Early Genetic Outcomes of American Black Bear Reintroductions in the Central Appalachians, USA

Early genetic outcomes of American black bear reintroductions in the Central Appalachians, USA

Sean M. Murphy1,7, John T. Hast1,2, Ben C. Augustine1,8, David W. Weisrock3,JosephD.Clark4, David M. Kocka5, Christopher W. Ryan6, Jaime L. Sajecki5, and John J. Cox1

1Department of Forestry and Natural Resources, University of Kentucky, Lexington, KY 40546, USA 2Kentucky Department of Fish and Wildlife Resources, Frankfort, KY 40601, USA 3Department of Biology, University of Kentucky, Lexington, KY 40506, USA 4United States Geological Survey, Northern Rocky Mountain Science Center, Southern Appalachian Research Branch, University of Tennessee, Knoxville, TN 37996, USA 5Virginia Department of Game and Inland Fisheries, Verona, VA 24482, USA 6West Virginia Division of Natural Resources, South Charleston, WV 25303, USA

Abstract: Habitat loss and overexploitation extirpated American black bears (Ursus americanus)from most of the Central Appalachians, USA, by the early 20th Century. To attempt to restore bears to the southwestern portion of this region, 2 reintroductions that used small founder groups (n = 27 and 55 bears), but different release methods (hard vs. soft), were conducted during the 1990s. We collected hair samples from black bears during 2004–2016 in the reintroduced Big South Fork (BSF) and Kentucky– Virginia populations (KVP), their respective Great Smoky Mountains (GSM) and Shenandoah National Park (SNP) source populations, and a neighboring population in southern West Virginia (SWV) to investigate the early genetic outcomes of bear reintroduction. Despite having undergone genetic bottlenecks, genetic diversity remained similar between reintroduced populations and their sources approximately 15 years after the founder events (ranges: AR = 4.86–5.61; HO = 0.67–0.75; HE = 0.65–0.71). Effective population sizes of the reintroduced KVP and BSF (NE = 31 and 36, respectively) were substantially smaller than their respective SNP and GSM sources (NE = 119 and 156, respectively), supporting founder effects. Genetic structure analysis indicated that the hard-released (i.e., no acclimation period) KVP founder group likely declined considerably, whereas the soft-released BSF founder group remained mostly intact, suggesting superior effectiveness of soft releases. Asymmetrical gene ﬂow via immigration from the SWV has resulted in the KVP recovering from the initial founder group reduction. Sustained isolation, small NE, and small population size of the BSF may warrant continued genetic monitoring to determine if gene ﬂow from neighboring populations is established or NE declines. For future bear reintroductions, we suggest managers consider sourcing founders from populations with high genetic diversity and soft-releasing bears to locales that are, if possible, within the dispersal capability of extant populations to mitigate the potential consequences of founder effects and isolation.

Key words: Appalachia, bear, bottleneck, effective population size, genetic diversity, genetic structure, Kentucky, reintroduction, Tennessee, Ursus americanus, Virginia, West Virginia

DOI: 10.2192/URSU-D-18-00011.1 Ursus 29(2):119–133 (2019)

Reintroductions have expedited large carnivore Wolf and Ripple 2018). Human population growth and restoration to historical ranges more rapidly than the rate associated habitat loss are expected to continue increas- at which natural recolonization typically occurs for these ing globally, likely heightening the importance of rein- taxa (Hayward and Somers 2009, Bruskotter et al. 2017, troducing large carnivores to remaining suitable habitats. This may be particularly relevant for ursids, given that 6 of the world’s extant bear species are globally or region- ally imperiled and all bear species occupy fractions of 7email: [email protected] 8Present address: Atkinson Center for a Sustainable Future, their historical ranges (International Union for Conser- Department of Natural Resources, Cornell University, Ithaca, vation of Nature and Natural Resources 2017). Indeed, NY 14850, USA multiple bear reintroductions are ongoing or planned on

119 r 120 GENETICS OF BLACK BEAR REINTRODUCTIONS Murphy et al. multiple continents (Kim et al. 2011, Tosi et al. 2015, BSF ࣈ 228 and KVP ࣈ 482 bears by 2012 and 2013, Ma et al. 2016). To inform future bear-restoration ef- respectively, which indicated short-term reintroduction forts and improve the potential for reintroduction success, success (Murphy 2016). Those studies assumed that pop- studies are needed that evaluate the outcomes of previous ulation growth was solely the product of the founder bear reintroductions (Clark et al. 2002, Clark 2009). groups, but it remains unclear if immigration from other Overexploitation and habitat destruction extirpated bear populations in the region was influential. Although American black bears (Ursus americanus; hereafter, both reintroductions occurred during a similar temporal ‘black bears’) from most of the Central Appalachian period and habitat conditions were similar in the release region of the eastern United States by the early 1900s areas, different release methods were used (hard vs. soft). (Hall 1981). Remnant isolated bear populations persisted The winter soft-release method for bears, which was used in parts of the most rugged and inaccessible mountains for the BSF reintroduction, involves translocating family of the region through the 1950s, including the areas groups (adult females with dependent cubs) from natural now known as Shenandoah National Park (SNP), Great dens to artificial dens during winter months (Eastridge Smoky Mountains National Park (GSM), and the Monon- and Clark 2001). This method is thought to have better gahela National Forest (Pursley 1974, Hall 1981, Virginia potential for population establishment than hard releases Department of Game and Inland Fisheries [VDGIF] because it mitigates the homing tendency of bears and fo- 2012, Unger et al. 2013). Black bears generally have low cuses on adult females, which are arguably the most crit- population growth rates and are considered poor colo- ical cohort to black bear population growth (Eastridge nizers (Clark et al. 2002); consequently, natural recol- and Clark 2001, Wear et al. 2005, Benson and Cham- onization of the Central Appalachians was unlikely to berlain 2007, Beston 2011). However, a direct compari- occur quickly because suitable habitat was patchy and son of the success of hard versus soft releases was pre- considerable distances separated remnant populations cluded, primarily because none of the KVP founders were (van Manen and Pelton 1997). radiomonitored; thus, the settling, survival, and disper- Beginning in the 1970s, state and federal wildlife man- sal rates that were estimated for the soft-released BSF agers and university researchers implemented several founders (Eastridge and Clark 2001) were not quantified black bear reintroductions in an attempt to reestablish for the hard-released KVP founders. bear populations in parts of the Central Appalachians. Population genetic data can elucidate patterns of bear From 1970 to 1997, >600 bears were translocated to population structure, isolation, and immigration and em- areas of the region that had remained devoid of resi- igration (Brown et al. 2009, Nowak et al. 2014). Sev- dent populations for decades. The majority of bears were eral studies have evaluated the genetic effects of natu- translocated from the remnant SNP population and hard- ral and anthropogenic barriers and human exploitation released (i.e., without an acclimation period) in north- on American black bear populations (Dixon et al. 2007; western (n = 300 bears) and south-central (n = 221 bears) Murphy et al. 2017, 2018; Pelletier et al. 2017). Ad- Virginia, along the state’s borders with West Virginia and ditionally, genetic structure analyses have been used to Tennessee, respectively (Fies et al. 1987, Comly-Gericke identify the source populations of natural recolonization and Vaughan 1997; Fig. 1). Two additional reintroduc- by bears (Onorato et al. 2007, Malaney et al. 2018). In tions used much smaller founder groups and occurred comparison, few studies have investigated the genetic concomitantly during the 1990s in the southwestern por- outcomes of bear reintroduction. Puckett et al. (2014) tion of the Central Appalachians. The first consisted of provided the most comprehensive evaluation, studying 55 bears (unknown sex or age ratios) translocated from American black bears in the Central Interior Highlands, SNP and hard-released along the Kentucky–Virginia bor- USA, to investigate the influence of genetic drift and ad- der during 1990–1997 (hereafter, the ‘Kentucky–Virginia mixture on genetic diversity and structure 8 bear gener- Population’ [KVP]; VDGIF 2008, Murphy 2016). The ations after reintroduction. Although the findings from second consisted of 14 adult females with 13 dependent that study were revealing, bear management and human cubs translocated from GSM and soft-released in the Big landscape alterations that could directly influence pop- South Fork area along the Kentucky–Tennessee border ulation genetics often occur at much shorter intervals during 1996–1997 (hereafter, the ‘Big South Fork Popu- than the 5 decades that elapsed between reintroduction lation’ [BSF]; Eastridge and Clark 2001, Murphy 2016). and that evaluation (Allendorf et al. 2008, Samarasin Murphy et al. (2015, 2016) found that both the BSF et al. 2016). Conservationists have urged that genetic and KVP populations grew rapidly since the founder monitoring of reintroduced populations should occur fre- events (λ = 1.14–1.18/yr), reaching population sizes of quently following founder events to afford managers the

