Mammalian Biology 90 (2018) 42–46

Contents lists available at ScienceDirect

Mammalian Biology
journal homepage: www.elsevier.com/locate/mambio

Short communication

Detection dogs allow for systematic non-invasive collection of DNA

samples from Eurasian lynx
Laura Hollerbach a,b,∗ , Marco Heurich c,d , Tobias Erik Reiners a , Carsten Nowak a
a
Senckenberg Research Institute and Natural History Museum Frankfurt, Conservation Genetics Group, Clamecystraße 12, 63571 Gelnhausen, Germany
b
Technische Universität Dresden, Chair of Forest Zoology, Faculty of Environmental Sciences, Pienner Straße 8, 01737 Tharandt, Germany
c
Bavarian Forest National Park, Department of Conservation and Research, Freyunger Straße 2, 94481 Grafenau, Germany
d
University of Freiburg, Chair of Wildlife Ecology and Management, Faculty of Environment and Natural Resources, Tennenbacher Straße 4, 79106
Freiburg, Germany

a r t i c l e i n f o a b s t r a c t

Article history: As Eurasian lynx (Lynx lynx) show signs of population recovery in parts of Central Europe, sound moni-
Received 7 December 2017 toring strategies are required to study population expansion, connectivity and genetic diversity. While
Accepted 14 February 2018 non-invasive DNA sampling strategies could serve this task, genetic samples of lynx are generally hard
Handled by Frank E. Zachos
to locate. To test the suitability of dog-based sampling we searched scat samples of lynx in the Bavarian
Available online 19 February 2018
Forest National Park, Germany, with two trained detection dog teams. In 44 grid cells of 2 × 2 km, dog
teams covered 440 km of predetermined forest road and hiking trail transects during the four week sur-
Keywords:
vey. A total of 169 collected samples resulted in 52 genetically confirmed lynx detections, of which 26
Detection dog
Lynx lynx
were assigned to 11 individuals. Using a single-season site occupancy model we found a detection prob-
Non-invasive DNA sampling ability of 0.13/km (SD = 0.02), with 10 km of dog search per grid cell required to get a 70 % probability to
Scat detect lynx presence. Our results show that detection dogs are an appropriate tool for systematic genetic
Bavarian Forest National Park lynx monitoring. We argue that detection dog-assisted genetic monitoring may supplement monitoring
strategies based on conventional camera trapping, especially when aiming to monitor genetic diversity
and population connectivity.
© 2018 Deutsche Gesellschaft für Säugetierkunde. Published by Elsevier GmbH. All rights reserved.

Large carnivores such as wolves (Canis lupus), brown bears
(Ursus arctos) and Eurasian lynx (Lynx lynx) are currently recolonis-
ing parts of their historic range in Western and Central Europe
(Chapron et al., 2014). This population expansion is intensively
monitored in these species, due to their high societal and envi-
ronmental impact. In the case of wolves and bears, traditional
monitoring methods are routinely complemented by the genetic
identification of species, populations and individuals based on non-
invasively collected samples (e.g. Barba et al., 2010; Lucchini et al.,
2002). Interestingly, only few studies have successfully applied
genetic approaches in monitoring programmes of Eurasian lynx
(Bull et al., 2016; Davoli et al., 2013). Lynx are elusive and single
individuals cover large home ranges, which poses a challenge to
monitoring schemes (Weingarth et al., 2012). Experts demand that
∗ Corresponding author at: Senckenberg Research Institute and Natural His-
tory Museum Frankfurt, Conservation Genetics Group, Clamecystraße 12, 63571
Gelnhausen, Germany.
E-mail addresses: laura.hollerbach@senckenberg.de (L. Hollerbach),
marco.heurich@npv-bw.bayern.de (M. Heurich), tobias.reiners@senckenberg.de
(T.E. Reiners), carsten.nowak@senckenberg.de (C. Nowak).
monitoring should be intensified and recommend genetic monitor-
ing alongside with camera trapping (Boitani et al., 2015; Chapron
et al., 2014). Currently, standardised lynx monitoring is mainly
conducted by camera trapping, which is a very efficient method
for detecting lynx presence and calculating population densities
(Weingarth et al., 2015). However, with this technique discrimina-
tion of single individuals requires distinct fur patterns, which are
not reliably existent in all subpopulations (Thüler, 2002). Further-
more, information regarding genetic population structure, genetic
diversity, inbreeding, ancestry and kinship cannot be acquired with
photo-monitoring.
