The translocation of wild animal species became a common practice worldwide to re-establish local populations threatened with extinction. Archaeological data confirm that chamois once lived in the Biokovo Mountain but, prior to their reintroduction in the 1960s, there was no written evidence of their recent existence in the area. The population was reintroduced in the period 1964–1969, when 48 individuals of Balkan chamois from the neighbouring mountains in Bosnia and Herzegovina were released. The main objective of this study was to determine the accuracy of the existing historical data on the origin of the Balkan chamois population from the Biokovo Mountain and to assess the genetic diversity and population structure of the source and translocated populations 56 years after reintroduction. Sixteen microsatellite loci were used to analyse the genetic structure of three source chamois populations from Prenj, Čvrsnica and Čabulja Mountains and from Biokovo Mountain. Both STRUCTURE and GENELAND analyses showed a clear separation of the reintroduced population on Biokovo from Prenj’s chamois and considerable genetic similarity between the Biokovo population and the Čvrsnica-Čabulja population. This suggests that the current genetic composition of the Biokovo population does not derive exclusively from Prenj, as suggested by the available literature and personal interviews, but also from Čvrsnica and Čabulja. GENELAND analysis recognised the Balkan chamois from Prenj as a separate cluster, distinct from the populations of Čvrsnica and Čabulja. Our results thus highlight the need to implement genetic monitoring of both reintroduced and source populations of endangered Balkan chamois to inform sustainable management and conservation strategies in order to maximise the chances of population persistence.

Biokovo, genetic structure, microsatellite, Prenj, translocation

The reintroduction and translocation of wild species for various purposes became a common practice worldwide and was used as a conservation tool for rescuing and re-establishing extirpated populations (Cullingham and Moehrenschlager 2013). Northern chamois (Rupicapra rupicapra L.) is one of the examples of successfully translocated species (Apollonio et al. 2014) in many areas of Europe (Crestanello et al. 2009; Martínková et al. 2012; Šprem and Buzan 2016), but also on other continents such as South America (Corlatti et al. 2011) and New Zealand (Christie 1964). Past translocations of chamois left a genetic signature in recent populations which can be now used for reconstructing undocumented events (Crestanello et al. 2009).

Today’s populations of the Northern chamois in northern Dinaric Mountains in Croatia are descendants of successfully translocated individuals captured on mountain areas in Bosnia and Herzegovina (Rupicapra rupicapra balcanica) and Slovenia (Rupicapra rupicapra rupicapra) (Apollonio et al. 2014; Šprem and Buzan 2016). Since different subspecies were involved in past reintroduction efforts, a contact zone was formed on the northern Velebit Mountains where these subspecies hybridise (Šprem and Buzan 2016).

The Balkan chamois (Rupicapra rupicapra balcanica) is one of the seven recognised subspecies of the Northern chamois. It is found both in the mountainous regions of Croatia and in the mountain ranges of the eight other countries of the Balkan Peninsula, from north to south: Bosnia and Herzegovina, Serbia, Montenegro, Kosovo, North Macedonia, Albania, Bulgaria, and Greece. The lack of continuity of these habitats and overhunting in the post-Neolithic period have severely fragmented the subspecies’ present distribution (Corlatti et al. 2022). In addition to low colonisation rates and reduced gene flow between isolated populations, which may lead to genetic differentiation due to the inbreeding effect and loss of allelic variants, the Balkan chamois is threatened by poaching (Papaioannou and Kati 2007), habitat change (Kavčić et al. 2019), unsustainable hunting (Šprem and Buzan 2016) and by the introduction of Alpine chamois subspecies (Iacolina et al. 2019). As a conservation measure, the Balkan chamois is listed in Annexes II and IV of the European Union Habitats Directive 92/43/EEC (OJ L 206, 22.7.1992) and in Appendix III of the Bern Convention (OJ L 38, 10.2.1982). The Balkan chamois is one of the most poorly studied subspecies of Northern chamois and knowledge on the genetic diversity and structure of the Balkan chamois population is limited and restricted to regional-local studies (Markov et al. 2016; Šprem and Buzan 2016; Papaioannou et al. 2019; Rezić et al. 2022).

