Received 1 October 2003 Accepted 1 December 2003 Published online 31 March 2004
Multiple causes of sexual segregation in European red deer: enlightenments from varying breeding phenology at high and low latitude Christophe Bonenfant1,4, Leif E. Loe2,3, Atle Mysterud2, Rolf Langvatn3, Nils Chr. Stenseth2*, Jean-Michel Gaillard1 and Franc¸ois Klein4 1
UMR 5558 ‘Biome´trie et Biologie Evolutive’, Universite´ Claude Bernard-Lyon 1, 43, Boulevard du 11 Novembre 1918, 69622 Villeurbanne Cedex, France 2 Centre for Ecological and Evolutionary Synthesis (CEES), Department of Biology, University of Oslo, PO Box 1050 Blindern, N-0316 Oslo, Norway 3 The University Centre in Svalbard (UNIS), N-9170 Longyearbyen, Spitsbergen, Norway 4 Centre National d’Etudes et de Recherches Applique´es Cervide´s-sanglier, BP 15 Gerstheim, F-67154 Erstein Cedex, France Sexual segregation outside the mating season occurs in most species of sexually dimorphic ungulates and has been extensively described in the literature, but the mechanisms causing segregation are still debated. The detailed pattern of sexual segregation throughout the year has rarely been presented for mammals, and no study, to our knowledge, has used latitudinal-related variation in breeding phenology to shed light on the underlying mechanisms. Recent methodological developments have made it possible to quantify separate components of segregation (social, habitat) and activity synchrony in animal groups, but these major improvements have so far been little used. We observed European red deer year round at two widely different latitudes (France and Norway) and tested three different mechanistic hypotheses of segregation related to: (i) predation risk; (ii) body-size-related forage selection; and (iii) activity budget. Habitat segregation peaked during calving in both populations and dropped rapidly after calving. Females with calves were more segregated from males than were females without calves, pointing to a key role of antipredator behaviour even though large predators are absent in France and extremely rare in Norway. However, at both sites individuals also grouped with their own sex within habitat types (i.e. social segregation), and individuals in mixed-sex groups were less synchronized in activity type than individuals in either unisex male or unisex female groups, suggesting that differences in activity budgets are involved. Social segregation peaked during calving and was lowest during the rut (indicating aggregation) in both populations; these activities occurred one month later in the Northern populations, corresponding well with known differences in breeding phenology. We conclude that latitude-dependent breeding phenology shapes the seasonal pattern of sexual segregation and that sexual segregation in ungulates has multiple causes. Keywords: social and habitat segregation; predation risk; activity pattern; habitat shelter; Cervus elaphus; biogeographical gradient
In an attempt at unification, Conradt (1998b) proposed a formal and objective definition of sexual segregation by defining segregation as the departure from random association and by allowing the quantification of the social and habitat components of sexual segregation. Three hypotheses of sexual segregation are generally recognized as the most likely (Main et al. 1996; Ruckstuhl & Neuhaus 2002). First, the ‘predation-risk hypothesis’ (Main et al. 1996; Ruckstuhl & Neuhaus 2002) states that anti-predator behaviour shapes sexual segregation because females and especially offspring are more vulnerable to predation than are larger males (Bowyer 1984; Miquelle et al. 1992; Bon & Campan 1996; Main & Coblentz 1996; Main et al. 1996; Bleich et al. 1997). Predictions are that females, especially those with offspring, should choose safe habitats (vegetation cover or escape habitat) even at the expense of forage quality. Segregation should then peak around calving time (table 1). Second, the ‘forage-selection hypothesis’ (CluttonBrock et al. 1982; Main et al. 1996; Ruckstuhl & Neuhaus
Sexual segregation is common in large herbivorous mammals outside the rutting season (Miquelle et al. 1992; Main et al. 1996; Bleich et al. 1997; Bowyer et al. 2002; Ruckstuhl & Neuhaus 2002), especially in temperate regions and when the sexes are dimorphic in size (Mysterud 2000; Ruckstuhl & Neuhaus 2002). However, the causes of sexual segregation are among the most controversial topics in the foraging literature, partly because a formal and consensual definition is lacking, and partly because reported results are highly dependent on the considered methodology and spatio-temporal scale (Bowyer et al. 1996; Main et al. 1996; Conradt 1998b). Sexual segregation can be split into social segregation (sexes live in different social groups) and habitat segregation (different use of habitats) (Bon & Campan 1996; Main et al. 1996).
