Influence of Social Status on Ovarian Function in Farmed Red Deer (Cervus elaphus)

Physiology & Behavior, Vol. 65, Nos. 4/5, pp. 691–696, 1999 © 1999 Elsevier Science Inc. Printed in the USA. All rights reserved 0031-9384/99 $–see front matter

PII S0031-9384(98)00229-7

Influence of Social Status on Ovarian Function in Farmed Red Deer (Cervus elaphus) N. GOODWIN,*1 V. HAYSSEN,* D. W. DEAKIN† AND A. P. F. FLINT* *School of Biology, University of Nottingham, Sutton Bonington Campus, Loughborough, Leics. LE12 5RD, UK †ADAS Rosemaund, Preston, Wynne, Hereford HR1 3PG, UK Received 19 September 1997; Accepted 23 July 1998 GOODWIN, N., V. HAYSSEN, D. W. DEAKIN AND A. P. F. FLINT. Influence of social status on ovarian function in farmed red deer (Cervus elaphus). PHYSIOL BEHAV 65(4/5) 691–696, 1999—This study examines the effects of behavioural and environmental factors on ovarian function in red deer hinds. Patterns of postovulatory luteal progesterone secretion were investigated in groups of farmed red deer hinds following oestrus induced by progesterone administration and withdrawal. Hinds were held under conditions leading to low (Experiment 1, held in paddocks: 0.29 per animal/h) or high (Experiment 2, held in pens: 1.11 per animal/h) interaction rates, and progesterone was measured in jugular venous blood samples obtained daily for 14 days after ovulation. Plasma progesterone levels tended to differ with respect to dominance status in both experiments (p , 0.1). Progesterone levels were substantially lower following prolonged hot dry weather (mean 0.67 6 SEM 0.03 ng/mL) than in a year of relatively typical climatic conditions (mean 1.63 6 0.04 ng/mL; p , 0.0005). Progesterone levels were not related to the body weight of the hinds, and were not affected by housing conditions leading to different interaction rates. We conclude that although there is a tendency for dominance status to affect ovarian function, as observed before, this relationship is obscured in farmed red deer where all animals in the herd enjoy a higher plane of nutrition and movements of individuals between groups confuse dominance relationships. © 1999 Elsevier Science Inc. Dominance

Progesterone

Red deer

However, these problems are eliminated in a captive population, where animals may be approached and handled with minimal disturbance to their established behaviour patterns. Preliminary data obtained in experiments at Whipsnade Wild Animal Park (Bedfordshire, UK) suggest that postovulatory progesterone profiles of red deer hinds differ according to dominance status (8). In this population, progesterone levels rose more rapidly after ovulation in dominant than in subordinate hinds: by 11 days, levels were 2.5-fold higher in dominants. The importance of progesterone in establishment and maintenance of pregnancy means that this difference could contribute to the reproductive trends described above. [In cattle, progesterone treatment immediately after ovulation increases conception rate (21).] However, these earlier studies used hinds kept at high densities in yards; this form of housing was relatively unnatural, and resulted in high rates of agonistic interaction. Dominance studies carried out in a captive red deer population at the present study site (N. Goodwin, unpublished data) show that agonistic interactions between ani-

WILD red deer (Cervus elaphus) hinds in social groups form well-established dominance hierarchies (4) based on dyadic agonistic interactions (23). Dominance status affects reproduction in wild red deer hinds in several ways. Dominant hinds calve more frequently than subordinates; they conceive and calve earlier in the respective seasons; and their offspring are heavier than those of subordinate hinds, with a higher probability of surviving to maturity (6). Furthermore, offspring sex ratios are skewed in favour of dominant hinds producing a greater proportion of male calves (5), fulfilling predictions arising from differential parental investment in male and female offspring in a polygynous species (24). Despite the accumulation of data illustrating these trends, there have been few investigations of the physiological mechanisms underlying them. Such studies are often impractical to carry out with wild populations due to difficulty of access and handling of animals. Physiological data on wild red deer, for instance, may be limited to what can be obtained from animals culled as part of a management procedure [e.g., (9)].

