Search
Author
Title
Vol.
Issue
Year
1st Page

Journal of Animal Science - Animal Production

Effects of varying floor space on aggressive behavior and cortisol concentrations in group-housed sows1

 

This article in JAS

  1. Vol. 94 No. 11, p. 4809-4818
     
    Received: Apr 27, 2016
    Accepted: Aug 26, 2016
    Published: October 13, 2016


    2 Corresponding author(s): phh@unimelb.edu.au
 View
 Download
 Share

doi:10.2527/jas.2016-0583
  1. P. H. Hemsworth 2*,
  2. R.S. Morrison,
  3. A. J. Tilbrook,
  4. K. L. Butler,
  5. M. Rice* and
  6. S. J. Moeller#
  1. * Animal Welfare Science Centre, Faculty of Veterinary and Agricultural Sciences, University of Melbourne, Parkville, VIC 3010, Australia
     Rivalea Australia, Corowa, NSW 2646, Australia
     Animal Welfare Science Centre, South Australia Research and Development Institute, University of Adelaide, Roseworthy, SA 5371, Australia
    § Biometrics Group, Department of Economic Development, Jobs, Transport & Resources, Hamilton, VIC 3030, Australia
    # Animal Welfare Science Centre, Department of Animal Science, The Ohio State University, Columbus, 43210

Abstract

Floor space is an important determinant of aggression and stress in group-housed sows, and the aim of the present experiment was to comprehensively examine the effects of floor space in the range of 1.45 to 2.90 m2/sow from mixing until 27 d after insemination on aggression, stress, and reproduction of group-housed sows. A previous experiment on the effects of floor space indicated spatial variability across and along the research facility in both sow aggression and stress. To minimize this spatial variability within the research facility, similar-sized pens but with varying groups sizes (10–20) in 4 separate blocks of 3 contiguous pens within each of 9 time replicates (180 sows/replicate) were used to examine 6 space allowances (1.45–2.9 m2/sow). Space treatments were appropriately randomized to pens. Although it may be argued that space allowance is confounded with group size in this design, there was no evidence in our previous experiment of group size effects, for pens of 10 to 80 sows, or appreciable interactions between space and group size on aggression, stress, and reproduction. In the present experiment, sows were introduced to treatments within 4 d of insemination and were floor fed 4 times per day (2.5 kg/sow per d). On both Days 2 and 26 after mixing, aggressive behavior (bites and knocks) at feeding and plasma cortisol concentrations were measured. Restricted maximum likelihood mixed model analyses were used to examine the treatment effect after accounting for replicate and random spatial location effects within replicate. There was a consistent linear effect of floor space allowance on aggression at feeding at Day 2 (P < 0.0001) and plasma cortisol concentrations at Day 2 (P = 0.0003), with aggression and stress declining with increasing space. However, there were no effects of space allowance on aggression and stress at Day 26 (P = 0.14 and P = 0.79, respectively). These results show that increased floor space in the immediate post-mixing period reduces aggression and stress and that sows may adapt to reduced floor space over time. A strategy of staged-gestation penning, with more space immediately after mixing and less space later in gestation, may address both animal welfare and economic considerations, but this clearly requires further examination.



INTRODUCTION

With legislative, consumer, and retailer pressure on the housing of sows in gestation stalls (Fraser, 2008; Matthews and Hemsworth, 2012), there is increasing use of group-housing systems for gestating sows. Sow aggression, especially if intense and prolonged, has obvious welfare implications, particularly for subordinate sows, because of injuries and stress (Mendl et al., 1992; Nicholson et al., 1993; Verdon et al., 2015, 2016) and their probable links to pain and fear (Hemsworth et al., 2015).

Floor space is an important determinant of aggression and stress in group-housed sows (see review by Verdon et al., 2015). Recent research shows that reducing floor space for sows within the range of 1.4 to 3.0 m2/animal increases aggression and stress, as assessed by aggressive behavior and plasma cortisol concentrations, respectively (Hemsworth et al., 2013). However, the space allowance at which aggressive behavior and cortisol concentrations are optimal (i.e., unaffected by increasing space) could not be determined by these results. Although the results were in accord with a linear decline in cortisol and aggression at Day 2 after mixing from 1.4 to 3.0 m2/sow, the results were also in accord with a decline in cortisol and aggression from 1.4 to 1.8 m2/sow and no further decline above 1.8 m2/sow. Examination of the sources of variation in this experiment showed that there was considerable spatial variability across and along the research facility in both aggression and stress, which led to insufficient statistical precision to determine the optimal space allowance.

