| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ANIMAL PRODUCTION |
Pork Industry Institute, Department of Animal and Food Sciences, Texas Tech University, Lubbock 79409
 |
Abstract |
|---|
 Top Abstract INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: alternative housing • feeding system • gestation • sow
|
INTRODUCTION |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). This common production system was implemented to optimize pig productivity, decrease labor, and increase disease control. There is growing concern about sow housing systems, and more emphasis is being placed on overall animal welfare (McGlone et al., 2004). Freedom to express normal behavior is an issue of much debate because the gestation crate (i.e., stall) does not allow expression of normal behavior (e.g., turning) and social interaction. When developing new sow housing systems, sow welfare as well as producer requirements should be met (den Hartog et al., 1993). A range of indicators may be used when evaluating the appropriate environment for an animal (Broom, 1997). Therefore, the objective of this study was to evaluate 2 types of penning systems, crates vs. group housing, and 2 feeding systems, a single drop vs. a trickle given over a 30-min period, using measures of sow and litter productivity and sow behavior and physiology. The trickle-feeding system (also called Biofix) was developed in Europe but has not yet been evaluated in North America. |
MATERIALS AND METHODS |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Penning Systems
Pens were constructed using the same fencing and gating materials as was used for the crates. The back two-thirds of each crate were opened to form a 3.07- x 2.16-m pen with four 0.8- x 0.61-m partial stalls, creating 5feeding stalls. This allowed sows or gilts about 6.6 m2 of space or about 1.32 m2 per pig. Crated sows had a 0.61 m wide stall that was 1.08 m tall and 2.1 m in length. This allowed crated sows 1.2 m2 of space in which they could not turn around. If <5 sows were available in a given block, filler (nonstudy) sows were included to allow for similar space allowances and social experiences. The flooring under the crates and pens was concrete slats. The concrete slats included a 15.2-cm slat with a 2.5-cm slot or the full 2.1-m length of the crate. Sows in the pens had the same slatted area that extended the full length of the pen.
Feeding Systems
The feeding tubes entering the crates or pens were identical between the 2 feeding systems. Drop-fed sows had their entire meal delivered in a single moment. Trickle-fed sows had their same daily portion of feed delivered slowly, in a trickle, over a 30-min period. The trickle-feeding system was designed and built by Automated Production Systems (Assumption, IL) as a way to limit-feed group-housed sows; however, it can be used for individually housed sows as well.
Animals and Productivity Measures
A total of one hundred sixty, 7-mo-old gilts (Camborough 22, PIC, Franklin, KY) were eligible to enter the study. The gilts came from a single farm and arrived in 2 groups of 80 gilts each. Gilts with known dates of estrus were moved into a randomly assigned treatment before their next estrus. Gilts were then checked daily for signs of estrus, and when they expressed standing estrus, they were bred by AI. Because the gilts were housed in a common large pen during isolation (i.e., acclimation), they were acquainted before being moved to the breeding and gestation barn.
Gilts and sows that completed gestation were transported about 300 m to common farrowing barns equipped with the previously described farrowing crates (Johnson et al., 2001). The farrowing rate was calculated as the number of gilts or sows that farrowed ÷ the number eligible to be bred x 100. If a gilt or sow did not express estrus or returned to estrus after being bred, she was removed from the study. This procedure yielded an apparently lower farrowing rate on a herd basis than when gilts or sows were rebred and placed in another breeding group. In this study, if gilts or sows were removed from a pen, another pregnant gilt or sow was added to the pen to maintain the same space allowance and group size. These filler sows were not included in the data set for measures of productivity, physiology, or behavior. Each treatment group had filler sows in parity 2 because the farrowing rates in parity 1 were <100%. Little antagonistic behavior was observed, probably because the gilts were reared in a common large pen before initiation of the study.
In the farrowing barn, feed intake of the sows was increased to ad libitum over the first 5 to 7 d postpartum. The diet contained 14% CP as-fed, corn, soybean meal, and vitamins and minerals. Sow lactation feed intake was recorded.
The sows were weighed at farrowing and weaning. The numbers of piglets born alive and dead and the number of pigs weaned were recorded. Piglet BW was determined at birth and weaning. The average weaning age was 22 d.
Productivity measures were taken to determine sow performance and occurrence and severity of lesions. Lesions were visually evaluated by a trained observer on the neck, shoulder, ear, side, back, ham, and feet of all sows and assigned a numerical value to indicate severity, where 1 = mild, 2 = medium, and 3 = severe. A mild lesion included superficial scratches that were not bleeding or infected. A medium lesion included small, deep lesions that had begun to heal with no signs of infection. A severe lesion included deep wounds that had not shown signs of healing and did show possible infection and, in some cases, bleeding. Lesions were assessed 1 wk after weaning when the sows were expected to have the most lesions. Backfat thickness was measured in millimeters at breeding, farrowing, and weaning.