DNA samples from lynx are notoriously difficult to obtain in
their natural habitat. Lynx scats are hard to find and cannot be
reliably identified in the field (Alda et al., 2008). There have been
approaches using hair trapping, but success rates were generally
low (Schmidt and Kowalczyk, 2006). Snow tracking is an option to
gain scat samples. The effort for this method is high and snow cover
is sparse in many areas, such as the Central European low mountain
region where the lynx is currently expanding its range. Especially
in recently populated areas with low lynx density, however, effec-
tive systematic monitoring strategies are required, including the

https://doi.org/10.1016/j.mambio.2018.02.003
1616-5047/© 2018 Deutsche Gesellschaft für Säugetierkunde. Published by Elsevier GmbH. All rights reserved.
 L. Hollerbach et al. / Mammalian Biology 90 (2018) 42–46
genetic assignment of dispersing animals to source populations as
well as estimating genetic connectivity between isolated popula-
tion fragments.
The use of trained dogs in wildlife management and conserva-
tion to obtain information about one or more target species has
gained increasing attention among researchers and conservation
managers during the last years (Dahlgren et al., 2012; Woollett
et al., 2014). Scat detection dogs have successfully located genetic
samples of closely related species, i.e. bobcat (Lynx rufus) (e.g. Clare
et al., 2015) and Canada lynx (Lynx canadensis) (Mumma et al.,
2015). To our knowledge, however, no study has been published
so far on scat detection dog use for surveying Eurasian lynx, which
occupy considerably larger home ranges than the new world lynx
species.
In this study we therefore investigated whether the use of
detection dogs is a feasible method for a systematic genetic lynx
assessment. We combined the detection dog survey with cam-
era trapping in the Bavarian Forest National Park (BFNP) (see
Fig. 1a), which hosts a well-monitored reintroduced lynx popula-
tion (Weingarth et al., 2015) to generate a sophisticated estimate
of detection probability for the dog-based approach.
The BFNP is Germany’s first national park founded in 1970. It
covers 240 km2 on the German side and is adjacent to the Czech
Šumava National Park. Elevation ranges from 600 m to 1453 m a.s.l.
and annual mean temperatures are from 2 ◦ C to 6.5 ◦ C. The forest is
dominated by Picea abies mixed with Fagus sylvatica and Abies alba
(Cailleret et al., 2014). The current lynx population in the German-
Czech border region is based on reintroductions in the 1970s and
1980s on both sides (Wölfl et al., 2001), with a current population
estimate of 59–83 individuals (Wölfl et al., 2015).
We set up a 2 × 2 km grid across the entire BFNP area. We
selected grid cells that are 80 % or more within the BFNP borders,
which resulted in a total number of 44 adjacent grid cells (Fig. 1b).
Two detection dog teams consisting of one handler, dog and an ori-
enteer each conducted scat surveys from 25.03.2017 to 21.04.2017.
While both handlers worked with their individual dogs, one addi-
tional person per dog team ensured navigation in the field. Both
teams surveyed on predetermined transects that were preferably
circular (Fig. 1a), for logistic reasons. Transects were mainly along
forest roads or paths. We targeted to cover transects only once in
one direction and only by one team. Each team covered 22 tran-
sects and was required to survey each of the 44 grid cells. Dogs and
orienteers were equipped with GPS devices (Garmin Alpha 100 T5
Bundle for team 1; Garmin GPSMAP 64 s and eTrex 10 for team
2) in order to track the distance covered. Dog 1 (female Labrador
Retriever) was 11 months old with seven months of previous scat
detection training. It mainly searched on a 15 m leash and indicated
a find through a bringsel, which is an item attached to the collar that
is taken into the muzzle by the dog in reaction to a find. The dog was
rewarded with food and a subsequent short game with a toy. Dog
2 (female Border Collie Golden Retriever cross) was four years old
with 14 months of training. It searched off-leash and indicated by
freezing and staring at the sample. This dog was rewarded through
playing with a ball. Both dogs, who have been living with their
handlers from puppy age on, were trained based on positive rein-
forcement by their handlers using samples from numerous captive
and free-ranging lynx (Wasser et al., 2004). Though blind testing
of dog teams had been included in training sessions, no formal
assessment of accuracy was conducted prior to this study. Conse-
quently, dog accuracy could potentially differ between dogs. After
a find, sample-related information including environmental data
and specific dog behaviour related to the detection were recorded.