The previous experiment by the authors (Hemsworth et al., 2013) indicated that reducing floor space within the range of 1.4 to 3.0 m2/animal increased aggression and stress at Day 2 after mixing but not at Days 9 and 51. The present experiment reexamined the effects of floor space in the range of 1.45 to 2.90 m2/sow from mixing until 27 d after insemination. The experiment was conducted in the same facility used in the previous experiment, but in order for the effects of space allowance to be examined with more precision, the contribution of spatial variability to experimental error was reduced by reducing the spatial area taken up by the experiment and by blocks of pens within the experimental area.


MATERIALS AND METHODS

Animals, Housing, and Experimental Design

This experiment was conducted between January 2013 and May 2014 in a gestation unit of a large commercial piggery in southern New South Wales, Australia, specifically renovated for this experiment. The 61- by 19-m building was equipped with adjustable blinds and overhead water sprinklers, covering the 50% slatted floor area of the pens, that were activated (3 min on and 15 min off) when the internal temperature exceeded 26°C.

All animal procedures were conducted with prior institutional ethical approval under the requirements of the New South Wales Prevention of Cruelty to Animals Act 1979 (NHMRC, 2013), in accordance with the National Health and Medical Research Council/Commonwealth Scientific and Industrial Research Organization/Australian Animal Commission Code of Practice for the Care and Use of Animals for Scientific Purposes (NHMRC, 2013).

A total of 1,620 Landrace × Large White sows were studied in 9 time replicates (i.e., 180 sows per replicate) in an experiment of 74 wk. The sows of mixed parity (2 and older) were of good health at the beginning of the experiment and were introduced to the post-insemination housing treatments within 4 d of insemination. From weaning until housing treatment imposition, sows were housed in mating stalls.

The sows were housed in 2 central blocks of 6 pens (total of 12 pens) within the facility and each pen had a floor area of 29 m2 (9.6 m long × 3.02 m wide; see Fig. 1). The floor space treatments imposed after insemination were

  1. floor space allowance of 1.45 m2/sow (20 sows in a group),

  2. floor space allowance of 1.61 m2/sow (18 sows in a group),

  3. floor space allowance of 1.81 m2/sow (16 sows in a group),

  4. floor space allowance of 2.07 m2/sow (14 sows in a group),

  5. floor space allowance of 2.42 m2/sow (12 sows in a group), and

  6. floor space allowance of 2.90 m2/sow (10 sows in a group).

Figure 1.
Figure 1.

Pen layout.

 

As indicated earlier, the present experiment was conducted in the same facility used in a similar experiment by Hemsworth et al. (2013), but in order for the effects of space allowance to be examined with more precision, the contribution of spatial variability to experimental error was reduced by reducing the spatial area taken up by the experiment and by blocks of pens. This was achieved by only using pens of 10 to 20 sows and using an incomplete block design that had 4 separate blocks of 3 contiguous pens within each time replicate. To avoid any confounding with group size, each space allowance treatment would need pens of different sizes. Consequently, in a randomized design, there is the practical problem of having to frequently modify the size of pens to achieve randomization of space treatments to pens. Therefore, space allowance in the present experiment was altered by varying group size from 10 to 20 sows. Although it may be argued that space allowance is confounded with group size in this design, in the previous experiment (Hemsworth et al., 2013), there was no evidence of group size effects or appreciable interactions between space and group size for pens having between 10 and 80 sows on aggression, stress, and reproduction. This finding provides confidence that the effect of space allowance in pens varying from 10 to 20 sows can be examined without appreciable confounding due to group size. In the previous experiment, there was often an appreciable effect of group size on skin injuries, and therefore, the effect of space allowance on skin injuries in pens with different number of sows cannot be examined without an expectation of confounding between group size and space allowance.

