- Email: [email protected]
Decrease of response to social separation in preparturient ewes
Behavioural Processes 40 (1997) 45 – 51
Decrease of response to social separation in preparturient ewes P. Poindron a,b,c,*, R. Soto b, A. Romeyer a b
a CIRA, CINVESTAV-UAT, AP 62, Tlaxcala, 90 000 TLAX., Me´xico FES Cuautitlan, UNAM, Carretera Me´xico-Teoloyucan, Km. 2.5, Cuautitlan-Izcalli, CP 54 700 Edo de Me´xico, Me´xico c URA CNRS 1291, INRA, 37 380 NOUZILLY, France
Received 15 May 1996; received in revised form 14 November 1996; accepted 15 November 1996
Abstract The response to social isolation was compared in multiparous non-pregnant (controls) and periparturient (experimental) Rambouillet ewes over three tests performed at times corresponding to 1119 31 h prepartum (148 days of pregnancy), 3.1 9 1.2 h prepartum (after onset of labour) and 1.3 9 0.2 h postpartum for the experimental group. Twenty-four and 15 ewes (test 1), 16 and 12 (test 2) and 7 and 10 (test 3) were tested in control and experimental groups respectively. Each test lasted 10 min (5 min in presence of conspecifics, 5 min in their absence). High-pitched bleats, locomotor activity, attempts at jumping out of the testing pen, eliminations and latency out of the pen were recorded. In both groups and in all tests there was a significant increase of agitation following removal of the conspecifics. Few differences were encountered between groups in the first test. On the other hand, in the second and third tests the agitation induced by social separation was significantly higher in dry ewes for the great majority of behaviours. We conclude that the reduction of response to social separation previously reported in postparturient ewes is not totally due to the maternal bonding to the neonate, but that before parturition, some physiological factors internal to the mother already reduce gregariousness. © 1997 Elsevier Science B.V. Keywords: Bonding; Maternal behaviour; Parturition; Sheep; Social behaviour
1. Introduction Sheep are very gregarious animals and the isolation of an individual from its conspecifics leads to the manifestation of well-defined distress be* Corresponding author. Centro de Neurobiologia en Quere´taro, UNAM, AP 70 228 Ciudad Universitaria, 04510 Me´xico DF., Me´xico.
haviours. Animals show an increase in locomotion, emit high pitched bleats and increase eliminative behaviours (Price and Thos, 1980; Romeyer and Bouissou, 1992; Le Neindre et al., 1993). On the other hand, there appears to be a drastic change in the social tendency of ewes as they give birth and bond to their lamb. For example, mother ewes separated from conspecifics
0376-6357/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. PII S 0 3 7 6 - 6 3 5 7 ( 9 6 ) 0 0 7 6 7 - X
46
P. Poindron et al. / Beha6ioural Processes 40 (1997) 45–51
in the first day following lambing do not respond to social isolation with the usual behavioural signs shown by non-pregnant non lactating ewes, as long as they are kept with their lamb (Poindron et al., 1994). However, it is not totally clear whether this reduction in gregariousness emerges before parturition or whether it depends on the presence of maternal behaviour and the bonding to the neonate. It has been reported that, in the field, sheep tend to isolate themselves before lambing which could suggest a prepartum reduction of gregariousness. Indeed, work by Arnold and Morgan (1975) and by Le´crivain and Janeau (1987) indicate that active isolation can be observed in more than half of the females giving birth. By contrast, in another study Stevens et al. (1981) reported that the majority of the females studied did not actively seek isolation at lambing but rather became isolated from the flock because of their physical incapacity to keep up with them. In fact, many factors can account for these differences including breed, experience, varying topography and paddock bush cover. In addition, the absence of actively seeking isolation does not exclude some reduction in social interest for adult conspecifics which would allow a better adaptation to prepartum isolation. Therefore, in the present experiment we investigated the possibility of a reduction of social tendency in pregnant ewes close to parturition by comparing their response to social isolation with that of non-pregnant, nonlactating ewes in standardized conditions, using the methods developed in a previous study (Poindron et al., 1994).
2. Animals, materials and methods
2.1. Animals and place of study The study was carried out in the experimental installations of the Veterinary School of the Facultad de Estudios Superiores de Cuautitlan (UNAM, State of Me´xico). Females were adult multiparous Rambouillet ewes, either dry (nonpregnant and non-lactating, N = 24-control group) or in their fifth month of pregnancy (N= 17-experimental group). Pregnant ewes had been
synchronized at oestrus so that the date of conception was known. Animals were kept in sheltered corrals. They were fed alfalfa hay and concentrate to cover their nutritional needs according to their physiological state (i.e. dry or pregnant) and had free access to water. Feeding occurred every day at 14:00 h in an attempt to synchronize activity and to obtain a majority of lambings during daytime hours (Gonyou and Cobb, 1986; Hudgens et al., 1986; J. Graber, personal communication).
2.2. Testing procedure The animals of the two groups were tested three times, corresponding respectively at day 148 9 1 days of pregnancy, at the first signs of lambing (repeated abdominal contractions or appearance of amniotic sac) and at 60 min postpartum for the experimental females. The exact latencies between testing and parturition in the experimental ewes were −111.39 31.2 h, − 3.19 1.4 h and + 1.390.2 h for test 1, 2 and 3 respectively. In addition, 10 additional dry ewes of the same flock were used as companions in the first part of each test. The testing procedure was derived from that of Poindron et al. (1994) and adapted to test periparturient ewes with minimal disturbances of the animals and differed slightly for the three tests. A testing pen of 4 m×4 m divided in four equal squares was erected with openwork panels in the living pen of the animals. In the first test (day 148 in pregnant ewes), all the ewes to be tested were penned behind the building in which they lived, and were introduced one at a time in the testing pen. During the first five minutes the behaviour of the female was observed while 10 conspecifics stayed in the living pen, surrounding the testing pen. The 10 conspecifics were then moved to another corral 20 m away. The procedure of moving the conspecifics away served the purpose of not inducing irrelevent disturbance in the observed animal by human manipulation and to mimic the type of separation occurring when the preparturient ewe becomes unable to keep up with the flock. For the second and third tests, ewes were not removed from their living pen previous to the test.