Ursus 29(2):119–133 (2019) r GENETICS OF BLACK BEAR REINTRODUCTIONS Murphy et al. 121

Fig. 1. Locations at which hair samples were collected from individual American black bears (Ursus americanus) in 5 populations in the Central Appalachians, USA, during 2004–2016. The reintroduced Big South Fork (BSF) and Kentucky–Virginia (KVP) populations were established during the 1990s using founders from the Great Smoky Mountains (GSM) and Shenandoah National Park (SNP) populations, respectively. We also sampled a neighboring population in southern West Virginia (SWV), but we did not sample 2 other reintroduced populations along the Virginia–West Virginia and Virginia–Tennessee borders.

opportunities to implement timely management actions, If either population is isolated for a prolonged period, the if necessary (Keller et al. 2012, Attard et al. 2016, La combination of human-caused mortality (e.g., legal har- Haye et al. 2017). vests, vehicle collisions, and illegal kills; Murphy 2016) Such monitoring may be particularly salient for the and genetic drift could cause erosion of genetic diver- reintroduced KVP and BSF bear populations because sity, effective population size, and thus population ﬁtness both populations are legally harvested and are relatively (Allendorf et al. 2008, Frankham et al. 2014, Puckett et al. small in number compared with most other bear popu- 2014). Conversely, gene ﬂow between the KVP and BSF lations in the eastern United States (Murphy 2016). Al- or from other populations in the region could prevent though both populations appeared to retain genetic diver- deleterious genetic effects from developing (Nathan et al. sity after the founder events (Murphy et al. 2015, 2016), 2017). those estimates were based on a small number of mi- We implemented a study to quantify contemporary ge- crosatellite markers (n = 7), which likely had low power netic characteristics of the reintroduced KVP and BSF to detect changes (Hale et al. 2012, Landguth et al. 2012). bear populations and to investigate the early genetic

Ursus 29(2):119–133 (2019) r 122 GENETICS OF BLACK BEAR REINTRODUCTIONS Murphy et al. outcomes of bear reintroduction. Specifically, our objec- bears were obtained using the following 20 microsatel- tives were to (1) compare genetic diversity and effective lite markers (Paetkau and Strobeck 1994, Taberlet et al. population sizes between reintroduced and source popu- 1997, Paetkau et al. 1998, Kitahara et al. 2000, Breen lations; (2) quantify genetic structure and differentiation et al. 2001, Proctor et al. 2002): CPH9, CXX20, CXX110, among populations; and (3) determine if the reintroduced G1A, G1D, G10B, G10C, G10H, G10J, G10L, G10M, populations are isolated or if admixture has resulted from G10P, G10U, G10X, MU23, MU50, MU51, MU59, gene flow. REN144A06, and REN145P07. To minimize genotyping error and mitigate incorrect identification of individuals, laboratory personnel discarded samples that failed Methods at >3 markers on the first pass of amplification. They re- Sample collection analyzed samples with 1–3 misidentified pairs, and dis- We collected hair samples from black bears ࣙ1 year carded samples that were without complete genotypes of age in 5 populations. Analyzing genetic data from for all markers. They completed error checking by rean- multiple populations with unequal sample sizes can pro- alyzing pairs of samples with genotypes mismatching at duce erroneous results from allele-frequency-based anal- 1–2 markers to determine if differences existed at each yses (Puechmaille 2016, Wang 2017). Therefore, we ran- locus (Paetkau 2003). In addition to the error checking domly selected samples from 20 to 30 individual bears performed by laboratory personnel, we used MICRO- in each population, which is within the recommended CHECKER v2.2.3 (Van Oosterhout et al. 2004) to test range of sample sizes needed for accurate parameter es- for null alleles, allelic dropout, and scoring errors. timation and comparisons if ࣙ20 microsatellite markers are used for genotyping (Hale et al. 2012, Landguth Population genetics et al. 2012, Puechmaille 2016). We used samples that We used the Program R statistical package genepop were collected from bears via noninvasive hair traps or (Rousset 2008, R Core Team 2017) to test for deviations when live-captured in the KVP (n = 30 bears) and BSF from Hardy–Weinberg equilibrium and to quantify link- (n = 30 bears) populations as part of demographic studies age disequilibrium, applying Bonferroni corrections for during 2010–2014 (Murphy 2016; Murphy et al. 2015, multiple comparisons (α < 0.002 and α < 0.0005, re- 2016). We acquired samples that were collected from spectively). For both tests, we ran 1,000 Markov chain bears via noninvasive hair traps in the GSM source pop- iterations for each of 100 batches. We used the R package ulation during a 2004 capture–mark–recapture study (n diveRsity (Keenan et al. 2013) to identify private alleles = 30 bears; Settlage et al. 2008). We also used samples (AP) and estimate allelic richness (AR) via rarefaction, collected from bears that were live-captured as part of observed (HO) and expected (HE) heterozygosity, and in- a demographic study in southern West Virginia (SWV; breeding coefficient (FIS). We calculated 95% confidence n = 30 bears; Ryan 2009) and as part of human–bear intervals using 1,000 bootstrap iterations. conflict management in the SNP source population in We followed the methods described by Waples et al. northern Virginia (n = 20 bears) during 2009 and 2016, (2014) to estimate effective number of breeders (NB) and respectively. Bears were live-captured using a combina- effective population size (NE) for iteroparous species (i.e., tion of Aldrich spring-activated foothold cable restraints overlapping generations) via the linkage disequilibrium (Johnson and Pelton 1980), culvert and box traps, and method. We first estimated NB using NEESTIMATOR free-range darting; detailed capture and immobilization v2.01 (Do et al. 2014), to which we applied a 2-vital-rate procedures are described by Ryan (2009) and Murphy adjustment formula to correct for bias caused by iteropar- (2016). All methods used for obtaining samples were ap- ity. We then used a separate 2-vital-rate adjustment for- proved by a University of Kentucky Institutional Animal mula to estimate NE from corrected NB (Waples et al. Care and Use Committee (protocol #00626A2003). 2014; Murphy et al. 2017, 2018). Age at maturity and adult life span (AL = maximum age − age at maturity Laboratory analysis + 1) are the 2 vital rates that explain a majority of varia- We sent all samples to Wildlife Genetics International tion in NB and NE (Waples 2016, Waples et al. 2014); we (Nelson, British Columbia, Canada) for DNA extraction, used population-specific vital rates in adjustment formu- polymerase chain reaction amplification, and microsatel- lae (Table S1). We used BOTTLENECK v1.2.02 (Piry lite genotyping. Data quality was managed by labora- et al. 1999) to investigate departure from mutation-drift tory personnel using the methods described by Paetkau equilibrium via a 2-phase model that incorporated 30% (2003) for bear hair samples. Genotypes for individual of multi-step mutations to account for uncertainty in the