Author for correspondence ([email protected]
Proc. R. Soc. Lond. B (2004) 271, 883–892 DOI 10.1098/rspb.2003.2661
2004 The Royal Society
884 C. Bonenfant and others Sexual segregation and breeding phenology in red deer Table 1. A review of the hypotheses and predictions of sexual segregation addressed in the current study. (Predictions that were supported by the current study are in bold while predictions that were rejected are in italics.) hypotheses
predictions A. habitat quality
1. predationrisk hypothesis
2. forageselection hypothesis
3. activitybudget hypothesis
B. annual pattern (except rut)
C. latitude: France versus Norway
females in lowerquality habitats
more around earlier in calving France than time, or in Norway, as long as since calving female with is earlier offspring males in lowersegregation no pattern quality habitats year round
segregation year round, as long as activity differs
2000, 2002) relates habitat segregation to the sex-specific nutritional requirements that are caused by sexual bodysize dimorphism (Watson & Staines 1978; Clutton-Brock et al. 1982). Based on interspecific allometric studies showing that large herbivores can survive on a lowerquality diet than can small ones, the Jarman–Bell principle (Bell 1971; Geist 1974; Jarman 1974), we predicted that males should occupy lower-quality habitats year round (Demment & Van Soest 1985; table 1). Another derived hypothesis, ‘the scramble-competition hypothesis’, focuses on the fact that there is no good reason for males to use low-quality habitat unless females are competitively superior and exclude males at high density (Clutton-Brock et al. 1987). Indeed, among grazing ungulates, the smaller sex (females) has a competitive advantage over the larger sex (males) when feeding on short grass swards. This is related to the lower allometric coefficient between incisor arcade breadth and body size compared with the coefficient linking energy demands and body size (Illius & Gordon 1987, 1990). Males may therefore be excluded from mutually preferred vegetation types and forced to feed on low-quality but more abundant diet at high density (Staines et al. 1982; Clutton-Brock et al. 1987; but see Conradt et al. 1999). However, such predictions are valid only when we assume that intraspecific allometry is equivalent to interspecific allometry. Third, the ‘activity-budget hypothesis’ (Ruckstuhl 1998; Ruckstuhl & Neuhaus 2000, 2002) states that between-sex differences in activity pattern may lead to social segregation (bighorn sheep, Ovis canadensis (Ruckstuhl 1998); red deer, Cervus elaphus (Conradt 1998a; Conradt & Roper 2000); see Ruckstuhl & Neuhaus (2002) for a review of 22 ruminant species). Indeed, differences in activity budget would make it difficult for females and males to stay in mixed-sex groups owing to the increased cost of synchrony to maintain group Proc. R. Soc. Lond. B (2004)
D. females with and without offspring
E. social segregation (within habitat)
F. male– female synchrony
more segregation for females with offspring
no difference (if similar sized), but lactating females need more energy
no difference if similar activity pattern; yes if different activity pattern
commonly single-sex groups
Bowyer (1984); Main et al. (1996)
Short (1963); McCullough (1979); CluttonBrock et al. (1982, 1987); Main et al. (1996) mixed-sex Conradt (1998a); groups less Ruckstuhl (1998, synchronous 1999); Ruckstuhl & Neuhaus (2002)
cohesion ( Jarman 1974; Coˆte´ et al. 1997). The ‘activitybudget hypothesis’ predicts: (i) that segregation should occur all year round (except during the rut); and (ii) that mixed-sex groups should be less synchronized in their activity than single-sex groups outside the rutting season (table 1). Although breeding phenology (i.e. the timing of calving and rutting) is likely to be the driving force behind the temporal pattern of sexual segregation, the interplay between latitude-dependent breeding phenology and sexual segregation has not yet, to our knowledge, been investigated. Mammalian breeding phenology is expected to be delayed at high latitudes (Bronson 1989) and the most likely mechanism for the delayed optimal time of calving (and rutting) in northern ungulates is low temperatures leading to later plant phenology (Albon & Langvatn 1992). Several studies dealing with sexual segregation have considered data from all seasons (e.g. Miquelle et al. 1992; Villaret et al. 1997; Cransac et al. 1998; Bon et al. 2001), but all except Cransac et al. (1998), who presented monthly estimates for mouflon sheep (Ovis gmelini ), provided the segregation estimates on a coarse time-scale. High-resolution temporal patterns of sexual segregation covering the full year have therefore rarely been presented for ungulates, and previous studies have never taken advantage of methodological improvements that allow separation of the social and habitat components of segregation (Conradt 1998b). In this study, we analyse social and habitat segregation in two red deer populations living at contrasting latitudes (49° N in France and 63° N in Norway) and therefore differing in their breeding phenology (a three to four week difference; R. Langvatn and F. Klein, unpublished results; figure 1) taking advantage of the recent methodological advancements (Conradt 1998b). The predictions that we aimed to test are summarized in table 1. Whatever the
Sexual segregation and breeding phenology in red deer C. Bonenfant and others 885
Norway N rut
F M A M
calving France France
D J 49˚ N
F M A M calving
Figure 1. (a) Location map and (b) breeding phenology of European red deer at high (Norway, 63° N) and low (France, 49° N) latitudes. Rutting and calving occur one month earlier in France than in Norway.
underlying mechanisms of sexual segregation (table 1), we expect that delayed breeding phenology at high latitude (Bronson 1989) should induce an equivalent temporal delay in the social and habitat segregation pattern. 2. MATERIAL AND METHODS
mid-May, while the rutting activity peaks in late September (F. Klein, unpublished results). Thus, the breeding phenology is approximately one month earlier in France than in Norway (figure 1). Natural predators are absent but a limited number of red deer are culled annually from 15 November to 30 January (Bonenfant et al. 2002).
