Abstract

Mountain hares (Lepus timidus hibernicus Bell) are extremely vagile and commute up to 3 km daily within their home ranges. However, observational and mark-recapture evidence suggests that they do not disperse far in their lifetime. Whether behavioral factors such as philopatry are promoting genetic structuring in this largely solitary species is not known. We examined genetic structure in the mountain hare using microsatellite markers from 321 mountain hares from across southern Ireland. Genetic differentiation ranged from low to moderate (pairwise FST = 0-0.116), and across distances of 200 km, FSt was not correlated with geographic distance, suggesting possible population fragmentation. Home ranges of males in this species are significantly larger than those of females during the breeding season, suggesting that dispersal may be male-biased. Mean corrected assignment index (mAIc) was lower in males (-0.280) than in females (0.403). FSt was lower among male cohorts of population samples (0.040) than among female cohorts (0.051), but the difference was not significant. These results are consistent with natal dispersal of nonresident males into the sampling areas.

The mountain hare (Lepus timidus L.) has a distribution extending across Eurasia from Norway to Japan. The species forms part of a Holarctic complex with L. arcticus Ross and L. othus Merriam in North America and Greenland. In Europe, its range extends across Fennoscandia, the Baltic States, and Russia and there are isolated glacial relict populations in Ireland (L. t. hibernicus), Scotland (L. t. scoticus), and the Alps (L. t. varronisAngerbjorn and Flux 1995). L. t. hibernicus Bell, 1837, is found on moorland and agricultural land at all elevations down to sea level in Ireland (Whelan 1985; Wolfe and Hay den 1996), in contrast to populations in Scotland and the Alps, which are limited to highland areas, possibly because of competitive exclusion by brown hares (Lepus europaeus Pallas—Thulin 2003). Although mountain hares in Ireland are widespread and locally common, studies have suggested that populations have declined in recent years (Dingerkus and Montgomery 2002; Preston et al. 2002), although examination of some data indicates that populations may be recovering (Tosh et al. 2004, 2005). Population declines may be related to changing agricultural practices, such as intensification and increased mechanization (Dingerkus 1997), as well as the destruction of hedgerows, which provide corridors for small mammals (Coffman et al. 2001) and daytime resting areas for mountain hares (Dingerkus and Montgomery 2001).

Previous molecular work on the mountain hare has focused on phylogeographic patterns across relictual subspecies in Europe (Hamill et al. 2006; Pierpaoli et al. 1999; Suchentrunk et al. 1999). Within subspecies, little is known about genetic structure. Home ranges of mountain hares normally include both a day-resting site and a separate night-feeding, area that may be situated 1–3 km from each other (Flux 1970; Hewson and Hinge 1990). Thus, the high vagility of mountain hares may be predicted to lead to high genetic connectivity between populations. However, mountain hares appear to be strongly philopatric. A mark-recapture study in Scotland showed that 1st-winter hares were more likely than adults to be recaptured within 100 m of their location the previous year, indicating infrequent juvenile dispersal (Hewson 1990a). In the same study, just 1 instance of long-distance dispersal was detected: a single male hare (of 494 marked) was recovered 12 km from the study site (Hewson 1990a), indicating that most individuals are relatively philopatric.

Such strong philopatry may be associated with reduced gene flow and local inbreeding. Factors such as the degree of territoriality, type of breeding system (Storz 1999), density-dependent population fluctuations (Angerbjorn 1983; Hewson 1985), habitat structure, and resource levels (Watson and Hewson 1973) determine whether individuals disperse or remain close to their natal site. These influences might act to reduce gene flow between populations. Where dispersal presents costs, it will tend to be optimized by natural selection (Favre et al. 1997) and the sex for which it is more costly is predicted to disperse less. Young mountain hares receive solely maternal care and home ranges of males are larger than those of females (Hewson and Hinge 1990), consistent with a polygy-nous or polygynandrous mating system (Storz 1999), although males do not hold territories and rarely defend females (Flux 1970). Sex-biased dispersal is a common reproductive strategy adopted by many species (Prugnolle and de Meeus 2002). In polygynous species of mammals, dispersal is predicted to be male-biased because of the greater benefits to females of remaining in familiar territory (Greenwood 1980), increased competition among related males for access to breeding partners (Dobson 1982; Hamilton 1972), inbreeding avoidance (Monard and Duncan 1996; Wolff 1993), or a combination of these factors (Favre et al. 1997).

Sex-biased dispersal among populations of hares has been investigated in few species and the results are equivocal. Both brown and snowshoe hares (Lepus americanus Erxleben) are polygynandrous (Burton and Krebs 2003; Hewson 1990b) and have complex mating behaviors and both are predicted to be male-biased in dispersal under each of the above models (Dobson 1982). Mitochondrial and microsatellite evidence identified male-biased dispersal in a population of brown hares (Fickel et al. 1999, 2005). By contrast, assignment indices of microsatellite data indicate that dispersal is equal among the sexes in snowshoe hares (Burton et al. 2002). Although mountain hares do not display such complex mating behavior as brown hares (Flux 1970), there is a degree of population structure in populations of mountain hares during the breeding season when home ranges of males expand by 50% (Hewson and Hinge 1990) and become significantly larger than those of females (Flux 1970), which expand by 10% (Hewson and Hinge 1990). Radiotracked females in Ireland had smaller home ranges than males during the breeding season (Wolfe and Hay den 1996), although the trend was not significant. This suggests a breeding system where males roam wider in order to secure matings, and females are philopatric. Natal dispersal of a subset of these males, that is, the permanent movement from the natal site to reproduce elsewhere, may be biasing gene flow.

Here we apply a panel of 7 polymorphic microsatellite loci originally developed for the European rabbit (Oryctolagus cuniculus L.) to a survey of population structure in populations of mountain hares in Ireland. We. quantified the level of gene flow among populations of mountain hares in Ireland, compared these levels with predicted structure based on observations in L. t. hibernicus and other subspecies, and examined whether there is a sex bias in dispersal among populations of this species. Our null hypothesis was complete panmixis between all sampling areas. Our 2nd hypothesis was strong site philopatry, which predicts high differentiation among populations and high likelihood of individuals being assigned to their own source population. Our 3rd hypothesis was high dispersal, predicting weak population differentiation and genetic isolation by geographical distance. In relation to how this gene flow is mediated, we hypothesized that genetic differentiation (FST) will be higher among female than male cohorts of the population and that females will have higher mean assignment indices to their source populations than males. Natural populations of mountain hares are likely to meet most of the simplifying assumptions about life-history parameters of such tests, including that there are nonoverlapping generations, dispersal occurs at the juvenile stage before reproduction, and sampling occurs postdispersal. In mountain hares, the generation time is usually 1 year, occasionally 2 years (Flux 1970), and all samples were taken during winter, immediately before and during the early part of the breeding season.

