Introduction

Species are distributed along environmental gradients (Begon et al. 1996), and gene frequencies may change locally when adaptations develop in populations by natural selection (Endler 1992; Strickberger 1996). However, locally adapted populations admix to an increasing degree because of range shifts in many species associated with the present use of land and recent climatic changes, especially in temperate areas (IPCC 2001, 2007). In addition, hybridisation rates have increased worldwide because of human-mediated translocations and habitat modifications, causing the extinction of native species, subspecies and locally adapted populations (Allendorf et al. 2001). When genetically different taxa hybridise, local adaptations may be lost from the native taxa by introgression of non-indigenous alleles and loss of local alleles and co-adapted gene complexes (Rhymer and Simberloff 1996; Barton 2001; Burke and Arnold 2001). Gene flow between populations in different environments may therefore constrain local adaptation and lower the short-term fitness of native populations (Storfer 1999). Alternatively, increasing levels of genetic variation from an isolate break (Hartl and Clark 1997) may have positive consequences for population viability through heterosis effects or reduced inbreeding depression (Frankham 1995; Coulson et al. 1998), depending on the genetic divergence of the hybridising taxa (Allendorf et al. 2001; Freeland 2005).

During recent centuries, game management has involved translocations of non-native individuals of many species into former habitats or native populations (Griffith et al. 1989; Hartl 1991; Wolf et al. 1996; Fischer and Lindenmayer 2000; DeYoung et al. 2003; Kruckenhauser and Pinsker 2004). In Europe translocations have been especially common among red deer (Cervus elaphus) populations to re-establish or reinforce local populations and avoid local extinction (Strandgaard and Simonsen 1993; Hartl et al. 1995; Zachos et al. 2003) or to transfer desirable traits for trophy hunters (Hartl et al. 2003). Many of these populations are morphologically different and have been described as separate subspecies (Lønnberg 1906; Whitehead 1972; Whitehead 1993). Many are genetically differentiated (Gyllensten et al. 1983; Kuehn et al. 2003), and even though some argue for one common European subspecies (Groves and Grubb 1987; Polziehn and Strobeck 2002), western and eastern (including Hungarian) European red deer form separate mtDNA haplogroup lineages (Ludt et al. 2004; Skog et al. 2009). The impact of such translocations should be evaluated, considering the growing knowledge of ungulate population genetics and phylogeography (Randi 2005).

The red deer have existed in Norway at least since the sub-boreal period (Ahlèn 1965) and low levels of genetic variation documented by allozyme and microsatellite analyses suggest long-time isolation and previous bottlenecks (Gyllensten et al. 1983; Røed 1998; Haanes et al. 2010). Historically, red deer were distributed across most of southern Norway (Friis 1874; Collett 1877) but after 1750 a major decline limited the population to only a few locations along the west coast (Collett 1909; Ingebrigtsen 1924). To our knowledge only one translocation of non-indigenous red deer into the Norwegian population has occurred in recent times. On the island Otterøya in the northern range of its distribution, the local red deer stock was almost extirpated in 1898, counting only 12–14 individuals including three or four stags (Collett 1909; Collett 1912). To avoid local extinction, 17 captive red deer of a cross breed between the German red deer and the plain—living Hungarian red deer (C. e. hippelaphus), including at least one stag, were translocated from the royal enclosure in Thüringen in central Germany to Otterøya and into the native Norwegian subspecies (C. e. atlanticus) from 1900 to 1903 (Die-Woche 1902; Collett 1909; Finsberg 1934). Among the potential parental origins for the Otterøya stock, the Norwegian, German and Hungarian red deer populations lie at very different latitudes (northern boundaries at 64.5, 54 and 48 N), and offer quite different climates and habitats (Mysterud et al. 2002; Welzholz et al. 2005; Matrai et al. 2004). Further, the western and eastern European red deer lineages are separated by the barriers provided by the Alps and the Carpathians (Hartl et al. 1995; Ludt et al. 2004; Skog et al. 2009), and the introduced German/Hungarian red deer are therefore presumably a mix of western and eastern lineage. Ten years after, the Otterøya population counted 100 individuals (Collett 1909) and since the 1930’s culls have increased considerably to an annual cull of 319 in 2006. Both Hungarian and German red deer are reported to have larger body and antler sizes than the Norwegian subspecies (Lønnberg 1906; Collett 1909; Haigh and Hudson 1993).

