Content uploaded by Denise O'Meara
Author content
All content in this area was uploaded by Denise O'Meara
Content may be subject to copyright.
ORIGINAL PAPER
A non-invasive approach to determining pine marten
abundance and predation
Emma Sheehy &Denise B. O’Meara &Catherine O’Reilly &
Anthony Smart &Colin Lawton
Received: 1 June 2013 /Revised: 2 October 2013 /Accepted: 4 October 2013
#Springer-Verlag Berlin Heidelberg 2013
Abstract A non-invasive approach was used to investigate
variation in pine marten (Martes martes) abundance between
the midlands and east of Ireland, and to determine the frequen-
cy of occurrence of squirrels and other small mammals in the
diet. Remotely plucked hair samples were genotyped to differ-
entiate between individual animals, and real-time polymerase
chain reaction (PCR) was used to identify predator and prey
DNA in scats. Macro analysis of prey remains was carried out
on a sub sample of scats and the results from both methods are
compared. Non-invasive techniques were successful in deter-
mining the presence and relative abundance of the pine marten
at woodland level. As expected, abundance was found to be
higher in the core population of the midlands than in the east.
Pine martens were found to reach higher numbers per km
2
of
forested habitat in Ireland than their British or European coun-
terparts. Both traditional hard part analysis and molecular die-
tary analysis of mammalian prey yielded similar results. We
provide the first evidence of the European pine marten
predating upon the North American grey squirrel (Sciurus
carolinensis) in its invasive range. While the grey squirrel
was not available as a prey item in any of the midlands sites,
it was available in the east, where it featured significantly more
frequently in the diet than the native red squirrel. In both the
midlands and the east the woodmouse is the most frequently
occurring mammal in the diet.
Keywords Pine marten .Squirrel .Hair sampling .
Genotyping .Macro faecal analysis .Prey DNA
Introduction
In the nineteenth and twentieth centuries, the European pine
marten (Martes martes) population in Ireland experienced
widespread decline as a result of habitat loss (large-scale
deforestation) and heavy persecution (O'Sullivan 1983).
Population censuses in the 1980s (O'Sullivan 1983) and again
in 2005 (O'Mahony D, O'Reilly C, Turner P 2006 National
pine marten survey of Ireland 2005)revealedthatpinemarten
distribution in Ireland is still mainly concentrated around core
populations in the west and midlands, along with several
smaller populations in the south west and south east of the
country. However, the pine marten population in the west and
midlands of Ireland has undergone a range expansion in recent
decades, as a result of increased habitat availability and con-
nectivity through afforestation, and importantly protection by
law (O'Mahony D, O'Reilly C, Turner P 2006 National pine
marten survey of Ireland 2005). The most recent population
estimate for the island of Ireland is 3060 individuals
(O'Mahony et al. 2012), although there is still relatively little
known about Irish pine marten population densities in the
westernmost part of their European range. The European pine
marten has traditionally been considered a forest specialist.
Zalewski and Jedrzejewski (2006) estimated that 2 km
2
is the
minimum area of forested habitat necessary to support an
adult pine marten in the temperate forest zone. Despite an
extremely fragmented forest landscape (with <11 % forested
area, Ireland represents the lowest forested land cover in their
range), previous population studies in Ireland (Lynch et al.
2006; Mullins et al. 2010) have found pine marten density can
be higher in Ireland than is typical throughout their British and
European ranges (Birks 2002; Zalewski and Jedrzejewski
Communicated by C. Gortázar
E. Sheehy (*):A. Smart :C. Lawton
Mammal Ecology Group, School of Natural Sciences, Ryan Institute,
National University of Ireland, Galway, Ireland
e-mail: emmasheehy@gmail.com
D. B. O’Meara :C. O’Reilly
Molecular Ecology Research Group, Waterford Institute of
Technology, Cork Road, Waterford, Ireland
Eur J Wildl Res
DOI 10.1007/s10344-013-0771-2
2006;Mergeyetal.2011; Caryl et al. 2012a). The Irish studies
were conducted on the smaller population pockets in the south
west (Killarney, Co. Kerry) and the south east (Portlaw, Co.
Waterford). A population density estimate for the core popu-
lation, and also the part of their range where they are believed
to be much less common (i.e., the east of the country), has yet
to be determined.
In Ireland and Scotland, it has been anecdotally suggested
that the recovering pine marten population may be inhibiting
the spread of the invasive North American grey squirrel
(Sciurus carolinensis ), and indeed that the grey squirrel pop-
ulation has retracted in the presence of pine martens in both
countries (Carey et al. 2007; Paterson and Skipper 2008).
Published information on red (Sciurus vulgaris) and grey
squirrel distribution in Ireland (Carey et al. 2007) and more
recent studies on squirrel distribution (unpublished data from
Sheehy and Lawton) have found the grey squirrel to be rare in
the midlands of Ireland, but their range potentially overlaps
with that of the pine marten in the east. It has been suggested
that the European pine marten preys preferentially on the
larger, less arboreal of the two squirrel species; however, there
has been no evidence to date of the European pine marten
predating upon the North American grey squirrel anywhere in
its invasive range, which only overlaps to a small extent with
that of the pine marten (see maps in Carey et al. 2007 and
O'Mahony et al. 2012).
Non-invasive genetic studies to identify species distribu-
tion and population size have become important tools to aid
the study of wild, and particularly elusive, carnivore popula-
tions such as martens (Mowat and Paetkau 2002; Williams
et al. 2009). Hair sampling is commonly used to non-
invasively survey for the presence of mammal species (e.g.,
Scotts and Craig 1988; Lindenmayer et al. 1999). Lynch et al.
(2006) found hair traps (fur-snagging devices) both quick and
reliable in detecting pine marten presence in broadleaved
woodlands. Subsequently, Mullins et al. (2010)optimiseda
panel of microsatellite loci to identify unique genotypes with-
in the Irish pine marten population, thus enabling distribution
and abundance to be established reliably through non-invasive
field studies such as hair trapping. Faecal analysis of scats is
used to determine species distribution (e.g., Palomares et al.
2002) and individual identity (e.g., Ruiz-González et al. 2013)
of carnivores. The use of molecular techniques in the analysis
of carnivore diet has also become popular in recent years
(Deagle et al. 2005;Dunshea2009; Shehzad et al. 2012), as
prey DNA found in scats can be identified to taxon and
species level and is not dependent on hard parts surviving
digestion. Molecular techniques have recently been optimised
to specifically detect the presence of mammalian prey in the
diet of the Irish pine marten (O'Meara et al. 2013).
Using these recently developed non-invasive techniques,
this study firstly aims to quantify pine marten abundance in
the fragmented forest habitat in their core range in the
midlands of Ireland and in the east of the country where they
are considered to be less common (O'Mahony D, O'Reilly C,
TurnerP2006NationalpinemartensurveyofIreland2005;
O'Mahony et al. 2012). Secondly, the study aims to quantify
the frequency of occurrence of small mammals in the diet with
emphasis on red and grey squirrels. In the process, we aim to
quantify scat density in the midlands and eastern regions and
to compare the findings of both molecular and macro dietary
analysis techniques.
Materials and methods
Field methods
Study area
The primary study area consisted of counties Laois and Offaly,
in the midlands of Ireland, and the secondary study area was
county Wicklow, in the east of the country, where the pine
marten population is considered to be less abundant
(O'Mahony et al. 2012)(Fig.1). Hair trapping sites (abun-
dance study, n=5) and scat based survey sites (dietary analy-
sis, n=23) are described below.