Ursus 29(2):119–133 (2019) r GENETICS OF BLACK BEAR REINTRODUCTIONS Murphy et al. 123

mutation process of microsatellites (Luikart et al. 1998b, networks using 3 filter thresholds: mD > 0.0, >0.3, and Peery et al. 2012). We ran 10,000 replicates and assessed >0.5 (Farrell et al. 2016). bottlenecks using Wilcoxon sign-rank tests (Cornuet and Luikart 1996, Luikart et al. 1998a, Peery et al. 2012). Results We detected deviation from Hardy–Weinberg equilib- Population structure and gene flow rium at locus G10B in the KVP, but no deviations were We analyzed population genetic structure using detected in any other populations. Nonrandom associa- Bayesian assignment in STRUCTURE v2.3 (Pritchard tion of alleles existed at 1.1% and 0.5% of 190 pairwise et al. 2000). Using the admixture model with correlated comparisons in the KVP and BSF, respectively. Null alle- allele frequencies, we conducted 10 Markov chain Monte les were suggested at locus REN144A06 in SWV, but we Carlo (MCMC) repetitions of each population cluster (K) found no evidence of null alleles, allelic dropout, or scor- between 2 and 10 for 1,000,000 iterations, and discarded ing errors in the other populations. No consistent patterns the first 100,000 iterations as burn-in. To identify the most of Hardy–Weinberg disequilibrium, linkage disequilib- probable K, we used CLUMPAK (Kopelman et al. 2015) rium, or null alleles were found at the same loci in >1 to calculate the rate of change in second-order derivatives population; therefore, we did not exclude any loci from of the log probability between successive K values (K; analyses (Morin et al. 2010). Evanno et al. 2005). We also used CLUMPAK to aver- Private alleles (AP) were detected in all 5 populations; age the repetitions within each value of K and to display GSM had the most (n = 9), whereas BSF and KVP had results. To estimate pairwise genetic differentiation (FST the least (n = 2; Table S2). Measures of genetic diversity [Weir and Cockerham 1984] and G’ST [Hedrick 2005]; (AR, HO, and HE) were similar among populations (Fig. Whitlock 2011), we used the fastDiv function in the R 2). Estimates of NB and NE for SNP,GSM, and SWV were package diveRsity and calculated 95% confidence inter- substantially larger (NE range = 82–156 bears) than for vals using 1,000 bootstrap iterations. We conservatively KVP and BSF (NE = 31 and 36, respectively). We found considered FST statistically significant if the lower bound strong support for genetic bottlenecks in all 5 popula- was ࣙ0.05 (Hartl and Clark 1997). tions (P = 0.00003–0.00129), but none of the popula- We used BAYESASS v3.0.4 (Wilson and Rannala tions had FIS estimates that were significantly different 2003) to estimate genetic migration rates (m) between from that expected under random mating (i.e., all 95% CIs populations, or the per-generation proportion of individ- included 0). uals in a population that are migrants from another popu- For K = 2 from STRUCTURE analysis, bears in SNP, lation (Rannala 2015, Samarasin et al. 2016). We adjusted SWV, and KVP grouped as a unique cluster, and bears in the mixing parameters for m, allele frequency, and FIS to BSF and GSM grouped as a unique cluster (Fig. 3). The 0.50, 0.70, and 0.70, respectively, to achieve optimal ac- SNP source and SWV each formed a unique cluster at ceptance rates (0.2–0.4; Rannala 2015). We performed K ࣙ 3, whereas the BSF and its respective GSM source 5 MCMC repetitions, each with a different starting seed represented a single cluster at K ࣙ 2. For K = 4 and between 1,000 and 3,000, and ran 2,000,000 iterations K = 5, additional clusters formed through further mixed with the first 200,000 discarded as burn-in (Meirmans assignment of individual bears, not individual sampling 2014). We used TRACER v1.6 (Rambaut et al. 2014) to locales as distinct clusters, which was indicative of over- assess MCMC convergence via trace plots and to cal- splitting of the data (Pritchard et al. 2000, Puechmaille culate effective sample size; we considered an effective 2016). Estimates of K supported K = 3 as most probable sample size of 400 as a minimum acceptable threshold. (Fig. S1), for which the KVP was admixed, primarily We calculated Bayesian deviance (D; Faubet et al. 2007) between SNP and SWV ancestry, with SWV ancestry for each of the 5 runs and produced m estimates from comprising the majority of KVP bears. Approximately the run with the lowest D (Meirmans 2014). We also 23% of bears in the KVP had at least partial ancestry used the divMigrate function in the R package diveRsity from the BSF–GSM cluster, whereas 10% of bears in the to estimate relative directional migration rates (mD) be- BSF had partial ancestry from the SNP or SWV clusters. tween populations, based on Alcala et al.’s (2014) effec- Both the reintroduced BSF and GSM source were mod- tive number of migrants statistic (Sundqvist et al. 2016). erately differentiated from KVP, SWV, and SNP (FST ࣙ This method produces reliable estimates for the sample 0.09; G’ST ࣙ 0.25), whereas the KVP was nominally dif- sizes that we used if global FST is moderate to high (i.e., ferentiated from the SNP source and neighboring SWV >0.05; Sundqvist et al. 2016). We constructed migration (FST = 0.04–0.05; G’ST = 0.13–0.15; Table 1). From

Ursus 29(2):119–133 (2019) r 124 GENETICS OF BLACK BEAR REINTRODUCTIONS Murphy et al.

Fig. 2. The number of private alleles (AP) and estimates of allelic richness (AR), observed (HO) and expected (HE) heterozygosity, effective number of breeders (NB), and effective population size (NE) for 5 American black bear (Ursus americanus) populations in the Central Appalachians, USA, from which hair samples were collected during 2004–2016. The reintroduced Big South Fork (BSF) and Kentucky–Virginia (KVP) populations were established during the 1990s using founders from the Great Smoky Mountains (GSM) and Shenandoah National Park (SNP) populations, respectively. We also sampled a neighboring population in southern West Virginia (SWV). Point estimates and 95% conﬁdence intervals are presented for AR, HO, HE, NB,andNE; inﬁnity is denoted by .

Ursus 29(2):119–133 (2019) r GENETICS OF BLACK BEAR REINTRODUCTIONS Murphy et al. 125

works produced similar results as BAYESASS: mD was highest between GSM and BSF, from SNP to KVP, and from SWV to KVP (Fig. 4).

Discussion Our findings reflect the considerable influence that human activities had on American black bears in the Central Appalachians during the 20th Century. The detection of genetic bottlenecks in the SNP and GSM source populations may confirm that the areas that are now Shenandoah and Great Smoky Mountains national parks served as isolated refugia for bears in a landscape with considerable forest loss and fragmentation (Hall 1981, Unger et al. Fig. 3. Results from genetic structure analysis of 2013). National parks and national forests were estab- 140 American black bears (Ursus americanus)that lished in the region by the 1930s, and game laws to protect were sampled during 2004–2016 in 5 populations in bears or regulate their harvest were implemented by the the Central Appalachians, USA, which were assigned 1950s (Unger et al. 2013). Most black bear mortality is to genetic clusters (K) based on genotypes at 20 generally caused by humans, so those federally protected microsatellite loci. The reintroduced Big South Fork (BSF) and Kentucky–Virginia (KVP) populations were areas were likely critical for allowing the SNP and GSM established using founders from the Great Smoky populations to grow in number over the past half-century. Mountains (GSM) and Shenandoah National Park This protection and subsequent growth likely retained ge- (SNP) populations, respectively, during the 1990s. netic diversity in the SNP and GSM populations (Puckett We also sampled a neighboring population in south- et al. 2014, Murphy et al. 2015). Indeed, our results sug- ern West Virginia (SWV) population. Each vertical bar represents an individual bear’s estimated ancestry gested that, despite being bottlenecked, both the SNP and from the discrete Ks (colored groupings); results for GSM source populations had genetic diversity (AR, HO, K = 3 were most probable. and HE) similar to large black bear populations elsewhere in North America, such as in Minnesota, USA, and por- tions of Canada (Puckett et al. 2014, Pelletier et al. 2017, the most supported run in BAYESASS (D = 11,776; Murphy et al. 2018). effective sample size = 809), m estimates were gener- Selecting populations that have high genetic diversity ally low among the 5 populations; however, m from SNP as sources for reintroduction can mitigate some of the ge- and SWV to KVP and from GSM to BSF were moderate netic consequences of reintroduction (Groombridge et al. (Table2). Significant asymmetrical m was supported from 2012). We found that 15 years (or approx. 2 bear gen- GSM to BSF, from SNP to KVP, from SWV to KPV, and erations; Onorato et al. 2004, Puckett et al. 2014) after from KVP to BSF. Relative directional migration net- the founder events, both the reintroduced KVP and BSF

Table 1. Estimates of genetic differentiation (FST [bottom diagonal] and G’ST [top diagonal]) among 5 American black bear (Ursus americanus) populations in the Central Appalachians, USA, from which hair samples were collected during 2004–2016. The reintroduced Big South Fork (BSF) and Kentucky–Virginia (KVP) populations were established during the 1990s using founders from the Great Smoky Mountains (GSM) and Shenandoah National Park (SNP) populations, respectively. We also sampled a neighboring population in southern West Virginia (SWV). Conﬁdence intervals (95%) are presented in parentheses and statistical signiﬁcance of FST (*) was assigned if the lower bound was ࣙ0.05.

