American black bear population size and genetic diversity at
Apostle Islands National Lakeshore
Jerrold L. Belant14, Julie F. Van Stappen2'5, and David Paetkau3'6
1NationalPark Service, PicturedRocks Science Center,Box 40, Munising,MI49862, USA
2NationalPark Service, Apostle Islands NationalLakeshore,Route 1, Box 4, Bayfield,WI54814, USA
610 RailwayStreet, Box 274, Nelson, BC V1L5P9, Canada
3Wildlife Genetics International,
Abstract: Effective management of American black bears (Ursus americanus) requires an understandingof populationdemographics.In 2002, we obtainedDNA fromhaircollected at barbed-wire
trapsto estimateblack bear populationsize and study populationgenetics on Stockton(4,069 ha) and
Sand (1,193 ha) islands at Apostle Islands NationalLakeshore(AINL), Wisconsin. Hair samples also
were collected from 2 nuisancebearson Oak Island.We analyzed372 hair samplesfrom Stocktonand
Sandislandscollectedon 4 occasionsat about14-dayintervals.Geneticanalysisof 6 microsatelliteDNA
markersresultedin 71 capturesof 26 individualson StocktonIslandand 13 capturesof 6 individualson
SandIsland.The estimatedbearpopulationson Stocktonand Sandislandswere 26 (SE - 0.54, 95% CI
[confidenceinterval]=26-26) and6 (SE=0.60, 95%CI=6-7) individuals,respectively.The estimated
density on StocktonIsland was 0.64 bears/km2and on Sand Islandwas 0.50 bears/km2.Genetic variation withinboth islandpopulationswas higher(meanHE> 0.77) thancould be maintainedby populations of this size in isolation,suggestingsubstantialimmigrationfromthe mainlandpopulationoccurred.
Genetic assignment testing using log genotype likelihoods demonstratedsufficient variation between bearpopulationson StocktonandSandislandsto permitidentificationof natalorigins.The 2 bears
from Oak Island were genetically intermediatebetween Stockton and Sand islands. Islands within
AINL contain small black bearpopulationsof high density that are geneticallydistinctand apparently
influencedby immigrationfrom the mainlandpopulation.
Key words: Americanblack bear,Apostle IslandsNationalLakeshore,genetic diversity,mark-recapture,population
estimates, Ursus americanus, Wisconsin
Ursus 16(1):85-92 (2005)
Knowledge of animalpopulationsizes and connectivity, and how they vary temporally, is necessary for
effective species management.The NationalParkService
(NPS) is mandatedto "... maintainas partsof the natural
ecosystems of parks all native plants and animals"
(NationalParkService2000). The NPS is also requiredto
monitor important resources that are necessary to
accomplish managementobjectives as requiredby law
or planning documents (National Park Service 2000).
Black bears (Ursus americanus) are an important
component of many units in the National Park system,
including Wisconsin's Apostle Islands National Lakeshore (AINL). At AINL, understanding black bear
abundanceand ecology is criticalbecause:(1) they have
[email protected]
[email protected]
[email protected]
85
an inherentrole in maintainingecological integrity,(2)
harvest of black bears is legal in AINL, and (3) of
potentialinteractionswith tourists.
Techniquesto estimatebearpopulationsthatcould be
used for long-termmonitoringareimpracticalin heavilyforested areas like AINL or are cost prohibitive.However, use of genetic markers to identify individual
animals has increased considerably in recent years
(Parkeret al. 1998). Small amountsof tissue, including
hair,arenow used frequentlyin genetic studiesto identify
individuals, as well as determine sex, species, and
genealogy (Taberletet al. 1993; Paetkauand Strobeck
1994, 1998; Foranet al. 1997; Haig 1998). Advances in
genotyping using non-invasive collection of tissue
samples (Woods et al. 1996, 1999) and mark-recapture
modeling(Whiteet al. 1982) have providedan additional
means to estimate animal populations, including estimatesof precision.Mark-recapturemodelingemploying
86
BLACKBEARDENSITY
ANDGENETICS
ATAPOSTLE
? Belant et al.
ISLANDS
Study area
The Apostle Islands Archipelagois
located in southwest Lake Superior
\);,/
(46?42'-47?05'N; 90?24'-91?03'W),
N
Wisconsin, USA (Fig. 1). Twentyone of the 22 islands are within
AINL. Islands range from 1.2-4,069
S
^? ~ ha. Mean maximum and minimum
daily temperatures are 25.5?C and
12?C in July and -4.5?C and -14?C
in January,respectively. Annual precipitation is about 75 cm, including
200 cm of snow. Before European
settlement,about 90% of the Apostle
Islandscontainedmixed conifer-hardwood forests (Frederickand Rakestraw 1976).