Abundance study
Five sites were selected, a broadleaved and a predominantly
coniferous woodland from each study area (with two
broadleaved woods examined sequentially in Co. Wicklow)
(Fig. 1). Site 1, Charleville Forest, Co. Offaly, is a mature
broadleaved wood (ca. 113 ha) in which oak (Quercus robur )
is the dominant tree species. Site 2, Clonad (ca. 143 ha), Co.
Offaly, is a mixed, mainly coniferous woodland situated
1.5 km from Charleville, where Norway (Picea abies )and
sitka spruce (Picea sitchensis) are the dominant tree species.
Site 3, Cloragh (ca. 160 ha) is located in Ashford, Co.
Wicklow. The dominant tree species present are sitka spruce
and Douglas fir (Pseudotsuga spp.). Site 4, Knocksink nature
reserve (ca. 60 ha), is located in Enniskerry, Co. Wicklow. It
consists of mature oak (Quercus petraea) and mixed wood-
land. Site 5, Tomnafinnoge (ca. 80 ha) in south Wicklow, is a
broadleaved woodland consisting mainly of mature oak
(Quercus petraea ) (Fig. 1). Sites 2 and 4 are considered
discrete woodlands as there is no forested habitat within
1 km of these sites. Site 1 is separated from adjacent forest
habitat by a new primary road. Site 3 is separated from
adjacent forestry by the Vartry river. Site 5 is not separated
from adjacent forestry by either natural or man-made bound-
aries, and as such represents the only site which is indiscrete in
this study (Fig. 1). Spring based hair traps as described by
Messenger and Birks (2000) were installed in sites 1 (n=5), 2
(n=7), 3 (n=8),and4(n=4) in March 2011 for a period of
Eur J Wildl Res
14 months, with the exception of site 4, where the hair traps
were moved to site 5 in October 2011. Each trap was checked
for samples and rebaited once per month with chicken and the
tree trunk was smeared with marmalade. Traps were positioned
450 m apart at a density of one trap per 20 ha throughout the
sites as previous home range sizes for the pine marten in Ireland
have been reported as ≥0.2 km
2
(Birks 2002). Animals that
were genetically identified in 3 or more months (including at
least 1 month between November 2011 and April 2012) were
assumed to be resident adults. Abundance values were then
obtained by applying the number of residents adults identified
at each site to the corresponding forested sampling area.
Dietary study
Scats were collected between March 2010 and August 2012
from 23 sites throughout the midlands and the east of Ireland
(Fig. 1) and stored at −20 °C. In order to ensure reliability and
validity of both scat density and dietary analysis, all scats
collected as part of this study were subjected to DNA analysis
in order to confirm they were of pine marten origin.
In order to determine whether a potential prey species was
being preyed upon, it was necessary to establish first of all that it
was available as a prey item. Of the 23 scat collection sites, 17
were classified as being either red squirrel or grey squirrel
positive sites in accordance with the findings of concurrent
squirrel distribution studies (Table 1) (unpublished data from
Sheehy and Lawton). A sample of scats from the west of
Ireland that were collected during the course of a red squirrel
population study (Waters and Lawton 2011) were also included.
Woodlands where both grey squirrels and pine marten were
confirmed as being present together were identified as key sites
and revisited where possible to increase the sample size of scats
collected from these zones. In August 2012, a scent detection
dog, specially trained to detect pine marten scat, carried out
searches in three woodlands (including abundance study site 5,
Tomnafinnoge and two further sites in the east). With the excep-
tion of the searches made by the scent detection dog, and a few
scats that were collected opportunistically, the distance covered
Fig. 1 Location of the study area in Ireland (inset ) and the locations of
the hair trap sites within the fragmented forested landscape. Non-forested
land cover is white , forested habitat is grey and hair trap sites are
represented in black . Scat survey sites are indicated by open diamonds.
The position of each hair trap within the study sites are also shown
Eur J Wildl Res
during the course of scat collection was used to estimate scat
density, where scat density = (no. of scats collected)/(distance
walked). All scats were collected by the lead author and transects
were walked once per month in the abundance study sites, and in
the rest of the sites either once or twice in total. A Fisher exact
test was used to test for an overall difference in accuracy in scat
identification in the field between the midlands and eastern sites,
and to test for a significant difference in the frequency of
occurrence of red and grey squirrels in the diet.
Laboratory methods
Abundance study
Molecular analysis
Genomic DNA was isolated from (n=158) hair samples
using The ZR Genomic DNA II Kit™(ZYMO Research,
California, USA) using the protocol for hair extraction
(ZYMO RESEARCH Cat no. D3040). The DNA was eluted
with 100 μl of deionised water. Real-time polymerase chain
reaction (PCR) was used for species (targeting mitochondrial
DNA) and sex determination (targeting ZFX and ZFY se-
quences on the X and Y chromosome) of the hair samples as
described by Mullins et al. (2010). The C
t
value for the ZFX
gene was used to screen the samples for genotyping suitability.
Samples with a C
t
value of less than 36 were deemed to
contain adequate quantities of nuclear DNA for genotyping.
Genotyping
Samples that were deemed suitable for genotyping were
screened in duplicate at seven loci (Ma2-mini, Mel1, Gg7-
mini, Mvi1341, Mvi1354, Mvis075, Ggu234) (Table 2). As
the samples used for genotyping came from a non-invasive
source (remotely plucked hairs), each sample was indepen-
dently genotyped twice. Scores were only recorded if they
were observed twice and exactly matched. Samples that were
not replicated after the first two PCRs were repeated. Details
of primers and multiplex setup are provided in Table 2.
Fragment Analysis was conducted on an ABI PRISM 310®
Genetic Analyser (Applied Biosystems) according to the man-
ufacturer's instructions with the standard run module. Alleles
were scored with GS500 LIZ™size standard using
GeneMapper software v3.7 (Applied Biosystems). Two au-
thors independently called alleles.
Data analysis
The two genotyping replicates were compared to assess the data
for genotyping errors including the presence of allelic drop out
and false alleles using GIMLET version 1.3.4 (Valière 2002).
PCR success rates were also calculated using GIMLET version
1.3.4. The occurrence of repeated genotypes was identified
using GENALEX version 6.4.1 (Peakall and Smouse 2006)
and the number of replicates or individual recaptures was re-
corded. GENALEX was also used to estimate the probability of
identity (PID). A final dataset was created with duplicated data
removed and MICROCHECKER version 2.2.3 (Van
Oosterhout et al. 2004) was used to further identify possible
genotyping errors, including the presence of null alleles, large
allele dropout, and scoring errors as a result of stutter peak
(using default settings).
Gametic phase linkage disequilibria by Fisher's method
(1,000 dememorizations and 5,000 iterations) and deviations
from Hardy–Weinberg equilibrium were assessed (default set-
tings, exact tests) using GENEPOP version 4.0.10 (Rousset
2009). Observed (H
O
) and expected (H
E
) heterozygosities and
thenumberofalleles(a), were calculated using GENALEX
version 6.4.1 (Peakall and Smouse 2006), and allelic richness
(R
s
) was estimated using FSTAT 2.94 (Goudet 1995).