1To

whom requests for reprints should be addressed. E-mail: [email protected]

691

692

GOODWIN ET AL.

mals occur at substantially different rates in different husbandry conditions. Specifically, hinds kept in indoor confinement exhibit interaction rates 3.8-fold higher than hinds kept outdoors in large paddocks (see later). The stress associated with frequent agonistic interactions may affect circulating progesterone levels. We, therefore, wished to determine whether the relationship between ovarian function and social status was independent of holding density, and carried out a behavioural and endocrinological study of hinds subjected to both possible sets of husbandry procedures during the rutting season. As status is related to body condition in wild red deer (6,23), and ovarian function is affected by food intake (18), the relationship between body weight and progesterone level was also examined. In addition, climatic differences in the two experimental years allowed us to investigate the impact of environmental conditions on progesterone secretion. METHODS

Study Animals and Management The hinds were part of the captive-born breeding herd at ADAS Rosemaund experimental farm, Preston Wynne, Herefordshire, UK. Unless otherwise noted, hinds were kept in paddocks approximately 100 3 300 m, sown with a mixture of rye grass and white clover, and providing ad lib water. Hind ages ranged from 3 to 7 years. Most hinds had produced a calf in the summer, and calves were weaned in mid-September. Rutting behaviour of hinds and stags typically started around the last week of September, consistent with wild British populations (4). In the UK, there was an exceptionally long hot summer in 1995, preceding Experiment 2. The supply of drinking water in the paddocks was unaffected, but mud wallows in which deer often bathe to cool themselves dried up, and the grass in the paddocks was scorched and generally shorter than in other summers. Possible implications of this are discussed later. Low Interaction Experiment 1 The outdoor experiment, providing low levels of agonistic interaction, was carried out in the Autumn of 1994. For management reasons hinds (n 5 66, Table 1) were divided into four groups: two each of 15 (groups 1 and 2) and two each of 18 (groups 3 and 4), and work began on these groups at different times. For groups 1 and 2, work commenced on 4 October; for group 3 on 4 November; and for group 4 on 14 November. To synchronize ovulation, hinds were administered progesterone by intravaginal CIDR (controlled internal drug release) devices (type G; Agricultural Division, CHH Plastic Products Group Ltd., Hamilton, New Zealand). Colour-coded plastic collars (modified cow collars; Dalton Suppliers Ltd., Nettle Bed, Oxfordshire, UK) were fitted to the hinds to aid identification in the field. CIDRs remained in place for 12 days. During this time, dyadic dominance interactions (23) were observed at various times between 0830 and 1730 h, while the deer were in the paddocks. Observations were continued until enough data had been gathered to allow accurate construction of dominance hierarchies. The four groups were observed for 19 h 20 min, 17 h 20 min, 29 h 10 min, and 28 h, respectively. CIDRs were removed at 0800–0900 h on the 12th day, and hinds were weighed to within 0.5 kg using a weigh crate. Two days after CIDR removal, when hinds were expected to come into heat, a mature stag was placed in the paddock with each group of hinds at 0800 h. Effectiveness of progesterone synchronization was confirmed by observing mating be-

tween 48 and 57 h after CIDR withdrawal; approximately 50% of hinds mated during this period. Beginning the following day, daily blood samples of 5–10 mL were taken from the hinds at times ranging from 0815 h to 1500 h for 14 consecutive days. Hinds were restrained in a drop-floor crush, and blood was collected by jugular venepuncture, either by heparinized vacutainer, or by syringe with immediate transfer to heparinized tubes. [Collection on ice was not necessary as progesterone degradation in heparinized blood at room temperature is negligible in the first 2 h (20)]. Within 2 h of collection, blood samples were centrifuged at 3000 r.p.m. for 10 min; plasma fractions were retained and stored at 2208C until assay. Individual hinds were sampled in random order with respect to dominance status. High Interaction Experiment 2 The indoor experiment, providing higher levels of agonistic interaction, was carried out in the Autumn of 1995. The above protocol was modified for this trial. Hinds (n 5 24, 15 of which had been used in the 1994 experiment) were divided into three groups of eight. After the experiment started, groups 5 and 6 were kept in straw-bedded pens 5 3 8 m in a shed normally used for housing the deer during winter. These hinds were fed silage ad lib. and given water ad lib. Group 7 hinds (outdoor controls) were kept in a paddock; for management reasons these animals were held along with 16 other hinds. The control hinds were distributed randomly throughout the hierarchy of this group of 24. The dimensions of this paddock shrank to approximately 100 3 150 m on 9 October, when a dividing fence was built across the middle, and the hinds were confined on one side of it. CIDRs were inserted into all hinds on 25 September. Identification collars were fitted to the control group and to the other hinds in the same paddock. Ear tags alone were sufficient for identifying hinds indoors; there was no apparent difference in the agonistic (or any other) behaviour of collared and uncollared hinds. CIDRs were in place for 14 days (replaced with fresh ones on day 7), after which time they were removed and hinds were weighed. While CIDRs were in place, dominance interactions were observed among the indoor hinds and among all hinds in the paddock, until enough data had been gathered to allow hierarchy calculation. The two indoor groups were observed simultaneously for a total of 12 h, 23 min; outdoor hinds were observed for a total of 6 h. Stags were not placed with the hinds, due to the impracticality of keeping a rutting stag in a confined space. Three days after the second CIDR removal, daily blood sampling (6 days per week) was commenced as in the first season. Samples were not taken on 23 October for all except four control hinds, and not for those four controls on 24 October due to constraints of other experimental work. On 2 days when insufficient staff were available to bring hinds in from the paddock (17 and 21 October), control hinds were housed in a pen 5 3 4 m (other conditions as above) over the previous night. Construction of Dominance Hierarchies Hierarchies were constructed by a method previously used for red deer (3). Within each group, a dominance index was calculated for each hind, according to the number of hinds she beat; the total number of hinds that those hinds beat; the number of hinds that beat her; and the total number of hinds that beat those hinds. The indices thus calculated were ranked from highest to lowest, giving a dominance hierarchy.