X--J ~ Black bear occurrence has been
documentedon 8 islands within AINL
(Belant and Van Stappen 2002). AlFig. 1. Location of Sand, Oak, and Stockton islands anid the Apostle thoughtiminganddurationof presence
Island Archipelago, Wisconsin, USA.
is generally unknown for the islands,
black bears apparentlymost recently
colonized StocktonIsland in the early
data from systematicallyobtained hair samples is now
1970s or later,andthe islandhad an estimatedpopulation
of 3 bears in 1984 (Fleming 1997; R.K. Anderson and
likely the most common method to estimate bear
et
al.
Poole
and
Strobeck
size
D. Trauba,1989, Black bear density on StocktonIsland
2000,
(Mowat
population
al.
et
Boersen
and mainland Wisconsin, University of Wisconsin,
2003).
2001,
As for connectivity, aquaticand other geographicor
Stevens Point, Wisconsin, USA). Bear huntinghas been
humancreatedbarrierssuch as roads can greatlyinhibit
allowed withinAINL except for 1983-94, duringa black
animal movement and dispersal (Schmiegelow and
bear research project when hunting was closed on
Nudds 1987, Proctor et al. 2002), increasing the risk
Stockton Island (Fleming 1997). Only 1 bear has been
of population extinction (Lande 1994). The NPS is
recordedas harvestedsince inception of AINL in 1970
of
full
the
"...
to
types
genetic
range
protect
obligated
(AINL, Bayfield, Wisconsin, USA, unpublisheddata).
(genotypes) of native plant and animal populations in
Numberof bear-humaninteractionswas roughlystable
the parksby perpetuatingnaturalevolutionaryprocesses
from 1990 through2002, withmost interactionsoccurring
and minimizing human interference with evolving
in mid-August. Until 2002, bear-humaninteractionsat
genetic diversity" (National Park Service 2000). UnAINLresultedin 1 beartranslocationevery 2-4 years.As
derstandingthe demographicand genetic connectivity
a result of increased bear-human interactions, more
of potentially isolated bear populations at AINL will
aggressive managementwas needed in 2002-03, resulthelp ensure appropriatemanagement of metapopulaing in the removalof 11 bears.
tions.
We conductedthis studyon StocktonandSandislands.
Our initial study goal was to estimate the population
islands were selected based on previous studies
These
size of bears on Stockton and Sand islands. Specific
(Trauba 1996, Fleming 1997) and observations that
objectiveswere to: (1) estimatebeardensity on Stockton
suggested viable populations were present. Stockton
and Sand islands and comparebear density on Stockton
Island(4,069 ha) is the largestisland withinAINL and is
Islandto previousestimates,(2) estimatethe sex ratioof
7.7 km from the mainland;elevations range from 183these island populations and compare it to previous
242 m. Logging began during the late 1800s and
estimates, and (3) conduct preliminaryassessment of
continuedthrough 1955, with extensive fires occurring
genetic variation within and between these island
duringthe 1950s (Rakestraw1976). Currently,deciduous
populations.
Ursus 16(1):85-92 (2005)
ANDGENETICS
BLACKBEARDENSITY
? Belant et al. 87
ATAPOSTLE
ISLANDS
forestscover 87%of this islandand aredominatedby red
(Acer rubrum)and sugar (A. saccharum) maple, with
trembling aspen (Populus tremuloides), white birch
(Betulapapyrifera),andredoak (Quercusrubra)(Trauba
1996). Forestedwetlandscomprise6% of the island with
overstory dominatedby black spruce (Picea mariana)
and white cedar (Thuja occidentalis). Ponds resulting
from beaver (Castor canadensis) activities (flowages)
comprised3% of the island area. Forests dominatedby
red (Pinus resinosa) andwhite (P. strobus)pine occurred
on 2% of the island.
Sand Island (1,193 ha) is located 2 km from the
mainland;elevationsrangefrom 183-203 m. SandIsland
was repeatedly logged, with 7 large timber harvests
occurring from the late 1800s until 1975. Presently,
mixed hardwood-coniferforests including yellow birch
(Betula alleghaniensis), balsam fir (Abies balsamea),
white birch,white cedar,and red maplecover 83%of the
island. Canada yew (Taxus canadensis) and mountain
maple (Acer spicatum) are prevalentin the understory.
Remaining habitat includes about 7% black sprucetamarack(Larixlaricina)swamp,6%semi-openbog, and
3% white pine-hemlock (Tsuga canadensis) forest.
Methods
Hair trapping
We used a systematicgrid (1.6 x 1.6 km) overlaidon
each island to distributetrappingeffort. Grid cells that
contained<50% land were includedwith adjacentcells.
We constructedhair traps as describedby Woods et al.