We have quantified the level of gene flow between native and introduced continental red deer originating from each of the two major lineages in the European red deer. The success of the German/Hungarian cross was evaluated from the genetic impact on the Otterøya population, and recent gene flow into the mainland population was estimated. We have also assessed the performance of red deer from the mixed stock on the island Otterøya compared with pure stocks from both the mainland and another island (Hitra) in the region. We used body mass as a proxy for performance, which for Norwegian red deer is closely correlated with age of first reproduction (Langvatn et al. 2004) and survival during the first critical winter (Loison et al. 1999).

Materials and methods

Study area

The island of Otterøya (64.5°N and 11.3°E, 143 km2) is separated from the mainland by 1–2 km wide sounds that narrows to a few hundred metres in some places. Annual mean precipitation is 1,440 mm and annual mean temperature is 3.7°C. A large area is situated above the tree line (~400 m) and an area of 98.5 km2 qualifies as suitable red deer habitat. Boreal rainforest characterised by Norway spruce (Picea abies) and Scots pine (Pinus Sylvestris) dominates and cultivated areas are scattered (~4 km2). The island constitutes winter pasture for 600 reindeer and has scarce stocks of roe deer (Capreolus capreolus) and moose (Alces alces).

Sampling and genetic analyses

Between 2001 and 2003 muscle tissue was sampled from 20 Hungarian red deer, 40 red deer from Otterøya and 136 red deer from adjacent (mainly mainland) areas in Norway. Since most animals were shot by local hunters, few reliable body mass estimates were available. We were unfortunately not able to attain any samples from Germany. The German population has been subject to bottlenecks and translocations (e.g. Kuehn et al. 2003; Zachos et al. 2007) and to obtain contemporary samples representative of the original indigenous German part of the introduced cross breed may not be possible. In Hungary, wild red deer were sampled in the north at Valkó close to Gödöllö (n = 6), and in the south at Hajósszentgyörgy close to the Gemenc forest (n = 7) and at Vajszló south of the city Pécs (n = 7, Fig. 1). The Valkó population is spatially separate from the two other sampling sites but 40 years ago several individuals were translocated to this area from the south (pers. comm. Làszló Szemethy, St. Stephen University, Hungary). Indigenous Norwegian red deer were sampled for genetic analyses in six municipalities adjacent to Otterøya (Fig. 1), which are 83–236 km (mean = 151; SE = 22) from Otterøya. The genetic variation of these Norwegian municipalities, except P4-1721, has been previously described (Haanes et al. 2010).

Fig. 1
figure 1

Genetic sampling localities of wild red deer (a), in Hungary (b) at Valkó (1), Hajósszentgyörgy (2) and Vajszló (3), and Otterøya and indigenous localities in Norway (c), sampled separately for either the genetic analysis (O), the body mass analysis (asterisk), or both analysis (filled circle)

Full size image

Genomic DNA was isolated from whole blood and muscle tissue (Qiagen DNeasy KIT). We selected 14 polymorphic microsatellite loci that show Mendelian heredity in Norwegian red deer (Haanes et al. 2005). These were CSSM03 (Moore et al. 1994), OarCP26 (Ede et al. 1995), RT5 (Wilson et al. 1997), SRCRSP10 (Bhebhe et al. 1994), NVHRT73 and NVHRT48 (Røed and Midthjell 1998), McM58 (Hulme et al. 1994), OarFCB193 and OarFCB304 (Buchanan and Crawford 1993), BM5004, BM888, BMC1009, BM4208 and BM4107 (Bishop et al. 1994). They were amplified on a GeneAmp PCR System 9600 (Applied Biosystems) in 10 μl reaction mixtures with 30–60 ng of genomic template DNA, 2 pmol of each primer, 50 mM KCl, 1.5 mM MgCl2, 10 mM Tris–HCl, 0.2 mM dNTP, and 0.5 U of AmpliTaq DNA polymerase (Applied Biosystems). After denaturation at 94°C for 5 min, 30 cycles of amplification with 1 min at 95°C, 30 s at 55°C and 1 min at 72°C were followed by 10 min extention at 72°C. The PCR products were then separated by size with capillary electrophoresis (ABI310) and electromorphs genotyped with GENOTYPER1.1.1 (both Applied Biosystems).