P. Poindron et al. / Beha6ioural Processes 40 (1997) 45–51
When a female showed signs of impending birth, she was introduced to the testing pen and the procedure was thereafter carried out as described above, the rest of the group (minimun 10 females) serving as conspecifics in the first part of the test. For practical reasons, all of the animals could not be used in all of the tests. The number of females tested on each instance in the control and experimental group were 24 and 15 on the first test, 16 and 12 for the second test, and 7 and 10 for the third test. Seven animals from group 1 and 8 from group 2 were tested on the three instances.
2.3. Recorded beha6iours The behaviour of the animals was noted on preformated paper sheets, by two observers. A third person manipulated the animals and the conspecifics. The following behaviours were recorded: Vocalizations (number of low-pitched and high-pitched bleats) Locomotor activity (number of squares crossed) Eliminations (number of micturitions and defecations) Number of attempts at jumping out of the pen or raising front feet on the panels Latency out of the pen. At the end of the second part of the test, the gate of the testing pen and of the pens leading to the access of the conspecifics were opened and the time taken by the ewe to go out of the testing pen was measured with a stopwatch. Animals which had not gone out of the pen within 1 min were given a latency of 61 s. Agitation index. This was constituted by adding the score of the behaviours generally considered as the most illustrative of agitation: number of high bleats, number of squares crossed, number of eliminations and and number of jumping attempts and raising front feet on the panels.
2.4. Statistical analyses Since preliminary analyses indicated a non normal distribution for the majority of the variables
47
(Lilliefors tests, PB 0.05), results were analysed with non parametric statistics (Siegel and Castellan, 1988), using the statistical package SYSTAT 5.0 (SPSS, Chicago). Mann-Whitney tests were used to compare control and experimental ewes on the three tests, and Friedman and Wilcoxon tests to compare within group changes between the three tests and between the two parts of each test. The main predictions were (i) that the effects of separation would be markedly less in periparturient ewes than in dry ones, (ii) that repeated testing may lead to a reduction of the response to separation and (iii) that activities indicative of agitation would be higher in the second part of the test. However this did not apply to lowpitched bleats, which are mainly associated with the presence of maternal behaviour (tests 2 and 3 in experimental ewes). Therefore in all instances where the direction of the differences could be predicted one tail probabilities were used at the level of P= 0.05 (Siegel and Castellan, 1988).
3. Results
3.1. Response to social separation In both groups, a significant increase of the behaviours reflecting distress was observed when comparing the first part of each test (with conspecifics) with the second part (without conspecifics). Results are summarized in Table 1. A significant increase in high-pitched bleats was found in all cases except for test 2 in experimental animals. Similarly, locomotion behaviour and jumping attempts also increased, although in a less consistant manner, especially in group 2. The general distressing effect of social separation was further emphazised when comparing the agitation indices which were always significantly higher in the second part of the test (Wilcoxon test, PB .0.02 in all cases). On the other hand eliminations were rare (0.759 0.4] ELIM] 0) and their frequency did not differ between the two parts of the test in either group.
** *
**
24.8 94.0b 21.8 9 3.2b 4.9 9 1.1a 52.4 9 4.7a
1.8 9 1.4b 6.1 9 1.7a 0.1 9 0.1b 8.5 9 2.7b
13.8 9 5.8a 10.3 9 2.2a 1.6 9 0.7a 26.5 9 6.8a
**
*
* *** ** **
5.89 3.8b 10.292.0b 0.89 0.6b 17.294.9b
30.195.6a 28.29 5.4a 4.29 1.2a 62.9910.1a
−Consp
7.69 3.4b 1.79 0.7a 0b 9.39 3.6a
0.69 0.6a 3.99 1.4a 0.39 0.2a 4.99 2.0a
+Consp
Test 3
*
**
** ** * **
19.3 98.6a 2.9 9 1.8b 0b 22.29 9.7b
18.49 6.4a 22.49 6.5a 2.3 9 0.7a 43.19 12.1a
−Consp
Locomotion: number of squares crossed; Agitation: agitation index ( =bleats+locomotion+jump attempts+eliminations). Values are mean frequencies 9S.E.M./5 min. Test 1: 111 h prepartum in group 2; Test 2: 3 h prepartum; Test 3: 1 h postpartum. +Consp: test with conspecifics; −Consp: test without conspecifics. Within group comparisons: Overall comparison between the 3 tests (Friedman test, two tails): The first symbol after each behaviour refers to the test with conspecifics, and the second symbol to the test without conspecifics. ns: no significant difference between the three tests; 1: 0.05\P\0.01, 2: 0.01\P\0.001. Between the two parts of a same test, *: 0.05\P\0.01, **: 0.01\P\0.001, ***: PB0.001 (Wilcoxon test, one tail). Between group comparisons: a variable with different letters in a same column differ significantly between groups 1 and 2 (Mann-Whitney test, one tail).