Tabl e 1 Scat surveys in sites where squirrel distribution studies had
taken place (unpublished data from Sheehy and Lawton; Waters and
Lawton 2011) resulted in a total of 361 scats for analysis of squirrel in
the diet
Site Location RS GS RS
site
scats
GS
site
scats
RS
prey
GS
prey
Charleville M Y N 162 4
Clonad M Y N 109 3
Abbeyleix Demesne M Y N 7 1
Birr Castle M Y N 1
Ballykilcavan M Y N 16
Cappard M Y N 6
Emo Park M Y N 9
Ballyteige M Y N 0
Garryhinch M Y N 11
Derryclare W Y N 6
Croneybyrne E Y N 1
Clara E Y N 1
Tomnafinnoge
a
ENY 20 5
Dollardstown
a
ENY 8
Ballygannon E N Y 4
Mullaghreelan
a
ENY 0
Oakpark E N Y 0
Jenkinstown E N Y 0
Total 329 32 8 5
2.4 %
FO
15.6 %
FO
RS red squirrel, GS grey squirrel, Mmidlands, Wwest, Eeast, Ysquirrel
species detected during field survey, Nsquirrel species not detected
during field survey
a
Site re-visited with scent detection dog
Eur J Wildl Res
Dietary study
Molecular analysis
Approximately 0.2 g of scat was used for DNA extraction as
described by O'Reilly et al. (2008), and using the ZR Genomic
DNA II Kit™(ZYMO Research). Pine marten DNAwas verified
as described above. All samples with a C
t
value lower than 32
were classified as pine marten and those with a greater C
t
value
were classified as non pine marten and excluded from further
analysis. To test for prey DNA in the confirmed pine marten
scats, species-specific Taqman assays designed to detect red and
grey squirrel DNA were used. All PCR reactions and probes
were as described by O'Meara et al. (2012). A sub-sample of 160
scats (80 each from sites 1 and 2) were also tested for small
mammal prey DNA; woodmouse (Apodemus sylvaticus), bank
vole (Myodes glareolus), pygmy shrew (Sorex minutus ), and
greater white toothed shrew (Crocidura russula)(O'Mearaetal.
2013). Samples with C
t
values of 36 or higher were discounted
and positive results were replicated for verification. Percentage
frequency of occurrence (%FO) in the diet for each prey species
was calculated as the number of scats in which the species' DNA
was amplified/total no. of scats tested × 100.
Macro analysis
A sub-sample of 110 scats was subjected to traditional hard part
analysis to identify mammalian prey using keys to identify
mammal bones (Yalden and Morris 1990) and hairs (Teerink
1991). The subsample of 110 scats comprised 40 scats from both
the Charleville and Clonad subsamples, respectively (which had
been tested for squirrel and other small mammalian prey DNA),
and a further 30 scats from the grey squirrel positive sites (which
had been tested for squirrel DNA only). The results from molec-
ular and macro analysis were compared and then combined to
determine an overall frequency of occurrence in the diet for each
prey species. A chi square test was used to investigate significant
differences in the frequency of occurrence of each species ac-
cording to molecular, macro and combined results. Regression
analysis was used to investigate whether a relationship exists
between the frequency of occurrence of prey items using molec-
ular and macro techniques. %FO for mammalian prey species
was calculated as in molecular analysis and percentage relative
biomass of prey ingested (%BPI) was calculated as: weight of
dried remains for each species/total weight of dried remains.
Previous studies investigating the contribution the main food
groups make in terms of biomass to the diet have used pre-
established correction factors in such estimations (Lynch and
McCann 2007; Caryl et al. 2012b); these correction factors were
derived from feeding trials in which the weight of the food item
eaten was divided by the dry weight of undigested matter later
identified in scats (Lockie 1961;Balharry1993; Jedrzejewska
and Jedrzejewski 1998; Lanszki et al. 2007). Individual correc-
tion factors for the mammalian prey species investigated in the
current study were not available as they are usually simply
grouped together in feeding trials as 'small mammals'.
Results
Density study
A total of 157 hair samples were collected out of 273 baited
hair traps. Sites 1 and 2 in the midlands yielded the highest
Tabl e 2 Microsatellite primers used in pine marten genotyping
Locus Primer sequence 5′–3′Size range Reference
Ma2-mini F: YAK-CCATGTACTTTTCCTATCTTTTAGGA
R: ATCTTGCATCAACTAAAAAT
131–141 O'Reilly (This study)
Davis and Strobeck (1998)
Mel1 F: FAM-CTGGGGAAAATGGCTAAACC
R: GCTCTTATAAATCTGAAAATTAGGAATTC
106–116 Bijlsma et al. (2000)
Mullins et al. (2010)
Gg7-mini F: FAM-GTTTTCAATTTTAGCCGTTCTG
R: GCTCTTCACTCTGTGTGGCATCTAC
132–140 Davis and Strobeck (1998)
O'Reilly (This study)
Mvi1341* F: PET-GTGGGAGACTGAGATAGGTCA
R: GTTTCTTGGCAACTTGAATGGACTAAGA
164–178 Vincent et al. (2003)
Mvi1354* F: FAM-CCAACTGGAGCAAGTAAAT
R: GTTTCTTCATCTTTGGGAAAGTATGTTT
200–212 Vincent et al. (2003)
Mvis075* F: FAM-GAAATTTGGGGAATGCACTC
R: GTTTCTTGGCAGGATAGGATGTGAGCT
145–155 Fleming et al. (1999)
Ggu234 F: PET-TTACTTAGAGGATGATAACTTG
R: GAACTCATAGGACTGATAGC
84–90 Duffy et al. (1998)
Reverse primers marked with an asterisk (*) were modified with a 5′sequence of GTTTCTT to promote non-templated nucleotide addition (Brownstein
et al. 1996). The Ma2-F and Gg7-R primers were redesigned to produce a smaller product. The primers were used in two multiplex mixes. Mix A contained
Gg7-mini and Mvi1354 and Mix B contained all the other primers. Each primer pair was at a final concentration of 0.5 μM. Microsatellite amplifications
were performed in a total volume of 10 μlwith4μl DNA extract, 1 μlprimermixand5μl GoTaq® Hot Start Green Master Mix (Promega). The PCR
conditions were 95 °C for 5 min followed by 30 cycles of 95 °C for 30 s, 60 °C for 90 s and 72 °C for 30 s, followed by 72 °C for 30 min
Eur J Wildl Res
success rates with 91 % and 78 %, respectively. In site 3, 37 %
of potential trapping events yielded hair samples. In site 4, one
hair sample was obtained out of a possible 24, and this was the
only hair sample in the study to test as negative for pine marten
DNA. Site 5 yielded nine hair samples, a success rate of
37.5 %. A further hair sample was collected from a roadkill
animal in July 2011, ca. 3 km from site 2, bringing the total
number of hair samples to 158.
Of the 158 hair samples, 157 were successfully genetically
identified as pine marten and 139 were successfully sex-typed.
109 samples had a ZFX C
t
value ≤36 and of these, 104 were
successfully genotyped (at six or more loci) (i.e., 95 % of
samples that passed the screening process, or 66 % of all hair
samples collected). This success rate varied between individ-
ual sites (Fig. 2).
A total of 25 individual genotypes were obtained from the
104 samples; one from the roadkill animal near site 2, and 24
from the abundance study sites (site 1, n=6 pine marten
detected with five hair traps; site 2, n=10pinemartendetected
with four to seven hair traps; site 3, n=6 pine marten detected
with eight hair traps; site 5, n=2 pine marten detected with
four hair traps). Pine marten abundance values ranged from 0
to 4.42 per km
2
including adult residents only (Table 3). Mean
abundance values for the midlands and east were 3.13 and
1.01, respectively, with an overall abundance value of 1.99
pine marten/km
2
(Table 3). The number of hair samples ge-
notyped from each individual ranged from 1 to 14 with a mean
value of 4.29 (±1.6, 95 % confidence interval [CI]) and the
number of months each animal was captured ranged from 1 to
7, with a mean value of 3.29 (±0.96, 95 % CI) (Table 4). One
individual (a male) was detected in both site 1 and site 2
(January 2012 and May 2011, respectively). These sites are
located 1.5 km apart (Fig. 1) and together they comprise less
than 3 km
2
of forested habitat. In total, eight adult residents
were detected in the two sites and a further eight that are
assumed to be either sub-adult or non-resident individuals.
