- Email: [email protected]
Stereotypic behavior, adrenocortical function, and open field behavior of individually confined gestating sows
Physiology & Behavior, Vol. 49, pp. 709-713. ©Pergamon Press plc, 1991. Printed in the U.S.A.
0031-9384/91 $3.00 + .00
Stereotypic Behavior, Adrenocortical Function, and Open Field Behavior of Individually Confined Gestating Sows EBERHARD
V O N B O R E L L 1 A N D J. F R A N K H U R N I K 2
Department of Animal and Poultry Science, University of Guelph Guelph, Ontario, Canada N I G 2W1 R e c e i v e d 19 D e c e m b e r 1990
VON BORELL, E. AND J. F. HURNIK. Stereotypic behavior, adrenocortical function, and open field behavior of individually confined gestating sows. PHYSIOL BEHAV 49(4) 709-713, 1991.--The adrenocortical response and open field behavior of a random sample of 37 individually confined gestating sows in different parities were tested around day 85 of pregnancy. The sows were classified as stereotyped [S] and nonstereotyped [N] based on 8-h individual behavioral observations during daytime. Behavioral analysis from time-lapse video recordings included percentage of time spent standing and sitting, as well as the duration spent performing elements of stereotyped actions. Blood samples were drawn for cortisol analysis by suborbital sinus puncture before and after adrenocortical stimulation with 200 IU ACTH. Locomotor activity in an open field, isolated visually and auditorily from other sows, was also studied. Seventeen sows exhibited stereotyped behavior for 54.9---4.8% of the 8-h observation period during daytime. The total time the sows spent standing and sitting was positively correlated with age and was significantly higher for IS] sows than for [N] sows. Sows in the [S] group exhibited a higher adrenocortical response to ACTH stimulation than [N] sows. Mean locomotor scores in the open field were higher for [S] than for [N] sows but these did not correlate with the physical activity and adrenocortical function in the home crate. Our results provide no evidence that the performance of stereotypies constitutes a mechanism that helps sows to cope in an environment of low complexity. Stereotypies Gestating sows
Adrenocortical function
Open field behavior
THE most apparent attribute in the various attempts to define stereotyped behavior is that these behaviors appear to serve no clear function in the context of their performance [see, e.g., (12, 15, 20)]. Other characteristics, such as operational uniformity and frequent repeatability, are not exclusive to stereotyped behaviors. They are also components of many other behavioral patterns such as locomotion, eating, drinking, etc. Stereotyped behavior often is shown by sows housed in an environment of low complexity and this behavior is thought to develop in chronic stress situations and to indicate past or current conflict or frustration (4, 7, 11, 29). Stereotypies may serve as a means of compensating for such an insufficiently stimulating environment. Some stereotyped activities, therefore, approximate responses to stimuli which are absent or severely limited in the present environment. Studies with tethered sows (4,5) and caged bank voles (14) have been interpreted as showing that the performance of stereotypies are associated with the release of opioids and, therefore, may function as an effective coping strategy. A more recent study indicates that there is no evidence that performance of behavioral stereotypies results in increased opioid activity (25). The predominant stereotyped behaviors of individually confined sows are oral
Chronic stress
Coping
Housing
activities which are either directed to the environment (i.e., chewing, licking and pushing the bars, feeder, drinker) or self-directed (tongue rolling, vacuum chewing or chomping). In addition to the limited opportunities for exploration in a restrictive confinement, feeding is reduced to one or two meals per day. High concentrate diets require a very short time for consumption, further increasing the sows' occupational deprivation. Investigating these factors led to the conclusion that stereotyped behavior may be adjunct to the feeding period (24) and caused by food restriction (1). Stereotyped activities in an environment of inadequate complexity could either be caused by under- or over-stimulation (15). The functional role of stereotypies in an environment of low complexity are interpreted differently. Thus Walsh and Cummins (30) suggest that an animal in a restricted environment becomes increasingly sensitive to stimulation in an attempt to achieve an optimal level of arousal. The supply of environmental stimuli significantly reduced stereotypies in caged canaries, suggesting that "stereotypies occur primarily to supply the birds with a source of stimulation" (13). On the other hand, Dantzer and Morm~de (8) suggest that the performance of stereotyped behavior has a de-
1Present address: Department of Animal Science, Iowa State University, 337 Kildee Hall, Ames, IA 50011. 2Requests for reprints should be addressed to J. Frank Hurnik, Department of Animal and Poultry Science, University of Guelph, Guelph, Ontario, Canada N1G 2W1.
709
710
VON BORELL AND HURNIK
arousal function which enables the animal either to dissipate tension, frustration or anxiety engendered by the situation with which it is confronted, or to switch attention away from the situation. This conclusion was based on a study in food-deprived pigs, where stereotyped repeated chain pulling between food deliveries was accompanied by decreased pituitary-adrenal activity. However, a more recent study with different schedules of food presentations indicated that stereotyped oral activities in pigs do not serve the same function in different situations (9). Stereotyped behavior can be induced in animals by restrictive housing conditions or by stimulant drugs (23). Artificially induced stereotypies in rats result in increased stereotyped locomotor activity when they are placed in an open field (28). From studies in laboratory animals it is well known that novelty (22) and other psychological stress increases the activity of the adrenal-pituitary axis (18). Animals that are reared in social isolation are generally more active in novel environments and exhibit more intense stereotyped behavior when treated by stimulants [i.e., amphetamine (26)]. In pigs, a high adrenocortical responsiveness was associated with high locomotor activity accompanied with frequent vocalizations and escape attempts in an open field (yon Borell and Ladewig, in preparation). So far, it is not known if animals that develop environmentally induced stereotypies differ in their adrenocortical response and open field behavior from those animals which are not exhibiting stereotyped activities. The objective of this study was to determine the relationship between stereotyped behavior, adrenal function and open field behavior of individually confined sows. Special emphasis was placed on the relationship between the occurrence and duration of stereotyped behavior and the physical activity (time spent standing and sitting) to the adrenocortical function, since elevated plasma corticosteroid levels and increased adrenocortical responses to exogenous stimuli were both shown to serve as indicators of chronic stress in intensively housed pigs (2,3) and laboratory animals (21,27). Moreover, the hypothesis was tested that sows performing stereotypies differ in their locomotor activity in an open field from sows that do not show stereotyped behavior.