Materials and Methods

DNA sampling.—We sampled populations of mountain hares that were caught under license for purposes of coursing (the practice of pursuing hares with muzzled dogs that follow by sight) at 9 locations in southwestern, central, and eastern Ireland in autumn and winter of 1999 and spring 2000. The sampled hares were caught using long nets, in the immediate radius (approximately 1–4 km) of sites marked in Fig. 1 in the month before each coursing meeting. Approximately 60–100 underhair follicles were plucked from each of 321 live individuals of L. t. hibernicus and stored dry at room temperature before DNA extraction. Guidelines of the American Society of Mammalogists (Animal Care and Use Committee 1998) were followed during tissue collection. The location of sample sites and sample identification codes may be seen in Fig. 1.

Fig. 1

Collection sites for samples of mountain hare (Lepus timidus hibernicus) were in the environs of the townlands indicated. Abbreviations for sampling areas are as follows: BDF = Bally duff, County Kerry; ABDY = Abbeydorney, County Kerry; LLX = Lixnaw, County Kerry; CSTL = Casteleisland, County Kerry; ENN = Ennis, County Clare; CLON = Kilsheelan, County Tipperary; KKY = Freshford, County Kilkenny; KCLN = Old Kilcullen, County Kildare; BBN = Balbriggan, County Dublin. Samples were collected in winter of 1999 and spring of 2000.

Fig. 1

Collection sites for samples of mountain hare (Lepus timidus hibernicus) were in the environs of the townlands indicated. Abbreviations for sampling areas are as follows: BDF = Bally duff, County Kerry; ABDY = Abbeydorney, County Kerry; LLX = Lixnaw, County Kerry; CSTL = Casteleisland, County Kerry; ENN = Ennis, County Clare; CLON = Kilsheelan, County Tipperary; KKY = Freshford, County Kilkenny; KCLN = Old Kilcullen, County Kildare; BBN = Balbriggan, County Dublin. Samples were collected in winter of 1999 and spring of 2000.

DNA extraction and polymerase chain reaction.—DNA was extracted by incubating 20–30 hair follicles from each animal with an extraction buffer comprising 81.5 µl polymerase chain reaction (PCR)-grade water, IX Reaction Buffer (Bioline, London, United Kingdom), 2.5 mM MgCl2, 0.5 µ1 Tween-20, 0.6 µ1 Tergitol Type NP-40 (liquid form after incubation for 10 min at 60°C), and 200 µg Proteinase K, for 45 minutes at 56°C and then at 95°C for 15 min. One microliter of the extraction buffer was used directly in a PCR. Seven microsatellite loci originally developed in the rabbit were optimized for mountain hare DNA and screened in all individuals. The PCR for all primers consisted of IX Reaction Buffer (Bioline), optimized MgCl2 concentrations, 0.2 mM each deoxynucleoside triphosphate, 0.4 mM Cy-5 labeled forward and 0.4 mM reverse primer, and 0.5 U BioTaq DNA polymerase (Bioline). Annealing temperature (TA) and MgCl2 concentration (Mg) are as follows: for the loci SOL33 (Mg = 1.5 mM, TA = 52°C), SOL30 (Mg = 1.5 mM, TA = 61°C—Rico et al. 1994), OCELAMB (Mg = 1.0 mM, TA = 61°C), OCLS1B (Mg = 1.5 mM, TA = 52°C), OCRLADF4 (Mg = 1.5 mM, TA = 61°C—van Haeringen et al. 1996), and D7UTR1 (Mg = 2.0 mM, TA = 61°C—Korstanje et al. 2003). PCR primers for D7UTR1 were forward-5′-ACACCTGGGGAATAAACAACAAG-3 ′ and reverse-5′-GAGGGAGGCAGAGGGATAAGA-3 '. Cycling conditions for these loci were as follows: 95°C for 3 min, then 7 cycles of 95°C for 15 s, variable annealing temperature for 15 s and 72°C for 25 s, and 25 cycles of 89°C for 15 s, variable annealing temperature for 15 s and 72°C for 25 s. Cycling conditions for S AT 12 (Mg = 1.5 mM, TA = 52°C) were according to Andersson et al. (1999). Products were separated on the Long Read Tower fluorescent sequencer (Visible Genetics Inc., Toronto, Ontario, Canada), with at least 2 internal lane size standards flanking the expected size range of each locus.

Because of time constraints, we could not determine the sex of individuals in the field. Instead, sex of individuals from 5 sites was determined by selective PCR amplification of a fragment of the Sry sex-determining gene on the Y-chromosome (size approximately 300 base pair [bp]) in duplex with a segment of the Transferrin gene (size approximately 500 bp) as per Wallner et al. (2001), and PCRs were subjected to electrophoresis on ethidium bromide-stained 2% agarose gels. Gel photographs were examined by eye to detect the number of amplicons for each individual. With every set of reactions carried out, 2 positive controls (1 male and 1 female) were included in the PCR and gel run. If these bands did not show up as expected, PCR was repeated. If an individual failed to produce any band or the result was not conclusive, PCR for that individual was repeated. For every set of reactions carried out, 5 individuals that had produced an interpretable result in the previous screening were rerun with the next experiment.

Analysis of genetic diversity.—Genotypic data for each individual were formatted for analysis using the Microsoft Excel add-in Microsatellite Toolkit (Park 2001). Estimation of pairwise linkage disequilibria for each pair of loci was calculated by LINKDOS (Garnier-Gere and Dillmann 1992). Allelic count and allelic richness were calculated in FSTAT, version 2.93.2 (Goudet 1995; J. Goudet, 2001, FSTAT, A Program to Estimate and Test Gene Diversities and Fixation Indices,http://www.unil.ch/izea/softwares/fstat.html). Total observed and expected heterozygosity and FIS were calculated for each sample and samples were tested for Hardy-Weinberg equilibrium using the Score test (Rousset and Raymond 1995) in GENEPOP, version 3.3 (Raymond and Rousset 1995a).

