Comparison of diet and prey selectivity of the Pyrenean desman and the Eurasian water shrew using next-generation sequencing methods

(1)

ContentslistsavailableatScienceDirect

Mammalian

Biology

j o ur na l ho me p ag e :w w w . e l s e v i e r . c o m / l o c a t e / m a m b i o

Original

investigation

Comparison

of

diet

and

prey

selectivity

of

the

Pyrenean

desman

and

the

Eurasian

water

shrew

using

next-generation

sequencing

methods

Marjorie

Bifﬁ

a,∗

_,

_Pascal

_Laffaille

a

_,

_Jérémy

_Jabiol

a

_,

_Adrien

_André

b

_,

_Franc¸

_ois

_Gillet

b

_,

Sylvain

Lamothe

a

_,

_Johan

_R.

_Michaux

b,c

_,

_Laëtitia

_Buisson

a

a_EcoLab,_Université_de_Toulouse,_CNRS,_INPT,_UPS,₁₁₈_route_de_Narbonne,₃₁₀₆₂_Toulouse_cedex_9,_France

b_Laboratoire_de_Biologie_Evolutive,_Unité_de_Génétique_de_la_{Conservation,}_Université_de_Liège,_Institut_de_Botanique_B22,_Quartier_Vallée_1,_Chemin_de_la

Vallée4,4000Liège,Belgium

c_CIRAD,_Agirs_Unit,_TA_C-_22/E-_Campus_{international}_de_Baillarguet,₃₄₃₉₈_Montpellier_Cedex_5,_France

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received9December2016 Accepted12September2017 HandledbyLauraIacolina Availableonline14September2017 Keywords: COI Dietaryoverlap Foragingstrategy Scatanalyses Semi-aquaticmammal

a

b

s

t

r

a

c

t

Inthisstudy,theinteractionsbetweentwosemi-aquaticmammals,theendangeredPyreneandesman GalemyspyrenaicusandtheEurasianwatershrewNeomysfodiens,wereinvestigatedthroughtheanalysis oftheirsummerdietusingnext-generationsequencingmethods,combinedwithanalysesofprey selec-tivityandtrophicoverlap.Thedietofthesepredatorswashighlydiverseincluding194and205genera forG.pyrenaicusandN.fodiensrespectively.Overall,bothspeciesexhibitedrathernon-selectiveforaging strategiesasthemostfrequentlyconsumedinvertebrateswerealsothemostfrequentandabundantin thestreams.ThissupportedageneralistforagingbehaviourforG.pyrenaicusandN.fodiensinthestudy area.ThePiankaindex(0.4)indicatedasigniﬁcantbutmoderatedietaryoverlapasG.pyrenaicusmostly reliedonpreywithaquaticstageswhereaspreyofN.fodiensweremainlyterrestrial.Moreover,no dif-ferenceinG.pyrenaicuspreyconsumptionwasfoundinpresenceorabsenceofN.fodiens.Adifferential useoftrophicresourcesthroughmechanismssuchasplasticfeedingbehaviourordifferencesinforaging micro-habitatarelikelytofacilitatethecoexistencebetweenthesetwomammalspecies.

Introduction

Biodiversityconservationrequiresathoroughknowledgeofthe

complexnatureofinteractions betweenspeciesand their

envi-ronment.Accordingtotheecologicalnichetheory(Hutchinson,

1957),sympatricspeciesareexpectedtoexhibitsomeniche

dif-ferentiationinpreyorhabitatusetocoexist(Pianka,1974).Under

limitingconditions,strongnicheoverlapbetweenspeciesmaylead

tostrongcompetitiveinteractions,andmayultimatelyresultinthe

competitiveexclusionoftheweakestcompetitorimplying

conse-quencesonitslocalandregionaldistribution(Wiszetal.,2013).The

studyofresourceuseandpotentialnicheoverlapwith

competi-torsseemsthuscrucialtoassessthevulnerabilityofspecieswith

Abbreviations:FOdiet,frequencyofoccurrenceofinvertebratetaxainthe

preda-tordiet(numberoffaecescontainingthetaxondividedbythetotalnumberof predatorfaeces); FOstream,frequencyofoccurrenceofinvertebratetaxainthe

streams(numberofSurbersampleswiththetaxadividedbythetotalnumberof Surbersamplescollectedinthestudysites).

∗ Correspondingauthor.

E-mailaddress:m.bifﬁ@live.fr(M.Bifﬁ).

conservationconcern.Thisisparticularlytruewhenfocusingon

specieslivinginecosystemsthatarevulnerabletoanthropogenic

alterations,suchasfreshwaterecosystems(Dudgeonetal.,2006).

Anyquantitativeor qualitativeshiftin theresource availability

and/ordiversity(e.g.preycommunity)followingdisturbance(e.g.

aquaticpollution,alterationofriverﬂow)mayexacerbatethe

com-petitiveinteractionsbetweenconsumers.

ThePyreneandesman,Galemyspyrenaicus(E.Geoffroy

Saint-Hilaire,1811,Talpidae)isasmallsemi-aquaticmammal,endemic

tothePyreneesMountainsandtheIberian Peninsula(northern

and central Spain, northern Portugal). The species is listed as

vulnerable by the IUCN (Fernandes et al., 2008)and is legally

protectedinallthecountriesencompassingitsdistributionarea.

ThealarmingdeclineofG.pyrenaicuspopulationsovertherecent

decadesacrossitswholerange(Charbonneletal.,2016;Gisbert

andGarcía-Perea,2014)hasencouragedlocal,nationaland

Euro-peanconservationinitiatives(e.g.,inFrance:Life+Desman,2013;

Némozetal.,2011).Yet,inspiteofanincreasingnumberof

stud-iesfocusingonthisspecies(e.g.,AymerichandGosàlbez,2015;Bifﬁ

etal.,2016;Charbonneletal.,2016;Escodaetal.,2017),the

respec-tiveinﬂuenceofprey,competitorsandpredatorsonitssurvivaland

distributionstillremainstobeexplored.

http://dx.doi.org/10.1016/j.mambio.2017.09.001

(2)

Modernadvancesinmoleculargeneticswererecentlyapplied

toinvestigate the diet of G. pyrenaicus in the French Pyrenees

(Gillet, 2015; Gilletet al.,2015).Theystressed thewide

diver-sityof prey(156genera) withhalfof them beingfoundinthe

dietofG.pyrenaicusfortheﬁrsttime.Besides,theseﬁrststudies

conﬁrmedthedietarypreferencesof G.pyrenaicustowards

Tri-choptera,EphemeropteraandPlecoptera(Insecta),aspreviously

knownfromtraditionalmethodsbasedongutcontentanalysisor

visualinspectionoffaecesremains(e.g.,Bertrand1994;Castiénand

Gosálbez1995;Santamarina1993;SantamarinaandGuitian1988).

Inaddition,Gillet(2015)emphasizedasubstantialconsumptionof

terrestrialprey.

Several aquatic or semi-aquatic animals are knownto prey

and use similar food and habitat resources as G. pyrenaicus

and act as potential competitors for resource acquisition: the

browntroutSalmotruttaandthedipperCincluscinclus(Bertrand,

1994; Santamarina, 1993; Santamarina and Guitian, 1988), the

EurasianwatershrewNeomysfodiens(CastiénandGosálbez1999;

Morueta-Holmeetal.,2010;SantamarinaandGuitian1988),the

MediterraneanwatershrewNeomysanomalus(Santamarina,1993)

and thePyreneanbrook newt Calotritonasper(Bertrand, 1994).

AmongthesespeciesN.fodiensexhibitssimilarhabitatpreferences

asG.pyrenaicus,i.e.swiftly-ﬂowingstreamswithnumerous

shel-ters(e.g.cavities)ontheriverbeds(Greenwoodetal.,2002;Keckel

etal.,2014).N.fodiensisalsoknownasanopportunisticfeeder

consumingbothaquatic(e.g.,crustaceans,insectlarvae)and

ter-restrial(e.g,coleopterans,gastropods,spiders,earthworms)prey

(CastiénandGosálbez,1999;Churchﬁeld,1985;Churchﬁeldand Rychlik,2006;Haberl,2002).Moreover,N.fodiensexhibitsa

simi-larpolyphasicactivitypatternasG.pyrenaicus(Meleroetal.,2014),

withactivityphases varyingacross seasons(Churchﬁeld, 1984;

Greenwoodetal.,2002;Keckeletal.,2014;Rychlik,2000).Despite

importantsimilaritiesintheirhabitat,resourcepreferencesandlife

style,fewstudieshavefocusedonthetrophicoverlapbetweenN.

fodiensandG.pyrenaicustodate.Thesestudies,limitedtosmall

samplessizes(e.g.,onlysixG.pyrenaicussamplesinSantamarina,

1993)andrelyingsolelyonthevisualinspectionofpreyremainsin

faecesorgutcontent,concludedthatcoexistenceanddiet

differ-entiationwerelikelytheresultofadifferentuseofmicro-habitats

(CastiénandGosálbez,1999).

