Introduction

In commercial pig production, access to rooting material such as straw improves pig welfare, as it increases the opportunities for pigs to perform important species-specific behaviours like foraging and exploration (SJVFS 2017). Roughage provision, in addition to straw, can further increase the time pigs spend eating and foraging (Olsen et al. 2000) and therefore meets the need of pigs to explore and forage to a greater extent (Høøk Presto et al. 2009). Pigs in conventional pig production systems have limited access to rooting material and are often kept in barren environments, which may redirect their motivation for foraging and exploration behaviour towards pen fittings and other pigs in the pen, sometimes leading to injurious behaviours (EFSA 2007). Provision of straw has been found to reduce such behaviour, while additional access to roughage increases foraging behaviour and reduces the incidence of aggressive behaviour even further (Høøk Presto et al. 2009; Presto et al. 2013).

In production systems where growing/finishing pigs are provided with roughage, e.g. organic production systems, this roughage is intended to supply nutrients and behavioural enrichment for the pigs (EC 2008). According to Wallenbeck et al. (2014), roughage can contribute some nutrients to growing/finishing pigs. If suitable roughage types are already grown on-farm, they can be preserved as silage and fed to pigs as a locally produced feedstuff, contributing to more self-sufficient and sustainable production (Wallenbeck et al. 2014). Forage often consists of grass and legume species and sometimes also herbaceous flowering plants. The most commonly grown legume species in Europe is clover, particularly red clover (Trifolium pratense) and white clover (Trifolium repens). Clover is commonly included in crop rotations due to its ability to fix nitrogen from the atmosphere and is often sown together with different grass species. Chicory (Cichorium intybus) can also be included in mixed leys. It is a perennial herb with a deep root system, which makes it persistent to drought, and it is high yielding and can be grown in temperate climates. This makes it a beneficial complement to other forage crops and gives a stable and persistent ley (Foster 1988). Chicory has a relatively high content of lysine in the leaves, but a lower amount of crude protein content than legumes (Crawley 2015). Inclusion of chicory in pig diets has been shown to favour Lactobacillus in the intestinal microflora, which may have a prebiotic effect (Ivarsson 2012).

Pigs in modern pig production, including the majority of organic pigs produced in Europe, are breed crosses, with a Landrace*Yorkshire crossbred sow and a Hampshire, Duroc or Piétrain sire. The dam breeds are selected for good reproduction in combination with high growth and the sire breeds are selected for high lean meat growth and effective feed conversion. Pigs have the capacity to digest and use nutrients from roughage to some extent, but inclusion of roughage in the diet of growing/finishing pigs can reduce the ileal digestibility of nutrients (Andersson and Lindberg 1997a, 1997b; Carlsson et al. 1999; Edwards 2003; Wallenbeck et al. 2014). In efforts to achieve sustainable organic pig production, it is important to evaluate the effect and validate the potential for using additional roughage, such as silage, as a nutrient source to modern crossbreed pigs.

The aim of the present study was to investigate the effect of access to chicory and red clover silage, in combination with a restricted feeding regime, on performance, time budgets and social behaviours in two pig crossbreeds used in commercial organic pig production in northern Europe.

Materials and methods

The experiment was performed at the Swedish Livestock Research Centre, Funbo Lövsta, Uppsala, Sweden, during July–August 2014. The study was approved by the Uppsala Ethics Committee on Animal Research (ethics approval number C86/14), which is in compliance with EC Directive 86/609/EEC on animal experiments.