7.8 93.4a 23.9 9 4.1a 0.99 0.3a 33.1 97.4a
High bleats ns,1 Locomotion 1,2 Jump attempts ns,ns Agitation 1,2
Group 2 (periparturient)
*** ***
***
34.3 9 3.6a 32.4 9 4.3a 5.3 9 1.3a 72.9 9 7.6a
8.5 91.7a 26.0 9 2.8a 1.3 90.5a 36.7 9 4.3a
+Consp
−Consp
+Consp High bleats ns,1 Locomotion,1,2 Jump attempts ns,ns Agitation2
Test 2
Test 1
Group 1 (dry)
Behaviour
Table 1 Behaviour of dry and periparturient ewes in response to social isolation from adult conspecifics
48 P. Poindron et al. / Beha6ioural Processes 40 (1997) 45–51
P. Poindron et al. / Beha6ioural Processes 40 (1997) 45–51
3.2. Comparison of the response to separation in dry and pregnant ewes In the first part of the test, a low level of activity was found in both groups and no significant differences were encountered between the two groups on the first test, which is reflected in the value of the agitation index (36.794.3 and 33.197.4 for controls and experimental animals respectively, Mann-Whitney U=149.5, P =0.38). On the other hand, during tests 2 and 3 some differences already emerged between the two groups, especially during test 2, with a difference indicating a higher level of agitation in the dry ewes (Table 1). On the third test (60 min after parturition in the experimental animals) highpitched and low-pitched bleats were more frequent in the experimental ewes (7.6 9 3.4 vs 0.69 0.6 and 6.492.4 vs 0, respectively; U = 17 and 7, P =0.02 and 0.002), while jumping attempts tended to be higher in the controls (U = 45, P=0.04). But when animals were deprived of conspecifics, the difference between the two groups in the response to social separation was even more marked. Agitation scores were significantly greater in the dry controls than in the periparturient ewes (Table 1) when examining the frequencies of high pitched bleats (tests 1 and 2), squares crossed and jumping attempts (tests 2 and 3). Conversly, latencies to leave the testing pen at the end of the test were always shorter in the controls (1.290.1 s, 1.490.2 s and 2.3 90.6 s vs 3.1 9 0.5 s, 18.99 7.5 s and 61 90 s for tests 1, 2 and 3 in control and experimental ewes respectively; Mann-Whitney, P B0.005). This was further illustrated when comparing the agitation indices which were always significantly higher in the control group than in the experimental one (except on test 1, P=0.19). This was also confirmed when comparing the magnitude of the change in the agitation index encountered for each test ((agitation index after social separation) − (agitation index before separation)). In all cases, the change between the two parts of the test was greater in the control than in the experimental ewes (36.2 9 6.8 vs 19.39 9.0, 36.4 9 8.8 vs 8.7 9 2.3 and 38.3911.0 vs 12.99 7.3), although the difference
49
did not reach significance on test 1 (Mann-Whitney, PB 0.02 for tests 2 and 3; P= 0.16 for test 1). Finally, low-pitched bleats were significantly more frequent in group 2 in test 3 (8.1 + 3.4 vs 0, U= 3.5, PB 0.001)
3.3. Changes in the response to social separation in relation to test repetition When considering the females in each group which had been tested on all instances, Friedman two-ways analyses of variance revealed differences in scores between the three tests in the two groups which indicates a general reduction of response to the experimental situation and to separation. In the controls, this concerned squares crossed and agitation indices in the two parts of the test (Friedman test, PB 0.002). This was also the case concerning high-pitched bleats in the second part of the test (Friedman test=6, P= 0.05). In the experimental ewes, the differences concerned an increase in low-pitched bleats in both parts of the tests, an increase of the latency out of the pen (Friedman test, PB0.01), as well as a decrease in the number of crossed squares and the agitation index in both parts of the tests (Friedman test, PB0.03), and a decrease in the number of highpitched bleats in the second part of the tests (Friedman= 7.31, P=0.03). Further analyses between two consecutive tests using Wilcoxon tests indicated that in the control group the agitation index decreased between each test in the first part of the tests (Wilcoxon, PB 0.02), but that the change occurred between the first test and the following ones in response to social isolation (PB0.025), while no significant difference was found between the second and third tests (z =0.255, P= 0.80). In the case of high-pitched bleats the only significant difference was between the second part of tests 1 and 3, with also a tendency between tests 2 and 3 (z = −1.52, P= 0. 07). Finally square crossing in the first part of the test decreased between each consecutive test (PB 0.03). However, when considering the situation without conspecifics, the reduction was mainly located between the first test and the following ones (PB 0.01).
50
P. Poindron et al. / Beha6ioural Processes 40 (1997) 45–51
In the experimental group, the general tendency was similar to that found in the controls, that is a major reduction of the agitation response to social separation between tests 1 and 2. In addition, a significant increase of the emission of low-pitched bleats was found between the second and third test for both parts of the test (Wilcoxon test, low bleats: PB0.01) while for high-pitched bleats a significant reduction was found in the absence of conspecifics between tests 1 and 2 (Wilcoxon test, z = −2.52, P = 0.01), followed by a rise between tests 2 and 3 (Wilcoxon test, z= 2.38, P = 0.02). Finally, the magnitude of the response to social isolation ((agitation test 2)−(agitation test 1)) did not change in either group with the repetition of the test. 4. Discussion The comparison of behavioural responses to social separation in non-pregnant, non-lactating ewes with that of ewes which are about to lamb clearly demonstrates a reduction of behavioural distress in response to social deprivation at the time of parturition. Since these differences can be found in preparturient ewes, it seems reasonable to assume that this translates as a prepartum reduction of gregariousness, which is in agreement with the tendency of preparturient ewes to isolate themselves (Arnold and Morgan, 1975; Le´crivain and Janeau, 1987; Gonyou and Stookey, 1985; Echeverri et al., 1992). It is interesting to note that these results were obtained in Merino ewes, despite the fact that previous reports suggest that this is a breed in which active prepartum isolation is not very marked (Stevens et al., 1981). Such a reduction of social tendency probably presents an important adaptative value. In field conditions, any parturient female is likely to become separated from the flock, which would result in high distress and a conflict situation for the mother in the absence of reduced gregariousness. This would in turn hamper the development of proper bonding to the lamb and therefore increase the probability of early postnatal mortality, a major cause of lamb death in field conditions, especially in litters (Stevens et al., 1982; Alexander et al., 1983; Alexander, 1988).