Duringthepastyears,moleculargeneticmethodsbasedon

fae-cesanalyseswereincreasinglyusedinsteadof‘traditional’methods

basedongutcontentanalysisorvisualinspectionoffaecesremains

(Pompanonetal.,2012).Theirmainadvantagesarethat(i)they

donot requestanimal sacriﬁcecomparedtogutcontent

analy-sis,(ii)theyidentifypreywithhightaxonomicresolution(genus

andspecieslevels),(iii)includinghighlydegradedorsoft-bodied

species(molluscs,earthworms)thatcannotbeidentiﬁed

morpho-logically,(iv)theyarelesstime consuming,and (v)theydonot

requireanytaxonomicalexpertiseoftherangeofpotentialpreyas

longastaxaarepresentingeneticdatabases(seePompanonetal.,

2012forareview).Todate,thoughthedietofG.pyrenaicushasbeen

recentlydescribed usingmolecularmethods(Gillet,2015;Gillet

etal.,2015;Bifﬁetal.,2017),suchdataaboutN.fodienshavenever

beengathered,makinganycomparisonoftherespectivedietsof

thesetwospeciesratherspeculative.

Inthisstudy,weaimedatdescribingthesummerdietofthese

twomammals,usingrecentmoleculargeneticmethods,inapart

oftheFrenchPyrenees(i.e.,theAriègedepartment)where they

areknowntoco-occur.Moreover,wecomparedthepreyofthese

predatorspecieswithstreambenthicinvertebratecommunities,on

whichG.pyrenaicusmostlyfeeds.Thisallowedusassessingtheprey

selectivityofthesetwomammalspeciesandthepotentialtrophic

overlapduringsummerinordertodiscussmechanismsthatcould

facilitatetheircoexistence.

Materialandmethods

Studyareaandsamplingsites

Samplingwas conducted in 65 sites spread over the Ariège

department,aFrenchadministrativeregioninthePyrenees

Moun-tains(Fig.1).ThisareaexhibitsrelativelyhighoccurrenceofG.

pyrenaicus(Bifﬁetal.,2016;Charbonneletal.,2015,2016),and

the presence of N. fodiens was recently reported (Charbonnel

et al., 2015). The mean elevation of the 65 sampling sites is

757.9±259.3mandvariesbetween375.2mand1755.6m.Mean

monthly streamﬂow equals to1.1±2.0m3_/s _with_a _maximum

of13.2m3_/s₍_Charbonnel_et_al.,₂₀₁₆_)._Natural_zones_with

herba-ceous or shrubby vegetation (52.1±36.1%), agricultural lands

(43.9±34.7%)andforests(40.6±32.5%;CorineLandCover©DB

2012)dominatethelandcover surroundingthe65sites.Inthis

mountainous area, the climate is cold (mean annual air

tem-perature=10.4±1.3◦C, SAFRAN © DB) and wet (mean annual

rainfall=1141.0±110.9mm,SAFRAN©DB).

Faecessamplingandmoleculargeneticsanalyses

FaecescollectionofG.pyrenaicusandN.fodienswereconducted

twicebetweenJuneandSeptember2015inthe65selectedsites

(Fig.1).Skilledobserversmeticulouslyinspectedeachemergent

item(i.e.,rock,treerootorbranch)alonga250mriverbed

tran-sect.Thislengthisacompromisebetweenthehomerange(HR)size

andtheaveragedistancetravelledalongstreamchannels(ADTS)

during24hforN.fodiens(narrowandlinearHRalongastream:

106–509m2_,_ADTS:₄₉_±₂₅_m;_Cantoni,_1993;_Lardet, ₁₉₈₈₎_and

G. pyrenaicus (linear HR: ≈500m; ADTSbetween resting sites:

≈250m;Meleroetal.,2012,2014).Oursamplingprotocolmeets

therecommendationsofParryetal.(2013)whofoundthat

repeat-ingvisitsinasinglesiteratherthanenlargingthesampledarea

allowedbetterdetectionprobabilitiesfortheEurasianotter.Itis

alsoinagreementwithCharbonneletal.(2014)whoobtained

rea-sonabledetectionprobabilitiesforG.pyrenaicus(0.58)inasurvey

of 100m-longsectionsof riversusingtemporal replicates.The

searchforfaecesisastandardandeffectiveprotocolfor

detect-ingthepresenceofG.pyrenaicus(Charbonneletal.,2014)andhas

alsoprovenefﬁciencyforN.fodiens(AymerichandGosálbez,2004).

AllputativeG.pyrenaicusorN.fodiensfaeces basedontheir

colour,size,smellandposition,werecollectedandanalysedwith

moleculargenetictoolsbothtoconﬁrmthespeciesidentityofthe

consumerandthepreyconsumed.Followingthemanufacturer’s

instructions, DNA wasextracted from faecal samplesusing the

StoolMiniKit(Qiagen Inc.,Hilden,Germany).PCRampliﬁcation

wasduplicated for each sample ona portionof the

mitochon-drialcytochromeoxydaseIgene(COI;fordetails,seeGilletetal.,

2015).NegativeDNAextractionandnegativePCRcontrolswere

includedintheprocedure.AgencourtAMPureXPbeads(Beckman

CoulterLifeSciences,IN,USA)and thenQuant-iTTM _PicoGreen®

dsDNAAssayKit(ThermoScientiﬁc,MA,USA)wereusedto

puri-fiedPCRproductsandtoquantifiedpurifiedampliconsrespectively.

Afterthequantiﬁcationstep,productswerepooledat

equimolar-ityandsenttotheGIGAGenomicsplatform(UniversityofLiège,

Belgium) for sequencing on an ILLUMINA MiSeq V2 benchtop

sequencer.Rawsequencesweresortedandﬁlteredusingascript

mixingFASTXToolkit(http://hannonlab.cshl.edu/fastxtoolkit;

23-09-16) and USEARCH (Edgar,2010)functions (seeAndré etal.,

2017fordetailsonbioinformatics).Sequencesoriginatinglikely

fromextractionorPCRcontaminantswereexcludedfromfurther

analyses.Theremainingsequenceswerethencomparedwith

pub-lished sequencesavailable intheonline BOLDdatabaseforCOI

(Ratnasinghamand Hebert,2007).Sequencesthat hadaunique

(3)

Fig.1. Locationofthestudyareaandsamplingsites(blackdots)inFrance.

score(i)higherorequalto90%fortheidentiﬁcationofpreytaxaat

thegenuslevel(whenpossible),and(ii)higherorequalto99%for

theidentiﬁcationofthepredatoratthespecieslevel.

PreytaxawerevalidatedaspotentiallyoccurringinFranceand

inthePyreneesusing theFrenchNationalInventoryof Natural

Heritage(Muséumnationald’Histoirenaturelle,2003–2017)and

theFrenchOfﬁceforInsectsandtheirEnvironment(OPIE-Benthos,

2017)onlinedatabases.Taxaidentiﬁedasendemicofotherpartsof

theworldwerekeptintheanalysisanddesignatedbyanasterisk

hereafter(*)astheymorelikelycorrespondtoageneticallysimilar

taxonpresentinthePyreneesbutabsentinthegeneticdatabases.

Whenpreytaxacouldnotbeidentiﬁedatthegenuslevel,theywere

groupedtogetheratahighertaxonomiclevel.

Aquaticmacroinvertebratessampling

Aquaticmacroinvertebrateswerecollectedin27outofthe65

sites,amongwhich N.fodiensandG.pyrenaicuswerepresentin

19and25sites,respectively(called“Neomyssites”and“Galemys

sites”,respectively).G.pyrenaicuswasdetectedaloneineightsites

(“Galemysalonesites”)andbothspecieswerefoundtoco-occurin

17sites(“Galemys+Neomyssites”).N.fodienswasdetectedalonein

twosites(“Neomysalonesites”).

The available habitats for aquatic macroinvertebrates were

describedaccordingto12categoriesofsubstratetype:bryophytes,

hydrophytes, helophytes, litter, twigs roots, algae,large stones

(>25cm),cobbles/pebbles(2.5–25cm),gravel(0.2–2.5cm),mud,

sand,bedrock;andfourcategoriesofcurrentvelocity:null,slow,

medium,fast.Thepercentagecoverofeachcombinationof

sub-stratetype/currentvelocitywasvisuallyestimatedineachsite.