(a) Study areas
(b) Red deer data
In Norway, the study area is situated in the western mid-part of the country (5156 km2; centred around 63° N), well within the core area for red deer (Mysterud et al. 2001), and consists of 12 municipalities in the counties Møre og Romsdal (four) and Sør-Trøndelag (eight). Red deer in this area comprise two of five defined populations of red deer (populations three and five; e.g. Mysterud et al. 2002). The deer habitat ranges in altitude from sea level to ca. 750 m and consists mainly of pine forest (Pinus sylvestris) with pockets of mixed deciduous forest (dominated by Betula sp., Populus tremula and Sorbus aucuparia), open heather (Calluna vulgaris) and agricultural land. Red deer occasionally also use areas above the tree line (mainly when foraging at night-time during the summer), which range in altitude between 500 and 700 m depending on the distance to the coast. The climate is strongly seasonal (mean January and July temperatures are –2.2 °C and 12.9 °C, respectively). Median calving date is 16 June and median date of ovulation is ca. 15 October (Langvatn et al. 2004; figure 1). Natural predators are very scarce and include golden eagles (Aquila chrysaetos) and lynx (Lynx lynx), both being capable of killing calves only. Human hunters annually cull red deer from 10 September to 15 November. The French study area is La Petite Pierre National Reserve, located in northeast France (28 km2; 49° N) in the Vosges Mountains, at a mean elevation of 300 m. The habitat is mainly forested, with trees covering ca. 85% of the total area. The heterogeneous canopy consists predominantly of a mixture of silver fir (Abies alba), Douglas fir (Pseudotsuga douglasii ), Norway spruce (Picea abies) and European beech (Fagus sylvatica). The climate is continental with oceanic influences (mean January and July temperatures are 0.6 °C and 18.4 °C, respectively) leading to warm winters and cool summers. Most calves are born in
In Norway, observations consist of three different types. In early spring (March and April), deer were observed primarily on farmland by research personnel in vehicles and by farmers (from observations on their properties). All year, research personnel on foot observed deer in all types of terrain, but with variable intensity. Hunters recorded groups of red deer in the period 10 September to 15 November (also in all types of terrain). Altogether, research personnel, farmers and hunters determined the age (according to three classes: calf, yearling and adult) and sex of red deer in 2045 groups (group size of between 1 and 56) from 4 January to 25 December in the years 1956–1984 and classified habitat into one of five types (table 2). In France, managers and research personnel from the Office National de la Chasse et de la Faune Sauvage performed all observations year round either on foot or by car. Altogether, 1912 groups (including observations of single animals) were recorded from 2 January to 30 December from 1980 to 1999. The sex and the age (calf versus adult) of each individual in a sighted group were determined. Field observers also recorded whether the females did or did not have a calf at heel (the status ‘unknown’ was assigned whenever the motherhood relationship was not formally established). Finally, the habitat occupied by each group was assigned to one of the seven different types (table 2).
Proc. R. Soc. Lond. B (2004)
(c) Statistical analyses (i) Definition of segregation, aggregation and synchronization coefficients The study of the interplay between habitat segregation and social segregation (grouping with their own sex regardless of habitat) has only recently been made possible. Conradt (1998b) developed an index of sexual segregation that is independent of
886 C. Bonenfant and others Sexual segregation and breeding phenology in red deer Table 2. Description of the habitat types found in Norway and France. habitat type
Norway pine forest
open pine (Pinus sylvestris) forest with blueberry (Vaccinium myrtillus) and heather (Calluna vulgaris) in the field layer a mix of deciduous (Betula sp., Populus tremula and Sorbus aucuparia), spruce (Picea abies) and pine (Pinus sylvestris) forest with juniper ( Juniperus communis) and willow (Salix sp.) in the field layer; dense forest type open dry glens (no trees) with Calluna vulgaris dominating the field layer treeless land at 500–700 m altitude; willow, blueberry and crowberry (Empetrum nigrum) in the field layer Timothy (Phleum pratense) and other grass species farmed as winter feed for cattle and sheep
heather land above tree line agricultural land France agricultural land meadows regeneration
plantation pole plantation timber-tree
open fields of corn, sugar beet or alfalfa; these standing crops are non-fenced and freely available to deer open fields of grass species farmed mainly for fodder production and summer feed for cattle typical colonizing vegetation with birch (Betula pendula), juniper ( Juniperus communis), hawthorn (Crataegus sp.) and brambles (Rubus sp.) following clear-cut events (exploitation or Lothar storm): less than 4 m in height but providing very good sheltering and resting habitats to deer young-stage (shoots less than 5 years old) and still-not-exploited deciduous (beech, hornbeam, oak), coniferous (spruce, fir) or mixed forest; semi-open to dense forest type closed uniformly distributed forest type consisting of Douglas fir (Pseudotsuga douglasii ) or Norway spruce (Picea abies) forest from 5 to 15 years old; poorly developed field and shrub layer a mix of deciduous oak (Quercus sp.), hornbeam (Carpinus betulus), beech (Fagus sylvatica) spruce and pine forest with juniper ( Juniperus communis); the field layer is composed of blueberry (Myrtillus sp.) and mosses, brambles or hairy wood-rush (Luzula pilosa) and bracken (Pteridium aquilinum) following the dominant timber-tree species; dense forest type identical to timber-tree habitat except for an additional and intermediate layer of young growing poles (mainly beech, hornbeam, spruce and fir); dense forest type
population sex ratio, population density and group size. We therefore calculated the magnitudes of both the social segregation, SCsocial (equation (2.1)), and the habitat segregation, SChabitat (equation (2.2)), components of sexual segregation following Conradt (1998b) for each month of the year (whenever sample sizes were greater than 30 group observations) and for each female reproductive status (with or without a calf at heel). All segregation coefficients can take values between 1 (total segregation) and –1 (total aggregation). Individuals of both sexes can be said to associate randomly if SCsocial is not statistically different from 0. We assessed synchrony within three group types (mixed groups, only-female groups and only-male groups) using a similar coefficient to the segregation coefficients proposed by Conradt (1998b) (equation (2.3)).
SCsocial = 1 ⫺
xi yi N . XY i = 1ni ⫺ 1
In equation (2.1), xi is the number of males in the ith group, yi is the number of females in the ith group, ni is the group size of the ith group (ni = xi ⫹ yi ), k is the number of groups with at least two animals, X is the total number of males sampled (excluding solitary animals), Y is the total number of females sampled (excluding solitary animals) and N is the total number of males and females sampled.
