SEARCH

SEARCH BY CITATION

Keywords:

  • body mass;
  • condition-dependent dispersal;
  • dispersal distance;
  • GPS ;
  • habitat heterogeneity;
  • roe deer;
  • ungulates

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  1. Natal dispersal is defined as the movement between the natal range and the site of first breeding and is one of the most important processes in population dynamics. The choice an individual makes between dispersal and philopatry may be condition dependent, influenced by either phenotypic attributes and/or environmental factors.
  2. Interindividual variability in dispersal tactics has profound consequences for population dynamics, particularly with respect to metapopulation maintenance. A better understanding of the mechanisms underlying this variability is thus of primary interest.
  3. We investigated the ranging behaviour of 60 juvenile European roe deer, Capreolus capreolus, monitored with GPS collars for 1 year prior to their first reproduction, from 2003 to 2010 in South-West France. Dispersal occurs across a spatial continuum so that dividing individuals into two categories (dispersers vs. philopatric) may lead to information loss. Therefore, to investigate condition-dependent dispersal more accurately, we developed an individual-based measure of dispersal distance, which took into account interindividual variation in ranging behaviour. We assessed the influence of body mass, the degree of habitat heterogeneity and sex on dispersal initiation date, dispersal propensity and distance.
  4. The overall population dispersal rate was 0·34, with a mean ± SD linear distance between natal and post-dispersal home ranges of 12·3 ± 10·5 km. Dispersal distances followed a classical leptokurtic distribution. We found no sex bias in either dispersal rate or distance. Forest animals dispersed less than those living in more heterogeneous habitats. Heavier individuals dispersed with a higher probability, earlier and further than lighter individuals. Our individual-based standardised dispersal distance increased linearly with body mass, with some suggestion of a body mass threshold of 14 kg under which no individual dispersed.
  5. Natal dispersal in roe deer was thus dependent on both phenotypic attributes and environmental context. Our results suggest that population connectivity can be altered by a change in average body condition and is likely higher in the rich and heterogeneous habitats typical of modern day agricultural landscapes.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Natal dispersal, the net movement between the natal area and the site of first breeding (Howard 1960), is one of the most important individual life-history traits affecting population dynamics and species evolution (Greenwood 1980; Clobert et al. 2001; Nathan 2001). Natal dispersal is a three-stage process involving the decision to leave the natal area, a transient phase, and finally settlement in the adult home range (Stenseth & Lidicker 1992). However, despite its importance, natal dispersal is still poorly understood, and the factors shaping variation in both dispersal rates and distances remain unknown in most cases (Ronce 2007). The main ultimate causes of natal dispersal involve inbreeding avoidance, competition (for mates or resources and kin interactions) and habitat instability (Bowler & Benton 2005).

In most populations, not all individuals disperse, and dispersing individuals are generally not a random subset of the population (Ronce 2007). Indeed, the decision an individual makes to stay in the natal area or leave it may be condition dependent (Bowler & Benton 2005). The term condition dependence encompasses the effects of both the individual's phenotype (e.g. fat reserves, body size or competitive ability) and the animal's environmental features (e.g. population density or habitat quality) on dispersal behaviour (Ims & Hjermann 2001). Both sets of factors can interact with one another in complex and subtle ways to determine dispersal outcomes. Indeed, environmental factors can affect dispersal through a direct pathway or indirectly, mediated by changes in phenotypic attributes (Ims & Hjermann 2001). Condition dependence could translate into individual differences in dispersal behaviour and can thereby indirectly generate variability in dispersal propensity among individuals, implying that dispersal costs and benefits differ among individuals (Bowler & Benton 2005). Indeed, the balance between the costs and benefits of dispersal depends on the internal state of the individual (Clobert et al. 2009), involving either fixed (e.g. sex) or time-dependent (e.g. body condition) traits. That is, dispersal may allow animals to obtain good quality ranges or to escape local competition, but since dispersal is costly (Nunes & Holekamp 1996; Dufty & Belthoff 2001), animals face a conflict the resolution of which may depend on environmental conditions or individual phenotypic attributes. For example, individuals that disperse before they have attained a certain threshold of body condition may increase the mortality risk associated with dispersal (Dufty & Belthoff 2001). As a general rule, dispersal costs, and therefore mortality risk, increase with dispersal distance, as has been demonstrated, for example, in American marten Martes americana (Johnson & Gaines 1990). After long being ignored, it is now recognised that interindividual differences in dispersal are important (Gibbs et al. 2009).

Dispersal is usually described as a dichotomous variable, opposing dispersing animals with philopatric ones. However, dispersal may be better described as a continuum, because the distinction between dispersers and philopatric individuals can be difficult to assess using some arbitrarily defined criteria. Some studies have used both a binomial variable (disperse vs. philopatric) and the dispersal distance to assess fine scale patterns of natal dispersal (Gaillard et al. 2008 in roe deer Capreolus capreolus; Selonen & Hanski 2010 in Siberian flying squirrels Pteromys volans L.; Long et al. 2005 in white-tailed deer Odocoileus virginianus). However, these studies did not account for among-individual differences in ranging behaviour prior to dispersal. In the present study, we analyse dispersal both as a Bernoulli (binomial) and a Gaussian (continuous) process. For this latter approach, we used an individual-based standardised dispersal distance (IBSDD) as a metric. This was defined as the raw dispersal distance weighted by the spatial extent of an individual's pre-dispersal range.

The distribution of natal dispersal distances shapes the speed of population spread, and potentially has thereby a strong impact on population persistence (Sutherland et al. 2000; Bowler & Benton 2005; Ronce 2007). However, despite this theoretical interest, empirical measures of interindividual variability in dispersal distances and identification of the factors affecting it are still scarce (Lowe 2010). Our empirical understanding of the causes and consequences of variation in dispersal distances is limited to some studies that assessed the impact of different factors on dispersal distance (climatic conditions in the Arctic tern Sterna paradisaea, Moller, Flensted-Jensen & Mardal 2006; wing length in female house sparrows Passer domesticus, Skjelseth et al. 2007; personality in the invasive mosquitofish Gambusia affinis, Cote et al. 2010). Although several theoretical studies have identified body condition as a potentially important predictor of dispersal distance and settlement success (Stamps, Krishnan & Reid 2005; Stamps 2006), we still lack empirical studies that have tested for condition dependence of dispersal rate and distance (Clobert et al. 2009). Heavy dominant individuals were found to disperse more than smaller ones when dispersal is energetically costly (Ims & Hjermann 2001; Bowler & Benton 2005), as reported in ground squirrels Spermophilus beldingi (Holekamp & Sherman 1989) and owls Bubo bubo (Delgado et al. 2010). In contrast, no effect of body condition on dispersal propensity was found in red deer Cervus elaphus (Loe et al. 2010) or in Siberian ground squirrels Pteromys volans (Selonen & Hanski 2010). Furthermore, at the interspecific level, Sutherland et al. (2000) demonstrated that median and maximum natal dispersal distances are correlated with species body mass in birds and mammals. Finally, in terms of environmentally driven condition-dependent dispersal, some studies have assessed the impact of population density and environment on the intensity of natal dispersal. For example, landscape fragmentation led to greater dispersal distance in white-tailed deer (Long et al. 2005) and in nuthatches Sitta europeae (Matthysen, Adriaensen & Dhondt 1995). We aimed to investigate whether natal dispersal is driven by phenotypic and/or environmental condition dependence in a large herbivore, focusing on (i) natal dispersal propensity and (ii) using an original IBSDD that accounts for among-individual variation in natal home range size.