Six Surber net samples (mesh size: 500␮m, sampled area:

0.04m2₎_were_conducted_in_each_site_according_to_a_stratiﬁed

sam-plinginthedominanthabitats(e.g.mainlycoarsemineralsubstrate

inhigh-ﬂowfacies)coveringmorethan5%ofeachstreamtransect

inordertoberepresentativeofthesite.Thesampled

macroinver-tebrateswerefrozenbeforebeingsorted,countedandidentiﬁedat

thefamilylevel(exceptforOligochaetaandHydracarina)following

Tachetetal.(2000)atthelaboratory.Themeandensity(numberof

individuals/m2_),_and_the_frequency_of_occurrence_(FO

stream:relative

numberofSurbersampleswiththetaxon)werethencalculatedfor

eachsiteandinvertebratetaxon.

Mean macroinvertebrates densities were compared with

Wilcoxonsign-rankedtestsandP-valuesadjustedwiththe

Bon-ferronicorrection,betweencategoriesofsites(i.e.,Galemysalone

sites,NeomysalonesitesandGalemys+Neomyssites).

Dietcompositionandcomparisonbetweenmammals

Presenceor absenceof prey in each faeces of G. pyrenaicus

andN.fodienswereusedtodescribethedietcompositionofboth

species.Thefrequencyofoccurrence(FOdiet,i.e.,thenumberof

fae-cescontainingthetaxondividedbythetotalnumberoffaeces)inG.

pyrenaicusandN.fodiensdietwascalculatedforeachorder,family

andgenusofinvertebratesandforthedifferenttypesofprey’s

habi-tat(i.e.,exclusivelyaquatic,exclusivelyterrestrialorwithaquatic

andterrestrialstages).Tocomparethefaecescompositionbetween

thetwo species,aCorrespondenceAnalysis(CA)wasappliedto

presence-absencedataofpreyinfaecesatthegenuslevel.The

coor-dinatesofeachfaecesontheﬁrstandsecondaxesoftheCAwere

comparedbetweenthetwospecieswithaStudenttestfor

homo-geneousvariances.Rarepreytaxa,withFOdietbelow5%bothinG.

pyrenaicusandN.fodiensfaeces,werenotincludedinthisanalysis.

ToassesspreyselectivityofG.pyrenaicusandN.fodiens,the

pres-enceofinvertebratesinthefaeceswereconsideredatthesitescale.

Whenaninvertebratetaxonwaspresentinatleastonefaecesof

themammalspeciesinasite,itwasconsideredasbelongingto

themammaldietinthissite.PreyFOdietinthefaeceswerethen

calculatedforthe19Neomyssitesandthe25Galemyssites.Data

onpreyavailabilityinthestreams(i.e.,FOstreampersitefromthe

Surbersamples)andpreyconsumedbythepredatorinthesite

(i.e.,FOdietpersiteinthesampledfaeces)atthefamilylevelwere

comparedusingtheIvlev’selectivityindex(Ivlev,1961).Thisindex

rangesfrom−1(i.e.,avoidanceofprey)indicatingthatthetaxon

isfrequentinstreamsbutneverfoundinfaeces,to1(i.e.,active

selection)whenthetaxonisrareinstreamsbutfoundinavery

highnumberoffaeces.Azerovalueindicatesthatconsumptionis

proportionaltotheamountofinvertebratesavailableinstreams.

(4)

5%inboththeSurbersamples(FOstream)andthefaecesofG.

pyre-naicusandN.fodiens(FOdiet)werenotconsideredinthisanalysis.

Trophicoverlapbetweenmammals

TodeterminethedegreeoftrophicoverlapbetweenG.

pyre-naicusandN.fodiensduringsummer,thePianka’sdietaryniche

overlapindex(Pianka,1973)wascalculatedfromthepreyFOdietin

thefaecesatthegenusandfamilylevels.Thisindexrangesfrom0

(notrophicresourceusedincommon)to1(fulldietaryoverlap).

Moreover,totestiftheco-occurrenceofbothmammalspecies

inasitecouldmodifythedietofG.pyrenaicus,thefaeces

composi-tioninpreywascomparedbetweentheGalemysalonesites(8)and

theGalemys+Neomyssites(17).Themeanfrequencyofoccurrence

ofeachpreytaxonfoundinthefaecesofG.pyrenaicus(atthefamily

level)werecomparedwithaWilcoxonsign-rankedtestbetween

thetwocategoriesofsites.ThedietofN.fodienscouldnotbe

com-paredbetweenNeomysalonesitesandGalemys+Neomyssitesdue

toaverysmallnumberofNeomysalonesites(2).

AllstatisticalanalyseswereconductedinR3.3.1(RCoreTeam,

2014)usingtheade4andspaapackages.

Results

Molecularidentiﬁcationofpredatorsproducingthefaecesand

preycontents

Atotalof464faeceswerecollectedfromthe65sampledsites

(7±2faecespersite)andanalysedusingmoleculargeneticstools.

AfterthetwoPCRampliﬁcations,atotalof2,160,447readswere

obtained.1,348,331readswerecorrectlyassignedto199faeces

(42.9%)thatbelongedtoG.pyrenaicus(3±2faecespersite)and

whosepresencewasconﬁrmedin58sites.Amongthem,itwas

possibletoidentifypreyin184faeces.Seventy-ninefaeces(17%

offaeces–463,706reads),including78faeceswithdiet

informa-tion,belongedtoN.fodiens(2±1faecespersite)whichconﬁrmed

itspresencein39sites.Fromtheremainingfaeces,51(11.0%)were

assignedto14otherhostspecies(mammalssuchasNeomys

anoma-lus,Apodemussp.,Sorexsp.;batsorbirds).Molecularidentiﬁcation

ofpredatorsandpreyfailedfor29.1%ofthesamplesdueto

insuf-ﬁcientDNAquantityordegradedsamples.

Overalldiversityofpreyinthepredatorfaeces

ThefaecesofG.pyrenaicusandN.fodienscontainedalmost

exclu-sivelyinvertebrate preyexcept oneamphibian foundinone G.

pyrenaicusfaeces.Preydiversitywashighwith10and9classes,

30and33orders(Table1),111and117families,and194and205

genera(AppendixAintheSupplementarymaterial)forG.

pyre-naicusandN.fodiens,respectively.Intotal,309differentgenera

wereidentiﬁedwhose178wereconﬁrmedtooccurinthe

Pyre-nees,95inFranceand9wereendemicofotherpartsoftheworld

(e.g.,Australia,America,Asia).Theremaining27taxacouldnotbe

identiﬁedatthegenuslevelinthedatabasesandtheirdistribution

rangeisunknown.

Forthetwomammalspecies,frequentlyconsumedprey(i.e.,

presentinmorethan5%ofthefaeces)representedasmall

propor-tionofthetotalpoolofpreyconsumed(12.9%and20.5%ofgenera

eatenbyG.pyrenaicusandN.fodiens,respectively).Thedominant

preyfoundinthefaecesofG.pyrenaicuswereInsecta(93.5%of

thefaeces),Malacostraca(23.4%)andDiplopoda(21.7%).Among

insects,G.pyrenaicuspreyedmorefrequentlyonEphemeroptera

(71.2%)whichincludedthemostfrequentfamily(Heptageniidae:

61.4%)butnotthemostfrequentgeneraHydropsyche(Insecta

Tri-choptera Hydropsychidae: 52.7%). N. fodiens seemed to have a

morediversedietwithDiplopoda(89.7%offaeces),Insecta(83.3%),

Arachnida(47.4%),Gastropoda(25.6%) andMalacostraca(24.3%)

occurringfrequentlyinthecollectedfaeces.Thepreyfoundthe

mostfrequentlyfor thisspeciesbelongedtothegenusGlomeris

(75.6%).

AbouthalfofG.pyrenaicuspreybelongedtoEphemeroptera,

Ple-copteraandTrichopteraorderswhileonly16.6%ofpreybelongedto

theseordersforN.fodiens.Theproportionofexclusivelyterrestrial

preywashigherforN.fodiens(70.8±29.4%)thanforG.pyrenaicus

(20.6_±30.5%),whiletheproportionofpreywithaquaticand

ter-restrialstageswashigherforG.pyrenaicus(74.6±31.4%)thanforN.

fodiens(26.1±28.3%).Exclusivelyaquaticpreyweremarginalfor

bothspecies.

Amongthe309identiﬁedgenerainthefaecesofbothmammals,

only90generawereconsumedbybothG.pyrenaicusandN.fodiens.