The fact that this reduction of social tendency was already observed before parturition indicates that it is not strictly dependent on maternal bonding to the lamb, but rather that it can be influenced by physiological factors of late pregnancy and parturition. Our results suggest also that this change of behaviour probably takes place very shortly before parturition. Several facts support such an hypothesis. First, although the agitation index of the pregnant females was slightly lower than that of the control in the first test, they did not differ significantly at this stage, which contrasts with what was found in the two following tests. A similar pattern was also found concerning the response to separation. This suggests that the reduction of agitation between the first test and the others was greater in preparturient females than in the dry ones. A second argument in favor of this interpretation comes from studying the correlation betwen the latency existing between the first test and parturition on the one hand and the agitation index on the other. The variability of pregnancy duration allows such an analysis: in average, ewes lambed 111 9 31 h after the first test, but some females were tested less than 48 h before parturition. Indeed, a significant Pearson correlation value is found between the agitation index in the second part of the test and the latency to parturition (r= 0.59, P= 0.03). Here again the results suggest that the closer to parturition, the lower the agitation will be in absence of conspecifics. On the other hand, our results do not allow a definate conclusion, due to the confounding effect of a general decrease of agitation in dry ewes with the repetition of the tests and the low number of animals that we could test on all instances. There is also another point worth mentioning that could mask such a decrease, which comes from the procedure we used to test response to social separation. In the second test, the ewe was moved to be introduced to the testing pen once labour had started. At this stage females may have already selected a lambing spot and moving them may have been a disturbing procedure thereby resulting in higher agitation scores than if they could have stayed in their selected lambing spot. Finally, the possibility that the differences found between dry and pregnant ewes
P. Poindron et al. / Beha6ioural Processes 40 (1997) 45–51
could be due to a confounding action of differing feeding demand of dry and pregnant ewes that would have influenced the behaviour of the animals during the test is very unlikely. Feeding adjustments were made so that the requirements of pregnant ewes were met at all times, and there were no signs of differences of body conditions between control and experimental animals. Also, in the case of a response due to some confounding effect of feeding motivation, the greater variation found in pregnant ewes between the three tests would be rather difficult to explain. To conclude, it can be emphazised that the reduction of social interest for adult conspecifics in post parturient ewes is not only due to the bonding process with the neonate, but already exists before lambing. Therefore this reduction in social tendency is likely to be related with the physiology of late pregnancy and parturition. Additional studies involving independent groups of animals at various times before lambing are necessary to define more precisely the dynamics of this change in gregariousness in the period immediatly prepartum, thus providing a better basis for the identification of the physiological factors involved. Finally it is interesting to note that, in sheep, the maternal affiliation to the neonate is preceded by a reduction of afiliation to adult conspecifics, therefore indicating that these two ‘social’ behaviours are not facilitated by the same physiological factors.
Acknowledgements This study was supported by CINVESTAV, the Universidad Autonoma de Tlaxcala and grant 4881-N of CONACyT. P. Poindron and A. Romeyer were supported by a Catedra Patrimonial de Excelencia para extranjeros of CONACyT. We are also grateful to Diego Rueda for the care . of the animals and to Marina Hernandez Calva for her help during the study.
51
References Alexander, G., 1988. What makes a good mother?: components and comparative aspects of maternal behaviour in ungulates. Proc. Austr. Soc. Anim. Prod., 17: 25 – 41. Alexander G., Stevens, D., Kilgour, R., de Langen, H., Mottershead, B.E. and Lynch, J.J., 1983. Separation of ewes from twin lambs: incidence in several sheep breeds. Appl. Anim. Ethol., 10: 301 – 317. Arnold, G.W. and Morgan, P.D., 1975. Behaviour of the ewe and lamb at lambing and its relationship to lamb mortality. Appl. Anim. Ethol., 2: 25 – 46. Echeverri, A.C., Gonyou, H.W. and Ghent, A.W., 1992. Preparturient behavior of confined ewes: time budgets, frequencies, spatial distribution and sequential analysis. Appl. Anim. Behav. Sci., 34: 329 – 344. Gonyou, H.W. and Cobb, A.R., 1986. The influence of time of feeding on the time of parturition in ewes. Can. J. Anim. Sci., 66: 569 – 574. Gonyou, H.W. and Stookey, J.M., 1985. Behavior of parturient ewes in group-lambing pens with and without cubicles. Appl. Anim. Behav. Sci., 14: 163 – 172. Hudgens, R.E, Albright, J.L. and Pennington, J.A., 1986. Influence of feeding time and diet on time of parturition in multiparous ewes. J. Anim. Sci., 63: 1036 – 1040. Le´crivain, E. and Janeau, G., 1987. Comportement d’isolement et de recherche d’abri de brebis agnelant en plein air dans un syste`me d’e´levage a` caracte`re extensif. Biol. Behav., 12: 127 – 148. Le Neindre, P., Poindron, P., Trillat, G. and Orgeur, P., 1993. Influence of breed on reactivity of sheep to humans. Genet. Sel. Evol., 25: 447 – 458. Poindron, P., Caba, M., Gomora, Arrati P., Krehbiel, D. and Beyer, C., 1994. Responses of maternal and non maternal ewes to social and mother-young separation. Behav. Proc., 31: 97 – 110. Price, E.G. and Thos, J., 1980. Behavioral responses to shortterm isolation in sheep and goat. Appl. Anim. Ethol., 6: 331 – 339. Romeyer, A. and Bouissou, M.F., 1992. Assesment of fear reactions in domestic sheep, and influence of breed and rearing conditions. Appl. Anim. Behav. Sci., 34: 93 – 119. Siegel, S. and Castellan, N.J., 1988. Non parametric statistics for the behavioral sciences, second edition. McGraw-Hill, Mexico. Stevens, D., Alexander, G. and Lynch, J.J., 1981. Do Merino ewes seek isolation or shelter at lambing? Appl. Anim. Ethol., 7: 149 – 155. Stevens, D., Alexander, G. and Lynch, J.J., 1982. Lamb mortality due to inadequate care of twins by Merino ewes. Appl. Anim. Ethol., 8: 243 – 252.