(1999), using 4-barbed, 2-strand wire to create an
enclosurearound>3 trees with wire about 50 cm above
ground.We placed hairtrapsin areaswe believed would
maximize black bear captures. Enclosures were baited
with 0.5 L fish oil pouredon a decayinglog positionedin
the center of the enclosure. Anise oil was applied to
vegetation about 2 m above ground at or near the
perimeter of the enclosure. Traps were rebaited and
reluredduringeach session. We established15 hairtraps
on Stockton (1 trap/2.7 km2) and 5 hair traps on Sand
Island(1 trap/2.4km2)andmonitoredthemfor 4 sessions
between 26 June and 13 August 2002. We attemptedto
remove hair from trapsat 2-week intervals.Duringeach
trap check, all hair from each barb was placed in
a separatepaper envelope, labeled, and air dried before
processing.Because of the comparativelyhigh densityof
hair traps(e.g., Mowat and Strobeck2000, Poole et al.
2001, Boersen et al. 2003), we did not move the traps
between sessions.
Ursus 16(1):85-92 (2005)
Home rangesof adultfemale and male black bearson
Stockton Island were 7.2- and 32.9 km2, respectively
(Trauba1996). Witha hair-trapdensityof 1 trap/2.7km2,
about3 hairtrapswere availableto each adultfemale and
12 trapswere availableto each adultmale. Home range
data were unavailablefor Sand Island.Otis et al. (1978)
recommended>4 traps/individualhome rangefor markrecapturestudies.
Population parameters
The studyareaswereislandsseparatedfromthe nearest
land by >2 km. Although bears have been observed
swimming between islands and between islands and the
mainland (Trauba 1996; AINL, unpublisheddata), we
assumed immigrationand emigrationwere low. During
a 12-year (1984-95) radiotelemetrystudyof black bears
on Stockton Island, only 4 occurrences of natural
emigrationand 1 instanceof immigrationwere observed
(Trauba1996, Fleming 1997). All of these movements
were temporary;only 2 occurredduring summerwhen
our study was conducted.Finally, mean annualsubadult
and adult mortality on Stockton Island was <10%
(Fleming 1997). Because of these factors and the short
durationof our study,we consideredour studyareasto be
geographicallyand demographicallyclosed. Therefore,
we used closed mark-recapturemodels availablein program CAPTURE (Otis et al. 1978, White et al. 1982,
Rexstadand Bumham 1991) to estimatepopulationsize
for each island. Model selection was based on statistical
tests conductedin CAPTUREandourknowledgeof bear
biology. We furtherverifiedpopulationclosureusing the
programCloseTest (Stanley and Richards2003), which
uses the Stanley and Bumhamclosure test (Stanley and
Bumham 1999), a chi-squarestatistic developed under
a null model allowing for time-specific variation in
captureprobabilitiesunder closure. This closure test is
most sensitive to permanentemigration,intermediately
sensitive to permanentor temporaryimmigration,and
least sensitive to temporary emigration (Stanley and
Burnham 1999). We also estimated if the sex ratio of
identifiedbearsdifferedfrom 1:1 using a chi-squaretest
for equal proportions.All tests were considered significant at alpha= 0.05.
DNA analyses
In addition to hair samples collected from traps, we
conducted DNA analyses on hair samples collected
from 2 black bears capturedin nuisance situations on
Oak Island (Fig. 1). For all samples we used 10 guard
hairs for extraction when possible to reduce the
probabilityof genotyping errors(Gossens et al. 1998).
88
BLACK BEAR DENSITYAND GENETICSAT APOSTLEISLANDS* Belant et al.
Table 1. American black bear hair capture results from Stockton (n = 15 trap sites) and Sand (n = 5 trap sites)
islands, Apostle Islands National Lakeshore, Wisconsin, USA, 2002.
Island
Stockton
Session
1
2
3
4
Grandmean
Total
Sand
Grandmean
Total
Trapduration
(days)
x
SD
14.2
13.6
13.6
14.6
14.0
2.4
1.3
1.7
1.4
0.5
withhair
samples
14
15
13
14
Hair
samples/trap
x
SD
7.8
7.6
4.3
3.2
5.7
2.9
4.1
2.5
1.7
2.3
56
1
2
3
4
12.4
15
14
18
14.9
0.5
0
0
0
2.4
3
4
3
4
3.3
3.8
1.7
4.5
3.3
1.5
2.8
0.6
3.9
1.2
12
We used 6 microsatelliteloci for analyses of individual
identity:G1OC,G1OM,G1OX,G1D, G10H, and MU59
(Paetkau et al. 1998a and references therein). Ten
additionalloci (CXX20, CXX 10, G1A, G1OB, G1OJ,
G1OL,GO1P,G1OU,MU50, and gender)from a sample
of each identified bear were analyzed to improve
population genetics analyses. Gender was determined
using size polymorphismin the amelogeningene (Ennis
and Gallagher 1994). We conducted DNA extractions
using QIAGEN DNeasy Tissue kits (Qiagen, Mississiauga, Ontario,Canada),following the manufacturer's
instructions.