Thepreytaxaconsumedbyonlyoneofthetwospeciesweremostly

exclusivelyterrestrialgenera(52and80%inG.pyrenaicusandN.

fodiensfaeces,respectively).

Preycompositionoffaeceswassigniﬁcantlydifferentbetween

G.pyrenaicusandN.fodiensontheﬁrstaxisoftheCA(Student:

t=14.2,df=149.2,p<0.001).Thisdifferencecouldbeduetothe

majorityofterrestrialpreyinN.fodiensfaecesincontrastwithprey

withbothaquaticandterrestrialstagesfoundmorefrequentlyin

G.pyrenaicusfaeces(Fig.2).

Aquaticmacroinvertebratesavailability

Inthe27 siteswherestreaminvertebratesweresampled,G.

pyrenaicusandN.fodiensconsumedrespectively81and88

inver-tebratefamiliesintotaloutofwhich40and25familiesincluded

invertebrateswithatleastoneaquaticstageintheirlifecycle.

Fromstreaminvertebratessamples,atotalof51different

inver-tebratefamilieswereidentiﬁed(27±5persite)amongwhich42

familiesofInsecta,1Malacostracafamily,1Hydracarinafamily,2

Gastropodafamilies,1Bivalviafamily,1ﬂatwormfamily,2

Clitel-latafamiliesand1Nematodesfamily.Insectataxahadthehighest

meandensitiesintheSurbersamples(Fig.3a).

Nineinvertebratefamilies(Rhagionidae,Caenidae,

Mermithi-dae, Planariidae, Leptoceridae, Gyrinidae, Calopterygidae,

Hydrophilidae,Psephenidae)wereexcludedfromfurtheranalyses

as they were found in the Surber samples and in faeces with

FOstream and FOdiet below 5%. Conversely, one-quarter of the

invertebratefamilieswithatleastoneaquaticstagefoundinthe

faecesofbothG.pyrenaicusandN.fodienswerenotfoundinthe

invertebratessamples.Thisconcerned9familiesoutof40forG.

pyrenaicusand5outof25familiesforN.fodiens.However,these

familieshada FOdietbelow10%inbothdiets,exceptthefamily

Anthomyiidae(20%inG.pyrenaicusdietand10.5% inN.fodiens

diet)andThaumaleidae(10.5%inN.fodiensdiet),andwerethus

alsoexcludedfromfurtheranalyses.

Densitiesofavailableinvertebrateswerenotsigniﬁcantly

differ-entbetweentheGalemysalonesites,theNeomysalonesitesandthe

Galemys+Neomyssites(Wilcoxonsign-rankedtests:alladjusted

p>0.05)allowingthecomparisonofinvertebratescontentin

fae-cescollectedinthedifferentcategoriesofsites.Forty-sixfamilies

and 50familieswereavailablein Galemysalone sitesand

Gale-mys+Neomyssitesrespectively,including45familiescommonto

bothtypesofsites.

Preyselectivitybetweenthetwomammals

Overall,themostfrequentlyconsumedpreybyG.pyrenaicus

corresponded to the most frequent and abundant invertebrate

taxa in streams (Fig. 3a and b). Seven out of the nine most

abundantpreywereconsumedinaccordancewiththefrequency

of occurrence and density estimated in the streams

(5)

Table1

Preytaxaidentiﬁedwithpositivematches(≥80%)from184faecesofG.pyrenaicusand78faecesofN.fodienscollectedinthestudyarea.Thefrequencyofoccurrence(%of faecescontainingtheprey)isdisplayed.SeeAppendixAinthesupplementarymaterialforthefulllistoftaxaatthefamilyandgenuslevel.*indicatesmisidentiﬁedtaxaby geneticdatabases.

Classes Frequencyofoccurrence(%) Orders Frequencyofoccurrence(%)

G.pyrenaicus N.fodiens G.pyrenaicus N.fodiens

Insecta 93.5 83.3 Ephemeroptera 71.2 32.1 Diptera 65.8 50.0 Trichoptera 63.6 30.8 Plecoptera 52.7 35.9 Coleoptera 14.1 43.6 Hemiptera 3.8 17.9 Lepidoptera 3.3 11.5 Hymenoptera 1.6 9.0 Odonata 1.1 2.6 Dermaptera 0.5 1.3 Neuroptera 0.0 5.1 Orthoptera 0.0 1.3 Mecoptera 0.0 3.8 Archaeognatha 0.0 1.3 Malacostraca 23.4 24.4 Amphipoda 20.7 19.2 Isopoda 3.3 6.4 Decapoda 0.5 0.0 Diplopoda 21.7 89.7 Glomerida 15.8 80.8 Polydesmida 7.1 53.8 Julida 4.3 43.6 Chordeumatida 2.7 16.7 Arachnida 8.2 47.4 Araneae 3.3 25.6 Opiliones 2.2 23.1 Sarcoptiformes 1.1 2.6 Trombidiformes 0.5 1.3 UndeterminedArachnida* 0.5 1.3 Mesostigmata 0.5 0.0 Clitellata 6.0 17.9 Haplotaxida 6.0 17.9 Gastropoda 4.3 25.6 Stylommatophora 3.3 25.6 Hygrophila 1.1 0.0 Collembola 1.1 19.2 Entomobryomorpha 1.1 17.9 Poduromorpha 0.0 2.6 Symphypleona 0.0 1.3 Amphibia 0.5 0.0 Anura 0.5 0.0 Chilopoda 0.5 9.0 Lithobiomorpha 0.5 7.7 Geophilomorpha 0.0 1.3 Maxillopoda 0.5 0.0 Sessilia 0.5 0.0

UndeterminedNemerta* 0.0 2.6 UndeterminedNemerta* 0.0 2.6

Fig.2.CorrespondenceAnalysis(CA)computedonthefrequentlyusedpreytaxa(FOdiet>5%)derivedfrom184G.pyrenaicusand78N.fodiensfaeces.(a)Projectionsofprey taxaontheﬁrst(8.4%)andsecond(4.2%)axisoftheCA:preytaxaaredepicteddifferentlyaccordingtothetypeofhabitatduringtheirlifecycle.(b)ProjectionsofG.pyrenaicus (“Galemys”)andN.fodiens(“Neomys”)faecesontheﬁrstfactorialplane.

(6)

Fig.3. PreyselectivityforG.pyrenaicus(Galemys;darkbarsfor25Galemyssites)andN.fodiens(Neomys,lightbarsfor19Neomyssites).(a)Meandensitiesofinvertebrate preyintheSurbersamplesforeachsiteinGalemysandNeomyssites.(b)Frequencyofoccurrenceofprey(FOdiet)incollectedfaeces.(c)Ivlev’selectivityindex.Theelectivity indexisbasedonthefrequencyofpreyinfaecesrelativetopreyavailableinstreamswherethefaeceswerecollected.Blacktrianglesidentifyexclusivelyaquaticinvertebrate familiesasopposedtofamilieswithouttrianglesthatincludeinvertebrateswithatleastoneaquaticstageandoneterrestrialstageintheirlifecycle.

Simulidae,Nemouridae;Fig.3c). N.fodiensconsumed

Gammari-daeaccordingtotheirrelativeamountinthestreamsandavoided

theothermostabundantfamilies.

However,mostinvertebratefamiliessampledinstreamswere

avoidedbyG.pyrenaicusandN.fodienswithnegativeIvlev’svalues

below −0.25 for 22 and 31 available taxa out of 41,

respec-tively(Fig.3c).BothG.pyrenaicusandN.fodiensseemedtoavoid

Brachycentridae and Elmidae prey in spite of their high

den-sities in streams. An active selection of invertebrate prey was

highlightedforsevenfamiliesforG.pyrenaicus(Cordulegastridae,

Blephariceridae,Tipulidae,Scirtidae,Limnephilidae,Psychodidae,

Ephemeridae)andﬁvefamiliesforN.fodiens(Cordulegastridae

Dix-idae,Limnephilidae,Scirtidae,Tipulidae).

ThedietaryoverlapbetweenG.pyrenaicusandN.fodienswas

moderateatthegenus(Piankaindex=0.4)andfamily(0.5)levels.

NodifferenceintheFOdietofthepreyfoundinG.pyrenaicus

faeces wasdetectedat thefamily level betweenthesites with

orwithoutthepresenceofN.fodiens(Wilcoxonsign-rankedtest:

V=2748,p=0.3).Whenconsideringallconsumedpreyfamilies,72

wereconsumedbyG.pyrenaicusinGalemysalonesites(38with

FOdiet>5%)and86inGalemys+Neomyssites(44withFOdiet>5%).