Animals, housing, diets and treatments

A total of 72 finishing pigs, 36 Yorkshire × Hampshire crosses (YxH) and 36 Yorkshire × Duroc crosses (YxD), from nine litters in one production batch, were included in the study. At weaning (5 weeks of age), the litters were split and the piglets were distributed to 18 pens, with four pigs per pen. At an age of approximately 67 days (mean 67 days ± standard deviation (SD) 1.2 days) and a live weight of approximately 30 kg (33.6 ± 4.1 kg), the pigs were moved from the farrowing unit to a growing/finishing unit. Each pen consisted of two females and two males (one sex of each cross type YxH and YxD) and these groups were kept together throughout the experimental period. The male pigs were immunocastrated with Improvac™, with their first injection at 77 days of age and their second at 105 days. The experimental period started when the pigs were at an approximate live weight of 80 kg (81.5 ± 8.0 kg) and an average age of 115 days (115 ± 1.2 days). The study continued until slaughter, which occurred at approximately 110 kg live weight (108.6 ± 10.0 kg) and an average age of 147 days (147 ± 1.2 days). The pens consisted of a concrete floor in the lying and feeding area and a slatted dunging area (one-third of pen area). Total pen area was 5.76 m2, giving a floor area of 1.44 m2 per pig. The lying area was 1.8 m × 2.2 m and the slatted area was 1.8 m × 1.0 m. The partitions between pens in the lying area were solid walls, with a rack for silage placed 0.75 m above the floor. The partitions between pens in the slatted area were gates with metal bars. There was a wet feed trough (1.8 m long) in the front of each pen, in the lying area. One water nipple was placed in the slatted area. All pigs had access to straw (approximately 0.5 kg per pen and day), all pigs were individually monitored for disease and injuries by the staff every day and all pens were cleaned daily.

The study included three treatments: chicory silage, red clover silage and a control, with six pens of four pigs (total 24 pigs) per treatment. All pigs were fed a wet commercial cereal-based diet three times a day (generally at 8–10, 15–16 and 17–22 h) in the feed troughs. Dry feed was mixed with water prior to feeding and the wet feed was distributed to each pen by an automatic computerised feeding system (Skiold Datamix AB).

The pigs in the control treatment were fed 100% commercial feed according to the standard feeding and nutrient recommendations for growing/finishing pigs (Andersson et al. 1997). During the experimental period, they were provided a maximum feed ration of 25.6 MJ NE/day. The pigs in the chicory and red clover treatments were given 80% commercial diet (energy basis), i.e. 20.5 MJ NE/day, and an ad libitum supply of whole-crop (stem, leaf and flower) chicory or red clover silage in the silage rack. The silage racks were filled with new silage three times per day. The silages were from the first cut, harvested in June 2014. Samples of the silages were pooled into bulk samples for analysis. The nutritional composition of the commercial diet and silages is presented in Table 1.

Table 1 Chemical composition and energy value of commercial feed and the silages
Full size table

Of the 72 pigs, one pig in the red clover treatment was tail-bitten 2 days prior to the start of the experimental period. This pig was removed from the pen and placed in a disease treatment pen, where it was kept until slaughter. It was excluded from the study.

Performance

Feed intake of the commercial diet for all pigs in the pen was recorded on a daily basis by the computerised feeding system and feed conversion ratio was calculated pen-wise. All pigs were weighed on the first day of the experimental period. When the pigs in each pen reached an average live weight of 108.6 kg (± 10.0 kg), their final weight was recorded and they were all sent to slaughter the following day. At slaughter, the carcass weight was recorded before cooling and lean meat content was determined with the Hennessy Grading Probe (Hennessy Grading Systems, Auckland, New Zealand; Sather et al. 1991). Daily growth from start of the experiment to slaughter was calculated as follows: (live weight on day before slaughter − weight at experiment start) / (slaughter date − start date).

Behaviour observations

All pens had a camera device installed over the slatted floor, which recorded the pigs day and night throughout the experiment. Behaviour observations were made based on the recorded video material, by the same person in every case. The pigs were observed for two consecutive days on two occasions (week 2 and week 4) during the experimental period, in order to estimate their time budgets and social interactions. To measure the time budgets of the pigs, instantaneous scan sampling was performed every hour during daytime (between 7.00 and 19.00 h). At each scanning, the activity (eating feed, eating silage, drinking, nosing/rooting pen floor, nosing/biting pen fittings, nosing/biting other pig and doing nothing), body posture (lying, sitting and standing) and location of the pigs in the pen (in lying area and in slatted area) were recorded as described in Table 2.

Table 2 Ethogram of pig behaviours recorded from the video material
Full size table

Social interactions were observed and recorded continuously for 5 min three times per day (at 11.15, 13.45 and 17.30 h) on the same days as the scan sampling. The social interactions were divided into pig performing an interaction and pig receiving an interaction. The definitions of the social interactions by performing pigs (nosing, nibbling/biting, tail biting, head knocking, climbing, lifting, pushing, belly massaging, biting on pen fittings and eating/rooting silage) are presented in Table 2. The receiving pig’s reaction to a social interaction was recorded as no reaction, avoid and return approach. Different social interactions were recorded as new events, regardless of whether they were performed by the same pigs or several different pigs. A new event was recorded as soon as new pigs interacted or when the interaction stopped for 3 s or more, and then started again.