We examined distributionof genotype similarity to
estimatethe probabilityof 2 or more sampledindividuals
havingidenticalgenotypesat the 6 loci thatwe examined.
The observednumbersof pairsof similargenotypeswere
used to estimatethe expectednumberof pairsof identical
genotypes (0 mismatchingmarker[MM] pairs;Paetkau
2003). The typicalpatternreflectsan orderof magnitude
decline with each successive decrease in number of
mismatchingmarkers.Thus, one would expect a single
errorfor every 10 1MM pairs (Paetkau2003). We used
the genetic assignmenttest (Paetkauet al. 1995, Waser
and Strobeck 1998), which calculatesthe probabilityof
observing specific individual genotypes in a series of
referencepopulations,based on observedallele frequencies in those populations.Log-likelihoodestimatesof the
cumulativeprobabilityof occurrenceof individualalleles
were used to determinegenetic similarity.We describe
genetic diversityusing the observednumberof alleles (A)
and expected heterozygosity (HE) because they are
metrics commonly used to describe genetic variability
No. hair
samples
No. bears
No. new
bears
109
114
56
45
17
22
17
15
17
7
2
0
324
71
26
10
15
5
18
4
4
2
3
4
2
0
0
48
13
6
(Paetkauand Strobeck 1998). We presentprobabilityof
identity (PID)values to demonstratethe power markers
have to identifyindividuals(PaetkauandStrobeck1998).
Results
Hair trapping
We collected 635 hair samples from both islands
during the study, 372 of which we considered had
adequateamountsof DNA for analyses.We obtained324
usablehairsamplesfrom56 trapvisits on StocktonIsland
and 48 usable samplesfrom 14 trapvisits on SandIsland
(Table 1). The number of traps with hair samples
remainedrelatively constant throughoutthe study. The
mean number of hair samples/trap declined during
successive sessions on Stockton Island. In contrast,
although samples sizes were low, the mean numberof
hair samples/trapwas greatest on Sand Island during
session 4. Mean intervalbetween trapsessions typically
was within 1 day of desired durationexcept between
sessions 3 and4 when inclementweatherdelayed access
to hairtrapson Sand Island.
DNA analyses
Of 372 samplesprocessed,256 (69%)were suitablefor
assignmentto individualblack bears. Of the remaining
samples, 22% were extracted but we were unable to
generatea reasonablegenotype,7%were likely a mixture
of DNA from2 bears,and2%lackedsuitablematerialfor
extraction. We found only 2 pairs of genotypes that
matchedat 4 of 6 markers(2MM pairs) and 1 pair that
matchedat all but 1 marker(1MM pair).This suggests it
Ursus 16(1):85-92 (2005)
ANDGENETICS
BLACKBEARDENSITY
ATAPOSTLE
ISLANDS
* Belant et al. 89
is highly unlikely that we sampledeven a single pair of
individualswith identical6-locus genotypes.
We identified32 individualbearsfromhairsamples;6
bears(1 male, 5 females) on SandIslandand26 bears(14
males, 12 females) on StocktonIsland.These individuals
were captured84 times; 71 captureson StocktonIsland
and 13 captures on Sand Island. For both islands
combined, we captured9 females during 2 sessions, 6
during 3 sessions, and 2 during4 sessions. Combining
males resultedin 4 capturedduring 1 session, 1 during2
sessions, 6 during3 sessions, and4 during4 sessions. The
numbers of individuals captured during sessions 1-4
were 21, 26, 19, and 18, respectively.No new individuals
were capturedon StocktonIsland aftersession 3 and on
Sand Islandaftersession 2.
Overall,25 of 32 (78%)bearswere capturedat > 1 hair
trap.Femaleson SandIslandwere capturedon averageat
2.5 (SD =1) traps; the male was capturedat 1 trap.
Females on StocktonIslandwere capturedon averageat
2.9 (SD- 1.7) traps;males were capturedat 4.6 (SD =
2.6) traps.Duringindividualtrapchecks from which we
obtainedgeneticallyidentifiablehairfrom > 1 individual
(n = 57), most (72%)containedhairfrom 1-3 bears;25%
contained hair from 4-6 bears, and 2% contained hair
from7 bears.No bearswere capturedat hairtrapson both
islands and no bears were removed from the islands
duringthe study for managementpurposes.
Population estimates
We conductedpopulationestimates for Stockton and
Sand Island from 8 closed mark-recapturemodels used
in CAPTURE. For Stockton Island, we used the null
estimator (Mo) because we did not detect variation in
capture due to heterogeneityamong individuals (X2
0.15, 1 df, P = 0.70), behavior response after initial
capture(X2= 0.09, 1 df, P = 0.76), or among trapping
sessions (X2= 6.31, 3 df, P = 0.10) using chi-square
goodness-of-fit tests provided in CAPTURE.Additionally, the Stanley and Burham closure test suggestedthe
populationwas closed (X2= 1.10, 2 df, P = 0.58). The
estimatedpopulationfor Stockton Island was 26 (SE
0.54) bears with 95% confidence interval(CI) range of
26-26. This represents a density of 0.64 bears/km2.