Thirty-seven Yorkshire sows in parity 1-5 were housed under commercial conditions in a breeding-gestation room. They were placed in individual stalls (108 stalls in 3 rows; each stall 0.66 x 2.10 m) on partially slatted concrete floor, each with a feeder in front. Each day at 0800 h they were fed automatically 2.0-2.5 kg of a commercial concentrate diet, formulated to meet the NRC energy requirements for gestating sows. Newly introduced sows were allocated a stall, by barn staff, according to availability and convenience. Lights were on between 0730 and 1730 h and room temperatures varied between 16 and 20°C.
to record the behavior of one sow a day. All occurrences and durations of standing, sitting, lying and stereotyped activities were observed, and each mutually exclusive behavioral action was later expressed as a percentage of total observation time. Physical activity was defined as the time spent standing and sitting, inactivity as the time spent lying. Stereotyped behavior was identified as regularly repeated, identical movements > 1 min in duration. Intervals of less than 10 s between bouts of stereotyped action were ignored. According to the definitions given by Cronin (4) the following elements of oral stereotyped behavior were observed: chewing, biting, licking, sucking, rooting, nosing, pressing and pushing, all self-directed or directed to objects in the environment, such as bars, the drinker, feeder and floor. Following these definitions and time limitations, 17 of 37 sows were classified as stereotyped [S] and 20 sows as nonstereotyped [N]. Nearly all stereotyped actions were performed while standing and sitting. Although the temporary sham-cbewing of two [S] sows was performed while lying down, this was included in the time spent standing and sitting. The open field tests were conducted in chronological order on sow number 16 through 37 (n= 22), which were later tested for their response to ACTH. They were moved after each 8-h observation period to a weighing/sorting room outside hearing range of their home stalls. The open field was a pen of 14 m 2 (5.3 × 2.65 m) bounded by vertical bars and with a solid concrete floor. A fluorescent white light above the pen provided normal working light. Each sow was monitored for 10 min by a video camera which was mounted on the ceiling directly above the field and connected to a video recorder in a separate room. The recordings immediately started when sows were placed in the open field. The image appearing on the screen was divided into 18 equal sized squares which were marked on the monitor screen (Staedler Lumocolor 317 pen) and each square represented an area of about 0.88 x 0.88 m. Locomotion-score in the open field was later determined by counting the number of squares which the right ear of the sow entered. Between 10.00 and 11.00 h on the day after behavioral observations, blood samples for cortisol analysis were drawn from 22 sows by suborbital sinus puncture, before and 60 min after adrenocortical stimulation with 200 IU ACTH IM (Acthar, USV Canada Inc.) dissolved in 5 ml saline (3). This blood collection method was reported to have no effect on plasma cortisol levels if the samples were drawn within 5 rain (10). Our samples were drawn within 2 min. After collection of blood in heparinized tubes, samples were centrifuged and plasma stored at - 2 0 ° C until analysis. Plasma cortisol concentrations were determined by radioimmunoassay using rabbit anticortisol-21-thyroglobulin (Lot No. 127F-4838, Sigma, St. Louis, MO). Details on the procedure were described by Liptrap and Raeside (16). The intra- and interassay coefficients of variation in the present experiment were 4.88% and 9.77% respectively.
Procedure
Statistical Analysis
A random sample of sows around day 85 of pregnancy were selected. This stage of pregnancy was chosen because the number of stereotyped elements of behavior and their duration increases to day 85 of the pregnancy (4). Behavior recorded continuously for 4 consecutive days around day 85 revealed very little day to day variation in the duration of stereotyped behavior within individuals (von Borell and Hurnik, in preparation). The classifications stereotypic [S] and nonstereotypic IN] were based on 8-h behavioral observations from 0830 to 1630. A video camera was placed in front of the individual stalls and a Panasonic 6050 time-lapse video recorder (12-h time mode) was used
Differences in behavioral activity and adrenocortical response between [S] and [N] sows were tested for significance by Student's t-test. Spearman's rank correlation test was used to compare behavioral traits with the adrenocortical response and age. Differences between groups were expressed as means "4-SEM and the level of significance was p<0.05.
METHOD
Animals and Housing
RESULTS
Seventeen of the 37 sows exhibited stereotyped behavior for an average of 54.9-+4.8% of the 8-h observation period. The
STEREOTYPIES AND ADRENAL FUNCTION IN SOWS
711
lOO
~100
9o
~_- 90
t'-
"o 80 e,-
ao
~e.-
"~ 70
*
-~ 60
60
~
50
!
20
~ lO 0
312
parity I
70
parity 2 parity 3
50
~ 40 e~
~
30
~ 2o "~ 10 ~
532
parity 4 parity5
FIG. 1. Mean percentage of time sows that exhibit stereotyped behavior [S] and sows that did not [N] spent standing and sitting in an 8-h period during parity 1-5. a, b: Means of [ALL] denoted by different letters over parity 1-5 are different (p<0.05, t-test). *Indicates difference between [S] and [N] within parity (p<0.05, t-test). Solid bars: All; heavily hatched bars: S; lightly hatched bars: N; cross-hatched bars: proportion of time performing stereotypies while standing and sitting.