The basic pen layout, prior to pre-randomization, is presented in Table 1. The 12 pens were in 4 groups of 3 pens defined by 2 rows and 2 columns, with each column being the width of 3 west-to-east pens. Therefore, there are 3 pens in each row by column combination (pen blocks). The experimental design allocation is constructed so that, within each time replicate, each treatment occurs once in each row and in each column. Additionally, every combination of 3 treatments occurs exactly twice within a time by row by column combination, with the exceptions of the 1.45, 1.81, and 2.42 combination and the 1.61, 2.07, and 2.90 combination. These 2 combinations never occur together in a time by row by column combination. Space allowance to pens from this design was then randomized by time replicate, by row within a time replicate, by column within a time replicate, and by pen within each row by column by time replicate combination. Introduction to the allocated housing treatment was considered Day 1 of treatment.


View Full Table | Close Full ViewTable 1.

Layout of treatments (m2/sow) prior to the randomization process. Within each time replicate, north is to the top of table.

 
Column
Time replicate Rows West East
1 1 1.45 1.61 1.81 2.07 2.42 2.9
2 2.07 2.42 2.9 1.45 1.61 1.81
2 1 1.45 1.61 2.07 1.81 2.42 2.9
2 1.81 2.42 2.9 1.45 1.61 2.07
3 1 1.45 1.61 2.42 1.81 2.07 2.9
2 1.81 2.07 2.9 1.45 1.61 2.42
4 1 1.45 1.61 2.9 1.81 2.07 2.42
2 1.81 2.07 2.42 1.45 1.61 2.9
5 1 1.45 1.81 2.07 1.61 2.42 2.9
2 1.61 2.42 2.9 1.45 1.81 2.07
6 1 1.45 1.81 2.9 1.61 2.07 2.42
2 1.61 2.07 2.42 1.45 1.81 2.9
7 1 1.45 2.07 2.42 1.61 1.81 2.9
2 1.61 1.81 2.9 1.45 2.07 2.42
8 1 1.45 2.07 2.9 1.61 1.81 2.42
2 1.61 1.81 2.42 1.45 2.07 2.9
9 1 1.45 2.42 2.9 1.61 1.81 2.07
2 1.61 1.81 2.07 1.45 2.42 2.9

Each time replicate (9 replicates each consisting of 12 pens housing a total of 180 sows) was introduced into the experiment over 2 wk as follows. Within each time replicate on each of 2 successive Wednesdays, 90 sows that had been inseminated within the previous 4 d were assigned to a (subreplicate) row. Sows were housed in mating stalls after weaning and twice inseminated in a morning/afternoon insemination routine before being first selected for study and then allocated to a pen within the row (6 group sizes).

Parity 2 to 6 sows were randomly assigned to their experimental pens while still in their mating stalls. Within 4 d of insemination, sows were moved to their allocated treatment pen after receiving their daily feed allocation, to avoid immediate competition for feed in their group, and then remained in their group until checked for pregnancy at about 27 d after insemination. Following the pregnancy check, pregnant sows were rehoused in post-treatment pens of up to 10 familiar sows with a floor space of at least 1.8 m2/sow.

The 12 experimental pens were 50% solid concrete floors with 50% slatted concrete at the rear of the pens. Feed was delivered using automated drop feeders that were evenly suspended across the width of each pen. Feed was delivered 4 times per day (hourly from 0700 h) to provide a total of 2.5 kg/sow per d of a commercial diet (13.1 MJ/kg DM and 12.8% CP). Feeders were adjusted according to treatment (number of sows in the pen) on the afternoon prior to mixing and 1 full drop of feed was on the floor in each pen at the time of mixing, providing an additional one-fourth of the daily feed allocation on the day of mixing. Feeders were also adjusted according to the number of sows remaining in the pen at Day 23 of treatment (prior to the Day 26 physiology measurements). Sows were provided ad libitum access to water via 3 nipple drinkers per pen attached to the back wall and located over the slatted flooring of each pen.

Measurements

Aggressive Behavior at Feeding.

To observe aggressive behavior at feeding, two 3.6-mm infrared closed-circuit television cameras were installed overhead to record aggressive behavior at the time that feed was distributed on the solid floor below the drop feeder. From the digital video recordings, continuous observations were conducted to measure the number of bouts of aggressive behavior in the 30 min following each of the 4 feed drops on Day 2 and the first feed drop on Day 26. Aggressive behavior on Day 26 was measured only following the first feed drop because of the substantial time required to conduct these measurements.