Only47 taxawereconsumed inboth categoriesof sites.Ratios

betweennumbersofrareprey(i.e.,FOdiet<5%)andfrequentprey

(i.e.,FOdiet>5%)inG.pyrenaicusdietweresimilar(around50%)in

GalemysalonesitesandinGalemys+Neomyssites.

Discussion

NovelinsightsintothedietofG.pyrenaicusandN.fodiensduring

summer

This study provides additional support to the usefulness of

moleculargenetictoolsindietanalyses,asitprovidesanenhanced

identiﬁcationofpreycontentfromindirectsignsofpresence(i.e.,

faeces)ofpredators.Theproportionoffaecesmisidentiﬁcationin

theﬁeld(i.e.,belongingtootherhostspecies:11%)orunidentiﬁed

faecesinthelaboratory(29%)areconsistentwithpreviousstudies

ofG.pyrenaicusdiet(Gillet,2015)conﬁrmingtheissuesarisingfrom

visualidentiﬁcationoffaecesandDNAextractionfromdegraded

faeces.

Fromthe194invertebrategeneraidentiﬁedinthefaecesofG.

pyrenaicusinthepresentstudy,only61werecommonwiththe156

generaidentiﬁedinapreviousstudyusingsimilargeneticmethods

butconductedinthewholeFrenchPyrenees(Gillet,2015;Bifﬁetal.,

2017).Newlyidentiﬁedpreyincludedanamphibian,aquaticand

terrestrialsnails,leeches,earwigs,Hydracarina,Chilopoda,

Maxil-lopodaandCollembola.

Thisstudyalsoprovidesunprecedentedinsightsintothe

sum-mertrophicnicheofN.fodiens,byidentifyingpreyidentityatthe

familyandgenuslevelsfortheﬁrsttime(butseeattheorderlevel:

Castiénand Gosálbez,1999; Churchﬁeld,1985;Churchﬁeld and Rychlik,2006).Inaccordancewithpreviousstudiesconductedat

ahighertaxonomiclevel,wefoundthatN.fodiensfedon

terres-trialpreyatalargerextentthanG.pyrenaicus,withadominance

(7)

Ephemeroptera,PlecopteraandTrichopterawaslow(about17%)

comparedtoG.pyrenaicus(55%).

Thedensity and species composition of streaminvertebrate

communitiesdescribedinthisstudyareconsistentwithprevious

reportsoffreshwaterinvertebratefaunaoftheFrenchPyrenees

(e.g.,Brownetal.,2006;Finnetal.,2013).BothG.pyrenaicusand

N.fodiensexhibitedrathernon-selectiveforagingstrategiesasthe

mostfrequentlyconsumedinvertebrateswerealsothemost

fre-quentandabundantinthestreams.Thissupportedageneralistdiet

forbothspeciesinthestudyareaalthoughasigniﬁcantnumberof

taxawereconsumedinlowerorhigherfrequenciesthanexpected

(Fig.3).

N.fodiensavoidedahighernumberofpreyandactivelyselected

asmallernumberofpreythanG.pyrenaicusintheaquatic

envi-ronment.Thissupportsthedifferenceintrophicnichesbetween

thetwospecies,withmoreterrestrialtaxainN.fodiensdiet

com-paredwith G.pyrenaicus. This latter species hasmorphological

features adapted to livein aquatic environments (e.g.,webbed

feet,Palmeirim,1983;Puisségur,1935;highdivingskills:1–4min,

Richard and Micheau, 1975 compared to 3–24s for N. fodiens,

Lardet,1988;Mendes-SoaresandRychlik,2009)thatlikelyresult

inabetterefﬁciencyinthecaptureofaquaticpreythanN.fodiens.

Foragingefﬁciencymaydeterminepreyselectivityinorderto

maximiseindividualsuccess.Itdependsonthebalancebetween

theenergyprovidedbypreyconsumptionandtheenergeticcosts

offoragingunderwater. Themostvaluableresourcesmaythus

correspond to easy-to-catch prey with low mobility (e.g.,

Tri-choptera)and/orhighabundance(e.g.,Gammaridae),soft-bodied

preythatcanbecompletelydigested(e.g.,incontrastwithsmall

and chitinous Hydracarina or Coleoptera,Costa et al., 2015) or

largeprey(Bertrand,1994).Inthisstudy,invertebrateswithhard

andchitinousbodies(e.g.Elmidae,HydraenidaeHydracarina)and

hard casesmadeof wood(Brachycentridae) or mineral

materi-als(Glossossomatidae,Sericostomatidae)wereavoidedbybothG.

pyrenaicusandN.fodiens.

SomeofourresultsaboutpreyselectivitybyG.pyrenaicus

con-tradictpreviousobservationsbyBertrand(1994)andSantamarina

(1993).Forinstance,Rhyacophilidaewereavoidedand

Limnephil-idaeactivelyselectedbyG.pyrenaicusinourstudywhereasthe

oppositewas observedby Bertrand(1994). Chironomidaewere

avoided and Sericostomatidae activelyselected in Santamarina

(1993)whereaswefoundtheoppositepatternforthesefamilies.

Thesedifferencesmaybeduetodifferentsamplingmethods

(stom-achanalysesvs.digestedremainsinfaecesvs.molecularanalyses)

aswellassamplesizes.Theymayalsoreﬂectthegeneralistdietof

G.pyrenaicusthatmayvarybetweenregions(i.e.,acatchmentof

Spainvs.thewholeFrenchPyreneesvs.asub-regionoftheFrench

Pyrenees)andseasons(Santamarina,1993).Finally,the

variabil-ityinlife-historytraitswithinandbetweeninvertebratespeciesof

thesamefamilyorgenus(e.g.,developmentstage,bodysize,local

adaptations)mayinﬂuencetheirexposuretopredation.

Thefaecesandaquaticinvertebratessamplingswereconducted

duringsummerwhenmany late instarsof aquaticinsects have

emergedfrommountainstreamsandlefttheaquaticenvironment

(Fürederet al.,2004).Invertebratecommunities are dominated

bysmall-sizedinvertebratesforthisperiod.Thisimpoverishment

inpreydiversityanddensitymayexacerbatethetrophicoverlap

betweentheirpredators,suchasG.pyrenaicusandN.fodiens,

com-paredtotherestoftheyear.Inspiteofthis,thePianka’sindex

oftrophicoverlapbetweenthetwospeciesequalled0.4

indicat-ingasigniﬁcantbutratherlowoverlapinthesummerdiets.This

valueisconsistentwithapreviousestimateofoverlap(Castiénand

Gosálbez,1999).

G.pyrenaicusconsumedalargernumberofdifferentpreyinthe

siteswhereN.fodienswasalsopresent.Thisincreasewasdrivenby

ahigherconsumptionofterrestrialtaxabutwasnotlinkedtoany

increaseinrarepreyconsumptionasratiosofrareprey/frequent

preyweresimilarinsiteswithorwithoutN.fodiens.However,no

signiﬁcantdifferencewasfoundindietcompositionofG.pyrenaicus

betweensiteswhereN.fodienswasdetectedornot.Alltheseresults

suggestnoevidenceofashiftofG.pyrenaicusdiettowards

sub-optimalpreyinthepresenceofanotherinsectivorespecieswith

similarfeedingstrategies.

Theabsenceof shiftin G.pyrenaicusdiet inthepresence of

N.fodienscouldresultfromthenon-limitingtrophicresourcesin

streamsand/oronterrestrialhabitatsduringsummer.Actually,the

measureofnicheoverlapmaymostlyreﬂectthepotential

compe-titioninthecaseoflimitingresources(Abrams,1980).

ItcouldalsobeduetoresourcepartitioningbetweenG.

pyre-naicusandN.fodiens.First,asynchronyofseasonalordailyactivity

periodsmayplayanimportantroleinresourcepartitioningand

facilitatecohabitation(e.g.,Harringtonet al.,2009).A temporal

shiftofnichehasbeenobservedinherbivorestoreducethelength

ofthetrophic overlapatwaterholesduringaridseasons(Valeix

etal.,2007).RadiotrackingconductedonafewindividualsofG.

pyrenaicusandN.fodiensseemstoindicatepolyphasicactivity

pat-terns(Lardet,1988;Meleroetal.,2014;Stone,1987)withasimilar

timingofactivityduringdayandnightperiods.Thissuggeststhat

thismechanismisprobablynotinvolvedintheresource

partition-ingbetweenthesetwospecies.