Immune Measures
For parity 1, 78 pregnant gilts were used to collect and analyze immune measures. Blood (20 mL) was drawn from the jugular vein, and sodium heparin as used as the anticoagulant. Immune measures included white blood cell counts and differential cell type determined using a Cell Dyne (Abbott Labs, Santa Clara, CA), neutrophil chemotaxis and chemokinesis (migration), and neutrophil phagocytosis.
Plasma cortisol levels were evaluated by RIA. Blood samples were collected from the external jugular vein with lithium heparin as the anticoagulant. Six hundred microliters of diluted plasma (1:2 dilution) was used to perform the assay. The Coat-a-Count cortisol kit (Diagnostic Products Corporation, Los Angeles, CA) was used, and all samples were analyzed in a single assay. The intraassay CV was 4.1%.
Neutrophil chemotaxis and chemokinesis were determined by methods previously described (McGlone et al., 1993; Morrow-Tesch et al., 1994). Briefly, neutrophils were isolated using 2 gradients of Histopaque 1077 and 1911 (Sigma-Aldrich, St. Louis, MO), and most of the erythrocytes were lysed using 10x PBS (Sigma-Aldrich) and distilled water. Isolated neutrophils were then re-suspended in RPMI medium (Sigma-Aldrich) with a concentration of 1 x 106 cells/mL for the chemotaxis, chemokinesis, and phagocytosis assays. A modified Boyden chamber (Neuro Probe, Cabin John, MD) was used to measure the migration of neutrophils toward media (chemokinesis) or toward 10–8 M recombinant human complement fragment (C5a, Sigma-Aldrich; chemotaxis). The medium and C5a were added to the bottom wells, and neutrophils were added to the top. The Boyden chambers were then incubated in a humidified CO2 chamber for 1 h at 37°C. The filter was removed, fixed, and stained using Hematoxalin stain I and II (Fisher Scientific, Houston, TX). Five fields per well were counted at 1,000x magnification, and the counters were blind to the treatments when the cells were counted.
Isolated neutrophils were also subjected to a phagocytosis assay to determine the percentage phagocytosis of opsonized latex beads (0.807-µL diameter, Sigma-Aldrich). To opsonize the latex beads, a 400-µL aliquot of sterile PBS was mixed with 50 µL of normal porcine serum (collected from healthy nursery pigs at the Texas Tech University Swine unit in New Deal, TX) and 250 µL of latex beads in a 50-mL conical tube. The bead and serum mix was incubated for 1 h in a humidified CO2 chamber at 37°C, and then the beads were washed twice with PBS, centrifuged, and resuspended to 1 x 107 beads/mL. The latex bead solution (500 µL) was added and mixed thoroughly in a conical tube containing each sample of 500 µL of isolated neutrophils to create a 1:5 cell-to-bead ratio. The neutrophil and bead mixture was then incubated for 10 min in a humidified CO2 chamber at 37°C. The tubes were centrifuged at 40 x g for 7 min at 8°C. The supernatant was removed with a pasteur pipette, taking care not to disturb the pellet of beads that had collected on the bottom. Then, 1 mL of media was added, and the solution was vortexed on high speed; this washing procedure was repeated until the solution was clear. The supernatant was removed, and 500 µL of the medium was added. The mixture was vortexed at high speed, ensuring that the pellet containing the beads was fully mixed into solution. A small volume (200 µL) of solution was obtained from the bead and cell mixture and centrifuged at 8,500 rpm and 1,100 x g (Cytofuge 2, model M801-22, StatSpin, Inc., Norwood, MA) for 2 min. A total of 100 neutrophils was counted. For each cell, the number of beads phagocytized was determined as 0, 1, 2, 3, 4, 5, or 
6, and the percentage of cells that phagocytized at least one bead and the average numbers of beads phagocytized were determined.