All samples that dogs indicated on were collected independently
of morphological characteristics. Samples were stored in DNA-free
50 ml plastic containers (Sarstedt, Mawson Lakes) with 30 ml of 96
% ethanol and were kept cool until further processing.
43
We set up one camera trap in every grid cell, for four consecutive
weeks. Spatial array of camera traps approximated the standard
monitoring design of the BFNP (Weingarth et al., 2012). Duration
of camera trapping was set to be broadly congruent to the use of
detection dogs. Cuddeback (De Pere) models Capture, C1 and Attack
were used in white flash mode. Cameras took one picture per trigger
with no pre-set delay. Recovery time was 1.5 s (C1) or 2.0 s (Cap-
ture, Attack) during daytime and 10 s during nights. Cameras were
attached to trees or wooden poles facing linear structures such as
forest roads (Weingarth et al., 2012), tested after setup and checked
weekly.
DNA extractions were performed using QIAcube automated
DNA extraction devices (Qiagen, Hilden) in a clean laboratory ded-
icated to the handling of non-invasively collected material. DNA
extraction from scat samples was conducted using the QIAamp
DNA Stool Kit and a final elution volume of 120 ␮l. DNA from
hair samples additionally found by dogs at scent marking sites
was extracted using the QIAamp DNA Investigator Kit (Qiagen,
Hilden) as described in Steyer et al. (2016), with two consecu-
tive 40 ␮l elution steps. Three sequence fragments targeting the
mitochondrial control region were used for species identification
and haplotyping (see Suppl. 1). A 180 bp stretch (primers L16782
and H16922, Gugolz et al., 2008), a lynx-specific 248 bp fragment
(Lynxfwd4 and Lynxrev5) as well as a 234 bp fragment (Hcarn200
and CanidC1) suitable for mammal species identification (Nowak
et al., 2014) were sequenced to reveal species identity and control
region haplotypes (see Suppl. 1 for reaction conditions). Sequences
were assigned to haplotypes described in Gugolz et al. (2008).
For confirmed lynx samples individual assignment was performed
using a set of 19 microsatellites and two sex markers (Suppl. 1) with
three PCR replicates per sample (see Suppl. 1 for details on frag-
ment length analysis). Consensus genotypes were derived using a
custom R script (R Core Team, 2016) based on the algorithms used
in GIMLET 3.3 (Valière, 2002). Consensus genotypes were assigned
to individuals using a customized R script under consideration of
gender, haplotype, sampling date and location. A maximum of three
mismatching loci was accepted to assign a sample to the same
individual due to high error rates in non-invasive samples. Ampli-
fication success, allelic dropout and false alleles were calculated
using GIMLET 3.3. (Valière, 2002). Only lynx samples with an ampli-
fication rate of >0.4 across loci and replicates were considered for
further processing.
We used a single-season site occupancy model to estimate the
power of the dog-based survey method. The model (Suppl. 2) was
implemented in WinBUGs using the package R2WinBugs accord-
ing to Kéry et al. (2012) with 20.000 iterations, 3 chains, a burn-in
of 4.000 and a thinning of 5. Furthermore, we extended the model
by a second detection process based on camera traps. Therefore
occupancy estimation was based on both camera traps and detec-
tion dog survey. The data was subdivided into spatial replicates
(1 km transect segments) for the detection dog survey and temporal
replicates (one week) for the camera traps.
On 27 days one or both teams conducted detection dog surveys,
resulting in 43 team search days during which 440.3 km of transects
were covered. Excluding breaks and data recording times, teams
searched for 149.2 h. For more details regarding total and daily
distances covered by handlers and dogs, total and daily time in the
field and actual search time see Table 1.