Second,successfulresourcepartitioningmayberelatedtohow

thepredatorsaccesstofoodresourcesaccordingtotheirforaging

strategies. Thesegregationof trophic nichesbased on

differen-tiation of foraging modes and foraging micro-habitats is well

documentedamongshrews(e.g.,ChurchﬁeldandRychlik,2006;

ChurchﬁeldandSheftel,199).First,G.pyrenaicusandN.fodienshave

differentbodysize(i.e.,10–15cmvs.7–9cmwithoutthetailforG.

pyrenaicusandN.fodiens,respectively)andmass(i.e.,50–60gvs.

8–17gforG.pyrenaicusandN.fodiens,respectively).Duetothis

differenceofsize,thetwospeciescouldtargetcontrasted

develop-mentstageandpreysizewithinagiventaxa,asitisknowntoallow

resource partitioning among shrew species (Churchﬁeld et al.,

1999).Second,theuseofdifferenthabitatstratawasalsoidentiﬁed

insmallmammalcommunitiestomodulatetheintensityof

inter-speciﬁccompetition(CastiénandGosálbez,1999;Churchﬁeldand

Sheftel,1994;Rychlik,1997).Specializationonparticular

micro-habitatsinstreamsisalsoknownforothersemi-aquaticmammals.

Forinstance,theplatypusOrnithorhynchusanatinusdoesnot

allo-cateanequalforagingefforttobenthicmacroinvertebratesacross

allhabitats,itspreferencegoingtopoolsandlittoralmarginsrather

thanrifﬂes(McLachlan-Troupetal.,2010).Theoveralllow

selec-tivity,activeselectionoravoidanceofdifferenttaxa,togetherwith

thedistincttrophicnicheweobservedinthewholestudyareain

G.pyrenaicusandN.fodiens,suggestthattheyusedistinctforaging

micro-habitats(CastiénandGosálbez,1999)withinthestreamand

riparianmosaic.Otheraquaticpredatorssuchasthebrowntrout

S.truttacouldhavehigherdietoverlapwithG.pyrenaicusdueto

theirbroaduseofhabitatswithinstreamsandhighlyopportunistic

feedingstrategies(GillerandGreenberg,2015).

Methodologicalconsiderationsandperspectives

Whilemolecularmethodsprovideanenhancedidentiﬁcationof

preycomparedtotraditionalmethods(i.e.,visualinspectionofprey

remainsingutsorfaeces),theyalsohavesomeshortcomings.First,

identifyingmanyadditionalpreytaxawithmoleculartools

com-paredtotraditionalmethodsisconsistentwithClareetal.(2014)

who suggest that preyoccurrence data obtainedfrom

(8)

preyandoverestimaterareprey(Clareetal.,2014;Krügeretal.,

2014).Nevertheless,20taxaidentiﬁed,suchasCollembolaor

sev-eralgenusoftheTachinidaefamily,areunlikelytobedirectpreyof

G.pyrenaicusorN.fodiens,astheyarepartofthesoilmicrofaunaor

areinvertebrateparasites.Othertaxa,suchassomeAnthomyiidae

(e.g.Polietes)orsmallterrestrialColeoptera,mayhavebeen

col-lectedwiththefaecesastheymaydevelopatthelarvalstageorfeed

onscat.Finally,taxaidentiﬁedaspreycouldhavebeenpassively

ingestedwiththeconsumptionofpredatororscavenger

inverte-brates(e.g.Trichoptera,Plecoptera,Odonata).Thiscontributesto

thedebateaboutthehighsensitivityofnext-generation

sequenc-ingmethodsandthedetectionofsecondarypredation(Sheppard

etal.,2005).

Second,moleculardatadonotgiveinformationonthestage

orsizeatwhichpreyareconsumed.Thisisaparticularlystrong

limitationwhen(i)feedingstrategiesareadapted tothemouth

morphologyandsizeofthepredatoror(ii)preyexhibit

impor-tantvariationinhabitatuseduringtheirlifecyclesuchasmany

aquaticinvertebratespecies.Giventherelativeshortnessofthe

terrestrialstageincomparison withtheaquaticstagefor many

invertebratespecies,theprobabilitythat invertebratesfoundin

thefaeceswereconsumedattheaquaticstage(oratthetimeof

emergence)remainshigh,whichissupportedbyprevious

morpho-logicalidentiﬁcationofpreyitemsinfaecalsamples(Trichoptera;

e.g.,Bertrand,1994;DuPasquierandCantoni,1992).The

combi-nationofmolecularandtraditionalmethodsofpreyidentiﬁcation

seemsthusimportanttobringdetailedinformationontheidentity,

sizeandstageofeatenprey.

Third,themoleculardataweuseddonotallowaquantitative

assessmentofthepreyconsumed(e.g.relativeandabsolute

abun-dance/biomass ofdifferenttaxa). Klareetal. (2011)stated that

suchqualitativeestimates(i.e.,presence-absenceandfrequencyof

occurrenceofprey)tendtooverestimatenichebreadthanddietary

overlapsbetweenspeciesleadingpotentiallytounreliable

conclu-sions.However,theyalsopointedoutthatsuchbiasremainslow

whenthedietsofthecomparedspeciesarecomposedofsimilar

taxonomicgroupsofprey,whichisthecaseforG.pyrenaicusandN.

fodiens.Moreover,thelikelybiasindietestimationduetotheuse

offrequenciesofoccurrenceissimilarbetweenG.pyrenaicusand

N.fodienswhichenablesareliablequalitativecomparisonofdiet

overlapandpreyselectionpatternsintermsofpreyidentity.Thisis

supportedbythePiankaindexoftrophicnicheoverlapquantiﬁed

inthisstudywhichisconsistentwithpreviousobservationsbased

onquantitativemethodsofpreyestimates.

Despitethoselimitations,molecularmethodsprovedefﬁciency

intheidentiﬁcationofahighlydiversedietforG.pyrenaicusand

N.fodiens that werehighlighted toexhibit a generalist feeding

behaviour.Thissuggeststhattheidentityofpreytaxaisnotthe

most important criteria for both mammals prey selection and

thattheymayexhibitsometolerancetovariationin prey

com-munitycomposition.Hence,stream communitychanges dueto

anthropogenicimpactsonriverecosystemsshouldhavemoderate

consequences(Costaetal.,2015)aslongasfoodresourcesremain

abundantenoughintheirrespectiveforaginghabitats.However,

pressuresonpreycommunitiescouldincreasethetrophicoverlap

betweenthetwospeciesaswellaswithotherspotential

competi-tors.

ThisstudyalsostressedthatthepresenceofN.fodiensdidnot

affectG.pyrenaicuspreyselection.However,thehigherproportion

ofexclusivelyterrestrialpreyinN.fodiensdietfoundhere

com-paredtootherdietarystudies(e.g.Churchﬁeld,1984;DuPasquier

andCantoni, 1992)maysuggesta shiftinthedietofN.fodiens

towardsamoreterrestrialnicheinpresenceofG.pyrenaicus.The

verysmallnumberofsiteshostingN.fodienswithoutG.pyrenaicus

didunfortunatelypreventustotestforthispotentialcompetitive

interaction.

Identifyingmoreﬁnelytheforagingmicro-habitatsofG.

pyre-naicusandN.fodiens,comparingthedietofN.fodiensinpresence

andabsenceofG.pyrenaicusandobservingtheirbehavioural

inter-actionsintheﬁeldalongtheyearshouldbefurtherinvestigatedin

thelightofterrestrialinvertebratessamplingandtheuseofdietary

quantitative data (e.g.abundance or biomass of prey in scats).

Thiswouldimproveourknowledgeaboutmechanisms

facilitat-ingtheircohabitationorcausingpotentialcompetitiveinteractions

andtheirvulnerabilitytohabitatandtrophicresourcealterations.

Acknowledgements

Wearegratefultoallpeoplewhohelpedcollectingdatainthe

ﬁeld:C.Dupuyds,M.Alvarez.,C.Lauzeral,F.Colas,F.JulienandV.

Lacaze.Wealsothankthe“Conservatoired’EspacesNaturels

Midi-Pyrénées”(CEN-MP),especiallyM.NémozandF.Blanc,fortheir

preciousadviceatmanystepsofthestudy.Thisstudywasfunded

byANRT(Cifren◦2011/1018),EDF(ElectricitédeFrance)andthe

EuropeanUnion(FEDER)inthecontextoftheLIFE+Nature

pro-grammedevotedtoG.pyrenaicus(LIFE13NAT/FR/000092).M.Bifﬁ

wassupportedbyaPhDfellowshipgrantedbythe“EcoleDoctorale

Sciencesdel’Univers,del’Environnementetdel’Espace”(SDU2E)

attheUniversityofToulouse.

AppendixA. Supplementarydata

Supplementarydataassociatedwiththisarticlecanbefound,in

theonlineversion,athttp://dx.doi.org/10.1016/j.mambio.2017.09.