Aggressive behavior was defined as slashes, butts, pushes, and bites, and these were distinguished from other tactile interactions with sows on the basis that the former were associated with avoidance or retaliation by one sow as a consequence of the interaction. The numbers of aggressive acts delivered by each individual sow during the observation period were continuously recorded by 1 observer. Because an aggressive behavior may result in a series of retaliations between 2 sows, during these fights, a bout criterion interval of 5 s was chosen to separate one bout of aggressive behavior from another bout of the same behavior by an individual sow (Hemsworth et al., 2013). Only aggressive interactions in which the head of the sow (defined as extending from the snout to the ears) displaying the aggressive behavior was clearly visible were recorded. The identity of each sow was not recorded because aggression at the level of the group was the main focus.

The number of sows in the field of view was recorded at regular intervals so that the number of bouts of aggression could be expressed on the basis of the average number of sows in the field of view during the observations. Instantaneous point sampling at 30-s intervals during each 5-min block of footage was used to count the number of sows at each sampling point, providing an estimate of the average number of sows in the field of view during each 5-min block of the observation period. Therefore, the frequency of bouts of aggression per sow in each pen, after each feed drop, was estimated by multiplying the number of aggressive interactions observed by the inverse of the estimated average number of sows in the field of view. Bouts of aggression per sow, in each pen, were estimated during the 30-min period following the first feed drops on Days 2 and 26 and during the 30-min period following feed drops 2 through 4 on Day 2.

Cortisol Concentrations.

Blood samples were collected via jugular venipuncture (10-mL lithium–heparinized tubes; BD Vacutainer; Becton, Dickinson and Company, Belliver Industrial Estate, Plymouth, UK) of sows restrained with a snout snare. Sampling by 3 technicians commenced at 1300 h and 10 sows from each pen (all sows in each pen with 10 sows and 10 sows in pens containing 12 or more sows) on Days 2 and 26. At the time a technician was ready to collect, sows were selected by the technician on the basis of the first sow alternately sighted in the central and peripheral areas of pens. Blood samples were collected within 2 min of snaring to avoid an acute stress response to handling influencing the basal plasma cortisol concentrations (Broom and Johnson, 1993) and all batches of 10 sows took less than 10 min to collect. Karlen et al. (2007) reported that repeated sampling of different sows within 3 groups of 85 over 30 min did not affect salivary cortisol concentrations. A 10-mL blood sample was collected from each of the 10 sows in each pen at Days 2 and 26 for subsequent analyses of plasma cortisol.

Blood samples for cortisol were centrifuged for 10 min at 1,912 × g at 4°C, with the plasma drawn off into individual microtubes and frozen. Samples from each pen were pooled using 200-μL aliquots from each individual sample and assayed for total and free cortisol concentrations. Plasma concentrations of cortisol were quantified using commercial RIA kits (Diasorin Australia Ltd., North Ryde, NSW, Australia). The sensitivity of the assay was 3.5 nmol/L. Mean intra-assay variation for low (21.7 nmol/L) and medium (174.3 nmol/L) plasma samples were 4.3 and 3.4%, respectively.

Sow Reproductive Performance.

Sows were routinely checked for return of estrus, throughout the study, using the back-pressure test in the close presence of a mature boar that was held in the aisle adjacent to the pens at the time of testing. All sows that returned to estrus were removed from treatment pens immediately on detection. The remaining sows were checked for pregnancy using ultrasound detection technology at about 27 d after mixing. Once confirmed pregnant, sows in their treatment groups were moved to a different housing facility and housed with a minimum of 1.8 m2/sow. All the sows farrowed in site-standard farrowing environment and data collected included farrowing rate (proportion of inseminated sows that farrowed) and litter size (number of piglets born alive, stillborn, and mummified). Stillborn piglets were judged on the basis that they were fully formed at farrowing, covered in fetal membrane, had fully formed eponychia on their hooves, and were located behind the sow. Data on sows confirmed pregnant that failed to farrow, abortions, and sows culled for injury were also collected.