Analysis of genetic structure.—We did an exact test for genotypic differentiation in FSTAT, version 2.93.2 (J. Goudet, 2001, FSTAT, A Program to Estimate and Test Gene Diversities and Fixation Indices,http://www.unil.ch/izea/softwares/fstat.html). An overall estimate of FST was obtained in GENEPOP. We used this value to estimate the number of migrants per population, that is, the product of migration rate and effective population size (Nm), under island-model assumptions using Wright's equation Nm = (1 — FSt)/4Fst (Wright 1969). Pairwise estimates of FST between each pair of samples were generated and tested for significance, using Arlequin, version 2.0 (Schneider et al. 2000). In order to obtain an individual-based estimate of how distinct the population samples were from each other, partially Bayesian assignment tests were done in GENECLASS (Cornuet et al. 1999). Two methods were employed; initially, we did direct assignment with the likelihood of belonging to each sample estimated using formula 9 of Rannala and Mountain (1997). Each individual was assigned to the sample in which its genotype had highest likelihood. We calculated the proportion of individuals assigned to their source sample. We also used the probability-of-belonging method and simulated 10,000 genotypes for each sample based on sample allele frequencies in GENECLASS (Cornuet et al. 1999). Each individual was assigned to the population in which its probability of belonging was highest, relative to the probability of belonging of the simulated genotypes. In order to eliminate bias, when genotypes were simulated, the focal genotype was excluded from the source sample. We also excluded population Abbeydorney (ABDY) from these tests because of small sample size. All samples in which the probability of belonging, exceeded 0.05 were counted, as an estimate of the admixture levels of the sample as a whole. We tested for an isolation-by-distance effect (Slatkin 1993) using the ISOLDE option in GENEPOP (Raymond and Rousset 1995a). A regression of matrices of genetic distance (FST/(1-FST) from GENEPOP) with log-transformed geographic distance in kilometers (measured in ArcMap [ESRI, Redlands, CA, USA] using the British national grid system) between sampling sites was carried out.

Tests for sex-biased dispersal.Goudet et al. (2002) proposed a number of tests for sex-biased dispersal, but the most powerful were mean assignment index (Paetkau et al. 1995) and differential FST (Weir and Cockerham 1984). The tests make use of randomization procedures for significance testing. Randomization proceeds by initially estimating the test statistic for the real data set, then randomly assigning a sex to each individual in the study, keeping sex ratio constant in each sample, then recalculating the statistic for the randomized data set. This was carried out 2,000 times and a null distribution was generated. The proportion of times the test statistic was exceeded by the randomized data set was the P-value for the test.

For each of the sex-typed individuals, the assignment index of Paetkau et al. (1995) was calculated and probabilities of assignment to source population were estimated using FSTAT (J. Goudet, 2001, FSTAT, A Program to Estimate and Test Gene Diversities and Fixation Indices,http://www.unil.ch/izea/softwares/fstat.html). The probability that the genotype of a given individual k, occurred by chance in population l, given the allelic frequencies in that population (calculated including the focal individual) is the assignment index (AI) of individual k to population l. These AI values are corrected (AIC) for variable genetic diversity across populations, by initially log-transforming them and then subtracting the average value of the population assignment index from the multilocus probability of each individual. The average AIC value of each population is then zero, regardless of the genetic diversity present in that population. Negative values indicate individuals less likely than average to have been born locally. We estimated the mean AIC for males and females in each population. Because immigrants tend to have lower assignment indices than residents, we postulated that male assignment indices would be significantly lower than those of the females in each population because of the presence of male immigrants in each population. A t-statistic corresponding to the difference between the AIC in the philopatric and dispersing cohorts divided by the square root of the sum of the total variance within each group was used (Goudet et al. 2002). We also examined the possibility of sex-biased dispersal by comparing among-population FST in male and female cohorts of the population. Under conditions of male-biased dispersal, it is predicted that FST will be greater among the female than the male cohorts of each sample, because some males will be immigrants and this will reduce the proportion of genetic variance attributable to differentiation into subpopulations in males. The test statistic used is (FSTF — FStm).

Results

Genetic diversity.—Complete genotypes at 7 loci (no non-amplifying individuals for any locus) were obtained for 321 individuals (see Fig. 1 for details of sample sizes and location code abbreviations). No significant linkage disequilibrium was found; therefore the 7 loci were treated as independent markers. All loci were polymorphic (2-13 alleles observed per locus) and genetic diversity in L. t. hibernicus was moderate compared to that in other mammalian species, as indicated by average locus specific heterozygosity. Forty-two alleles were observed in total (Table 1). Mean allele count per sample per locus was 4.4 and average expected heterozygosity across loci was 0.518 per sample. A test for departure from Hardy-Weinberg equilibrium in the overall subspecies sample was significant (P = 0.001), therefore the individual samples of the study were assessed for goodness-of-fit to Hardy-Weinberg equilibrium. Fisher's Global test was applied across loci, and, after Bonferroni correction (Rice 1989), only 1 sample displayed significant departure from Hardy-Weinberg equilibrium (Ballyduff [BDF]; P = 0.0015 across pooled loci).

Table 1

Genetic diversity indices of samples of Lepus timidus hibernicus collected in Ireland in winter 1999 and spring 2000. Sample locations are shown in Fig. 1. N = number of individuals sampled per location; A = mean number of alleles observed per location per locus; AR = allelic richness, as calculated in FSTAT, version 2.93.2 (J. Goudet, 2001, FSTAT, A Program to Estimate and Test Gene Diversities and Fixation Indices,http://www.unil.ch/izea/softwares/fstat.html) estimated for minimum location sample size n = 19; HE = expected heterozygosity; Ho = observed heterozygosity, calculated in GDA (P. O. Lewis and D. Zaykin, 2001, Genetic Data Analysis: Computer Program for the Analysis of Allelic Data,http://lewis.eeb.uconn.edu/lewishome/software.html); and HWE = combined P-value estimate for departure from Hardy-Weinberg equilibrium for all loci using Fisher's method, calculated in GENEPOP (Raymond and Rousset 1995a).

 Ar Overall FIs HWE (P -value) 
Sample N A (n = 19) HE Ho 
BDF 53 4.9 4.3 0.514 0.496 0.035 0.002 
ABDY 10 4.3 — 0.535 0.471 0.125 0.468 
L1X 48 4.7 4.3 0.548 0.5 0.089 0.057 
CSTL 50 5.3 4.6 0.584 0.54 0.076 0.758 
ENN 20 3.9 3.8 0.469 0.436 0.072 0.515 
CLON 25 4.1 4.1 0.533 0.543 -0.02 0.074 
KKY 19 4.3 4.3 0.496 0.451 0.092 0.154 
KCLN 48 4.6 4.2 0.509 0.452 0.112 0.043 
BBN 48 3.7 0.472 0.461 0.023 0.989 
X 35.7 4.5 4.2 0.518 0.483 0.067 — 
L. t. hibernicus 321 — 0.538 0.488 0.092 0.004 
 Ar Overall FIs HWE (P -value) 
Sample N A (n = 19) HE Ho 
BDF 53 4.9 4.3 0.514 0.496 0.035 0.002 
ABDY 10 4.3 — 0.535 0.471 0.125 0.468 
L1X 48 4.7 4.3 0.548 0.5 0.089 0.057 
CSTL 50 5.3 4.6 0.584 0.54 0.076 0.758 
ENN 20 3.9 3.8 0.469 0.436 0.072 0.515 
CLON 25 4.1 4.1 0.533 0.543 -0.02 0.074 
KKY 19 4.3 4.3 0.496 0.451 0.092 0.154 
KCLN 48 4.6 4.2 0.509 0.452 0.112 0.043 
BBN 48 3.7 0.472 0.461 0.023 0.989 
X 35.7 4.5 4.2 0.518 0.483 0.067 — 
L. t. hibernicus 321 — 0.538 0.488 0.092 0.004 