Behavioral Measures
Behaviors were recorded for 24 h (Dailey and McGlone, 1997b) using time-lapse videography (0.8 frames/s) between d 50 and 75 of gestation. Sow behavior was observed using the Observer 5.1 behavior program (Noldus, Leesburg, VA). The recorded behaviors were drinking, eating, oral-nasal-facial (ONF), sitting, standing, and lying. The social behaviors were antagonistic, mounting, and nonaggressive interactions. From these behaviors, total active behaviors were calculated by subtracting lying behavior. All behavioral measures are defined in Table 1.
|
Chi square was used to evaluate behavioral sequences based on a Markov Chain analysis of the transition matrix using an advanced technique that builds on simple transitional analysis (Dailey and McGlone, 1997a). Briefly, behavioral sequences were counted and entered into a transitional matrix; rows and columns represented the first and second behaviors found in each sequence. This modified Markov Chain method accounts for structural zeros (Goodman, 1983). As an additional set, significant (P < 0.05) sequences were identified and placed in a second Chi-square table that included each sequence in rows and treatments in columns. Behavioral sequences were identified that were found at greater or lower frequencies than expected because of random chance. Behavioral trills were identified in this analysis; a trill is a repeating sequence of behavior (e.g., A-B-AB-A-B, where A and B are unique behaviors), which indicates a linked behavioral mechanism. Behaviors were then placed in a diagram that indicated which behavioral sequences were found at greater or lower than expected levels as assessed by Chi-square analysis.
|
RESULTS |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). During parity 1, gilts in pens had a tendency (P < 0.10) to have a lower farrowing rate (penned gilts had an 11% lower farrowing rate than crated gilts). In parities 1 and 2, gilts and sows fed by drop or trickle systems did not differ in farrowing rate. In parity 1, the relationship between penning and feeding system was significant in the Chi-square analysis (P < 0.05). Gilts in crates that were trickle fed had the greatest farrowing rate (7 to 15% greater than other treatment groups). In parity 2, no significant differences among penning or feeding system were identified in farrowing rate (Table 2).
|
). At weaning, sows housed in pens during gestation and all sows lactating in farrowing crates had 12% greater (P < 0.05) back-fat thickness than those sows housed in crates (Table 3). Body weight was different (P < 0.05) among treatments at farrowing (Figure 1). Farrowing weight was lower (P < 0.05) in drop-fed crated sows compared with trickle-fed crated sows, but there was no difference between drop-fed penned sows and trickle-fed penned sows. Body weights of sows that farrowed showed a significant parity by feeding system effect (P < 0.01). However, the means were very similar, and the effect was probably not biologically important (<4 kg difference among treatment groups).
|
|
), and the overall level of skin lesions and wounds was low. All other productivity measures were not significantly different among treatments.
Immune Measures
Immunological response was not greatly influenced by housing or feeding systems (Table 4). However, the percentage of phagocytosis by neutrophils was lower (P < 0.05) for the trickle-fed penned sows and the drop-fed crated sows vs. the drop-fed penned sows and the trickle-fed crated sows (Figure 2). Plasma cortisol was not different among treatments (Table 3), and mean values in the treatment groups were very close to each other.
|
|
). Low frequencies and durations of agonistic behaviors were observed in all treatment groups. Significant (P < 0.05) interactions were observed among the 4-h periods and treatments (Figure 3). During the 0800 period, drop-fed crated sows displayed more (P < 0.05) ONF and active behaviors than did gilts and sows in other treatment groups.
|
|
). However, 1 h after feeding, there were no differences for pen entering and stall exiting among feeding systems.
|
, the level of frequency of a given sequence is greater, lower, or not different than expected because of random chance in the Chi-square analysis. Thus, the thickness of the arrows in Figure 5 indicates relative differences in number of behavioral sequences observed among treatment groups.
|
Another way to evaluate sequencing results is to compare the industry standard system (drop feeding and crate housing) to other groups. In parity 1, pregnant gilts in the drop-fed, pen housing treatment had high frequencies of the ONF behavior-drink-stand triangle. In parity 2, crated sows had a high level of ONF behavior-drink trill; sows on the drop-fed, pen housing treatment and on the trickle-fed, pen housing treatment had fewer ONF behavior-drink trills and more stand-ONF behavior trills (trickle feeding and pen housing) and more ONF behavior-stand trills (drop feeding and pen housing).
|
DISCUSSION |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Varley and Stedman, 1994). From an evolutionary perspective, females would have reduced fitness if severely stressed, and this would lower reproductive rates in an attempt to limit the population growth until more favorable times. Langendijk et al. (2000) suggested that group housing and the associated social stress could affect the onset of ovulation and behavior related to estrus. Reduced sow welfare and reduced fitness would be associated with either a reduced farrowing rate or reduced litter size. Because reproductive rates are sensitive to stress, they are an important measure of sow welfare.