Decrease of response to social separation in preparturient ewes P. Poindron a,b,c,*, R. Soto b, A. Romeyer a b
a CIRA, CINVESTAV-UAT, AP 62, Tlaxcala, 90 000 TLAX., Me´xico FES Cuautitlan, UNAM, Carretera Me´xico-Teoloyucan, Km. 2.5, Cuautitlan-Izcalli, CP 54 700 Edo de Me´xico, Me´xico c URA CNRS 1291, INRA, 37 380 NOUZILLY, France
Received 15 May 1996; received in revised form 14 November 1996; accepted 15 November 1996
Abstract The response to social isolation was compared in multiparous non-pregnant (controls) and periparturient (experimental) Rambouillet ewes over three tests performed at times corresponding to 1119 31 h prepartum (148 days of pregnancy), 3.1 9 1.2 h prepartum (after onset of labour) and 1.3 9 0.2 h postpartum for the experimental group. Twenty-four and 15 ewes (test 1), 16 and 12 (test 2) and 7 and 10 (test 3) were tested in control and experimental groups respectively. Each test lasted 10 min (5 min in presence of conspecifics, 5 min in their absence). High-pitched bleats, locomotor activity, attempts at jumping out of the testing pen, eliminations and latency out of the pen were recorded. In both groups and in all tests there was a significant increase of agitation following removal of the conspecifics. Few differences were encountered between groups in the first test. On the other hand, in the second and third tests the agitation induced by social separation was significantly higher in dry ewes for the great majority of behaviours. We conclude that the reduction of response to social separation previously reported in postparturient ewes is not totally due to the maternal bonding to the neonate, but that before parturition, some physiological factors internal to the mother already reduce gregariousness. © 1997 Elsevier Science B.V. Keywords: Bonding; Maternal behaviour; Parturition; Sheep; Social behaviour
1. Introduction Sheep are very gregarious animals and the isolation of an individual from its conspecifics leads to the manifestation of well-defined distress be* Corresponding author. Centro de Neurobiologia en Quere´taro, UNAM, AP 70 228 Ciudad Universitaria, 04510 Me´xico DF., Me´xico.
haviours. Animals show an increase in locomotion, emit high pitched bleats and increase eliminative behaviours (Price and Thos, 1980; Romeyer and Bouissou, 1992; Le Neindre et al., 1993). On the other hand, there appears to be a drastic change in the social tendency of ewes as they give birth and bond to their lamb. For example, mother ewes separated from conspecifics
0376-6357/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. PII S 0 3 7 6 - 6 3 5 7 ( 9 6 ) 0 0 7 6 7 - X
46
P. Poindron et al. / Beha6ioural Processes 40 (1997) 45–51
in the first day following lambing do not respond to social isolation with the usual behavioural signs shown by non-pregnant non lactating ewes, as long as they are kept with their lamb (Poindron et al., 1994). However, it is not totally clear whether this reduction in gregariousness emerges before parturition or whether it depends on the presence of maternal behaviour and the bonding to the neonate. It has been reported that, in the field, sheep tend to isolate themselves before lambing which could suggest a prepartum reduction of gregariousness. Indeed, work by Arnold and Morgan (1975) and by Le´crivain and Janeau (1987) indicate that active isolation can be observed in more than half of the females giving birth. By contrast, in another study Stevens et al. (1981) reported that the majority of the females studied did not actively seek isolation at lambing but rather became isolated from the flock because of their physical incapacity to keep up with them. In fact, many factors can account for these differences including breed, experience, varying topography and paddock bush cover. In addition, the absence of actively seeking isolation does not exclude some reduction in social interest for adult conspecifics which would allow a better adaptation to prepartum isolation. Therefore, in the present experiment we investigated the possibility of a reduction of social tendency in pregnant ewes close to parturition by comparing their response to social isolation with that of non-pregnant, nonlactating ewes in standardized conditions, using the methods developed in a previous study (Poindron et al., 1994).
2. Animals, materials and methods
2.1. Animals and place of study The study was carried out in the experimental installations of the Veterinary School of the Facultad de Estudios Superiores de Cuautitlan (UNAM, State of Me´xico). Females were adult multiparous Rambouillet ewes, either dry (nonpregnant and non-lactating, N = 24-control group) or in their fifth month of pregnancy (N= 17-experimental group). Pregnant ewes had been
synchronized at oestrus so that the date of conception was known. Animals were kept in sheltered corrals. They were fed alfalfa hay and concentrate to cover their nutritional needs according to their physiological state (i.e. dry or pregnant) and had free access to water. Feeding occurred every day at 14:00 h in an attempt to synchronize activity and to obtain a majority of lambings during daytime hours (Gonyou and Cobb, 1986; Hudgens et al., 1986; J. Graber, personal communication).
2.2. Testing procedure The animals of the two groups were tested three times, corresponding respectively at day 148 9 1 days of pregnancy, at the first signs of lambing (repeated abdominal contractions or appearance of amniotic sac) and at 60 min postpartum for the experimental females. The exact latencies between testing and parturition in the experimental ewes were −111.39 31.2 h, − 3.19 1.4 h and + 1.390.2 h for test 1, 2 and 3 respectively. In addition, 10 additional dry ewes of the same flock were used as companions in the first part of each test. The testing procedure was derived from that of Poindron et al. (1994) and adapted to test periparturient ewes with minimal disturbances of the animals and differed slightly for the three tests. A testing pen of 4 m×4 m divided in four equal squares was erected with openwork panels in the living pen of the animals. In the first test (day 148 in pregnant ewes), all the ewes to be tested were penned behind the building in which they lived, and were introduced one at a time in the testing pen. During the first five minutes the behaviour of the female was observed while 10 conspecifics stayed in the living pen, surrounding the testing pen. The 10 conspecifics were then moved to another corral 20 m away. The procedure of moving the conspecifics away served the purpose of not inducing irrelevent disturbance in the observed animal by human manipulation and to mimic the type of separation occurring when the preparturient ewe becomes unable to keep up with the flock. For the second and third tests, ewes were not removed from their living pen previous to the test.