SChabitat = 1 ⫺
zi wi M . ZWi = 1mi ⫺ 1
In equation (2.2), zi is the number of males in the ith habitat type, wi is the number of females in the ith habitat type, mi is the number of males and females in the ith habitat type (mi = zi ⫹ wi ), l is the number of habitat types that are used by Proc. R. Soc. Lond. B (2004)
at least two animals, Z is the total number of males sampled, W is the total number of females sampled and M is the total number of males and females sampled. Contrary to the calculation of social segregation, when calculating SChabitat solitary individuals are included.
Nh Nh ⫺ 1 h ah, j rh, j . N AhRh j = 1nh, j ⫺ 1 h = 08.00 20.00
SynC = 1 ⫺
In equation (2.3), N is the total number of animals observed, Nh is the number of animals observed in hour h of the day, Ah is the number of animals observed foraging in hour h of the day, Rh is the number of animals observed resting in hour h of the day, nh,j is the total number of animals observed in hour h of the day in the jth group, ah,j is the number of animals observed foraging in hour h of the day in the jth group and rh,j is the number of animals observed resting in hour h of the day in the jth group. Standard errors are not readily available from the original definition of the segregation coefficients by Conradt (1998a,b). Consequently, we could not perform basic statistical inference and hypothesis testing. We thus computed standard errors by bootstrapping the segregation-coefficient estimators (Efron & Tibshirani 1993). We replicated the analysis 1000 times with randomly selected subsamples of our data and used the bootstrap mean as an unbiased estimator of segregation coefficients (Efron & Tibshirani 1993). Under the assumption that the SC distribution is correctly approximated by a normal distribution, computation of standard errors allowed us to test biological hypotheses and to compare observed segregation coefficients with a theoretical value of 0 (i.e. the null hypothesis that males and females aggregate at random). A null hypothesis was not
Sexual segregation and breeding phenology in red deer C. Bonenfant and others 887
rejected when two 95% confidence intervals (CI) overlapped or when the theoretical value lay within the 95% CI.
(ii) Testing the seasonal pattern of segregation between groups Comparisons of temporal patterns were investigated with simple correlations of monthly SC estimates in the two populations. According to the recommendation of Conradt (1998b) to use non-parametric statistics we used Spearman’s rank correlation test (Sokal & Rohlf 1995) to compare within-year variations in SC between Norway and France and to test for a difference in the variation of SC over time between females with and without a calf at heel. Similarly, we compared the mean SCs of red deer in the two countries with a Mann–Whitney U-test (Sokal & Rohlf 1995).
(iii) Differentiating habitat use by sex Between-sex differences in habitat use were investigated within a generalized linear model framework (McCullagh & Nelder 1989). Because the response variable is the number of individuals counted, we used a Poisson regression and a gamma error distribution. Factors entered in the model were ‘habitat type’ (a five- and seven-modality factor for Norway and France, respectively), ‘time’ (12 levels from January to December) and ‘sex’ (two modalities: male and female). As we were interested in the relative use of each habitat type by each sex, i.e. the proportion of males and females found in each habitat over time, third- and second-order interactions including the factor ‘habitat type’ were the only terms of interest. All statistical analyses were performed by using the S-plus statistical package (Venables & Ripley 1999).
3. RESULTS Social segregation was higher overall than habitat segregation, indicating that individuals grouped with their own sex within large-scale habitat types (monthly estimates of SCsocial ⬎ SChabitat: Norway: 2 = 8.67, d.f. = 1, p = 0.003; and France: 2 = 8.67, d.f. = 1, p = 0.003; figure 2). Habitat segregation was most pronounced and significantly different from 0 only during calving time (figure 2b). Habitat-use analyses selected the same model in both countries including the third-order interaction among habitat, time and sex (Norway: 2 = 87.984, d.f. = 39, p ⬍ 0.0001; France: 2 = 307.877, d.f. = 72, p ⬍ 0.0001). At this time ( June) in Norway, adult females stayed in sheltered deciduous forest (64%, n = 69, agricultural land excluded; figure 3a), while males occupied high-altitude non-forested areas recently free from snow (57%, n = 21, agricultural land excluded; figure 3b). Correspondingly, female deer in France stayed in closed and mostly sheltered habitat (timber-tree 81%, n = 132; figure 3c), while the males used open habitat such as agricultural lands and meadows (20%, n = 13) or plantation (37%, n = 25) during the calving time (May) (figure 3d ). Social segregation varied extensively throughout the year, showing a peak during the calving time immediately followed by a return to values similar to those prior to calving and then a drop below 0 during the rut (indicating aggregation) in both red deer populations (figure 2a). Sexes were consistently more segregated in France than in Norway outside the rutting season (September and October; figure 2a). Females with a calf at heel were more segregated from Proc. R. Soc. Lond. B (2004)