Genetic structure.—An exact test for overall genie differentiation (Raymond and Rousset 1995b) across all samples was significant (P < 0.001), but an overall estimate of genetic differentiation was quite low (FSt = 0.035). This value would correspond to about 6.9 migrants per sample and per year under the island model of gene flow (Wright 1931), indicating high migration between subsamples. Pairwise FSt was low to moderate (Table 2); estimates between sampling areas were variable and ranged from <0.001 to 0.11. Neighboring samples were generally not significantly differentiated (Table 2). The Balbriggan (BBN) sample (most outlying by geographic distance; Fig. 1) was significantly different from all other samples in pairwise exact tests of genotypic differentiation. Additionally, pairwise FST values between Balbriggan and all other samples averaged 0.073, in comparison to 0.029 among significant estimates between other sample pairs.

Table 2

Genetic divergence (FSt) between samples of Lepus timidus hibernicus collected in Ireland in winter 1999 and spring 2000 (below diagonal) and P -values from pairwise exact tests for genie differentiation (above diagonal). Sample locations are shown in Fig. 1. FST was calculated in Arlequin, version 2.0 (Schneider et al. 2000). Significance (*) at the nominal P -value (sequential Bonferroni corrected) for genie differentiation was calculated in FSTAT, version 2.93.2 (J. Goudet, 2001, FSTAT, A Program to Estimate and Test Gene Diversities and Fixation Indices,http://www.unil.ch/izea/softwares/fstat.html). NS = not significant.

 BDF ABDY LIX CSTL ENN CLON KKY KCLN BBN 
BDF  NS NS  
ABDY 0.034  NS NS NS NS NS NS 
LIX 0.002 0.017  NS NS NS NS 
CSTL 0.021 -0.003 0.011     
ENN 0.036 -0.006 0.015 0.022  NS 
CLON 0.041 0.006 0.027 0.026 0.041  NS  
KKY 0.046 0.006 0.039 0.039 0.044 0.002  NS 
KCLN 0.018 0.013 0.006 0.031 0.022 0.015 0.015  
BBN 0.056 0.091 0.063 0.081 0.116 0.073 0.063 0.041  
 BDF ABDY LIX CSTL ENN CLON KKY KCLN BBN 
BDF  NS NS  
ABDY 0.034  NS NS NS NS NS NS 
LIX 0.002 0.017  NS NS NS NS 
CSTL 0.021 -0.003 0.011     
ENN 0.036 -0.006 0.015 0.022  NS 
CLON 0.041 0.006 0.027 0.026 0.041  NS  
KKY 0.046 0.006 0.039 0.039 0.044 0.002  NS 
KCLN 0.018 0.013 0.006 0.031 0.022 0.015 0.015  
BBN 0.056 0.091 0.063 0.081 0.116 0.073 0.063 0.041  

Table 3 shows that in the direct Bayesian assignment test (Rannala and Mountain 1997), just 32% of individuals were assigned to their source samples (i.e., had higher probability of belonging to this than any other sample), but 53% of individuals were assigned to samples within 50 km of their source sample, indicating similar genotypic distributions at this scale. The sample with the highest rate of self-assignment was Balbriggan (BBN); 68% of Balbriggan individuals were self-assigned with highest likelihood (Table 3). We pooled all individuals into 50-km areas, that is, we included Ballyduff (BDF), Lixnaw (LIX), and Castleisland (CSTL); County Kerry samples; and the Clonmel (CLON) and Kilkenny (KKY) samples (into the Southeast sample); and then carried out the Bayesian simulation analysis (Cornuet et al. 1999). We plotted the proportion of individuals assigned to each sample. Fig. 2 shows that the majority of individuals in 4 of the 5 sampling areas were self-assigned. In the Kerry sampling area, which had the largest sample size, 80% of the sample was self-assigned, and 56% of the Balbriggan sample was self-assigned. The Kilcullen, County Kildare (KCLN), individuals were the least self-assigned group, with just 20% self-assigned, equivalent to random assignment.

Fig. 2

Classification analysis, using the partially Bayesian approach, with samples of Lepus timidus hibernicus pooled into 50-km areas. Samples were collected in winter 1999 and spring 2000. Locations of samples are shown in Fig. 1; pooling of samples is described in the text.

Fig. 2

Classification analysis, using the partially Bayesian approach, with samples of Lepus timidus hibernicus pooled into 50-km areas. Samples were collected in winter 1999 and spring 2000. Locations of samples are shown in Fig. 1; pooling of samples is described in the text.

Table 3

Proportion of mountain hares (Lepus timidus hibernicus) classified to their own and other samples using direct assignment (Rannala and Mountain 1997) in GENECLASS (Cornuet et al. 1999).a Sample locations are shown in Fig. 1.

Sample n Self-assigned Assigned to samples within-50 km of sample site 
BDF 53 0.28 BDF, ABDY, LIX, CSTL 0.68 
ABDY 10 BDF, ABDY, LIX, CSTL 0.30 
LIX 48 0.1 BDF, ABDY, LIX, CSTL 0.5 
CSTL 50 0.35 BDF, ABDY, LIX, CSTL 0.7 
ENN 20 0.5 No sample within 50 km 0.5 
CLON 20 0.36 CLON and KKY (16) 0.8 
KKY 19 0.15 CLON and KKY (7) 0.37 
KCLN 48 0.21 No sample within 50 km 0.21 
BBN 48 0.69 No sample within 50 km 0.69 
Mean 35.1 0.29  0.53 
Sample n Self-assigned Assigned to samples within-50 km of sample site 
BDF 53 0.28 BDF, ABDY, LIX, CSTL 0.68 
ABDY 10 BDF, ABDY, LIX, CSTL 0.30 
LIX 48 0.1 BDF, ABDY, LIX, CSTL 0.5 
CSTL 50 0.35 BDF, ABDY, LIX, CSTL 0.7 
ENN 20 0.5 No sample within 50 km 0.5 
CLON 20 0.36 CLON and KKY (16) 0.8 
KKY 19 0.15 CLON and KKY (7) 0.37 
KCLN 48 0.21 No sample within 50 km 0.21 
BBN 48 0.69 No sample within 50 km 0.69 
Mean 35.1 0.29  0.53 
a

Classification tests applied Bayesian direct assignment, incorporating the leave-one-out approach.