time spent standing and sitting was positively correlated with age (rs = .43, n = 37, p<0.01) and parity (Fig. 1). Thus parity 4 and 5 sows exhibited more standing and sitting than parity 1-3 sows (p<0.05). Differences in the time [S] and IN] sows spent standing and sitting were highly significant (73.2-+5.1% vs. 46.1 -+ 3.9%, t = 4 . 2 9 , p<0.001). These differences were also significant within parities 2 and 3 (t=2.32/2.51, p<0.05, Fig. 1). The low number of sows within parities 1, 4 and 5 does not allow meaningful statistical analysis. The most obvious differences in daytime activity patterns between [S] and [N] sows occurred at midday (12.00 h), when about two-thirds of the [S] sows, but only one-third of the [N] sows were standing and sitting (Fig. 2). Stereotyped behavior was self-directed 60% of the time and 40% of the time was directed toward objects in the environment (Fig. 3). However, there was no relationship between time spent in the two types of behavior and age or parity. The adrenocortical response to ACTH was higher for IS] ( n = 9 ) than for IN] ( n = 1 3 ) sows (154.5-+25.7 ng/ml vs. 105.1---13.2 ng/ml cortisol, t = 1.86, p = 0 . 0 3 9 ; Fig. 4), but [S] and [N] sows did not differ in the basal plasma cortisol values (15.5---4.6 ng/ml vs. 10.4 -+ 1.8 ng/ml, t = 1.16, p = 0 . 1 3 1 ) , although the relative extent of cortisol release before and after stimulation was significantly correlated (rs = .79, p<0.01). Open field locomotor scores differed significantly between [S] and [N] sows (188.3-+ 26.2 vs. 146.4-+ 11.2, t = 1.75,p=0.049), despite a high individual variation. Locomotor scores ranged from 95 to 260 line crossings for [S] and 67 to 194 for [N] sows. The scores were not correlated either to the time spent standing and sitting in the home pen (rs = . 17, p>0.05) or to the adrenocortical response to ACTH (rs = - .25, p>0.05). DISCUSSION
The current study supports the concept that stereotypies in individually housed sows are indicative of a higher level of either
0
r
i
9
10
i
11
12 13 14 observation time
15
16
17
FIG. 2. Percentage of the numbers of sows that exhibit stereotyped behavior [S] and sows that did not [N] spent standing and sitting between 0830 and 1630 h.
behavioral or adrenocortical arousal. Behaviorally, this high level of excitability in [S] sows was not only linked to the housing environment but also to the unfamiliar situation in an open field, since these animals showed higher locomotor activities. Thus the increased locomotor response to the open field by sows with "natural" stereotypies appears similar to that observed in rats with stimulant-induced stereotypies (28). The high proportion of observation time the [S] sows spent on stereotyped behavior seemed to account for the significantly higher percentage of time spent standing and sitting compared to [N] sows. The mean proportion of observation time that sows performed stereotypies was within the same range as in Cronin's (4) study (43.6---31.5%, mean_+SD), observing sows between 10.00 and 14.00 h. The mean ratio of time between self-directed and object-directed ste-
CS 59.3
CFL CF 1.1 0
9.2
CB 7.1
CD 6.5
y 5.9 RB 5.9
FIG. 3. Mean percentage of time sows that exhibit stereotyped behavior spent on different stereotyped elements in relation to the total time spent on stereotypies. Cross-hatched section: self-directed; hatched section: directed to objects. Action: C = chewing, biting, licking, sucking; R = rooting, nosing, pressing, pushing. Substrate: B = bars, D = drinker, F = feeder, FL = floor, S = sham.
712
VON BORELL AND HURNIK
200 180 160
14o
I
o. 120
c 100 o
._~
80
o o
60 40 20
basal
ACTH-induced
FIG. 4. Mean adrenocortical response of [S] and [N] sows before and 60 min after stimulation with ACTH. *Indicates difference between [S] and IN] sows (p<0.05, t-test).
reotypies in the present study was about 60 to 40%. Differences in this ratio between sows of different parities could not be tested statistically due to the low number of sows in higher parities. Changes from predominantly environment-directed to mostly selfdirected stereotypies in high parity sows was reported from studies monitoring the development of stereotypies over long periods of time (4,29). The qualitative change between object- and selfdirected stereotypies over time, may be a result of enhanced short-duration elements of stereotyped actions leaving less time for contact with the initially chosen objects (7). It was suggested that after initial environmentally induced activation, stereotypies would function as opioid-mediated coping mechanisms (5,14). The progressive increase in physical activity with age and parity, combined with a high performance rate of stereotypies and enhanced adrenocortical sensitivity shown by [S] sows in our study, do not support this concept. Additionally, there is no evidence that performance of stereotypies in sows results in increased opioid activity (25). Dantzer (7) proposed that stereotypies simply gain strength because of the positive feedback effect of sensory stimulation on their underlying control system, which leads to a progressive sensitization of these neural systems. The question why most sows in the restrictive environment did not develop stereotyped behavior remains unanswered. The
lower adrenocortical response to stimulation in [N] sows suggest that individuals react differently to the same "stressful" condition, depending on their ability to cope with the condition. It has been proposed that differences in the amount of, or vulnerability to, prior stress experienced by rats and humans may account for the extreme variability in response to a stimulant that induces stereotypies (17). Adrenocortical hyperreactivity has been demonstrated in mice (21) and rats (27) under chronic stress conditions. Similar responses were shown by tethered pigs (3) or pigs housed at high stocking densities (19). Thus the stereotyped activities could be either interpreted as an indicator of the experienced chronic stress, or of genetic differences between sows in the sensitivity of the adrenal cortex. Apart from the psychophysiological aspect it could well be that the increased adrenal function reflects an adaptive mechanism to fulfil the higher energy requirements of the body. Corticosteroids are known as mobilizing hormones which are necessary for adjusting the body to stressful environmental conditions (i.e., acclimation to cold temperatures). This could explain why [N] sows in the same housing system, spending 27.1% less time standing and sitting, had a lower adrenocortical response to ACTH. Cronin et al. (6) found that sows with a high level of stereotypies produced 35.7% more heat than those sows with a low level. The increased activity-related heat production was mostly due to the performance of stereotypies. The present study has shown that stereotyped behavior in gestating sows is correlated with high physical activity and an increased adrenocortical sensitivity to acute stimulation. On the basis of our results and those from the growing body of literature on stereotypies in pigs we cannot conclude that the performance of stereotypies constitutes a mechanism that helps animals to cope in an environment of low complexity. A coping response is directly linked to the maintenance of homeostasis. The consequence of such a coping response would include an adjustment to a new state of homeostasis. But since the frequency and duration of stereotypies increase with age and duration of confinement there seem to be no indications of a coping response in those animals. However, the lower adrenocortical response and the failure to develop stereotyped behavior in [N] sows does not necessarily mean that these sows are not affected by their housing conditions. ACKNOWLEDGEMENTS The authors gratefully acknowledge the financial support received from the German Research Council (Deutsche Forschungsgemeinschaft), Bonn, Germany and from the Natural Sciences and Engineering Research Council of Canada (NSERCC). The authors wish to thank the Ontario Ministry of Agriculture and Food for provision of experimental animals and housing facilities, and R. Liptrap for the hormone analysis. We also thank J. Rushen for his useful comments and discussions.