BSF GSM KVP SWV SNP BSF – 0.12 (0.06–0.17) 0.30 (0.23–0.37) 0.41 (0.36–0.46) 0.33 (0.28–0.38) GSM 0.04 (0.02–0.07) – 0.25 (0.20–0.31) 0.30 (0.25–0.35) 0.27 (0.22–0.33) KVP 0.09 (0.07–0.11)* 0.11 (0.08–0.13)* – 0.13 (0.08–0.18) 0.15 (0.09–0.21) SWV 0.15 (0.13–0.17)* 0.10 (0.08–0.13)* 0.04 (0.02–0.06) – 0.20 (0.15–0.25) SNP 0.12 (0.10–0.15)* 0.10 (0.08–0.12)* 0.05 (0.03–0.07) 0.08 (0.06–0.10)* –

Ursus 29(2):119–133 (2019) r 126 GENETICS OF BLACK BEAR REINTRODUCTIONS Murphy et al.

Table 2. Genetic migration rates (m) among 5 American black bear (Ursus americanus) populations in the Cen- tral Appalachians, USA, from which hair samples were collected during 2004–2016. Estimates of m represent the per generation proportion of individuals in row populations that are migrants from column populations. The reintroduced Big South Fork (BSF) and Kentucky–Virginia (KVP) populations were established during the 1990s using founders from the Great Smoky Mountains (GSM) and Shenandoah National Park (SNP) populations, respectively. We also sampled a neighboring population in southern West Virginia (SWV). Bayesian credible intervals (95%) are presented in parentheses and signiﬁcant estimates (i.e., CIs > 0) are indicated with an asterisk.

BSF GSM KVP SWV SNP BSF – 0.28 (0.22–0.34)* 0.02 (0.00–0.04) 0.02 (0.00–0.04) 0.03 (0.00–0.07) GSM 0.01 (0.00–0.03) – 0.01 (0.00–0.03) 0.01 (0.00–0.03) 0.01 (0.00–0.03) KVP 0.05 (0.01–0.09)* 0.01 (0.00–0.03) – 0.19 (0.13–0.25)* 0.10 (0.06–0.14)* SWV 0.01 (0.00–0.03) 0.01 (0.00–0.03) 0.01 (0.00–0.03) – 0.01 (0.00–0.03) SNP 0.01 (0.00–0.03) 0.02 (0.00–0.06) 0.02 (0.00–0.06) 0.02 (0.00–0.06) –

populations retained genetic diversity similar to that of Results from our genetic structure analysis suggest their respective SNP and GSM source populations. Prior that population growth was likely the outcome of differ- studies that used a small number of microsatellite mark- ent processes in each reintroduced population. Multiple ers (n = 7) posited that the retention of genetic diver- bears in the KVP population had complete ancestry from sity in the KVP and BSF populations was inﬂuenced by the SNP source, suggesting those bears were either sur- rapid population growth after the founder events (Mur- viving founders or ﬁrst-generation offspring of founders; phy et al. 2015, 2016). Based on a substantially larger approximately 60% of KVP bears also had at least par- number of markers (n = 20), and thus improved reso- tial ancestry from the SNP. Surprisingly, however, bears lution and power, our results provide further support for with partial ancestry from the neighboring SWV popu- this conclusion. lation comprised the majority of the contemporary KVP

Fig. 4. Relative directional migration rate (mD) networks among 5 American black bear (Ursus americanus) populations in the Central Appalachians, USA, from which hair samples were collected during 2004–2016. The reintroduced Big South Fork (BSF) and Kentucky–Virginia (KVP) populations were established during the 1990s using founders from the Great Smoky Mountains (GSM) and Shenandoah National Park (SNP) populations. We also sampled a neighboring population in southern West Virginia (SWV). Filter thresholds were mD >0.00 (A), >0.30 (B), and >0.50 (C). Lighter gray and thinner arrow bars denote lower mD, whereas darker and wider arrow bars denote higher mD.

Ursus 29(2):119–133 (2019) r GENETICS OF BLACK BEAR REINTRODUCTIONS Murphy et al. 127 population. This suggests that considerable immigration tions, such gene flow may be presently undetectable using of bears from SWV into the KVP occurred, which was microsatellite markers. strongly supported by m and mD estimates. In contrast, The methods used to release founders for popula- the BSF population remains primarily the product of the tion establishment can also influence reintroduction out- GSM-sourced founders. Although 10% of bears in the comes. For example, hard-released bears often attempt BSF population had partial ancestry (ࣘ50%) from the to return to their original capture location, exhibit long- SWV or SNP populations, this was most likely the result distance dispersals, and have low survival rates, all of of 2 male bears that were translocated to BSF from the which can reduce founder group size in reintroductions admixed KVP for human–bear conflict management (J.T. (Comly-Gericke and Vaughan 1997, Mukesh et al. 2015, Hast and J.H. Plaxico, Kentucky Department of Fish and Milligan et al. 2018). The KVP founders were hard- Wildlife Resources, unpublished data). Thus, growth of released, but a smaller proportion of ancestry from the the BSF population has likely resulted solely from the SNP source population exists in the contemporary KVP founder group, whereas growth of the KVP population population than ancestry from the neighboring SWV pop- was influenced by immigration of bears from the SWV ulation. Conversely, the BSF population was primarily population. This may help explain why the KVP popu- established using the winter soft release of fewer bears, lation has an estimated population size that is twice as but this population remained largely identical to its GSM large as the BSF population. Although estimated popu- source for reintroduction. Approximately 2 bear gener- lation growth rates were similar between the 2 popula- ations have elapsed since the KVP founder event and, tions, those calculations were based on the number of considering the low differentiation between the SNP and founders released and resulting population sizes 14–16 SWV populations, genetic swamping is unlikely to have years later; thus, the intrinsic rate of growth for the BSF occurred. Thus, the lower proportion of SNP ancestry founder group was approximately 28% higher than for compared with SWV ancestry in the KVP population sug- the KVP, in part because no or limited immigration oc- gests that its founder group likely declined more than the curred to bolster the smaller BSF founder group (Mur- approximately 33% reduction (from 27 to 18 founders) phy et al. 2015, 2016). These findings collectively sug- that the BSF founder group experienced (Eastridge and gest that spatial proximity of release locations to extant Clark 2001, Murphy et al. 2015). Assuming the KVP populations may be an important consideration for bear founders had a low survival probability similar to the reintroductions. hard-released founders of the other reintroductions in Vir- Bears were also reintroduced to northwestern and ginia (S = 0.23; Comly-Gericke and Vaughan 1997), as south-central Virginia, along the state’s borders with West few as approximately 13 of the 55 founders may have Virginia and Tennessee, neither of which we sampled nor persisted in the KVP reintroduction area. Our findings, included in our analyses (Fies et al. 1987, Comly-Gericke therefore, provide genetic evidence to support the greater and Vaughan1997). The distances separating those popu- effectiveness of soft-releasing bears for reintroduction. lations from the KVP and BSF populations (approx. 120– Founder groups for both recolonization and reintro- 400 km) are within the dispersal capability of black bears duction often succumb to genetic bottlenecks that re- (Stratman et al. 2001, Liley and Walker 2015), but the ex- duce NE because only a subset of alleles from the source pansive and mostly non-forested Great Appalachian Val- population is retained (i.e., founder effect; Puckett et al. ley (James 1920) likely represents a formidable barrier 2014, Malaney et al. 2018, Murphy et al. 2018). Both to natural bear movement from those populations and the the KVP and BSF populations were established using GSM population to the KVP and BSF populations (van small founder groups and consequently had substan- Manen and Pelton 1997). Additionally, because the rein- tially smaller NE than their respective SNP and GSM troduced populations in northwestern and south-central sources. We importantly note that the linkage disequilib- Virginia were also established using founders from the rium method that we used to estimate NB and NE results SNP population, ancestry from those populations in the in negatively biased point estimates with poor precision KVP or BSF likely would be indistinguishable from an- when population abundance is large (ࣙ1,000 individuals; cestry of the KVP founders (Puckett et al. 2014, Mur- Waples and Do 2010). Demographic data indicate that the phy et al. 2018). Consequently, we do not exclude the SNP and GSM populations are indeed large in number possibility that some bears from the other reintroduced (Settlage et al. 2008; VDGIF 2012; U.S. National Park populations in Virginia may have emigrated to the KVP Service 2015, 2017), so our NB and NE estimates for those population, but the potential thereof was likely low and, 2 source populations are likely underestimates, which is given the common source population used for reintroduc- further supported by the lack of confidence interval upper