Populationestimatesfor males (n = 14) and females (n =
12) also were identical to the known number of
individuals obtained from DNA analyses. The mean
captureprobabilityamong sessions was 0.68. The M:F
sex ratio of 1.16:1 was similar to 1:1
(X2 0.15, 1 df,
P = 0.70).
For Sand Island, we also used the null estimator
(Mo)
because of lack of variationin capturedue to behavior
Ursus 16(1):85-92 (2005)
Table 2. Variability of genetic markers used to
identify individual American black bears, Stockton
Island, Apostle Islands National Lakeshore, Wisconsin, USA, 2002. HE= expected heterozygosity, Ho =
observed heterozygosity, A = mean observed number of alleles, and PID= probability of identity.
Marker
HE
MU59
0.84
G1D
0.83
G1OX
0.80
G10H
0.77
G10M
0.75
G10C
0.63
Mean
0.77
Overallprobability
of identity
Ho
A
PID
0.92
0.89
0.85
0.89
0.81
0.69
0.84
8
8
9
9
7
7
8
0.05
0.05
0.06
0.08
0.10
0.21
2E-07
response afterinitial capture(X2= 2.28, 1 df, P = 0.13)
and among trappingsessions (X2= 2.06, 3 df, P = 0.56).
Sample size was inadequateto test for heterogeneityof
trappingprobabilitiesamong individualsin the population. Additionally,the Stanley and Burham closure test
suggestedthe populationwas closed (X2= 0.99, 2 df, P =
0.61). The estimatedpopulationfor Sand Island was 6
(SE = 0.60) bears with 95% CI range of 6-7. This
representsa density of 0.50 bears/km2.Mean capture
probabilityamong sessions was 0.54. We also conducted
an analysisusing Chao's heterogeneitymodel (Mh)in the
event heterogeneity of trappingprobabilitiesoccurred
and we were unableto detectit. Model Mhperformswell
under strong trappingheterogeneity(Mowat and Strobeck 2000). Using Mh,we deriveda similarestimateof 6
(SE = 0.54) bears(95% CI 6-9).
Genetic variation among islands
Considerablegenetic variation was observed within
and among islands at the 6 markersthat were used to
identifyindividuals.The meanHe was 0.84 for Stockton
Island with the probability of 2 randomly drawn,
unrelatedindividuals with identical genotypes of 2 x
10-7 (Table 2). Although sample size was limited for
Sand Island,mean Ho for markersused was 0.94, which
suggests black bears on Sand Island possess higher
genetic variabilitythan black bears on Stockton Island.
Using all 15 markers,25 of 30 single-locusgenotypesfor
the 2 Oak Islandbearswere heterozygous
(He = 0.83).
Geneticassignmenttesting demonstratedconsiderable
variation between Sand and Stockton Island bear
populations (Fig. 2). A notable exception was a male
capturedon Stocktonthathad a probabilityof occurrence
9 ordersof magnitudelower (10-35) thanthatof the next
lowest probability observed for an individual
(10-26)
90
* Belant et al.
BLACKBEARDENSITY
ANDGENETICS
ATAPOSTLE
ISLANDS
A
-16'I._
--210
CO
A
A
AA
A
A
,
A
0
K
XO
0)1
y
0
.1
A
-36
-31
-36
-26
-21
-16
Log likelihood Sand Island
0 Sand X Oak A Stockton -
Fig. 2. Assignment testing based on log genotype
likelihood probabilities for black bears on Stockton,
Sand, and Oak islands, Apostle Islands National
Lakeshore, Wisconsin, USA, 2002.
sampledon StocktonIsland.The 2 individualsfrom Oak
Islandwere intermediateto black bearson Stocktonand
Sand islands.
Discussion
We report the first bear population estimates in
Wisconsin with data from nonharvestedanimals that
includedestimatesof precision.The density of bears on
Stockton Island (0.64 bears/km2)remains the highest
reported in Wisconsin (Kohn 1982, Storlid 1995,
Fleming 1997) and among the highest reportedin North
America (Garshelis 1994). The density of bears on
StocktonIsland was estimatedpreviously at 0.07 bears/
km2 in 1984 and increased to 0.76 bears/km2in 1994
before declining to 0.61 bears/km2in 1996 (Fleming
1997). Methods between studies are not directly
comparable;however, previous and current estimates
suggest the population has remained approximately
stable during 1990-2002. Similarly, the 1:1 sex ratio
we observed for StocktonIsland is comparableto those
reportedduringearlierstudies (Trauba1996). Although
the
population trends are unknown for Sand Island,
and
sex
1:1
ratio,
stable
size,
population
apparently
approximatelystable age distributionduring the early
1990s (Fleming 1997) on Stockton Island suggests that
the black bear populationhas reachedthe demographic
effective populationsize (Caughleyand Sinclair 1994).