Genetic distance was significantly correlated with increasing geographic distance within the subspecies (R2 = 0.23, P = 0.007; Mantel test). However, because the Balbriggan (BBN) sample was both the most genetically distinct and was also at one extreme of the geographic range (Fig. 1), we repeated the regression excluding this sample and found the trend was much lower and no longer significant (R2 = 0.07, P = 0.24; Mantel test). This indicates that genetic structure does not really follow an isolation-by-distance model; instead the data are showing a bias due to pseudoreplication because each sample is involved in multiple nonindependent comparisons. It is possible that we have identified 2 demes that are significantly differentiated from each other, and moderately diverged (FST = 0.073). At the scale of 200 km, that is, within demes, it is possible that an island model is appropriate to our data. Alternatively, there may be recent population fragmentation, and the genetic similarity is due to limited genetic drift occurring since then.

Tests for sex-bias in individual dispersal.—The sex of 156 individuals from 5 sites was determined by selective PCR amplification. We identified more males (92) than females (64). However, this probably does not mean there is an unequal sex ratio. PCRs with DNA from females resulted in the amplification of a single band. Sometimes if the single band was weak and a faint band indicating males was undetectable by eye (the Tf band was brighter than the Sry band), we excluded that data point.

We found that mean assignment index of the sampled mountain hares was higher for females than males, as predicted by our hypothesis. The 64 females in the study had a positive mean assignment index (AIC) of 0.403 in contrast to the negative mean assignment index of -0.280 for the 92 males. The difference was significant at the nominal level (P = 0.032; 2,000 randomizations), but not after strict Bonferroni correction for multiple testing. The component of overall FSt contributed by males was 0.0397 and was lower than the overall FSt between female cohorts in each sample (0.0505), but this difference was not significant at the nominal level. Expected heterozygosity was significantly higher at the nominal level among male cohorts of the sample than among females (0.5395 for males compared to 0.4923 for females; P = 0.048, 2,000 randomizations) as predicted by our hypothesis, but were not significant after correction for multiple testing. As predicted, observed heterozygosity did not significantly differ between males and females (0.4814 compared to 0.4911). These tests for biased dispersal probably reveal a tendency toward higher dispersal in males.

Discussion

Population structure of mountain hares in Ireland.—The analysis of highly polymorphic molecular markers can allow the rapid estimation of population structure and dispersal levels—data that can take years to acquire by conventional mark-recapture or radiotracking methods. Our data set paints a picture of moderately diverse, weakly differentiated populations of mountain hares across southern Ireland and suggests the presence of 2 genetically divergent demes that may be considered management units. Overall genetic diversity was moderate (Haiti and Clark 1997) with heterozygosity of about 0.54 and 4 or 5 alleles per locus. This was lower than the mean observed in other European subspecies of the mountain hare genotyped at the same loci (Hamill et al. 2006), indicating that populations from Ireland have maintained smaller effective population sizes. Although both sample size and locus polymorphism were variable in this study, it is likely that these are reliable estimates of diversity in this subspecies, because these loci have the potential to have higher diversity in other samples, for example, heterozygosity of 0.8-0.9 was observed with the same loci in Fennoscandian samples; in all but 1 of our samples, sample size was more than double the number of alleles detected; there was no correlation between number of alleles per sample and sample size (data not shown); and the incidence of private alleles was extremely low (4 private alleles in 4 different samples with cumulative frequency of <0.01).

Using the panel of 7 polymorphic microsatellite markers, we were able to reject the null hypothesis of panmixis across the study range. However, we could also reject the 2nd hypothesis of extreme philopatry because samples were weakly differentiated and no significant differences were found between samples taken 10–30 km apart. Instead, we identified low but significant genetic differentiation among the samples, supporting our 3rd hypothesis. Gene flow was high between neighboring sample areas, as evidenced by nonsignificant pairwise comparisons between geographically close samples, but at wider scales, several pairwise comparisons across the southern part of the range of mountain hares in Ireland indicated significant differences between genotypic distributions among sampling areas. A simple model of isolation by geographic distance did not fit the data, so natural populations are possibly fragmented and patchy, rather than continuous. However, population fragmentation is likely to be recent because overall genetic differentiation, in terms of FST, was low between most population sample pairs, with the exception of Balbriggan (BBN), which was significantly differentiated at a moderate level (Wright 1978) from the other samples.

Individual-based Bayesian assignment tests indicated considerable misassignment of individuals to nonsource samples across most of the subspecies range, because of the similarity of the allelic distributions in all samples. Direct comparison with parallel mark-recapture studies shows that assignment indices are very reliable indicators of the natal site of individual dispersers provided sufficient samples per loci are screened (Berry et al. 2004; Favre et al. 1997). It seems likely that the addition of further loci would increase the proportion of samples that were self-assigned. However, previous work involving empirical data and simulation has found that assignment accuracy may be better improved by increasing the number of individuals screened (Berry et al. 2004), rather than number of loci. Such samples will be more likely to include rare alleles and to sample individual dispersers, where dispersal is rare. This may explain why only the largest sample tested (the pooled Kerry sample of 120 individuals) achieved self-assignment rates as high as 80%.

Comparison with mark-recapture studies indicates that highly accurate individual assignment to natal site can also depend on other factors; for example, greater accuracy is achieved where source populations are genetically well differentiated (Berry et al. 2004; Cornuet et al. 1999). Here, populations on the whole were relatively undifferentiated, with FSt being <0.05 between most sampling areas—the sample that had greatest proportion of individuals self-assigned in our direct assignment tests was the Balbriggan (BBN) sample, which was the most divergent gene pool in the study. Accurate individual assignment to natal site also may be affected by polymorphism of the loci (Berry et al. 2005). High precision and frequency of correct assignments has been achieved with data sets that applied highly polymorphic markers (hetero-zgyosities of 0.7-0.8—Berry et al. 2005; Manel et al. 2002). In contrast, our data set was moderately polymorphic with heterozygosity of approximately 0.55. The simulation analysis of Cornuet et al. (1999) also can be used to exclude particular populations as potential source populations. Simulations were used to statistically exclude potential source samples by permitting assignment of individuals to samples only where they had assignment indices within- the 0.95 range of the 10,000 simulated genotypes run in GENECLASS. None of the samples in the study area were excluded as potential sources for 37% of individuals tested. This is illustrative of the high degree of misclassification among samples due to the low genetic differentiation identified. Additionally, populations were not sampled continuously across Ireland, therefore analyses that seek to assign individuals to neighboring sample sites may not achieve high power.