With the exception of the one animal there was no further
crossover of individuals detected between these two wood-
lands, despite their close proximity to one another and the lack
of surrounding forested habitat.
Assessing genotyping errors
The proportion of positive PCRs ranged from 90 % to 100 %
across loci and from 86 % to 100 % across samples. Analysis
of genotyping error revealed the presence of allelic dropout
rates of 0.08 at locus Ma2-mini, 0.10 at locus Ggu234 and
0.46 at locus Mvis1354, with no false alleles detected. The
overall allelic dropout error rate across all loci was 0.09, 0.32
across all samples, and 0.08 across all PCRs. We found that
there was no systemic evidence of scoring errors and the data
was not shown to be affected by the systemic presence of null
alleles or large allelic dropout. The cumulative PID was
PI=1.9×10
−3
, which is sufficient for the estimation of popu-
lation size (0.01) (Mills et al. 2000).
Genetic variability
The number of alleles was low ranging from 2 at Ggu234 and
Mvis075 to 4 at Gg7-mini (Table 5). Low levels of allelic
richness per locus and per sample were also observed from
2 at Ggu234 and Mvis075 to 3.69 at Gg7-mini. Expected
levels of heterozygosity averaged 0.386 and ranged from
0.106 at Mvis1354 to 0.570 at Ma2-mini. Observed levels of
heterozygosity averaged 0.462 and ranged from 0.111 at
Mvis1354 to 0.708 at Gg7-mini and Ma2-mini (Table 5).
There were no significant deviations from Hardy–Weinbe rg
expectations at any loci.
Dietary study
A total of 517 scats were collected between March 2010 and
August 2011. Four hundred of these were collected in the
midlands (n=9 sites) and 117 were collected in the eastern
sites (n=14 sites) (Fig. 1). Overall, 86 % of scats collected in
the midlands tested positive for pine marten DNA compared
to 39 % in the east. As such, accuracy in the field was found to
be significantly lower in the eastern region (p<0.001, Fisher
exact). The majority of scats were included in scat density
calculations as distance walked was known (333 out of 344
and 38 out of 46 in midlands and eastern sites, respectively).
Thus scat density was estimated to be 1.745 scats/km in the
midlands and 0.221 scats/km in the east. Although scat den-
sity was found to be higher in areas of higher pine marten
occupancy, no statistically significant relationship was found
between scat density and pine marten abundance. The scent
detection dog succeeded in collecting seven pine marten scats
from two of three woods visited over a 2-day period. When
scats were categorised into red squirrel and grey squirrel sites
(as per unpublished data from Sheehy and Lawton) a total of
329 were classified as coming from red squirrel positive sites,
and 32 from grey squirrel positive sites (Table 1). Squirrels
appeared in the diet as prey items at low frequencies during
9 months of the year, spring and early summer being most
common (Table 6).
Molecular and macro analysis
Regression analysis found a linear relationship to exist be-
tween the %FO of mammalian prey items as detected by
molecular and macro analyses (y=−1.694+ 1.969x,R
2
=
0.895, p<0.05) (Fig. 3). The woodmouse featured more fre-
quently in the diet than any other mammal species in both the
molecular (χ
2
=41.17) and the macro (χ
2
=67.58) analyses
(df =4, p<0.01) (Fig. 4). Grey squirrels featured more fre-
quently than red squirrels as prey items in both analyses,
Eur J Wildl Res
significantly so in the molecular analysis, despite the consid-
erable difference in sample size (p<0.05, Fisher exact).
When results from the reduced sample (n=110) that was
the subject of both molecular and macro analyses were com-
bined, the %FO increased for all species; however, there was
very little effect on the low frequency prey items (red squirrel,
bank vole and pygmy shrew). The %FO for woodmouse and
grey squirrel increased more considerably when results from
both methods were combined, but not significantly so (Fig. 5).
%BPI also found the woodmouse to be the most important
prey species in the diet of the pine marten, followed by the
grey squirrel where it was available, although ca. 30 % of
mammalian remains could not be identified to species level
(Fig. 6).
Discussion
Pine marten abundance
This study has provided an index of abundance for the pine
marten population in both their core range and a considerably
less populated part of their range in the east of Ireland.
Abundance values are not directly comparable to European
studies where density values were obtained from radio-
tracking or snow tracking (e.g., Zalewski and Jedrzejewski
2006;Mergeyetal.2011), but are comparable to Irish studies
that have used a combination of hair trapping and live-
trapping todetermine population density estimates.The abun-
dance valuesobtained in thisstudy suggest that the population
in the midlands of Ireland (mean value of 3.13 adult residents/
km
2
) is currently living at a higher density than previously
reported for the species in Europe (0.01–1.75 per km
2
)
(Zalewski and Jedrzejewski 2006) or Ireland (0.5–2per
km
2
)(Lynchetal.2006; Mullins et al. 2010) and thus quite
possibly represents the highest density in their natural range. It
is unclear why the European pine marten reaches these rela-
tively high numbers in Ireland, particularly when their
favoured habitat is so sparse and fragmented. In a review of
available literature on European pine marten densities,
Zalewski and Jedrzejewski (2006) found that between 41°
and 68°N densities declined exponentially with decreasing
winter temperature and increasing seasonality, and suggested
that both winter severity and availability of rodents are
Fig. 2 The number of successful
hair trapping events and the
portion of hair samples
successfully genotyped from each
site
Tabl e 3 The total number of pine marten identified by unique genotypes at each site, in total (All), in sites 1 and 2 combined (Midlands) and in sites 3, 4
and 5 combined (East)
Site name Site area Total PM identified Mean captures No. adult residents No. adult residents per km
2
Midlands 1. Charleville 1.13 km
2
6 4.67 (0.98) 5 4.42
2. Clonad 1.43 km
2
10 2.8 (0.79) 3 2.10
East 3. Cloragh 1.58 km
2
6 2.83 (0.75) 2 1.27
4. Knocksink 0.6 km
2
0 0.00 0 0.00
5. Tomnafinnoge 0.8 km
2
2 2.5 (1.5) 1 1.25
Total All 5.54 km
2
24 3.29 (0.46) 11 1.99
Midlands 2.56 km
2
16 3.56 (0.63) 8 3.13
East 2.98 km
2
8 2.75 (0.62) 3 1.01
Mean Captures = mean number of times each animal was captured per site (Std Er). No. adult residents = number of animals that were detected in
≥3 months including at least 1 month during November 2011 and April 2012
Eur J Wildl Res
limiting factors on populations. Thus it is possible that Ireland's
relative lack of seasonality and mild winters (the moderating
influence of the Atlantic gulf stream results in mean minimum
winter temperatures of 2–6 °C) (MetEireann 2012) contribute
to the observed high pine marten abundance. Other contribu-
tory factors may include lack of competition and lack of
predators, with the red fox (Vulpes vulpes)representingthe
pine marten's only real competitor or predator in Ireland.