REFERENCES 1. Appleby, M. C.; Lawrence, A. B. Food restriction as a cause of stereotypic behaviour in tethered gilts. Anim. Prod. 45:103-110; 1987. 2. Barnett, J. L.; Winfield, C. G.; Cronin, G. M.; Hemsworth, P. H.; Dewar, A. M. The effect of individual and group housing on behavioural and physiological responses related to the welfare of pregnant pigs. Appl. Anim. Behav. Sci. 14:149-161; 1985. 3. von Borell, E.; Ladewig, J. Altered adrenocortical response to acute stressors or ACTH(1-24) in intensively housed pigs. Dom. Anim. Endocrinol. 6:299-309, 1989. 4. Cronin, G. M. The development and significance of abnormal stereotyped behaviours in tethered sows. Ph.D. Thesis. The Agricultural University of Wageningen, The Netherlands; 1985. 5. Cronin, G. M.; Wiepkema, P. R.; van Ree, J. M. Endorphins implicated in stereotypies of tethered sows. Experientia 42:198-199; 1986. 6. Cronin, G. M.; van Tartwijk, J. M. F. M.; van der Hel, W.; Verste-
7. 8. 9. 10.
gen, M. W. A. The influence of degree of adaptation to tether-housing by sows in relation to behaviour and energy metabolism. Anim. Prod. 42:257-268; 1986. Dantzer, R. Behavioral, physiological and functional aspects of stereotyped behavior: a review and reinterpretation. J. Anim. Sci. 62: 1776-1786; 1986. Dantzer, R.; Mormtde, P. De-arousal properties of stereotyped behaviour: evidence from pituitary-adrenal correlates in pigs. Appl. Anim. Ethol. 10:233-244; 1983. Dantzer, R.; Gonyou, H. W.; Curtis, S. E.; Kelley, K. W. Changes in serum cortisol reveal functional differences in frustration-induced chain chewing in pigs. Physiol. Behav. 39:775-777; 1987. Evans, F. D.; Christopherson, R. J.; Aheme, F. X. Development of the circadian rhythm of cortisol in the gilt from weaning until puberty. Can. J. Anim. Sci. 68:1105-1111; 1988.
S T E R E O T Y P I E S A N D A D R E N A L F U N C T I O N IN S O W S
11. Fraser, D. The effect of straw on the behaviour of sows in tether stalls. Anita. Prod. 21:59--68; 1975. 12. Hurnik, J. F.; Webster, A. B.; Siegel, P. B. Dictionary of farm animal behaviour. University of Guelph, Guelph: Print Publication Services; 1985. 13. Keiper, R. R. Studies of stereotypy function in the canary (Serinus canarius). Anim. Behav. 18:353-357; 1970. 14. Kennes, D.; Odberg, F. O.; Bouquet, Y.; De Rycke, P. H. Changes in naloxone and haloperidol effects during the development of captivity-induced jumping stereotypy in bank voles. Eur. J. Pharmacol. 153:19-24; 1988. 15. Kiley-Worthington, M. Stereotypies. In: Behavioural problems of farm animals. London: Oriel Press; 1977:74-79. 16. Liptrap, R. M.; Raeside, J. I. Effect of cortisol on the response to gonadotrophin releasing hormone in the boar. J. Endocrinol. 97:7581; 1983. 17. MacLennan, A. J.; Maier, S. F. Coping and the stress-induced potentiation of stimulant stereotypy in the rat. Science 219:1091-1093; 1983. 18. Mason, J. W. A historical view of the stress field, Part I and II. J. Hum. Stress 1:6-12, 22-36; 1975. 19. Meunier-Salaun, M. C.; Vantrimponte, M. N.; Raab, A.; Dantzer, R. Effect of floor area restriction upon performance behavior and physiology of growing-- finishing pigs. J. Anim. Sci. 64:1371-1377; 1987. 20. Odberg, F. O. Abnormal behaviours: (stereotypies). In: Proc. 1st World Congress on Ethology Applied to Zootechnics. Madrid: 1978: 475-480. 21. Ortiz, R.; Armario, A.; Castellanos, J. M.; Balasch, J. Post-wean-
713
22. 23.
24. 25. 26.
27.
28. 29.
30.
ing crowding induces corticoadrenal hyperreactivity in male mice. Physiol. Behav. 34:857--860; 1985. Pfister, H. P. The glucocorticosterone response to novelty as a psychological stressor. Physiol. Behav. 23:649--652; 1979. Randrup, A.; SCrensen, G.; Kobayashi, M. Stereotyped behaviour in animals induced by stimulant drugs or by a restricted cage environment: relation to disintegrated behaviour, brain dopamine and psychiattic disease. Jpn. J. Psychopharmacol. 8:313-327; 1988. Rushen, J. Stereotyped behaviour, adjunctive drinking and the feeding periods of tethered sows. Anim. Behav. 32:1059-1067; 1984. Rushen, J.; de Passill6, A. M. B.; Schouten, W. Stereotypic behavior, endogenous opioids, and posffeeding hypoalgesia in pigs. Physiol. Behav. 48:91-96; 1990. Sahakian, B. J.; Robbins, T. W.; Morgan, M. J.; Iversen, S. D. The effect of psychomotor stimulants on stereotypy and locomotor activity in socially-deprived and control rats. Brain Res. 84:195-205; 1975. Sakellaris, P. C.; Vernikos-Danellis, J. Increased rate of response of the pituitary-adrenal system in rats adapted to chronic stress. Endocrinology 97:597-602; 1975. Schi~rring, E. An open field study of stereotyped locomotor activity in amphetamine-treated rats. Psychopharmacology (Berlin) 66:281287; 1979. Stolba, A.; Baker, N.; Wood-Gush, D. G. M. The characteristics of stereotyped behaviour in stalled sows by informational redundancy. Behaviour 87:157-182; 1983. Walsh, R. N.; Cummins, R. A. Mechanisms mediating the production of environmentally induced brain changes. Psychol. Bull. 82: 986-1000; 1975.