We obtained a limited amount of data to develop
a populationestimatefor Sand Island (n = 13 samples).
However, because of geographic and demographic
closure,high captureprobability,and high meannumber
of trapsites individualswere capturedat, we believe the
population estimate for Sand Island is reasonable.
Although the density on Sand Island was less than
StocktonIsland,the density of females (0.42 bears/km2)
was greaterthan the female density on Stockton Island
(0.29 bears/km2).Greaterdensityof femalebearson Sand
Islandcould be becauseof greaterfood abundance,closer
proximityto the mainlandpopulation,or a combination
of these and otherfactors.
There are at least 2 possible explanationsfor the low
numberof males (n= 1) detectedon SandIsland.The first
is thatwe wereunableto capturehairfromothermales on
the island. We believe this is unlikely because of our
overall high mean capture probability. It is possible,
however, that additionalmales immigratedto the island
temporarily,particularlyduringthe breedingseason, and
these nonresident individuals may not have been
detected. It could also be that only one male bear was
present.Because of its small size (<12 km2), resources
(e.g., food, territory)on SandIslandmay be too limitedto
supportadditionalmale bears. This male bearmay have
used Sand Island as only partof its home range, as has
been reportedin other island studies (Garshelis 1994),
includingStocktonIsland(Trauba1996, Fleming 1997).
In a previous study of mammalian biogeography at
AINL, we demonstrateda significantpositive relationship betweenislandsize and body mass of mammalsthat
could be supported(Belantand Van Stappen2002).
Although the populations on Stockton and Sand
islands were small, the high genetic variabilityof bears
in this study was similar to other mainland bear
populations (Paetkau et al. 1998b). These population
characteristicsare incompatible unless a high rate of
immigrationhas occurred.Forexample,when population
estimatesareavailablefor 2 populations,andan estimate
of heterozygosityis availablefor 1 of those populations,
the heterozygosity for the second population can be
estimatedassumingthat both populationsare at equilibrium for immigration, mutation, and genetic drift
(Paetkauet al. 1998b).Using datafromthe isolatedblack
bearpopulationon the island of Newfoundland(approximateN = 6,400, HE = 0.41) and the populationestimate
for StocktonIsland,we expect heterozygosityto be 0.3%
in the latterpopulation,assumingthatthis populationwas
isolatedandatequilibrium.An indistinguishableestimate
Ursus 16(1):85-92 (2005)
BEARDENSITY
ANDGENETICS
BLACK
ATAPOSTLE
ISLANDS
* Belant et al. 91
is obtained when data from Kodiak brown bears (U.
arctos) are used (Paetkauet al. 1998b). We know from
historic data that the Stockton Island population was
founded too recently to be at equilibrium.Additionally,
the high levels of genetic variationare incompatiblewith
any significantperiod of isolation.
Considering the inescapably low sample sizes, the
genetic data were remarkablysuccessful in clustering
bearsbasedon islandof origin.The exceptionalStockton
Island male was clearly not bor on Stockton Island,
providingdirectevidence of populationconnectivityand
one of the first examples of using genetic assignmentof
natal origins to directly study animal movements from
a single capturelocation.This individualmay have been
bor on the mainlandand immigratedto StocktonIsland.
The placementof the 2 individualsfrom Oak Island
intermediateto Sand and Stockton Island populations
corresponds with their respective distances from the
mainlandand furtherdemonstratesthat a high level of
genetic differentiationacross islands exists. A general
trend in genetic similarityas a function of distance has
also been demonstratedin mainlandblack bear populations (Miller et al. 1998). We conclude that AINL
contains bear populations small enough to cause rapid
genetic drift but possesses a level of intra-population
variabilitythat could only be maintainedby gene flow
from the mainlandpopulation. Also, based on genetic
variability observed and apparent immigration, we
suggest that inbreeding depression is unlikely to be
a significantthreatto these populations.
Future studies quantifyinglevels of genetic differentiation among bear populationson additionalislands at
AINL and mainland Wisconsin are warranted.Such
research could aid in determining dispersal patterns
among islands that resulted in the relatively high
variabilitywe observed. Also, because of the range of
distances among islands and between islands and the
mainland,we could gain insightinto sex-specificratesof
movement. Paetkau et al. (1998a) demonstratedfor
brown bears that oceanic water crossings of 3-4 km
eliminated female-mediatedgene flow, whereas male
dispersalcontinued at this level but was eliminatedby
8-km watercrossings.