The genetic structure of the Balkan chamois population on the Biokovo has been studied only by Šprem and Buzan (2016), and few other ecological studies have included this population in a population density estimation (Kavčić et al. 2021a) and rutting behaviour (Kavčić et al. 2021b). The study of phylogenetic relationships in Šprem and Buzan (2016) revealed the existence of endemic Balkan haplotypes in the Prenj and Biokovo Mountains and a genetic richness of the historically viable Prenj population comparable to Alpine chamois studied in Buzan et al. (2013) from the south-eastern Alps. The paleontological findings in the Baba cave, which are more than ten thousand years old, confirm that chamois once lived in Biokovo (Šabić 2011) but, before the reintroduction in the 1960s, there was no written evidence of the recent existence of chamois in this area. According to historical records, the chamois on Biokovo Mt. are descendants of individuals translocated from the “Prenj” hunting district in Bosnia and Herzegovina established in 1961 (Rapaić and Kunovac 2020) which included both Prenj and Čvrsnica massifs. This hunting district had, at that time, stable and numerous chamois populations and was used for many reintroduction programmes in the Balkans (Rapaić and Kunovac 2020). The reintroduction of chamois on Biokovo Mt. was the result of the planned introduction on this area by the Union of Hunting Association, the municipality of Makarska and the Union of Hunting Association of Imotski, mainly with the aim of increasing the population size for hunting purposes (Šabić 2011). Prior to the reintroduction from the “Prenj” hunting district, assessment of the suitability of the Biokovo habitat was made and, after a positive evaluation, a first release of 7 individuals (3 males and 4 females) took place on 1 November 1964 (Šabić 2014). A total of 48 chamois was successfully reintroduced in the period between 1 November 1964 and 23 October 1969 (Šabić 2014). The success of the reintroduction of the Balkan chamois in the Biokovo Mt. is reflected by the latest population size estimate (Kavčić et al. 2021a), according to which this area is now inhabited by at least 600 individuals.

The main objective of this study was to determine the accuracy of historical data on the origin of chamois in Biokovo, and to assess and document the genetic status of both the source and translocated populations, 56 years after reintroduction, by using microsatellite markers.

The sixteen microsatellite loci yielded a total of 95 alleles, which varied between 2 (for locus ETH10 and SR-CRSP-6) and 10 (for locus BM1258) with an average value of 5.937 alleles per locus (see Suppl. material 1: Table S1).

The values of null allele frequencies were low for most analysed loci, except for loci ETH10, SY434, TGLA53, and SR-CRSP-6, whose frequencies were estimated to be r ≥ 0.20 (see Suppl. material 1: Table S2). The presence of null alleles was found in all three populations from Bosnia and Herzegovina. The various factors caused by natural population mechanisms, such as disassortative mating, bottleneck, fluctuations in population size, can cause heterozygote deficit that can be interpreted as false positive presence of null alleles (Dąbrowski et al. 2014), which led to the decision to retain all analysed loci.

The Prenj population deviated from Hardy-Weinberg equilibrium (HWE) but the deviation was significant at the 0.05 level only for locus SY434 after applying sequential Bonferroni adjustment (Table 1). This may be a consequence of the recent severe bottleneck in this population, which was previously stable. Gafić and Džeko (2009) noted that the population of Balkan chamois in the Prenj hunting district, which included both the Prenj and Čvrsnica massifs, was approximately 4,000 individuals in 1966, but due to the civil war in the 1990s, the population was greatly reduced by illegal hunting, with up to 95% of the population lost. Since no deviation from HWE was observed at this locus in other populations, we retained it in all subsequent analyses. After applying the sequential Bonferroni correction to the linkage disequilibrium results, no significant value was observed.