Statistical Analyses

The unit of analysis was always the group of animals from a single treatment pen within a time replicate. Each measurement was analyzed using REML models with random effects for time replicate, row within time replicate, column within time replicate, and row–column combinations within time replicate. These random effects reflect the randomization of space allowances to the unrandomized pen configuration (Table 1). As the design is orthogonal to time replicate, row within time replicate, and column within time replicate, but not row–column combinations within time replicate, this analysis can be described as an analysis with recovery of information from row–column combinations (pen blocks). Prior to analyses, average aggression on Day 2, feed drop 1 aggression on Day 2, and cortisol on Day 2 were logarithmically transformed; average of feed drops 2 to 4 aggression on Day 2 was log(y + 1) transformed; cortisol on Day 26 was transformed using the negative reciprocal (−1/y) transformation; feed drop 1 aggression at Day 26 and number of stillborn and number of mummified piglets were square root transformed; and farrowing rate was transformed using a log(1.1 − y) transformation, so that the residual variation was homogeneous between replicates and space allowances. The analyses for number of piglets born alive per litter, number of stillborn piglets per litter, number of mummified piglets per litter, and total number of piglets born per litter were variance weighted by the number of sows farrowing from each treatment pen, because the residual variance appeared inversely proportional to the number of sows farrowing from a treatment pen. To mimic the standard approach used in multi-strata ANOVA, when a variance component was estimated to be negative, this was allowed to stand. Only 2 outliers, which were judged to be anomalous using graphs of fitted values versus residuals, were deleted from analyses. These outliers were a low value for cortisol at Day 2 and a low value for cortisol at Day 26.

Tests for a linear response to space allowance, on the transformed scale, were calculated by comparing a model with a fixed linear response to space allowance with a model with no fixed effects, using a Wald F test (has 1 numerator degrees of freedom). Tests for any additional response to space allowance (tests if a linear response to the space per sow is adequate for describing the response to sow) were calculated by comparing a model with a fixed 6-level factor for space allowance with a model with a fixed linear response to space allowance, using a Wald F test (has 4 numerator degrees of freedom).

Response curves to space allowance were estimated by fitting a model with, on the transformed scale, a fixed linear response to space allowance (space per sow). Predicted values, for each of the 6 space allowances, were fitted with a fixed 6-level factor for space allowance. Response curves and predicted values were calculated on the transformed scale, back-transformed to the original scale, and then graphically presented. All statistical analyses were performed using the GenStat 17 statistical package (Payne, 2014).


RESULTS

Aggressive Behavior at Feeding

There was a consistent linear effect of floor space on average aggression after the 4 feed drops on Day 2 (P < 0.0001; Table 2; Fig. 2A). There were similar relationships between floor space and average aggression after feed drops 2 to 4 (P = 0.0007; Table 2; Fig. 2B) and aggression after the first feed drop on Day 2 (P < 0.0001; Table 2; Fig. 2C). In contrast, there was no relationship between floor space and aggression after the first feed drop on Day 26 (P = 0.14; Table 2; Fig. 2D).


View Full Table | Close Full ViewTable 2.

Tests for effects of space on output measurements (P < 0.05 in bold)

 
P-value
Measurement Transformation Weighting Linear response to space per sow Any additional space effect
Aggression
    Day 2 average of 4 drops Logarithm 0.0000089 0.47
    Day 2 drop 1 Logarithm 0.000070 0.12
    Day 2 average of drops 2 to 4 Log(y + 1) 0.00072 0.47
    Day 26 drop 1 Square root 0.14 0.60
Cortisol
    Day 2 Logarithm 0.00030 0.20
    Day 26 Negative reciprocal (1/y) 0.79 0.75
Reproduction
    Farrowing rate log(1.1 − y) 0.29 0.028
    Born alive Number of sows farrowing 0.96 0.72
    Stillborn Square root Number of sows farrowing 0.73 0.15
    Mummies Square root Number of sows farrowing 0.90 0.21
    Total born Number of sows farrowing 0.78 0.77
Figure 2.
Figure 2.

Predicted values of aggression (frequency per sow) as affected by floor space allowance at (A) Day 2 (average across all 4 feed drops), (B) Day 2 (average of feed drops 2 to 4), (C) Day 2 (first feed drop only), and (D) Day 26 (first feed drop only).

 

Cortisol Concentrations

There was also a linear effect of floor space on plasma cortisol concentrations on Day 2 (P = 0.0003; Table 2; Fig. 3A). However, as seen with aggression, there was no relationship found between floor space and plasma cortisol concentrations on Day 26 (P = 0.79; Table 2; Fig. 3B).

Figure 3.
Figure 3.

Predicted values of cortisol (ng/mL) of sows at (A) Day 2 and (B) Day 26 as affected by floor space allowance.