Chemical analyses

Feed samples were freeze-dried, milled through a 1-mm sieve and then analysed for dry matter (DM) content by drying at 103 °C for 16 h. Ash content was analysed by combustion at 550 °C for 3 h. Nitrogen (N) was determined by the Kjeldahl method (Nordic Committee on Food Analysis 1976) using a 2020 Digestor and a 2400 Kjeltec Analyser Unit (FOSS Analytical A/S Hilleröd, Denmark). Correction of the crude protein (CP) content for losses of N at freeze-drying was performed according to NorFor (Nordic feed evaluation system). Neutral detergent fibre (NDF) was determined with ND solution (100%), amylase and sulphite (Chai and Udén 1998). Crude fat (EG fat) was determined according to the method described in the Official Journal of the European Communities (EU 1984), using a 1047 Hydrolysing Unit and a Soxtec System HT 1043 Extraction Unit (FOSS Analytical A/S). Crude fibre content was analysed with the SLL method according to conventional Swedish analytical methods for feed and plant material (Jennische and Larsson 1990). Gross energy (GE) content was measured with an Isoperibol bomb calorimeter (Parr 6300, Parr Instrument Company, Moline, IL, USA).

Statistical analyses

Statistical analyses were performed with the Statistical Analysis System, version 9.4 (SAS 2017). The effect of treatment on performance was evaluated with Proc MIXED. The statistical model included treatment (chicory silage, red clover silage or control), genotype (YxD or YxH) and gender (female or vaccinated entire male pigs) as fixed effects. Pen and litter were included as random effects in the model. Two-way interactions were tested and included in the model if found to be significant (P < 0.05). Pig was the experimental unit for carcass and all performance variables except feed consumption and feed conversion ratio, where pen was regarded as experimental unit. The model for the (pen-wise) feed consumption and feed conversion ratio included only treatment (chicory silage, red clover silage or control) as a fixed factor. Initial weight was included in the model as a covariate for the analysis of daily weight gain, while carcass weight was included in the model as a covariate for the analysis of lean meat content.

Descriptive statistics on behaviour variables were calculated using Proc FREQ and MEANS. The effect of treatment on time budgets and social interactions was evaluated using Proc MIXED. Differences between treatments in pigs’ time budget in terms of body posture, pen location and activity (percent of time) and social interactions with other pigs (average number of interactions in the pen per observation occasion) were analysed with pen average as the experimental unit. The statistical model included the fixed effects of treatment (chicory silage, red clover silage or control) and observation occasion (week 2 or week 4), and the random effect of pen. For the variable eating roughage, only the chicory silage and red clover silage treatments were included in the fixed effect of treatment. Residuals of the variables were examined for normal distribution using Proc UNIVARIATE, where the Shapiro-Wilks test for normality and normal probability was considered, and all were found to be normally distributed. Results are expressed as least square mean, unless otherwise stated.

Results

Performance

There was a significant effect of treatment on growth rate of the pigs (Table 3). Pigs fed chicory had 15% lower daily growth rate and pigs fed red clover 10% lower daily growth rate than pigs in the control treatment. The allowance of the commercial feed was 20% lower (energy basis) in the silage treatments and consumption of the commercial feed was correspondingly 20% lower for pigs in pens with silage treatments than for pigs in pens with the control treatment (2.0 kg feed/day for pigs in both the chicory and red clover treatments compared with 2.5 kg feed/day for pigs in the control treatment). Feed conversion ratio based on intake of the commercial feed was lower among pigs in the red clover treatment than among pigs in the chicory and control treatments (Table 3). Silage consumption and silage residues were not recorded, so estimation of feed conversion ratio based on total feed and silage intake was not possible. Pigs in the chicory treatment had the numerically lowest carcass weight, followed by pigs in the red clover treatment and pigs in the control treatment (77.6, 81.5 and 83.3 kg, respectively), with the difference between the chicory treatment and the control being statistically significant (P = 0.036). A similar pattern was found for dressing percentage (73.5, 74.3 and 76.0%, respectively; significant difference between chicory treatment and control (P = 0.018)). Lean meat content did not differ between treatments (Table 3).