The Prenj population had the highest values of observed (0.636) and expected (0.637) heterozygosity, and allelic richness (2.517). A similar pattern was recorded in the study of Šprem and Buzan (2016) where the Prenj population had the highest values of allelic richness, observed, and expected heterozygosity and significantly deviated from HWE. In the Šprem and Buzan (2016) study, the Biokovo population had the lowest allelic richness. The observed number of alleles (A) varied from 3.375 in the Čabulja population to 4.625 in the Biokovo population. All populations had private alleles (N_pr) and the highest number of private alleles (11) was observed in the Prenj population from Bosnia and Herzegovina (Table 1).

Effective population size was estimated for three sampled sites, excluding the Čabulja population due to small sample size (Table 1). Čvrsnica had the lowest results for N_e, although very similar to those estimated for Prenj. The higher estimates for the Biokovo population should be taken with caution, considering our results suggest the presence of multiple funding sources. Additionally, Do et al. (2014) showed that microsatellite loci could lead to a slight upward bias for the linkage disequilibrium method when the critical value is set to p = 0.02.

The lowest F_ST value was found between Čvrsnica and Čabulja (F_ST = 0.024), while the highest and significant F_ST value (0.084) was observed between two neighbouring populations from Bosnia and Herzegovina (Prenj and Čvrsnica; Table 2). The global F_ST values were 0.067 (CI = 0.031–0.108) without using the correction method and 0.071 (CI = 0.038–0.112) with the ENA correction method for null alleles. The Wilcoxon Two Sample test showed no significant differences between the corrected and original F_ST values (p = 0.734), indicating that the presence of putative null alleles did not affect the analysis (see Suppl. material 1: Fig. S1).

The algorithm developed by Evanno et al. (2005) identified K = 3 as the optimal number of ancestral genotypes detected by STRUCTURE analysis for four analysed populations and detected another peak at K = 5 suggesting a possible further genetic structure subdivision within the populations (see Suppl. material 1: Fig. S2). According to STRUCTURE results, populations from Prenj, Čvrsnica, and Biokovo had high proportions of genomes from a single ancestral genotype (q ≥ 0.88 in all three populations), which for Prenj differ from Čvrsnica and Biokovo, while in the population from Čabulja the highest proportion of same ancestral genotype as in Čvrsnica was lower (q = 0.63; Fig. 1a). The difference in genetic composition between Balkan chamois from Prenj and other populations is likely due to a barrier to gene flow between the studied populations, but also probably a consequence of recent bottleneck effect, due to extirpation of chamois during the Balkan civil war (Frković 2008) and population local adaptation. Čvrsnica, Čabulja, and Prenj Mts. are restricted to habitat patches, particularly by the steep river canyon of Neretva, which might lead to their fine-scale fragmentation. The Neretva River, one of the largest rivers in the eastern part of the Adriatic Basin, separates the Prenj Mt. from the Čvrsnica and Čabulja Mts. and forms the official border between the “Čvrsnica” and “Prenj” game reserves (Jurić 1998). Landscape features of the Neretva River valley, as well as the adjacent road M17, built as a part of the European Route E73, can act as effective barriers for the species natural dispersal, and hinder gene exchange between the studied populations (Soglia et al. 2010; Buzan et al. 2013). Additionally, the population decline in the last Balkan civil war could have contributed by isolating small groups of chamois to restricted habitat patches, although geographically very close to each other. Individual proportions of ancestral genotypes assigned by STRUCTURE support the higher levels of admixture in Čabulja when compared to the other populations (Fig. 1b). According to the q values, only one individual from Čabulja had a q value above the threshold of 0.75 sharing similar genotype as individuals from the Čvrsnica population, while all others had admixed genotype (q < 0.70). One individual from Čvrsnica stood out from the others with a q = 0.23 proportion of ancestral genotype that was present mostly in Prenj population. As previously mentioned, natural migration between individuals from Prenj and Čvrsnica is currently unlikely, due to the presence of barriers, however this might not have been the case before the construction of the road infrastructure. STRUCTURE indicated that the reintroduced Balkan chamois population in the Biokovo Mt. is genetically more similar to the population from Čabulja, suggesting that the reintroduced individuals in the Biokovo Mt. may originate from this area, as well as from Čvrsnica, whereas animals translocated from Prenj did not leave detectable genetic signatures. According to Jurić (1998), 992 individuals of Balkan chamois were counted in the Čvrsnica hunting ground which also includes the territories of the neighbouring Čabulja Mt. It is possible that populations Čvrsnica and Čabulja were connected in the past and formed a single population, as can be inferred from the results of STRUCTURE. Due to the civil war in the Balkans in the 1990s and continued illegal hunting and poaching, this single population has dwindled in numbers and become fragmented and isolated in the high mountain habitats (statements of local people). The divergence between Prenj and Biokovo from Čvrsnica and Čabulja populations may be due to the historical founder effect and more recent genetic drift due to isolation, and local adaptation. Despite decades of long unsustainable hunting, predation, poaching and natural events (Šprem and Buzan 2016) the inbreeding levels in the analysed populations are still moderate. The results of genetic composition of Biokovo population can influence population viability through time and it is very important to monitor genetic parameters of the reintroduced population to prevent a loss of genetic diversity due to inbreeding and genetic drift (DeMay et al. 2017).