Ursus 29(2):119–133 (2019) r 128 GENETICS OF BLACK BEAR REINTRODUCTIONS Murphy et al.

bounds. Nonetheless, NE of the KVP and BSF populations population genetics and demographics to identify a link were perhaps smaller than necessary to prevent inbreed- between inbreeding and the decline in ࣙ1 vital rate that ing depression (NE ࣙ 50–100; Jamieson and Allendorf influences fitness (e.g., survival or fecundity; Hedrick and 2012, Frankham et al. 2014). Although FIS estimates did Garcia-Dorado 2016). Given the long generation times not support nonrandom mating in either population, given of bears, multiple decades of such monitoring may be the small founder group sizes, inbreeding may have oc- needed to collect sufficient data (Pemberton 2004, 2008; curred but be difficult to detect so few generations after Hasselgren et al. 2018); thus, perhaps it is not surpris- the founder events (Keller and Waller 2002, Hasselgren ing that inbreeding depression has not been documented et al. 2018). Because of the relatively moderate rate of im- in wild bear populations. However, inbreeding depres- migration from the SWV population, we suspect NE of the sion has been confirmed in small populations of other KVP population likely will improve over time if gene flow large carnivores that had NE similar to the KVP and BSF is sustained. However, genetic diversity and NE can pre- populations (e.g., wolves [Canis lupus; Åkesson et al. cipitously erode in isolated populations as a consequence 2016] and cougars [Puma concolor; Johnson et al. 2010]). of heightened genetic drift (Nei et al. 1975, Jamieson and Therefore, we suggest that implementing early and fre- Allendorf 2012, Frankham et al. 2014). Our collective quent monitoring of reintroduced bear populations at pre- results suggest that the BSF population likely remains defined time intervals is prudent and would likely allow isolated, which, given the smaller number of bears in this managers to detect genetic deficiencies that may warrant population, could potentially become problematic in the conservation action. Given the small NE and isolation of future if sufficient gene flow from other populations does the BSF population, we suggest that managers consider not occur (Dixon et al. 2007, Murphy et al. 2018). implementing genetic monitoring at 5–7-year intervals, or approximately once every bear generation, to detect if gene flow into the BSF occurs or NE declines further. Management implications Our genetic results support that survival and settling rates of founder groups and subsequent popula- Acknowledgments tion growth are improved if bears are soft-released to This study was primarily funded by U.S. Federal Aid in areas with sufficient habitat of high quality; therefore, Wildlife Restoration administered by Kentucky Depart- we suggest that soft releases of founders are preferable ment of Fish and Wildlife Resources. Supplemental funds to hard releases for bear reintroduction. Regardless of the were provided by University of Kentucky, University of method used to release founders, subsequent augmenta- Tennessee, U.S. Forest Service, U.S. Geological Survey, tion may be necessary for genetic restoration if popula- U.S. National Park Service, Virginia Department of Game tions remain isolated for multiple generations (Murphy and Inland Fisheries, and West Virginia Division of Natu- et al. 2018). Bears exhibit male-biased dispersal; there- ral Resources. We thank L. Harris, D. Paetkau, and other fore, selecting release locations that are within the disper- personnel at Wildlife Genetics International for expedi- sal capabilities of male bears in extant populations may ent genotyping. We are grateful for the helpful comments facilitate natural gene flow. Implementing reintroduction and suggestions provided by J. E. Swenson, F. T. van Ma- in this stepping-stone process, if possible, could provide nen, R. K. Rowe, the Associate Editor, and anonymous the opportunity for bears to overcome the consequences reviewers. Any use of trade, firm, or product names is for of founder group reductions or isolation via immigration descriptive purposes only and does not imply endorse- without the need for additional translocations (Karaman- ment by the U.S. Government. Author J. T. Hast was an lidis et al. 2018, Murphy et al. 2018). Nevertheless, we employee of the primary funding entity; to the best of our knowledge, no other potential conflict of interest— caution that immediately reduced NE should be expected in reintroduced bear populations even if genetic diver- financial, relational, or other—exists. sity is retained and population growth is rapid, which has implications for inbreeding depression (Jamieson and Al- Literature cited ÅKESSON,M.,O.LIBERG,H.SAND,P.WABAKKEN,S.BENSCH, lendorf 2012, Keller et al. 2012, Frankham et al. 2014). To AND Ø. FLAGSTAD. 2016. Genetic rescue in a severely inbred our knowledge, inbreeding depression has not been doc- wolf population. Molecular Ecology 25:4745–4756. umented in any wild bear populations, though it has not ALCALA,N.,J.GOURDET, AND S. VUILLEUMIER. 2014. On the been explicitly investigated either (but see Laikre et al. transition of genetic differentiation from isolation to pan- [1996] for an example in a captive bear population). Such mixia: What we can learn from GST and D. Theoretical Pop- an investigation would require concurrent monitoring of ulation Biology 93:75–84.

Ursus 29(2):119–133 (2019) r GENETICS OF BLACK BEAR REINTRODUCTIONS Murphy et al. 129