0031-9384/91 $3.00 + .00
Stereotypic Behavior, Adrenocortical Function, and Open Field Behavior of Individually Confined Gestating Sows EBERHARD
V O N B O R E L L 1 A N D J. F R A N K H U R N I K 2
Department of Animal and Poultry Science, University of Guelph Guelph, Ontario, Canada N I G 2W1 R e c e i v e d 19 D e c e m b e r 1990
VON BORELL, E. AND J. F. HURNIK. Stereotypic behavior, adrenocortical function, and open field behavior of individually confined gestating sows. PHYSIOL BEHAV 49(4) 709-713, 1991.--The adrenocortical response and open field behavior of a random sample of 37 individually confined gestating sows in different parities were tested around day 85 of pregnancy. The sows were classified as stereotyped [S] and nonstereotyped [N] based on 8-h individual behavioral observations during daytime. Behavioral analysis from time-lapse video recordings included percentage of time spent standing and sitting, as well as the duration spent performing elements of stereotyped actions. Blood samples were drawn for cortisol analysis by suborbital sinus puncture before and after adrenocortical stimulation with 200 IU ACTH. Locomotor activity in an open field, isolated visually and auditorily from other sows, was also studied. Seventeen sows exhibited stereotyped behavior for 54.9---4.8% of the 8-h observation period during daytime. The total time the sows spent standing and sitting was positively correlated with age and was significantly higher for IS] sows than for [N] sows. Sows in the [S] group exhibited a higher adrenocortical response to ACTH stimulation than [N] sows. Mean locomotor scores in the open field were higher for [S] than for [N] sows but these did not correlate with the physical activity and adrenocortical function in the home crate. Our results provide no evidence that the performance of stereotypies constitutes a mechanism that helps sows to cope in an environment of low complexity. Stereotypies Gestating sows
Adrenocortical function
Open field behavior
THE most apparent attribute in the various attempts to define stereotyped behavior is that these behaviors appear to serve no clear function in the context of their performance [see, e.g., (12, 15, 20)]. Other characteristics, such as operational uniformity and frequent repeatability, are not exclusive to stereotyped behaviors. They are also components of many other behavioral patterns such as locomotion, eating, drinking, etc. Stereotyped behavior often is shown by sows housed in an environment of low complexity and this behavior is thought to develop in chronic stress situations and to indicate past or current conflict or frustration (4, 7, 11, 29). Stereotypies may serve as a means of compensating for such an insufficiently stimulating environment. Some stereotyped activities, therefore, approximate responses to stimuli which are absent or severely limited in the present environment. Studies with tethered sows (4,5) and caged bank voles (14) have been interpreted as showing that the performance of stereotypies are associated with the release of opioids and, therefore, may function as an effective coping strategy. A more recent study indicates that there is no evidence that performance of behavioral stereotypies results in increased opioid activity (25). The predominant stereotyped behaviors of individually confined sows are oral
Chronic stress
Coping
Housing
activities which are either directed to the environment (i.e., chewing, licking and pushing the bars, feeder, drinker) or self-directed (tongue rolling, vacuum chewing or chomping). In addition to the limited opportunities for exploration in a restrictive confinement, feeding is reduced to one or two meals per day. High concentrate diets require a very short time for consumption, further increasing the sows' occupational deprivation. Investigating these factors led to the conclusion that stereotyped behavior may be adjunct to the feeding period (24) and caused by food restriction (1). Stereotyped activities in an environment of inadequate complexity could either be caused by under- or over-stimulation (15). The functional role of stereotypies in an environment of low complexity are interpreted differently. Thus Walsh and Cummins (30) suggest that an animal in a restricted environment becomes increasingly sensitive to stimulation in an attempt to achieve an optimal level of arousal. The supply of environmental stimuli significantly reduced stereotypies in caged canaries, suggesting that "stereotypies occur primarily to supply the birds with a source of stimulation" (13). On the other hand, Dantzer and Morm~de (8) suggest that the performance of stereotyped behavior has a de-
1Present address: Department of Animal Science, Iowa State University, 337 Kildee Hall, Ames, IA 50011. 2Requests for reprints should be addressed to J. Frank Hurnik, Department of Animal and Poultry Science, University of Guelph, Guelph, Ontario, Canada N1G 2W1.
709
710
VON BORELL AND HURNIK
arousal function which enables the animal either to dissipate tension, frustration or anxiety engendered by the situation with which it is confronted, or to switch attention away from the situation. This conclusion was based on a study in food-deprived pigs, where stereotyped repeated chain pulling between food deliveries was accompanied by decreased pituitary-adrenal activity. However, a more recent study with different schedules of food presentations indicated that stereotyped oral activities in pigs do not serve the same function in different situations (9). Stereotyped behavior can be induced in animals by restrictive housing conditions or by stimulant drugs (23). Artificially induced stereotypies in rats result in increased stereotyped locomotor activity when they are placed in an open field (28). From studies in laboratory animals it is well known that novelty (22) and other psychological stress increases the activity of the adrenal-pituitary axis (18). Animals that are reared in social isolation are generally more active in novel environments and exhibit more intense stereotyped behavior when treated by stimulants [i.e., amphetamine (26)]. In pigs, a high adrenocortical responsiveness was associated with high locomotor activity accompanied with frequent vocalizations and escape attempts in an open field (yon Borell and Ladewig, in preparation). So far, it is not known if animals that develop environmentally induced stereotypies differ in their adrenocortical response and open field behavior from those animals which are not exhibiting stereotyped activities. The objective of this study was to determine the relationship between stereotyped behavior, adrenal function and open field behavior of individually confined sows. Special emphasis was placed on the relationship between the occurrence and duration of stereotyped behavior and the physical activity (time spent standing and sitting) to the adrenocortical function, since elevated plasma corticosteroid levels and increased adrenocortical responses to exogenous stimuli were both shown to serve as indicators of chronic stress in intensively housed pigs (2,3) and laboratory animals (21,27). Moreover, the hypothesis was tested that sows performing stereotypies differ in their locomotor activity in an open field from sows that do not show stereotyped behavior.