Implications for conservation managers.—Are mountain hare populations in Ireland declining or becoming fragmented? The low genetic divergence observed over large areas of Munster and southern Leinster indicates that populations have until recently experienced high gene flow. However, significant differences were identified between genotypic distributions of several populations and the data did not fit an isolation-by-distance model. This may indicate recent fragmentation of populations, consistent with declining habitat connectivity across southern Ireland. When populations become fragmented, for example, because of changing patterns of agricultural practice, there is generally a lag in the time between the fragmentation event with the accompanying cessation in gene flow, and the time when resulting drift in the genotypic distributions can be detected (Honnay et al. 2005). Such populations can become vulnerable even before a great deal of genetic structuring is observed, because they cannot exchange migrants with other populations (Hale et al. 2001). Thus, although genetic diversity is moderate and divergence between populations in Ireland is not high, the significant fragmentation of the populations may be a cause for concern among conservation managers. The Balbriggan (BBN) sample was the most genetically divergent sample. Because this population lies at a geographic extreme of the studied range (Fig. 1), it is possible that we have sampled individuals that are part of a separate breeding unit or deme. There are no significant natural barriers in Leinster to isolate this population but human activity and infrastructure may have contributed to a recent fragmentation of natural populations of mountain hares. However, the genetic distinctiveness of the hares in this particular area has been recognized in the past. On the basis of a buff coloration of the fur, Barrett-Hamilton in the early 20th century ascribed the name L. timidus lutescens (Barrett-Hamilton, 1912—cited in Corbet and Harris 1991) to hares from the Balbriggan area— a name now arranged as a synonym of hibernicus. The population genetics appears to suggest a degree of differentiation, which may merit further study and requirement for conservation as a separate management unit. Alternatively, there may be a deme extending across the northeastern part of the country, of which the Balbriggan sample is a part. Further sampling from more northerly areas would clarify the relationships of this population.

Is there a sex-bias in individual dispersal in mountain hare populations?—Although data are limited, previous research has shown that the home ranges of male mountain hares are larger than those of females during the breeding season (Hewson and Hinge 1990). The longest (observed) dispersers have been males (Hewson 1990a), suggesting that females may be more philopatric than males in this species. Recent genetic work has demonstrated the utility of microsatellite markers in elucidating patterns of individual dispersal that differ between the sexes (Austin et al. 2003; Blundell et al. 2002; Spong and Creel 2001). Direct comparison with parallel mark-recapture studies has shown that assignment indices can be very reliable indicators of the dispersing status of individuals. There is suggestive evidence from our microsatellite data that male mountain hares disperse further than females from their natal site. Expected heterozygosity is also expected to be higher in the longer-dispersing sex, but observed heterozygosity is not expected to differ between the sexes. We observed this pattern with tests for male-biased dispersal. Additionally, mean assignment index was significantly different at the nominal level between males and females (although not after Bonferroni correction). Males were negative on average, and females were positive, indicating that sampled females were more likely to have been born in the sampling area than sampled males. The most likely explanation of this tendency is that male cohorts of each sample include immigrants into the sampling area with extraneous genotypes and that more males are moving further from their natal site than are females. In other lagomorph species in which dispersal for reproductive purposes is sex-biased, the males disperse, for example, brown hares and rabbits (Fickel et al. 1999; Surridge et al. 1999). Because females care for offspring, a male mountain hare may disperse to avoid mating with his mother (Kerth et al. 2002; Wolff 1993) or to avoid competition with close relatives for mates. A female may require a good knowledge of the resources of her territory in order to provide optimal care for her offspring, hence remaining within natal territories more frequently than males.

Although our evidence suggests that males disperse farther than females, not all of the predicted indices were significant. Although the trend was in the predicted direction, we found no significant difference in FST between male and female cohorts, in contrast to the significant differences in mean assignment index and expected heterozygosity. This may be due to a lack of sufficient power with just 8 loci and relatively small sample sizes. However, in a simulation analysis (Goudet et al. 2002) similar to our study design, that is, a sample of 24 individuals in each of 6 populations screened for 8 loci, both assignment index (mAIc) and FSt had high power (>80%) for a wide range of dispersal rates ranging from 4% to 50%. The present loci displayed mean heterozygosity levels of only 0.55 and this may have resulted in less power. It might be necessary to increase the number of marker loci in order to get a more powerful estimate of biased dispersal with populations that are just moderately diverse. However, there are other potential reasons for the FSt test not being significant, even if there was a real effect. The geographic scale examined here was larger than many other studies of this type (Favre et al. 1997). At greater geographic distances, FST between males and females will tend to converge, whereas the assignment tests are carried out within samples and thus may be more appropriate to this study. The FSt test also assumes that all potential source populations for immigrants have been sampled, which is not assumed by assignment tests; again these are perhaps more appropriate for our data set. Intrinsic factors such as the proportion of migrants and the degree of sex bias in dispersal also may have affected the statistical power. Simulation studies have shown that sex bias is most detectable with microsatellite arrays when it is strong (>70% biased) and when dispersal rates are intermediate (4-30% per capita dispersal rate), rather than very high (Goudet et al. 2002). Additionally, when population differentiation is very low, any bias becomes more difficult to detect, because of the similarity of allelic arrays in all subpopulations (Goudet et al. 2002).

Conclusion

Although there have been changes in the agricultural landscape in Ireland that could cause population fragmentation, the genetic evidence presented indicates that mountain hares currently constitute a relatively panmictic population across southern Ireland. However, because the loci screened were not highly polymorphic, and because there can be a time lag between a fragmentation event and its associated genetic signature, we suggest that population genetic structure in mountain hares in Ireland should continue to be monitored. Although most samples were not significantly differentiated from each other, there was 1 particularly divergent sample. Significant differences in allelic distributions were identified between mountain hares from Balbriggan (BBN), County Dublin (the easternmost sample), and the other samples of the study and genetic distances were high, with FST values reaching 0.112. Mountain hares from this local area may form part of a separate deme. Sampling of areas north and west of Balbriggan could help determine whether the genetic break we identified is due to a localized population, or the presence of 2 or more large demes in southern Ireland. Based on behavioral observations, we tested a hypothesis of sex-biased dispersal between male and female cohorts of the population. Evidence from Bayesian assignment tests using microsatellite data suggests that female mountain hares tend to be more philopatric than males and trends in heterozygosity and differential FST also suggest a bias.

Acknowledgments

We thank the Irish Coursing Club for granting us permission to sample at coursing meetings. We are grateful to T. O'Donoghue, National Parks and Wildlife Service, for assistance with sampling. Thanks to D. Kidd for advice on creating Fig. 1 and thanks also to M. Ritchie, S. Duke, and N. LeBas for thoughtful comments on the manuscript.