We investigated the proximal causes underlying natal dispersal in a roe deer population living in a spatially heterogeneous agricultural landscape. Roe deer are medium-sized, slightly dimorphic and weakly polygynous mammalian herbivores that are widely spread across Europe and have markedly expanded their range since the 1960s (Andersen, Duncan & Linnell 1998). Individual body mass is quite stable over the lifetime (Hewison et al. 2011) and provides a good proxy of individual quality (Toigo et al. 2006), with higher probability for heavy adult females to reach old age (Gaillard et al. 2000) and higher reproductive success among heavy males (Vanpé et al. 2010). To investigate condition-dependent dispersal, we quantified the effects of body mass and of the degree of habitat heterogeneity on dispersal initiation, propensity and our continuous dispersal distance metric. As dispersal is a costly process (Ronce 2007), heavier animals should be better able to cope with dispersal costs (Bowler & Benton 2005). From this, heavier than average roe deer should be more likely to disperse and should disperse earlier and further. Landscape heterogeneity is known to affect several aspects of roe deer ecology. In particular, deer in more open habitats have larger home ranges (Cargnelutti et al. 2002), higher diet quality (Abbas et al. 2011) and are thus heavier as adults (Hewison et al. 2009). Given this, we also expected landscape structure to exert an influence on dispersal propensity and distance in the studied population, with individuals from the more open areas more likely to disperse and to travel further. Based on previously published findings (Coulon et al. 2006; Gaillard et al. 2008), we expected no differences in dispersal patterns to occur between sexes.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Study area

The study was conducted in a hilly (260–380 m asl) and heterogeneous agricultural landscape in the Comminges region of South-West France (N 43°13′, E 0°52′) covering around 10 000 ha. The climate is oceanic, with an average annual temperature of 11–12 °C and 800 mm precipitation, mainly as rain. The area is a mixed landscape of open fields and small woodland patches (average size of 3 ha) dominated by oak Quercus spp., with two larger forest patches of 672 and 460 ha (see Hewison et al. 2009 for further details). We identified two sectors of contrasting landscape structure based on woodland extent (Fig. 1). The first sector included two forest blocks (sector 1 ‘closed’: with 100% woodland), and the second sector was composed of a more open landscape of fragmented woodland (sector 2 ‘open’: with 23·7% woodland, 36·1% meadows, 32·1% cultivated fields and 4·3% hedgerows).

image

Figure 1. Map of the study site with the main habitat types and the limits of the two sectors.

Download figure to PowerPoint

Human presence is high, with small villages and farms distributed along the extensive road network. The roe deer population is hunted by drive hunts with dogs from September to February, but also in summer by stalking (June–August). Deer density was estimated using radiotracking and direct observation (Hewison et al. 2007) to average 9·3 deer per 100 ha (±1·32, min = 6·6, max = 10·9) in the fragmented landscape, but to be 2–3 times higher in the two forest blocks (unpublished data).

Capture, handling and monitoring

Roe deer were caught from 2002 to 2011 during winter (from 16 November to 27 March) using large-scale drives of between 30 and 100 beaters and up to 4 km of long nets positioned at one of 11 capture sites. For each captured animal, we recorded its body mass to the nearest 0·1 kg, its sex, and we attributed an age class before fitting it with a collar and releasing it on site. Juveniles (<1 year old) are distinguishable from older deer by the presence of a tri-cuspid third premolar milk tooth (Ratcliffe & Mayle 1992). During the eight winters of sampling (2002–2010), 88 juveniles and 132 adults were captured and equipped with a Lotek 3300 GPS or a Lotek Small WildCell GSM collar. Collars were programmed to obtain a fix of the roe deer's location every 4 h in 2002–2004 (first two winters) or every 6 h (following winters) over approximately 11 months. We performed differential correction to improve fix accuracy (Adrados et al. 2003). From 2002 to 2010, GPS data from 79 juveniles and 128 adults were recovered, although body mass was not available for one of these juveniles. The collars of the nine remaining juveniles were never recovered because of Very High Frequency transmitter failure (= 2), drop-off failure (= 4) or loss of contact during the transient phase of dispersal (= 3). In the latter case, a slight underestimation of dispersal rate at the population level might result.

Natal dispersal of roe deer is a rather stereotyped process (Strandgaard 1972). Only juvenile roe deer disperse (Pettorelli et al. 1993) and do so in spring, when the mother–juvenile bond breaks down before the mother gives birth to her next litter (between late April and mid-June at this latitude, Linnell, Wahlstrom & Gaillard 1998). The timing of dispersal is thus highly synchronised (Wahlstrom & Liberg 1995; Linnell, Wahlstrom & Gaillard 1998; Van Moorter et al. 2008) so that the dispersal status of juveniles can be reliably determined by the end of July. Of the 79 juveniles captured and for which data were recovered, only those 60 individuals that were monitored from their capture in winter to at least the following August 1st and had more than 50 locations during the summer period were included in the analyses. Dispersal outcomes could not be determined for the remaining 19 individuals because of early death (= 10) or loss of contact (= 9) prior to autumn. Note that as the juvenile stage only spans 1 year, all individuals only appeared once in the data set (i.e. no repeated measurements). As capture and handling might induce transient modification of roe deer behaviour, the location data for the first week after release were excluded from the analysis following Morellet et al. (2009)'s recommendations.

Statistical analysis

Defining philopatry vs. natal dispersal

Natal dispersal was defined as permanent emigration from the natal range (pre-dispersal home range) to a distinct adult range (post-dispersal home range), such that pre-dispersal locations did not overlap post-dispersal ones (Kenward et al. 2002). We defined two periods describing the juvenile range (1 January to 31 March, mean ± SD: 219 ± 86 locations/individual), which consistently corresponded to the natal range, and the adult range (1 July to 30 September, mean ± SD: 277 ± 90 locations/individual), which corresponded to the natal range (or part of it) for philopatric roe deer only, to describe pre- and post-dispersal home ranges, respectively. Based on previous observations in the studied population and the available literature, all dispersal events of roe deer occur between these two periods. Home ranges were estimated using the 90% fixed kernel method with an ad hoc method used for the smoothing parameter (Worton 1989; Borger et al. 2006). The annual range was also calculated by combining juvenile and adult ranges for each animal (average annual adult home range size = 154·9 ha, corresponding to a 1·40-km-diameter circle). We used the range stability index proposed by Roshier & Reid (2003) to discriminate dispersers from philopatric animals:

  • display math

Range stability indices were calculated for both juvenile and adult ranges. An average annual index was also calculated as:

  • display math

Philopatric animals use annual ranges that encompass both juvenile and adult home ranges (Fig. 2b), so that the value of the average annual indices is close to one. In contrast, when the juvenile and adult home ranges differ markedly (Fig. 2a), that is, in dispersing animals, values of the annual indices are low (Roshier & Reid 2003). Individuals were considered to have dispersed when their average annual index was equal to 0·5, corresponding to an absence of overlap between the juvenile and adult home ranges, or philopatric otherwise.

image

Figure 2. Typical space use of a disperser (a) and of a philopatric (b) juvenile roe deer. Adult locations and home ranges are in dark grey, and juvenile locations and home ranges are in light grey. The two large white dots indicate the centre of gravity of the corresponding home range.