The two detection dog teams found 165 scat and four hair
samples. Fifty scat and two hair samples (30.8 %) showed lynx hap-
lotypes, while 78 (46.2 %) yielded DNA of other species. For 23.1 %
of all samples no species could be identified. The dogs were trained
for and successfully detected scats of two further species, namely
wolf (Canis lupus) and European wildcat (Felis silvestris). Therefore,
false positive indication rate (=42.6 %, mainly fox (Vulpes vulpes)
44 L. Hollerbach et al. / Mammalian Biology 90 (2018) 42–46

Fig. 1. a) Permanent lynx distribution (Chapron et al., 2014), BFNP location in red, BFNP with transects, all samples found and genetically confirmed lynx samples; b) All grid
cells and genetically confirmed lynx samples, grid cells where lynx presence was confirmed by dogs only, camera traps only, and by camera traps and dogs. (For interpretation
of the references to colour in this figure legend, the reader is referred to the web version of this article.)
 L. Hollerbach et al. / Mammalian Biology 90 (2018) 42–46
Table 1
Details on time and spatial parameters of the detection dog survey in the BFNP.
Parameter Result
days on which searches were conducted 27
total team search daysa , b 43
total distance covered by handlersa [km] 440.3
daily distance covered by handlerb [km] 2.0–19.6 (mean = 10.2)
total distance covered by dogsa [km] 598.6
daily distance covered by dogb [km] 2.2–28.6 (mean = 13.9)
total time in the fielda [h] 192.6
daily time in the fieldb [h] 1.0–9.5 (mean = 4.5)
total search timea [h] 149.2
daily search timeb [h] 0.9–6.7 (mean = 3.5)
a
Both teams/handlers/dogs.
b
Individual team/handler/dog.
scats) was calculated excluding samples of other species the dogs
were trained to detect.
Lynx presence was confirmed by detection dogs in 21 of the 44
grid cells (47.8 %). In 14 of these cells camera traps did not detect
lynx. Camera traps confirmed lynx presence in 16 grid cells, of
which nine yielded no genetically confirmed lynx samples (Fig. 1b).
Based on our model, the chance for successful dog-based detec-
tion of genetically confirmed lynx samples is 0.13/transect km
(SD = 0.02). Due to imperfect detection we calculated that 10 km
of transect search per grid cell are required to obtain a 70 % proba-
bility to detect lynx presence. This probability was based on 1,000
simulations of the Bernoulli sampling process while referring to the
detection probability estimate given by the model.
Genotypes from 26 samples were suitable for individual dis-
crimination using microsatellite markers (allelic dropout = 31 %,
false alleles = 3.3 %) and yielded 11 individuals, with six females,
four males and one unknown sex (see Suppl. 3). Five individuals
were detected multiple times, with up to nine detections.
Our results show that detection dogs are a suitable tool to gain
samples for a systematic genetic lynx monitoring. Two dog teams
found 52 genetically confirmed lynx samples during four weeks.
Eleven individual genotypes could be identified from these sam-
ples. However, we found that a high number of searched kilometres
are required per grid to achieve a 70 % detection rate. In addition,
false positive samples were an issue in our study for both dogs.
Genetic analyses in our study indicated the presence of both lynx
and fox DNA in seven scat samples (data not shown). A morpho-
logical inspection suggested that those scats were likely from lynx.
Contamination with fox DNA presumably occurred through over-
marking of lynx scats with fox urine (Wikenros et al., 2017). This
contamination is not detectable for the dog handler in the field,
who rewards the dog for indicating on a scat that is morphologi-
cally classified as lynx. It is therefore possible that the dogs picked
up fox scent as “new target” from contaminated scats. Duggan et al.
(2011) and Goodwin et al. (2010) stress that minimising the rate
of missed targets at the cost of accepting high false positive rates
should be prioritised when working with cryptic and rare species. In
contrast, costs of genetic analysis can be high when false samples
are included. We decided to genetically analyse all samples that
the dogs indicated on, independent of visual assessment through
dog handlers. Costs of genetic analyses can be reduced by only
analysing promising samples or at least excluding obvious false
positives. However, chances are high that this would decrease the
total number of genetically confirmed lynx samples.
Our genotyping success was moderate but comparable to
respective success rates from lynx scat samples reported elsewhere
(Bull et al., 2016). Notably, we analysed all samples found by the
dogs irrespective of their morphological characterisation, leading
to unsuccessful analyses of old and weathered samples, which
would likely be sorted out a priori in other studies.