RED DEER SOCIAL STATUS

693

Progesterone Assays Plasma progesterone concentrations were determined by solvent-extracted radioimmunoassay, as previously validated for deer (15). Sensitivity of the assay was 0.22 ng/mL (calculated as 2 3 standard deviation below mean total binding). The interassay coefficients of variation were 22.4% for concentrations of mean 1.3 ng/mL and 22.4% for concentrations of mean 4.8 ng/mL. The intraassay coefficients of variation at these concentrations were 15.7 and 7.9%, respectively. Analysis of Data To compare data from dominant and subordinate hinds, each hierarchy was split into two halves: animals in the top half were considered “dominant”, and animals in the bottom hand were considered “subordinate”. (The groups of 15 were arbitrarily split into eight dominant and seven subordinate members.) This method was favoured over scalar measures because of the nonlinearity of hierarchy ranks, and the nonquantitative nature of the phenomenon of dominance (22). Differences in progesterone profiles of dominant and subordinate hinds were tested using repeated-measures analysis of variance (Genstat 5, Numerical Algorithms Group, Oxford, UK). Correlation of progesterone levels with hind weight was performed by linear regressions of progesterone levels at day 10 and day 16 after CIDR removal against hind weight. RESULTS

Dominance Hierarchies In Experiment 1 mean interaction rate was 0.29 per animal/h. In Experiment 2, mean indoor interaction rate was 1.11 per animal/h; mean outdoor interaction rate was 0.48 per animal/h. Although not all possible dyadic interactions took place in all groups, sufficient displacements were recorded to allow construction of a dominance hierarchy. As with many dominance studies, hierarchies were not perfectly linear, but the consistency of the observed dyads with the calculated rank orders (an indication of the accuracy of the assumed hierarchies) varied between 68 and 97%, mean 86% (Table 1). Note that the data in Table 1 and the method of rank calculation do not include repeated observations of the same dyadic interaction. Hind Weights Hinds used in 1995 were generally of lower weight (mean 90.5 6 SEM 1.6 kg) than those used in 1994 (mean 95.7 6 1.0

