Impact of Deer Browsing on
Regeneration in Mixed Stands in
Southern New England
David B. Kittredge, Department of Forestry and Wildlife
Management, Holdsworth Hall, University of Massachusetts,
Amherst,MA 01003; and P. Mark S. Ashton, Yale Schoolof
Forestry and Environmental Studies,Marsh Hall, 360 Prospect
Street, New Haven, CT 06511.
ABSTRACT. Browsingpreferencesby white-taileddeer were evaluatedfor 6 tree speciesin northeastern
Connecticut.
Deerdensity
averaged
23/mile
2.Deerexhibited
nospecies-specific
preferences
forseedlings
greater
than19 in. For seedlingslessthan19 in., hemlockandblackbirchwerepreferred.Redmaple,sugarmaple,and
white pine seedlingswere avoided.Red oak seedlingswere neitherpreferred nor avoided.A much higher
proportionof seedlings
greaterthan19.7 in. in heightwasbrowsed,regardlessof species.Browsingpreferences
for speciesin thesmallerseedlingclass,combinedwith a lackofpreferencefor speciesin thelarger classmay
resultinfuturestands
withlessdiverse
treespecies
composition.
Deerdensities
in excess
of23/mile
2 maybe
incompatible
withregenerationof diverseforestsin southernNewEngland.North.J. Appl.For. 12(3):115-120.
Studies
have
demonstrated
that
deer
can
negatively
affect
theestablishment
of naturalregeneration
onforestedlandsin
thenortheastern
United States[Stoeckeleret al. 1957,Tierson
etal. 1966,Marquis1981,FrelichandLorimer1985,Tilghman
1989,MetropolitanDistrictCommission
(MDC) 1989].Deer
densities
of 60-80/mile2 on theQuabbin
Reservation
in
centralMassachusetts
result in virtually no regeneration
(MDC 1989). Behrendet al. (1970) and Tilghman(1989)
estimated
thatdensities
of approximately
17-18/mile
2 are
compatiblewith adequatehardwoodregenerationin the
Adirondack region of northernNew York State and the
Alleghenyregionof Pennsylvania.
Preferentialbrowsingby deerof sometreespeciesalters
the short-andlong-termdevelopment
andfuturecomposition of a stand(Marquis1981,Whitney 1984,Frelichand
Lonmer1985,Tilghman1989).Furthermore,
commercially
desirablespecies,suchasredoakandsugarmaple,areknown
to be dependent
on an adequatestockingof advancedreproduction in the understoryif they are to be successfully
regenerated
(Marquisetal. 1992,Sander1977,Hornbeckand
Leak 1992,Kelty 1988,Leak et al. 1987).Browsingby deer
could thereforesignificantlyimpact the establishmentof
certainspeciesin the understory,andconsequent
composition anddevelopmentof futurestands.
Most northeastern
studieshaveinvestigated
the relationsh•pbetweendeer browsingand regenerationin northern
hardwood,hemlock,or Alleghenyhardwood(cherry-maple
type)standsin New York or Pennsylvania
(Stoeckeler
et al.
1957,Tiersonetal. 1966,Marquis1981,FrelichandLorimer
1985, Tilghman1989). Therehavebeenno suchinvestigationspublishedfor standslocatedin thetransitionhardwood
regionof southernNew England(Connecticut,Massachusetts,andRhodeIsland).With a differentmixtureof species
(standsfrequentlydominatedby oak), deer browsepreferencesandimpacton regeneration
may be different.
In southern
New Englandmostforestsconsistof a mosaic
of naturallyregenerated
even-agedmixed-species
standsin
the60-80 yr ageclass,resultingfroma varietyof large-scale
naturalandhuman-induced
disturbances,
includingagriculturalabandonment,
clearcutting
for fuelwoodor pinetimber,
andthehurricaneof 1938.Theprevalentspeciescomposition
of this regionis referredto as the "transitionhardwoodswhitepine-hemlockforestvegetationzone"by Westveldet
al. (1956), andtypicallyincludessuchspeciesas red oak,
blackoak,whiteoak, red maple,sugarmaple,blackbirch,
white birch, white ash,severalspeciesof hickory,eastern
white pine, andeasternhemlock.