males than were females without a calf at heel in both populations especially during the calving season (Wilcoxon signed-rank test for paired data: Norway: Z = ⫺2.096, p = 0.036, n = 6; and France: Z = ⫺2.392, p = 0.016, n = 12; figure 4). Mixed-sex groups were less synchronized in activity than only-male and only-female groups outside the rutting season (figure 5). The pattern of sexual segregation (both social and habitat) was delayed by one month in the northern populations, corresponding well with the known delay in breeding phenology (the correlations between social segregation of both populations without and with correction for the time delay in breeding phenology were, respectively: r 2 = 0.713, Z = 2.354, p = 0.018, n = 11; and r 2 = 0.867, Z = 2.864, p = 0.004, n = 11; figure 1). In the rutting season, males and females in France aggregated only weakly and less than in Norway ( 2 = 10.83, d.f. = 1, p = 0.001; figure 2a). French and Norwegian stags held on average the same number of hinds during the rutting season (mean ± s.e. of 1.6 ± 0.08 and 1.5 ± 0.07 females per male, respectively). 4. DISCUSSION We have shown that the annual cycle of sexual segregation is closely related to breeding phenology by reporting for the first time, to our knowledge, a temporal delay in annual sexual-segregation pattern in the same species at two different latitudes. Although the annual cycle was predicted by the predation-risk hypothesis, several other predictions of the activity-budget hypothesis were also supported. Therefore, the pattern of sexual segregation has multiple causes and the different hypotheses are not mutually exclusive, but rather explain different aspects of the complex pattern of sexual segregation (summarized in table 1). (a) The predation hypothesis and evidence based on breeding phenology The predation-risk hypothesis has received massive support in explaining sexual segregation in ungulates: females with young often select safer habitat even at the expense of forage quality (Bergerud et al. 1984; Francisci et al. 1985; Shank 1985; Festa-Bianchet 1988; Berger 1991; Miquelle et al. 1992; du Toit 1995; Main & Coblentz 1996; Main et al. 1996; Bleich et al. 1997; Villaret et al. 1997). In our study, social segregation and habitat segregation peaked when calves were newborn and probably most prone to predation, and was delayed in Norway compared with France, demonstrating the utility of accounting for breeding phenology in sexual-segregation studies. Females stayed in more sheltered habitat, while at the same time males used habitat of superior nutritional quality in Norway (Albon & Langvatn 1992) and France. It is somewhat surprising that the predation-risk hypothesis was the main determinant of sexual segregation, as red deer predators (mainly wolves and bears) have been scarce or completely absent both in east France (Breitenmoser 1998) and Norway (Swenson et al. 1995, 1998) during the past 100 years, or approximately 13 red deer generations (generation time of 7 years; Clutton-Brock et al. 1982). Witnessing the ghost of past predation after so long suggests that evolution of optimal behaviour in ungulates has been very strongly influenced by predators (Byers 1997).
888 C. Bonenfant and others Sexual segregation and breeding phenology in red deer
Feb Mar Apr May Jun Jul Aug Sep time (months)
Oct Nov Dec
Feb Mar Apr May Jun Jul Aug Sep time (months)
Oct Nov Dec
Figure 2. Monthly sexual segregation in Norwegian (N) (open circles) and French (F) (filled circles) red deer (mean ± 95% CI) as calculated by the segregation coefficients in Conradt (1998b) and disentangled into the components of (a) social and (b) habitat segregation.
(b) Females with and without young: a role for lactation or predation? Females caring for young may segregate from males as well as other females according to all the three segregation hypotheses, but the detailed temporal pattern can be used to distinguish between the three possibilities. Energetic needs are highest when the females lactate because the production of milk is extremely costly (Sadleir 1969; Loudon 1985; Robbins 1993). Lactating females can therefore have a different activity budget from other females (Ruckstuhl & Neuhaus 2002). Females with calves were more socially segregated from males than were females without calves. In addition, earlier studies reported that females with young tend to form groups on their own in several ungulate species (bison, Bison bison (Rutberg 1984; Komers et al. 1993), and chamois, Rupicapra rupicapra (Pe´rez-Barberı´a & Nores 1994)). But, if sexual segregation is caused by females grazing on high-quality pastures to Proc. R. Soc. Lond. B (2004)
compensate for energy loss to lactation, then segregation should be high (though gently decreasing) during the entire period of intensive lactation (ca. 80 days for red deer; Clutton-Brock et al. 1982) and not only during calving. Both social and habitat segregation, as well as the segregation difference between reproducing and nonreproducing females, dropped dramatically the month after calving, long before females stop lactating, which is supposed to last for at least six months in red deer (Clutton-Brock et al. 1982). In addition, reproductive females in charge of their offspring have been shown to use low-quality habitat in several different species (Edwards 1983; Bowyer et al. 1999). Therefore the sexual-segregation peak we observed in spring in both populations (figure 2) cannot be accounted for by sexual differences in forage selection or activity budget caused by lactation alone, and is likely to be related to the antipredator behaviour.
Sexual segregation and breeding phenology in red deer C. Bonenfant and others 889
proportion of time feeding
0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
proportion of time feeding
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Figure 3. Proportions of different habitat types used each month by male and female red deer in Norway and France. (a) Norway habitat use, females; (b) Norway habitat use, males (dark blue, pine forest; dark red, mixed forest; green, heather land; orange, above tree line; mid-blue, agricultural land); (c) France habitat use, females; and (d ) France habitat use, males (dark blue, agricultural land; dark red, human buildings; green, meadows; orange, regeneration; mid-blue, plantation; brown, pole plantation; mid-red, timber-tree; light blue, regenerating timber-tree).