Literature Cited

Andersson
A. C
Thulin
C. G.
Tegelstrom
H.
.
1999
.
Applicability of rabbit microsatellite primers for studies of hybridisation between an introduced and a native hare species
.
Hereditas
 
130
:
309
315
.
Angerbjorn
A.
1983
.
Proximate causes of mountain hare population crashes: a review
.
Finnish Game Research
 
41
:
29
38
.
Angerbjorn
A.
Flux
J. E. C.
.
1995
.
Lepus timidus, L
.
Mammalian Species
 
495
:
1
11
.
Animal Care and Use Committee
.
1998
.
Guidelines for the capture, handling, and care of mammals as approved by the American Society of Mammalogists
.
Journal of Mammalogy
 
79
:
1416
1431
.
Austin
J. D.
Davila
J. A.
Lougheed
S. C.
Boag
P. T.
.
2003
.
Genetic evidence for female-biased dispersal in the bullfrog, Rana catesbeiana (Ranidae)
.
Molecular Ecology
 
12
:
3165
3172
.
Barrett-Hamilton
G. E. H.
1912
.
A history of British mammals
 .
Gurney and Jackson
,
London, United Kingdom
.
Berry
O.
Tocher
M.
Gleeson
D.
Sarre
S.
.
2005
.
Effect of vegetation matrix on animal dispersal: genetic evidence from a study of endangered skinks
.
Conservation Biology
 
19
:
855
864
.
Berry
O.
Tocher
M.
Sarre
S.
.
2004
.
Can assignment tests measure dispersal?
Molecular Ecology
 
11
:
551
561
.
Blundell
G. M.
Ben-David
M.
Groves
P.
Bowyer
R. T.
Geffen
E.
.
2002
.
Characteristics of sex-biased dispersal and gene flow in coastal river otters: implications for natural recolonization of extirpated populations
.
Molecular Ecology
 
11
:
289
303
.
Burton
C
Krebs
C. J.
.
2003
.
Influence of relatedness on snowshoe hare spacing behavior
.
Journal of Mammalogy
 
84
:
1100
1111
.
Burton
C
Krebs
C.
Taylor
E.
.
2002
.
Population genetic structure of the cyclic snowshoe hare (Lepus americanus) in southwestern Yukon, Canada
.
Molecular Ecology
 
11
:
1689
1701
.
Coffman
C. J.
Nichols
J. D.
Pollock
K. H.
.
2001
.
Population dynamics of Microtus pennsylvanicus in corridor-linked patches
.
Oikos
 
93
:
3
21
.
Corbet
G. B.
Harris
S.
.
1991
.
The handbook of British mammals
 .
Blackwell
,
London, United Kingdom
.
Cornuet
J. M.
Piry
S.
Luikart
G.
Estoup
A.
Solignac
M.
.
1999
.
New methods employing multilocus genotypes to select or exclude populations as origins of individuals
.
Genetics
 
153
:
1989
2000
.
Dingerkus
S. K.
1997
.
The distribution and ecology of the Irish hare, Lepus timidus hibernicus, in Northern Ireland
.
Ph.D. dissertation
 ,
Queen's University of Belfast
,
Belfast, Northern Ireland
.
Dingerkus
S. K
Montgomery
W. I.
.
2001
.
The diet and landclass affinities of the Irish hare, Lepus timidus hibernicus
.
Journal of Zoology (London)
 
253
:
233
240
.
Dingerkus
S. K
Montgomery
W. I.
.
2002
.
A review of the status and decline in abundance of the Irish hare, L. t. hibernicus, in Northern Ireland
.
Mammal Review
 
32
:
1
16
.
Dobson
F.
1982
.
Competition for mates and predominant juvenile dispersal in mammals
.
American Naturalist
 
30
:
1183
1192
.
Favre
L.
Balloux
F.
Goudet
J.
Perrin
N.
.
1997
.
Female-biased dispersal in the monogamous mammal Crocidura russula: evidence from field data and microsatellite patterns
.
Proceedings of the Royal Society of London, B. Biological Sciences
 
264
:
127
132
.
Fickel
J.
et al
.
2005
.
Genetic structure of populations of European brown hare: implications for management
.
Journal of Wildlife Management
 
69
:
760
770
.
Fickel
J.
Lieckfeldt
D.
Pitra
C.
.
1999
.
Analysis of genetic diversity and structure in neighboring populations of the European brown hare (Lepus europaeus, Pallas 1778)
.
Zeitschrift für Jagdwissenschaft
 
45
:
230
237
.
Flux
J. E. C.
1970
.
Life history of the mountain hare (Lepus timidus scoticus, Hilz.) in north-east Scotland
.
Journal of Zoology
 
161
:
75
123
.
Garnier-Gere
P.
Dillmann
C.
.
1992
.
A computer program for testing pairwise linkage disequilibria in subdivided populations
.
Journal of Heredity
 
83
:
239
.
Goudet
J.
1995
.
FSTAT (version 1.2): a computer program to calculate F-statistics
.
Journal of Heredity
 
86
:
485
486
.
Goudet
J.
Perrin
N.
Waser
P.
.
2002
.
Tests for sex-biased dispersal using bi-parentally inherited genetic markers
.
Molecular Ecology
 
11
:
1103
1114
.
Greenwood
P.
1980
.
Mating systems, philopatry and dispersal in birds and mammals
.
Animal Behaviour
 
28
:
1140
1162
.
Hale
M.
Lurz
P.
Shirley
M.
Rushton
S.
Fuller
R.
Wolff
K.
.
2001
.
The impact of landscape management in the genetic structure of red squirrel populations
.
Science
 
293
:
2246
2248
.
Hamill
R. M.
Doyle
D.
Duke
E. J.
.
2006
.
Spatial patterns of genetic diversity across European subspecies of the mountain hare, Lepus timidus L
.
Heredity
 
97
:
355
365
.
Hamilton
W.
1972
.
Altruism and related phenomena, mainly in social insects
.
Annual Review of Ecology and Systematics
 
13
:
1
21
.
Hartl
D. L.
Clark
A. G.
.
1997
.
Principles of population genetics
 .
3
rd ed.
Sinauer Associates, Inc., Publishers
,
Sunderland Massachusetts
.
Hewson
R.
1985
.
Long-term fluctuations in populations of mountain hares (Lepus timidus L.)
.
Journal of Zoology (London)
 
206
:
269
273
.
Hewson
R.
1990a
.
Behavior, population-changes and dispersal of mountain hares (Lepus timidus L.) in Scotland
.
Journal of Zoology (London)
 
220
:
287
309
.
Hewson
R.
1990b
.
Interactions between mountain hares (Lepus timidus L.) and other mammals
.
Journal of Zoology (London)
 
221
:
302
305
.
Hewson
R.
Hinge
M. D. C.
.
1990
.
Characteristics of the home range of mountain hares, Lepus timidus L
.
Journal of Applied Ecology
 
27
:
651
666
.
Honnay
O.
Jacquemyn
H.
Bossuyt
B.
Hermy
M.
.
2005
.
Forest fragmentation effects on patch occupancy and population viability of herbaceous plant species
.
New Phytologist
 