For a subsample of each population (n = 18 and n = 16 and from Hungary and Otterøya, respectively, and n = 17 from Norway, including two areas south of the sample range) a 463 base pair region of the mitochondrial D-loop adjacent to the tRNA pro gene was amplified using the primers 5′-AATAGCCCCACTATCAGCACCC (L15394) and 5′-TATGGCCCTGAAGTAAGAACCAG (H15947) (c.f. Flagstad and Roed 2003). Thirty-five cycles of amplification with 30 s at 94°C, 30 s at 60°C and 45 s at 72°C were preceded by a 2 min pre-denaturation step and followed by a final 7 min extension. Amplifications were performed in 10 μl volumes containing 1.5 mM MgCl2, 200 μM of each dNTP, 4 pmol of each primer and 0.5 units of AmpliTaq DNA polymerase (Applied Biosystems). PCR products were purified using ExoZapitTM (Amersham Biosciences). Sequencing was performed using BigDye terminator cycle sequencing chemistry on an ABI 3100 instrument, and sequences aligned manually using SeqScape version 2.0 (Applied Biosystems). The sequences have been reported to Genbank (accession nos. GU369821–GU369871).

Statistical analyses of genetic data

Random mating within populations was assessed by exact tests of Hardy–Weinberg equilibrium across the 14 microsatellite loci using GENEPOP 3.4 with default settings (Raymond and Rousset 1995). Significance levels were sequentially Bonferroni adjusted for repeated tests (Rice 1989). To investigate genetic variation we calculated the number of private alleles, allele richness (El Mousadik and Petit 1996) and gene diversity (Nei 1987) in each population across loci using FSTAT 2.9.3 (Goudet 2001). Genetic variation in the control region of mtDNA was expressed as haplotype number, haplotype diversity (h) and nucleotide diversity (Π), calculated using ARLEQUIN 2.000 (Schneider et al. 2000). The observed sequences were blasted in Genbank and compared with sequences from red deer sampled in Germany (DQ520209, DQ520210, DQ520214 and DQ520226; Skog et al. 2009) and Hungary (EU4367XX–EU4368XX; personal communication, Dr. Zsolt Pádár, Institute of Forensic Sciences, Ministry of Justice and Law Enforcement, Hungary), and a neighbour joining tree was built using MEGA 4.0 (Tamura et al. 2007).

For the microsatellite data, genetic structure within (Fis) and among (Fst) populations (Weir 1996) was assessed using FSTAT with sequential Bonferroni adjustment. Genetic distances, DA (Nei et al. 1983), among the populations (with Norwegian localities separated) were calculated and a neighbour-joining tree built with 1,000 bootstraps on loci using POPULATIONS (Langella 2002). The tree was visualised using TREEVIEW1.6.6 (Page 1996).

To address the degree of interbreeding between the continental red deer and the native island population at the time of the introduction we used the microsatellite data and estimated the proportionate admixture from two of the parental populations into the Otterøya population using present day Norwegian and Hungarian populations as representatives. We used ADMIX1.0 (Bertorelle and Excoffier 1998) to calculate bootstrap estimates of two admixture estimators (1,000 replicates), the allele frequency based mC (Chakraborty et al. 1992) and the coalescent based mY (Bertorelle and Excoffier 1998). To include the possible affect of genetic drift on the admixture estimates we also estimated admixture using LEA (Chikhi et al. 2001) with 200,000 Monte Carlo Marcov Chain (MCMC) iterations.

To assess the genetic structure of the data set, and to address recent gene flow between the Norwegian mainland and Otterøya, we used the microsatellite data and Bayesian individual assignment as implemented in STRUCTURE2.3 (Pritchard et al. 2000). For each of a different number of genetic clusters (K ∈ [1,11]), a model with uniform priors, an admixture model (α = 1, αmax = 50), correlated allele frequencies (Falush et al. 2003), allowing missing values to be interpreted as null alleles (RECESSIVE ALLELES = 1), 100,000 burnins cycles and 500,000 MCMC iterations was run 10 times. Delta K, which is negatively related to the increasing variance among repeated runs and often increasing posterior probabilities with higher K values, was used to determine the main genetic structure of the data set (Evanno et al. 2005). Dispersers were identified from individuals assigned to clusters other than the one corresponding with where they had been sampled.