Thirty-seven Yorkshire sows in parity 1-5 were housed under commercial conditions in a breeding-gestation room. They were placed in individual stalls (108 stalls in 3 rows; each stall 0.66 x 2.10 m) on partially slatted concrete floor, each with a feeder in front. Each day at 0800 h they were fed automatically 2.0-2.5 kg of a commercial concentrate diet, formulated to meet the NRC energy requirements for gestating sows. Newly introduced sows were allocated a stall, by barn staff, according to availability and convenience. Lights were on between 0730 and 1730 h and room temperatures varied between 16 and 20°C.
to record the behavior of one sow a day. All occurrences and durations of standing, sitting, lying and stereotyped activities were observed, and each mutually exclusive behavioral action was later expressed as a percentage of total observation time. Physical activity was defined as the time spent standing and sitting, inactivity as the time spent lying. Stereotyped behavior was identified as regularly repeated, identical movements > 1 min in duration. Intervals of less than 10 s between bouts of stereotyped action were ignored. According to the definitions given by Cronin (4) the following elements of oral stereotyped behavior were observed: chewing, biting, licking, sucking, rooting, nosing, pressing and pushing, all self-directed or directed to objects in the environment, such as bars, the drinker, feeder and floor. Following these definitions and time limitations, 17 of 37 sows were classified as stereotyped [S] and 20 sows as nonstereotyped [N]. Nearly all stereotyped actions were performed while standing and sitting. Although the temporary sham-cbewing of two [S] sows was performed while lying down, this was included in the time spent standing and sitting. The open field tests were conducted in chronological order on sow number 16 through 37 (n= 22), which were later tested for their response to ACTH. They were moved after each 8-h observation period to a weighing/sorting room outside hearing range of their home stalls. The open field was a pen of 14 m 2 (5.3 × 2.65 m) bounded by vertical bars and with a solid concrete floor. A fluorescent white light above the pen provided normal working light. Each sow was monitored for 10 min by a video camera which was mounted on the ceiling directly above the field and connected to a video recorder in a separate room. The recordings immediately started when sows were placed in the open field. The image appearing on the screen was divided into 18 equal sized squares which were marked on the monitor screen (Staedler Lumocolor 317 pen) and each square represented an area of about 0.88 x 0.88 m. Locomotion-score in the open field was later determined by counting the number of squares which the right ear of the sow entered. Between 10.00 and 11.00 h on the day after behavioral observations, blood samples for cortisol analysis were drawn from 22 sows by suborbital sinus puncture, before and 60 min after adrenocortical stimulation with 200 IU ACTH IM (Acthar, USV Canada Inc.) dissolved in 5 ml saline (3). This blood collection method was reported to have no effect on plasma cortisol levels if the samples were drawn within 5 rain (10). Our samples were drawn within 2 min. After collection of blood in heparinized tubes, samples were centrifuged and plasma stored at - 2 0 ° C until analysis. Plasma cortisol concentrations were determined by radioimmunoassay using rabbit anticortisol-21-thyroglobulin (Lot No. 127F-4838, Sigma, St. Louis, MO). Details on the procedure were described by Liptrap and Raeside (16). The intra- and interassay coefficients of variation in the present experiment were 4.88% and 9.77% respectively.
Procedure
Statistical Analysis
A random sample of sows around day 85 of pregnancy were selected. This stage of pregnancy was chosen because the number of stereotyped elements of behavior and their duration increases to day 85 of the pregnancy (4). Behavior recorded continuously for 4 consecutive days around day 85 revealed very little day to day variation in the duration of stereotyped behavior within individuals (von Borell and Hurnik, in preparation). The classifications stereotypic [S] and nonstereotypic IN] were based on 8-h behavioral observations from 0830 to 1630. A video camera was placed in front of the individual stalls and a Panasonic 6050 time-lapse video recorder (12-h time mode) was used
Differences in behavioral activity and adrenocortical response between [S] and [N] sows were tested for significance by Student's t-test. Spearman's rank correlation test was used to compare behavioral traits with the adrenocortical response and age. Differences between groups were expressed as means "4-SEM and the level of significance was p<0.05.
METHOD
Animals and Housing
RESULTS
Seventeen of the 37 sows exhibited stereotyped behavior for an average of 54.9-+4.8% of the 8-h observation period. The
STEREOTYPIES AND ADRENAL FUNCTION IN SOWS
711
lOO
~100
9o
~_- 90
t'-
"o 80 e,-
ao
~e.-
"~ 70
*
-~ 60
60
~
50
!
20
~ lO 0
312
parity I
70
parity 2 parity 3
50
~ 40 e~
~
30
~ 2o "~ 10 ~
532
parity 4 parity5
FIG. 1. Mean percentage of time sows that exhibit stereotyped behavior [S] and sows that did not [N] spent standing and sitting in an 8-h period during parity 1-5. a, b: Means of [ALL] denoted by different letters over parity 1-5 are different (p<0.05, t-test). *Indicates difference between [S] and [N] within parity (p<0.05, t-test). Solid bars: All; heavily hatched bars: S; lightly hatched bars: N; cross-hatched bars: proportion of time performing stereotypies while standing and sitting.