Zalewski and Jedrzejewski (2006)estimatedthat2km
2
is
the minimum area of forested habitat necessary to support an
adult pine marten in the temperate forest zone. However, this
does not appear to apply to the core Irish pine marten popu-
lation and both midlands sites, Charleville (1.13 km
2
)and
Clonad (1.43 km
2
), sustain relatively high pine marten num-
bers in comparison to other extensively forested parts of
Europe (Zalewski and Jedrzejewski 2006). The European pine
marten has traditionally been considered a forest specialist;
however, recent studies have found the species to be less
Tabl e 4 The site each pine marten was recorded at and the sex assigned
through DNA analysis
Animal Site Sex GT Months
CVF01 1. Charleville Female 1 1
CVF02 3 3
CVF03 10 7
CVF04 6 4
CVM01 Male 9 7
CVM02 14 6
CDM05
a
11
CDF01 2. Clonad Female 4 4
CDF02 1 1
CDF03 3 3
CDF04 1 1
CDM01 Male 1 1
CDM02 3 2
CDM03 1 1
CDM04 12 8
CDM05
a
11
CDM06 7 6
WWF01 3. Cloragh Female 2 2
WWF02 3 3
WWM01 Male 7 5
WWM02 1 1
WWM03 6 5
WWM04 1 1
TFF01 5. Tomnafinnoge Female 4 4
TFF02 Male 1 1
Roadkill 1
Total 104
GT the number of hair samples successfully genotyped for each animal,
Months the number of months each animal was identified
a
Animal was captured at two sites
Tabl e 5 Descriptive statistics for microsatellite analysis of pine martens in four study sites in Ireland
Ggu234 Mel1 Gg7-mini Ma2-mini Mvis075 Mvi1341 Mvi1354 Mean
N24 24 24 24 24 24 18 23.14
a23 4 3 2 3 3 2.86
R
s
2.0 2.99 3.69 3.0 2.0 2.75 3.0 2.78
a
s
87–93 108–116 132–142 131–137 151–153 170–180 200–208
H
E
0.353 0.379 0.548 0.570 0.305 0.442 0.106 0.386
H
O
0.458 0.417 0.708 0.708 0.292 0.542 0.111 0.462
HW 0.145 0.861 0.775 0.889 0.834 0.632 0.996
Ndenotes the number of individuals that successfully amplified at each locus, ais the number of alleles per locus, R
s
is the allele size range, A
s
is the
allele size, H
E
is the expected heterozygosity, H
O
is the observed heterozygosity. There were no significant deviations from Hardy–Weinb erg
equilibrium
Tabl e 6 Squirrels as detected as prey items in molecular and macro
analysis of pine marten scats, including the date and site at which the
scat was collected
Species Date Site Molecular Macro
RS 2010-03-22 Abbeyleix Y
RS 2010-11-16 Clonad Y N
RS 2011-03-31 Charleville Y Y
RS 2011-04-01 Charleville Y
RS 2011-05-11 Charleville Y
RS 2011-06-07 Clonad Y Y
RS 2011-06-07 Clonad N Y
RS 2011-09-30 Charleville Y
GS 2011-10-01 Tomnafinnoge Y
GS 2012-02-03 Tomnafinnoge N Y
GS 2012-03-13 Tomnafinnoge Y Y
GS 2012-04-18 Tomnafinnoge N Y
GS 2012-05-18 Tomnafinnoge Y N
RS red squirrel, GS grey squirrel, Ypositive, Nnegative (blank = not tested)
Eur J Wildl Res
restricted to large forests than previously believed and
highlighted the importance of the surrounding landscape not
only in providing habitat corridors but also in providing
essential food resources and den sites throughout fragmented
landscapes (Clevenger 1994;Pereboometal.2008; Mergey
et al. 2011; Caryl et al. 2012a). In the current study, hair traps
were only placed within the forested habitat and thus no data
was obtained on the use of the surrounding landscape, or in
the case of site 5, the adjacent forestry. Abundance values per
km
2
are thus only applicable to forested area as use of sur-
rounding landscape is not accounted for with this sampling
technique. The fact that only one animal was detected both in
sites 1 and 2 despite their close proximity to one another,
supports the theory that these small woodlands can be
Fig. 3 A linear relationship was
found to exist between the
frequency of occurrence of
mammalian prey items as
determined using molecular and
macro analyses (y=−1.694+
1.969x,R
2
=0.895, p<0.05)
n=329
n=32
n=160
n=160 n=160
n=80
n=30
n=110
n=80 n=110
0
5
10
15
20
25
30
35
Red Squirrel Grey squirrel Woodmouse Bank vole Py
g
my shrew
Molecular
Macro
Fig. 4 Overall results for
molecular and macro analysis of
mammalian prey species in pine
marten diet (n=numberofpine
marten scats tested). Those subject
to macro analysis are a random sub-
sample of the molecular samples
except the sample from the grey
squirrel positive sites. The
woodmouse features significantly
more frequently than any other prey
species (molecular and macro
analysis) and macro analysis
detected significantly more
woodmouse in the diet than
molecular analysis (p<0.01, Fisher
exact). Grey squirrels were more
frequently detected than red
squirrels as prey items (molecular
analysis: p<0.05, Fisher exact)
Eur J Wildl Res
considered as relatively discrete in terms of pine marten
occupancy; however, it is likely that surrounding non-
forested landscape is also used to some extent.
Non-invasive techniques
When compared to other methods of determining pine marten
density, such as snow-tracking (e.g., Zalewski 1999) (not feasible
in Ireland), camera-trapping (e.g., Manzo et al. 2012) and radio-
tracking, there are certain benefits and limitations to genetic
tagging through hair samples. Genetic tagging does not provide
specific biological information such as weight, reproductive
activity and condition, nor does it provide detailed spatial infor-
mation on home range size or territoriality. It is however the only
non-invasive method that confidently differentiates between in-
dividual animals. The data collection process was non intrusive
to the animal and time-efficient in terms of data collection. The
hair traps themselves are inexpensive to construct and maintain.
However it is clear from our data that hair trap density should be
increased to optimise information gathered on the population. In
the midlands sites there were more individuals detected per site
over the study period than there were hair traps available per
Fig. 5 Comparison of results
from both molecular and macro
analysis of pine marten scats,
where all samples were subjected
to both techniques (n=sample
size). Frequency of occurrence
was higher for all species detected
when results were combined, and
significantly so for the
woodmouse (χ
2
=6.04, df =1,
p<0.05)
Fig. 6 The percentage relative
biomass (%BPI) of mammalian
prey items ingested as determined
by macro analysis of pine marten
scats. WM woodmouse, Rrat, BV
bank vole, PS pygmy shrew, RS
red squirrel, GS grey squirrel
Eur J Wildl Res
month; in site 1, six pine marten were detected using five hair
traps and in site 2, ten pine marten were detected with as few as
four hair traps (over the course of the sampling period, hair traps
were persistently stolen from this site). It is possible that this
resulted in an underestimation of the abundance value at this site
as the number of hair traps available was not sufficient to give
each individual present the opportunity to use a hair trap each
month, in particular during the winter months when residency
was being determined.
The proportion of hair samples collected that were success-
fully genotyped (66 %) could also be improved in future
studies. A relationship has been found to exist between the
number of hairs in a sample, and the probability it will be
successfully genotyped (Mowat and Paetkau 2002). In the
majority of hair trapping events in the current study more than
ten hairs were captured, thus providing a relatively high
amount of DNA. However, the samples were left in situ for
a period of up to 1 month, which may have caused the DNA to
degrade due to relatively high ambient temperatures and hu-
midity. Lynch et al. (2006) suggested a survey period of 6 days
is sufficient to detect pine marten presence (in lowland broad-
leaf woods). Screening the quantity of nuclear DNA in the
samples using the sex typing assay prior to genotyping helped
to increase the genotyping success (95 %) as the samples that
were deemed to have insufficient high quality DNA did not
proceed to the genotyping stage. This also helped reduce the
overall cost and this technique combined with shorter sam-
pling periods could substantially help improve the overall
success rate in future studies.