Farrowing rate (as measured in the present study) in the pig is influenced by a combination of expression of estrus and maintenance of pregnancy. If stress impacted either the expression of estrus or the maintenance of pregnancy, then farrowing rate would be impacted. Thus, examination of farrowing rate among production systems may indicate whether a given system is stressful. Farrowing rate was influenced by treatment in parity 1, but treatment did not significantly affect farrowing rate in parity 2 (Table 2). Sows had reduced farrowing rates in parity 2, which is a common observation in the field (reduced primiparous sow farrowing rates). In parity 2, sow farrowing rate was reduced regardless of treatment. This result may imply that sows had adapted to the environments by the second parity. Alternatively, gilts that failed to attain or maintain pregnancy were dropped from the study by the second parity; therefore, there were fewer litters farrowed in the second parity. The data indicate that among gilts, the trickle feeding system improved farrowing rate of crated sows by 7% compared with the more common drop feeding system. Overall, gilts in pens had a tendency (P < 0.10) to have 11% lower farrowing rates than did gilts in gestation crates, but no difference was observed in farrowing rates among second parity sows in the treatments evaluated.
The level of backfat thickness in this study was generally low (10 to 13 mm) compared with industry standards. Gilts and sows kept in pens during gestation had greater backfat thickness than gilts and sows in crates. Pregnant females are able to effectively huddle in pens. Because treatment groups were fed identical amounts of feed, the more effective huddling among penned gilts and sows might have conserved heat and lead to greater backfat thickness. If this hypothesis holds, then it might be possible to feed group-penned sows slightly less feed than crated sows to maintain the same body condition.
The number of pigs born per litter ranged from 10.8 to 11.3 (not a significant difference). Still, because subtle differences in birth weights and numbers born per litter may influence or exacerbate numbers and weights of piglets at weaning, we considered these measures relevant. The average weaning weight was lower for piglets from trickle-fed gilts and sows than from drop-fed gilts and sows (Table 3). This effect was smaller (P = 0.06 instead of P = 0.03) when the average number of pigs weaned per litter was included as a covariate. Slowing down the rate of eating (taking 30 min instead of 10 to15 min to eat a meal) might have subtle metabolic consequences.
We selected one time point to determine immune measures during gestation because a previous study of ours (McGlone and Fullwood, 2001) found no interaction between stage of gestation and treatments on immune measures. In addition, we focused on measures that are more responsive to stress in the pig. Neutrophil function, including phagocytosis, is a sensitive measure of stress (Salak et al., 1993; Salak-Johnson et al., 1997). Among crated gilts and sows, trickle feeding increased neutrophil phagocytosis; among penned sows, trickle feeding reduced neutrophil phagocytosis.
Sows in the drop-fed, penned housing treatment group had elevated farrowing weights and elevated neutrophil phagocytosis compared with sows in the drop-fed, crated housing treatment group (Figures 1 and 2). Among trickle-fed sows, the findings were opposite; trickle-fed penned sows had lower farrowing weights and lower neutrophil phagocytosis than did trickle-fed crated sows. Differential weight gain while on the same controlled-calorie diet may indicate the relative stressfulness of an environment (among other possible explanations) because nutrient utilization is negatively affected by stress (McGlone and Curtis, 1985). Similarly, neutrophil phagocytosis is expected to be suppressed when sows are stressed (Salak et al., 1993). Therefore, based on these 2 diverse measures, penned sows are expected to be more stressed when drop-fed, and crated sows are expected to be more stressed when trickle-fed. However, many measures were not influenced by any of the treatments.
The overall proportion of sows that displayed stereotypies was reported to be lower among group-housed sows compared with individually housed sows in some studies (Broom et al., 1995; Vieulle-Thomas et al., 1995). We did not find this effect for trickle-fed sows in pens or crates, but it was the case during one period for crated, drop-fed sows. The occurrence of ONF and active behaviors increased at 0800 (postfeeding period) for those sows housed in individual stalls compared with the other treatments (Figure 3). However, for the overall 24-h period, stereotypies did not differ among treatments. This finding was consistent with the results of Backus (1997), who showed that stalled sows and trickle-fed groups of 6 to 8 sows showed similar ONF activities during parity 1.