Download figure to PowerPoint

Dispersal initiation

Dispersal initiation was defined as the last date a deer was located inside its juvenile range without ever returning. For this analysis only, we included juvenile animals captured during the winter 2010–2011 and for which dispersal fates were already known by the spring of 2011 (= 5), but for which the collar had not yet been recovered by the time of data analysis.

Individual-based standardised dispersal distance

Natal dispersal can be categorised using a Bernoulli process (philopatry vs. dispersal) reflecting the decision whether or not to leave the natal area, or quantified as a Gaussian distribution of the Euclidean distances between pre- and post-dispersal home ranges reflecting the length of the transience phase. However, to distinguish dispersers from philopatric individuals most often requires the choice of an arbitrary threshold value, which can lead to a mis-classification of some individuals. Furthermore, dispersal is a movement process across a spatial continuum, where some individuals may move long distances, while others just shift their home range a few hundred metres. Hence, because categorising individuals in two types also leads to information loss (Jin, Breitbart & Muoh 2009), we used an IBSDD to assess condition-dependent dispersal in roe deer.

We constructed a continuous metric based on dispersal distance weighted by interindividual variation in ranging behaviour. We calculated the Euclidean (straight line) distance between juvenile (pre-dispersal home range) and adult (post-dispersal home range) centres of gravity for all individuals. As the Euclidean distance between the two home ranges was positively correlated with the size of the individual's home range, we generated an IBSDD metric calculated as the residual value from the linear regression between Euclidean distance and juvenile (i.e. pre-dispersal) home range size (McCullagh & Nelder 1989), hereafter defined as our IBSDD. We used the juvenile home range size because it was usually larger than the adult range (Borger et al. 2006) and because post-dispersal home ranges of dispersers may not yet be stable.

Assessing condition-dependent dispersal

To explore condition-dependent dispersal, we assessed the influence of body mass and the degree of habitat heterogeneity on dispersal propensities. Juvenile roe deer usually are still growing during their first winter (Hewison et al. 2002), but a study conducted in the same area did not find any detectable variation in winter body mass of juvenile roe deer in relation to Julian date (Hewison et al. 2009), so date of capture was not included in the models. First, we used multiple linear regression to assess whether temporal variation in dispersal initiation (= 28) varied in relation to sex, landscape sector and body mass. Secondly, we used logistic regression to determine whether natal dispersal status, a binary response variable (disperser or philopatric, = 55), varied between sexes, landscape sectors or in relation to body mass. Finally, condition-dependent dispersal was explored using a mixed linear model to investigate how variation in the IBSDD (dependent variable) was related to sex, landscape and body mass (explanatory variables); the factor year was included as a random factor to control for the potential nonindependence of dispersal distance within a year.

Five groups of mixed linear models were identified, corresponding to specific hypotheses (Table 1). The first group of three models (‘Envir’) tested the hypothesis of strict environmental condition-dependent dispersal, including landscape sector as a factor (with or without an interaction with sex). The second group of four models (‘Pheno’) tested the hypothesis of strict phenotypic condition-dependent dispersal, with body mass as a factor (with or without an interaction with sex). For this second group, body mass was considered as either a simple linear effect or a threshold effect, that is, with no effect until the threshold value and with a linear effect above the threshold value. The threshold value was chosen by selecting the best threshold model from a range of values varying from the minimum to the maximum observed body mass in our sample. The third group of eleven models (‘Both’) tested the hypothesis of simultaneous phenotypic and environmental condition-dependent dispersal, including the two-way interactions in the candidate models. One model (‘Sex-biased’) allowed us to test the hypothesis of sex-biased dispersal. Finally, a constant model (‘Null’) corresponded to the hypothesis of no condition-dependent dispersal.

Table 1. Candidate mixed linear models for explaining variation in dispersal initiation date (= 28), dispersal propensity (binary variable) (= 55) and individual-based standardised dispersal distance (Gaussian variable) (= 55) of juvenile roe deer and their corresponding hypotheses
HypothesisModelsInitiationRateDistance
  1. The models include body mass as a simple linear effect (BM) or with a threshold effect at 14 kg (BMth_14 kg) as a covariate, degree of habitat heterogeneity (Sector), sex (Sex) and their two-way interactions (Sector/BM, BM/Sex and Sector/Sex) as fixed factors, and year as a random factor.

NullConstant×××
Sex-biasedSex×××
PhenoBM×××
PhenoBMth_14 kg×
PhenoBM + Sex×××
PhenoBM + Sex + BM/Sex××
EnvirSector×××
EnvirSector + Sex×××
EnvirSector + Sex + Sector/Sex××
BothBMth_14 kg + Sector×
BothBM + Sector×××
BothBM + Sector + Sex×××
BothBM + Sector + Sector/BM××
BothBM + Sector + Sex + Sector/Sex××
BothBM + Sector + Sex + BM/Sex××
BothBM + Sector + Sex + Sector/BM××
BothBM + Sector + Sex + Sector/Sex + BM/Sex××
BothBM + Sector + Sex + Sector/Sex + Sector/BM××
BothBM + Sector + Sex + BM/Sex + Sector/BM××
BothBM + Sector + Sex + BM/Sex + Sector/BM + Sector/Sex××

First, a set of eight candidate models for dispersal initiation, grouped in five categories (‘Pheno’, ‘Envir’, ‘Both’, ‘Sex-biased’ and ‘Null’) based on the hypotheses described above, were compared (Table 1). Then, in the analysis of dispersal propensity (dispersal as a binary variable), a set of 18 candidate models were compared (Table 1). Finally, in the analysis of the IBSDD (dispersal as a quantitative variable), a set of 20 candidate models, grouped in the same five categories, were compared (Table 1). We retained the model with the lowest AICc value (i.e. AIC corrected for small sample size), reflecting the best compromise between precision and complexity of the model (Burnham & Anderson 1998). According to the rule of parsimony, when the AICc of two competing models differed by <2, we retained the simplest model. We also calculated AICc weights as a measure of the likelihood that a given model was the best among the set of fitted models. All statistical analyses were performed with R software version 2.12.1 (R Development Core Team 2008).