Data on red deer body mass

The data on red deer performance derive from a separate data set from the annual Norwegian autumn harvest lasting from 10 September to 15 November in the period 1965 to 2006, for which we do not have genetic data. The date and location (municipality) of harvest together with biological information on sex, body mass and the mandibles for each of 20,161 deer were provided by hunters. Body mass was recorded as dressed mass, which is live mass minus head, skin, viscera, bleedable blood and metapodials, constituting about 58% of live mass and highly correlated with total mass. Using the mandibles, calves and yearlings were aged based on patterns of tooth eruption (Loe et al. 2004). Older animals were aged using annuli in the cementum of the first incisor (Hamlin et al. 2000). The data used are a subset of a larger dataset from the whole of the southwest coast (e.g., Mysterud et al. 2001a, b; Pettorelli et al. 2005). A subset of nine municipalities was selected according to distance from the focal area of Otterøya (Fig. 1). Harvest data may be prone to bias in some cases due to hunter selectivity (see in depth discussion in Mysterud et al. 2008). However, as the tradition for hunting is similar between our focal areas, this is unlikely to be important for the spatial contrast of focus here.

We have good knowledge of the performance of these populations and how they have been affected by density dependence and by climate (the North Atlantic Oscillation (NAO); Mysterud et al. 2001a). As a measure for density, we used the proxy “number of harvested animals” per km2 of red deer habitat (Table S1). Despite this being a crude index, the increase in harvest has been fivefold due to a huge density increase, and over time the measure has shown to correlate well with other direct measures of density (cfr. Mysterud et al. 2007). As an index of climate, we used the station-based winter (Dec-Mar) index of Hurrell (1995).

Statistical analyses of body mass data

We analysed variation in (ln) body mass of red deer with a combination of additive and linear models. Based on previous results, we ran separate models on males and females due to their strongly different life histories (Mysterud et al. 2001b). In addition, it is particularly important to model the age effect correctly, since so much of the variation is found in that parameter. Therefore, and since we are not directly interested in the age effect, we tried both age as a class variable (0, 1, 2, 3, 4, ≥5 years; providing good fit for females, Mysterud et al. 2001b) and modelled the age effect with smoothing splines (for males; Yoccoz et al. 2002), which are very flexible, using the library (mgcv) in (Wood 2006). Similarly, for any growth or decay during the autumn, we used a spline for the “date of harvest” effect to make sure this did not bias our results. We allowed the program to choose the appropriate number of spline knots. Due to age-dependent effort in the rutting of males, we ran this as an interaction between age and date of harvest (cfr. Yoccoz et al. 2002).

Our focal factor is the spatial contrasts between Otterøya and adjacent areas. We therefore used “Treatment” contrasts, i.e., comparing levels of a factor with one specific level—a reference level (the Otterøya stock). However, environmental conditions may not be comparable, so that any spatial difference may not be due to genetic effects, but rather reflect either density or habitat quality. We therefore entered the density index (described above) and the NAO (climate index) to control for annual fluctuations.

Results

Genetic variation

In each investigated population all microsatellite loci were in Hardy–Weinberg equilibrium after sequential Bonferroni adjustment, except BM4208 in the Hungarian population and BM5004 in the Hungarian and Otterøya populations. Across the 14 microsatellite loci, 151 alleles were found in the three investigated populations. Among these, 73 alleles were population specific for either the indigenous Norwegian (7), Otterøya (13) or Hungarian (56) population (Tables 1 and 2). Another 25 microsatellite alleles found in the Otterøya population were shared with only either the Hungarian (16) or the Norwegian (9) population (Table 1), strongly suggesting inheritage from both populations. The Hungarian red deer had by far the highest gene diversity, allele richness and number of private alleles, the Otterøya population was intermediate, while the indigenous Norwegian population was the least genetically variable (Table 2).