The overall number of alleles and levels of heterozygosity
in this study were very low. Mullins et al. (2010)alsorecorded
low levels of genetic diversity (average H
E
=0.35 and H
O
=
0.34) in the Irish pine marten population, using a larger
microsatellite panel than the current study. However, Mullins
et al. (2010) used samples from a wider geographic range in
Ireland than the current study. The low levels of genetic
variability in both studies could be due to the low number of
individuals that the current pine marten population have re-
established themselves from. Furthermore, there has only
been one mitochondrial DNA haplotype found in the contem-
porary Irish population (Jordan et al. 2012). The long-term
effects of such low levels of genetic diversity in an expanding
population are not known. However, the low diversity found
in this study may also be partially due to the microsatellites
used, as they were originally developed for use with other
mustelids. This was also discussed as a reason for lower levels
of genetic variability in the Iberian pine marten population by
Ruiz-González et al. (2013).
An alternative form of quality screening to the method used
in the current study involves preliminary analysis with a sub-
group of microsatellites, as was undertaken by Ruiz-González
et al. (2013). Samples that amplified well (>50 % positive
PCRs) with the sub-group were then taken to the next stage of
analysis. The approach taken in the current study may be more
useful as data not used for genotyping at least provides further
information on species and sex. The pre selection of DNA
samples for genotyping removes samples that are unlikely to
replicate or may cause a higher occurrence of genotyping error
(Zhan et al. 2010) and thus may be more efficient.
Low genotyping error rates were reported in this study, and
were at the lower end of the level of error when compared to
other non invasive genetic studies using DNA extracted from
hair (Broquet et al. 2007). Genotyping errors are easier to
control and account for in small studies with fewer samples
(Zhan et al. 2010). Mullins et al. (2010), also working with a
relatively small dataset, similarly reported low genotyping
errors. The high number of recaptures reported in the current
study further supports the low occurrence of genotyping er-
rors, and helps to validate the genotyping results. If an erro-
neous individual had been detected within the dataset, this
individual would not affect the overall abundance estimates,
as only animals identified within a site on at least three
separate months were included in the abundance estimates.
Scat density
There are inherent problems with surveying for pine marten
scats in areas of low population density (Birks et al. 2005);
most notably misidentification of scats in the field, even by
experienced surveyors. This problem is addressed in modern
surveys by the use of genetic tests to confirm pine marten
origin (O'Reilly et al. 2008; Balestrieri et al. 2011; Caryl et al.
2012b). The current study found that in the east of Ireland,
where pine marten abundance is lower, a significantly lower
portion of the scats collected were confirmed as being of pine
marten origin than those collected in the midlands, where
marten abundance is higher. A factor that may have contrib-
uted to this result is the likelihood that the surveyor was less
discriminate about which scats were collected in the lower scat
density sites. In areas where scats are more abundant, key
features such as smell and shape are more easily taken into
account, and the surveyor is likely to be more critical regard-
ing the quality of the scat collected for the survey.
Furthermore, in areas of low pine marten population den-
sity, territorial scent marking behaviour may be greatly re-
duced (Macdonald et al. 1998). Lockie (1964) was the first to
suggest that a relationship exists between the number of scats
and pine marten abundance; however, in a review of nine
previous scat surveys in the UK and Spain, Birks et al.
(2005) found that the field relationship between scat abun-
dance on transects and marten numbers was yet to be
established. Indeed, whilst the current study found scat density
to be higher in areas of higher pine marten abundance, regres-
sion analysis failed to define this possible relationship.
Scent detection dogs are increasingly being used in the
study of elusive carnivores (Smith et al. 2003;Longetal.
Eur J Wildl Res
2007; Reed et al. 2011), and have been found to have a
superior detection rate to that of humans. In the current study,
the scent detection dog was used over a 2-day period, and
succeeded in collecting a total of 11 scats, seven of which
were confirmed as pine marten through molecular analysis.
Those that tested negative for pine marten DNA also tested
negative for fox DNA (the species that pine marten scat is
most likely to be misidentified as in Ireland), which suggests
that the quality of the DNA in those samples was too degraded
for genetic species identification. As such, it is not possible to
determine whether the scats detected by the dog that tested
negative for pine marten DNA were true or false negatives.
The lead author only detected one pine marten scat during the
2-day survey without the aid of the dog, suggesting that the
use of scent detection dogs in areas of low pine marten and
low scat density can greatly improve sampling efficiency.
Dietary analysis techniques
In this study, both molecular and macro analyses detected prey
species in similar proportions; therefore, molecular techniques
can be accepted as a reliable method to detect mammals as
prey items in pine marten diet. This is a useful tool in deter-
mining the small mammal composition of carnivore diet and
also the spread (and possible decline) of both invasive and
native mammal species in Ireland. However the macro analy-
sis was significantly more sensitive in the detection of the
woodmouse, which was the most frequently consumed mam-
mal in the diet. It is recommended that any study aiming to
determine exact frequencies of a species in the diet(as distinct
from determining prey species presence or absence) be vali-
dated with traditional hard part (macro) analysis. In this study,
a standard DNA extraction for both the species and dietary
analysis was used (a cost effective strategy). However, to
improve the molecular dietary detection of prey DNA, future
molecular studies might increase the detection rate by sam-
pling a larger amount of scat, extracting multiple samples
from the same scat, or homogenising the scat prior to DNA
extraction (see King et al. 2008). The woodmouse was found
to occur in 31.8 % of scats tested, which is similar to the
frequencies found in Northern Spain and Tuscany (De Marinis
and Masetti 1995). Previous studies in Ireland have found the
woodmouse to occur at around 13 % frequency (Lynch and
McCann 2007) and 14.7 % (O'Meara et al. 2013)inpine
marten scats.
Biomass or %BPI values could be better estimated for both
macro and molecular analyses iffeeding trials were conducted
with captive pine marten to determine the appropriate correc-
tion factors for (a) the detection rates of the various mamma-
lian prey species DNA after known amounts have been con-
sumed and (b) the relationship between weight of dried re-
mains and fresh weight ingested for red and grey squirrels as
distinct from each other and from the 'small mammal'
grouping.
Pine marten predation on squirrels
The absence of grey squirrel in the diet of the pine marten in
the midlands most likely reflects their lack of availability as a
prey item (Carey et al. 2007; unpublished data from Sheehy
and Lawton). In Ireland and Scotland it has been speculated
that the pine marten population has inhibited the grey squirrel
population from spreading, and has even caused the grey
squirrel population to crash in areas where they were once
established (Carey et al. 2007; Caryl 2008; Paterson and
Skipper 2008). No grey squirrel control measures had been
carried out in any of the Irish midlands sites surveyed since the
1990s; therefore, human management of the alien squirrel
population is not an explanatory factor in their rarity. Habitat
is not a factor either, as red squirrel populations are found in
the woodland, as until relatively recently were high numbers
of grey squirrels. Whether predation was a factor in the
retraction of the grey squirrel range historically is not possible
to determine in retrospect, but evidence of predation on the
alien squirrel species in the east confirms that the pine marten
will indeed prey upon the grey squirrel, where it is available.
Molecular and macro analysis produced an overall frequen-
cy of occurrence for grey squirrel of 9.4 % and 10 %, respec-
tively, in sites where grey squirrels are known to be present,
that increased to 15.6 % when results from molecular and
macro analysis were combined. The relative biomass of grey
squirrels gave a similar estimate (13 % BPI). These figures are
based upon a relatively small sample size however, and must
be interpreted with caution as small sample sizes can cause a
prey item to be either under or over represented in dietary
analysis (Trites and Joy 2005). However, they do confirm that
the North American grey squirrel forms part of the European
pine marten diet when the two species' ranges overlap.