ALLENDORF, F.W., P.R. ENGLAND,G.LUIKART,P.A.RITCHIE, Re-implementation of software for the estimation of con- AND N. RYMAN. 2008. Genetic effects of harvest on wild temporary effective population size (Ne) from genetic data. animal populations. Trends in Ecology and Evolution 23: Molecular Ecology Resources 14:209–214. 327–337. EASTRIDGE,R.,AND J.D. CLARK. 2001. Evaluation of 2 soft- ATTARD, C.R.M., L.M. MÖLLER,M.SASAKI,M.P.HAMMER, release techniques to reintroduce black bears. Wildlife So- C.M. BICE,C.J.BRAUER,D.C.CARVALHO,J.O.HARRIS, ciety Bulletin 29:1163–1174. AND L.B. BEHEREGARAY. 2016. A novel holistic framework EVANNO,G.,S.REGNAUT, AND J. GOUDET. 2005. Detecting for genetic-based captive-breeding and reintroduction pro- the number of clusters of individuals using the software grams. Conservation Biology 30:1060–1069. STRUCTURE: A simulation study. Molecular Ecology 14: BENSON, J.F., AND M.J. CHAMBERLAIN. 2007. Space use, 2611–2620. survival, movements, and reproduction of reintroduced FARRELL, E.D., J.E.L. CARLSSON, AND J. CARLSSON. 2016. Louisiana black bears. Journal of Wildlife Management Next gen pop gen: Implementing a high-throughput ap- 71:2393–2403. proach to population genetics in boarfish (Capros aper). BESTON, J.A. 2011. Variation in life history and demography of Royal Society Open Science 3:160651. the American black bear. Journal of Wildlife Management FAUBET,P.,R.S.WAPLES, AND O.E. GAGGIOTTI. 2007. Evalu- 75:1588–1596. ating the performance of a multilocus Bayesian method for BREEN,M.,S.JOUQUAND,C.RENEIR,C.S.MELLERSH, the estimation of migration rates. Molecular Ecology 16: C. HITTE,N.G.HOLMES,A.CHERON,N.SUTER,F. 1149–1166. VIGNAUX,A.E.BRISTOW,C.PRIAT,E.MCCANN,C. FIES, M.L., D.D. MARTIN, AND G.T. BLANK. 1987. Movements ANDRE,S.BOUNDY,P.GITSHAM,R.THOMAS,W.L. and rates of return of translocated black bears in Virginia. In- BRIDGE,H.F.SPRIGGS,E.J.RYDER,A.CURSON,J.SAMP- ternational Conference on Bear Research and Management SON,E.A.OSTRANDER,M.M.BINNS, AND F. GALIBERT. 7:369–372. 2001. Chromosome-specific single-locus FISH probes al- FRANKHAM, R., C.J.A. BRADSHAW, AND B.W. BROOK. 2014. low anchorage of an 1800-marker integrated radiation- Genetics in conservation management: Revised recommen- hybrid/linkage map of the domestic dog genome to all chro- dations for the 50/500 rules, Red List criteria and pop- mosomes. Genome Research 11:1784–1795. ulation viability analyses. Biological Conservation 170: BROWN, S.K., J.M. HULL,D.R.UPDIKE,S.R.FAIN, AND H.B. 56–63. ERNEST. 2009. Black bear population genetics in Califor- GROOMBRIDGE, J.J., C. RAISIN,R.BRISTOL, AND D.S. nia: Signatures of population structure, competitive release, RICHARDS. 2012. Genetic consequences of reintroduc- and historical translocation. Journal of Mammalogy 90: tions and insights from population history. Pages 395– 1066–1074. 440 in J.G. Ewin, D.P. Armstrong, K.A. Parker, and P.J. BRUSKOTTER, J.T., J.A. VUCETICH,M.J.MANFREDO,G.R. Seddon, editors. Reintroduction biology: Integrating sci- KARNS,C.WOLF,K.ARD,N.H.CARTER,J.V.LÓPEZ-BAO, ence and management. Wiley–Blackwell, Oxford, England, G. CHAPRON,S.D.GEHRT, AND W.J. RIPPLE. 2017. Mod- UK. ernization, risk, and conservation of the world’s largest car- HALE, M.L., T.M. BURG, AND T.E. STEEVES. 2012. Sampling nivores. BioScience 67:646–655. for microsatellite-based population genetic studies: 25 to 30 CLARK, J.D. 2009. Aspects and implications of bear reintroduc- individuals per population is enough to accurately estimate tion. Pages 126–145 in M.W. Hayward and M.J. Somers, allele frequencies. PLoS ONE 7:e45170. editors. Reintroduction of top-order predators. Blackwell, HALL, E.R. 1981. The mammals of North America. Second Hoboken, New Jersey, USA. edition. John Wiley and Sons, New York, New York, USA. ———, D. HUBER, AND C. SERVHEEN. 2002. Bear reintroduc- HARTL, D.L., AND A.G. CLARK. 1997. Principles of populations: Lessons and challenges. Ursus 13:335–345. tion genetics. Third edition. Sinauer Associates, Sunderland, COMLY-GERICKE, L.M., AND M.R. VAUGHAN. 1997. Survival Massachusetts, USA. and reproduction of translocated Virginia black bears. Bears: HASSELGREN,M.,A.ANGERBJÖRN,N.E.EIDE,R.ERLANDS- Their Biology and Management 9:113–117. SON,Ø.FLAGSTAD,A.LANDA,J.WALLÉN, AND K. NORÉN. CORNUET, J.M., AND G. LUIKART. 1996. Description and power 2018. Genetic rescue in an inbred Arctic fox (Vulpes analysis of two tests for detecting recent population bottle- lagopus) population. Proceedings of the Royal Society B necks from allele frequency data. Genetics 144:2001–2014. 285:20172814. DIXON, J.D., M.K. OLI,M.C.WOOTEN,T.H.EASON,J.W. HAYWARD, M.W., AND M.J. SOMERS. 2009. Reintroduction of MCCOWN, AND M.W. CUNNINGHAM. 2007. Genetic con- top-order predators: Using science to restore one of the sequences of habitat fragmentation and loss: The case of the drivers of biodiversity. Pages 1–9 in M.W. Hayward and Florida black bear (Ursus americanus floridanus). Conser- M.J. Somers, editors. Reintroduction of top-order predators. vation Genetics 8:455–464. Blackwell, Hoboken, New Jersey, USA. DO,C.,R.S.WAPLES,D.PEEL,G.M.MACBETH,B.J. HEDRICK, P.W. 2005. A standardized genetic differentiation TILLETT, AND J.R. OVENDEN. 2014. NEESTIMATOR v2: measure. Evolution 59:1633–1638.

Ursus 29(2):119–133 (2019) r 130 GENETICS OF BLACK BEAR REINTRODUCTIONS Murphy et al.

———, AND A. GARCIA-DORADO. 2016. Understanding in- for identifying clustering modes and packaging population breeding depression, purging, and genetic rescue. Trends in structure inferences across K. Molecular Ecology Resources Ecology and Evolution 31:940–952. 15:1179–1191. INTERNATIONAL UNION FOR CONSERVATION OF NATURE AND LA HAYE, M.J.J., T.E. REINERS,R.RAEDTS,V.VERBIST, AND NATURAL RESOURCES. 2017. The IUCN Red List of threat- H.P. KOELEWIJN. 2017. Genetic monitoring to evaluate rein- ened species, version 2017–3. https://www.iucnredlist.org. troduction attempts of a highly endangered rodent. Conser- Accessed 16 Mar 2018. vation Genetics 18:877–892. JAMES, H.F. 1920. The geography of a portion of the Great LAIKRE,L.,R.ANDRÉN,H.-O.LARSSON, AND N. RYMAN. 1996. Appalachian Valley and selected adjacent regions. Thesis, Inbreeding depression in brown bear Ursus arctos. Biolog- University of Wisconsin, Madison, Wisconsin, USA. ical Conservation 76:69–72. JAMIESON, I.G., AND F.W. ALLENDORF. 2012. How does the LANDGUTH, E.L., B.C. FEDY,S.J.OYLER-MCCANCE,A.L. 50/500 rule apply to MVPs? Trends in Ecology and Evolu- GAREY,S.L.EMEL,M.MUMMA,H.H.WAGNER,M.-J. tion 27:578–584. FORTIN, AND S.A. CUSHMAN. 2012. Effects of sample size, JOHNSON, K.G., AND M.R. PELTON. 1980. Prebaiting and snar- number of markers, and allelic richness on the detection ing techniques for black bears. Wildlife Society Bulletin of spatial genetic pattern. Molecular Ecology Resources 8:46–54. 12:276–284. JOHNSON, W.E., D.P. ONORATO,M.E.ROELKE,E.D.LAND,M. LILEY, S.G., AND R.N. WALKER. 2015. Extreme movement by CUNNINGHAM,R.C.BELDEN,R.MCBRIDE,D.JANSEN,M. an American black bear in New Mexico and Colorado. Ursus LOTZ,D.SHINDLE,J.HOWARD,D.E.WILDT,L.M.PEN- 26:1–6. FOLD,J.A.HOSTETLER,M.K.OLI, AND S.J. O’BRIEN. 2010. LUIKART,G.,F.W.ALLENDORF,J.M.CORNUET, AND W.B. Genetic restoration of the Florida panther. Science 329: SHERWIN. 1998a. Distortion of allele frequency distributions 1641–1645. provides a test for recent population bottlenecks. Journal of KARAMANLIDIS, A.A., T. SKRBINŠEK,M.DE GABRIEL HE- Heredity 89:238–247. MANDO,L.KRAMBOKOUKIS,V.MUNOZ-FUENTES,Z.BAI- ———, W.B. SHERWIN,B.STEELE, AND F.W. ALLENDORF. LEY,C.NOWAK, AND A.V. STRONEN. 2018. History-driven 1998b. Usefulness of molecular markers for detecting popu- population structure and asymmetric gene flow in a recov- lation bottlenecks via monitoring genetic change. Molecular ering large carnivore at the rear-edge of its European range. Ecology 7:963–974. Heredity 120:168–182. MA,K.,D.LIU,R.WEI,G.ZHANG,H.XIE,Y.HUANG,D.LI,H. KEENAN,K.,P.MCGINNITY,T.F.CROSS,W.W.CROZIER, AND ZHANG, AND H. XU. 2016. Giant panda reintroduction: Fac- P.A. P RODOHL. 2013. diveRsity: An R package for the es- tors affecting public support. Biodiversity and Conservation timation and exploration of population genetics parameters 25:2987–3004. and their associated errors. Methods in Ecology and Evolu- MALANEY, J.L., C.W. LACKEY,J.P.BECKMANN, AND M.D. tion 4:782–788. MATOCQ. 2018. Natural rewilding of the Great Basin: KELLER, L.F., I. BIEBACH,S.R.EWING, AND P.E.A. H OECK. Genetic consequences of recolonization by black bears 2012. The genetics of reintroductions: Inbreeding and (Ursus americanus). Diversity and Distributions 24: genetic drift. Pages 362–393 in J.G. Ewin, D.P. Arm- 168–178. strong, K.A. Parker, and P.J. Seddon, editors. Reintroduc- MCLEAN, P.K.1991. The demographic and morphological char- tion biology: Integrating science and management. Wiley– acteristics of black bears in the Smoky Mountains. Disserta- Blackwell, Oxford, England, UK. tion, University of Tennessee, Knoxville, Tennessee, USA. ———, AND D.M. WALLER. 2002. Inbreeding effects in MEIRMANS, P.G. 2014. Nonconvergence in Bayesian estima- wild populations. Trends in Ecology and Evolution 17: tion of migration rates. Molecular Ecology Resources 14: 230–241. 726–733. KIM, Y.-K., Y.-J. HONG,M.-S.MIN,K.S.KIM,Y.-J.KIM,I. MILLIGAN,S.,L.BROWN,D.HOBSON,P.FRAME, AND G. VOLOSHINA,A.MYSLENKOV, G.J.D. SMITH,N.D.CUONG, STENHOUSE. 2018. Factors affecting the success of griz- H.H. THO,S.-H.HAN,D.-H.YANG,C.-B.KIM, AND H. zly bear translocations. Journal of Wildlife Management 82: LEE. 2011. Genetic status of Asiatic black bear (Ursus thi- 519–530. betanus) reintroduced into South Korea based on mitochon- MORIN, P.A., K.K. MARTIEN,F.I.ARCHER,F.CIPRIANO,D. drial DNA and microsatellite loci analysis. Journal of Hered- STEEL,J.JACKSON, AND B.L. TAYLOR. 2010. Applied con- ity 102:165–174. servation genetics and the need for quality control and re- KITAHARA,E.,Y.ISAGI,Y.ISHIBASHI, AND T. SAITOH. 2000. porting of genetic data used in fisheries and wildlife man- Polymorphic microsatellite DNA markers in the Asiatic agement. Journal of Heredity 101:1–10. black bear Ursus thibetanus. Molecular Ecology 9:1661– MUKESH, S.L.K., S.A. CHAROO, AND S. SATHYAKUmar. 2015. 1662. Conflict bear translocation: Investigating population genet- KOPELMAN, N.M., J. MAYZEL,M.JAKOBSSON,N.A.ROSEN- ics and fate of bear translocation in Dachigam National Park, BERG, AND I. MAYROSE. 2015. CLUMPAK: A program Jammu and Kashmir, India. PLoS ONE 10:e0132005.