P. Poindron et al. / Beha6ioural Processes 40 (1997) 45–51
When a female showed signs of impending birth, she was introduced to the testing pen and the procedure was thereafter carried out as described above, the rest of the group (minimun 10 females) serving as conspecifics in the first part of the test. For practical reasons, all of the animals could not be used in all of the tests. The number of females tested on each instance in the control and experimental group were 24 and 15 on the first test, 16 and 12 for the second test, and 7 and 10 for the third test. Seven animals from group 1 and 8 from group 2 were tested on the three instances.
2.3. Recorded beha6iours The behaviour of the animals was noted on preformated paper sheets, by two observers. A third person manipulated the animals and the conspecifics. The following behaviours were recorded: Vocalizations (number of low-pitched and high-pitched bleats) Locomotor activity (number of squares crossed) Eliminations (number of micturitions and defecations) Number of attempts at jumping out of the pen or raising front feet on the panels Latency out of the pen. At the end of the second part of the test, the gate of the testing pen and of the pens leading to the access of the conspecifics were opened and the time taken by the ewe to go out of the testing pen was measured with a stopwatch. Animals which had not gone out of the pen within 1 min were given a latency of 61 s. Agitation index. This was constituted by adding the score of the behaviours generally considered as the most illustrative of agitation: number of high bleats, number of squares crossed, number of eliminations and and number of jumping attempts and raising front feet on the panels.
2.4. Statistical analyses Since preliminary analyses indicated a non normal distribution for the majority of the variables
47
(Lilliefors tests, PB 0.05), results were analysed with non parametric statistics (Siegel and Castellan, 1988), using the statistical package SYSTAT 5.0 (SPSS, Chicago). Mann-Whitney tests were used to compare control and experimental ewes on the three tests, and Friedman and Wilcoxon tests to compare within group changes between the three tests and between the two parts of each test. The main predictions were (i) that the effects of separation would be markedly less in periparturient ewes than in dry ones, (ii) that repeated testing may lead to a reduction of the response to separation and (iii) that activities indicative of agitation would be higher in the second part of the test. However this did not apply to lowpitched bleats, which are mainly associated with the presence of maternal behaviour (tests 2 and 3 in experimental ewes). Therefore in all instances where the direction of the differences could be predicted one tail probabilities were used at the level of P= 0.05 (Siegel and Castellan, 1988).
3. Results
3.1. Response to social separation In both groups, a significant increase of the behaviours reflecting distress was observed when comparing the first part of each test (with conspecifics) with the second part (without conspecifics). Results are summarized in Table 1. A significant increase in high-pitched bleats was found in all cases except for test 2 in experimental animals. Similarly, locomotion behaviour and jumping attempts also increased, although in a less consistant manner, especially in group 2. The general distressing effect of social separation was further emphazised when comparing the agitation indices which were always significantly higher in the second part of the test (Wilcoxon test, PB .0.02 in all cases). On the other hand eliminations were rare (0.759 0.4] ELIM] 0) and their frequency did not differ between the two parts of the test in either group.
** *
**
24.8 94.0b 21.8 9 3.2b 4.9 9 1.1a 52.4 9 4.7a
1.8 9 1.4b 6.1 9 1.7a 0.1 9 0.1b 8.5 9 2.7b
13.8 9 5.8a 10.3 9 2.2a 1.6 9 0.7a 26.5 9 6.8a
**
*
* *** ** **
5.89 3.8b 10.292.0b 0.89 0.6b 17.294.9b
30.195.6a 28.29 5.4a 4.29 1.2a 62.9910.1a
−Consp
7.69 3.4b 1.79 0.7a 0b 9.39 3.6a
0.69 0.6a 3.99 1.4a 0.39 0.2a 4.99 2.0a
+Consp
Test 3
*
**
** ** * **
19.3 98.6a 2.9 9 1.8b 0b 22.29 9.7b
18.49 6.4a 22.49 6.5a 2.3 9 0.7a 43.19 12.1a
−Consp
Locomotion: number of squares crossed; Agitation: agitation index ( =bleats+locomotion+jump attempts+eliminations). Values are mean frequencies 9S.E.M./5 min. Test 1: 111 h prepartum in group 2; Test 2: 3 h prepartum; Test 3: 1 h postpartum. +Consp: test with conspecifics; −Consp: test without conspecifics. Within group comparisons: Overall comparison between the 3 tests (Friedman test, two tails): The first symbol after each behaviour refers to the test with conspecifics, and the second symbol to the test without conspecifics. ns: no significant difference between the three tests; 1: 0.05\P\0.01, 2: 0.01\P\0.001. Between the two parts of a same test, *: 0.05\P\0.01, **: 0.01\P\0.001, ***: PB0.001 (Wilcoxon test, one tail). Between group comparisons: a variable with different letters in a same column differ significantly between groups 1 and 2 (Mann-Whitney test, one tail).