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Dispersal propensity

No dispersal event was recorded among adult deer. Of the 60 juveniles for which dispersal fate was known, 19 dispersed and 37 were philopatric, while the four others left their juvenile area, travelling an average of 12·6 (±8·6) km, but then returned sometime later (average 7·5 ± 8·1 weeks), during summer (i.e. pseudo-dispersal events). Although these four individuals left their juvenile area during the natal dispersal period, they were not considered as dispersers because their first breeding event likely took place within their natal home range; as a consequence, these four individuals were excluded from subsequent analyses. The overall population dispersal rate was therefore estimated at 0·34 (= 56). In the most parsimonious mixed generalised linear model (AICc weight = 13%; ∆AICc to best model = 1·42; ∆AICc to constant model = 6·76), dispersal propensity was affected by body mass only, but the model with the highest support (AICc weight = 27%; ∆AICc to best model = 0; ∆AICc to constant model = 8·18) included the effect of both body mass and landscape sector on dispersal propensity, indicating some support for effects of both phenotypic attributes and environmental context. The probability of dispersing increased markedly with body mass (slope ± SE: 0·60 ± 0·23) and was higher in the open landscape sector (= 40; dispersal rate ± SE = 0·42 ± 0·078) than in the forest sector (n = 16; dispersal rate ± SE = 0·12 ± 0·083). As expected, no sex bias in dispersal propensity was observed in the studied population (AICc weight = 0%; ∆AICc to best model = 8·73; ∆AICc to constant model = 1·97), with a dispersal rate of 0·29 ± 0·099 for males (= 21) and of 0·37 ± 0·082 for females (= 35).

Dispersal initiation

Dispersal initiations (= 28), including pseudo-dispersal events, were synchronised in spring, with a peak at the end of April. The earliest departure occurred on the 30th of March and the latest on the 27th of May. The most parsimonious model included the effects of body mass (AICc = 0), indicating that departure date was not influenced by sex or landscape sector, but was negatively related to body mass (Table 2). That is, there was clear evidence that heavier animals left their pre-dispersal range earlier than lighter individuals (estimated initiation date: 28th April for an individual weighing 16 kg vs. 20th April for an individual weighing 18 kg).

Table 2. Performance of the nine candidate mixed linear models explaining variation in dispersal initiation date for juvenile roe deer (= 28) using the Akaike's Information Criterion corrected for small sample size (AICc)
ModelsAICc∆AICcAICcWt K
  1. The full model includes an effect of body mass (BM), the degree of habitat heterogeneity (Sector), and sex (Sex) and year as a random factor. K refers to the number of model parameters.

BM226·1200·654
Sector + BM229·052·930·155
BM + Sex229·163·040·145
Sector + BM + Sex232·396·270·036
Constant233·066·940·023
Sex235·279·150·014
Sector235·629·50·014
Sector + Sex238·2012·0705

Individual-based standardised dispersal distance

The pattern of dispersal distance followed a classic leptokurtic distribution, with a mean linear distance between natal and post-dispersal home ranges of 4·68 ± 8·44 km considering all juveniles and of 12·85 ± 10·52 km when considering dispersers only (Fig. 3).

image

Figure 3. Distribution of the Euclidean distances between juvenile and adult range centres of gravity for all juveniles (= 56).

Download figure to PowerPoint

Of the candidate models describing condition-dependent dispersal using our IBSDD, two had similar AIC values (AICc < 2) and the same number of parameters (Table 3). These models both indicated that the IBSDD between juvenile and adult home ranges varied in relation to body mass and landscape sector in an additive way. The best model (AICc weight = 29%) assumed that the IBSDD increased linearly above a certain body mass threshold, whereas the second best model (AICc weight = 22%) assumed a simple linear relationship between these variables (Table 3). Hence, the models with the greatest support indicated both phenotypic and environmental condition dependence in dispersal distance of juvenile roe deer. In particular, first, IBSDD was greater in the open landscape sector compared to the forest sector (intercept ± SE: −0·5 ± 0·5 in the open sector and −1·75 ± 0·46 in the forest sector), and secondly, the IBSDD increased linearly with body mass from a threshold value of 14 kg (slope ± SE: 0·36 ± 0·15). Finally, the slope of the relationship between IBSDD and body mass did not vary between landscape sectors (no sector-body mass interaction term in the best models), indicating that roe deer dispersed further for a given body mass in the open landscape sector compared to animals in pure forest habitat (Fig. 4). Our analysis also indicated, in line with our prediction, that dispersal distance was not sex-biased in our population (Table 3), with similar mean distances between juvenile and adult home ranges for males (4·05 ± 6·74 km) and females (4·99 ± 9·52 km). Indeed, the models incorporating a difference between the sexes in IBSDD all had higher AICc values (AICc > 4).

image

Figure 4. Individual-based standardised dispersal distance between the centres of gravity of the juvenile and adult home ranges for each juvenile, corrected for variation in pre-dispersal home range size, in relation to its body mass (= 55). Grey lines represent the relationship predicted by the selected linear model (BMth_14 kg + Sector;  = 0·25) (light grey for the closed sector, dark grey for the open sector).

Download figure to PowerPoint

Table 3. Performance of the 20 candidate linear models for explaining variation in individual-based standardised dispersal distance of juvenile roe deer (= 55), including body mass, either as a simple linear effect (BM) or with a threshold at 14 kg (BMth_14 kg) as a covariate; degree of habitat heterogeneity (Sector), sex (Sex) and the two-way interactions between body mass and sex, sector and body mass, and sector and sex as fixed factors; and year as a random factor
HypothesisModelsAICc∆AICcAICcWt K
BothBMth_14 kg + Sector215·5500·295
BothBM + Sector216·140·590·225
BothBM + Sector + Sector/BM217·732·180·16
EnvirSector218·653·100·064
BothBM + Sector + Sex218·663·110·066
PhenoBM219·053·500·054
PhenoBMth_14 kg219·063·510·054
BothBM + Sector + Sex + Sector/Sex220·074·520·037
BothBM + Sector + Sex + Sector/BM220·374·820·037
EnvirSector + Sex220·665·110·025
PhenoBM + Sex220·845·290·025
BothBM + Sector + Sex + BM/Sex221·205·650·017
BothBM + Sector + Sex + Sector/Sex + BM/Sex222·036·480·018
EnvirSector + Sex + Sector/Sex222·056·500·016
BothBM + Sector + Sex + Sector/Sex + Sector/BM222·346·790·018
BothBM + Sector + Sex + BM/Sex + Sector/BM222·987·430·018
PhenoBM + Sex + BM/Sex223·337·780·016
BothBM + Sector + Sex + BM/Sex + Sector/BM + Sector/Sex224·488·9309
NullConstant226·4710·9203
Sex-biasedSex228·7213·1704

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Natal dispersal was observed in one-third of the juvenile roe deer and, in line with previous reports for this species, was not sex-biased. However, as expected, natal dispersal outcomes were dependent on both the phenotypic attributes of individuals (body mass) and environmental factors (habitat openness). Heavier individuals had a higher probability of dispersing and our IBSDD metric increased linearly with increasing body mass, with some support for a body mass threshold of 14 kg under which no dispersal occurred. The degree of habitat heterogeneity in the natal home range also influenced dispersal behaviour: individuals born in more open areas dispersed more and travelled further than individuals from closed habitats. Our study provides a rare example of multifactorial condition-dependent dispersal in a large herbivore and highlights the complexity of dispersal mechanisms, with several conditions or cues operating simultaneously to determine the dispersal decisions of individual animals (Clobert et al. 2009).