The presence of high levels of stereotypies may be an indicator of an inadequate social or physical environment (Barnett et al., 2001) or elevation in brain stress neuropeptides (Salak-Johnson et al., 1997). Alternatively, pregnant sows may have a strong motivation to express oral behavior regardless of environment (Dailey and McGlone, 1997b). Von Borrell et al. (1992) classified stereotypies as repetitive behavioral actions of unusual form or function characterized by sequences of repetitive movements. Stereotypies are, by definition, trills; however, not all trills are stereotypies. Quantitative behavioral sequence data are often not summarized in studies of farm animal behavior. However, behavioral sequences may provide insights into complex effects of environments on brain and behavior. Behavioral sequences associated with stress in other models are being investigated, for example among young pigs, injection of corticotrophin-releasing factor (CRF) into the lateral cerebral ventricle elicits extensive ONF behaviors (Salak-Johnson et al., 1997). Also, Von Borrell and Hurnik’s study (1991) found that sows stimulated with adrenocorticotropin releasing hormone increase stereotypical behaviors. Recently, it has been shown that the trill ONF behavior-stand is a reliable indicator of activation of brain CRF (Salak-Johnson et al., 2004). We have recently completed studies that showed that central CRF also causes exaggerated ONF behaviors among crated sows (unpublished observations). First parity drop-fed gilts had an elevated ONF behavior-stand trill compared with all other treatments, but by the second parity, trickle-fed penned sows displayed increased ONF behavior-stand sequences compared with crated sows. Based on the behavioral bioassay developed by brain injections of CRF in young pigs and sows, we hypothesized that the gilts and sows in the drop-fed pen and sows in the trickle-fed pen might have experienced elevated brain CRF. By this possible behavioral bioassay for central CRF, penned gilts and sows [compared with those in stalls and second parity sows (in all treatments)] might have had small increases in brain CRF. These hypotheses will have to be tested in other studies. Interestingly, gilts and sows in stalls that were drop-fed expressed longer durations of ONF behaviors without the increase in the ONF behavior-stand sequence. Gilts were so concentrated on ONF behaviors that they had fewer sequences of entering and exiting from this behavior as compared with the ONF behavior-stand sequence. These results lead to another hypothesis, i.e., that the brain mechanisms for expression of enhanced ONF may be different than for enhanced expression of the ONF behavior-stand sequence.
Meunier-Salaün and Dantzer (1990) reported that pigs exhibited a threshold to adaptive capabilities in their physical and social environments. Once a threshold was reached, behavioral alterations appeared with possible negative effects on health and performance. When feeding and penning systems were previously evaluated (Backus, 1997; Boyle et al., 2000), sows housed in groups had an increased occurrence of locomotive disorders and a greater occurrence of lesions after weaning. Van Putten and van de Burgwal (1990) suggested that stressful situations may be one of several factors associated with agonistic behaviors (and wounds). The mixing of sows can lead to concerns of increased aggression, although sow dominance order is quickly established within the group. Reproductive performance within group-housed sows can be equal to that of individually crated sows with good management practices (Arey and Edwards, 1998).
No differences in skin lesions were observed after farrowing and weaning. This study was unique in that gilts were previously acclimated to social groups for a minimum of 2 mo before the study started. In this model, the mixing of gilts and even second parity sows was associated with no measurable agonistic behavior and very few wounds, lesions, or injures. This study provided breeding-eligible gilts and regrouped sows the least stressful experience that is practically possible.
Overall productivity was not greatly influenced by housing or feeding system. This observation is consistent with other reports (England and Spurr, 1969; Brouns and Edwards, 1992; McGlone et al., 2004), suggesting that housing systems in general have few consistent effects on overall productivity.
Von Borrell et al. (1992) suggested that sow health and production were not compromised by housing sows in small social groups. This study showed that the immune measures were comparable for the crated and penned sows, indicating that no benefits or disadvantages with immune response occurred with either housing or feeding system for parity 1 sows, which is in agreement with the results from Nind et al. (1997). The suppressed percentage of phagocytosis observed with the trickle feeding, pen housing treatment might have been due to an increase in competition during feeding. Cortisol levels are often used as an indicator of stress. There were no differences in cortisol levels among treatments. Barnett et al. (1985) also reported that there were no differences in cortisol levels among different sow housing systems; this may be indicative of a lack of severe stress expressed by the sow during the gestation period.
In conclusion, the environments evaluated did not have major effects on gilt or sow physiology or reproductive performance. Crated, drop-fed sows expressed more stereotyped ONF behaviors during the 4-h postprandial period but had few physiological or performance signs of stress. Penned sows, which may experience low levels of social stress, had elevated ONF behavior-stand trills that may indicate low level activation of central CRF but few other physiological or performance signs of stress.
|
Footnotes |
|---|
2 Corresponding author: john.mcglone{at}ttu.edu
Received for publication October 13, 2004. Accepted for publication November 29, 2005.
|
LITERATURE CITED |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. D. Schneider, M. D. Tokach, S. S. Dritz, J. L. Nelssen, J. M. DeRouchey, and R. D. Goodband Effects of feeding schedule on body condition, aggressiveness, and reproductive failure in group-housed sows J Anim Sci, December 1, 2007; 85(12): 3462 - 3469. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