Braun(1950) referredto theforestcomposition
of southernNew Englandasbeingin eitherthehemlock-whitepinenorthernhardwoodsregionor the glaciatedsectionof the
oak-chestnut
forestregion.The boundarybetweenthesetwo
regionsrunsroughlyeast-westthroughnorthernConnectiNJAF12(3)1995 115
cut, but is indefinite. Westveld et al. (1956) stated that
hardwoodstypical of more northernforestsand southern
forestsmeet in a complexmosaicof covertypes.With the
exceptionof the coastalpitchpineforestsof CapeCod and
thehigherelevationnorthernhardwoodforestsof theBerkshireHills region,the forestsof the tri-statesouthernNew
Englandregioncanbe considered
similarin overallcomposition.
Becausesomanyof theforestsof thisregionareevenaged
and in a sawtimbersize class(i.e., commerciallymature;
Brookset al. 1993),it is appropriate
to considerthepotential
for future establishment
of their naturalregeneration.Our
objectivesin this studywere to surveyregeneration
in the
transitionhardwoodtypeto characterize
the overallprevalence of deer browsing,browsepreferences,and potential
impactof thesepreferences
on regeneration.
Study Area
The studywas conductedon the 7800 ac Yale-Myers
Forest,locatedin thenortheasterncornerof Connecticut.It is
largelycomposed
of a mosaicof even-agedstandsoriginating fromnaturalandhuman-induced
disturbances
typicalof
the region. Becausethe standswere too young for final
commercialharvest,few regenerationharvestshave been
conducted
overthepast50 yr, butthinningsor somelevel of
intermediatesilviculturaltreatmenthavebeenimplemented
throughout
theproperty.The goalof thesethinningswasto
improvethe qualityof residualstandsby removinginferior
individualsor speciesof low commercialvalue (e.g., red
maple,blackbirch),withoutalteringtheeven-aged
structure
of the stands(0liver 1978).
Methods
In 1982, 30 transectswere randomlylocatedthroughout
theYale-MyersForest.Ten 13.2ft radiusplotsweresystematicallylocatedalongeachtransect,
atintervalsof 66 ft. Each
transectwas alignedin a north-southdirection.The plots
were revisitedoccasionallyin subsequent
years,but dueto
incompletedata,no analysiscanbe made.In summer1986,
pelletgroupcountsandregeneration
surveyswereconducted
on thesetransects.Resultsand conclusionsreportedin this
paperarebasedsolelyon 1986 data.
The meannumberof pellet groups/plotwas determined
for eachcovertype.Estimatesof deerdensitywerederived
by the techniquedescribedby Bennettet al. (1940) and
EberhardtandVan Etten(1956) usinga defecationrateof 13
pelletgroupsperday,anestimatedperiodof deposition
(time
sincethe plots had last been visited and all previouspellet
groupsremoved),and13.2-ftradiusplots.
Fuller(1991) indicatedthelimitationsof thepelletgroup
methodto accuratelyindex deer densitiesor population
changes.
Ourestimates
werefurthercomplicated
by thefact
thatpelletgroupcountsweremadein the summer,afterthe
springtimebirth of fawns. As a result, estimatesof deer
densitymaybelow,sincefawnswereaddedtothepopulation
in May, andwerenot depositing
pelletgroupsoverthe full
deposition
period.For thesereasons,we reportthemeanand
116
NJAF12(3)1995
standarderrorof the numberof pelletgroups/plot
by cover
type,in additionto the derivedestimateof deerdensity.
The three predominantoverstorycover typesin which
samplingoccurred
werehardwood(largelycomposed
of red
oak,redmaple,sugarmaple,andblackbirch),n = 136plots,
hemlock/hardwood,n = 110 plots;andeasternwhite pine/
hemlock/hardwood,n = 30 plots.The remaining24 plots
occurredin eitherpurehemlockstandsor wetlandsandwere
notincludedin thestudydueto thesmallsamplesizeof each
condition.Becauseof theirrandomassignment,
thedistribution of transectsvariedby covertypes.
All seedlings
werecountedby species,
andassigned
to one
of two size classes(i.e., lessthan 19.7 in., and 19.7 to 36 in.,
hereafterreferredto asthesmallandlargesizeclasses).