Table 3 Performance and carcass quality of pigs fed 100% commercial feed (control) or 80% (energy basis) commercial feed with ad libitum access to either chicory or red clover silage. Data are presented as least square means, pooled standard error (SE) and F-value. N = 71 pigs (chicory silage = 24 pigs, red clover silage = 23 pigs, control = 24 pigs). Treatment means within rows with different superscripts are significantly different (P < 0.05)
Full size table

No difference in performance was found between genotypes (YxH or YxD) or between genders (female or vaccinated entire male pigs), except for lean meat content and dressing percentage. The YxH pigs had significantly higher lean meat content (59.7%) than the YxD pigs (57.6%) (P = 0.003, SE = 0.50%, F = 10.2) but lower dressing percentage (73.8 and 75.3%, respectively; P = 0.048, SE = 0.57%, F = 4.1). Female pigs had lower lean meat content (58.2%) than vaccinated entire male pigs (59.1%) (P = 0.020, SE = 0.41%, F = 5.9) and lower dressing percentage (73.7 and 75.4%, respectively; P = 0.007, SE = 0.52%, F = 8.1).

Time budgets

Time spent doing nothing was most frequent among pigs in the control treatment (77.4%) and differed significantly from pigs in the chicory and red clover treatments (71.5 and 67.2%, respectively) (P = 0.001, SE = 1.547%, F = 10.9, N = 36 observations unless otherwise stated). The pigs spent on average 70.5 ± 6.0% (mean ± SD) of their time lying down. Pigs in the chicory and red clover treatments spent a significantly lower proportion of their time lying down (69.3 and 67.0%, respectively) than pigs in the control treatment (75.2%) (P = 0.026, SE = 1.98%, F = 4.6). There was no difference in time spent sitting between treatments (2.7, 2.6 and 2.4%, for chicory, red clover and control, respectively; P = 0.920, SE = 0.66%, F = 0.1). However, pigs in the chicory treatment tended to spend more time standing (27.9% of observation time) and pigs in the red clover treatment spent significantly more time standing (30.3%) than pigs in the control treatment (22.3%) (P = 0.084 for chicory vs. control; P = 0.018 for red clover vs. control, SE = 3.06%, F = 3.6), see Fig. 1.

Fig. 1
figure 1

Proportion (%) of time (least square means ± pooled standard error) spent on different behaviours by pigs in the chicory silage, red clover silage and control treatments. Treatment means with different superscripts are significantly different (P < 0.05). N = 36 except for eating silage, where N = 24

Full size image

Pigs in the chicory and red clover treatments spent significantly more time in the lying area (64.4 and 67.0%, respectively) than pigs in the control treatment (54.2%) (P = 0.007, SE = 2.63%, F = 6.6). Thus, they spent correspondingly significantly less time in the slatted area (35.6 and 33.1%, respectively) than pigs in the control treatment (45.8%) (P = 0.007, SE = 2.63%).

The time spent eating commercial feed did not differ between treatments (5.5, 6.3 and 5.3% for chicory, red clover and control, respectively) (P = 0.405, SE = 0.06%, F = 1.0). The frequency of drinking was also similar between treatments (1.2, 1.2 and 1.1%, for chicory, red clover and control, respectively) (P = 0.984, SE = 0.44%, F = 0.02), see Fig. 1.

There was no significant difference in time spent eating/rooting silage between pigs in the chicory and red clover treatments (4.3 and 4.7%, respectively; P = 0.560, SE = 0.70%, F = 27.0; N = 24 observations). Pigs in the chicory treatment tended to spend more time nosing/rooting on the pen floor (12.3% of observation time) and pigs in the red clover treatment spent significantly more time (14.5%) in this activity than pigs in the control treatment (7.6%) (P = 0.095 for chicory vs. control; P = 0.019 for red clover vs. control, SE = 2.67%, F = 3.5, N = 36 observations). Pigs in the chicory and red clover treatments spent a lower proportion of time biting pen fittings (1.1 and 1.5%, respectively) than pigs in the control treatment (3.8%) (P = 0.001, SE = 0.43%, F = 11.2, N = 36 observations), see Fig. 1.

The proportion of time spent in different body postures (lying down, sitting and standing) did not differ between the first and second observation occasion (week 2 and week 4) (P > 0.001 for all). The pigs spent more time in the lying area in week 4 than in week 2 (67.1 and 56.6% of time, respectively) and correspondingly less time in the slatted area (32.9 and 43.4%, respectively; P = 0.001, SE = 1.71%, F = 43.8, N = 36 observations for both).