time spent standing and sitting was positively correlated with age (rs = .43, n = 37, p<0.01) and parity (Fig. 1). Thus parity 4 and 5 sows exhibited more standing and sitting than parity 1-3 sows (p<0.05). Differences in the time [S] and IN] sows spent standing and sitting were highly significant (73.2-+5.1% vs. 46.1 -+ 3.9%, t = 4 . 2 9 , p<0.001). These differences were also significant within parities 2 and 3 (t=2.32/2.51, p<0.05, Fig. 1). The low number of sows within parities 1, 4 and 5 does not allow meaningful statistical analysis. The most obvious differences in daytime activity patterns between [S] and [N] sows occurred at midday (12.00 h), when about two-thirds of the [S] sows, but only one-third of the [N] sows were standing and sitting (Fig. 2). Stereotyped behavior was self-directed 60% of the time and 40% of the time was directed toward objects in the environment (Fig. 3). However, there was no relationship between time spent in the two types of behavior and age or parity. The adrenocortical response to ACTH was higher for IS] ( n = 9 ) than for IN] ( n = 1 3 ) sows (154.5-+25.7 ng/ml vs. 105.1---13.2 ng/ml cortisol, t = 1.86, p = 0 . 0 3 9 ; Fig. 4), but [S] and [N] sows did not differ in the basal plasma cortisol values (15.5---4.6 ng/ml vs. 10.4 -+ 1.8 ng/ml, t = 1.16, p = 0 . 1 3 1 ) , although the relative extent of cortisol release before and after stimulation was significantly correlated (rs = .79, p<0.01). Open field locomotor scores differed significantly between [S] and [N] sows (188.3-+ 26.2 vs. 146.4-+ 11.2, t = 1.75,p=0.049), despite a high individual variation. Locomotor scores ranged from 95 to 260 line crossings for [S] and 67 to 194 for [N] sows. The scores were not correlated either to the time spent standing and sitting in the home pen (rs = . 17, p>0.05) or to the adrenocortical response to ACTH (rs = - .25, p>0.05). DISCUSSION
The current study supports the concept that stereotypies in individually housed sows are indicative of a higher level of either
0
r
i
9
10
i
11
12 13 14 observation time
15
16
17
FIG. 2. Percentage of the numbers of sows that exhibit stereotyped behavior [S] and sows that did not [N] spent standing and sitting between 0830 and 1630 h.
behavioral or adrenocortical arousal. Behaviorally, this high level of excitability in [S] sows was not only linked to the housing environment but also to the unfamiliar situation in an open field, since these animals showed higher locomotor activities. Thus the increased locomotor response to the open field by sows with "natural" stereotypies appears similar to that observed in rats with stimulant-induced stereotypies (28). The high proportion of observation time the [S] sows spent on stereotyped behavior seemed to account for the significantly higher percentage of time spent standing and sitting compared to [N] sows. The mean proportion of observation time that sows performed stereotypies was within the same range as in Cronin's (4) study (43.6---31.5%, mean_+SD), observing sows between 10.00 and 14.00 h. The mean ratio of time between self-directed and object-directed ste-
CS 59.3
CFL CF 1.1 0
9.2
CB 7.1
CD 6.5
y 5.9 RB 5.9
FIG. 3. Mean percentage of time sows that exhibit stereotyped behavior spent on different stereotyped elements in relation to the total time spent on stereotypies. Cross-hatched section: self-directed; hatched section: directed to objects. Action: C = chewing, biting, licking, sucking; R = rooting, nosing, pressing, pushing. Substrate: B = bars, D = drinker, F = feeder, FL = floor, S = sham.
712
VON BORELL AND HURNIK
200 180 160
14o
I
o. 120
c 100 o
._~
80
o o
60 40 20
basal
ACTH-induced
FIG. 4. Mean adrenocortical response of [S] and [N] sows before and 60 min after stimulation with ACTH. *Indicates difference between [S] and IN] sows (p<0.05, t-test).
reotypies in the present study was about 60 to 40%. Differences in this ratio between sows of different parities could not be tested statistically due to the low number of sows in higher parities. Changes from predominantly environment-directed to mostly selfdirected stereotypies in high parity sows was reported from studies monitoring the development of stereotypies over long periods of time (4,29). The qualitative change between object- and selfdirected stereotypies over time, may be a result of enhanced short-duration elements of stereotyped actions leaving less time for contact with the initially chosen objects (7). It was suggested that after initial environmentally induced activation, stereotypies would function as opioid-mediated coping mechanisms (5,14). The progressive increase in physical activity with age and parity, combined with a high performance rate of stereotypies and enhanced adrenocortical sensitivity shown by [S] sows in our study, do not support this concept. Additionally, there is no evidence that performance of stereotypies in sows results in increased opioid activity (25). Dantzer (7) proposed that stereotypies simply gain strength because of the positive feedback effect of sensory stimulation on their underlying control system, which leads to a progressive sensitization of these neural systems. The question why most sows in the restrictive environment did not develop stereotyped behavior remains unanswered. The
lower adrenocortical response to stimulation in [N] sows suggest that individuals react differently to the same "stressful" condition, depending on their ability to cope with the condition. It has been proposed that differences in the amount of, or vulnerability to, prior stress experienced by rats and humans may account for the extreme variability in response to a stimulant that induces stereotypies (17). Adrenocortical hyperreactivity has been demonstrated in mice (21) and rats (27) under chronic stress conditions. Similar responses were shown by tethered pigs (3) or pigs housed at high stocking densities (19). Thus the stereotyped activities could be either interpreted as an indicator of the experienced chronic stress, or of genetic differences between sows in the sensitivity of the adrenal cortex. Apart from the psychophysiological aspect it could well be that the increased adrenal function reflects an adaptive mechanism to fulfil the higher energy requirements of the body. Corticosteroids are known as mobilizing hormones which are necessary for adjusting the body to stressful environmental conditions (i.e., acclimation to cold temperatures). This could explain why [N] sows in the same housing system, spending 27.1% less time standing and sitting, had a lower adrenocortical response to ACTH. Cronin et al. (6) found that sows with a high level of stereotypies produced 35.7% more heat than those sows with a low level. The increased activity-related heat production was mostly due to the performance of stereotypies. The present study has shown that stereotyped behavior in gestating sows is correlated with high physical activity and an increased adrenocortical sensitivity to acute stimulation. On the basis of our results and those from the growing body of literature on stereotypies in pigs we cannot conclude that the performance of stereotypies constitutes a mechanism that helps animals to cope in an environment of low complexity. A coping response is directly linked to the maintenance of homeostasis. The consequence of such a coping response would include an adjustment to a new state of homeostasis. But since the frequency and duration of stereotypies increase with age and duration of confinement there seem to be no indications of a coping response in those animals. However, the lower adrenocortical response and the failure to develop stereotyped behavior in [N] sows does not necessarily mean that these sows are not affected by their housing conditions. ACKNOWLEDGEMENTS The authors gratefully acknowledge the financial support received from the German Research Council (Deutsche Forschungsgemeinschaft), Bonn, Germany and from the Natural Sciences and Engineering Research Council of Canada (NSERCC). The authors wish to thank the Ontario Ministry of Agriculture and Food for provision of experimental animals and housing facilities, and R. Liptrap for the hormone analysis. We also thank J. Rushen for his useful comments and discussions.