7.8 93.4a 23.9 9 4.1a 0.99 0.3a 33.1 97.4a
High bleats ns,1 Locomotion 1,2 Jump attempts ns,ns Agitation 1,2
Group 2 (periparturient)
*** ***
***
34.3 9 3.6a 32.4 9 4.3a 5.3 9 1.3a 72.9 9 7.6a
8.5 91.7a 26.0 9 2.8a 1.3 90.5a 36.7 9 4.3a
+Consp
−Consp
+Consp High bleats ns,1 Locomotion,1,2 Jump attempts ns,ns Agitation2
Test 2
Test 1
Group 1 (dry)
Behaviour
Table 1 Behaviour of dry and periparturient ewes in response to social isolation from adult conspecifics
48 P. Poindron et al. / Beha6ioural Processes 40 (1997) 45–51
P. Poindron et al. / Beha6ioural Processes 40 (1997) 45–51
3.2. Comparison of the response to separation in dry and pregnant ewes In the first part of the test, a low level of activity was found in both groups and no significant differences were encountered between the two groups on the first test, which is reflected in the value of the agitation index (36.794.3 and 33.197.4 for controls and experimental animals respectively, Mann-Whitney U=149.5, P =0.38). On the other hand, during tests 2 and 3 some differences already emerged between the two groups, especially during test 2, with a difference indicating a higher level of agitation in the dry ewes (Table 1). On the third test (60 min after parturition in the experimental animals) highpitched and low-pitched bleats were more frequent in the experimental ewes (7.6 9 3.4 vs 0.69 0.6 and 6.492.4 vs 0, respectively; U = 17 and 7, P =0.02 and 0.002), while jumping attempts tended to be higher in the controls (U = 45, P=0.04). But when animals were deprived of conspecifics, the difference between the two groups in the response to social separation was even more marked. Agitation scores were significantly greater in the dry controls than in the periparturient ewes (Table 1) when examining the frequencies of high pitched bleats (tests 1 and 2), squares crossed and jumping attempts (tests 2 and 3). Conversly, latencies to leave the testing pen at the end of the test were always shorter in the controls (1.290.1 s, 1.490.2 s and 2.3 90.6 s vs 3.1 9 0.5 s, 18.99 7.5 s and 61 90 s for tests 1, 2 and 3 in control and experimental ewes respectively; Mann-Whitney, P B0.005). This was further illustrated when comparing the agitation indices which were always significantly higher in the control group than in the experimental one (except on test 1, P=0.19). This was also confirmed when comparing the magnitude of the change in the agitation index encountered for each test ((agitation index after social separation) − (agitation index before separation)). In all cases, the change between the two parts of the test was greater in the control than in the experimental ewes (36.2 9 6.8 vs 19.39 9.0, 36.4 9 8.8 vs 8.7 9 2.3 and 38.3911.0 vs 12.99 7.3), although the difference
49
did not reach significance on test 1 (Mann-Whitney, PB 0.02 for tests 2 and 3; P= 0.16 for test 1). Finally, low-pitched bleats were significantly more frequent in group 2 in test 3 (8.1 + 3.4 vs 0, U= 3.5, PB 0.001)
3.3. Changes in the response to social separation in relation to test repetition When considering the females in each group which had been tested on all instances, Friedman two-ways analyses of variance revealed differences in scores between the three tests in the two groups which indicates a general reduction of response to the experimental situation and to separation. In the controls, this concerned squares crossed and agitation indices in the two parts of the test (Friedman test, PB 0.002). This was also the case concerning high-pitched bleats in the second part of the test (Friedman test=6, P= 0.05). In the experimental ewes, the differences concerned an increase in low-pitched bleats in both parts of the tests, an increase of the latency out of the pen (Friedman test, PB0.01), as well as a decrease in the number of crossed squares and the agitation index in both parts of the tests (Friedman test, PB0.03), and a decrease in the number of highpitched bleats in the second part of the tests (Friedman= 7.31, P=0.03). Further analyses between two consecutive tests using Wilcoxon tests indicated that in the control group the agitation index decreased between each test in the first part of the tests (Wilcoxon, PB 0.02), but that the change occurred between the first test and the following ones in response to social isolation (PB0.025), while no significant difference was found between the second and third tests (z =0.255, P= 0.80). In the case of high-pitched bleats the only significant difference was between the second part of tests 1 and 3, with also a tendency between tests 2 and 3 (z = −1.52, P= 0. 07). Finally square crossing in the first part of the test decreased between each consecutive test (PB 0.03). However, when considering the situation without conspecifics, the reduction was mainly located between the first test and the following ones (PB 0.01).
50
P. Poindron et al. / Beha6ioural Processes 40 (1997) 45–51
In the experimental group, the general tendency was similar to that found in the controls, that is a major reduction of the agitation response to social separation between tests 1 and 2. In addition, a significant increase of the emission of low-pitched bleats was found between the second and third test for both parts of the test (Wilcoxon test, low bleats: PB0.01) while for high-pitched bleats a significant reduction was found in the absence of conspecifics between tests 1 and 2 (Wilcoxon test, z = −2.52, P = 0.01), followed by a rise between tests 2 and 3 (Wilcoxon test, z= 2.38, P = 0.02). Finally, the magnitude of the response to social isolation ((agitation test 2)−(agitation test 1)) did not change in either group with the repetition of the test. 4. Discussion The comparison of behavioural responses to social separation in non-pregnant, non-lactating ewes with that of ewes which are about to lamb clearly demonstrates a reduction of behavioural distress in response to social deprivation at the time of parturition. Since these differences can be found in preparturient ewes, it seems reasonable to assume that this translates as a prepartum reduction of gregariousness, which is in agreement with the tendency of preparturient ewes to isolate themselves (Arnold and Morgan, 1975; Le´crivain and Janeau, 1987; Gonyou and Stookey, 1985; Echeverri et al., 1992). It is interesting to note that these results were obtained in Merino ewes, despite the fact that previous reports suggest that this is a breed in which active prepartum isolation is not very marked (Stevens et al., 1981). Such a reduction of social tendency probably presents an important adaptative value. In field conditions, any parturient female is likely to become separated from the flock, which would result in high distress and a conflict situation for the mother in the absence of reduced gregariousness. This would in turn hamper the development of proper bonding to the lamb and therefore increase the probability of early postnatal mortality, a major cause of lamb death in field conditions, especially in litters (Stevens et al., 1982; Alexander et al., 1983; Alexander, 1988).