Each
seedlingwasassessed
for signsof browsingthathadoccurred
sincethe previousgrowingseason.If a seedlinghadoneor
more twigs browsed,it was consideredbrowsed.No finer
resolutionof browseimpact(e.g., the numberof browsed
twigs on an individualseedling)wasmade.The numberof
seedlings
peracreandtheincidenceof browsingwerecalculated for the six principalspeciespresent:red oak, black
birch,redmaple,whitepine,sugarmaple,andeastern
hemlock.
Chi-square
analysis
wasused
todetect
browse
p.reference
by speciesandsizeclass,by comparingavailabilityof seedlings(i.e., the proportionof seedlingsby speciesavailable)
with theproportionof browsedseedlings
by species(White
andGarrott1990,186-191).If deerexhibitednopreference,
browsedseedlingsof a given specieswould appearat the
sameproportionasthat speciesappearsin the mix of availableseedlings.
Highly significantdifferencesindicateeither
a strongpreferencefor or avoidanceof a given species
Preferenceor avoidancewas determinedby comparingthe
availabilityof a particularspecieswith theconfidenceinterval around the estimate of its use. Confidence
limits were
calculatedby the methodsuggested
by White and Garrott
(1990, 188), using Bonferroninormal statistics.A significance level of 0.10 was used.
Results
Deer Density
Estimatesof deerdensityrangedfrom 21.8 to 26.7 deer/
mile2withincover
types
(Table1).Meandeerdensity
across
covertypeswas23.3deer/mile
2.
SeedlingAvailability and Occurrenceof Browsing
Resultsof theregeneration
surveywerereported(Kittredge
andAshton,1990).The overwhelmingmajorityof seedlings
Table 1. Pellet group counts (mean number of pellet groups/plot
andstandarderrors)
anddeerdensityestimates(meannumber
of deer/mile'-, and standard errors), by cover type.
Pellet
groups/
Covertype
plot
Deed
S.E
mile2
S.E.
Hardwood
0.985
0.153
21.8
3.4
Hemlock/hardwood
Pine/hem/hardwood
1.109
1,200
0.215
0.281
24.6
26.7
4.7
6.2
Overall forest
1.058
O.118
23.3
2.6
Tabla 2. Mean number of seedlingsper acre in the small size class(lessthan 19.7in. in height) and mean percentagebrowsed,by species
and cover type (standard errors in brackets).
Hardwood
Species
No.
Red Oak
185
%
35.0
[32]
Black birch
Red maple
White pine
Sugarmaple
Hemlock
Hemlock-hardwood
[4.9]
714
53.0
[187]
[4.3]
No.
87
Pine-hemlock-hardwood
%
25,0
[29]
[6,5]
942
36,0
[200]
No.
222
%
46.0
[48]
2359
[9.4]
44.0
[5,5]
[801 ]
[9.8]
931
14.0
656
8.5
1183
23.2
[100]
[2.4]
[116]
[2,9]
[356]
[6.2]
40
36.0
74
12.0
1407
26.6
[11]
[7.8]
[29]
[6.8]
[550]
[7.7]
272
16.0
49
0
323
23.0
[88]
[4.2]
[19]
0
[156]
[11.6]
212
58.2
[39]
[5.2]
590
52.9
[98]
[4.9]
observedwas in the small size class(Tables 2 and 3), for all
species,in all covertypes.Incidenceof browsingon seed-
lings in the large size classwas generallyhigherthan on
seedlingsin the small size class.
BrowsePreferencesby Species
Thereweresignificantdifferences
betweentheproportion
of 5 of 6 speciesof smallseedlingsavailableandthe proportion browsed(Table 4). Deer preferentiallybrowsedeastern
hemlockandblackbirch seedlingsin all covertypes.They
avoidedredmaple,whitepine,andsugarmaple.Redoakwas
browsedat ratessimilarto its availability.
Little differenceexistedbetweenthe proportionof larger
seedlingsavailable and the proportionbrowsed(Table 5)
suggesting
little speciespreferencefor seedlings> 19.7 in.