Dispersal rate and distance

Although dispersal rate is highly variable among roe deer populations, the rate of 33·9% that we observed in our study conforms to values previously reported (Wahlstrom & Liberg 1995; Gaillard et al. 2008). Compared to other large herbivores, this rate of dispersal is quite low for males, because 52% of juvenile white-tailed deer (Nixon et al. 2007) and 68·8% of male red deer (Loe et al. 2010) were reported to disperse. There are fewer studies of dispersal distances for large herbivores in general, although the estimate of 4·7 ± 8·4 km for our study site lies within the reported range for roe deer (1·1 ± 0·1 to 7·6 ± 3·0 km: Wahlstrom & Liberg 1995; Gaillard et al. 2008). The distribution of IBSDDs that we observed followed a classic leptokurtic or ‘tick-tailed’ distribution (Paradis et al. 1998), with relatively few individuals moving long distances and most moving shorter distances (Johnson & Gaines 1990; Bowler & Benton 2005; Ronce 2007), indicating a strong possibility for condition-dependent dispersal in our data set (see below). Indeed, leptokurtosis is thought to be driven by intrapopulation variation among individuals in dispersal tactic (Fraser et al. 2001) and by variation in habitat structure (Morales & Ellner 2002). This intrapopulation variation can be attributed to different ultimate causes, which may result in different optimal dispersal distances. For example, in white-tailed deer, longer dispersal distances seem necessary for avoiding inbreeding rather than mate competition (Long et al. 2008).

Dispersal initiation

The date of dispersal initiation was highly synchronised among individuals, as previously reported in roe deer (Linnell, Wahlstrom & Gaillard 1998). There was clear evidence that heavier roe deer disperse earlier, a finding consistent with studies on male Belding's ground squirrels (Nunes & Holekamp 1996) and red foxes Vulpes vulpes (Gosselink et al. 2010). Early dispersal may be advantageous because early-dispersing individuals may have more time to locate, and/or may arrive first on, a vacant, high-quality home range. However, this relationship is unlikely to involve differences in maturation among juveniles. Indeed, births are highly synchronised in roe deer (Linnell, Wahlstrom & Gaillard 1998), with 80% occurring within 3 weeks, so that the impact of birth date on juvenile body mass at dispersal (10–12 months of age) is likely weak. While birth date influences both early growth and survival in roe deer (F. Plard et al., unpublished data), fawns that survive to their first winter seem able to compensate for a bad start because no relationship occurred between early growth and mass at 8–10 months of age (Gaillard, Delorme & Jullien 1993) in a population not markedly limited by food resources, as is the case in our study population.

Interestingly, we observed ‘pseudo-dispersal’ events in four males who left their natal range during the dispersal period, travelling a considerable distance (between 5·2 and 25 km), before returning several weeks later during summer to their natal range. Speculatively, we suggest that this potentially costly behaviour occurs when male juveniles are unable to locate a suitable vacant home range and so are forced to return to their natal range because of antagonistic social interactions with territorial adult males (Wahlstrom 1994). In roe deer, territoriality governs male–male interactions but not female–female or male–female ones (Hewison, Vincent & Reby 1998), and, indeed, we did not observe any similar pseudo-dispersal events among females in our population.

Sex-biased dispersal

As expected, on average, male and female juvenile roe deer initiated dispersal events at approximately the same time, dispersed in similar proportions and travelled similar distances during the transience phase. This result is consistent with both genetic studies (in this same population: Coulon et al. 2006; see also Bonnot et al. 2011), but also with studies based on direct observations (Wahlstrom & Liberg 1995; Gaillard et al. 2008). The low sexual size dimorphism, the mating tactic of resource defence and the low level of polygyny of this species (Vanpé et al. 2008) may explain this absence of sex bias in roe deer dispersal (Gaillard et al. 2008). However, although dispersal outcomes are clearly similar in male and female roe deer, it is likely that the proximate mechanisms underlying dispersal decisions differ between the sexes. Indeed, Wahlstrom (1994) reported that the number of antagonistic interactions experienced by male yearling roe deer was positively correlated with antler size, and he suggested that these social interactions were the proximate cause of natal dispersal in juvenile males. As there is a strong allometric relationship linking antler length with body mass in this species (Vanpé et al. 2007), and in other deer (Plard, Bonenfant & Gaillard 2011), bigger juvenile males are predicted to suffer more male–male aggression, leading to the observed pattern of mass-dependent dispersal. However, this mechanism clearly does not hold for females who are not territorial (Hewison, Vincent & Reby 1998), hence, the cue for initiating dispersal of heavier juveniles (see below) likely differs between sexes.

Phenotypic condition-dependent dispersal: the role of body mass

Condition-dependent dispersal tactics should perform better than unconditional fixed tactics because they allow individuals to respond to variation in the costs and benefits of dispersal over the short term (Bowler & Benton 2005). We found that individual body mass played a crucial role in determining dispersal rate and distance, with heavier juveniles dispersing more frequently and travelling further and with some support for a body mass threshold under which roe deer juveniles cannot sustain the energetic costs of dispersal. Indeed, none of the eight roe deer weighing less than 14 kg dispersed. Similar results were found in Belding's ground squirrels (Holekamp 1986). However, this pattern contrasts with a study on red deer that reported no relationship between body mass and male dispersal propensity (Loe et al. 2010). Red deer and roe deer markedly differ in many life-history tactics. The red deer is a highly dimorphic and polygynous species (Clutton-Brock, Guinness & Albon 1982) and a grazer (Hofmann 1989); females are close to the capital end of the continuum of energy allocation to reproduction. In contrast, the roe deer is a weakly dimorphic species with a low level of polygyny (Vanpé et al. 2008) and a browser (Hofmann 1989); females are close to the income end of the continuum of energy allocation. It is thus not surprising that the pattern of natal dispersal also differs markedly between these two related species.