In week 4, the pigs spent more time (8.1% of time) eating commercial feed, i.e. with their snout in the feeding trough, than in week 2 (3.4%) (P = 0.001, SE = 0.46%, F = 52.4). However, among the pigs fed silage, the proportion of time spent eating silage did not differ between the observation occasions (P = 0.690). There was a slight tendency for a higher proportion of time spent nosing/biting on pen fittings in week 2 compared with week 4 (2.6 and 1.7%, respectively; P = 0.078, SE = 0.35%, F = 3.5). The proportion of time spent doing nothing, i.e. having the snout in the air, decreased over time (73.8% of time in week 2 compared with 70.3% in week 4; P = 0.026, SE = 1.15%, F = 6.0).

Social interactions

Differences between treatments as regards performance of social interactions are presented in Table 4. Nosing other pigs occurred more frequently among pigs in the control treatment than pigs in the chicory and red clover treatments (P = 0.012). Biting/nibbling tended to be performed to a greater extent among pigs in the control and chicory treatments than pigs in the red clover treatment (P = 0.073). Pigs in the chicory treatment also performed more head knocking behaviour than pigs in the control and red clover treatments (P = 0.003). The occurrence of other social interactions such as climbing, lifting, pushing and belly massaging did not differ between treatments (P > 0.200 for all). The frequency of rooting/eating silage among pigs in the chicory and red clover treatments did not differ (P = 0.738). Pigs in the control treatment were observed biting the pen fittings more frequently than pigs in the chicory and red clover treatments (P = 0.002).

Table 4 Average number of social interactions per pig and observation occasion for pens in the chicory silage, red clover silage and control treatments. Total time of continuous observations was 1 h/pen (5 min × 3 times a day × 2 days on 2 occasions). Data are presented as least square means, pooled standard error (SE) and F-value. N = 36 observations. Treatment means within rows with different superscripts are significantly different (P < 0.05)
Full size table

There were no significant differences in the reactions of receiving pigs (i.e. avoiding an approach or returning an approach) between the different treatments (P > 0.200 for both). However, receiving pigs in the chicory treatment tended to respond with no reaction less often (12.4 times) than pigs in the red clover and control treatments (15.4 and 15.8 times, respectively) (P = 0.080, SE = 1.08 times, F = 2.9).

The number of social interactions by performing pigs did not differ over time (P > 0.10 for all) except for climbing, which occurred more frequently in week 4 than week 2 (0.96 times and 0.31 times, respectively; P = 0.015, SE = 0.21 times, F = 7.3). The reaction of receiving pigs was not influenced by observation occasion (P > 0.10 for all).

Discussion

This study investigated the effect of access to chicory and red clover silage on the performance, time budget and social interactions of finishing pigs. Organic pig production can benefit from using ley crops, as inclusion of legumes in the crop rotation results in fixation of atmospheric nitrogen. Leys where several plant species are grown simultaneously also have a positive impact on biodiversity. Moreover, inclusion of nutritionally high-quality silage in the diet of growing/finishing pigs could increase the use of locally produced feed materials, replacing other protein sources, and result in more sustainable use of arable land (Hermansen et al. 2017). At the same time, the legal requirement on giving organic pigs’ access to roughage can be fulfilled. This study was performed under experimental conditions and did not fulfil all the requirements for organic pig production, e.g. organically produced feed and access to outdoor areas, because we wanted to evaluate the influence of silage in a controlled study. However, the number of pigs per pen was less than in commercial systems, and the space allowance (1.44 m2/pig) was similar to the organic requirement of 1.3 m2/pig under EU organic regulations and 0.8–1.0 m2/pig (for pigs of 80–110 kg live weight) under Swedish animal welfare law (EC 2008; SJVFS 2017). Overall, the results from the present study provide a better understanding of the benefits and disadvantages of feeding silage to finishing pigs.