Table 1 Observed allele frequencies (a-r) for 14 microsatellite loci (L) in Hungarian (H), a hybrid population at Otterøya (O) and in the indigenous Norwegian red deer (N)
Full size table
Table 2 Genetic variation in microsatellites and mtDNA sampled from red deer of the Hungarian, indigenous Norwegian (No) and Otterøya populations
Full size table

All together, 13 different mtDNA haplotypes were found; two on Otterøya, five in the indigenous Norwegian population and seven in the Hungarian population (Table S2). The nucleotide diversity was intermediate in the Otterøya population while the haplotype diversity was the lowest (Table 2). The two haplotypes in the Otterøya stock were evenly distributed (nA = 9, nB = 7) and both most probably belong to the western mtDNA lineage of European red deer. One of the Otterøya haplotypes was also found in the municipality P3-1563 in the indigenous Norwegian population (nA = 4). It matched a Genbank entry (AF291888), and only differed in one base compared to another indigenous Norwegian haplotype. The other Otterøya haplotype was indigenous to the island, but differed only in two succeeding bases (G–T vs. T–G) from a German sequence (DQ520226) and in two other bases from an observed Norwegian haplotype. Even though it also only differed in two separate base pairs from one Hungarian sequence (EU436831), it differed from all the other Hungarian haplotypes in more than seven different base pairs. Several matches were found in Genbank for one of the other Norwegian haplotypes (e.g. EF636730), as well as for four of the Hungarian haplotypes (e.g.; EU436833, EU436837, EU436838, EU436788). All sampled Hungarian haplotypes except one had from one to three Thymine inserts and grouped out with high bootstrap support (99) in a neighbour-joining tree (not shown), separate from all the haplotypes sampled in North West Europe (including both Otterøya sequences), among which there was no obvious pattern.

The microsatellite data demonstrated limited gene flow and strong genetic structure from significant Fst values between the Otterøya population and both the Hungarian (0.13) and the indigenous Norwegian population (0.19). Similarly, long genetic distances (DA) of 0.40 and 0.33, respectively, showed that the Otterøya population was genetically different and intermediate to both investigated parental populations (Fig. 2). By comparison, a higher Fst value (0.23) and a longer genetic distance (0.48) was found between the Norwegian and Hungarian populations. Within the indigenous Norwegian population much shorter genetic distances (Fig. 2) and lower Fst values (mean = 0.03, SD = 0.02) indicate low genetic structure. An inbreeding coefficient that was not significantly different from zero indicated that there was little genetic structure within the Otterøya population.

Fig. 2
figure 2

Unrooted Neighbour Joining tree based on pairwise DA-distances between Hungarian, indigenous Norwegian and red deer of a possible hybrid population at Otterøya. Bootstrap value of main branching is 100% (1000 replicates)

Full size image

The bootstrap estimate of admixture based on microsatellite allele frequencies indicated an even admixture between indigenous Norwegian and Hungarian red deer into the Otterøya population with proportions of mC = 0.55 (SE = 0.06) and mC = 0.45 (SE = 0.06) admixed from the Norwegian and Hungarian populations, respectively. The coalescent based estimator indicated a higher proportion admixed from the Norwegian population with mY = 0.69 (SE = 0.05) and a lower proportion from the Hungarian population with mY = 0.31 (SE = 0.04). The admixture estimate including genetic drift showed more skewed proportions, with 0.85 from the Norwegian (P1) and 0.1 from the Hungarian population (Fig. 3).

Fig. 3
figure 3

Frequency histogram of LEA results of proportion of admixture (P1) of Hungarian red deer (a) and Norwegian red deer (b) into the hybrid population on Otterøya with 500000 iterations

Full size image

In the STRUCTURE analysis, a pronounced higher delta K value for K = 2 suggest that the main genetic structure of the data set is a dichotomy (Table S3), with the Hungarian and Otterøya samples assigned to one cluster and the indigenous Norwegian samples to the other cluster. A slightly higher posterior probability for K = 4 reflects an underlying lower hierarchical structure between; 1) the Otterøya hybrid stock and the Norwegian population, and 2) within the Norwegian population (Table S3 and Fig. 4). The STRUCTURE analyses (K = 2) show a very limited recent gene flow between the Norwegian mainland and Otterøya populations, with all individuals assigned with high probabilities to the cluster corresponding to where they had been sampled, except for three individuals. Two of the individuals sampled on Otterøya were assigned to the Norwegian cluster (0.55 < q < 0.99) while one individual sampled in the Norwegian population was assigned to the Otterøya/German/Hungarian cluster (P = .69). These are most probably first or second generation migrants.