 

Sow Reproductive Performance

The percentages of sows detected in estrus by 27 d after mixing in the 6 space allowance treatments of 1.45, 1.61, 1.81, 2.07, 2.42, and 2.90 m2/sow were 6.9, 7.7, 3.1, 5.2, 4.2, and 0%, respectively. As shown in Table 2 and Fig. 4A to 4D, there was no indication of a space effect on farrowing rate (P = 0.29), litter size (born alive; P = 0.96), numbers stillborn (P = 0.73), number of mummies (P = 0.090), and litter size (total born; P = 0.78). There was no indication of a space effect on sows removed for injury (Table 3; Fig. 5)

Figure 4.
Figure 4.

Predicted values of (A) farrowing rate (proportion of inseminated sows that farrowed), (B) litter size (piglets born alive per litter), (C) stillborn (per litter), and (D) mummified piglets (per litter) as affected by floor space allowance.

 

View Full Table | Close Full ViewTable 3.

Percentage of sows removed for nonreproductive reasons in each treatment

 
Measurement Floor space, m2/sow
Space allowance 1.45 1.61 1.81 2.07 2.42 2.90
Sows removed, % 6.1 6.5 3.8 5.2 3.7 5.6
Figure 5.
Figure 5.

Sows removed for injuries as a function of floor space allowance (each dot represents a group of sows from a single treatment pen within a time replicate).

 


DISCUSSION

The present results show a consistent linear effect of floor space on both aggression and cortisol at Day 2 after mixing but not Day 26: both aggression at feeding and plasma cortisol concentrations at Day 2 declined with increasing floor space from 1.45 to 2.9 m2/sow. Although increased space reduced aggression and stress in the immediate post-mixing period, there were no effects of space on aggression and plasma cortisol concentrations at Day 26, indicating that sows adapted to reduced space over time. These results agree with the findings of Hemsworth et al. (2013) that increasing floor space for sows within the range of 1.4 to 3.0 m2/sow reduces aggression and plasma cortisol concentrations in the immediate postmixing period (Day 2 in both experiments) but not later (Day 26 and Days 9 and 51 in the present and previous experiments, respectively). In contrast to the previous experiment, there was no effect of floor space on farrowing rate of sows in the present experiment.

Earlier studies have shown similar effects of space in gilts and sows. In gilts, for example, aggression was generally higher at a space allowance of 1.0 than at 3.0 m2/gilt (Barnett et al., 1992; Barnett, 1997) and plasma cortisol concentrations were higher at 1.0 than at 1.4, 2.0, or 3.0 m2/gilt (Hemsworth et al., 1986; Barnett et al., 1992; Barnett, 1997). However, unlike the present experiment and the previous experiment by Hemsworth et al. (2013) on sows, the effects of space on aggression and cortisol concentrations in gilts were still apparent up to 11 wk after mixing (Hemsworth et al., 1986; Barnett et al., 1992; Barnett, 1997). The effects of space later in gestation may be due to failure of gilts to adapt to the restricted space allowance of 1.0 m2/gilt. Other experiments on sows have shown that aggression is generally higher at 6 and 7 d after mixing at 2.0 m2/sow than at 2.4, 3.6, or 4.8 m2/sow (Weng et al., 1998) and at 3 and 8 d after mixing at 2.25 m2/sow than at 3.0 m2/sow (Remience et al., 2008). Effects of floor space on skin injuries have been observed in some of these studies (Weng et al., 1998; Salak-Johnson et al., 2007; Remience et al., 2008) but not in others (Barnett et al., 1992; Barnett, 1997; Hemsworth et al., 2013). These observations on space, particularly those on aggression and stress early after mixing, indicate that a space allowance for sows of 1.4 m2/animal is likely too small early after mixing and that reductions in aggression and plasma cortisol concentrations are likely to be achieved with space allowances for sows of 2.0 m2/animal or more.