Therewerenospecies
preferences
expressed
in thehardwood
andhemlock-hardwood
covertypes.Redoakandredmaple
weremodestlypreferredin thepine-hemlock-hardwood
cover
type. In general, the lack of a consistentpreference(as
570
52.7
[188]
[9.6]
comparedto thesmallsizeclass,acrosscovertypes),andthe
higherratesof browsefor largesizeclassseedlings
(Table3)
suggesta muchhigherprobabilityof a large seedlingbeing
browsed,regardlessof its speciesor cover type. However,
becausewe measuredonly frequencyof browsing,rather
than percentof stemsbrowsed,we cannotstateunequivocallythattheimpactof browsingis greateronthetallerclass
of seedlings.
Discussion
Browseof Seedlingsin the Large and Small Size Classes
Regenerationguidelinesfor uplandoaks (Sanderet al.
1984)requirea minimumof oneseedoriginstemgreaterthan
4.5 ft in heighton 59% of the surveyplots(1/735-ac) in a
stand(i.e., 434/ac).Fewerstemsmaybe compensated
for by
stumpsprouts.For northernhardwoods,Leak et al. (1987)
recommenda minimumof onestemat least3 ft in heighton
65% of surveyed
milacreplotsin anuneven-aged
stand(i.e.,
Table 3. Mean number of seedlings per acre in the large size class (greater than 19.7 inches in height) and mean percentage of seedlings
browsed, by species and cover type (standard errors in brackets).
Hardwood
Species
No.
Red Oak
20
[8]
Black birch
142
[61 ]
Red maple
White pine
Sugarmaple
Hemlock
Hemlock-hardwood
%
No.
%
No.
%
100.0
62.0
7
20.0
40
[12.9]
[4]
[19.9]
[26]
55.0
[7.9]
83
39.0
[30]
[1.1]
0.0
21
50.0
3
[7]
[13.5]
[2]
5
79.0
1
[2]
[39.3]
[1]
74
57.0
1
[26]
[10.4]
[1]
100.0
3
[2]
3
[2]
Pine-hemlock-hardwood
100.0
81
33.0
[32]
[16.6]
54
94.0
[40]
[13.9]
3
100.0
[2]
0.0
50.0
[25.0]
96
45.1
[78[
[47.6]
0
NJAF12(3)1995 117
Table 4. Proportionof seedlingsin the small sizeclassavailable,and confidenceintervalaroundthe proportionof seedlingsbrowsed
(P = 0.10 levelof significance),Chi-squarevalueindicatingthe degreeof difference,andpreference(P),avoidance(A),or indifference(I)
Hardwood cover type
Species
Confidence
interval,
proportion
browsed Proportion
available
Red oak
Black birch
0.06156< P <0.11044
0.45741 < P <0.54459
0.079
0.303
0.45369
97.98902
I
P
Red maple
White pine
Sugarmaple
0.14003< P <0.20597
0.00710< P <0.03090
0.03762< P <0.07838
0.395
0.017
0.116
94.45434
0.17442
22.20103
A
I
A
Hemlock
0.13080<
0.090
45.36451
P
0.036
0.393
2.00983
9.18304
I
P
103.23072
A
P <0.19520
Hemlock-hardwood
Red oak
Black birch
0.01496<
0.41605<
P <0.04504
P <0.50395
Red maple
0.05263< P <0.09937
0.274
White pine
0.00240< P <0.02160
0.031
7.88416
0.020
14.80000
0.246
88.43433
P
0.037
0.389
2.85289
45.15839
I
P
Sugarmaple
Hemlock
0.37845<
P <0.46555
A
Hemlock-hardwood-pine
Red oak
Black birch
0.03610<
0.45426<
P <0.05790
P <0.50574
Red maple
White pine
Sugarmaple
0.10985< P <0.14415
0.15351 < P <0.19249
0.02466< P <0.04334
0.195
0.232
0.053
59.30590
30.33563
10.56836
A
A
A
Hemlock
0.12118<
0.094
57.65969
P
P <0.15682
650 stems/ac).Prior to a shelterwoodremovalcutin an even-
agedstand,theyrecommend
roughly5000 stems/ac
of estab-
developbeyondthe seedlingphase.Basedon the browsing
pressureappliedby estimateddeer densitiesof approm-
lished 3-4 ft trees.
mately23/mile
2,whenseedlings
arelessthan19.7in.deer
Based on these recommendations
and the documented
scarcityof regeneration
greaterthan19.7in., theremay not
be a sufficientnumberof seedlings
presentin theunderstory
to adequatelystocka futurestand,If continuallysuppressed
below 19.7 in. by browsing,at leastsomespecieswill not
preferto browsecertainspecies.However,thereis a greater
likelihood that seedlingsover 19.7 in. in height will be
browsed,irrespectiveof species.Given the smallnumberof
seedlingsper acrehigherthan 19.7 in., it is plausiblethat a
relativelysmalldeerpopulationcouldmaintainsuppressed
Table 5. Proportion of seedlings in the large size class available, confidenceinterval around the proportion of seedlingsthat were
browsed (P= 0.10 level of significance), Chi-square value indicating the degree of difference, end preference (P), avoidance (A), or
indifference (I).