Throughout the course of the current study, the grey squirrel
was only confirmed as an available prey item in areas of low
scat density, which made scat collection for dietary analysis in
these areas very challenging.A larger sample size of scats will
allow for a more robust dietary analysis in areas of low
density. In contrast, the sample size of scats collected where
red squirrels were confirmed as present was adequateto detect
with confidence the frequency at which the red squirrel occurs
inthediet.Redsquirrelsandpinemartenhaveco-existedin
Ireland and many other parts of Europe over many millennia,
and the red squirrel has also appeared only as a very low
frequency prey item in previous Irish pine marten dietary
studies (Warner and O'Sullivan 1982; Lynch and McCann
2007; O'Meara et al. 2013). However, red squirrels have been
recorded at higher frequencies in Russia and Sweden where
other small mammal prey are less abundant (De Marinis and
Masetti 1995). It is possible that the red squirrel's lower
Eur J Wildl Res
frequency of occurrence in pine marten diet than that of the
alien grey squirrel is a result of differences between red and
grey squirrel ecology. Red squirrels live at lower densities
(0.3–1.5 per ha) than grey squirrels (2–16 per ha) (Gurnell
1987) and would therefore be numerically less available as
prey items. They are also lighter than the grey squirrel, capa-
ble of reaching the outermost branches of trees, and spend the
vast majority of their foraging time in the canopy, whereas
grey squirrels spend a larger proportion of their foraging time
on the ground (Kenward and Tonkin 1986). This study intro-
duces the possibility that there could be some form of density
dependent effect on the grey squirrel population, where areas
of high predator abundance might be discouraging the nor-
mally invasive grey squirrel from remaining in, or establishing
in a woodland in the first place. The relationship between the
native squirrel predator and the alien squirrel species is yet to be
defined however, and what effect an increase in pine marten
numbers in the east of Ireland will have on grey squirrel
distribution and abundance merits further investigation.
Acknowledgements Emma Sheehy is funded by the Irish Research
Council and Denise B. O’Meara by the Embark Strand 1 Initiative.
Molecular work was supported by a grant from the European Squirrel
Initiative. We thank Louise Wilson and Luna from Conservation Dogs
and all of the woodland owners and managers especially David Hutton
Bury, Donal Whelan and the Irish Forestry Unit Trust, Geoffrey Totten-
ham, Eamon Doran, National Parks and Wildlife Service (Department of
Arts, Heritage and the Gaeltacht) and Michael Carey. Thanks to David
O'Neill for guidance in the laboratory, the volunteers who helped in the
field throughout and two anonymous reviewers for helpful advice on an
earlier version of this manuscript. The authors acknowledge support of
the HEA under PRTLI4 for licensing OSI Digital Imagery through the
Ryan Institute.
References
Balestrieri A, Remonti L, Ruiz-González A, Vergara M, Capelli E,
Gómez-Moliner B, Prigioni C (2011) Food habits of genetically
identified pine marten (Martes martes) expanding in agricultural
lowlands (NW Italy). Acta Theriol 56:199–207
Balharry D (1993) Factors affecting the distribution and population
density of pine martens (Martes martes L.) in Scotland. PhD thesis,
University of Aberdeen, Aberdeen, Scotland
Bijlsma R, Van De Vliet M, Pertoldi C, Van Apeldoorn R, Van de Zande
L (2000) Microsatellite primers from the Eurasian badger, Meles
meles . Mol Ecol 9:2215–2216
Birks J (2002) The pine marten. SP Press, London, The Mammal
Society
Birks J, Messenger J, Braithwaite T, Davison A, Brookes R, Strachan C
(2005) Are scat surveys a reliable method for assessing distribution
and population status of pine martens? Martens and Fishers
(Martes) in Human-Altered Environments:235–252
Broquet T, Ménard N, Petit E (2007) Noninvasive population genetics: a
review of sample source, diet, fragment length and microsatellite
motif effects on amplification success and genotyping error rates.
Conserv Genet 8:249–260
Brownstein MJ, Carpten JD, Smith JR (1996) Modulation of non-tem-
plated nucleotide addition by Taq DNA polymerase: primer
modifications that facilitate genotyping. Biotechniques 20:1004–
1006, 1008–1010
Carey M, Hamilton G, Poole A, Lawton C (2007) Irish squirrel survey.
COFORD, Dublin
Caryl F (2008) Pine marten diet and habitat use within a managed
coniferous forest. PhD dissertation University of Stirling, Scotland
Caryl FM, Quine CP, Park KJ (2012a) Martens in the matrix: the impor-
tance of nonforested habitats for forest carnivores in fragmented
landscapes. J Mammal 93:464–474. doi:10.1644/11-mamm-a-149.1
Caryl FM, Raynor R, Quine CP, Park KJ (2012b) The seasonal diet of
British pine marten determined from genetically identified scats. J
Zool 288:252–259
Clevenger AP (1994) Habitat characteristics of Eurasian pine martens
Martes martes in an insular Mediterranean environment. Ecography
17:257–263
Davis C, Strobeck C (1998) Isolation, variability, and cross-species
amplification of polymorphic microsatellite loci in the family
Mustelidae. Mol Ecol 7:1776–1778
De Marinis AM, Masetti M (1995) Feeding habits of the pine marten
(Martes martes Lineaus 1758) in Europe: a review. Hystrix 7:143–150
Deagle BE, Tollit DJ, Jarman SN, Hindell MA, Trites AW, Gales NJ
(2005) Molecular scatology as a tool to study diet: analysis of prey
DNA in scats from captive Steller sea lions. Mol Ecol 14:1831–1842
Duffy A, Landa A, O'Connell M, Stratton C, Wright J (1998) Four polymor-
phic microsatellites in wolverine, Gulo gulo . Anim Genet 29:63
Dunshea G (2009) DNA-based diet analysis for any predator. PLoS ONE
4:e5252
Fleming MA, Ostrander EA, Cook JA (1999) Microsatellite markers for
American mink (Mustela vison) and ermine (Mustela erminea). Mol
Ecol 8:1352–1355
Goudet J (1995) FSTAT (version 1.2): a computer program to calculate
F-statistics. J Hered 86:485–486
Gurnell J (1987) The natural history of squirrels. Christopher Helm, London
Jedrzejewska B, Jedrzejewski W (1998) Predation in vertebrate commu-
nities: the Bialowieza Primeval Forest as a case study, vol 135.
Ecological Studies, Springer, Berlin
Jordan N, Messenger J, Turner P, Croose E, Birks J, O'Reilly C (2012)
Molecular comparison of historical and contemporary pine marten
(Martes martes ) populations in the British Isles: evidence of differ-
ing origins and fates, and implications for conservation manage-
ment. Conserv Genet 13:1195–1212
Kenward R, Tonkin J (1986) Red and grey squirrels: some behavioural
and biometric differences. J Zool 209:279–304
King R, Read D, Traugott M, Symondson W (2008) Molecular analysis
of predation: a review of best practice for DNA–based approaches.