001.

References

Abrams,P.,1980.Somecommentsonmeasuringnicheoverlap.Ecology61,44–49. André,A.,Mouton,A.,Millien,V.,Michaux,J.,2017.Livermicrobiomeof

Peromyscusleucopus,akeyreservoirhostspeciesforemerginginfectious diseasesinNorthAmerica.Infect.Genet.Evol.52,10–18.

Aymerich,P.,Gosàlbez,J.,2015.Evidenciasderegresiónlocaldeldesmánibérico (Galemyspyrenaicus)enlosPirineosmeridionales.Galemys27,31–40. Aymerich,P.,Gosálbez,J.,2004.Laprospeccióndeexcrementoscomometodología

paraelestudiodeladistribucióndelosmusga ˜nos(Neomyssp.).Galemys16, 83–90.

Bertrand,A.,1994.Répartitiongéographiqueetécologiealimentairedudesman desPyrénées,Galemyspyrenaicus(Geoffroy,1811)danslesPyrénées franc¸aises.Thèsededoctorat.UniversitéPaulSabatierdeToulouse(50pp). Bifﬁ,M.,Charbonnel,A.,Buisson,L.,Blanc,F.,Némoz,M.,Laffaille,P.,2016.Spatial

differencesacrosstheFrenchPyreneesintheuseoflocalhabitatbythe endangeredsemi-aquaticPyreneandesman(Galemyspyrenaicus).Aquat. Conserv.Mar.Freshw.Ecosyst.26,761–774.

Bifﬁ,M.,Gillet,F.,Laffaille,P.,Colas,F.,Aulagnier,A.,Blanc,F.,Galan,M., Tiouchichine,M.-L.,Némoz,M.,Buisson,L.,Michaux,J.R.,2017.Novelinsights intothedietofthePyreneandesman(Galemyspyrenaicus)using

next-generationsequencingmolecularanalyses.J.Mammal.,http://dx.doi.org/ 10.1093/jmammal/gyx070.

Brown,L.E.,Milner,A.M.,Hannah,D.M.,2006.Stabilityandpersistenceofalpine streammacroinvertebratecommunitiesandtheroleofphysicochemical habitatvariables.Hydrobiologia560,159–173.

Cantoni,D.,1993.Socialandspatialorganizationoffree-rangingshrews,Sorex coronatusandNeomysfodiens(Insectivora,Mammalia).Anim.Behav.45, 975–995.

Castién,E.,Gosálbez,J.,1995.DietofGalemyspyrenaicus(Geoffroy,1811)inthe NorthoftheIberianpeninsula.Neth.J.Zool.45,422–430.

Castién,E.,Gosálbez,J.,1999.Habitatandfoodpreferencesinaguildof insectivorousmammalsintheWesternPyrenees.ActaTheriol.44,1–13. Charbonnel,A.,D’Amico,F.,Besnard,A.,Blanc,F.,Buisson,L.,Némoz,M.,Laffaille,

P.,2014.Spatialreplicatesasanalternativetotemporalreplicatesfor occupancymodellingwhensurveysarebasedonlinearfeaturesofthe landscape.J.Appl.Ecol.51,1425–1433.

Charbonnel,A.,Buisson,L.,Bifﬁ,M.,D’Amico,F.,Besnard,A.,Aulagnier,S.,Blanc,F., Gillet,F.,Lacaze,V.,Michaux,J.R.,Némoz,M.,Pagé,C.,Sanchez-Perez,J.M., Sauvage,S.,Laffaille,P.,2015.Integratinghydrologicalfeaturesandgenetically validatedoccurrencedatainoccupancymodellingofanendemicand endangeredsemi-aquaticmammal,Galemyspyrenaicus,inaPyrenean catchment.Biol.Conserv.184,182–192.

Charbonnel,A.,Laffaille,P.,Bifﬁ,M.,Blanc,F.,Maire,A.,Némoz,M.,Sanchez-Perez, J.M.,Sauvage,S.,Buisson,L.,2016.Canrecentglobalchangesexplainthe

(9)

dramaticrangecontractionofanendangeredsemi-aquaticmammalspeciesin theFrenchPyrenees?PLoSOne11,e0159941.

Churchﬁeld,S.,Rychlik,L.,2006.DietsandcoexistenceinNeomysandSorexshrews inBiałowie ˙zaforest,easternPoland.J.Zool.269,381–390.

Churchﬁeld,S.,Sheftel,B.I.,1994.Foodnicheoverlapandecologicalseparationina multi-speciescommunityofshrewsintheSiberiantaiga.J.Zool.234,105–124. Churchﬁeld,S.,Nesterenko,V.A.,Shvarts,E.A.,1999.Foodnicheoverlapand

ecologicalseparationamongstsixspeciesofcoexistingforestshrews (Insectivora:soricidae)intheRussianFarEast.J.Zool.248,349–359. Churchﬁeld,S.,1984.Aninvestigationofthepopulationecologyofsyntopic

shrewsinhabitingwater-cressbeds.J.Zool.Lond.204,229–240.

Churchﬁeld,S.,1985.ThefeedingecologyoftheEuropeanwatershrew.Mammal Rev.15,13–21.

Clare,E.L.,Symondson,W.O.C.,Broders,H.,Fabianek,F.,Fraser,E.E.,MacKenzie,A., Boughen,A.,Hamilton,R.,Willis,C.K.R.,Martinez-Nu ˜nez,F.,Menzies,A.K., Norquay,K.J.O.,Brigham,M.,Poissant,J.,Rintoul,J.,Barclay,R.M.R.,Reimer,J.P., 2014.ThedietofMyotislucifugusacrossCanada:assessingforagingqualityand dietvariability.Mol.Ecol.23,3618–3632.

Costa,A.,Salvidio,S.,Posillico,M.,Matteucci,G.,DeCinti,B.,Romano,A.,2015. Generalisationwithinspecialization:inter-individualdietvariationintheonly specializedsalamanderintheworld.Sci.Rep.5,13260.

DuPasquier,A.,Cantoni,D.,1992.Shiftsinbenthicmacroinvertebratecommunity andfoodhabitsofwatershrew,Neomysfodiens(Soricidae,Insectivora).Acta OEcologicaOEcal.Gener.13,81–99.

Dudgeon,D.,Arthington,A.H.,Gessner,M.O.,Kawabata,Z.-I.,Knowler,D.J., Lévêque,C.,Naiman,R.J.,Prieur-Richard,A.-H.,Soto,D.,Stiassny,M.L.J., Sullivan,C.A.,2006.Freshwaterbiodiversity:importance,threats,statusand conservationchallenges.Biol.Rev.81,163–182.

Edgar,R.C.,2010.SearchandclusteringordersofmagnitudefasterthanBLAST. Bioinforma.Oxf.Engl.26,2460–2461.

Escoda,L.,González-Esteban,J.,Gómez,A.,Castresana,J.,2017.Usingrelatedness networkstoinfercontemporarydispersal:applicationtotheendangered mammalGalemyspyrenaicus.Mol.Ecol.,http://dx.doi.org/10.1111/mec. 14133.

Füreder,L.,Wallinger,M.,Burger,R.,2004.Longitudinalandseasonalpatternof insectemergenceinalpinestreams.Aquat.Ecol.39,67–78.

Fernandes,M.,Herrero,J.,Aulagnier,S.,Amori,G.,2008.GalemysPyrenaicus.The IUCNRedListofThreatenedSpecies.Version2014.2(<www.iucnredlist.org> Downloadedon1June2014)http://dx.doi.org/10.2305/IUCN.UK.2008.RLTS. T8826A12934876.en.

Finn,D.S.,Khamis,K.,Milner,A.M.,2013.Lossofsmallglacierswilldiminishbeta diversityinPyreneanstreamsattwolevelsofbiologicalorganization.Glob. Ecol.Biogeogr.22,40–51.

Giller,P.,Greenberg,L.,2015.Therelationshipbetweenindividualhabitatuseand dietinbrowntrout.Freshw.Biol.60,256–266.

Gillet,F.,Tiouchichine,M.-L.,Galan,M.,Blanc,F.,Némoz,M.,Aulagnier,S., Michaux,J.R.,2015.AnewmethodtoidentifytheendangeredPyrenean desman(Galemyspyrenaicus)andtostudyitsdiet,usingnextgeneration sequencingfromfaeces.Mamm.Biol.80,505–509.

Gillet,F.,2015.GénétiqueetbiologiedelaconservationdudesmandesPyrénées (Galemyspyrenaicus)enFrance.Thèsededoctorat,UniversitéPaulSabatierde Toulouse(France)/UniversitédeLiège(Belgique),228pp.