(c) Social segregation Sexual segregation was long thought to be a by-product of different habitat selection by females and males (Clutton-Brock et al. 1987), even though some studies reported segregation despite small differences in habitat use between the sexes (Cransac et al. 1998; Ruckstuhl 1998). Sexual segregation regardless of habitat (i.e. social segregation; Conradt 1998a) has since been documented conclusively (Conradt 1999; Conradt et al. 1999, 2001), providing an intuitively appealing mechanism (Conradt 1998a; Ruckstuhl 1998) and taking advantage of recent advances in methodology (Conradt 1998b). Our study confirms that a large proportion of sexual segregation is accounted for by social segregation (figure 2) and that mixed-sex groups were less synchronized than only-male and only-female groups outside the rutting season (figure 5). Different optimal activity budgets in males and females is the most likely mechanism explaining the component of sexual segregation that is not explained by habitat segregation (Ruckstuhl & Neuhaus 2002). (d ) Lack of evidence for the ‘forage-selection’ hypothesis We found no support for the ‘forage-selection’ hypothesis in accordance with other recent results (Conradt 1999; Conradt et al. 1999), although red deer are one of the most dimorphic ungulate species (more dimorphic Proc. R. Soc. Lond. B (2004)
than the 20% threshold needed for sexual segregation to occur; Ruckstuhl & Neuhaus 2002) and several empirical studies provide evidence for males having higher energy needs than females. However, the Jarman–Bell principle was coined at the interspecific level to explain variations in energy requirements and differences in diet selection between the African elephant (Loxodonta africana) and dik-dik (Madoqua kirki ) and more recently between the cow (Bos taurus) (750 kg) and Thomson’s gazelle (Gazella thomsoni ) (20 kg) (Wilmshurst et al. 2000). The reasoning might be ill founded to explain sexual segregation at an intraspecific level, with some 50 kg weight difference (Conradt et al. 1999). At this level, local environmental conditions and individual quality rather than evolutionary constraints might shape energy needs. (e) Geographical variation in the red deer social system The latitudinal-related variation in segregation, probably linked to breeding phenology, was not the only geographical variation in the social organization of red deer. Sexes were generally more socially segregated in France than in Norway outside the rutting season. Norwegian red deer’s extensive use of agricultural land may lead to the formation of mixed foraging groups in a small space, which split up once they are back in their natural habitat. This will decrease the social-segregation estimates
890 C. Bonenfant and others Sexual segregation and breeding phenology in red deer
0.6 0.4 0.2 0 –0.2 –0.4 Jan
Feb Mar Apr May Jun
Oct Nov Dec
0.2 0 –0.2 –0.4 –0.6 –0.8 Apr
Jul time (months)
Figure 4. Monthly sexual segregation (social component, mean ± 95% CI) in red deer hinds with (filled circles) and without (open circles) a calf at heel in (a) France and (b) Norway.
compared with those for France, where much less farmland is available for the deer in the forested area studied. Our study of red deer in France and Norway also contrasts with the most well-known one on the Isle of Rum, Scotland, where prime-aged males hold large harems in open terrain (Clutton-Brock et al. 1982). Mating systems in mammals are related to the spatio-temporal distribution of oestrus females (Emlen & Oring 1977; Clutton-Brock 1989), which in turn is closely linked to food-resource distribution. The small number of hinds held by stags in our study compared with red deer on Rum (Clutton-Brock et al. 1982) implies a largely different mating system in forested areas from that in more open areas (e.g. Calef & Van Camp 1987). The common male mating strategy in our two forested study areas seems to be the tending of one or two females. The weaker aggregation in France than in Norway during the rut suggests that French males moved between female groups more often than did Proc. R. Soc. Lond. B (2004)
Norwegian males (and therefore were observed more often alone). Indeed, male mating strategy in European red deer seems to be highly flexible (Carranza et al. 1995).
5. CONCLUSION We conclude that latitude-dependent breeding phenology drives an annual cycle of sexual segregation, which peaks during calving and drops during the rut, and that the segregation peak in spring is probably caused by a shelter-seeking anti-predator strategy in female European red deer at both high and low latitudes. Sexual segregation is mainly driven by social segregation and not by habitat segregation. We further suggest that a sexual difference in optimal activity budgets is a likely mechanism for the social part of sexual segregation in red deer, and that sexual segregation in ungulates hence has multiple causes.
degree of activity synchronization
Sexual segregation and breeding phenology in red deer C. Bonenfant and others 891
1.0 0.9 0.8 0.7 0.6 0.5 0.4 MM
FF group type
Figure 5. Synchronization coefficients for Norway (mean ± 95% CI) according to the kind of group: onlyfemale (FF), only-male (MM) and mixed-sex (FM) groups. Mean values (with associated s.d.): MM: 0.850 (0.079); FF: 0.758 (0.027); and FM: 0.558 (0.054).
We gratefully acknowledge the financial support of the Research Council of Norway to L.E.L., A.M. and N.C.S. C.B. has received grants from the Office National de la Chasse et de la Faune Sauvage (ONCFS) and the Centre National de la Recherche Scientifique (CNRS). We thank Larissa Conradt for explaining details about the segregation coefficients. Part of this work was initiated as a result of the financial support of the ‘Aurora Programme’ by the French Ministry for Foreign Affairs.