MolEcol17:947–963
Lanszki J, Zalewski A, Horváth G (2007) Comparison of red fox Vulp es
vulpes and pine marten Martes martes food habits in a deciduous
forest in Hungary. Wildl Biol 13:258–271
Lindenmayer DB, Incoll RD, Cunningham RB, Pope ML, Donnelly CF,
MacGregor CI, Tribolet C, Triggs BE (1999) Comparison of
hairtube types for the detection of mammals. Wildl Res 26:745–
753. doi:10.1071/WR99009
Lockie J (1961) The food of the pine marten Martes martes in west Ross-
shire, Scotland. J Zool 136:187–195
Lockie JD (1964)Distribution and fluctuations of the pinemarten, Martes
martes (L.),inScotland.JAnimEcol33:349–356
Long RA, Donovan TM, Mackay P, Zielinski WJ, Buzas JS (2007)
Comparing scat detection dogs, cameras, and hair snares for survey-
ing carnivores. J Wildl Manage 71:2018–2025
Lynch Á, Brown M, Rochford J (2006) Fur snagging as a method of
evaluating the presence and abundance of a small carnivore, the pine
marten (Martes martes). J Zool 270:330–339
Lynch Á, McCann Y (2007) The diet of the pine marten (Martes martes )
in Killarney National Park. Biology & Environment: Proceedings of
the Royal Irish Academy 107:67–76
Eur J Wildl Res
Macdonald DW, Mace G, Rushton S (1998) Proposals for future moni-
toring of British mammals. Department of the Environment,
Transport and the Regions, London, UK
Manzo E, Bartolommei P, Rowcliffe J, Cozzolino R (2012) Estimation of
population density of European pine marten in central Italy using
camera trapping. Acta Theriol 57:165–172
Mergey M, Helder R, Roeder J-J (2011) Effect of forest fragmentation on
space-use patterns in the European pine marten (Martes martes ). J
Mammal 92:328–335
Messenger JE, Birks JDS (2000) Monitoring the very rare pine marten
populations in England and Wales. Management and conservation
aspects of small carnivores: human interactions. Backhuys, Leiden,
the Netherlands
MetEireann (2012) Temperature in Ireland. http://www.met.ie/climate/
temperature.asp?prn=1.Accessed23May2013
Mills L, Citta J, Lair K, Schwartz M, Tallmon D (2000) Estimating animal
abundance using noninvasive DNA sampling: promise and pitfalls.
Ecol Appl 10:283–294
Mowat G, Paetkau D (2002) Estimating marten Martes americana pop-
ulation size using hair capture and genetic tagging. Wildl Biol 8:
201–109
Mullins J, Statham M, Roche T, Turner P, O'Reilly C (2010) Remotely
plucked hair genotyping: a reliable and non-invasive method for
censusingpinemarten(Martes martes, L. 1758) populations. Eur J
Wildl Res 56:443–453
O'Mahony D, O'Reilly C, Turner P (2006) National pine marten survey of
Ireland (2005) Coford Connects. Environ Sci Technol 7:1–8
O'Mahony D, O'Reilly C, Turner P (2012) Pine marten (Martes martes)
distribution and abundance in Ireland: a cross-jurisdictional analysis
using non-invasive genetic survey techniques. Mamm Biol 77:351–
357
O'Meara D, Sheehy E, Turner P, O'Mahony D, Harrington A, Denman H,
Lawton C, MacPherson J, O'Reilly C (2013) Non-invasive multi-
species monitoring: real-time PCR detection of small mammal and
squirrel prey DNA in pine marten (Martes martes) scats. Acta
Theriol:1–7. doi:10.1007/s13364-013-0155-8
O'Meara D, Turner P, Coffey L, O'Reilly C (2012) TaqMan assays for
species identification of the red squirrel (Sciurus vulgaris) and the
grey squirrel (Sciurus carolinensis). Conserv Genet Resour 4:603–
604
O'Reilly C, Statham M, Mullins J, Turner P, O'Mahony D (2008) Efficient
species identification of pine marten (Martes martes) and red fox
(Vulpes vulpes) scats using a 5′nuclease real-time PCR assay.
Conserv Genet 9:735–738
O'Sullivan PJ (1983) The distribution of the pine marten (Martes martes)
in the Republic of Ireland. Mammal Rev 13:39–44
Palomares F, Godoy JA, Piriz A, O'Brien SJ (2002) Faecal genetic
analysis to determine the presence and distribution of elusive carni-
vores: design and feasibility for the Iberian lynx. Mol Ecol 11:2171–
2182. doi:10.1046/j.1365-294X.2002.01608.x
Paterson WD, Skipper G (2008) The diet of pine martens (Martes
martes) with reference to squirrel predation in Loch lomond and
the Trossachs National Park. University of Glasgow, Glasgow,
Scotland. BSc dissertation
Peakall ROD, Smouse PE (2006) genalex 6: genetic analysis in Excel.
Population genetic software for teaching and research. Mol Ecol
Notes 6:288–295
Pereboom V, Mergey M, Villerette N, Helder R, Jean-Francois G, Lod T
(2008) Movement patterns, habitat selection, and corridor use of a
typical woodland-dweller species, the European pine marten
(Martes martes ), in fragmented landscape. Can J Zool 86:
983–991
Reed SE, Bidlack AL, Hurt A, Getz WM (2011) Detection distance and
environmental factors in conservation detection dog surveys. J Wildl
Manage 75:243–251
Rousset F (2009) genepop'007: a complete re-implementation of the
genepop software for Windows and Linux. Mol Ecol Resour 8:
103–106
Ruiz-González A, Madeira MJ, Randi E, Urra F, Gómez-Moliner BJ
(2013) Non-invasive genetic sampling of sympatric marten species
(Martes martes and Martes foina): assessing species and individual
identification success rates on faecal DNA genotyping. Eur J Wildl
Res 59:371–386
Scotts D, Craig S (1988) Improved hair-sampling tube for the detection of
rare mammals. Wildl Res 15:469–472
ShehzadW,RiazT,NawazMA,MiquelC,PoillotC,ShahSA,
Pompanon F, Coissac E, Taberlet P (2012) Carnivore diet analysis
based on next-generation sequencing: application to the leop-
ard cat (Prionailurus bengalensis) in Pakistan. Mol Ecol 21:
1951–1965
Smith DA, Ralls K, Hurt A, Adams B, Parker M, Davenport B, Smith
MC, Maldonado JE (2003) Detection and accuracy rates of dogs
trained to find scats of San Joaquin kit foxes (Vulpes macrotis
mutica ). Anim Conserv 6:339–346
Teerink BJ (1991) Hair of west-European mammals, 1st edn. Cambridge
University Press, Cambridge
Trites AW, Joy R (2005) Dietary analysis from fecal samples: how many
scats are enough? J Mammal 86:704–712
Valière N (2002) GIMLET: a computer program for analysing genetic
individual identification data. Mol Ecol Notes 2:377–379
Van Oosterhout C, Hutchinson WF, Wills DPM, Shipley P (2004) micro-
checker: software for identifying and correcting genotyping errors in
microsatellite data. Mol Ecol Notes 4:535–538
Vincent I, Farid A, Otieno C (2003) Variability of thirteen microsatellite
markers in American mink (Mustela vison). Can J Anim Sci 83:
597–599
Warner P, O'Sullivan P (1982) The food of the pine marten Martes martes
in Co. Clare Transactions of the International Union of Game
Biology 14:323–330
Waters C, Lawton C (2011) Red squirrel translocation in Ireland. Irish
Wildlife Manuals, vol 51. National Parks and Wildlife Service,
Department of the Environment, Heritage and Local Government,
Dublin, Ireland
Williams BW, Etter DR, Linden DW, Millenbah KF, Winterstein SR,
Scribner KT (2009) Noninvasive hair sampling and genetic
tagging of co-distributed fishers and American martens. J
Wildl Manage 73:26–34
Yalden DW, Morris PA (1990) Analysis of owl pellets. The Mammal
Society, London
Zalewski A (1999) Identifying sex and individuals of pine marten using
snow track measurements. Wildl Soc Bull 27:28–31
Zalewski A, Jedrzejewski W (2006) Spatial organisation and dynamics of
the pine marten Martes martes population in Bialowieza Forest (E
Poland) compared with other European woodlands. Ecography 29:
31–43
Zhan X, Zheng X, Bruford MW, Wei F, Tao Y (2010) A new method for
quantifying genotyping errors for noninvasive genetic studies.
Conserv Genet 11:1567–1571
Eur J Wildl Res