To improve the previous analyses, the spatial context of individuals was taken into consideration and tested with GENELAND. This analysis revealed a similar pattern of clustering of individuals as STRUCTURE, but suggested an additional fourth spatial cluster along the MCMC chain (see Suppl. material 1: Fig. S3). GENELAND detected two spatial clusters among the three analysed populations in Bosnia and Herzegovina, while one spatial cluster corresponded to the population in Biokovo (Fig. 1). The distinction between Prenj and Biokovo was also detected by Šprem and Buzan (2016), who used the BAPs algorithm for spatial clustering of groups and showed the separation of the Balkan chamois population of the Prenj Mt. from the Biokovo Mt. population. The fourth cluster revealed by GENELAND spatial model was a so-called “ghost cluster” (Frantz et al. 2009) since no individual was assigned to it. Ghost clusters are not uncommon but are still a poorly understood phenomenon that can be caused by a heterogeneous distribution of samples (Aziz et al. 2018). It is possible that all the clusters identified by GENELAND represent a true genetic subdivision, but the degree of differentiation between them was too low for the clustering to be consistent (Frantz et al. 2009). Another possibility is that the GENELAND model might overestimate the number of genetic clusters when analysing populations which are affected by isolation by distance (Frantz et al. 2009).

Future studies will need to incorporate non-invasive genetic sampling, telemetry and behavioural patterns to confirm possible migration and gene flow between these populations. In the available literature there is no indication of the exact location where the animals released on Biokovo were caught. It is only known that the individuals came from the Prenj hunting district, which included two game reserves called “Čvrsnica” and “Prenj” (established in 1893 by the Austro-Hungarian Empire) and which were declared protected areas (Rapaić and Kunovac 2020). Jurić (1998) stated in his master’s thesis that the translocation of Balkan chamois from the game reserve “Čvrsnica” began in 1962 and lasted until 1970. During this period, a total of 101 Balkan chamois was translocated to different areas in the Balkans, but the author did not write any additional information about the location where the animals were caught or the places where they were translocated. It is not yet known whether these data exist. Therefore, it is very important to establish standards for documenting and monitoring species translocation projects.

Non-invasive monitoring of genetic parameters of both reintroduced and source populations of endangered Balkan chamois, together with demographic monitoring, is crucial for sustainable management practices and improving conservation strategies to maximise the chances of population persistence. Our genetic diversity results show that the Balkan chamois population from Biokovo can serve as a potential source for future translocations, especially to the source habitats, Čvrsnica and Čabulja, that are currently threatened by loss of genetic diversity due to unsustainable hunting and poaching, leading to inbreeding and genetic drift.

We thank Luca Corlatti for permission to use his drawing (the silhouettes of the chamois yearling). This work was supported by the Croatian Science Foundation [grant number IP 2016-06-5751] and the ResBios European Union’s Horizon 2020 Research and Innovation Program [grant number 872146].