45
We are aware that with the detection rates found in this study,
successful scat detection is highly dependent on sample size and
therefore on monitoring effort. We used a very simplistic single-
season site-occupancy model, and we did not account for potential
density differences in the study area. Furthermore, we assumed a
closed population and we did not consider covariates such as spa-
tial distance, time, altitude, snow cover or distance to settlements.
In the framework of a more comprehensive study involving a sec-
ond study site, the current model will be extended to an n-mixture
model and spatial mark-recapture model including covariates.
Overall our study documents the great potential of detection
dogs to obtain sufficiently large samples sizes for an effective
genetic monitoring of the Eurasian lynx. As a complementary tool to
well-established camera trapping, this approach will help to study
and monitor the ongoing range dynamics of reintroduced lynx pop-
ulations in Central Europe. Rapid genetic erosion and inbreeding are
considered major threats for the long-term success of lynx rein-
troduction in this region. The overall goal is therefore to establish
a metapopulation with dispersal corridors effectively connecting
the different subpopulations in Germany, France, the Alpine region
and the eastern populations (Boitani et al., 2015). A sound genetic
monitoring strategy will allow to survey genetic diversity over time
and provide crucial data for the assessment of the population sta-
tus and implementation of conservation measures for landscape
connectivity.
Funding
Method development was supported by the Hessisches Lan-
desamt für Naturschutz, Umwelt und Geologie, Gießen, Germany.
Conflicts of interest
None.
Acknowledgements
Special thanks to Milena Bös for providing enormous support
in terms of dog training. We thank the BFNP administration for
granting permission to conduct this research and for providing
accommodation and equipment. We are thankful to Elena Jess for
leading dog team 2. Particular thanks go to Martin Gahbauer, Lilli
Evert, Tamara Gramlinger, Lea Wirk, Regine Frank, and the BFNP
volunteers who facilitated fieldwork. We are grateful to Berardino
Cocchiararo for conducting parts of the genetic analyses. Thanks to
Hannah Jüngling and Yvonne Puder for carrying out a large propor-
tion of lab work. Furthermore, we wish to express our appreciation
to the zoos and wildlife parks that provided training samples. And
finally: Good job, Nara and Maple!
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at https://doi.org/10.1016/j.mambio.2018.02.
003.
References
Alda, F., Inogés, J., Alcaraz, L., Oria, J., Aranda, A., Doadrio, I., 2008. Looking for the
Iberian lynx in central Spain: a needle in a haystack? Anim. Conserv. 11 (4),
297–305.
Barba, M., de Waits, L.P., Garton, E.O., Genovesi, P., Randi, E., Mustoni, A., Groff, C.,
2010. The power of genetic monitoring for studying demography, ecology and
genetics of a reintroduced brown bear population. Mol. Ecol. 19 (18),
3938–3951.
Boitani, L., Alvarez F., Anders, O., Andren, H., Avanzinelli, E., Balys, V., Blanco, J.C.,
Breitenmoser, U., Chapron, G., Ciucci, P., Dutsov, A., Groff, C., Huber, D., Ionescu,
O., Knauer, F., Kojola, I., Kubala, J., Kutal, M., Linnell, J., Majic, A., Mannil, P.,
46 L. Hollerbach et al. / Mammalian Biology 90 (2018) 42–46
Manz, R., Marucco, F., Melovski, D., Molinari, A., Norberg, H., Nowak, S., Ozolins,
J., Palazon, S., Potocnik, H., Quenette, P.-Y., Reinhardt, I., Rigg, R., Selva, N.,
Sergiel, A., Shkvyria, M., Swenson, J., Trajce, A., Arx, M., von, Wölfl, M.,
Wotschikowsky, U., Zlatanova, D., 2015. Key actions for large carnivore
populations in Europe. Report to DG Environment, European Commission,
Bruxelles. Institute of Applied Ecology, 120pp. (http://ec.europa.eu/
environment/nature/conservation/species/carnivores/pdf/key actions large
carnivores 2015.pdf. Accessed 8 November 2017).
Bull, J.K., Heurich, M., Saveljev, A.P., Schmidt, K., Fickel, J., Förster, D.W., 2016. The
effect of reintroductions on the genetic variability in Eurasian lynx
populations: the cases of Bohemian-Bavarian and Vosges-Palatinian
populations. Conserv. Genet. 17 (5), 1229–1234.