REFERENCES 1. Appleby, M. C.; Lawrence, A. B. Food restriction as a cause of stereotypic behaviour in tethered gilts. Anim. Prod. 45:103-110; 1987. 2. Barnett, J. L.; Winfield, C. G.; Cronin, G. M.; Hemsworth, P. H.; Dewar, A. M. The effect of individual and group housing on behavioural and physiological responses related to the welfare of pregnant pigs. Appl. Anim. Behav. Sci. 14:149-161; 1985. 3. von Borell, E.; Ladewig, J. Altered adrenocortical response to acute stressors or ACTH(1-24) in intensively housed pigs. Dom. Anim. Endocrinol. 6:299-309, 1989. 4. Cronin, G. M. The development and significance of abnormal stereotyped behaviours in tethered sows. Ph.D. Thesis. The Agricultural University of Wageningen, The Netherlands; 1985. 5. Cronin, G. M.; Wiepkema, P. R.; van Ree, J. M. Endorphins implicated in stereotypies of tethered sows. Experientia 42:198-199; 1986. 6. Cronin, G. M.; van Tartwijk, J. M. F. M.; van der Hel, W.; Verste-
7. 8. 9. 10.
gen, M. W. A. The influence of degree of adaptation to tether-housing by sows in relation to behaviour and energy metabolism. Anim. Prod. 42:257-268; 1986. Dantzer, R. Behavioral, physiological and functional aspects of stereotyped behavior: a review and reinterpretation. J. Anim. Sci. 62: 1776-1786; 1986. Dantzer, R.; Mormtde, P. De-arousal properties of stereotyped behaviour: evidence from pituitary-adrenal correlates in pigs. Appl. Anim. Ethol. 10:233-244; 1983. Dantzer, R.; Gonyou, H. W.; Curtis, S. E.; Kelley, K. W. Changes in serum cortisol reveal functional differences in frustration-induced chain chewing in pigs. Physiol. Behav. 39:775-777; 1987. Evans, F. D.; Christopherson, R. J.; Aheme, F. X. Development of the circadian rhythm of cortisol in the gilt from weaning until puberty. Can. J. Anim. Sci. 68:1105-1111; 1988.
S T E R E O T Y P I E S A N D A D R E N A L F U N C T I O N IN S O W S
11. Fraser, D. The effect of straw on the behaviour of sows in tether stalls. Anita. Prod. 21:59--68; 1975. 12. Hurnik, J. F.; Webster, A. B.; Siegel, P. B. Dictionary of farm animal behaviour. University of Guelph, Guelph: Print Publication Services; 1985. 13. Keiper, R. R. Studies of stereotypy function in the canary (Serinus canarius). Anim. Behav. 18:353-357; 1970. 14. Kennes, D.; Odberg, F. O.; Bouquet, Y.; De Rycke, P. H. Changes in naloxone and haloperidol effects during the development of captivity-induced jumping stereotypy in bank voles. Eur. J. Pharmacol. 153:19-24; 1988. 15. Kiley-Worthington, M. Stereotypies. In: Behavioural problems of farm animals. London: Oriel Press; 1977:74-79. 16. Liptrap, R. M.; Raeside, J. I. Effect of cortisol on the response to gonadotrophin releasing hormone in the boar. J. Endocrinol. 97:7581; 1983. 17. MacLennan, A. J.; Maier, S. F. Coping and the stress-induced potentiation of stimulant stereotypy in the rat. Science 219:1091-1093; 1983. 18. Mason, J. W. A historical view of the stress field, Part I and II. J. Hum. Stress 1:6-12, 22-36; 1975. 19. Meunier-Salaun, M. C.; Vantrimponte, M. N.; Raab, A.; Dantzer, R. Effect of floor area restriction upon performance behavior and physiology of growing-- finishing pigs. J. Anim. Sci. 64:1371-1377; 1987. 20. Odberg, F. O. Abnormal behaviours: (stereotypies). In: Proc. 1st World Congress on Ethology Applied to Zootechnics. Madrid: 1978: 475-480. 21. Ortiz, R.; Armario, A.; Castellanos, J. M.; Balasch, J. Post-wean-
713
22. 23.
24. 25. 26.
27.
28. 29.
30.
ing crowding induces corticoadrenal hyperreactivity in male mice. Physiol. Behav. 34:857--860; 1985. Pfister, H. P. The glucocorticosterone response to novelty as a psychological stressor. Physiol. Behav. 23:649--652; 1979. Randrup, A.; SCrensen, G.; Kobayashi, M. Stereotyped behaviour in animals induced by stimulant drugs or by a restricted cage environment: relation to disintegrated behaviour, brain dopamine and psychiattic disease. Jpn. J. Psychopharmacol. 8:313-327; 1988. Rushen, J. Stereotyped behaviour, adjunctive drinking and the feeding periods of tethered sows. Anim. Behav. 32:1059-1067; 1984. Rushen, J.; de Passill6, A. M. B.; Schouten, W. Stereotypic behavior, endogenous opioids, and posffeeding hypoalgesia in pigs. Physiol. Behav. 48:91-96; 1990. Sahakian, B. J.; Robbins, T. W.; Morgan, M. J.; Iversen, S. D. The effect of psychomotor stimulants on stereotypy and locomotor activity in socially-deprived and control rats. Brain Res. 84:195-205; 1975. Sakellaris, P. C.; Vernikos-Danellis, J. Increased rate of response of the pituitary-adrenal system in rats adapted to chronic stress. Endocrinology 97:597-602; 1975. Schi~rring, E. An open field study of stereotyped locomotor activity in amphetamine-treated rats. Psychopharmacology (Berlin) 66:281287; 1979. Stolba, A.; Baker, N.; Wood-Gush, D. G. M. The characteristics of stereotyped behaviour in stalled sows by informational redundancy. Behaviour 87:157-182; 1983. Walsh, R. N.; Cummins, R. A. Mechanisms mediating the production of environmentally induced brain changes. Psychol. Bull. 82: 986-1000; 1975.