Performance

Daily weight gain was lower in pigs in the chicory and red clover silage treatments than in pigs in the control treatment. The conventional feed allowance was 20% lower (on energy basis) in the silage treatments, and the lower weight gain indicates that the capacity of these pigs to consume and digest silage was too low to supply this nutrient deficit. Similar findings have been reported in previous studies (Edwards 2003; Hansen et al. 2006; Bikker et al. 2014; Wallenbeck et al. 2014). However, the growth rate was not as pronounced as expected based on the reduced energy allowance, which suggests that chicory and red clover silage contributed some energy and nutrients to the finishing pigs. Carlsson et al. (1999) concluded that clover-grass silage, and other roughage-based diets, are consumed well by pigs, give satisfactory digestibility values and can provide energy for growing pigs. In the present study, the pigs were supplied with intact silage ad libitum, although the exact amount of chicory and red clover silage consumed was not recorded. Previous studies have shown that both clover-grass and chicory forages are palatable and well consumed by pigs (Carlson et al. 1999; Ivarsson et al. 2012). Restricted feeding is believed to result in higher foraging motivation (Day et al. 1995; Beattie and O’Connell 2002). However, limited capacity for intake and hindgut fermentation of fibrous feedstuffs in young growing pigs can have a negative effect on performance (Edwards 2003). The low silage consumption in the present study may also have been due to warm weather, and corresponding high indoor temperatures, during the experimental period. Voluntary feed intake is affected by thermal environment factors such as air temperature and humidity (Bruce and Clark 1979; Verstegen et al. 1987; Massabie et al. 1997). Moreover, microbial hindgut fermentation of fibrous feedstuffs to produce short-chain fatty acids also produces heat which, together with high ambient temperatures, may have affected the intake of silage (Nyachoti et al. 2004). Stahly and Cromwell (1986) reported that dietary addition of 10% alfalfa meal depressed daily gain by only 1% in a cold environment (10 °C), compared with 3 and 5% in 22.5 and 35 °C environments, respectively.

In the present study, there were no substantial differences in the energy content and chemical composition of the two silages (Table 1). Pigs in the red clover treatment had numerically higher daily growth than pigs in the chicory treatment, which may indicate higher consumption of red clover silage. Moreover, feed conversion ratio based on commercial feed intake was lower for the pigs fed red clover than for those fed chicory and the control diet, which suggests that the red clover silage contributed some nutrients to the finishing pigs. Data on digestibility coefficients for forage crops are lacking, but in vitro and in vivo studies on the organic matter (OM) digestibility of different forms of forages (meal, fresh or silage) have reported digestibility values within the range 40–80% and have concluded that growing/finishing pigs have the capacity to digest roughage to some extent (Håkansson and Malmlöf 1984; Andersson and Lindberg 1997a, 1997b; Carlsson et al. 1999; Jørgensen et al. 2012). However, a number of factors would affect the digestibility value: choice of plant species, plant stage at harvest, dietary NDF content, whether the forage is dried, milled, chopped, ensiled or served fresh, and the ability of pigs to sort out more desirable parts such as leaves and stems (Andersson and Lindberg 1997b; Carlsson et al. 1999; Edwards 2003).

Inclusion of fibre-rich feed stuffs is often associated with increased gut fill and increased gastrointestinal organ weight (Pluske et al. 2003; Wellock et al. 2008). This in turn lowers the dressing percentage (Heyer et al. 2006; Bikker et al. 2014; Wallenbeck et al. 2014). Dressing percentage of pigs in the chicory and the red clover treatments in the present study was numerically lower (by 2.5 and 1.7 percentage units, respectively) than for pigs in the control, although the difference was only significant between the chicory treatment and the control. However, the NDF content of the chicory and red clover silages in this study was similar and, because actual silage consumption was not recorded, it is difficult to explain this result. Ivarsson et al. (2012) found that inclusion of chicory forage stimulated hindgut development and resulted in a heavier colon, which might explain the lower dressing percentage among pigs in the chicory treatment. The lower dressing percentage might be the reason for the significantly lower carcass weight found among these pigs. As shown in other studies (Danielsen et al. 1999; Hansen et al. 2006; Bikker et al. 2014; Wallenbeck et al. 2015), lean meat content was not affected by the dietary treatment. Any energy intake above maintenance and beyond the required level for maximum protein deposition will be used for lipid synthesis (Van Milgen et al. 2000). However, the pigs in the chicory and red clover treatments did not exceed the maximum protein deposition and thus energy was not used for lipid synthesis and fat deposition. These pigs probably did not consume enough feed to fulfil their energy requirement, resulting in poorer growth than in pigs in the control treatment.