166
:
723
736
.
Kerth
G.
Mayer
F.
Petit
E.
.
2002
.
Extreme sex-biased dispersal in the communally breeding, nonmigratory Bechtstein's bat (Myotis bechsteini)
.
Molecular Ecology
 
11
:
1491
1498
.
Korstanje
R.
et al
.
2003
.
Mapping of rabbit microsatellite markers using chromosome-specific libraries
.
Journal of Heredity
 
94
:
161
169
.
Manel
S.
Berthier
P.
Luikart
G.
.
2002
.
Detecting wildlife poaching: identifying the origin of individuals with Bayesian assignment tests and multilocus genotypes
.
Conservation Biology
 
16
:
650
659
.
Monard
A.
Duncan
P.
.
1996
.
Consequences of natal dispersal in female horses
.
Animal Behaviour
 
52
:
565
579
.
Paetkau
D.
Calvert
W.
Stirling
I.
Strobeck
C.
.
1995
.
Microsatellite analysis of population-structure in Canadian polar bears
.
Molecular Ecology
 
4
:
347
354
.
Park
S.
2001
.
Microsatellite toolkit
 .
Department of Genetics, Trinity College
,
Dublin, Ireland
.
Pierpaoli
M.
Riga
F.
Trocchi
V.
Randi
E.
.
1999
.
Species distinction and evolutionary relationships of the Italian hare (Lepus corsicanus) as described by mitochondrial DNA sequencing
.
Molecular Ecology
 
8
:
1805
1817
.
Preston
J.
Prodohl
P.
Portig
A.
Montgomery
W. I.
.
2002
.
The Northern Ireland hare survey 2002
 .
Queen's University of Belfast
,
Belfast, Northern Ireland
.
Prugnolle
F.
de Meeus
T.
.
2002
.
Inferring sex-biased dispersal from population genetic tools: a review
.
Heredity
 
88
:
161
165
.
Rannala
B.
Mountain
J. L.
.
1997
.
Detecting immigration by using multilocus genotypes
.
Proceedings of the National Academy of Sciences
 
94
:
9197
9201
.
Raymond
M.
Rousset
F.
.
1995a
.
GENEPOP (version 1.2): population genetics software for exact tests and ecumenicism
.
Journal of Heredity
 
86
:
248
249
.
Raymond
M.
Rousset
F.
.
1995b
.
An exact test for population differentiation
.
Evolution
 
49
:
1280
1283
.
Rice
W. R.
1989
.
Analyzing tables of statistical tests
.
Evolution
 
43
:
223
225
.
Rico
C
RIco
I.
Webb
N.
Smith
S.
Bell
D.
Hewitt
G.
.
1994
.
Four polymorphic microsatellite loci for the European wild rabbit, Oryctolagus cuniculus
.
Animal Genetics
 
25
:
367
.
Rousset
F.
Raymond
M.
.
1995
.
Testing heterozygote excess and deficiency
.
Genetics
 
140
:
1413
1419
.
Schneider
S.
Roessli
D.
Excoffier
L.
.
2000
.
Arlequin: a software for population genetics data analysis
.
Version 2.000
 .
Genetics and Biometry Lab, Department of Anthropology, University of Geneva
,
Geneva, Switzerland
.
Slatkin
M.
1993
.
Isolation by distance in equilibrium and non-equilibrium populations
.
Evolution
 
47
:
264
279
.
Spong
G.
Creel
S.
.
2001
.
Deriving dispersal distances -from genetic data
.
Proceedings of the Royal Society of London, B. Biological Sciences
 
268
:
2571
2574
.
Storz
J.
1999
.
Genetic consequences of mammalian social structure
.
Journal of Mammalogy
 
80
:
553
569
.
Suchentrunk
F.
et al
.
1999
.
Spatial partitioning of allozyme variability in European mountain hares (Lepus timidus): gene pool divergence across a disjunct distributional range?
Zeitschrift für Säugetierkunde
 
64
:
308
318
.
Surridge
A. K.
Bell
D. J.
Hewitt
G. M.
.
1999
.
From population structure to individual behaviour: genetic analysis of social structure in the European wild rabbit (Oryctolagus cuniculus)
.
Biological Journal of the Linnean Society
 
68
:
57
71
.
Thulin
C.
2003
.
The distribution of mountain hares, Lepus timidus, in Europe: a challenge from brown hares, L. europaeus
.
Mammal Review
 
33
:
29
42
.
Tosh
D.
Towers
R.
Preston
J.
Portig
A.
McDonald
R.
Montgomery
W.
.
2004
.
The Northern Ireland hare survey 2004
 .
Queen's University of Belfast
,
Belfast, Northern Ireland
.
Tosh
D.
Towers
R.
Preston
J.
Portig
A.
McDonald
R.
Montgomery
W. I.
.
2005
.
Northern Ireland Irish hare survey 2005
 .
Queen's University of Belfast
,
Belfast, Northern Ireland
.
van Haeringen
W. A.
den Bieman
M.
van Zutphen
L. F. M.
van Lith
H. A.
.
1996
.
Polymorphic microsatellite DNA markers in the rabbit (Oryctolagus cuniculus)
.
Journal of Experimental Animal Science
 
38
:
49
57
.
Wallner
B.
Huber
S.
Achmann
R.
.
2001
.
Non-invasive PCR sexing of rabbits (Oryctolagus cuniculus) and hares (Lepus europaeus)
.
Mammalian Biology
 
66
:
190
192
.
Watson
A.
Hewson
R.
.
1973
.
Population densities of mountain hares on western Scottish and Irish moors and on Scottish hills
.
Journal of Mammalogy
 
170
:
151
159
.
Weir
B. S.
Cockerham
C. C.
.
1984
.
Estimating F-statistics for the analysis of population structure
.
Evolution
 
38
:
1358
1370
.
Whelan
J.
1985
.
The population and distribution of the mountain hare (Lepus timidus L.) on farmland
.
Irish Naturalists' Journal
 
21
:
532
534
.
Wolfe
A.
Hayden
T. J.
.
1996
.
Home range sizes of Irish mountain hares on coastal grassland
.
Biology and Environment: Proceedings of the Royal Irish Academy
 
96B
(
3
):
141
146
.
Wolff
J.
1993
.
What is the role of adults in mammalian juvenile dispersal?
Oikos
 
68
:
173
176
.
Wright
S.
1931
.
Evolution in Mendelian populations
.
Genetics
 
16
:
97
159
.
Wright
S.
1969
.
Evolution and the genetics of populations
 .
University of Chicago Press
,
Chicago, Illinois
.
Wright
S.
1978
.
Variability within and among natural populations
 . Vol.
4
.
University of Chicago Press
,
Chicago, Illinois
.

Author notes

Present address of RMH: Teagasc, Ashtown Food Research Centre, Ashtown, Dublin 15, Ireland
Associate Editor was Eric A. Rickart.