Fig. 4
figure 4

Individual posterior probabilities (y-axis) of Bayesian assignment (STRUCTURE) with two to four clusters (K ∈ [2, 4]; different colours) among Hungarian (Hun), Otterøya hybrids (P4) and red deer from six Norwegian indigenous localities (separated by vertical lines)

Full size image

Body mass variation

The models of female (r2 = .853) and male (r2 = .724) body mass variation gave largely similar results concerning the spatial contrast, after controlling for the strong effects of age, date of harvest, density and the NAO (Table 3). Deer from Otterøya were larger than deer from coastal municipalities in Sør-Trøndelag (population P3; municipality no. 1,612, 1,613 and 1,622) including the island of Hitra (population P5), but not larger than all the inland municipalities (municipality no. 1,635 and 1,636). Body mass of red deer from Otterøya was comparable to body mass of those from the mainland municipalities (municipality no. 1,714 and 1,721) in Nord-Trøndelag (population P4). The interaction term between age and municipality could not be entered due to unbalanced data, however, similar differences between municipalities were obtained when only analysing variation for a specific age class (yearlings).

Table 3 Analysis of body mass of red deer from Sør-Trøndelag (mainland-P3; island Hitra-P5) and Nord-Trøndelag (mainland P4, island Otterøya) using a model with both linear and additive (spline; df = 3) terms
Full size table

Discussion and conclusions

This study demonstrates that the German/Hungarian red deer introduced onto Otterøya around 1900 interbred with the resident native Norwegian population. Variation in both microsatellite and mtDNA demonstrate that it is a genetic intermediate with heritage from both the native Norwegian and the Hungarian population. Extensive interbreeding between these presumed subspecies (Lønnberg 1906; Whitehead 1972; Whitehead 1993) was evident from the estimators of admixture and from the similar frequencies of the two haplotypes found on Otterøya. However, any negative effects of introgression that may have been expected were not observed in either population growth or body mass. Rather the hybrid stock has grown tremendously in numbers during recent decades and the observed body mass of red deer on Otterøya was similar or larger than those of indigenous Norwegian inland and coastal localities.

Potential limitations and biases on the admixture estimates

For the microsatellite data, the allele frequency-based estimator yielded close to even proportions of admixture from the indigenous Norwegian and Hungarian populations. However, frequency-based estimators are often biased towards even proportions compared to the coalescent-based estimator MY, which also incorporates molecular divergence between alleles and parental populations (Bertorelle and Excoffier 1998; Wang 2003). The coalescence-based estimator yielded skewed proportions of admixture, as could be expected from influence by the unaccounted German part of the introduced red deer cross. Considering that the resident population on Otterøya had a similar size and sex ratio as the group of introduced German/Hungarian red deer (Collett 1909; Ingebrigtsen 1924; Finsberg 1934), the German part would constitute one quarter of admixture with free interbreeding. This is roughly proportionate with and could help explain the deviations from even admixture in the coalescence-based estimator and the genetic drift model (LEA). The private alleles found only on Otterøya probably represent the German contribution but, considering the relatively low sample size, could also be of Hungarian origin. The private alleles and the alleles shared only by Otterøya and Hungarian samples may also explain why these cluster together with K = 2 in the STRUCTURE analysis, since private alleles and large allele frequency differences have substantial effects on the clustering of STRUCTURE (Rosenberg et al. 2001; Musani et al. 2006). The even frequencies of the two mitochondrial haplotypes on Otterøya, which were of Norwegian and non-indigenous origin, provided support for an even admixture. The non-indigenous haplotype was relatively similar to both known German and Hungarian sequences, but since it fell together with the previously described West European haplogroup rather than the East European haplogroup (Ludt et al. 2004; Skog et al. 2009), it most likely represents a German lineage in the introduced cross breed. The even admixture in both mtDNA and microsatellite data further indicates similar contributions from both sexes. In spite of high Fst values and long genetic distances between the parental populations, our data support an even admixture and free interbreeding between the introduced German/Hungarian red deer and the native Norwegian population.