Cailleret, M., Heurich, M., Bugmann, H., 2014. Reduction in browsing intensity may
not compensate climate change effects on tree species composition in the
Bavarian Forest National Park. For. Ecol. Manage. 328, 179–192.
Chapron, G., Kaczensky, P., Linnell, J.D.C., Arx, M., von Huber, D., Andrén, H.,
López-Bao, J.V., Adamec, M., Álvares, F., Anders, O., Balčiauskas, L., Balys, V.,
Bed"o, P., Bego, F., Blanco, J.C., Breitenmoser, U., Brøseth, H., Bufka, L., Bunikyte,
R., Ciucci, P., Dutsov, A., Engleder, T., Fuxjäger, C., Groff, C., Holmala, K., Hoxha,
B., Iliopoulos, Y., Ionescu, O., Jeremić, J., Jerina, K., Kluth, G., Knauer, F., Kojola, I.,
Kos, I., Krofel, M., Kubala, J., Kunovac, S., Kusak, J., Kutal, M., Liberg, O., Majić, A.,
Männil, P., Manz, R., Marboutin, E., Marucco, F., Melovski, D., Mersini, K.,
Mertzanis, Y., Mysłajek, R.W., Nowak, S., Odden, J., Ozolins, J., Palomero, G.,
Paunović, M., Persson, J., Potočnik, H., Quenette, P.-Y., Rauer, G., Reinhardt, I.,
Rigg, R., Ryser, A., Salvatori, V., Skrbinšek, T., Stojanov, A., Swenson, J.E.,
Szemethy, L., Trajçe, A., Tsingarska-Sedefcheva, E., Váňa, M., Veeroja, R.,
Wabakken, P., Wölfl, M., Wölfl, S., Zimmermann, F., Zlatanova, D., Boitani, L.,
2014. Recovery of large carnivores in Europe’s modern human-dominated
landscapes. Science 346 (6216), 1517–1519.
Clare, J.D.J., Anderson, E.M., MacFarland, D.M., Sloss, B.L., 2015. Comparing the
costs and detectability of bobcat using scat-detecting dog and remote camera
surveys in Central Wisconsin. Wildl. Soc. Bull. 39 (1), 210–217.
Dahlgren, D.K., Elmore, D.R., Smith, D.A., Hurt, A., Arnett, E.B., Connelly, J.W., 2012.
Use of dogs in wildlife research and management. In: Silvy, N.J. (Ed.), The
Wildlife Techniques Manual, vol. 1, 7th ed. Johns Hopkins University Press,
Baltimore, pp. 140–153.
Davoli, F., Schmidt, K., Kowalczyk, R., Randi, E., 2013. Hair snaring and molecular
genetic identification for reconstructing the spatial structure of Eurasian lynx
populations. Mamm. Biol. 78 (2), 118–126.
Duggan, J.M., Heske, E.J., Schooley, R.L., Hurt, A., Whitelaw, A., 2011. Comparing
detection dog and livetrapping surveys for a cryptic rodent. J. Wildl. Manage.
75 (5), 1209–1217.
Goodwin, K.M., Engel, R.E., Weaver, D.K., 2010. Trained dogs outperform human
surveyors in the detection of rare spotted knapweed (Centaurea stoebe). Invas.
Plant Sci. Manag. 3 (2), 113–121.
Gugolz, D., Bernasconi, M.V., Breitenmoser-Würsten, C., Wandeler, P., 2008.
Historical DNA reveals the phylogenetic position of the extinct Alpine lynx. J.
Zool. 275 (2), 201–208.
Kéry, M., Schaub, M., Beissinger, S.R., 2012. Bayesian Population Analysis Using
WinBUGS: A Hierarchical Perspective, 1st ed. Academic Press, Boston (537 pp.).
Lucchini, V., Fabbri, E., Marucco, F., Ricci, S., Boitani, L., Randi, E., 2002. Noninvasive
molecular tracking of colonizing wolf (Canis lupus) packs in the western Italian
Alps. Mol. Ecol. 11 (5), 857–868.
Mumma, M.A., Zieminski, C., Fuller, T.K., Mahoney, S.P., Waits, L.P., 2015.