Time budgets

It is well established that pigs of modern breeds housed indoors in pens are inactive, lying down resting and sleeping for most of the day (Morrison et al. 2007). This was also found in the present study, with an inactivity level of 70% on average. The experimental period had very warm summer weather and the ventilation system did not manage to cool the finishing unit sufficiently, so the indoor temperature was not as low as desirable. This might have affected the proportion of inactive time among the pigs. While the activity level in pigs kept indoors is low, it has been shown that various rooting materials can increase their exploratory behaviour to differing degrees (Beattie et al. 1996; Olsen et al. 2000; Guy et al. 2002a, 2002b; Long 2002; Olsen et al. 2002). Jensen and Pedersen (2007) concluded that straw is a good rooting material, but that some other materials are preferred by pigs. Foraging behaviour is stimulated when the enrichment material is also edible (Studnitz et al. 2007). According to Olsen et al. (2000), pigs prefer roughage with a low DM content and the amount of time pigs spend manipulating roughage depends on the specific characteristics of the roughage offered, such as DM content, complexity, texture, smell and taste. In the present study, pigs with access to silage were more active and spent less time lying down than pigs in the control treatment without silage. As shown previously by Høøk Presto et al. (2009) and Presto et al. (2013), access to silage stimulated the pigs to be more active than when only straw was provided as an enrichment (control treatment). The pigs provided with chicory and red clover silage spent a higher proportion of their time in the pen lying area and up to twice as much time nosing/rooting on the pen floor than the pigs in the control treatment. This was probably because the silage racks were placed in the lying area and some of the silage ended up on the pen floor when the pigs were eating, which indicates that the pigs spent time exploring the silage on the floor.

The higher activity level and the lower proportion of time spent biting at pen fittings among pigs in the chicory and red clover treatments than pigs in the control treatment suggest that access to silage motivated the pigs to perform foraging and rooting behaviour.

Social interactions

There were fewer behaviours directed at other pigs and pen fittings among the pigs provided with silage. Pigs deprived of opportunities to forage and explore their surroundings may redirect those behaviours towards other pigs in the pen and pen fittings (Lyons et al. 1995; Beattie et al. 2000; Presto et al. 2013). In agreement with this, performing behaviour such as nosing at other pigs also occurred less frequently among pigs in the chicory and red clover treatments than pigs in the control. Taken together with the fact that the pigs in the silage treatments also spent more time nosing/rooting on the pen floor and less time on biting pen fittings (both regarding time budgets and social interactions), this indicates that the roughage occupied the pigs. Several studies have concluded that provision of straw and other rooting materials reduces unwanted and aggressive behaviours (Beattie et al. 2000; Kelly et al. 2000; Beattie et al. 2001; Olsen 2001; Guy et al. 2002b; Long 2002; Olsen et al. 2002; Bolhuis et al. 2005). However, as found by Olsen et al. (2000) and Høøk Presto et al. (2009), in the present study, providing access to silage seemed to stimulate the pigs even more than providing straw only.

The pigs in the chicory treatment performed more head knocking behaviour than those in the other two treatments. This was unexpected and could relate to those pigs being more unsatisfied due to hunger, as suggested by Jensen and Pedersen (2010). The general impression among the staff was that there were more silage residues in pens where pigs were fed chicory silage. However, those pigs spent a similar proportion of time eating/rooting silage as the pigs fed red clover silage. Numerically, pigs fed chicory silage had the poorest daily growth of all treatments, which might suggest that they did not fulfil their energy and nutrient requirements and thus felt hunger.

Conclusions

Finishing pigs fed 80% commercial cereal-based feed (energy basis) and chicory or red clover silage ad libitum showed lower daily weight gain than pigs in a control group fed 100% commercial cereal-based feed (100% of nutrient recommendation). Conversion of the commercial cereal-based feed to lean meat growth was more efficient for pigs fed red clover silage than pigs fed chicory silage and pigs fed only commercial feed (control), indicating that the pigs utilised the nutrients in clover silage. Moreover, pigs fed chicory or red clover silage were more active and performed more feed-directed behaviours and less behaviours directed towards other pigs and pen fittings, i.e. expressed more natural behaviour. Provision of a silage supplement can thus meet the goal in organic production of further improving animal welfare. There is potential to use silage as a locally produced feedstuff for finishing pigs, but a better understanding is needed of nutrient intake, feeding strategies and production systems that promote utilisation of silage in pig production.