REFERENCES Albon, S. D. & Langvatn, R. 1992 Plant phenology and the benefits of migration in a temperate ungulate. Oikos 65, 502–513. Bell, R. H. V. 1971 A grazing ecosystem in the Serengeti. Sci. Am. 225, 86–89. Berger, J. 1991 Pregnancy incentives, predation constraints and habitat shifts: experimental and field evidence for wild bighorn sheep. Anim. Behav. 41, 61–77. Bergerud, A. T., Butler, H. E. & Miller, D. R. 1984 Antipredator tactics of calving caribou: dispersion in mountains. Can. J. Zool. 62, 1566–1575. Bleich, V. C., Bowyer, R. T. & Wehausen, J. D. 1997 Sexual segregation in mountain sheep: resources or predation? Wildl. Monogr. 134, 3–50. Bon, R. & Campan, R. 1996 Unexplained sexual segregation in polygamous ungulates: a defense of an ontogenetic approach. Behav. Proc. 38, 131–154. Bon, R., Rideau, C., Villaret, J.-C. & Joachim, J. 2001 Segregation is not only a matter of sex in Alpine ibex, Capra ibex ibex. Anim. Behav. 62, 495–504. Bonenfant, C., Gaillard, J.-M., Klein, F. & Loison, A. 2002 Sex- and age-dependent effects of population density on lifehistory traits of red deer Cervus elaphus in a temperate forest. Ecography 25, 446–458. Bowyer, R. T. 1984 Sexual segregation in southern mule deer. J. Mamm. 65, 410–417. Bowyer, R. T., Kie, J. G. & Van Ballenberghe, V. 1996 Sexual segregation in black-tailed deer: effects of scale. J. Wildl. Mngmt 60, 10–17. Bowyer, R. T., Van Ballenberghe, V., Kie, J. G. & Maier, J. A. K. 1999 Birth-site selection by Alaskan moose: maternal strategies for coping with a risky environment. J. Mamm. 80, 1070–1083. Proc. R. Soc. Lond. B (2004)
Bowyer, R. T., Stewart, K. M., Wolfe, S. A., Blundell, G. M., Lehmkuhl, K. L., Joy, P. J., McDonough, T. J. & Kie, J. G. 2002 Assessing sexual segregation in deer. J. Wildl. Mngmt 66, 536–544. Breitenmoser, U. 1998 Large predators in the Alps: the fall and rise of man’s competitors. Biol. Conserv. 83, 279–289. Bronson, F. H. 1989 Mammalian reproductive biology. University of Chicago Press. Byers, J. A. 1997 American pronghorn: social adaptations and the ghosts of predators past. University of Chicago Press. Calef, G. W. & Van Camp, J. 1987 Seasonal distribution, group size and structure, and movements of bison herds. Can. Wildl. Service Occasional Pap. 63, 15–20. Carranza, J., Garcia-Mun˜oz, A. J. & de Dios Vargas, J. 1995 Experimental shifting from harem defence to territoriality in rutting red deer. Anim. Behav. 49, 551–554. Clutton-Brock, T. H. 1989 Mammalian mating systems. Proc. R. Soc. Lond. B 236, 339–372. Clutton-Brock, T. H., Guinness, F. E. & Albon, S. D. 1982 Red deer: behaviour and ecology of two sexes. Edinburgh University Press. Clutton-Brock, T. H., Iason, G. R. & Guinness, F. E. 1987 Sexual segregation and density-related changes in habitat use in male and female red deer (Cervus elaphus). J. Zool. Lond. 211, 275–289. Conradt, L. 1998a Could asynchrony in activity between the sexes cause intersexual social segregation in ruminants? Proc. R. Soc. Lond. B 265, 1359–1363. (DOI 10.1098/rspb. 1998.0442.) Conradt, L. 1998b Measuring the degree of sexual segregation in group-living animals. J. Anim. Ecol. 67, 217–226. Conradt, L. 1999 Social segregation is not a consequence of habitat segregation in red deer and feral Soay sheep. Anim. Behav. 57, 1151–1157. Conradt, L. & Roper, T. J. 2000 Activity synchrony and social cohesion: a fission–fusion model. Proc. R. Soc. Lond. B 267, 2213–2218. (DOI 10.1098/rspb.2000.1271.) Conradt, L., Clutton-Brock, T. H. & Thomson, D. 1999 Habitat segregation in ungulates: are males forced into suboptimal foraging habitats through indirect competition by females? Oecologia 119, 367–377. Conradt, L., Gordon, I. J., Clutton-Brock, T. H., Thompson, D. & Guinness, F. E. 2001 Could the indirect competition hypothesis explain intersexual site segregation in red deer (Cervus elaphus L.)? J. Zool. 254, 185–193. Coˆte´, S. D., Schaefer, J. A. & Messier, F. 1997 Time budgets and synchrony of activities in muskoxen: the influence of sex, age, and season. Can. J. Zool. 75, 1628–1635. Cransac, N., Gerard, J. F., Maublanc, M. L. & Pepin, D. 1998 An example of segregation between age and sex classes only weakly related to habitat use in mouflon sheep (Ovis gmelini ). J. Zool. 244, 371–378. Demment, M. W. & Van Soest, P. J. 1985 A nutritional explanation for body size patterns of ruminant and non-ruminant herbivores. Am. Nat. 125, 641–672. du Toit, J. T. 1995 Sexual segregation in kudu: sex differences in competitive ability, predation risk or nutritional needs? S. Afr. J. Wildl. Res. 25, 127–134. Edwards, J. 1983 Diet shifts in moose due to predator avoidance. Oecologia 60, 185–189. Efron, B. & Tibshirani, R. J. 1993 An introduction to the bootstrap. London: Chapman & Hall. Emlen, S. T. & Oring, L. W. 1977 Ecology, sexual selection and the evolution of mating systems. Science 197, 215–223. Festa-Bianchet, M. 1988 Seasonal range selection in bighorn sheep: conflicts between forage quality, forage quantity, and predator avoidance. Oecologia 75, 580–586.