The fact that this reduction of social tendency was already observed before parturition indicates that it is not strictly dependent on maternal bonding to the lamb, but rather that it can be influenced by physiological factors of late pregnancy and parturition. Our results suggest also that this change of behaviour probably takes place very shortly before parturition. Several facts support such an hypothesis. First, although the agitation index of the pregnant females was slightly lower than that of the control in the first test, they did not differ significantly at this stage, which contrasts with what was found in the two following tests. A similar pattern was also found concerning the response to separation. This suggests that the reduction of agitation between the first test and the others was greater in preparturient females than in the dry ones. A second argument in favor of this interpretation comes from studying the correlation betwen the latency existing between the first test and parturition on the one hand and the agitation index on the other. The variability of pregnancy duration allows such an analysis: in average, ewes lambed 111 9 31 h after the first test, but some females were tested less than 48 h before parturition. Indeed, a significant Pearson correlation value is found between the agitation index in the second part of the test and the latency to parturition (r= 0.59, P= 0.03). Here again the results suggest that the closer to parturition, the lower the agitation will be in absence of conspecifics. On the other hand, our results do not allow a definate conclusion, due to the confounding effect of a general decrease of agitation in dry ewes with the repetition of the tests and the low number of animals that we could test on all instances. There is also another point worth mentioning that could mask such a decrease, which comes from the procedure we used to test response to social separation. In the second test, the ewe was moved to be introduced to the testing pen once labour had started. At this stage females may have already selected a lambing spot and moving them may have been a disturbing procedure thereby resulting in higher agitation scores than if they could have stayed in their selected lambing spot. Finally, the possibility that the differences found between dry and pregnant ewes
P. Poindron et al. / Beha6ioural Processes 40 (1997) 45–51
could be due to a confounding action of differing feeding demand of dry and pregnant ewes that would have influenced the behaviour of the animals during the test is very unlikely. Feeding adjustments were made so that the requirements of pregnant ewes were met at all times, and there were no signs of differences of body conditions between control and experimental animals. Also, in the case of a response due to some confounding effect of feeding motivation, the greater variation found in pregnant ewes between the three tests would be rather difficult to explain. To conclude, it can be emphazised that the reduction of social interest for adult conspecifics in post parturient ewes is not only due to the bonding process with the neonate, but already exists before lambing. Therefore this reduction in social tendency is likely to be related with the physiology of late pregnancy and parturition. Additional studies involving independent groups of animals at various times before lambing are necessary to define more precisely the dynamics of this change in gregariousness in the period immediatly prepartum, thus providing a better basis for the identification of the physiological factors involved. Finally it is interesting to note that, in sheep, the maternal affiliation to the neonate is preceded by a reduction of afiliation to adult conspecifics, therefore indicating that these two ‘social’ behaviours are not facilitated by the same physiological factors.
Acknowledgements This study was supported by CINVESTAV, the Universidad Autonoma de Tlaxcala and grant 4881-N of CONACyT. P. Poindron and A. Romeyer were supported by a Catedra Patrimonial de Excelencia para extranjeros of CONACyT. We are also grateful to Diego Rueda for the care . of the animals and to Marina Hernandez Calva for her help during the study.
51
References Alexander, G., 1988. What makes a good mother?: components and comparative aspects of maternal behaviour in ungulates. Proc. Austr. Soc. Anim. Prod., 17: 25 – 41. Alexander G., Stevens, D., Kilgour, R., de Langen, H., Mottershead, B.E. and Lynch, J.J., 1983. Separation of ewes from twin lambs: incidence in several sheep breeds. Appl. Anim. Ethol., 10: 301 – 317. Arnold, G.W. and Morgan, P.D., 1975. Behaviour of the ewe and lamb at lambing and its relationship to lamb mortality. Appl. Anim. Ethol., 2: 25 – 46. Echeverri, A.C., Gonyou, H.W. and Ghent, A.W., 1992. Preparturient behavior of confined ewes: time budgets, frequencies, spatial distribution and sequential analysis. Appl. Anim. Behav. Sci., 34: 329 – 344. Gonyou, H.W. and Cobb, A.R., 1986. The influence of time of feeding on the time of parturition in ewes. Can. J. Anim. Sci., 66: 569 – 574. Gonyou, H.W. and Stookey, J.M., 1985. Behavior of parturient ewes in group-lambing pens with and without cubicles. Appl. Anim. Behav. Sci., 14: 163 – 172. Hudgens, R.E, Albright, J.L. and Pennington, J.A., 1986. Influence of feeding time and diet on time of parturition in multiparous ewes. J. Anim. Sci., 63: 1036 – 1040. Le´crivain, E. and Janeau, G., 1987. Comportement d’isolement et de recherche d’abri de brebis agnelant en plein air dans un syste`me d’e´levage a` caracte`re extensif. Biol. Behav., 12: 127 – 148. Le Neindre, P., Poindron, P., Trillat, G. and Orgeur, P., 1993. Influence of breed on reactivity of sheep to humans. Genet. Sel. Evol., 25: 447 – 458. Poindron, P., Caba, M., Gomora, Arrati P., Krehbiel, D. and Beyer, C., 1994. Responses of maternal and non maternal ewes to social and mother-young separation. Behav. Proc., 31: 97 – 110. Price, E.G. and Thos, J., 1980. Behavioral responses to shortterm isolation in sheep and goat. Appl. Anim. Ethol., 6: 331 – 339. Romeyer, A. and Bouissou, M.F., 1992. Assesment of fear reactions in domestic sheep, and influence of breed and rearing conditions. Appl. Anim. Behav. Sci., 34: 93 – 119. Siegel, S. and Castellan, N.J., 1988. Non parametric statistics for the behavioral sciences, second edition. McGraw-Hill, Mexico. Stevens, D., Alexander, G. and Lynch, J.J., 1981. Do Merino ewes seek isolation or shelter at lambing? Appl. Anim. Ethol., 7: 149 – 155. Stevens, D., Alexander, G. and Lynch, J.J., 1982. Lamb mortality due to inadequate care of twins by Merino ewes. Appl. Anim. Ethol., 8: 243 – 252.