kg) (two-tailed t-test p , 0.01). Subordinate hinds tended to be lighter than dominant hinds (1994: subordinate mean weight 93.2 6 1.3 kg; dominant mean weight 98.0 6 1.5 kg; two-tailed t-test, p , 0.05; 1995: subordinate mean weight 87.3 6 2.4 kg; dominant mean weight 93.6 6 1.9 kg; p 5 0.05). Progesterone Profiles In 1994, 94% of hinds, and in 1995, 88% of hinds, exhibited a rise in progesterone, beginning slowly and accelerating around 6 days after CIDR removal. The profiles showed no evidence of premature luteolysis. One hind in group 4 showed delayed signs of oestrus, and mated 5 days after CIDR removal: her progesterone profile showed a drop 2 days prior to mating, consistent with this observation. Data for this hind were included in the analysis, as they would not be expected to contribute to any difference between groups. Profiles for groups 1–4 showed similar differences between dominant and subordinate hinds, and as there was no effect of time of season on progesterone slope (two-tailed t-test, p . 0.1), data from these groups were pooled for analysis, as were data from groups 5 and 6. Characteristics of the progesterone rise differed according to year of experiment and dominance status of the hinds. Overall progesterone levels differed in 1994 and 1995 (Fig. 1); thus, mean progesterone concentrations of all the samples in the two experiments differed significantly (1994: 1.63 6 0.04 ng/mL; 1995: 0.67 6 0.03 ng/mL; two-tailed t-test, p , 0.0005). In 1994 and in the 1995 indoor groups there were interactions between dominance status and time after CIDR removal that approached significance at the 5% level, indicating differences in slopes according to status [1994: F(2.4, 125.1) 5 2.27, 0.1 . p . 0.05, Fig. 2; 1995 indoors, F(1, 10.6) 5 3.76, 0.1 . p . 0.05 Fig. 3a; 1995 outdoors, F(1.6, 8.7) 5 0.62, p . 0.1, Fig. 3b]. In the 1995 trial, progesterone profiles of hinds kept indoors did not differ significantly from those of hinds kept outdoors [there was no interaction between indoor/outdoor treatment and time after CIDR removal: F(2.5, 48) 5 1.08, p . 0.1]. Mean progesterone levels of dominant hinds exceeded an arbitrary 1-ng/mL threshold 1 day earlier than those of subordinates in 1994, and 2 days earlier in 1995 (indoor groups). Progesterone concentrations were not significantly correlated with body weight in either experiment, although there was a positive association in all cases [1994 day 10: F(1, 62) 5 0.46, p . 0.1; day 16: F(1, 62) 5 0.70, p . 0.1; 1995 day 10: F(1, 20) 5 0.44, p . 0.1; day 16: F(1, 18) 5 0.30, p . 0.1].

TABLE 1 CONSISTENCY OF OBSERVED DOMINANCE DYADS WITH CALCULATED RANK ORDERS Group

1 2 3 4 5 6 7*

No. of Hinds

Experiment Started

Housing

No. Dyads Observed

Total Possible Dyads

No. Dyads Consistent With Calculated Rank

% Dyads Consistent With Calculated Rank

15 15 18 18 8 8 8

04 Oct 95 04 Oct 94 04 Nov 94 14 Nov 94 25 Sep 95 25 Sep 95 25 Sep 95

Out Out Out Out In In Out

39 30 83 67 18 28 49

105 105 153 153 28 28 276

35 29 72 57 17 19 41

90 97 87 85 94 68 84

*1 16 other hinds.

694

GOODWIN ET AL.

FIG. 1. Summary of progesterone data for all hinds in both experiments (1994 and 1995). j, 1994 hinds (groups 1–4); m, 1995 indoor hinds (groups 5 and 6); d, 1995 outdoor control hinds (group 7). Points represent means 6 SEM. (Some time points are not represented due to sampling constraints).

DISCUSSION

Circulating progesterone profiles differed at the 10% level with dominance status, and therefore, there was a trend for dominant hinds to show higher rates of increase in progesterone levels than subordinates. This occurred under two different sets of conditions with a 3.8-fold difference in agonistic interaction rate, and is consistent with data obtained at Whipsnade Wild Animal Park (8). Overall progesterone levels were apparently independent of holding density. The fact that a marked difference in interaction rate made little difference to the comparative rate of increase of progesterone between high- and low-ranking hinds suggests that dominance status, as measured in this case by agonistic interactions, may not be the most important factor influencing ovarian function. Studies of wild deer on Rhum have made extensive use of dominance as a variable due to the ease of collecting data on

FIG. 2. Progesterone rise in dominant (j) and subordinate (m) hinds in the 1994 experiment (groups 1–4; all hinds outdoors).

FIG. 3. Progesterone rise in dominant (j) and subordinate (m) hinds in: (a) the 1995 indoor experiment (groups 5 and 6); (b) the 1995 outdoor control hinds (group 7).

agonistic interactions, compared with the relative difficulty of obtaining information on physical parameters such as body weight (S.D. Albon, personal communication). Where such data do exist, however, important correlations are present: live weights of hinds immobilised in the Rhum studies were strongly correlated with their social status (6). It is not surprising, therefore, that dominant wild hinds, being heavier and in better condition than subordinates, show the patterns of greater reproductive success described in the introduction. Hinds in the present study, however, exhibited unstable dominance relationships (due to movements between groups), the structures of which were often poorly correlated with live weight and condition (12). This may tend to obscure patterns of reproduction that are expected to differ according to dominance status of the hinds. Hind weight shows a degree of association with progesterone levels at two different times after CIDR removal in both experiments, although the link is not significant. However, it should be borne in mind that hinds at Rosemaund were generally on a high plane of nutrition, and differences in body condition are small, possibly resulting in a “ceiling effect” in