Hardwood cover type
Species
Confidence
interval,
proportion
browsed Proportion
available
Red oak
Black birch
0.02909<
0.42237<
Red maple
White pine
Sugarmaple
Hemlock
P <0.13691
P <0.61763
0.075
0.536
0.01333
0.02827
0.02014< P <0.11986
0.00000< P <0.05867
0.19226< P <0.36774
0.079
0.019
0.279
0.18750
0.33333
0.00024
0.00000<
0.012
0.75000
0.071
0.847
0.48400
0.08312
P <0.04736
Hemlock-hardwood
Red oak
Black birch
0.00000<
0.77012<
P <0.11596
P <1.01588
Red maplea
White pine
0.00000< P <0.09358
Sugarmaplea
Hemlock
0.00000<
P <0.11790
0.031
1,10000
0.010
0.90000
0.010
0.40000
0.031
0.14545
9.64065
10.21914
A
I
Hemlock-hardwood-pine
Red oak
Black birch
0.16363< P <0.32437
0.09471 < P <0.23329
0.146
0.296
Red maple
White pine
Sugarmaple
Hemlockb
0.22345< P <0.39655
0.00000< P <0.04288
0.18151< P <0.34649
0.197
0.011
0.350
0.000
a Browsing
notobserved.
b Nohemlock
seedlings
> 19.7in.observed.
118
NJAF12(3)1995
9.87805
1.22500
2.76050
P
A
P
I
A
regenerationby continuedbrowsing.Even •f deer density
declined,regeneration
releasecouldbestifledunlessreduced
densitywas maintained.Stoeckeleret al. (1957) concluded
that deer populationsneededto be kept low for 6-8 yr in
northernWisconsinin order to permit successful
regeneration to become established.
The paucityof seedlingsin thelargesizeclasscompared
to the small size class is similar to patternsdocumented
elsewherein theNortheast.In a studyusingfencedexclosures
wahin clearcutsin the Alleghenyregionof Pennsylvania,
Marquis (1981) reportedthat after 22 yr in an area with
approximately
25 deer/mile
2, the totalnumberof stems
•ns•deandoutsideexclosures
wasnot significantlydifferent.
Therewere,however,differencesin speciescompositionand
heightgrowth.Therewere significantlyfewer black cherry,
sugarmaple,redmaple,yellowpoplar,andredoakseedlings
outsidetheexclosures,
especiallyfor stemsgreaterthan59 in.
T•erson et al. (1966) reported a similar pattern in the
Adirondacks,
withadensity
ofapproximately
21deer/mile
2.
Inside exclosures,there were over 5261 stems/acof sugar
maple,white ash,andyellow birchover 35 in. in height,but
outsideexclosurestherewere no yellow birch over 35 in.
present,and only 162 sugarmaple stems/acover 35 in.
Tdghman (1989) similarly reporteda differencein ratesof
seedlingheight growth. In her study,regenerationbecame
establishedundervirtually all deer densities(rangingfrom
10-80deer/mile2),
yetthenumber
thatdeveloped
intotaller
heightclasseswaslimitedunderhigherdensities(i.e., greater
than40deer/mile2).
Trumbull
etal.(1989)reported
asimilar
pattern of fewer taller seedlingsoutside exclosuresthan
w•thinexclosures,
in a studyon theAlleghenyplateauwhere
deerdensity
approximated
22deer/mile
2.
BrowsePreferenceby Species
The analysisindicatesa browsepreferencein the smaller
seedlingclass for hemlock and black birch (Table 4). If
relative deer densitiesremain approximatelythe same as
thoseat the time of the study,thepossibilityof changein the
futureforestcompositionexists.