Ursus 29(2):119–133 (2019) r GENETICS OF BLACK BEAR REINTRODUCTIONS Murphy et al. 131

MURPHY, S.M. 2016. Ecology of two reintroduced black tests for detecting recent population declines. Molecular bear populations in the Central Appalachians. Dissertation, Ecology 21:3403–3418. University of Kentucky, Lexington, Kentucky, USA. PELLETIER, A., M.E. OBBARD,M.HARNDEN,S.MCCONNELL, ———, B.C. AUGUSTINE,W.A.ULREY,J.M.GUTHRIE,B.K. E.J. HOWE,F.G.BURROWS,B.N.WHITE, AND C.J. KYLE. SCHEICK,J.W.MCCOWN, AND J.J. COX. 2017. Conse- 2017. Determining causes of genetic isolation in a large car- quences of severe habitat fragmentation on density, genetics, nivore (Ursus americanus) population to direct contempo- and spatial capture–recapture analysis of a small bear pop- rary conservation measures. PLoS ONE 12:e0172319. ulation. PLoS ONE 12:e0181849. PEMBERTON, J. 2004. Measuring inbreeding depression in the ———, J.J. COX,B.C.AUGUSTINE,J.T.HAST,J.M.GUTHRIE,J. wild: The old ways are the best. Trends in Ecology and Evo- WRIGHT,J.MCDERMOTT,S.C.MAEHR, AND J.H. PLAXICO. lution 19:613–615. 2016. Characterizing recolonization by a reintroduced black ———. 2008. Wild pedigrees: The way forward. Proceedings bear population using genetic spatial capture–recapture. of the Royal Society B 275:613–621. Journal of Wildlife Management 80:1390–1407. PIRY,S.,G.LUIKART, AND J.-M. CORNUET. 1999. BOTTLE- ———, ———, J.D. CLARK,B.C.AUGUSTINE,J.T.HAST,D. NECK: A program for detecting recent effective popula- GIBBS,M.STRUNK, AND S. DOBEY. 2015. Rapid growth tion size reductions from allele data frequencies. Journal of and genetic diversity retention in an isolated reintroduced Heredity 90:502–503. black bear population in the Central Appalachians. Journal PRITCHARD, J.K., M. STEPHENS, AND P. DONNELLY. 2000. In- of Wildlife Management 79:807–818. ference of population structure using multilocus genotype ———, J.S. LAUFENBERG,J.D.CLARK,M.DAVIDSON,J.L. data. Genetics 155:945–959. BELANT, AND D.L. GARSHELIS. 2018. Genetic diversity, PROCTOR, M.F., B.N. MCLELLAN, AND C. STROBECK. 2002. effective population size, and structure among black bear Population fragmentation of grizzly bears in southeastern populations in the Lower Mississippi Alluvial Valley, USA. British Columbia, Canada. Ursus 13:153–160. Conservation Genetics 19:1055–1067. PUCKETT, E.E., T.V. KRISTENSEN,C.M.WILTON,S.B.LYDA, NATHAN, L.R., Y. KANNO, AND J.C. VOKOUN. 2017. Popula- K.V. NOYCE,P.M.HOLAHAN,D.M.LESLIE,J.BERINGER, tion demographics influence genetic responses to fragmen- J.L. BELANT,D.WHITE, AND L.S. EGGERT. 2014. Influence tation: A demographic assessment of the ‘one migrant per of drift and admixture on population structure of American generation’ rule of thumb. Biological Conservation 210: black bears (Ursus americanus) in the Central Interior High- 261–272. lands, USA, 50 years after translocation. Molecular Ecology NEI,M.,T.MARUYAMA, AND R. CHAKRABORTY. 1975. The 23:2414–2427. bottleneck effect and genetic variability in populations. Evo- PUECHMAILLE, S.J. 2016. The program STRUCTURE does lution 29:1–10. not reliably recover the correct population structure when NOWAK,C.,C.DOMOKOS,A.DUTSOV, AND C. FROSCH. 2014. sampling is uneven: Subsampling and new estimators al- Molecular evidence for historic long-distance translocations leviate the problem. Molecular Ecology Resources 16: of brown bears in the Balkan region. Conservation Genetics 608–627. 15:743–747. PURSLEY, D. 1974. History and status of black bear in West Vir- ONORATO, D.P., E.C. HELLGREN,R.A.VA N D E N BUSSCHE, AND ginia. Thesis, West Virginia University, Morgantown, West D.L. DOAN–CRIDER. 2004. Phylogeographic patterns within Virginia, USA. a metapopulation of black bears (Ursus americanus)inthe RCORE TEAM. 2017. R: A language and environment for sta- American Southwest. Journal of Mammalogy 85:140–147. tistical computing. R Foundation for Statistical Computing, ———, ———, ———, ———, AND J. R. SKILES. 2007. Ge- Vienna, Austria. http://www.cran.r-project.org/. Accessed 5 netic structure of American black bears in the desert south- Jul 2017. west of North America: Conservation implications for re- RAMBAUT,A.,M.A.SUCHARD,D.XIE, AND A.J. DRUMMOND. colonization. Conservation Genetics 8:565–576. 2014. Tracer v1.6. http://www.beast.bio.ed.ac.uk/Tracer/. PAETKAU, D. 2003. An empirical exploration of data quality Accessed 10 Mar 2017. in DNA-based population inventories. Molecular Ecology RANNALA, B. 2015. BayesAss edition 3.0 user’s manual. Uni- 12:1375–1387. versity of California Davis, Davis, California, USA. _____, AND C. STROBECK. 1994. Microsatellite analysis of ge- ROUSSET, F. 2008. GENEPOP’007: A complete re- netic variation in black bear populations. Molecular Ecology implementation of the GENEPOP software for Win- 3:489–495. dows and Linux. Molecular Ecology Resources 8: _____, G.F. SHIELDS, AND C. STROBECK. 1998. Gene flow 103–106. between insular, coastal and interior populations of brown RYAN, C.W. 2009. Population ecology, residents’ attitudes, bears in Alaska. Molecular Ecology 7:1283–1292. hunter success, economic impact, modeling options and PEERY, M.Z., R. KIRBY,B.N.REID,R.STOELTING,E.DOUCET- retention time of Telazol of West Virginia black bears. BËER,S.ROBINSON,C.VÁSQUEZ-CARRILLO,J.N.PAULI, Dissertation, West Virginia University, Morgantown, West AND P.J. PALSBØLL. 2012. Reliability of genetic bottleneck Virginia, USA.