Dispersal is known to be a risky behaviour (Ronce 2007), and costs increase with increasing dispersal distance (Rousset & Gandon 2002; see Johnson et al. 2009 for a study case on American martens). Our study suggests that (i) there is a threshold of 14 kg minimum mass for an individual to be able to cope with the costs of dispersal, and (ii) the observed relationship between dispersal distance and body mass similarly suggests that only the heaviest juveniles are able to offset the costs of long-distance dispersal. The higher rate of movement necessary for dispersal could imply increased energetic expenditure. Indeed, we found evidence that dispersers moved greater distances per time unit during the dispersal event compared to the distances travelled over the same period of time by nondispersing individuals during their normal activities within their home range (= 23, = 34, = 0·051). Evidently, for dispersal to evolve, it must also generate some benefits that, over the long term, compensate these costs. For example, more female dispersers attained dominant status than their philopatric counterparts in red fox V. vulpes (Soulsbury et al. 2008). While we are unable to conclude on the nature of the benefits obtained by dispersing roe deer in our study, we speculate that inbreeding avoidance is an important consideration in view of the lifelong sedentary nature of adult roe deer and the social system based on small family units (Hewison, Vincent & Reby 1998), which likely leads to substantial opportunity for inbred matings. Furthermore, Vanpé et al. (2009b) showed that roe deer fawns born from closely related parents survived less well over their first summer than those with unrelated parents. In general, the dispersal distance necessary to avoid kin competition or inbreeding is much longer and requires greater movement ability than that required to escape competition with nonrelatives (Ronce 2007; Long et al. 2008 in white-tailed deer).

In our population, not all heavy animals dispersed, suggesting that there was a choice available to disperse or not and that several factors were involved in that choice. Competitive ability may influence whether an individual disperses or not (Ims & Hjermann 2001). In this context, two contrasting hypotheses were proposed by Bowler & Benton (2005): first, heavier animals are more competitive than lighter ones, hence lighter animals are forced to disperse to avoid competition with heavy, more competitive individuals; alternatively, larger individuals may be more prone to disperse if they are more capable of immigrating into a new competitive patch successfully or if dispersal requires a certain amount of energy reserves. Roe deer seems to fit better with this latter scenario. For example, we have previously shown that only particularly heavy roe deer bucks are able to establish their first territory at 2 years of age (Vanpé et al. 2009a), and we speculatively suggest that high body mass may also be important in primiparous females for the acquisition of a high-quality fawning range. Moreover, as inbreeding has a cost in terms of fawn survival in roe deer (Vanpé et al. 2009b), dispersal could allow heavier individuals to increase their offspring survival.

Environmental condition-dependent dispersal: the impact of habitat heterogeneity

In our study, individuals inhabiting more open habitats dispersed more frequently and further than individuals living in more forested habitats. In open habitats, individual phenotypic quality, as indexed by body mass, is generally higher (Hewison et al. 2009); however, individuals in more open habitats dispersed more irrespective of body mass. This suggests that the degree of habitat heterogeneity could markedly impact dispersal propensity in this large herbivore. This pattern of habitat-dependent dispersal distance may be a general feature of heterogeneous landscapes, as mean dispersal distance of nuthatches was several times greater in a highly heterogeneous landscape compared to more densely forested landscapes (Matthysen, Adriaensen & Dhondt 1995), while dispersal distances of juvenile male white-tailed deer were greater in habitats with less forest cover (Long et al. 2005).