In our studythisbrowsepreferencewasnot expressed
in
seedlingsin thelargesizeclass.The incidenceof browsewas
considerablyhigher for this height class, for all species.
Seedlingsthat attainedthisheightwere browsed,regardless
of species.
We postulatethat historic deer densitiesin this region
were low, accountingfor the currentdominanceof hemlock
•n the canopy.This may also be attributableto preferential
t•mberharvestingpractices.Our findingsalsosuggestthatif
currenttrendsin deerpopulationdensitiespersist,the future
forestcompositioncould be one dominatedby red maple
(becausea lack of browsepreferencein the smallsizeclass
andits vigoroussproutingability) andblackbirch(because
of relatively high numbersof seedlingspresent). Future
manipulation
of theoverstoryanditsinfluenceonlightlevels
•n the understorywill play an importantrole in how these
standsdevelop.Red maple is consideredto be tolerantof
shade,especiallyin the seedlingstage,while blackbirchis
consideredmore intolerant [Burns and Honkala 1990].
Severebrowsingof regeneration
couldmutestanddevelopmentpatternsthat promotespeciesstratification.In extremecases,forestscouldresembleopen,monostoried
woodlands,dominatedby relativelyfewertreespecies,butwith an
extensiveherbaceous
or woodybut unpalatableunderstory
(Tilghman1989). This patternhas also beenobservedin
central Massachusetts
on the QuabbinReservation(MDC
1989).
Application
Managersinterestedin regenerating
standsshouldconsiderthebrowsingpressure
thatdeerpopulations
canimpose
on seedlings.Given the differentialbrowsepreferenceexpressed
for smallseedlings
andthealmostuniversalbrowse
pressureplacedon large seedlingswhen deer population
densities
areapproximately
23/mile
2,it isclearthatabsolute
numbersof seedlings
presentalonewouldbe inadequate
for
assessing
theregeneration
potentialof a givenstand.Marquis
et al. (1992) accountfor the influenceof deerwhen consideringregeneration
in theAlleghenyhardwoodregion.Their
deerimpactindexacknowledges
theeffectof deerbrowseon
regeneration,
basedon the amountof "deerfood" available
andthe deerdensity.They estimatethat in an areawith 24
deer/mile2 and low-to-mediumamountsof deer food available,regeneration
couldundergoa shift in composition,or
even trend towards a monoculture.
Conditions of low deer
foodavailabilityaresubjectivelyestimatedby circumstances
suchasthepresenceof a distinctbrowseline, highnumbers
of browsedstems,andunderstories
dominatedby speciesthat
are not preferredby deer suchas fern.
Theresultsof thisstudyenableforesters
to appreciate
deer
browseimpactin forestsof southernNew England.At an
approximate
densityof 23 deer/mile
2, a shiftin species
compositionof the futurestandcouldoccurunlessstepsare
taken,suchasthosesuggested
by Marquiset al. (1992) (e.g.,
fencing,accelerateddeerharvest,fertilizationto accelerate
regeneration
development,thinningsthroughoutthe areato
providea greaterquantityof deerfood andreducethe deer
impactindexor pressure).
Finally, theseresultsprovide an initial benchmarkfor
assessingdeer impact on regenerationin southernNew
England,analogousto similarstudiesconductedelsewhere
in the northeast(Lake States,Adirondacks,Allegheny region).Suchinformationwill be usefulin thefuturedevelopment of a deer impactindex, similar to the one reportedby
Marquis et al (1992).
Literature
Cited
BEHREND,
D.F., G.F. MATrFELD,
W.C. TIERSON,
andJ.E. WILEYliI. 1970. Deer
densitycontrolfor comprehensive
forestmanagement.
J. For. 68: 695700.
BENNETT,
L.J., P.F. ENGLISH,
andR. McCAnN.1940.A studyof deerpopulationsby useof pellet-group
counts.J. Wildl. Manage.4: 398-403.
B R•UN,E.L. 1950. Deciduousforestsof easternNorth America. B lakistonCo.
Philadelphia.596 p.
BROOKS,
R.T., D.B. K•rmEr)GE,
andC.L. AI•EmCH.
1993. Forestresourcesof
southernNew England.USDA For. Serv.Res.Bull. NE-127.71 p.
BURNS,R.M., and B.H. HONV.