Gisbert,J.,García-Perea,R.,2014.Historiadelaregresióndeldesmánibérico Galemyspyrenaicus(É.GeoffroySaint-Hilaire,1811)enelSistemaCentral (PenínsulaIbérica).Conservationandmanagementofsemi-aquaticmammals ofsouthwesternEurope.MunibeMonogr.Nat.Ser.,3.,pp.19–35.

Greenwood,A.,Churchﬁeld,S.,Hickey,C.,2002.Geographicaldistributionand habitatoccurrenceofthewatershrew(Neomysfodiens)intheWealdof South-EastEngland.MammalRev.32,40–50.

Haberl,W.,2002.Foodstorage,preyremainsandnotesonoccasionalvertebrates inthedietoftheEurasianwatershrew,Neomysfodiens.FoliaZool.51,93–102. Harrington,L.A.,Harrington,A.L.,Yamaguchi,N.,Thom,M.D.,Ferreras,P.,

Windham,T.R.,Macdonald,D.W.,2009.Theimpactofnativecompetitorsonan alieninvasive:temporalnicheshiftstoavoidinterspeciﬁcaggression?Ecology 90,1207–1216.

Hutchinson,G.E.,1957.Concludingremarks.ColdSpringHarb.Symp.Quant.Biol. 22,415–427.

Ivlev,V.S.,1961.ExperimentalEcologyofFeedingFishes.YaleUniversityPress, NewHaven,Connecticut(302pp).

Keckel,M.R.,Ansorge,H.,Stefen,C.,2014.Differencesinthemicrohabitat preferencesofNeomysfodiens(Pennant1771)andNeomysanomalusCabrera 1907inSaxony.Germany.ActaTheriol.59,485–494.

Klare,U.,Kamler,J.F.,Macdonald,D.W.,2011.Acomparisonandcritiqueof differentscat-analysismethodsfordeterminingcarnivorediet.Mammal Review41,294–312.

Krüger,F.,Clare,E.L.,Greif,S.,Siemers,B.M.,Symondson,W.O.C.,Sommer,R.S., 2014.Anintegrativeapproachtodetectsubtletrophicnichedifferentiationin thesympatrictrawlingbatspeciesMyotisdasycnemeandMyotisdaubentonii. Mol.Ecol.23,3657–3671.

Lardet,J.-P.,1988.Spatialbehaviourandactivitypatternsofthewatershrew Neomysfodiensintheﬁeld.ActaTheriol.33,293–303.

Life+Desman,2013.TechnicalApplicationForms–ConservationoftheFrench PopulationsofGalemysPyrenaicusandItsPopulationsontheFrenchPyrénées (LIFE13NAT/FR/000092).274p.

McLachlan-Troup,T.A.,Dickman,C.R.,Grant,T.R.,2010.Dietanddietaryselectivity oftheplatypusinrelationtoseason,sexandmacroinvertebrateassemblages.J. Zool.280,237–246.

Melero,Y.,Aymerich,P.,Luque-Larena,J.J.,Gosálbez,J.,2012.Newinsightsinto socialandspaceusebehaviouroftheendangeredPyreneandesman(Galemys pyrenaicus).Eur.J.Wildl.Res.58,185–193.

Melero,Y.,Aymerich,P.,Santulli,G.,Gosálbez,J.,2014.Activityandspacepatterns ofPyreneandesman(Galemyspyrenaicus)suggestnon-aggressiveand non-territorialbehaviour.Eur.J.Wildl.Res.60,707–715.

Mendes-Soares,H.,Rychlik,L.,2009.Differencesinswimminganddivingabilities betweentwosympatricspeciesofwatershrews:Neomysanomalusand Neomysfodiens(Soricidae).J.Ethol.27,317–325.

Morueta-Holme,N.,Fløjgaard,C.,Svenning,J.-C.,2010.Climatechangerisksand conservationimplicationsforathreatenedsmall-rangemammalspecies.PLoS One5,e10360.

Muséumnationald’Histoirenaturelle[Ed].2003–2017.InventaireNationaldu PatrimoineNaturel.https://inpn.mnhn.fr.Accessed15May2016. Némoz,M.,Bertrand,A.,Sourie,M.,Arlot,P.,2011.Afrenchconservationaction

planforthepyreneandesmanGalemyspyrenaicus.Galemys23,47–50. OPIE-Benthos,2017.OfﬁcePourlesInsectesetleurEnvironnement

(http://www.opie-benthos.fr/opie/insecte.php.Accessed15May2016). Palmeirim,J.M.,1983.Galemyspyrenaicus.Mamm.Species207,1–5. Parry,G.S.,Bodger,O.,McDonald,R.A.,Forman,D.W.,2013.Asystematic

re-samplingapproachtoassesstheprobabilityofdetectingottersLutralutra usingspraintsurveysonsmalllowlandrivers.Ecol.Inform.14,64–70(The analysisandapplicationofspatialecologicaldatatosupporttheconservation ofbiodiversity).

Pianka,E.R.,1973.Thestructureoflizardcommunities.Annu.Rev.Ecol.Syst.4, 53–74.

Pianka,E.R.,1974.Nicheoverlapanddiffusecompetition.Proc.Natl.Acad.Sci.71, 2141–2145.

Pompanon,F.,Deagle,B.E.,Symondson,W.O.C.,Brown,D.S.,Jarman,S.N.,Taberlet, P.,2012.Whoiseatingwhat:dietassessmentusingnextgeneration sequencing.Mol.Ecol.21,1931–1950.

Puisségur,C.,1935.RecherchessurleDesmandesPyrénées.Bull.Soc.Hist.Nat. Toulouse67,163–227.

RCoreTeam,2014.R:ALanguageandEnvironmentforStatisticalComputing.R FoundationforStatisticalComputing,Vienna,Austriahttp://www.R-project. org/.

Ratnasingham,S.,Hebert,P.D.N.,2007.BOLD:thebarcodeoflifedatasystem (www.barcodinglife.org).Mol.Ecol.Notes7,355–364.

Richard,P.B.,Micheau,C.,1975.Lecarrefourtrachéendansl’adaptationdudesman desPyrénées(Galemyspyrenaicus)àlaviedulc¸aquicole.Mammalia39, 467–478.

Rychlik,L.,1997.Differencesinforagingbehaviourbetweenwatershrews:Neomys anomalusandNeomysfodiens.ActaTheriol.42,351–386.

Rychlik,L.,2000.Habitatpreferencesoffoursympatricspeciesofshrews.Acta Theriol.45(Suppl.1),173–190.

Santamarina,J.,Guitian,J.,1988.Quelquesdonnéessurlerégimealimentairedu desman(Galemyspyrenaicus)danslenord-ouestdel’Espagne.Mammalia52, 302–307.

Santamarina,J.,1993.Feedingecologyofavertebrateassemblageinhabitinga streamofNWSpain(Riobo;Ullabasin).Hydrobiologia252,175–191. Sheppard,S.K.,Bell,J.,Sunderland,K.D.,Fenlon,J.,Skervin,D.,Symondson,W.O.C.,

2005.DetectionofsecondarypredationbyPCRanalysesofthegutcontentsof invertebrategeneralistpredators.Mol.Ecol.14,4461–4468.

Stone,R.D.,1987.TheactivitypatternsofthePyreneandesman(Galemys pyrenaicus)(Insectivora:talpidae),asdeterminedundernaturalconditions.J. Zool.213,95–106.

Tachet,H.,Richoux,P.,Bournaud,M.,Usseglio-Polatera,P.,2000.Invertébrésd’eau douce.Systématique,Biologie,Ecologie,CNRSEditions,Paris.

Valeix,M.,Chamaillé-Jammes,S.,Fritz,H.,2007.Interferencecompetitionand temporalnicheshifts:elephantsandherbivorecommunitiesatwaterholes. Oecologia153,739–748.

Wisz,M.S.,Pottier,J.,Kissling,W.D.,Pellissier,L.,Lenoir,J.,Damgaard,C.F., Dormann,C.F.,Forchhammer,M.C.,Grytnes,J.-A.,Guisan,A.,Heikkinen,R.K., Høye,T.T.,Kühn,I.,Luoto,M.,Maiorano,L.,Nilsson,M.-C.,Normand,S., Öckinger,E.,Schmidt,N.M.,Termansen,M.,Timmermann,A.,Wardle,D.A., Aastrup,P.,Svenning,J.-C.,2013.Theroleofbioticinteractionsinshaping distributionsandrealisedassemblagesofspecies:implicationsforspecies distributionmodelling.Biol.Rev.88,15–30.