RED DEER SOCIAL STATUS

695

factors such as endocrine profiles, which may be strongly influenced by nutritional status. As in this study, there was no correlation between initial body weight and rate of rise in circulating progesterone in the earlier Whipsnade study (8). The large difference between progesterone levels in the 1994 and 1995 outdoor experiments, in which husbandry conditions were almost identical, shows that there appears to be an additional, confounding effect. As there was no effect of time of year on profiles, it is unlikely that this difference arose from the slightly earlier time at which hinds were treated in 1995. However, a possible reason for this difference between years is the adverse weather conditions preceding the second experiment (see the Methods section), which may have contributed to the body weight differences observed between 1994 and 1995. Such environmental conditions occurring while hinds were still suckling their calves may have resulted in greater energy loss as previously demonstrated in red deer (16) and cattle (17), causing hinds to be in poorer condition at the start of the 1995 experiment. Heat stress has been linked with poor fertility in cattle and other farmed animals during summer months (10,26). It is worth noting that calving rates at Rosemaund were poorer in 1996 than in previous years, and this may have been caused by a detrimental effect of climatic factors during the previous summer. As the adrenal gland is a source of progesterone in red deer (13), and adrenocortical function is increased by stress, differences between groups in terms of the stress to which they were subjected may have affected circulating progesterone levels. However, this does not appear to be an important factor, as levels in 1995—when heat stress may have been significant—were lower than in 1994; and subordinate hinds—which might be expected to be more susceptible to stressful interactions—nevertheless, had lower levels than dominant hinds.

The relationship in wild deer between body condition, dominance status, and fertility has been noted (1,6), and in farmed deer, stress has been linked to poor conception rates (2). Given that luteal progesterone forms an important indicator of ovarian function, it is unsurprising that hinds that may be under stress—whether due to environmental factors or due to bullying associated with subordinate social status—should exhibit lower concentrations of plasma progesterone. Evidence obtained from endocrine studies on sheep (25) and cattle (7,14) supports this correlation. Low levels of progesterone might interfere with reproductive efficiency in several ways. Failure to produce sufficient progesterone during early conceptus development may lead to pregnancy loss, and might be the cause of failure to breed in subordinate animals, or of a delay in establishment of pregnancy. Through its effect on uterine function, differential progesterone secretion may promote selection of conceptus gender, based on sexually dimorphic characteristics of the blastocyst. It has been postulated previously that differing progesterone profiles contribute to a mechanism of offspring sex selection according to female dominance status (8,11,19). Despite the possible obscuring influences of the high planes of nutrition at Rosemaund, trends have been observed in this study that are consistent with such a theory. ACKNOWLEDGEMENTS

We thank M. H. Davies for permission to carry out the work at ADAS Rosemaund. D. G. Chapple and other Rosemaund staff gave invaluable assistance with animal handling and management. We are also grateful to D. R. J. Brainbridge and S. D. Albon for practical assistance and helpful discussion, and to J. Craigon for statistical advice. N. Goodwin was supported by a postgraduate studentship from the University of Nottingham.

REFERENCES 1. Albon, S. D.; Mitchell, B.; Huby, B. J.; Brown, D.: Fertility in female red deer (Cervus elaphus): The effects of body composition, age and reproductive status. J. Zool. Lond. (A) 209:447–460; 1986. 2. Asher, G. W.; Fisher, M. W.; Fennessy, P. F.: Environmental constraints on reproductive performance of farmed deer. Anim. Reprod. Sci. 42:35–44; 1996. 3. Clutton–Brock, T. H.; Albon, S. D.; Gibson, R. M.; Guinness, F. E.: The logical stag: Adaptive aspects of fighting in red deer (Cervus elaphus L.). Anim. Behav. 27:211–225; 1979. 4. Clutton–Brock, T. H.; Albon, S. D.; Guinness, F. E.: Red deer: Behaviour and ecology of two sexes. Edinburgh: Edinburgh University Press; 1982. 5. Clutton–Brock, T. H.; Albon, S. D.; Guinness, F. E.: Maternal dominance, breeding success and birth sex ratios in red deer. Nature 308:358–360; 1984. 6. Clutton–Brock, T. H.; Albon, S. D.; Guinness, F. E.: Great expectations: Dominance, breeding success and offspring sex ratios in red deer. Anim. Behav. 34:460–471; 1986. 7. Diskin, M. G.; Srennan, J. M.: Progesterone and embryo survival in the cow. In: Sreenan, J. M.; Diskin, M. G., eds. Embryonic mortality in farm animals. Dordrecht: Martinus Nijhoff Publishers; 1986:142–158. 8. Flint, A. P. F.; Albon, S. D.; Loudon, A. S. I.; Jabbour, H. N.: Behavioral dominance and corpus luteum function in red deer (Cervus elaphus). Horm. Behav. 31:296–304; 1997. 9. Flint, A. P. F.; Albon, S. D.; Jaffar, S. I.: Blastocyst development and conceptus sex selection in red deer Cervus elaphus: Studies of a free-living population on the Isle of Rum. Gen. Comp. Endocrinol. 106:374–383; 1997.