ALA.Technicalcoordinators.1990. Silvics of
NorthAmerica.Vol. 2, Hardwoods.USDA Ag. Handb.654. 877 p.
NJAF12(3)1995 119
EBERHARDT,
L., and R.C. VANETTEN.1956. Evaluationof the pellet-group
countasa deercensus
method.J. Wildl. Manage.20: 70-74.
FRELICH,
L.E., and e.G. LORIMER.
1985. Currentand predictedlong-term
effectsof deerbrowsingin hemlockforestsin Michigan,USA. Biol.
Conserv. 34: 99-120.
FULLER,
T.K. 1991.DO pelletcountsindexwhite-taileddeernumbersand
population
change?
J. Wildl. Manage.55(3):393-396.
HORNBECK,
J.W.,andW.B. LEAK.1992.Ecologyandmanagement
of northern
hardwood
forestsinNewEngland.USDA For.Serv.Gen.Tech.Rep.NE159.44 p.
KELTY,
M.J. 1988.Sources
of hardwood
regeneration
andfactorsthatinfluencethesesources.
P. 17-30 in Guidelinesfor regenerating
Appalachian
hardwoodstands.W. Va. Univ. Books,Morgantown.
KITIILEDGE,
D.B. andP.M.S. ASHTON.
1990.Naturalregeneration
patternsin
even-aged
mixedstands
in southern
NewEngland.North.J.Appl.For.7:
163-168.
LEAK,W.B., D.S. SOLOMON,
andP.S. DEBALD.1987.Silviculturalguidefor
northernhardwoodtypesin the northeast
(rev.).USDA For. Serv.Res.
Pap.NE-603.36 p.
MARQUIS,
D.A. 1981. Effect of deer browsingon timber productionin
Alleghenyhardwood
forestsof northwestern
Pennsylvania.
USDA For.
Serv.Res.Pap.NE-475.10 p.
MARQUIS,
D.A., R.L. ERNST,
andS.L. STOUT.
1992.Prescribingsilvicultural
treatments
inhardwood
stands
oftheAlleghenies
(rev.).USDAFor.Serv.
Gen.Tech.Rep.NE-96. 101p.
METROPOLrrAN
DISTRlCT
COMMISSION.
1989. NaturalResourceManagement
120
NJAF12(3)1995
ReNew Panelreport.Belchertown,MA. 44 p.
OLIVER,
C.D. 1978.The developmentof northernred oak in mixed standsin
centralNew England.Yale Univ.Sch.For.andEnviron.Stud.Res.Bull
91.63 p.
SANDER,
I.L. 1977.Manager'shandbookfor oaksin thenorthcentralstates
USDA For.Serv.Gen.Tech.Rep.NC-37.35 p.
SANDER,
I.L., P.S. JOHNSON,
and R. ROGERS.
1984.Evaluatingoak advance
reproduction
intheMissouriOzarks.USDA For.Serv.Res.Pap.NC-251
16p.
STOECKELER,
J.H., R.O. STROTIIMANN,
andL.W. KREFrrNG.
1957. Effect of deer
browsingon reproduction
in the northernhardwood-hemlock
type m
northeastern
Wisconsin.J. Wildl. Manage.21: 75-80.
TIERSON,
W.C., E.F. PATRIC,
and D.F. BEHR_END.
1966. Influence of whitetaileddeerontheloggednorthernhardwoodforest.J. For. 64:801-805
TILGHMAN,
N.G. 1989. Deer densitiesand forest regeneration.J. W•ldl
Manage. 53(3): 524-532.
TRUMBULL,
V.L., E.J.ZIELINSKI,
andE.C. AHARRAH.
1989.The impactof deer
browsingontheAlleghenyforesttype.North.J. Appl.For.6:162-165
WESTVELD,
M., et al. 1956.Naturalforestvegetation
zonesof New England
J. For. 54(5): 332-338.
WHITE,G.C., andR.A. GARROTI'.
1990.Analysisof wildliferadio-trackmg
data.AcademicPress,SanDiego,CA. 383 p.
WHITNEY,
G.G. 1984.Fifty yearsof changein the arborealvegetationof
Heart'sContent,an old-growthhemlock-whitepine-northern
hardwood
stand.Ecology65(2):403-408.