892 C. Bonenfant and others Sexual segregation and breeding phenology in red deer Francisci, F., Focardi, S. & Boitani, L. 1985 Male and female alpine ibex: phenology of space use and herd size. In The biology and management of mountain ungulates (ed. S. Lovari), pp. 124–133. Kent, Australia: Croom Helm Ltd. Geist, V. 1974 On the relationship of social evolution and ecology in ungulates. Am. Zool. 14, 205–220. Illius, A. W. & Gordon, I. J. 1987 The allometry of food intake in grazing ruminants. J. Anim. Ecol. 56, 989–999. Illius, A. W. & Gordon, I. J. 1990 Variation in foraging behaviour in red deer and the consequences for population biology. J. Anim. Ecol. 59, 89–101. Jarman, P. J. 1974 The social organisation of antelope in relation to their ecology. Behaviour 48, 215–266. Komers, P. E., Messier, F. & Gates, C. C. 1993 Group structure in wood bison: nutritional and reproductive determinants. Can. J. Zool. 71, 1367–1371. Langvatn, R., Mysterud, A., Stenseth, N. C. & Yoccoz, N. G. 2004 Timing and synchrony of ovulation in red deer constrained by short northern summers. Am. Nat. (In the press.) Loudon, A. S. I. 1985 Lactation and neonatal survival of mammals. Symp. Zool. Soc. Lond. 54, 183–207. McCullagh, P. & Nelder, J. A. 1989 Generalized linear models. London: Chapman & Hall. McCullough, D. R. 1979 The George Reserve deer herd: population ecology of a K-selected species. Ann Arbor, MI: The University of Michigan Press. Main, M. B. & Coblentz, B. E. 1996 Sexual segregation in Rocky Mountain mule deer. J. Wildl. Mngmt 60, 497–507. Main, M. B., Weckerly, F. W. & Bleich, V. C. 1996 Sexual segregation in ungulates: new directions for research. J. Mamm. 77, 449–461. Miquelle, D. G., Peek, J. M. & Van Ballenberghe, V. 1992 Sexual segregation in Alaskan moose. Wildl. Monogr. 122, 1–57. Mysterud, A. 2000 The relationship between ecological segregation and sexual body-size dimorphism in large herbivores. Oecologia 124, 40–54. Mysterud, A., Yoccoz, N. G., Stenseth, N. C. & Langvatn, R. 2001 The effects of age, sex and density on body weight of Norwegian red deer: evidence of density-dependent senescence. Proc. R. Soc. Lond. B 268, 911–919. (DOI 10.1098/rspb.2001.1585.) Mysterud, A., Langvatn, R., Yoccoz, N. G. & Stenseth, N. C. 2002 Large-scale habitat variability, delayed density effects and red deer populations in Norway. J. Anim. Ecol. 71, 569–580. Pe´rez-Barberı´a, F. J. & Nores, C. 1994 Seasonal variation in group size of Cantabrian chamois in relation to escape terrain and food. Acta Theriol. 39, 295–305.
Proc. R. Soc. Lond. B (2004)
Robbins, C. T. 1993 Wildlife feeding and nutrition. San Diego, CA: Academic. Ruckstuhl, K. E. 1998 Foraging behaviour and sexual segregation in bighorn sheep. Anim. Behav. 56, 99–106. Ruckstuhl, K. E. 1999 To synchronise or not to synchronise: a dilemma for young bighorn males? Behaviour 136, 805–818. Ruckstuhl, K. E. & Neuhaus, P. 2000 Sexual segregation in ungulates: a new approach. Behaviour 137, 361–377. Ruckstuhl, K. E. & Neuhaus, P. 2002 Sexual segregation in ungulates: a comparative test of three hypotheses. Biol. Rev. 77, 77–96. Rutberg, A. T. 1984 Birth synchrony in American bison (Bison bison): response to predation or season? J. Mamm. 65, 418–423. Sadleir, R. M. F. S. 1969 The ecology and reproduction in wild and domestic mammals. London: Methuen. Shank, C. C. 1985 Inter- and intra-sexual segregation of chamois (Rupicapra rupicapra) by altitude and habitat during summer. Z. Saugetierkd. 50, 117–125. Short, H. L. 1963 Rumen fermentations and energy relationships in white-tailed deer. J. Wildl. Mngmt 27, 184–195. Sokal, R. R. & Rohlf, F. J. 1995 Biometry. New York: Freeman. Staines, B. W., Crisp, J. M. & Parish, T. 1982 Differences in the quality of food eaten by red deer (Cervus elaphus) stags and hinds in winter. J. Appl. Ecol. 19, 65–77. Swenson, J. E., Wabakken, P., Sandegren, F., Bja¨rvall, A., Franze´n, R. & So¨derberg, A. 1995 The near extinction and recovery of brown bears in Scandinavia in relation to the bear management policies of Norway and Sweden. Wildl. Biol. 1, 11–26. Swenson, J. E., Sandegren, F. & So¨derberg, A. 1998 Geographic expansion of an increasing brown bear population: evidence for presaturation dispersal. J. Anim. Ecol. 67, 819–826. Venables, W. N. & Ripley, B. D. 1999 Modern applied statistics with S-plus. New York: Springer. Villaret, J. C., Bon, R. & Rivet, A. 1997 Sexual segregation of habitat by the alpine ibex in the French Alps. J. Mamm. 78, 1273–1281. Watson, A. & Staines, B. W. 1978 Differences in the quality of wintering areas used by male and female red deer (Cervus elaphus) in Aberdeenshire. J. Zool. 186, 544–550. Wilmshurst, J. F., Fryxell, J. M. & Bergman, C. M. 2000 The allometry of patch selection in ruminants. Proc. R. Soc. Lond. B 267, 345–349. (DOI 10.1098/rspb.2000.1007.)
As this paper exceeds the maximum length normally permitted, the authors have agreed to contribute to production costs.