Obtainingpopulationestimates from mark-recapture
studiesusingDNA fromhairsamplesis a methodsuitable
for long-term monitoring at AINL. Other techniques,
including mark-resight estimates (e.g., Miller et al.
1997), arenot feasible becauseof theircomparativehigh
cost and forest cover which limits visibility. Population
estimators involving captured animals, however, can
provide additionalinformation,such as age, reproducUrsus 16(1):85-92 (2005)
tion, and survival, which are also critical demographic
metrics that cannot be obtained using techniques we
employed. High mean individualbear captureprobabilities and the low 95% CI we observed suggests that the
numberof hairtrapsat StocktonIslandcould be reduced.
Simulationmodeling with existing data from Stockton
Island could refine the experimentalapproachfor future
studies. We do not recommendreducingthe density of
hairtrapson smallerislands such as Sand Islandbecause
of the low numberof trapsthatwould be deployedandthe
limited numberof bears likely present.We recommend
periodic monitoring of bear populations on these and
other islands at AINL relative to bear nuisance activity,
harvestmanagement,and to populationconnectivity.
Acknowledgments
We thank U. Gafvert, S. Johnson, J. Kroll, and H.
Quint for field assistance. S. Waterhousewas responsible for the genetic analyses. Primary funding was
provided by the National Park Service's Natural
Resource Preservationand Protectionprogram.
Literaturecited
BELANT,J.L., AND J.F. VAN STAPPEN.
2002. Island bio-
geographyof mammalsin Apostle Islands National
USA. NaturalAreasJournal22:180-185.
Lakeshore,
ANDT.L. KING.2003. Estimating
BOERSEN,
M.R., J.D. CLARK,
black bear populationdensity and genetic diversityat
TensasRiver,Louisiana
DNAmarkers.
usingmicrosatellite
WildlifeSocietyBulletin31:197-207.
CAUGHLEY,
1994. Wildlife ecology
G., ANDA.R.E. SINCLAIR.
andManagement.
BlackwellScience,Malden,Massachusetts,USA.
ENNIS,D.R., ANDT.F. GALLAGHER.
1994. PCR based sex de-
termination
assayin cattlebasedon thebovineamelogenin
locus.AnimalGenetics25:840-847.
K.C. 1997. A demographiccomparisonof a hunted
FLEMING,
and an unhuntedpopulationof black bearsin northern
Wisconsin.Thesis,Universityof Wisconsin,StevensPoint,
Wisconsin,USA.
FORAN, D.R., S.C. MINTA, AND K.S. HEINEMEYER.
1997. DNA-
basedanalysisof hairto identifyspeciesandindividuals
for
populationresearchand monitoring.Wildlife Society
Bulletin25:840-847.
FREDERICK,
D.J., ANDL. RAKESTRAW.
1976. A forest type map
of PicturedRocksandApostleIslandsNationalLakeshore
from the original GeneralLand Office survey notes.
NationalParkService,Munising,Michigan,USA.
D.L. 1994. Density-dependentpopulationregulaGARSHELIS,
tion of black bears.Pages 3-14 in M. Taylor,editor.
Density-dependent
populationregulationin black,brown,
92
* Belant et al.
BLACKBEARDENSITY
ANDGENETICS
ATAPOSTLE
ISLANDS
and polarbears.InternationalConferenceon Bear Research
and ManagementMonographSeries 3.
GOSSENS,
1998. Plucked hair
B., L.P. WAITS,ANDP. TABERLET.
samples as a source of DNA: reliability of dinucleotide
microsatellitegenotyping.MolecularEcology 7:1237-1241.
HAIG,S.M. 1998. Molecular contributionsto conservation.
Ecology 79:413-425.
KOHN,B. 1982. Status and management of black bears in
Wisconsin. Wisconsin Departmentof Natural Resources
Technical Bulletin 129. Madison, Wisconsin, USA.
R. 1994. Risk of populationextinctionfrom fixationof
LANDE,
new deleteriousmutations.Evolution 48:1460-1469.
M.R. VAUGHAN,AND J.W.
MILLER,D.A., E.M. HALLERMAN,
KASBOHM.
1998. Genetic variation in black bear populations fromLouisianaand Arkansas:examiningthe potential
influence of reintroductionsfrom Minnesota. Ursus 10:
335-341.
MILLER,S.D., G.C. WHITE,R.A. SELLERS,H.V. REYNOLDS,
K. TITUS,V.G. BARNES,JR., R.B. SMITH,R.R.
J.W. SCHOEN,
NELSON, W.B. BALLARD, AND C.C. SCHWARTZ.1997. Brown
and black bear density estimation in Alaska using radiotelemetryand replicatedmark-resighttechniques.Wildlife
Monograph133.
MOWAT, G., AND C. STROBECK. 2000. Estimating population
size of grizzly bearsusing haircapture,DNA profiling,and
mark-recaptureanalysis. Journalof Wildlife Management
64:183-193.