Insufficient floor space can prolong aggression by affecting a sow’s ability to avoid or escape others and, hence, the formation of a stable hierarchy (Lindberg, 2001). Although there are other stressors associated with mixing such as handling and moving sows to a new pen (Barnett et al., 1981, 2001), the present results and those by Hemsworth et al. (2013) show that increased floor space in the immediate post-mixing period reduces aggression and stress. Indeed, the effects of space on aggression and stress in these 2 experiments were most pronounced early after grouping, suggesting that sows in static groups may adapt over time to a reduced space of 1.4 m2/sow or more. Indeed, adaptation may occur as the dominance hierarchy forms because it functions to reduce the need for aggression (Lindberg, 2001). As Hemsworth et al. (2013) noted, there is evidence in rodents and sheep of a dampening of the hypothalamo–pituitary–adrenal axis’ response to stressors as gestation proceeds and during lactation (Lightman et al., 2001; Tilbrook et al., 2006). Moreover, increasing density (increasing animals/unit of space) increased plasma corticosterone concentrations in male mice at Days 1 and 7, but not at Day 14, after grouping (Peng et al., 1989). Although the development of the social hierarchy over time may assist in reducing aggression and stress with overcrowding, habituation to spatial restriction may also be implicated. Using an operant conditioning apparatus, Faure (1991) found that laying hens kept in a small space can habituate to this space restriction within 6 wk. Apart from a few studies similar to those of Peng et al. (1989) and Faure (1991), there is little evidence in the literature of habituation to spatial restriction. Although sows in the present experiment were experienced with group housing, they were more likely to have required more space early after mixing to establish their social hierarchy than later in gestation when the hierarchy had been established. Some of the sows in some groups may have been familiar from previous gestations, and although the evidence of previous familiarity on sow aggression after mixing is contradictory (see review by Verdon et al., 2015), sows familiar from a previous gestation may also have required more space early after mixing to re-establish their social hierarchy.

Since the effects of space on aggression and stress in the present experiment and that by Hemsworth et al. (2013) were most pronounced early after grouping, there is a need to examine the effects of reducing space during gestation in static groups because this effect may offer the opportunity for staged-gestation penning to provide increased space immediately after insemination. Nevertheless, these results highlight the importance of sufficient space to reduce aggression first, and stress at mixing and second, the sow’s requirement for space appears to be less once the group is well established.

In comparison with the previous experiment by Hemsworth et al. (2013), more precision was achieved in examining the effects of space in the present experiment by reducing the spatial area taken up by each pen block to reduce the contribution of spatial variability to experimental error. These results show that the relationship between space and both aggression and cortisol on Day 2 after mixing was linear, suggesting that both aggression and stress are likely to continue to decline as space is increased beyond 2.9 m2/sow. It is perhaps not surprising that aggression and stress on Day 2 are likely to decline as space is increased beyond 2.9 m2/sow. For example, Edwards et al. (1986) observed that in a situation of unrestricted space in a straw yard, sows were pursued following aggressive interactions over distances that varied from 0 to 20 m. Indeed, irrespective of space in the immediate post-mixing period, unfamiliar sows are likely to attempt to establish the dominance hierarchy in the group, and therefore, aggression and stress are likely to occur soon after mixing. Therefore, identifying the space allowance at which aggressive behavior and cortisol concentrations are optimal (i.e., unaffected by increasing space) is an unrealistic objective. Since there is evidence sows may adapt to reduced space, albeit with risks to reproductive performance as shown by Hemsworth et al. (2013), the optimal strategy from both animal welfare and economic perspectives may be staged-gestation penning for static groups of sows to provide increased space immediately after insemination. This increased space early after mixing should allow the dominance hierarchy in the group to be established quickly with minimum aggression and stress, and, thus, minimum risk to reproduction, and then reduce space for the remainder of gestation. But obviously, as discussed earlier, there is a need to examine the subsequent effects of reducing space on aggression and stress in the remainder of gestation.

In fact, staged-gestation penning to provide increased space for recently inseminated sows is one of the common features often recommended in a dedicated mixing pen (Edwards et al., 1993; Arey and Edwards, 1998). Increased floor space along with other features, such as visual barriers, feeding stalls, straw or bedding, and ad libitum feeding, have been advocated as ways of reducing aggression, stress, and injuries early after mixing (Edwards et al., 1993; Arey and Edwards, 1998; Barnett et al., 2001). However, there has been little research on this topic, particularly the long-term effects when sows are subsequently placed in gestation group systems at lower space allowances.