Ursus 29(2):119–133 (2019) r 132 GENETICS OF BLACK BEAR REINTRODUCTIONS Murphy et al.

SAMARASIN, P., B.J. SHUTER,S.I.WRIGHT, AND F.H. RODD. ited: The importance of adult mortality. Heredity 117: 2016. The problem of estimating recent genetic connectivity 241–250. in a changing world. Conservation Biology 31:126–135. ———, T. ANTAO, AND G. LUIKART. 2014. Effects of over- SETTLAGE, K.E., F.T. VA N MANEN,J.D.CLARK, AND T.L. KING. lapping generations on linkage disequilibrium estimates of 2008. Challenges of DNA-based mark–recapture studies effective population size. Genetics 197:769–780. of American black bears. Journal of Wildlife Management ———, AND C. DO. 2010. Linkage disequilibrium estimates 72:1035–1042. of contemporary Ne using highly variable genetic markers: SRATMAN, M.R., C.D. ALDEN,M.R.PELTON, AND M.E. SUN- A largely untapped resource for applied conservation and QUIST. 2001. Long distance movement of a Florida black evolution. Evolutionary Applications 3:244–262. bear in the Southeastern Coastal Plain. Ursus 12:55–58. WEAR, B.J., R. EASTRIDGE, AND J.D. CLARK. 2005. Factors SUNDQVIST,L.,K.KEENAN,M.ZACKRISSON,P.PRODOHL, AND affecting settling, survival, and viability of black bears rein- D. KLEINAHNS. 2016. Directional genetic differentiation and troduced to Felsenthal National Wildlife Refuge, Arkansas. relative migration. Ecology and Evolution 6:3461–3475. Wildlife Society Bulletin 33:1363–1374. TABERLET,P.,J.-J.CAMARRA,S.GRIFFIN,E.UHRES,O.HAN- WEIR, B.S., AND C.C. COCKERHAM. 1984. Estimating F- OTTE,L.P.WAITS,C.DUBOIS-PAGANON,T.BURKE, AND J. statistics for the analysis of population structure. Evolution BOUVET. 1997. Noninvasive genetic tracking of the endan- 38:1358–1370. gered Pyrenean brown bear population. Molecular Ecology WHITLOCK, M.C. 2011. G’ST and D do not replace FST.Molec- 6:869–876. ular Ecology 20:1083–1091. TOSI,G.,R.CHIRICHELLA,F.ZIBORDI,A.MUSTONI,R.GIO- WILSON, G.A., AND B. RANNALA. 2003. Bayesian inference of VANNINI,C.GROFF,M.ZANIN, AND M. APOLLONIO. 2015. recent migration rates using multilocus genotypes. Genetics Brown bear reintroduction in the Southern Alps: To what 163:1177–1191. extent are expectations being met? Journal for Nature Con- WOLF,C.,AND W.J. RIPPLE. 2018. Rewilding the world’s large servation 26:9–19. carnivores. Royal Society Open Science 5:172235. UNGER, D.E., J.J. COX,H.B.HARRIS,J.L.LARKIN,B.AUGUS- TINE,S.DOBEY,J.M.GUTHRIE,J.T.HAST,R.JENSEN,S. MURPHY,J.PLAXICO, AND D.S. MAEHR. 2013. History and Received: May 2, 2018 current status of the black bear in Kentucky. Northeastern Accepted: December 3, 2018 Naturalist 20:289–308. Associate Editor: E. Bellemain U.S. NATIONAL PARK SERVICE. 2015. Black bear management in Great Smoky Mountains National Park. https://www.nps.gov/grsm/learn/nature/dff309-focusnps1. htm. Accessed 12 Mar 2018. Supplemental material ———. 2017. American black bear—Shenandoah Na- Table S1. Population-speciﬁc vital-rate estimates tional Park. https://www.nps.gov/shen/learn/nature/black- that were used in adjustment formulae to correct for bear.htm. Accessed 12 Mar 2018. overlapping generations and estimate effective num- VA N MANEN, F.T., AND M.R. PELTON. 1997. Procedures to en- ber of breeders and effective population sizes of 5 hance the success of a black bear reintroduction program. In- ternational Conference on Bear Research and Management American black bear (Ursus americanus) populations 9:67–78. in the Central Appalachians, USA, via the linkage VAN OOSTERHOUT,C.,W.F.HUTCHINSON, D.P.M. WILLS, AND disequilibrium method. Vital rates were maximum P. SHIPLEY. 2004. MICRO-CHECKER: Software for identi- known age (ω), age at maturity (α; i.e., primipar- fying and correcting genotyping errors in microsatellite data. ity), and adult life span (AL) for the reintroduced Molecular Ecology Resources 4:535–538. Big South Fork (BSF) and KentuckyÐVirginia (KVP) VIRGINIA DEPARTMENT OF GAME AND INLAND FISHERIES populations, their respective Great Smoky Mountains [VDGIF]. 2008. VDGIF bear relocation report 2008. Vir- (GSM) and Shenandoah National Park (SNP) source ginia Department of Game and Inland Fisheries, Henrico, populations, and a neighboring population in south- Virginia, USA. ern West Virginia (SWV). Hair samples were col- ———. 2012. Virginia black bear management plan 2012– lected from individual bears in each population dur- 2021. Virginia Department of Game and Inland Fisheries, Henrico, Virginia, USA. ing 2004Ð2016 for genetic analyses. WANG, J. 2017. The computer program STRUCTURE for as- Table S2. The number of private alleles (AP) signing individuals to populations: Easy to use but easier to and estimates of allelic richness (AR), observed het- misuse. Molecular Ecology 17:981–990. erozygosity (HO), expected heterozygosity (HE), ef- WAPLES, R.S. 2016. Life-history traits and effective popula- fective number of breeders (NB), effective population tion size in species with overlapping generations revis- size (NE), and inbreeding coefﬁcient (FIS) with 95%

Ursus 29(2):119–133 (2019) r GENETICS OF BLACK BEAR REINTRODUCTIONS Murphy et al. 133 conﬁdence intervals for 5 American black bear (Ur- Fig. S1. Results for the rate of change in second- sus americanus) populations in the Central Appalachi- order derivatives of the log probability between suc- ans, USA: the reintroduced Big South Fork (BSF) cessive values (K) for identifying the most prob- and KentuckyÐVirginia (KVP) populations, their able number of genetic clusters (K) from STRUC- respective Great Smoky Mountains (GSM) and TURE analysis of American black bears (Ursus Shenandoah National Park (SNP) source populations, americanus) in 5 populations in the Central Ap- and a neighboring population in southern West Vir- palachians, USA. Based on analysis of genotypes ginia (SWV). Based on analysis of genotypes obtained obtained from hair samples that were collected from hair samples that were collected from individual from individual bears in each population during bears in each population during 2004Ð2016. 2004Ð2016.

Ursus 29(2):119–133 (2019)