10. Fuquay, J. W.: Heat stress as it affects animal production. J. Anim. Sci. 52:164–174; 1981. 11. Goodwin, N.; Hayssen, V.; Flint, A. P. F.: Effect of ovarian luteinization rate on offspring gender in dominant and subordinate red deer. J. Reprod. Fertil. Abst. 17:57; 1996. 12. Goodwin, N.; Hayssen, V.; Deakin, D. W.; Flint, A. P. F.: Instability of dominance relationships in farmed red deer. Proc. 32nd Congr. Int. Soc. Appl. Ethol. 149; 1998. 13. Jopson, N. B.; Fisher, M. W.; Suttie, J. M.: Plasma progesterone concentrations in cycling and in ovariectomised red deer hinds: The effect of progesterone supplementation and adrenal stimulation. Anim. Reprod. Sci. 23:61–73; 1990. 14. Lamming, G. E.; Darwash, A. O.; Back, H. L.: Corpus luteum function in dairy cows and embryo mortality. J. Reprod. Fertil. Suppl. 27:245–252; 1989. 15. Loudon, A. S. I.; McLeod, B. J.; Curlewis, J. D.: Pulsatile secretion of LH during the periovulatory and luteal phases of the oestrus cycle in the Pere David’s deer hind (Elaphurus davidianus). J. Reprod. Fertil. 89:663–670; 1990. 16. Loudon, A. S. I.; McNeilly, A. S.; Milne, J. A.: Nutrition and lactational control of fertility in red deer. Nature 302:145–147; 1983. 17. MacMillan, K. L.; Lean, I. J.; Westwood, C. T.: The effects of lactation on the fertility of dairy cows. Aust. Vet. J. 73:141–147; 1996. 18. Mitchell, B.; Lincoln, G. A.: Conception dates in relation to age and condition in two populations of red deer in Scotland. J. Zool. Lond. 171:141–152; 1973. 19. Pratt, N. C.; Lisk, R. D.: Role of progesterone in mediating stressrelated litter deficits in the golden hamster (Mesocricetus auratus). J. Reprod. Fertil. 92:139–146; 1991.

696 20. Reimers, T. J.; McCann, J. P.; Cowan, R. G.: Effects of storage times and temperatures on T3, T4, LH, prolactin, insulin, cortisol and progesterone concentrations in blood samples from cows. J. Anim. Sci. 57:683–691; 1983. 21. Robinson, N. A.; Leslie, K. E.; Walton, J. S.: Effect of treatment with progesterone on pregnancy rate and plasma concentrations of progesterone in Holstein cows. J. Dairy Sci. 72:202–207; 1989. 22. Rushen, J.: Should cardinal dominance ranks be assigned? Anim. Behav. 32:932–933; 1984. 23. Thouless, C. R.; Guinness, F. E.: Conflict between red deer hinds: The winner always wins. Anim. Behav. 34:1166–1171; 1986.

GOODWIN ET AL. 24. Trivers, R. L.; Willard, D. E.: Natural selection of parental ability to vary the sex ratio of offspring. Science 179:90–92; 1973. 25. Wilmut, I.; Ashworth, C. J.; Sales, D. I.: The influence of progesterone profile on embryo survival in ewes. In: Sreenan, J. M.; Diskin, M. G., eds. Embryonic mortality in farm animals. Dordrecht: Martinus Nijhoff Publishers; 1986:135–141. 26. Wolfenson, D.; Flamenbaum, I.; Berman, A.: Hyperthermia and body energy store effects on estrous behaviour, conception rate, and corpus luteum function in dairy cows. J. Dairy Sci. 71:3497– 3504; 1988.