Performance of the hybrid stock

Hybridisation among subspecies may have different outcomes. It is rare in some species (Hansen 2002; Lorenzen and Siegismund 2004; Storfer et al. 2004), but happen to various degrees in others (Jensen et al. 2005). Hybridisation is common within the red deer species complex (Hartl et al. 1995; Tate et al. 1998; Hartl et al. 2003), but negative effects are known between genetically and morphologically different taxa like the wapiti and the red deer (Asher et al. 2005). Examples from other species involve loss of local native genetic variation (Shaw et al. 1992; Latch et al. 2006), loss of native adaptations (Storfer 1999), and reduced fitness (Allendorf et al. 2004). However, intentional translocations of species for conservation and management reasons have been common, and have yielded positive results in many species (Griffith et al. 1989; Fischer and Lindenmayer 2000), for example in the Scandinavian wolf (Vila et al. 2003), wild boar (Vernesi et al. 2003) and the Florida Panther (Hedrick 2001). The Norwegian, German and Hungarian red deer populations are located at different latitudes (northern boundaries at 64.5, 54 and 48°N) and genetic variation indicates that the indigenous Norwegian population has been isolated for a long time (Gyllensten et al. 1983; Røed 1998). It therefore seems reasonable to assume that these populations may have developed different local adaptations. However, the negative effects on population viability that may be expected when taxa with different local adaptations hybridise (Rhymer and Simberloff 1996; Allendorf et al. 2001; Burke and Arnold 2001) were not observed in either the population growth or the body mass of the Otterøya hybrid stock. Even though this natural experiment does not offer an adequate evolutionary time frame, our results support the high phenotypic plasticity suggested for red deer (Geist 1998; Lister 2004), rather than different local adaptations in different environments. Ten years after the introduction the Otterøya population counted 100 individuals (Collett 1909). Since then annual culls have increased considerably, reaching 319 in 2005 and 2006. This follows the general trend of expansion of the Norwegian population (Forchhammer et al. 1998; Langvatn, 1998), which is partly explained by climatic variation (Forchhammer et al. 1998; Mysterud et al. 2001a), changes in human harvesting (Langvatn and Loison 1999) and the altered use of agricultural land (Ahlèn 1965; Mysterud et al. 2002). The increase in population size of the hybrid stock could also reflect positive effects on population viability, as could be expected from heterosis and reduced inbreeding involved with hybridisation of more closely related taxa (Haig 1998; Freeland 2005). The increased level of genetic variation in the small initial founding population on Otterøya may thus have prevented the negative effects of inbreeding and counteracted loss of genetic variation from random genetic drift. In red deer, heterosis effects have been documented as both increased lifetime reproductive success and calf body mass (Coulson et al. 1998; Slate et al. 2000), and could explain the heavier body mass on Otterøya compared to most indigenous localities. However, German and Hungarian red deer have a relatively larger body mass than Norwegian red deer (Lønnberg 1906; Collett 1909; Haigh and Hudson 1993) and may indicate effects from additive genetic variation. On the other hand, much geographic variation in body size is attributable to phenotypic plasticity affected by habitat and nutrition (Lister 1984; Geist 1998), as demonstrated by the huge increase in body size and antlers of west European red deer after translocation to New Zealand (Huxley 1931). Further, red deer body mass is generally strongly negatively related to density (Mysterud et al. 2001b) and the higher density on Otterøya may have obscured differences to the inland localities with apparently similar body mass. These comparisons were difficult because of the lack of adequate data on habitat quality, as density relative to resource levels is expected to determine body mass.

Implications for management

Generally, management is concerned with conservation of local biodiversity and indigenous genetic variation (Rhymer and Simberloff 1996; Storfer 1999; Allendorf et al. 2001). Even though no first-generation migrants from the Otterøya population were detected in the mainland population in 2001 and 2002, we observed very low frequencies of some alleles that were only common in the Hungarian and Otterøya populations. Otterøya is separated from the mainland by only a 200-300 meter wide sound, and these alleles are probably the result of introgression into the mainland population during 30 generations. Considering the recent range shifts of many species (IPCC 2001; IPCC 2007), and the population expansion of the Norwegian red deer last century (Forchhammer et al. 1998; Langvatn 1998), some dispersal from the Otterøya seems very likely. Until the effects of heterosis (positive) on Otterøya have been further addressed, it will be quite challenging to decide how to manage the Otterøya hybrid stock. On the one hand it can be argued that the genetically different hybrid population on Otterøya, with its higher genetic diversity, should be allowed to expand and interbreed with the native mainland population to increase its level of genetic variation. On the other hand, the Norwegian population might be one of the few European populations not affected by translocations, and it is probably better to be safe than sorry by preventing any spreading and interbreeding with the indigenous Norwegian population.