Evaluating noninvasive genetic sampling techniques to estimate large
carnivore abundance. Mol. Ecol. Resour. 15 (5), 1133–1144.
Nowak, C., Büntjen, M., Steyer, K., Frosch, C., 2014. Testing mitochondrial markers
for noninvasive genetic species identification in European mammals. Conserv.
Genet. Resour. 6 (1), 41–44.
R. Core Team, 2016. A language and environment for statistical computing. R
Foundation for Statistical Computing, Vienna, Austria https://www.R-project.
org/.
Schmidt, K., Kowalczyk, R., 2006. Using scent-marking stations to collect hair
samples to monitor Eurasian lynx populations. Wildl. Soc. Bull. 34 (2), 462–466.
Steyer, K., Kraus, R.H.S., Mölich, T., Anders, O., Cocchiararo, B., Frosch, C., Geib, A.,
Götz, M., Herrmann, M., Hupe, K., Kohnen, A., Krüger, M., Müller, F., Pir, J.B.,
Reiners, T.E., Roch, S., Schade, U., Schiefenhövel, P., Siemund, M., Simon, O.,
Steeb, S., Streif, S., Streit, B., Thein, J., Tiesmeyer, A., Trinzen, M., Vogel, B.,
Nowak, C., 2016. Large-scale genetic census of an elusive carnivore, the
European wildcat (Felis s silvestris). Conserv. Genet. 17 (5), 1183–1199.
Thüler, K., 2002. Spatial and temporal distribution of coat patterns of Eurasian
Lynx (Lynx lynx) in two re-introduced populations in Switzerland. KORA
Bericht 13 e. KORA, 35pp. https://www.kora.ch/fileadmin/file sharing/5
Bibliothek/52 KORA Publikationen/520 KORA Berichte/KORA 13 E 2002 Coat
Patterns of Eurasian Lynx.pdf. (Accessed 19 September 2017).
Valière, N., 2002. GIMLET: A computer program for analysing genetic individual
identification data. Mol. Ecol. Notes 2 (3), 377–379.
Wölfl, M., Bufka, L., Červený, J., Koubek, P., Heurich, M., Habel, H., Huber, T., Poost,
W., 2001. Distribution and status of lynx in the border region between Czech
Republic, Germany and Austria. Acta Theriol. 46 (2), 181–194.
Wölfl, S., Mináriková, T., Poledník, L., Bufka, L., Wölfl, M., Engleder, T., Belotti, E.,
Gahbauer, M., Heurich, M., Schwaiger, M., Poledníková, K., Volfová, J., Strnad,
M., 2015. Status and distribution of the transboundary lynx population of
Czech Republic, Bavaria and Austria in the lynx year 2014. Report of
Trans-Lynx Project (12 pp. Accessed 2 November 2017).
Wasser, S.K., Davenport, B., Ramage, E.R., Hunt, K.E., Parker, M., Clarke, C.,
Stenhouse, G., 2004. Scat detection dogs in wildlife research and management:
application to grizzly and black bears in the Yellowhead Ecosystem, Alberta,
Canada. Can. J. Zool. 82 (3), 475–492.
Weingarth, K., Heibl, C., Knauer, F., Zimmermann, F., Bufka, L., Heurich, M., 2012.
First estimation of Eurasian lynx (Lynx lynx) abundance and density using
digital cameras and capture–recapture techniques in a German national park.
Anim. Biodivers. Conserv. 35 (2), 197–207.
Weingarth, K., Zeppenfeld, T., Heibl, C., Heurich, M., Bufka, L., Daniszová, K., Müller,
J., 2015. Hide and seek: extended camera-trap session lengths and autumn
provide best parameters for estimating lynx densities in mountainous areas.
Biodivers. Conserv. 24 (12), 2935–2952.
Wikenros, C., Jarnemo, A., Frisén, M., Kuijper, D.P.J., Schmidt, K., 2017.
Mesopredator behavioral response to olfactory signals of an apex predator. J.
Ethol. 35 (2), 161–168.
Woollett, D.A., Hurt, Aimee, Richards, Ngaio L., 2014. The current and future roles
of free-ranging detection dogs in conservation efforts. In: Gompper, M.E. (Ed.),
Free-ranging Dogs and Wildlife Conservation. Oxford Univ. Press, Oxford, pp.
239–264.