Implications of condition-dependent dispersal

Our results provide compelling empirical evidence for condition-dependent dispersal in a large herbivore, indicating that high phenotypic quality is a critical prerequisite to disperse successfully. Dispersing individuals are thus not a random subset of the population. We showed that dispersers are heavier than philopatric individuals, suggesting that immigrants to a given area may be more competitive than the philopatric individuals already present. Condition-dependent dispersal can have profound consequences for population and metapopulation dynamics (Clobert et al. 2001; Bonte & de la Pena 2009). For example, a change in average body condition can alter connectivity between populations and consequent gene flow (Bohonak 1999). In a simulation study, Bonte & de la Pena (2009) suggested that body condition-dependent dispersal tactics affect population dynamics and induce evolutionary rescue mechanisms in spatially structured populations. In particular, when dispersal is modelled as a condition-dependent tactic, local metapopulation extinction rates are always close to zero (Bonte & de la Pena 2009). A better understanding of the mechanisms involved in natal dispersal, such as condition dependence, will thus help us to understand the evolution of this behaviour, as well as providing a basis for better prediction of metapopulation functioning.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We thank two anonymous referees and the associate editor for their constructive comments on a previous version of this paper. We also would like to thank the local hunting associations, the Fédération Départementale des Chasseurs de la Haute Garonne, as well as numerous co-workers and volunteers for their assistance and, in particular, J.M. Angibault, J. Merlet, D. Picot, J.L. Rames, J. Joachim, H. Verheyden and N. Cebe.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Abbas, F., Morellet, N., Hewison, A.J.M., Merlet, J., Cargnelutti, B., Lourtet, B., Angibault, J.-M., Daufresne, T., Aulagnier, S. & Verheyden, H. (2011) Landscape fragmentation generates spatial variation of diet composition and quality in a generalist herbivore. Oecologia, 167, 401411.
  • Adrados, C., Verheyden-Tixier, H., Cargnelutti, B., Pepin, D. & Janeau, G. (2003) GPS approach to study fine-scale site use by wild red deer during active and inactive behaviors. Wildlife Society Bulletin, 31, 544552.
  • Andersen, R., Duncan, P. & Linnell, J.D.C. (1998) The European Roe Deer: The Biology of Success. Scandinavian University Press, Oslo.
  • Bohonak, A.J. (1999) Dispersal, gene flow, and population structure. Quarterly Review of Biology, 74, 2145.
  • Bonnot, N., Gaillard, J.M., Coulon, A., Galan, M., Cosson, J.F., Delorme, D., Klein, F. & Hewison, A.J.M. (2011) No difference between the sexes in fine-scale spatial genetic structure of roe deer. PLoS ONE, 5, e14436.
  • Bonte, D. & de la Pena, E. (2009) Evolution of body condition-dependent dispersal in metapopulations. Journal of Evolutionary Biology, 22, 12421251.
  • Borger, L., Franconi, N., Ferretti, F., Meschi, F., De Michele, G., Gantz, A. & Coulson, T. (2006) An integrated approach to identify spatiotemporal and individual-level determinants of animal home range size. American Naturalist, 168, 471485.
  • Bowler, D.E. & Benton, T.G. (2005) Causes and consequences of animal dispersal strategies: relating individual behaviour to spatial dynamics. Biological Reviews, 80, 205225.
  • Burnham, K.P. & Anderson, D.R. (1998) Model Selection and Inference, 2nd edn. Springer-Verlag, New York.
  • Cargnelutti, B., Reby, D., Desneux, L., Angibault, J.M., Joachim, J. & Hewison, A.J.M. (2002) Space use by roe deer in a fragmented landscape some preliminary results. Revue d'Ecologie (La Terre et la Vie), 57, 2937.
  • Clobert, J., Danchin, E., Dhondt, A.A. & Nichos, J.D. (2001) Dispersal. Oxford University Press, New York.
  • Clobert, J., Le Galliard, J.F., Cote, J., Meylan, S. & Massot, M. (2009) Informed dispersal, heterogeneity in animal dispersal syndromes and the dynamics of spatially structured populations. Ecology Letters, 12, 197209.
  • Clutton-Brock, T.H., Guinness, F.E. & Albon, S.D. (1982) The Red Deer. Behavior and Ecology of Two Sexes. University of Chicago Press, Chicago.
  • Cote, J., Fogarty, S., Weinersmith, K., Brodin, T. & Sih, A. (2010) Personality traits and dispersal tendency in the invasive mosquitofish (Gambusia affinis). Proceedings of the Royal Society B: Biological Sciences, 278, 16701678.
  • Coulon, A., Cosson, J.F., Morellet, N., Angibault, J.M., Cargnelutti, B., Galan, M., Aulagnier, S. & Hewison, A.J.M. (2006) Dispersal is not female biased in a resource-defence mating ungulate, the European roe deer. Proceedings of the Royal Society B: Biological Sciences, 273, 341348.
  • Delgado, M.D., Penteriani, V., Revilla, E. & Nams, V.O. (2010) The effect of phenotypic traits and external cues on natal dispersal movements. Journal of Animal Ecology, 79, 620632.
  • Dufty, A.M. & Belthoff, J.R. (2001) Proximate mechanisms of natal dispersal: the role of body condition and hormones. Dispersal (eds J. Clobert, E. Danchin, A.A. Dhondt & J.D. Nichos), pp. 217229. Oxford University Press, New York.
  • Fraser, D.F., Gilliam, J.F., Daley, M.J., Le, A.N. & Skalski, G.T. (2001) Explaining leptokurtic movement distributions: intrapopulation variation in boldness and exploration. American Naturalist, 158, 124135.
  • Gaillard, J.M., Delorme, D. & Jullien, J.M. (1993) Early growth and body-weight at the onset of the winter in roe deer (Capreolus capreolus). Mammalia, 57, 359366 (In French).
  • Gaillard, J.M., Festa-Bianchet, M., Delorme, D. & Jorgenson, J. (2000) Body mass and individual fitness in female ungulates: bigger is not always better. Proceedings of the Royal Society B: Biological Sciences, 267, 471477.
  • Gaillard, J.M., Hewison, A.J.M., Kjellander, P., Pettorelli, N., Bonenfant, C., Van Moorter, B., Liberg, O., Andren, H., Van Laere, G., Klein, F., Angibault, J.M., Coulon, A. & Vanpé, C. (2008) Population density and sex do not influence fine-scale natal dispersal in roe deer. Proceedings of the Royal Society B: Biological Sciences, 275, 20252030.
  • Gibbs, M., Saastamoinen, M., Coulon, A. & Stevens, V.M. (2009) Organisms on the move: ecology and evolution of dispersal. Biology letters, 6, 146148.
  • Gosselink, T.E., Piccolo, K.A., van Deelen, T.R., Warner, R.E. & Mankin, P.C. (2010) Natal dispersal and philopatry of red foxes in urban and agricultural areas of Illinois. Journal of Wildlife Management, 74, 12041217.
  • Greenwood, P.J. (1980) Mating systems, philopatry and dispersal in birds and mammals. Animal Behaviour, 28, 11401162.
  • Hewison, A.J.M., Vincent, J.P. & Reby, D. (1998) Social organisation of European roe deer. The European Roe Deer: The Biology of Success (eds R. Andersen, P. Duncan & J. Linnell), pp. 189219. Scandinavian University Press, Oslo.
  • Hewison, A.J.M., Gaillard, J.M., Angibault, J.M., Van Laere, G. & Vincent, J.P. (2002) The influence of density on post-weaning growth in roe deer Capreolus capreolus fawns. Journal of Zoology, 257, 303309.
  • Hewison, A.J.M., Angibault, J.M., Cargnelutti, B., Coulon, A., Rames, J.L., Serrano, E., Verheyden, H. & Morellet, N. (2007) Using radio-tracking and direct observation to estimate roe deer Capreolus capreolus density in a fragmented landscape: a pilot study. Wildlife Biology, 13, 313320.
  • Hewison, A.J.M., Morellet, N., Verheyden, H., Daufresne, T., Angibault, J.M., Cargnelutti, B., Merlet, J., Picot, D., Rames, J.L., Joachim, J., Lourtet, B., Serrano, E., Bideau, E. & Cebe, N. (2009) Landscape fragmentation influences winter body mass of roe deer. Ecography, 32, 10621070.
  • Hewison, A.J.M., Gaillard, J.M., Delorme, D., Van Laere, G., Amblard, T. & Klein, F. (2011) Reproductive constraints, not environmental conditions, shape the ontogeny of sex-specific mass-size allometry in roe deer. Oikos, 120, 12171226.
  • Hofmann, R.R. (1989) Evolutionary steps of ecophysiological adaptation and diversification of ruminants - A comparative view of their digestive systems. Oecologia, 78, 443457.