In contrast to the present experiment, Hemsworth et al. (2013) found a significant negative relationship between space allowance and farrowing rate. There is no obvious explanation for these conflicting results. Although there was no replicate effect on aggression and stress in the previous experiment, the fertility of sows in summer was more susceptible to reduced space than in other replicates (Hemsworth et al., 2013). Unlike the previous experiment, sows in the present experiment, once confirmed pregnant at about 27 d after insemination, were housed in their treatment groups with a minimum of 1.8 m2/sow. Providing more space after Day 27 for sows initially housed at 1.45 and 1.61 m2/sow in the present experiment may have ameliorated spatial restriction effects on reproduction, particularly in summer. Furthermore, although it is generally accepted that stress impairs reproduction, Turner et al. (2005) concluded that a proportion of female pigs appear to be resistant to the effects of prolonged stress or sustained increased cortisol. Group housing during gestation was introduced into the experimental facilities used in the present experiment and that of Hemsworth et al. (2013) in 2009, and therefore, all sows in both experiments had prior experience of group housing during gestation, but older parity sows in the previous experiment most likely had less experience of group housing than those older sows in the present experiment. Furthermore, management of sows in groups is likely to improve over time as stockpeople become more familiar and experienced with group housing. Therefore, differences in seasonal effects, experience of both sows and stockpeople, and/or variation in the susceptibility of sows to stress may be implicated in these conflicting findings on reproduction.

Because there is evidence that sows may adapt to reduced floor space, staged-gestation penning to provide increased floor space for recently inseminated sows immediately after mixing may be a practical solution to reduce aggression and stress. Providing more floor space at the upper end of the range studied (1.45 to 2.9 m2/sow) should allow the dominance hierarchy in the group to be established quickly with minimal aggression and stress, and thus, minimum risk to reproduction, with the opportunity to reduce for floor space for the remainder of gestation.

The feed intake of gestating sows is commonly restricted to reduce farrowing and locomotion problems (Meunier-Salaün et al., 2001) and there are a number of feeding methods available to allocate feed to group-housed gestating sows (Bench et al., 2013). Automated overhead drop feeders were used in the previous (Hemsworth et al., 2013) and present experiments to restrictively feed the experimental sows. Drop feeding is common in many countries (Marchant-Forde, 2009), and a number of authors have reported that group-housed sows consume their feed well within 30 min of the feed drop (Csermely and Woodgush, 1990; Brouns and Edwards, 1994). It is, therefore, important to consider the implications of the present results for group housing with other feeding systems. As noted by a number of authors, conventional restricted feeding in group-housed sows increases hunger and, in turn, competition for feed or access to feeding areas (Barnett et al., 2001; Bench et al., 2013; Hemsworth et al., 2013; Verdon et al., 2015). Floor feeding of sows is competitive, but so is accessing feeding stalls or electronic sow feeder stalls. For example, with the provision of feeding stalls in group pens, floor space, either total or outside the feeder, affects aggression and plasma cortisol concentrations (Barnett et al., 1992; Barnett, 1997). Therefore, the results of the present experiment on floor space have implications for group housing with other feeding systems in which sows are restrictively fed.

At the time of conducting this experiment, the recommended minimum floor space allowance for group-housed gestating sows in codes of practice in some countries was 1.4 m2/sow (e.g., the Australian Model Code of Practice for the Welfare of Animals – Pigs [Primary Industries Ministerial Council, 2007]). Furthermore, the major aim of the present experiment was to determine the space allowance at which aggressive behavior and cortisol concentrations are optimal (i.e., unaffected by increasing space). The present results show a linear effect of floor space on both aggression and stress at Day 2 after mixing but not Day 26.

Questions regarding acceptable animal use practices are ethical ones in which animal welfare implications are considered as well as other wider societal interests such as human health, economic, environmental, and social implications (Fisher and Mellor, 2008; Mellor and Bayvel, 2008). Indeed, there will continue to be trade-offs between the attendant conditions, such as profitable animal production as well as human health and social implications, and compromises, such as moral and environmental costs with many farm animal uses and, indeed, other animal uses outside agriculture. The present results and those of Hemsworth et al. (2013) indicate that a strategy of staged-gestation penning, with more space immediately after insemination and less space later in gestation, may address both animal welfare and economic considerations, but clearly, further research is required to examine the effects of reducing space later in gestation.

 

References

Footnotes


Comments
Be the first to comment.



Please log in to post a comment.
*Society members, certified professionals, and authors are permitted to comment.