2000. Management policies 2001. US
PARKSERVICE.
NATIONAL
Departmentof the Interior,National Park Service, Washington, D.C., USA.
OTIS,D.L., K.P. BURNHAM, G.C. WHITE, AND D.R. ANDERSON.
1978. Statistical inference from capture data on closed
animalpopulations.Wildlife Monographs62.
D. 2003. An empiricalexplorationof data quality in
PAETKAU,
DNA-based population inventories. Molecular Ecology
12:1375-1387.
, AND C. STROBECK. 1994. Microsatellite analysis of
genetic variation in black bear populations. Molecular
Ecology 3:489-495.
. 1998. Ecological genetic studiesof bears
, AND
microsatellite
analysis. Ursus 10:299-306.
using
1995.
AND C. STROBECK.
, W. CALVERT,I. STIRLING,
Microsatelliteanalysis of populationstructurein Canadian
polar bears. MolecularEcology 4:347-354.
1998a. Gene flow
, G.F. SHIELDS,AND C. STROBECK.
between insular,coastal and interiorpopulationsof brown
bears in Alaska. MolecularEcology 7:1283-1292.
E. VYSE,
L. CRAIGHEAD,
,L.P. WAITS,P.L. CLARKSON,
1998b. Variation in genetic
R. WARD, ANDC. STROBECK.
diversityacross the range of North Americanbrown bears.
ConservationBiology 12:418-429.
PARKER, P.G., A.A. SNOW,M.D. SCHUG, G.C. BOOTON,AND
P.A. FUERST.1998. What molecules can tell us about
populations: choosing and using a molecular marker.
Ecology 79:361-382.
POOLE, K.G., G. MOWAT,AND D.A. FEAR. 2001. DNA-based
population estimate for grizzly bears Ursus arctos in
northeasternBritish Columbia, Canada. Wildlife Biology
7:105-115.
2002.
PROCTOR, M.F., B.N. MCCLELLAN, AND C. STROBECK.
Populationfragmentationof grizzly bears in southeastern
British Columbia,Canada.Ursus 13:153-160.
RAKESTRAW,L. 1976. Forest and culturalhistory in Apostle
Islands and Pictured Rocks National Lakeshore, Lake
Superior. Michigan Technological University, Houghton,
Michigan, USA.
REXSTAD, E., AND K. BURNHAM. 1991.
User's guide for
interactive program CAPTURE. Colorado Cooperative
Fish and Wildlife Research Unit, Fort Collins, Colorado,
USA.
SCHMIEGELOW,
F.K.A., AND T.D. NUDDS. 1987. Island bio-
geographyof vertebratesin GeorgianBay IslandsNational
Park. CanadianJournalof Zoology 65:3041-3043.
STANLEY, T.R., ANDK.P. BURNHAM.1999. A closure test for
time-specific capture-recapturedata. Environmentaland
Ecological Statistics6:197-209.
2003. CloseTest: A program for
, ANDJ.D. RICHARDS.
testing capture-recapturedata for closure. US Geological
Survey, Fort Collins Science Center, Fort Collins, Colorado, USA.
S.A. 1995. Spring and summerhabitatuse and food
STORLID,
habits of black bears in northern Wisconsin. Thesis,
Universityof Wisconsin, Stevens Point, Wisconsin, USA.
TABERLET,P., H. MATTOCK,C. DUBOIS-PAGANON,AND J. BOUVET.
1993. Sexing free-rangingbrown bears Ursus arctos using
hairs found in the field. MolecularEcology 2:399-403.
D.R. 1996. Black bear population dynamics, home
TRAUBA,
range,andhabitatuse on an islandin Lake Superior.Thesis,
Universityof Wisconsin, Stevens Point, Wisconsin, USA.
1998. Genetic signaturesof
P.M., ANDC. STROBECK.
WASER,
in Ecology and Evolution
Trends
interpopulationdispersal.
13:43-44.
ANDD.L. OTIS.
K.P. BURNHAM,
WHITE,G.C., D.R. ANDERSEN,
1982. Capture-recaptureand removal methods for sampling closed populations.Los Alamos NationalLaboratory
LA-8787-NERP.Los Alamos, New Mexico, USA.
AND
C. STROBECK,
D. PAETKAU,
WOODS,J.G., B.L. MCLELLAN,
1996. DNA fingerprintingapplied to markM. PROCTOR.
recapturebear studies. InternationalBear News 5:9-10.
M. PROCTOR,
D. LEWIS,B.N. MCLELLAN,
, D. PAETKAU,
1999. Genetic tagging of free-ranging
AND C. STROBECK.
black and brown bears. Wildlife Society Bulletin 27:
616-627.
Received: 25 February 2004
Accepted: 3 December 2004
Associate Editor: G. Shields
Ursus 16(1):85-92 (2005)