in the pelleted silage, leading to a longer period of satiety. The group size and animal density for pigs in this study was lower than what is common in most commercial production systems. It is possible that the positive effect of intact and pelleted silage on injurious behaviour would be even larger in such systems.

In conclusion, pigs fed chopped or intact silage in addition to straw in the present study were more active and spent a larger proportion of their time foraging than pigs fed only pelleted feed and straw. Moreover, provision of silage was found to influence the social behaviour of pigs. Pigs in groups fed chopped or intact silage expressed a lower proportion of severe reactions to social interactions and pigs fed intact silage had lower number of wounds on their bodies. Taken together, the results indicate that in housing environments already enriched with straw, provision of silage has the potential to further improve pig welfare.

Acknowledgements

This study was funded by the Swedish Farmers' Foundation for Agricultural Research. The authors thank the staff at the Lövsta pig research unit at the Swedish University of Agricultural Sciences.