  • Holekamp, K.E. (1986) Proximal causes of natal dispersal in beldings ground-squirrels (Spermophilus-Beldingi). Ecological Monographs, 56, 365391.
  • Holekamp, K.E. & Sherman, P.W. (1989) Why male ground-squirrels disperse. American Scientist, 77, 232239.
  • Howard, W.E. (1960) Innate and environmental dispersal of individual vertebrates. American Midland Naturalist, 63, 152161.
  • Ims, R.A. & Hjermann, D.O. (2001) Condition-dependent dispersal. Dispersal (eds J. Clobert, E. Danchin, A.A. Dhondt & J.D. Nichos), pp. ????????. Oxford University Press, New York.
  • Jin, R.M., Breitbart, Y. & Muoh, C. (2009) Data discretization unification. Knowledge and Information Systems, 19, 129.
  • Johnson, M.L. & Gaines, M.S. (1990) Evolution of dispersal – theoretical-models and empirical tests using birds and mammals. Annual Review of Ecology and Systematics, 21, 449480.
  • Johnson, C.A., Fryxell, J.M., Thompson, I.D. & Baker, J.A. (2009) Mortality risk increases with natal dispersal distance in American martens. Proceedings of the Royal Society B: Biological Sciences, 276, 33613367.
  • Kenward, R.E., Rushton, S.P., Perrin, C.M., Macdonald, D.W. & South, A.B. (2002) From Marking to Modelling: Dispersal Study Techniques for Land Vertebrates. Blackwell Publishing, Maldel.
  • Linnell, J., Wahlstrom, K. & Gaillard, J.M. (1998) From birth to independence. The European Roe Deer: The Biology of Success (eds R. Andersen, P. Duncan & J.D.C. Linnell). Scandinavian University Press, Oslo.
  • Loe, L.E., Mysterud, A., Veiberg, V. & Langvatn, R. (2010) No evidence of juvenile body mass affecting dispersal in male red deer. Journal of Zoology, 280, 8491.
  • Long, E.S., Diefenbach, D.R., Rosenberry, C.S., Wallingford, B.D. & Grund, M.R.D. (2005) Forest cover influences dispersal distance of white-tailed deer. Journal of Mammalogy, 86, 623629.
  • Long, E.S., Diefenbach, D.R., Rosenberry, C.S. & Wallingford, B.D. (2008) Multiple proximate and ultimate causes of natal dispersal in white-tailed deer. Behavioral Ecology, 19, 12351242.
  • Lowe, W.H. (2010) Explaining long-distance dispersal: effects of dispersal distance on survival and growth in a stream salamander. Ecology, 91, 30083015.
  • Matthysen, E., Adriaensen, F. & Dhondt, A.A. (1995) Dispersal distances of nuthatches, sitta-europaea, in a highly fragmented forest habitat. Oikos, 72, 375381.
  • McCullagh, P. & Nelder, J.A. (1989) Generalized Linear Models. Chapman and Hall, London.
  • Moller, A.P., Flensted-Jensen, E. & Mardal, W. (2006) Dispersal and climate change: a case study of the Arctic tern Sterna paradisaea. Global Change Biology, 12, 20052013.
  • Morales, J.M. & Ellner, S.P. (2002) Scaling up animal movements in heterogeneous landscapes: the importance of behavior. Ecology, 83, 22402247.
  • Morellet, N., Verheyden, H., Angibault, J.M., Cargnelutti, B., Lourtet, B. & Hewison, A.J.M. (2009) The effect of capture on ranging behaviour and activity of the European roe deer Capreolus capreolus. Wildlife Biology, 15, 278287.
  • Nathan, R. (2001) The challenges of studying dispersal. Trends in Ecology & Evolution, 16, 481483.
  • Nixon, C.M., Mankin, P.C., Etter, D.R., Hansen, L.P., Brewer, P.A., Chelsvig, J.E., Esker, T.L. & Sullivan, J.B. (2007) White-tailed deer dispersal behavior in an agricultural environment. American Midland Naturalist, 157, 212220.
  • Nunes, S. & Holekamp, K.E. (1996) Mass and fat influence the timing of natal dispersal in Belding's ground squirrels. Journal of Mammalogy, 77, 807817.
  • Paradis, E., Baillie, S.R., Sutherland, W.J. & Gregory, R.D. (1998) Patterns of natal and breeding dispersal in birds. Journal of Animal Ecology, 67, 518536.
  • Pettorelli, N., Gaillard, J.M., Duncan, P., Maillard, D., Van Laere, G. & Delorme, D. (1993) Age and density modify the effects of habitat quality on survival and movements of roe deer. Ecology, 84, 33073316.
  • Plard, F., Bonenfant, C. & Gaillard, J.M. (2011) Revisiting the allometry of antlers among deer species: male-male sexual competition as a driver. Oikos, 120, 601606.
  • R Development Core Team (2008) R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria.
  • Ratcliffe, P.R. & Mayle, B. (1992) Roe deer biology and management. Forestry Commision Bulletin, 105, 128.
  • Ronce, O. (2007) How does it feel to be like a rolling stone? Ten questions about dispersal evolution. Annual Review of Ecology Evolution and Systematics, 38, 231253.
  • Roshier, D.A. & Reid, J.R.W. (2003) On animal distributions in dynamic landscapes. Ecography, 26, 539544.
  • Rousset, F. & Gandon, S. (2002) Evolution of the distribution of dispersal distance under distance-dependent cost of dispersal. Journal of Evolutionary Biology, 15, 515523.
  • Selonen, V. & Hanski, I.K. (2010) Condition-dependent, phenotype-dependent and genetic-dependent factors in the natal dispersal of a solitary rodent. Journal of Animal Ecology, 79, 10931100.
  • Skjelseth, S., Ringsby, T.H., Tufto, J., Jensen, H. & Saether, B.E. (2007) Dispersal of introduced house sparrows Passer domesticus: an experiment. Proceedings of the Royal Society B: Biological Sciences, 274, 17631771.
  • Soulsbury, C.D., Baker, P.J., Iossa, G. & Harris, S. (2008) Fitness costs of dispersal in red foxes (Vulpes vulpes). Behavioral Ecology and Sociobiology, 62, 12891298.
  • Stamps, J.A. (2006) The silver spoon effect and habitat selection by natal dispersers. Ecology Letters, 9, 11791185.
  • Stamps, J.A., Krishnan, V.V. & Reid, M.L. (2005) Search costs and habitat selection by dispersers. Ecology, 86, 510518.
  • Stenseth, N.C. & Lidicker, W.Z. (1992) Animal Dispersal: Small Mammals as a Model. Chapman and Hall, London.
  • Strandgaard, H. (1972) The roe deer (Capreolus capreolus) population at Kalø and the factors regulating its size. Danish Review of Game Biology, 7, 1205.
  • Sutherland, G.D., Harestad, A.S., Price, K. & Lertzman, K.P. (2000) Scaling of natal dispersal distances in terrestrial birds and mammals. Conservation Ecology, 4, 16.
  • Toigo, C., Gaillard, J.M., Van Laere, G., Hewison, M. & Morellet, N. (2006) How does environmental variation influence body mass, body size, and body condition? Roe deer as a case study. Ecography, 29, 301308.
  • Van Moorter, B., Gaillard, J.M., Hewison, A.J.M., Said, S., Coulon, A., Delorme, D., Widmer, O., Cargnelutti, B. & Angibault, J.M. (2008) Evidence for exploration behaviour in young roe deer (Capreolus capreolus) prior to dispersal. Ethology Ecology & Evolution, 20, 115.
  • Vanpé, C., Gaillard, J.M., Kjellander, P., Mysterud, A., Magnien, P., Delorme, D., Van Laere, G., Klein, F., Liberg, O. & Hewison, A.J.M. (2007) Antler size provides an honest signal of male phenotypic quality in roe deer. American Naturalist, 169, 481493.
  • Vanpé, C., Kjellander, P., Galan, M., Cosson, J.F., Aulagnier, S., Liberg, O. & Hewison, A.J.M. (2008) Mating system, sexual dimorphism, and the opportunity for sexual selection in a territorial ungulate. Behavioral Ecology, 19, 309316.
  • Vanpé, C., Gaillard, J.M., Morellet, N., Kjellander, P., Liberg, O., Delorme, D. & Hewison, A.J.M. (2009a) Age-specific variation in male breeding success of a territorial ungulate species, the European roe deer. Journal of Mammalogy, 90, 661665.
  • Vanpé, C., Kjellander, P., Gaillard, J.M., Cosson, J.F., Galan, M. & Hewison, A.J.M. (2009b) Multiple paternity occurs with low frequency in the territorial roe deer, Capreolus capreolus. Biological Journal of the Linnean Society, 97, 128139.
  • Vanpé, C., Gaillard, J.M., Kjellander, P., Liberg, O., Delorme, D. & Hewison, A.J.M. (2010) Assessing the intensity of sexual selection on male body mass and antler length in roe deer Capreolus capreolus: is bigger better in a weakly dimorphic species? Oikos, 119, 14841492.
  • Wahlstrom, L.K. (1994) The significance of male male-aggression for yearling dispersal in roe deer (Capreolus-Capreolus). Behavioral Ecology and Sociobiology, 35, 409412.
  • Wahlstrom, L.K. & Liberg, O. (1995) Patterns of dispersal and seasonal migration in roe deer (Capreolus capreolus). Journal of Zoology, 235, 455467.
  • Worton, B.J. (1989) Kernel methods for estimating the utilization distribution in home-range studies. Ecology, 70, 164168.